Pharmacokinetics and Pharmacodynamics of Nemonoxacin against Streptococcus pneumoniae in an In Vitro Infection Model

Abstract

The aim of this paper was to investigate the pharmacokinetics (PK) and pharmacodynamics (PD) of nemonoxacin, a novel nonfluorinated quinolone, against Streptococcus pneumoniae in vitro. A modified infection model was used to simulate the pharmacokinetics of nemonoxacin following scaling of single oral doses and multiple oral dosing. Four S. pneumoniae strains with different penicillin sensitivities were selected, and the drug efficacy was quantified by the change in log colony counts within 24 h. A sigmoid maximum-effect (E_max) model was used to analyze the relationship between PK/PD parameters and drug effect. Analysis indicated that the killing pattern of nemonoxacin shows a dualism which is mainly concentration dependent when the MIC is low and that the better PK/PD index should be the area under the concentration-time curve for the free, unbound fraction of the drug divided by the MIC (fAUC_0–24/MIC), which means that giving the total daily amount of drug as one dose is appropriate under those conditions. When the MIC is high, the time (T) dependency is important and the valid PK/PD index should be the cumulative percentage of a 24-h period in which the drug concentration exceeds the MIC under steady-state pharmacokinetic conditions (f%T>MIC), which means that to split the maximum daily dose into several separate doses will benefit the eradication of the bacteria. To obtain a 3-log₁₀-unit decrease, the target values of fAUC_0–24/MIC and f%T>MIC are 47.05 and 53.4%, respectively.

INTRODUCTION

Nemonoxacin, a novel nonfluorinated quinolone, exerts broad antibacterial activity by interrupting DNA synthesis in prokaryotic organisms (1). The antimicrobial spectrum includes Gram-positive and Gram-negative bacteria as well as atypical pathogens such as penicillin-resistant Streptococcus pneumoniae and methicillin- and vancomycin-resistant Staphylococcus aureus (2, 3). The antibacterial activity of nemonoxacin is stronger than that of levofloxacin and moxifloxacin against most Gram-positive cocci (4).

Completed phase I clinical trials in China have demonstrated that nemonoxacin is well tolerated within the dose range of 125 to 1,000 mg and shows linear pharmacokinetics. The absorption peak appeared at 1 to 2 h, and the elimination half-life was 10 to 12 h (5). In a phase II clinical trial, the effect of nemonoxacin (500 or 750 mg daily for 7 to 10 days) was similar to that of 500 mg levofloxacin for treating community-acquired pneumonia (CAP) in adults (6). A phase III clinical trial has just completed. However, few studies have investigated the killing pattern and pharmacokinetic/pharmacodynamic (PK/PD) characteristics of this drug, which are needed for the dose regimen design.

Compared to animal models, the PK process of the drug in humans could be well simulated in the in vitro models without considering species differences between human and animals (7, 8). Meanwhile, in vitro models allow determination of time-kill behavior and identification and optimization of PK/PD indices and breakpoints (9). In the present work, we used an in vitro PK/PD infection model to observe the antimicrobial effect of nemonoxacin on S. pneumoniae strains with different penicillin sensitivities (sensitivity, intermediate sensitivity, and resistance), which commonly cause CAP. The bacterial killing pattern and the PK/PD parameters associated with the antimicrobial effect were identified. The results can be used for the future design of dose regimens of this drug.

MATERIALS AND METHODS

Strains.

We used four S. pneumoniae strains (10-w12-27, 10-w2-5, ET-17, and 10-w8-75) that were isolated from clinical specimens. Both 10-w12-27 and 10-w2-5 are penicillin-sensitive strains, and the two strains have penicillin MICs (MIC_PEN) of 0.015 and 2 μg/ml, respectively; ET-17 is a strain with intermediate penicillin sensitivity and a MIC_PEN of 4 μg/ml; and 10-w8-75 is a penicillin-resistant strain with a MIC_PEN of 8 μg/ml. The multilocus sequence typing information for these strains is shown in Table S1 in the supplemental material. All strains were stored at −80°C in Mueller-Hinton broth containing 20% glycerol. Before each experiment, the strains were streaked on a Mueller-Hinton agar plate with 5% defibrinated sheep blood and grown for 16 h at 35°C with 5% CO₂.

Antibiotic.

