The AAPS Journal ( # 2017) DOI: 10.1208/s12248-017-0055-y

Research Article Theme: Revisiting IVIVC (In Vitro-In Vivo Correlation) Guest Editors: Amin Rostami Hodjegan and Marilyn N. Martinez

Exploring Canine-Human Differences in Product Performance. Part II: Use of Modeling and Simulation to Explore the Impact of Formulation on Ciprofloxacin In Vivo Absorption and Dissolution in Dogs M. N. Martinez,1,7 B. Mistry,1 V. Lukacova,2 K. A. Lentz,3 J. E. Polli,4 S. W. Hoag,4 T. Dowling,5 R. Kona,6 and R. M. Fahmy1

Received 6 December 2016; accepted 2 February 2017

This study explored the in vivo performance of three oral ciprofloxacin Abstract. formulations (oral solution, fast, or slow dissolving tablets) in beagle dogs. The in vivo absorption and dissolution behaviors, estimated with in silico mechanistic models, were compared to the results previously published in human volunteers. Six normal healthy male beagle dogs (five to completion) received three oral formulations and an intravenous infusion in a randomized crossover design. Plasma ciprofloxacin concentrations were estimated by tandem mass spectrometry detection. A mechanistic absorption model was used to predict the in vivo dissolution and absorption characteristics of the oral formulations. Canine ciprofloxacin absorption was constrained to the duodenum/jejunum. This absorption window was far narrower than that seen in humans. Furthermore, while substantial within-individual variability in drug absorption was seen in human subjects, a greater magnitude of variability was observed in dogs. For three sets of data, a lag time in gastric emptying was necessary to improve the accuracy of model-generated in vivo blood level profile predictions. In addition to species-associated dissimilarities in drug solubilization due to human versus canine differences in gastrointestinal fluid compositions, the far more rapid intestinal transit time and potential segmental differences in drug absorption needed to be considered during human-canine extrapolation of oral drug and drug product performance. Through the use of mechanistic models, the data generated in the human and canine studies contributed insights into some aspects of the interspecies differences to be considered when extrapolating oral bioavailability/formulation effect data between dogs and humans. KEY WORDS: canine; ciprofloxacin; interspecies extrapolation; mechanistic model; pharmacokinetics.

INTRODUCTION Electronic supplementary material The online version of this article (doi:10.1208/s12248-017-0055-y) contains supplementary material, which is available to authorized users. 1

The Food and Drug Administration, Rockville, Maryland 20855, USA. 2 Simulations Plus, Inc., 42505 10th Street West, Lancaster, California 93534, USA. 3 Pharmaceutical Candidate Optimization, Metabolism and Pharmacokinetics, Bristol-Myers Squibb, Wallingford, Connecticut, USA. 4 Department of Pharmaceutical Sciences, University of Maryland Baltimore, Baltimore, Maryland 21201, USA. 5 Department of Pharmacy Practice, Ferris State University, Big Rapids, Michigan 49307, USA. 6 Division of Formulation Development, Actavis Inc., Parsippany, New Jersey 07054, USA. 7 To whom correspondence should be addressed. (e-mail: [email protected])

The dog is frequently used as a preclinical animal model to support human drug product development (1,2). Similarly, human drug pharmacokinetic (PK) information is frequently considered during the development of canine oral drug formulations. For that reason, it is important to appreciate the species-specific physiological attributes that can result in dissimilarities in the in vivo dissolution and absorption of oral drug products. For example, as compared to the human gastrointestinal (GI) tract, dogs tend to have a higher and more variable fasted gastric pH, a shorter intestinal residence time, a smaller gastric fluid volume, greater gastric crushing force, higher fasted bile salt content, larger intestinal intercellular pore diameters (contributing to a greater magnitude of paracellular absorption), and a tendency for larger particles to leave the stomach under nonphase IV intervals associated with the interdigestive migrating motor complex (3–7). 1550-7416/17/0000-0001/0 # 2017 American Association of Pharmaceutical Scientists

