Pharmacokinetics and pharmacodynamics integration of danofloxacin against Eschrichia coli in piglet ileum ultrafiltration probe model

Improper use of antibiotics results in poor treatment and severe bacterial resistance. In this study, ultrafiltration probes were successfully placed in the ileum of piglets with the aid of anesthetic. After the fluoroquinolone antimicrobial drug danofloxacin (DAN) was intramuscularly administered, blood and ileum ultrafiltrate were collected at different time points and then determined by High Performance Liquid Chromatography (HPLC). Pharmacokinetics (PK) parameters for plasma and ileum ultrafiltrate were calculated by WinNonlin software. The DAN concentration in ileum ultrafiltrate was much higher than that in plasma during the period 1.2–48 h. The DAN concentration in plasma reached its maximum at 1.10 ± 0.03 h, but reached at 6.00 ± 0.00 h in the ileum ultrafiltrate. The mean Cmax of the ileum is 13.59 times that of plasma. The elimination half-life (T1/2β) in the ileum ultrafiltrate (6.84 ± 1.49 h) was shorter than those in plasma (7.58 ± 3.20 h). The MIC, MBC and MPC of DAN in MH broth against Escherichia coli (O158) were 0.5 µg/mL, 0.5 µg/mL and 4 µg/mL, respectively. Both in vitro and ex vivo kill curves indicated that the killing mechanism of DAN against E. coli is concentration-dependent. The AUC/MPC ratio is 21.33 ± 2.14. Mean PK/PD index (AUC24h/MIC) for ileum ultrafiltrate that achieved bacteriostatic, bactericidal, and eradication were 99.85, 155.57, and 218.02 h, respectively. Three different dosages (1.49 mg/kg, 2.42 mg/kg, and 3.24 mg/kg) were calculated respectively based on AUC24h/MIC ratio above, which might provide a novel approach to the rational design of dosage schedules.

www.nature.com/scientificreports/ drug resistant bacteria appear in the early stage of clinical application of many new drugs 7 . Therefore, a rational dosage schedule must be designed to reduce the emergence of drug resistance. The design of rational dosage schedule is dependent on (i) the linkage of Pharmacokinetics (PK) data to ex vivo or in vivo pharmacodynamic (PD) data generated in animal models or in clinical trials and (ii) The killing mechanism of the drug against bacteria. The integrated PK-PD parameters the area under curve (AUC 24h )/MIC and maximum concentration of drug (C max )/MIC are therefore closely related to a successful treatment outcome 3 . However, few previous publications have described the PK of DAN in animal model, especially in piglets. The objectives of this study were (1) to implant an ultrafiltration probe into the ileum of pig to continuously collect intestinal fluid; (2) to establish plasma and ileum PK data for DAN in piglet after intramuscular (i.m.) treatment; (3) to determine integrated PK-PD parameters (AUC 24h /MIC and C max /MIC) for DAN in vivo; (4) to establish the AUC/MIC ex vivo that produce bacteriostatic, bactericidal, and eradication by using the inhibitory effect sigmoid E max equation; and (5) to calculate respectively different dosages based on AUC 24h /MIC ratio that might provide a novel approach to the rational design of dosage schedules which provide maximal efficacy and minimizes opportunity for the emergence of resistant strains.