Nemonoxacin malate (lot PTOPO-060201A, containing 71.3% nemonoxacin) was provided by TaiGen Biopharmaceuticals Co., Ltd. (Beijing, China). The stock solution was prepared with sterile distilled water and stored at −80°C for a maximum of 2 months before use.

MIC determination.

The MICs of nemonoxacin against all S. pneumoniae strains were determined using the microdilution method according to the Clinical and Laboratory Standards Institute guidelines (10). S. pneumoniae ATCC 49619 served as the control strain. Bacteria from two or three clones were picked from freshly streaked culture plates and incubated in cation-adjusted Mueller-Hinton (CAMH) broth supplemented with 5% lysed horse blood for 8 h. The bacterial culture was then diluted with double-concentrated CAMH broth containing 10% lysed horse blood to a final concentration of 5 × 10⁵ CFU/ml. Nemonoxacin solutions were prepared by diluting the stock solution with sterile distilled water. The diluted bacterial suspension (50 μl) was added to an equal volume of drug solution in a single microwell and incubated for 24 h at 35°C with 5% CO₂. The minimum concentration in the microwells with no visible bacterial growth was defined as the MIC. All MIC determinations were repeated twice on different days.

In vitro pharmacokinetic model.

A two-compartment model without an absorption delay was developed for the nemonoxacin PK data obtained from our previous research (details are provided in Table S2 in the supplemental material). We constructed the in vitro PK/PD infection model as described by Grasso et al. (11) and Sevillano et al. (12). The system consists of a fresh medium reservoir, absorption compartment, central compartment, and liquid waste storage compartment. These components are connected in tandem with silicone tubing (Fig. 1). The medium is pumped from the fresh medium reservoir to the absorption compartment by a peristaltic pump (Masterflex L/S; Cole-Parmer Co. Ltd.), which is digitally controlled by a computer using WinLIN 3.2 software (Cole-Parmer Co. Ltd.). A cellulose ester membrane (0.45-μm pore size) is placed at the bottom of central compartment to prevent bacteria from flowing out with the medium. Above the membrane, a magnetic stir bar is fixed by an axle to a bearing on the top of the central compartment. This stirs the medium and prevents the membrane pores from being blocked by bacteria. Outside air was not allowed into the absorption and central compartments. For the experiments, the entire system, other than the computer-controlled peristaltic pump, was placed in an incubator at 35°C with 5% CO₂.

Fig 1.

Equipment for oral dosing in vitro PK/PD model.

The medium used in this model was Todd-Hewitt broth supplemented with 0.5% yeast extract (13). It had been shown that the MICs of nemonoxacin in Todd-Hewitt broth with yeast extract were identical to that in CAMH broth with lysed horse blood in the preliminary experiment. The absorption compartment holds 35.06 ml, and the central compartment holds 250 ml. Fresh medium pumped into the absorption compartment displaces an equal volume of liquid, which flows into the central compartment. Likewise, the same volume of liquid flows out through the outlet of the central compartment and enters the liquid waste storage container. To simulate two-compartment model pharmacokinetics with this system, the flow rate was 1.39 ml/min from 0 to 2.3 h. From 2.3 to 13 h, the flow rate decreased stepwise from 1.13 to 0.33 ml/min, with a decrement of 0.05 ml/min. From 13 to 24 h, the flow rate was maintained at 0.33 ml/min. The drug was introduced into the system through the dosing port at the zero time point.

Validation of the in vitro pharmacokinetic model.

Pharmacokinetics in healthy Chinese volunteers receiving oral doses of nemonoxacin (100, 250, 500, and 750 mg) within 24 h were simulated in the model (14). The serum protein-binding rate was previously reported as 16% (15). The doses added to the absorption compartment were 0.329, 0.822, 1.644, and 2.466 mg as a 0.5-ml solution. After dosing, a 250-μl sample was obtained from the central compartment to determine the nemonoxacin concentrations at the following time points: 0.25, 0.5, 0.75, 1, 1.5, 2, 4, 6, 8, 10, 12, and 24 h. Each situation was repeated in triplicate, and the average concentration at each point was plotted on a graph with the target PK curve.