Martinez et al. Despite these dissimilarities, interspecies extrapolations remain a valuable tool for pharmacologists and formulation chemists. Therefore, it is necessary to understand the interrelationship between drug physiochemical characteristics, GI physiological attributes, and product formulation. Achieving this insight is predicated upon opportunities to differentiate between product in vivo dissolution and absorption processes. Therefore, aligned with that goal, an examination of species-by-formulation interactions was generated using ciprofloxacin as the targeted compound due to the challenges it presents from an in vivo dissolution and absorption perspective. Ciprofloxacin is a zwitterion, exhibiting biphasic solubility (highest at low and high pH, poorest solubility at neutral pH values) (1). The GastroPlus™ software (Simulations Plus, Inc.) was used for modeling the blood level profiles generated in vivo following the administration of an IV and three oral formulations (two tablet formulations plus an oral solution) to dogs. Factors that could influence drug permeability (e.g., intestinal influx and efflux transporters) were accounted for through the use of the absorption scaling factor (ASF). One set of ASF values was used across the three oral formulations (oral solution, a fast dissolving tablet, and a slowly dissolving tablet) for each dog. The data from each animal was modeled individually to facilitate an understanding of between-animal variations. Mechanistic absorption models serve as a tool for translating blood level data to species-by-formulation differences in absorption versus dissolution processes (8–10). Such models are structured so that the systems data (physiology attributes of the subject population), drug data [the physicochemical characteristics of the active pharmaceutical ingredient (API)], and subject-specific PK information are integrated into model estimates describing the absorption and dissolution of the API as it traverses the GI tract. Through the use of intravenous (IV) data to define species-specific drug clearance and distribution attributes and an oral solution to characterize segmental drug permeability, these models enable investigators to explore interspecies differences in segmental events that influence in vivo product performance. In our first publication (11), a mechanistic absorption model was used to examine the in vivo dissolution and absorption characteristics of three oral ciprofloxacin formulations in normal healthy human volunteers. An IV infusion in each individual was used to characterize subject-specific clearance and distribution attributes. In that investigation (hereon referred to as Bstudy 1^), an apparent Babsorption window^ was observed in most human volunteers such that if the drug was not presented in a solubilized form in the upper small intestine, there was difficulty achieving a high level of oral bioavailability. In many (but not all) of these human subjects, the majority of solubilized drug presented within this window was absorbed. However, we also observed several subjects (which were classified as Bpoor absorbers^) in which dissolved drug was not efficiently absorbed within this window, resulting in marked Bapparent^ drug precipitation as dissolved drug moved into the lower intestinal segments. In the presence of a much-reduced magnitude of absorption below the jejunum (based upon in silico fitted estimates), there was either difficulty achieving further drug dissolution

or a decrease in milligrams of dissolved ciprofloxacin. In addition, in the majority of subjects, two sets of counterclockwise relationships were observed in the modeled in vivo dissolution versus absorption profiles: one occurred shortly after gastric emptying and a second event occurred as the drug moved beyond the jejunum. The question was whether a similar phenomena would be encountered when these same formulations were administered to dogs. The current study, which was conducted in normal healthy fasted beagle dogs, employed the identical formulations as those used in study 1 (an IV injection plus three oral formulations). The study was conducted using a fourperiod crossover design. As the human and canine investigations were conducted in parallel, identical batches of formulations were employed, thereby eliminating the potential for batch-to-batch variability that could confound our interspecies comparisons. Accordingly, differences between the investigations were primarily attributable to canine-human differences in in vivo dissolution and absorption processes. In this regard, it should be noted that the model used for fitting the data both in human and canine subjects did not address the issue of first-pass extraction. Through the use of mechanistic models, the data generated in the human and canine studies provided insights into aspects of the interspecies differences to be considered when extrapolating oral bioavailability/formulation effect data between dog and man. MATERIALS AND METHODS Production of Ciprofloxacin Tablets As summarized (in greater detail) in study 1 (11), intragranular components were blended in a twin-shell blender (Paterson-Kelley Company, East Stroudsburg, PA), and the powder blend was compacted on a roller compactor (Model WP 120 V Pharma, Alexanderwerk Inc., Horsham, PA). Although both solid oral dosage forms were considered immediate release tablets, they were manufactured to have fast or slow release. The fast and slow formulations selected for use in the clinical trials differed in their respective amounts of hydroxypropyl cellulose (HPC) (w/w = 2 for fast and 6 for slow), the amount of starch (w/w = 5 for the fast tablets and 4 for the slow tablets), and the feed screw feed speed/roller speed ratio (FSS/RS), which was 3 for the fast tablets and 5 for the slow tablets. Tablets were manufactured under current good manufacturing practices (GMP) within the University of Maryland, School of Pharmacy. In Vitro Dissolution In vitro dissolution methods, the corresponding ciprofloxacin assay, and the corresponding in vitro dissolution profiles are described in study 1. Briefly, the in vitro dissolution studies were evaluated using USP apparatus II (paddle) at 50 rpm. Temperature was maintained at 37 ± 0.5°C. Studies were carried out at pH 2, 4.5, and 5.5 and at

Exploring Canine-Human Differences in Product Performance ionic strengths (IS) of 0.01, 0.05, and 0.2. The amount of ciprofloxacin was analyzed using a validated highperformance liquid chromatography (HPLC) as described in the USP ciprofloxacin monograph (12). In Vivo Bioavailability Study The study involved six normal healthy male beagle dogs, with starting weights of 8.1 to 10.2 kg (7.9 to 10 kg finishing weights). Weight change between the first dose (IV) and the last dose was less than 8%. One of the six dogs needed to be eliminated from the study due to a massive histamine release during the ciprofloxacin IV infusion. Dogs were ambulatory and manually restrained. The oral doses were administered to the back of the throat. Dogs were encouraged to swallow by gently rubbing the throat. The study involved four ciprofloxacin formulations: a commercially available IV solution (200 mg ciprofloxacin free base/20 mL) that was used both for the IV infusion and as an oral solution (sol) [10 mg/mL free base oral solution (sol)], and two tablet formulations [fast release tablets (fast) and slow release tablets (slow)], each containing 200 mg ciprofloxacin as the HCl salt (172 mg as the free base).1 These were the identical tablets and batches as was used for human administration in study 1 (11). Doses administered were as follows:

&

IV: 50 mg (10 mg/mL). All dogs received the IV 5-min infusion during period 1. & Sol: 75 mg per dog (7.5 mL of the 10-mg/mL IV solution). & Fast tablets: 200 mg ciprofloxacin (172 mg free base) + 10 mL water flush. & Slow tablets: 200 mg ciprofloxacin (172 mg free base) + 10 mL water flush. Prior to drug administration, all dogs were subject to an overnight fast (at least 16 h). A standard canine diet was provided 2 h after dosing following IV dose and 4 h following oral doses. The oral dosage forms were randomized into one of three sequences: Sequence A: Sequence B: Sequence C:

sol, slow, fast: dogs 1 and 2 slow, fast, sol: dog 3 fast, sol, slow: dogs 4 and 5

A 7-day washout interval separated the study periods. Only one case of emesis (yellow foam) occurred 1 h and 20 min after drug administration. Venous blood (0.3 mL) was collected from the cephalic or saphenous vein into sodium EDTA tubes. The corresponding blood sampling schedule was as follows: IV: 0 m, 5 m, 10 m, 15 m, 30 m, 45 m, 1 h, 2 h, 4 h, 6 h, 8 h, 12 h, 24 h PO: 0 m, 15 m, 30 m, 45 m, 1 h, 1.5 h, 2 h, 4 h, 6 h, 8 h, 12 h, 24 h

Samples were centrifuged, and plasma was transferred to polypropylene tubes for storage at −20°C until analyzed by LC/MS/MS. Institutional Animal Care and Use Committee (IACUC) approval was obtained prior to initiating the investigation. The study was not conducted in accordance with GLP regulations. Analytical Methods A bioanalytical method utilizing liquid chromatography separation followed by tandem mass spectrometry detection (LC/MS/MS) was developed for the analysis of ciprofloxacin. The plasma standard curve and quality control (QC) samples were prepared in male beagle dog plasma (sodium EDTA as anticoagulant) by spiking an appropriate amount of stock solution concentration of approximately 4 mg/mL prepared in 50% acetonitrile in water. Ciprofloxacin hydrochloride monohydrate (Quimica, Madrid, Spain) was used as the reference standard. Lomefloxacin hydrochloride (internal standard) stock solution concentration of approximately 1 mg/mL was prepared in 50% acetonitrile in water. The ciprofloxacin standard curves in dog plasma had a dynamic range of 2.4 to 20,000 ng/mL. For the extraction of analytes, an aliquot of 40 μL of plasma standard curve, QC, or study sample was mixed with 200 μL of working internal standard solution (500 ng/mL) in acetonitrile. The sample mixture was gently vortexed to precipitate plasma protein and centrifuged at 1000 rpm for 2 min. The supernatant organic liquid was collected and dried under nitrogen at 40°C, and the resultant residues were reconstituted in 125 μL of 10% acetonitrile in water. An aliquot of 5 μL final extract was injected into equilibrated HPLC system equipped with Phenomenex Luna C-18(2)-HST (Torrance, CA) analytical column maintained at 60°C and detected using tandem mass spectrometry detection (API 4000 Qtrap (303A)) with electrospray (Sciex, Framingham, MA). The mobile phase A contained 1% formic acid in deionized water and mobile phase B contained 1% formic acid in acetonitrile. The analytes were eluted at a flow rate of 0.4 mL/min using a gradient elution method by changing organic mobile phase B percentages [5% (0 to 1.1 min), 95% (1.1–1.4 min), and back to 5% (1.5 to 1.8 min)]. Detection of ciprofloxacin was performed using selected reaction monitoring mode. Ions representing the precursor (M+H)+ species of ciprofloxacin were selected in quadrupole 1 and collisionally dissociated with N2 to generate specific product ions that were subsequently monitored by quadrupole 3. An ion transition of (M+H)+: 332.3 → 231.4 for ciprofloxacin and (M+H)+ 352.2 → 237.2 for lomefloxacin was monitored. The unknown study samples were quantitated using standard curves (ranging from 2.4 to 20,000 ng/mL) fitted using linear regression with a weighting factor of 1/Concentration2. The typical coefficient of determination (R2) for regression was >0.99. The accuracy of all standards was within 22% of the standard curve and QCs prepared from a separate weighing were within 32% of the standard curve. In Silico Model

1

Please note that in the write-up of study 1, we inadvertently labeled the base drug amount as 180.2 mg. The correct milligram free base amount of ciprofloxacin is 172 mg, as indicated in this write-up.

Physicochemical Characteristics The ADMET Predictor™ version 7.2 (Simulations Plus™, Lancaster, CA) was used to estimate pKa, solubility,

Martinez et al. logP, molecular weight, Peff (the effective passive human intestinal permeability expressed in terms of 10−4 cm/s), bile salt effects, and diffusion coefficients. Within the constraints of the available information, we defaulted to a model that defined solubility on the basis of human bile salt effects. The ADMET Predictor™-generated ciprofloxacin physicochemical characteristics were originally provided in Table II in study 1. For this canine study, a copy of that information is provided in Supplemental Table 1.