Materials and methods
Ethical statement. All  Pathogenicity test. The laboratory preserved strains (HP65, HP189, HP232 and HP501) were conveniently selected for pathogenicity test. Pathogenicity test was performed with 39 Kunming mice (weighted 32 ± 3 g, specific pathogen free grade) purchased from Experimental Animal Centre of Harbin Medical University. Every three mice were intraperitoneal administrated with 200µL E. coli suspension (10 7 , 10 8 , and 10 9 CFU/mL). The blank control group (n = 3) were intraperitoneal administrated with 200 µL sterile broth. The mice were observed every 3 h after the injection up to 72 h. If there were three mice dead in a group, the inoculant given to this group was considered to be of high pathogenicity.
Bacterial strain, antimicrobials, and chemicals. According  Animals and implantation of an ultrafiltration probe. Six healthy males, castrated crossbred piglets (Duroc × Landrace), weighing 21-35 kg, were obtained from Harbin breed swine farm. They were housed in individual cages and allowed a 7 day acclimation period prior to the initiation of the study. Piglets had free access to water and fed antibiotic free feed twice a day.
Ultrafiltration probes were implanted into the ileum of piglets according to the previous study with modified 8 . Several BASi accessories were used for in vivo ultrafiltration sampling. These include: in vivo ultrafiltration sampling probes (MF-7023); flanged tubing connectors (MD-1510); vacuum vial needles (MD-1320) and the vacuum needle holder (MD-1322); vacutainers (MF-7024). Feed and water were withheld from the piglets for 12 h prior to surgery. Surgical procedures have been performed two days with three surgeries done each day. The piglets were pre-medicated with intramuscular administration of 0.02 mg/kg atropine sulfate prior to anesthesia. Approximately 15 min later, anesthesia was induced with 5 mg/kg propofol injectable emulsion intravenously through the ear vein. Each piglet was placed in left lateral recumbency on the surgical table and manually intubated orotracheally (Fig. 1a). General anesthesia was maintained by inhalation of 1.5-3.0% isoflurane in 100% oxygen on a circle circuit with mechanical ventilation. During surgery, the steers were given a continuous infusion of physiological saline at 10 mL/kg/h. The electrocardiogram, oxygen saturation, and non-invasive blood pressure via doppler were monitored throughout surgery (Fig. 1f).
The abdominal hair was clipped and partially disinfected with iodophor. A vertical 10 cm skin incision was made approximately 10 cm above the midline of the abdomen and approximately 5 cm from the last rib (Fig. 1b). Sharp dissection continued through the external and internal abdominal oblique muscles, transverse abdominal muscle, and peritoneum to enter the peritoneal cavity. The cecum was identified and exteriorized caudally to expose the ileocecal fold 8 . The ileocecal fold was used to identify the ileum, the location of the probe was 10 cm away from the cecum. A stab incision was made with a scalpel blade in the ileum, and an introducer needle (MR-5313, BASi, USA) was placed through the stab incision into the lumen of the ileum (Fig. 1c). The collecting end of the ultrafiltration probe was inserted through the introducer needle aborally toward the cecum so that the entire collecting apparatus was within the lumen 8 . The introducer needle was fed off the probe, and nodular suture of the ileum incision with 2-0 absorbable surgical suture obtained from Shanghai Pudong Jinhuan Medical Supplies Co., Ltd. (Fig. 1d, e). The free ends of the probes were exteriorized to the skin incision approximately 25 cm above the midline of the abdomen by using the introducer needle to create a tunnel from the abdominal cavity through the muscle layers 8  www.nature.com/scientificreports/ the free end of the probe could be fed through to the outside of the body. The abdominal incision was closed in three layers. The first layer included the peritoneum, transverse