The determination of nemonoxacin levels was performed using a liquid chromatography-tandem mass spectrometry (LC-MS/MS) method according to our previous work (5). In detail, the LC-MS/MS system included a Waters 2690 module (Waters Co. Ltd., Milford, MA) in tandem with the TSQ Quantum Discovery Max mass spectrometer (Thermo Finnigan Co. Ltd., San Jose, CA). Both the analyte and the internal standard (gatifloxacin) were ionized by electrospray and detected in the positive monitoring mode. Chromatography was performed with a Waters Symmetry Shield RP C₁₈ column (Waters Co. Ltd.) (50 mm by 2.1 mm; 5-μm pore size), and the mobile phase was acetonitrile–0.1% formic acid (15:85). A 150-μl sample was mixed with 10 μl of the internal standard (gatifloxacin [0.025 μg/ml]) and then was extracted with 1 ml of ethyl acetate-isopropanol (3:1 [vol/vol]) for 10 min on a rotator. After centrifugation (8,000 × g for 5 min), 800 μl of the top layer (organic phase) was transferred to a 2-ml polyethylene tube and evaporated to dryness under a stream of nitrogen gas at 40°C. The extract was dissolved in 150 μl of the mobile phase, and 5 μl of the extract solution was injected into the LC-MS/MS system. The lower limit of quantitation was 0.005 μg/ml, and the linear range was 0.005 to 1 μg/ml.

In vitro PK/PD study of nemonoxacin.

An overnight culture of bacteria was added to the central compartment (1 × 10⁶ CFU/ml) and incubated at 35°C for 1 h (13). After this preculture step, different doses of nemonoxacin or a control (sterile distilled water) were injected into the absorption compartment, and the peristaltic pump was turned on. A 200-μl sample was obtained for the PD analysis through the sampling port at the following time points: 0, 2, 4, 6, 8, 10, 12, and 24 h. Samples were properly diluted with MH broth, and then 100-μl aliquots of the diluted samples were poured onto MH agar plates supplemented with 5% defibrinated sheep blood and incubated at 35°C for 24 h. The log-transformed colony counts (y axis) were plotted against time (x axis) for the time-kill curve. To monitor drug concentrations, 250-μl samples were also obtained at 1, 6, 12, and 24 h and centrifuged at 4°C and 8,000 × g for 10 min. The supernatant (200 μl) was stored at −80°C until the concentration was determined. All experiments were repeated on different days.

To observe the effect of splitting the doses, the dose regimens of 125 mg every 6 h and 500 mg every day were carried out in parallel within the infection models of 10-w12-27 and 10-w2-5 strains. For this experiment, the flow rate setting for the simulation of 125 mg every 6 h was repeated every 6 h and that of 500 mg every day was the same as that described above. The sampling points were every 2 h from 0 to 12 h and every 3 h after 12 h. The treatment of the samples was performed as described above.

PK/PD analysis.

The pharmacokinetic data were analyzed using Phoenix WinNonlin 6.0 software (Pharsight Co. Ltd.), and fC_max and fAUC_0–24, where f is the unbound fraction of nemonoxacin with value of 0.84, C_max is the maximum concentration of drug in serum, and fAUC_0–24 is the area under the concentration-time curve for the free, unbound fraction of nemonoxacin from h 0 to 24, were calculated. The PK/PD parameters fAUC_{0–24 h}/MIC and fC_max/MIC and the cumulative percentage of a 24-h period in which the drug concentration exceeds the MIC under steady-state pharmacokinetic conditions (f%T>MIC) were determined using the pharmacokinetic values and MIC data in each experiment. The in vitro drug effect was quantified by calculating changes in log colony counts between 24 and 0 h (ΔLog CFU₂₄) (16). A sigmoid maximum-effect (E_max) model was constructed to describe the relationship between the PK/PD parameters and the drug effects. The formula can be described as follows: ΔLog CFU₂₄ = E₀ + (E_max − E₀) · X^H/(EC₅₀^H + X^H), where X represents the PK/PD parameter (fAUC_0–24/MIC, fC_max/MIC, or f%T>MIC), H represents the Hill coefficient, EC₅₀ represents the value of PK/PD parameters when the effect is 50% of the E_max, and E₀ is the baseline value. All of the experimental data were used in the analysis. The PK/PD target value for a 3-log-unit decrease was then calculated (17).

RESULTS

MIC of nemonoxacin against S. pneumoniae strains.