The Mechanistic Model The default fasted beagle intestinal physiology in the GastroPlus software (V9.0) (10) was used for modeling in vivo product performance. This analysis provided a mechanism for evaluating a best-fit estimate of the concentration of dissolved drug and the milligram amount of drug (percent total dose) absorbed from each of the intestinal segments. The mechanistic modeling of the canine data was effectively identical to that used for the human dataset. Empirical compartmental PK models (two or three compartments) provided the basis for obtaining subject-specific estimates of the volumes of distribution (Vd) and total systemic clearance (CL) from the plasma ciprofloxacin concentration versus time profile observed following the IV dose. Based upon an objective weighting function of 1/Yhat2, the Akaike information criterion (AIC) was used for selecting the best-fit compartmental model for each study subject. All dogs were best fit by a three-compartment model. The dog-specific regional ASFs were obtained by fitting the oral solution data. The resulting subject-specific PK parameter values were used for predicting the oral in vivo dissolution and absorption profiles based upon the ciprofloxacin systemic drug concentrations obtained with the fast and slow dissolving tablets. The ASF is a multiplier used to scale the effective permeability to account for variations in surface-to-volume ratio, pH effects, and other passive absorption-rate-determining effects that differ between intestinal compartments. The segmental passive absorption is normally defined by the segmental ASF value × Peff. However, ciprofloxacin absorption can be influenced both by influx (13) and efflux (14) transporters in addition to passive diffusion. Therefore, as was the case for the human simulations generated in study 1, the user-defined ASF values were estimated so that it accounted for regional changes in absorption rate due to regional differences in passive diffusion as well as influx and/or efflux transporters (15). A copy of the user-defined ASF values is provided in Supplemental Table 2. Although some of the optimized values are large, they nevertheless are consistent with a compensation for transporter function and are comparable to that reported elsewhere in the literature (16). User-defined estimates of ASF were obtained using the optimization module (weighted sums of squares) and an objective weighting function of 1/Yhat2 and the Hooke and Jeeves pattern search method. Optimally, the ASF values are estimated on the basis of an oral solution, thereby separating absorption from dissolution processes. For four of the five dogs, the solution-based segmental ASF estimates were used when modeling the tablets (fast and slow) without the need for any further modification. However, as was the case for some of the human subjects, the rapid absorption of the oral solution necessitated ASF optimization as a two-step process

in one of the dogs in order to obtain ASF estimates in the lower small intestine and in the large intestine. When a twostep process was employed, the solution-based ASF values were used in the duodenum and jejunum, and the slow dissolving tablets were used to estimate the ASF values in the ileal segments, cecum, and ascending colon. For consistency in the physiological parameters used within a given dog, a single set of ASF values was used when generating the final model estimates of in vivo dissolution and absorption across the three oral formulations. The final step in the fitting procedure was the determination of the drug in vivo dissolution. For the fast dissolving tablets, the dissolution was first modeled using the Johnson dissolution model (17), with in vivo solubility and effective diffusion coefficient adjusted for the bile salt effect (18). The diffusion layer was adjusted with changing radius up to a maximum diffusion layer thickness of 30 μm (default value). A mean particle size radius of 25 μm (default value) was used as input for the dissolution model. However, for four of the five dogs, both the fast and slow dissolving tablets needed to be modeled as a dispersed controlled release formulation (CR dispersed). The in vivo dissolution profile was fitted to a double Weibull function using the optimization module with an objective weighting function of 1/Yhat2. For all dogs, the lag time (Tlag) was included as one of the Weibull parameters that was estimated. As indicated in the BRESULTS^ section, on three occasions, the gastric emptying time of fast or slow dissolving tablets was adjusted within the bounds of 0.25 to 0.75 h to optimize the model fit. The basics and scientific underpinnings of our semimechanistic modeling approach were detailed in the supplemental material provided in study 1. Noncompartmental Parameter Estimation For each dog and formulation, the area under the curve from time zero to 24 h (AUC0–24), the AUC from time zero to time infinity (AUCinf), peak drug concentrations (Cmax), and the time to Cmax (Tmax) were calculated for both the observed and modeled plasma ciprofloxacin concentration versus time profiles as part of the data output generated by the GastroPlus software. Using the observed data, total bioavailability (F) was defined by the AUCinf values calculated after oral (po) or IV administration (AUCinf-po/AUCinf-IV), where the respective AUCinf values are corrected for differences in administered dose. RESULTS In Vitro Profiles and In Vivo/In Vitro Relationships During the in vitro dissolution study, the fast tablets consistently released more than 85% of its ciprofloxacin contents within 10 min, and this release was relatively unaffected by pH within the range of 2 to 5.5 or by the IS of the dissolution medium. In contrast, the slow release tablets were highly sensitive both to pH and IS in vitro. At any pH, the slow tablet dissolution rate was greater under conditions with the higher IS (Fig. 1a). Since the slow tablets were more influenced by the conditions of the in vitro dissolution evaluation, the mean