abdominal, and internal abdominal oblique muscles. This layer was closed in a simple continuous pattern using 2-0 absorbable surgical suture. The external abdominal oblique muscle was also closed with 2-0 absorbable surgical suture in a simple continuous pattern. Skin layer sutured with 3-0 absorbable surgical suture. The vacuum tube holder and needle are sutured to the skin (Fig. 1g). After recovery from anesthesia, they were housed in individual cages and received drug treatment to prevent infection by intramuscular injection of penicillin (1,000,000 IU/kg) and application of iodophor over the wound twice a day for 3 days.
Sample collection and HPLC analysis. DAN was intramuscularly injected at 2.5 mg/kg body weight in each piglet after 72 h probe insertion. Blood samples (5 mL) from the brachiocephalic vein were collected into EDTA dipotassium salt tubes at 0, 0.25, 0.5, 1, 2, 4, 6, 8, 12, 24, 26, 30, and 48 h after drug administration 3 . To collect ultrafiltrate from the gastrointestinal tract, a 3 mL vacutainer ( Fig. 1d) was then inserted into the needle of the vacuum vial needle holder. Approximately 300-400 µL of ultrafiltrate is collected each hour, and the probes will continuously collect ultrafiltrate. The timed interval samples were collected at 0, 1, 2, 4, 6, 8, 12, 24, 26, 30, and 48 h after drug administration by changing the vacutainer tubes. Plasma was separated by centrifugation at 3000 g for 10 min and then the supernatant was transferred into a fresh tube. All plasma and ultrafiltrate samples stored at − 80 °C till analysis. The method for the analysis of DAN concentration was modified from that described by Schrickx and Garcia 9,10 . The HPLC system Waters 2695 was connected to a Waters 2475 fluorescence detector (λ ex = 280 nm and λ em = 450 nm) with a mixture of acetonitrile and aqueous solution (15:85, v/v) as the mobile phase. The aqueous solutions were prepared by dissolving potassium dihydrogenophosphate (0.020 M), phosphoric acid (0.006 M), and tetraethylammonium bromide (0.012 M) in water. The pH of the mobile phase was adjusted to 3.0 by addition of 2 N NaOH. The flow rate was set at 1.0 mL/min; A Waters C18 reverse phase column C18 (250 mm × 4.6 mm I.D.; particle size, 5 μm) was used to perform HPLC at 30 °C; and the injection volume was 10 μL.
Samples were thawed at room temperature, and 10 μL of 50 μg/mL ciprofloxacin was added to plasma (490 μL) and ileum ultrafiltrate dilution (50 μL ultrafiltrate and 440 μL mobile phase) as the internal standard. After adding 3 mL of acetonitrile, the plasma samples were shaken at 220 oscillations/min for 15 min and centrifuged at 12,000 g for 10 min. The organic layer was transferred into a fresh tube and dried at 40 °C under nitrogen stream. The residue was dissolved in the mobile phase (0.5 mL), and 10 µL injected for HPLC analysis. The ultrafiltrate sample was diluted 10 times with mobile phase and analyzed directly by HPLC without extraction after filtered.
The limit of detection (LOD) was 0.005 µg/mL and the limit of quantification (LOQ) was 0.01 µg/mL, respectively. Standard curves were linear from 0.01 to 1.5 µg/mL in plasma (R 2 = 0.9999) and ileum ultrafiltrate (R 2 = 0.9998). The intra-day and inter-day variation for determination in plasma ranged from 0.39 to 1.95% and 0.18 to 4.74%. The intra-day and inter-day variation for determination in ileum ultrafiltrate ranged from 0.47 to 1.42% and 0.26 to 1.86%, respectively. The recovery of DAN in plasma ranged from 88.50 to 104.80%. www.nature.com/scientificreports/ Pharmacokinetic analysis. PK parameters for DAN concentration-time data for plasma, ileum ultrafiltrate were calculated by using the WinNonlin software (version 5.2.1, Pharsight Corporation, USA). Minimum Akaike Information Criterion (AIC) was applied to discriminate the model with the best fit 11 . The data for plasma of all the six piglets were best fitted to a two-compartmental model. The data for ileum ultrafiltrate of six piglets were best described by a non-compartment model.