The MIC values of nemonoxacin against the S. pneumoniae strains were 0.06 μg/ml (strains 10-w8-75 and ET-17), 0.25 μg/ml (10-w12-27), and 2 μg/ml (10-w2-5).

In vitro pharmacokinetic model.

The in vivo pharmacokinetics of nemonoxacin was well simulated by the model (Fig. 2A). All monitoring data were well fitted to the target curve, with the relative deviation below 10% (Fig. 2B). The average relative deviations of fC_max/MIC and fAUC_0–24/MIC were 5.1% and 3.3%, respectively. Meanwhile, the time profiles of the in vitro concentration were also reproduced by the two-compartment model, indicating that multiphase modulation of flow rate performed well to simulate nemonoxacin pharmacokinetics in vivo.

Fig 2.

Pharmacokinetic simulation of nemonoxacin in the in vitro PK model. (A) Pharmacokinetic simulation in model validation. (B) Pharmacokinetic simulation monitoring in the PK/PD experiment. Dashed lines represent the concentration data in the in vitro PK model (means ± standard deviations [SD]; n = 3 [A] and n = 8 [B]); the solid lines represents the target pharmacokinetic curve. p.o., orally.

Time-kill kinetics of nemonoxacin in the in vitro PK/PD model.

At the MIC of 0.06 μg/ml, 100 to 750 mg nemonoxacin every 24 h completely killed the bacteria (without regrowth) in 24 h (Fig. 3). When the MIC increased to 0.25 μg/ml, 500 to 750 mg nemonoxacin every 24 h effectively killed bacteria in 24 h, whereas 100 to 250 mg only inhibited bacterial growth within 12 h. At the MIC of 2 μg/ml, 500 to 750 mg nemonoxacin every 24 h produced a 3-log-unit decrease at 8 to 10 h, but regrowth was observed at 24 h; 100 to 250 mg every 24 h inhibited growth only weakly. The sensitivities of the S. pneumoniae strains to penicillin were unrelated to the killing effects of nemonoxacin.

Fig 3.

Time-kill curves of nemonoxacin administered with single scaling doses in the in vitro PK/PD model (means ± SD; n = 2). The horizontal line at 1 log CFU/ml represents the lower limit of detection for the assay. NM, nemonoxacin; PEN, penicillin.

During the dose regimen comparison of 125 mg every 6 h and 500 mg every 24 h, the relationship of the end point effects of two regimens was dissimilated with the MIC variance. When the MIC was 0.25 μg/ml, both regimens killed the bacteria thoroughly and without regrowth in 48 h, and the end point effects were similar. When the MIC was 2 μg/ml, neither regimen inhibited the growth in 24 h, but the regimen of 125 mg administered every 6 h had a lower Log CFU at the end time. However, when the effect is measured with the AUBKC (area under bacterial killing curve), the dose regimen of 500 mg every 24 h has a lesser value than that of 125 mg every 6 h in both strains (Fig. 4).

Fig 4.

Dose regimen comparison. The total daily dose of 500 mg was administered as one dose or fractionated into four doses. Triangles and squares represent the 10-w2-5 and 10-w12-27 strains, respectively. The solid line stands for daily dosing administered as one dose; the dotted line stands for fractionated dosing. The horizontal line at 1 Log CFU/ml represents the lower limit of detection for the assay.

PK/PD analysis.

The f%T>MIC and fAUC_0–24/MIC have similar correlations with the drug activity, and the R-squared values are 0.9753 and 0.9708 (Fig. 5). The correlation between f%T>MIC and ΔLog CFU₂₄ (R² = 0.975) is less than that reported above, with the R-squared value equal to 0.9582. The estimates of E₀, E_max, EC₅₀, and the Hill coefficient for each parameter are listed in Table 1. The inferred 3-log₁₀-unit decreased target values for f%T>MIC, fAUC_0–24/MIC, and fC_max/MIC were 53.4%, 47.05 and 5.07, respectively.

Fig 5.

Correlations between drug effect (ΔLog CFU_{24 h}) and PK/PD parameters of nemonoxacin. Diamonds and solid lines indicate measured values and simulation values based on the sigmoid E_max model.

Table 1.