Exploring Canine-Human Differences in Product Performance fraction dissolved in vitro, across each of the test conditions, was compared to the predicted in vivo fraction dissolved (Fig. 1b). From Fig. 1b, it is evident that the in vitro dissolution was not indicative of in vivo events. This is consistent with what was observed in people. In Vivo Analysis The IV data were best described as a three-compartment body model for all dogs (R2 > 0.99 for each dog). However, the inclusion of a third slowly depleting terminal phase resulted in a long terminal elimination half-life (T½) estimate (mean = 73.8 h, %CV = 94 h). When modeled as a two-compartment model, the average terminal phase T½ was 2.87 h (%CV = 12.9). The average clearance per kilogram (0.53 L/h/kg, 19.4%CV) was similar to that estimated for humans in study 1 (0.44, 13%CV). While the estimated canine volume of the central compartment (0.29 L/kg, 47.1%CV) was similar to that estimated in humans

Fraction dissolved

a

(0.22 L/kg, 69%CV), the volumes of the second and third compartments in dogs were substantially larger on a liter per kilogram basis (1.35 L/kg, 22.86%CVand 2.12 L/kg, 99.9%CV for V2 and V3, respectively) than those estimated in people (0.60 L/ kg, 45%CV and 1.31 L/kg, 70%CV for V2 and V3, respectively). While the absolute bioavailability of the oral solution was similar in dogs and humans, the dogs exhibited far greater difficulty in absorbing the slow dissolving tablets (lower F, Fig. 2). Due to the very different number of subjects in the human versus canine study (n = 16 humans, n = 5 canine), the comparisons were not statistically compared. The trend toward a lower absorption in dogs versus humans was also seen for the fast tablet formulation, even though the difference was smaller than for the slow tablet formulation. One of the most outstanding characteristics of the canine study results was the very large intersubject variability in the formulation-associated ciprofloxacin in vivo bioavailability. The range of observed absolute bioavailability was 4–73% for

Invitrodissolution(fast and slowtablets)

F pH2, 0.01 IS

1

F pH2, 0.05 IS

0.9

S pH2, 0.01 IS

0.8

S pH2, 0.05 IS

0.7 F acetate 4.5, 0.05 IS

0.6 0.5

F acetate 4.5, 0.2IS

0.4

S acetate 4.5, 0.05 IS

0.3

S acetate 4.5, 0.2 IS

0.2 F acetate 5.5, 0.05 IS

0.1

F acetate 5.5, 0.2IS

0 0

40

80 Time (min)

120

S acetate 5.5, 0.05 IS S acetate 5.5, 0.2 IS

b

Fig. 1. In vitro dissolution results. a In vitro dissolution release profiles for the fast (F) and slow (S) dissolving tablets as a function of pH and IS. b Relationship between the mean fraction dissolved in vitro across each of the test conditions was compared to the predicted in vivo fraction dissolved. Comparisons reflect in vitro and in vivo dissolved at a specified time (postdose or of in vitro testing)

Martinez et al.

Fig. 2. Comparison of observed ciprofloxacin absolute bioavailability in the three formulations: dogs (n = 5) versus human (n = 16)

slow dissolving tablet, 28–108% for fast dissolving tablet, and 49–100% for solution administration. Moreover, the rank ordering was not consistent across formulations (i.e., the animal with high bioavailability for one formulation may not have exhibited high bioavailability for the other two formulations) (Table I). The modeled mean AUC (AUC0–24, AUC0-inf) and Cmax for the solution and slow tablets were within 10% of the observed values. The modeled mean AUCs for the fast tablet and the mean Cmax for the fast and slow tablets had prediction error in the range of 18–24%. These higher prediction errors were, however, the result of a single outlier. In particular, there was difficulty associated with modeling the profile of the fast tablets for dog 2. This was not surprising given its observed absolute bioavailability of 108%, suggesting that some change was associated with the clearance of that dog during that particular study period. When excluded from the mean, the error in the AUC and Cmax values was less than 8%. As evidenced by the PK parameter values in Table I, although there were only five dogs included in this investigation, instances of discordant formulation-associated profiles were observed. For dog 1, improvement of the fit of the fast tablets was achieved by the introduction of a 0.25-h delay in gastric emptying of the tablet. In the case of dog 2, the bioavailability of the fast tablets was substantially greater than that of even the oral solution. Moreover, for dog no. 2, there was a clear delay between dosing of the slow tablets versus when the ciprofloxacin first appeared in the blood. This necessitated the introduction of a 0.75-h lag time in gastric emptying. With respect to dog 4, the slow dissolving tablets resulted in higher ciprofloxacin exposure as compared to that of the fast tablets, even though most dogs exhibited very poor ciprofloxacin bioavailability when the slow dissolving formulation was administered. In fact, for dog 4, the observed Cmax values approached that of the dose-normalized oral solution. Dog 4 also necessitated the introduction of a lag time (0.25 h) in the slow dissolving tablets. As a result of these irregularities, situations were encountered whereby it was difficult to