Pharmacodynamics analysis. Determination of MIC, MBC, and MPC for E. coli in vitro. The MIC was
determined by broth microdilution testing in accordance with the guidelines in CLSI 12 . The MBC was performed according to the guidelines in CLSI 13 . The MPC was conducted based on the previous study 14 .
In vitro time-kill curves. For  Calculation of the administered dosage. The dose was calculated by the using the following formula 15 : where AUC/MIC is the targeted endpoint for optimal efficacy; CL is clearance per day; MIC is minimum inhibitory concentration ; fu is free fraction of drug in ileum ultrafiltrate (from 0 to 1); F is the bioavailability factor (from 0 to 1). In this study, fu is 1 because the ileum ultrafiltrate collected by the ultrafiltration probe contained only the free state of DAN.
The dose was calculated respectively by the formula above, based on the PK/PD parameters AUC 24h /MIC and C max /MIC required to achieve the bacteriostatic action, bactericidal action and bacterial elimination.

Statistical analyses. Statistical analysis was undertaken by analysis of variance using GraphPad Software
Prism 6 (version 6.01). Differences between the means of the plasma group and ileum ultrafiltrate group were assessed using two-tailed paired Student's t test. All data were presented as the mean ± SD, where P < 0.05 was considered a statistically significant difference.

Results
Pathogenicity test. The results of E. coli virulence test for mice were presented in Table 1 Table 2, the DAN con-  (Table 3).

PK-PD integration for DAN in ileum ultrafiltrate. The PK-PD integration parameters for the in vivo
PK data and the MICs determined in vitro for ileum ultrafiltrate are presented in  Fig. 4 by the inhibitory effect sigmoid E max equation. Both the AUC 24h /MIC and C/MIC had a good correlation with antimicrobial efficacy (R 2 = 0.99). The estimated E max (Log 10 CFU/mL), E 0 (Log 10 CFU/mL), EC 50 (h), and N (slope) are showed in Table 5. According to the inhibitory effect sigmoid E max equation:E = 2.32 − 6.67·C 4.51 114.78 4.51 +C 4.51 , the calculated mean AUC 24h /MIC, is the C in the inhibitory effect sigmoid E max equation, for ileum ultrafiltrate that produced bacteriostasis, bactericidal activity, and elimination of bacteria were 99.85, 155.57, and 218.02 h, respectively (Table 5); the calculated mean C max /MIC for ileum ultrafiltrate that produced bacteriostasis, bactericidal activity, and elimination of bacteria were 4.14, 6.49, and 9.05, respectively ( Table 5).
Calculation of the administered dosage. Three different dosages (1.49 mg/kg, 2.42 mg/kg, and 3.24 mg/kg) were calculated respectively based on AUC 24h /MIC ratio for bacteriostatic, bactericidal and bacterial eradication in Table 6.  www.nature.com/scientificreports/  Table 5. The values of PK/PD parameters AUC 24h /MIC and C max /MIC required to achieving various levels of antibacterial effect of DAN. E 0 and E max are the changes in log 10 difference in bacterial count between 0 and 24 h in the control sample and in the DAN containing samples, respectively; EC 50 is the AUC 24h /MIC or C max /MIC producing 50% of the maximal antibacterial effect; C is the AUC 24h /MIC or C max /MIC in the effect compartment; N is the Hill coefficient, which describes the steepness of the (AUC 24h /MIC)-effect curve or C max /MIC-effect curve.