E_max model parameters characterizing the relationship between ΔLogCFU_24h and PK/PD indices of nemonoxacin

PK/PD index	E0a	Emaxa	EC50	H
fAUC/MIC	2.835	−6.126	38.60	3.154
fCmax/MIC	2.553	−5.897	4.826	13.15
f %T>MIC	2.827	−7.290	47.63	2.683

^aUnits are in log CFU/ml.

DISCUSSION

Complex pharmacokinetics such as two- or three-compartment model or extravascular-administration pharmacokinetics are often simplified as intravenous or oral one-compartment models when simulated in vitro (7, 18). One potential consequence of this is that the simplification may result in an unrealistic drug PK/PD profile. Sevillano and colleagues first simulated intravenous two-compartment models using computer-controlled in vitro systems with biphasic or triphasic flow rates (12, 19). Simulating extravascular-administration two- or three-compartment PK in vitro remains a difficult problem. By using the basic equipment for oral-dosing one-compartment model simulation and adopting a multiphasic flow rate pattern, we have realized the in vitro simulation of the oral-dosing two-compartmental model in this study. The in vitro PK of nemonoxacin was similar to the in vivo PK. In our experiments, the relative deviations of the data obtained at 1 and 6 h did not exceed 5%, and those of data obtained at 12 and 24 h did not exceed 10% (data not shown), indicating the high precision of fC_max and fAUC_0–24 simulation. To the best of our knowledge, this is the first report describing an oral-dosing two-compartment model simulation in the in vitro one-compartment PK model.

S. pneumoniae is a species that frequently causes CAP, which is often treated with penicillin and fluoroquinolones (20, 21). However, some strains have developed resistance to these drugs (22, 23). Nemonoxacin is a new drug that is effective against S. pneumoniae, including strains resistant to penicillin or ciprofloxacin (13). So far, few studies have evaluated the killing pattern or PK/PD characteristics of this drug. In this study, we aimed to answer these questions using an in vitro PK/PD model of infection. From the dosage scaling data, we found that the antibacterial effects obviously increased when the dosages were enlarged in strains 10-w12-27 and 10-w2-5. Although the end point effects of all dosages were the same in strains of 10-w8-75 and ET-17, we found that the antibacterial effects measured with the AUBKC were still increased. That means that the antibacterial effect of this drug has a good correlation with its concentration, which corresponds to the PK/PD analysis result showing that the fAUC_0–24/MIC correlated well with the drug activity.

In the PK/PD analysis, we found that the f%T>MIC has the best correlation with the drug effect. We cannot exclude the possibility that it is a pseudomorph, because few data following fractionated dosing were included in this study and the existing data were intensively plotted at both ends. From the regimen comparison data, we have found that the end effect in multiple dosing was better than that in single dosing, which means that the time during which the concentration is above the MIC is an important factor, especially when the MIC is high. Giving the total daily dose as one dose can bring faster eradication of bacteria when the MIC is low, but the risk of bacterial regrowth is bigger than the risk seen with fractionated dose regimen at the same MIC. Hence, we concluded that the time dependency is a basic and more generalized characteristic for nemonoxacin with respect to killing the bacteria and that the dose regimen based on it will cover a wider MIC range. The concentration dependency is a narrowly defined characteristic for this drug, and it should be emphasized when the MIC is low.

With respect to the dualism in the antibacterial effect of nemonoxacin, we suggest that the dose regimen should be adjusted according to the MIC level. When the MIC is low, the dose regimen should consist of giving the total daily drug amount in a single dose, which will lead to a more rapid eradication of the bacteria than multiple dosing. When the MIC is high, especially when it is near the intermediate sensitivity level, eradication of the bacteria is more important than the killing speed, so the full daily dose should have a proper split in order to maintain the concentration at a level higher than the MIC for enough time. The MICs for clinical isolates of S. pneumoniae were nearly all less than 0.25 μg/ml (2, 4, 24), suggesting that the killing effect of nemonoxacin can be defined as concentration dependent. In a sense, this is consistent with many reports of classic fluoroquinolones such as levofloxacin and moxifloxacin (25, 26). Hence, the regimens of 500 mg administered every 24 h (q24h) and 750 mg (q24h) not only provide favorable bacteriological efficacy but also have good clinical feasibility. However, these regimens should be further certified from the perspectives of clinical results and resistance development.