fit all three formulations in the same dogs (see Fig. 3a–e). In contrast, the model fit achieved for the IV infusion data was very closely aligned with the observed values (Fig. 3f). As previously observed in human subjects, a counterclockwise relationship was seen between the predicted milligram in vivo dissolution versus absorption (bottom graphs, Fig. 4a–c). However, in dogs, this was less pronounced than what was seen in humans. Furthermore, in contrast to humans, this event occurred primarily after the drug left the stomach. Also, in dogs, this counterclockwise relationship was primarily associated with the oral solution, being seen after administration of the fast and slow dissolving tablets in only one dog. In all dogs, the absorption profile lagged behind the in vivo dissolution profile for approximately 2 h after tablet administration, which was also the time when absorption reached a plateau (top graphs, Fig. 4a–c). In contrast, in humans, although most of the absorption also occurred within 2 h postdose, ciprofloxacin absorption continued for about 3–4 h after tablet administration (11). As compared to dogs, the combination of a longer duration of absorption and the markedly different shape in the in vivo dissolution/time profile resulted in a lag time between human in vivo dissolution and absorption that was substantially wider than what was predicted in dogs. Canine-human differences are evident when the predicted segmental absorption values are considered. For both species, the primary locations of absorption were the duodenum and jejunum, with negligible absorption was predicted beyond the jejunum both in humans and dogs (Fig. 5). Nevertheless, from a qualitative perspective, differences could be observed. On the average, dogs absorbed more of the administered dose of the oral solution and of the fast tablets in the duodenum as compared to that seen in humans, with effectively no difference in fraction of administered dose absorbed in the duodenum for the slow formulation. Yet, in the upper jejunum, a greater fraction of the administered dose absorbed was already apparent in humans as compared to dogs across all dosage forms.

Exploring Canine-Human Differences in Product Performance Table I. Canine Pharmacokinetic Observed and Predicted Parameter Values

Units AUC0–24 (M) AUC0–24 (O) AUCinf (M) AUCinf (O) Cmax (M) Cmax (O) Tmax (M) Tmax (O) Modeled Fa Observed F IV parameters

μg h/mL

μg/mL h

Oral solution Mean 10.73 11.53 10.82 11.74 2.69 2.65 0.55 0.60 74.82 77.55

Mean CL (M) L/h/kg 4.97 CL/kg (M) L/h/kg 0.53 Vc/kg (M) L/kg 0.29 V2/kg (M) 1.35 V3/kg (M) 2.12 T½ (deep—M) h 73.83 Absolute bioavailability (F, obs) by dog Dog number Oral solution Fast tablet 1 0.93 0.88 2 0.65 1.08 3 1.00 0.28 4 0.80 0.43 5 0.49 0.37

%CV 23.62 21.99 23.57 20.19 18.43 19.12 53.86 47.51 31.98 26.42

Fast tablet Mean %CV 16.03 39.03 19.81 49.37 16.18 39.19 21.24 55.46 3.40 44.99 4.10 49.96 0.87 62.97 1.20 45.17 47.88 42.09 60.67 57.77

Fast tablet w/o subject 2 Mean %CV 16.01 45.10 17.21 52.77 16.13 45.37 17.37 53.18 3.46 50.96 3.62 55.63 0.95 63.33 1.31 42.24 47.90 48.59 48.91 54.73

Slow tablet Mean %CV 8.92 108.13 9.04 104.49 8.98 107.62 9.10 103.91 1.91 111.26 2.12 116.86 1.10 47.07 1.20 45.17 27.27 114.86 26.48 109.09

%CV 12.01 19.40 47.10 22.86 99.90 94.08 Slow tablet 0.04 0.35 0.06 0.73 0.14

Moreover, in dogs, the absorption in the ileum and below was far less than that associated with the estimates generated on the basis of the human dataset, which is consistent with the very short length of the canine ileal segment. In fact, in only one of the five dogs was drug absorption predicted to have occurred in ileum 1 (dog 1, fast and slow tablets). In all the other dogs, the model predicted no drug absorption beyond jejunum 2. It is also informative to examine the differences in withinanimal variability in formulation effects across the five study subjects (Fig. 6). For example in dog 5, absorption was constrained primarily to the duodenum and jejunum 1. Although far more absorption occurred in the duodenum versus jejunum 1 for the solution, a greater fraction of the administered ciprofloxacin was absorbed in jejunum 1 versus duodenum for the fast and slow tablets. This is consistent with dissolution limitations in dog 5. In dog 3, jejunum 2 was an important site of absorption for the oral solution and the fast tablets, but showed similarly compromised absorption across all other intestinal segments after administration of the fast or slow tablets. For dog 4, a similar fraction of dose absorbed occurred with the slow tablets and solution, but absorption was compromised for the fast tablets. Although this indicates problems dissolving the fast tablet in dog 4, the reason why the slow tablets performed better than the fast tablets in that dog is unclear. Importantly, it points to the high variability that occurs when low solubility compounds are administered to dogs. Although the difference in ability to dissolve and absorb the slow tablets and location of absorption within the gastrointestinal tract was an important observation, the most