Discussion
DAN is a fluoroquinolones antibacterial that has wide spectrum of antimicrobial activity against most gramnegative bacteria and some gram-positive bacteria 3 . Improper use of antibiotics that inappropriate dosing, dosing interval, dosing time, dosing route, dosing conditions lead to poor treatment effect and serious bacterial resistance 16 . To date, PK-PD modeling has been widely used to improve dosing regimens of approved drugs and individualized treatment 17 . Therefore, it was investigated that the PK, PD, and PK-PD integration of DAN in piglet model against E. coli in this study. PK data of DAN in plasma has been described in several species, including cattle, cows, ewes, sheep, goat, camel, brown trout, rabbits, chicken, turkeys, ducks, donkeys, horses and pigs 3 . After i.m. administration of DAN at a dose of 2.5 mg/kg body weight in piglets, the PK parameters including that the time to reach to maximum concentration (T max ) in plasma, the peak drug concentration (C max ), the area under the curve (AUC 0-t ), and the elimination half-life (T 1/2β ) do not differ to a large extent from previous studies [18][19][20] .
Previous publications were mainly focused on PK data of DAN in plasma but ignored the concentrations of antimicrobial agents in their target sites. Therefore, the PK data in both plasma and ileum were analyzed in our study. Luminal gastrointestinal tract (GIT) fluid could be collected by a wide variety of approaches. For instance, euthanasia of multiple animals to collect GIT contents at multiple time points, requiring a large number of trial animals and yielding a relatively low number of animals per time point 21 . Several cannulas and tubes have been used to collect GIT fluid from ileum, rumen and duodenal cannulas in animals 15,22 . These cannulas must be anchored to the abdominal wall which results in disrupting normal intestinal motility, carrying inherent risks of leakage, dislodgement, and peritonitis and significant welfare concerns. In addition, antimicrobial drugs such as fluoroquinolones can be extensively bound to the intestinal contents, but intestinal concentrations should take into account only the unbound and active drug for subsequent PD studies. Ultrafiltration probe technology can solve this problem well in consideration of animal welfare. Ultrafiltration probes have been used to collect interstitial fluid (ISF) in a number of species 8,23,24 . In this study, ultrafiltration sampling probes were successfully placed in the ileum of piglets with the aid of anesthetic, and ileum ultrafiltrate contained only free drugs was collected at different time points. These free drug concentrations could finally give a value of an exposure parameter for the digestive tract (AUC ileum ) that, when combined with the relevant MIC in ileum ultrafiltrate, would predict the efficacy of the drug dosing regimen on the intestinal bacterial population.
A study described the PK variables of DAN in gastrointestinal content from ileum, jejunum, caecum and colon in healthy and in salmonella typhimurium infected pigs 25  Rational use of antibiotics requires extensive knowledge of the infectious bacteria and the PK and PD of the drugs used 27 . The efficacy of antimicrobial agents against bacterial pathogens is classically predicted from PK/ PD studies. In general, to be efficacious against the pathogens responsible for an infection, drug dosing regimens should ensure that the optimal value of the PK/PD index correlated with the antibacterial activity of the drug is achieved 28 . For fluoroquinolones, a C max /MIC ≥ 10 or an AUC 24h /MIC ≥ 125 is the PK/PD index that reach antimicrobial efficacy 3,29,30 . This finding was also true in this study (Table 5), both in vitro kill curves indicated that the higher the concentration of DAN, the better the bactericidal effect. Moreover, both the AUC 24h /MIC and C max / MIC had a good correlation with antimicrobial efficacy (R 2 = 0.99) by the inhibitory effect sigmoid E max equation. The calculated mean AUC 24h /MIC for ileum ultrafiltrate that achieved bacteriostatic, bactericidal, and eradication were 99.85, 155.57, and 218.02 h. These results are consistent with other studies which AUC 24h /MIC ratios for the quinolones should be at least 125 for optimal bactericidal efficacy against gram-negative pathogens 31 .
Previous studies have found that the increase in antibiotic resistance in the intestinal flora was mainly due to the abuse of antimicrobials in unsuitable routes of administration 32,33 . Therefore, three different doses (1.49 mg/ kg, 2.42 mg/kg and 3.24 mg/kg) were calculated respectively based on the PK/PD index (AUC 24h /MIC ratio) for bacteriostatic, bactericidal and bacterial eradication in the present study. However, it is necessary to design experiments with either naturally diseased animals or disease models to establish the lowest AUC/MICs that lead to the bacterial elimination in vivo 3 . Such experiments will further provide a more rational basis for the selection of optimal dosage schedules for DAN.

Conclusion
In this study, ultrafiltration probes were successfully placed in the ileum of piglets to collect ileum ultrafiltrate. The mean PK/PD index (AUC 24h /MIC) for ileum ultrafiltrate that achieved bacteriostatic, bactericidal, and eradication were 99.85, 155.57, and 218.02 h, respectively. Three different dosages (1.49 mg/kg, 2.42 mg/kg, and 3.24 mg/kg) were calculated respectively based on AUC 24h /MIC ratio above, which might provide a novel approach to the rational design of dosage schedules.