outstanding feature of the dog data was the intersubject diversity in the drug bioavailability. Even though the human subjects could be categorized as Bgood,^ Bpoor,^ and Berratic^ absorbers, the magnitude of intrasubject differences in canine ciprofloxacin tablet bioavailability exceeded that observed in humans. To emphasize this point, the relationship between fraction absorbed by subject and formulation in humans (predicted values from study 1) is compared to the observed canine data (Fig. 7). While the oral bioavailability of the oral solution was not remarkably different from that observed in the human subjects, there was far greater within-dog difference in the relative magnitude of oral absorption of the two tablet formulations or of the tablets relative to the oral solution. For example, the slow tablets had less than 10% bioavailability in two dogs, but in another dog, the bioavailability of the slow dissolving tablet approached that of the oral solution, far exceeding the absolute bioavailability of the fast tablet for that animal (since the model did not account for first-pass extraction, estimates of absorption and bioavailability were equivalent). DISCUSSION A semimechanistic modeling approach was used to support an evaluation of the in vivo behavior of ciprofloxacin oral tablets in dogs and to compare these results to those previously published for human volunteers (study 1). The maximum fitted estimates of in vivo dissolution and absorption were more rapidly achieved in dogs as compared to humans (Fig. 8). These interspecies differences were most

Martinez et al. c

Observed vs modeled profiles, Dog 1

8

Observed vs modeled profiles, Dog 3 8

SLTN (N)

7 Fast

6

Slow

5

Obs SLTN (N)

4

Obs Fast

3

Obs Slow

2

SLTN (N)

Concentration (mcg/mL)

Concentration (mcg/mL)

a

1

7

Obs SLTN (N)

4

Obs Fast

3

Obs Slow

2 1 0

0

b

2

4

6 Time (hr)

8

10

12

0

2

d

Observed vs modeled profiles, Dog 2

4

6 Time (hr)

8

10

12

Observed vs modeled profiles, Dog 4 8

8 SLTN (N)

7

Concentration (mcg/mL)

Concentration (mcg/mL)

Slow

5

0

Fast

6

Slow

5

Obs SLTN (N)

4

Obs Fast

3

Obs Slow

2 1

SLTN (N)

7

Fast

6

Slow

5

Obs SLTN (N)

4

Obs Fast

3

Obs Slow

2 1 0

0

e

2

4

6 Time (hr)

8

10

f

Observed vs modeled profiles, Dog 5 8 7

SLTN (N)

6

Fast

5

Slow

4

Obs SLTN (N)

3

Obs Fast Obs Slow

2

0

12

1

Concentration (mcg/mL)

0

Concentration (mcg/mL)

Fast

6

2

4

6 Time (hr)

8

10

12

Mean (stdev) observed versus modeled IV concentration-time profile Pred Obs

5

0.5

0 0

2

4

6 Time (hr)

8

10

12

0

2

Time (hr)

4

6

Fig. 3. Relationship between observed and predicted ciprofloxacin plasma concentrations as a function of dog and formulation (a–e). Mean observed and predicted IV blood level profile across all five dogs (f). (Note that the scale of the X and Y axes has been truncated so that any difference between the lines describing the observed and predicted values, as well as the corresponding standard deviations, is visible). SLTN(N) indicates that the oral solution plasma concentrations have been normalized to that of the tablets in order to facilitate profile comparisons

pronounced with the slow dissolving tablets reflecting the canine earlier and narrower absorption window, shorter GI tract, and the likely effect of differences in dog-human fasted GI fluid composition. In this regard, there are three interspecies-specific GI attributes that may have contributed to these human-canine disparities (3,4,19) (Fig. 9). 1. Transit time and length of the GI tract: The dog intestinal transit time is far more rapid than that of humans, necessitating that in vivo dissolution occurs within a shorter timeframe. 2. Absorptive surface area: The shape and surface area of the various intestinal segments of the dog differ from those of humans. In addition, there are likely to be differences in the location and abundance of transporters (influx and efflux) that need to be considered. Based upon the information provided in the review by Kararli (19), the intestinal absorptive surface area is about two- to threefold smaller than that of humans. 3. The gastric pH of dogs is typically higher than that seen in humans, rendering ciprofloxacin less soluble in the

canine versus human gastric fluids. Furthermore, differences in intestinal pH and bile salt composition likewise differ, impacting the in vivo dissolution of low solubility compounds (and particularly that of zwitterions). Differences in the dog versus human physiological model that could influence tablet dissolution and absorption are shown in Supplemental Figures 1, 2, and 3. Please note that with the exception of canine gastric emptying time [which can be highly variable in dogs (11)], GastroPlus default canine GI parameter values were used when fitting the data generated with the oral dosage forms. Furthermore, there appeared to be some dogs in which the dosage form failed to disintegrate, remaining within the stomach. These three cases necessitated the introduction of a lag time in gastric emptying. The difference in tablet bioavailability may also be a reflection of differences in in vivo dissolution due to corresponding dog-human differences in the amount of water administered upon dosing. In the current investigation, tablets were administered with a 10-mL water flush, while in study 1, the tablets were administered to human volunteers

Exploring Canine-Human Differences in Product Performance

Fig. 4. Deconvoluted dissolution versus absorption after administration fast tablets (a), slow tablets (b), and oral solution (c). The graphs for each formulation include dissolution versus time and absorption versus time profiles (mean ± SD), model transit time in the canine intestine, and deconvoluted in vivo dissolution versus in vivo absorption profile (each estimated at the same time after administration). Error bars are represented by the blue (dissolution) or red (absorption) vertical lines. The bottom graph within each dosage form provides the in vivo dissolution versus absorption profiles for each of the canine study subjects

with 240 mL of water. Whether or not the small fluid volume contributed to the instances of delayed gastric emptying of the tablet is unclear but does suggest a potential source for the higher within- and between-subject variability in oral dosage form performance in dogs. The issue of gastric fluid volume was further confounded by the use of identical tablet batches in humans and dogs. This decision was grounded in an effort to avoid the introduction of dissolution variability due to the manufacture of different tablet sizes. Consequently, the milligram per kilogram dose used in dogs was greater than that in humans (172 mg free base/9.5 kg, average dog weight, study 2 versus 172/73.8 kg, average human weight, study 1). At least in part, this greater variability in dogs versus humans may be attributable to the more rapid transit through the intestine of the dog as compared to that of humans (particularly when considering the ileum where the modelbased transit times are 0.06 versus 1.29 h for dogs and humans, respectively). Furthermore, the variability in the gastric pH of normal healthy dogs appears to be very wide,

with normal fasted values ranging from 1.2 to over 6. In fact, dogs appear to exhibit periodic spikes in their gastric pH under fasted conditions (20). Particularly for a drug such as ciprofloxacin, this within- and between-animal variability could have a large influence on the dissolution behavior of a dosage form or on the precipitation of ciprofloxacin when administered as an oral solution. Given the differences in milligram per kilogram dose administered to humans and dogs, the potential model overestimation of formulation differences in canine oral drug absorption due to saturable elimination or to first pass drug loss needed to be considered. Upon examination of information obtained from other published ciprofloxacin studies, the issue of saturable elimination does not appear to be a concern as the canine IV CL and V2 values estimated in our data analysis were nearly the same as those reported by Papich (21) following an IV injection of 9.8 mg/kg dose (to six normal healthy beagle dogs). While the IV dose administered in our investigation was only 50 mg/dog and, therefore, does not necessarily cover the potential for saturable clearance

Martinez et al.

Fig. 4. continued.

Exploring Canine-Human Differences in Product Performance

Fig. 5. Mean fraction (stdev) of administered dose absorbed as a function of formulation across the intestinal segments of dogs (a) and humans (b)

processes if most of the 172-mg dose was absorbed (equaling a dog dose of about 18 mg/kg), we believe that it is unlikely that saturable elimination or first pass metabolism influenced our study results because of the following reasons:

&

Reported dose proportional PK was observed when dogs were administered IV doses in the range of 5 to 10 mg/kg (22). & Those same authors reported that while linear PK was associated with the oral absorption of ciprofloxacin at a dose of 10 to 20 mg/kg to four fasted beagle dogs, greater than dose proportional drug exposure was seen at oral doses of 40 mg/kg (23). Mean plasma ciprofloxacin concentrations following a 20-mg/kg oral dose in that published investigation was consistent with the plasma drug concentrations seen following administration of the fast tablets in the current study. & With regard to first pass drug loss, given that the limited ciprofloxacin loss by first-pass metabolism in humans is considered to be of negligible clinical significance (24), we consider it unlikely that a saturable first-pass effect was responsible for the exceptionally high exposure seen in some dogs (e.g., dog 2). Thus, concluding that neither saturable drug clearancen or first pass metabolism detracted from the applicability of

our mechanistic absorption model as a tool for examining the data generated in the current investigation, we focused the remainder of our assessments on the absorption and dissolution processes. In terms of the possibility of saturable absorption processes, whilethe bioavailability of the oral solution exceeded that of the tablets in most dogs, the milligram per kilogram dose of the oral solution was lower than that of the tablets. Nevertheless, when considering Fig. 7 (where we see that differences in formulation effects across dogs ranged from the fast tablets being more bioavailable than the oral solution in dog 2 to markedly lower tablet versus oral solution bioavailability in dog 3, and to dog 4 where the fraction absorbed in the slow tablets and oral solution were similar), it appears that factors other than saturable absorption was responsible for the tremendous within- and between-dog variability that was observed. This variability far exceeded what was observed with the identical formulations when administered to human volunteers. One factor that may have contributed to the greater within- and between-subject variability in dogs appears to be their narrow absorption window, leading to a higher degree of sensitivity to variability in tablet dissolution. This emerges as a likely contributing factor when comparing the graphs in Fig. 4 (dogs) to the corresponding figures in humans (Fig. 3 of study 1 and also provided as Figure 1 Supplemental Material in this report) where the magnitude of the gap between

Martinez et al.

Fig. 6. Model-predicted segmental drug absorption as a function of dog and formulation

in vivo dissolution and absorption events was wider in humans than in dogs.

Although the smaller intestinal absorptive surface associated with dogs versus humans (19) could have contributed to a lower bioavailability, it would not lead to the distinct absorption window observed in our fitted absorption estimates. In that regard, while an absorption window appears to be present in the upper small intestine of humans and dogs, that window was narrower (but possibly more efficient) in dogs. Influx across the intestinal mucosa of rats appears to involve Oatp1a5, and in humans, the transporter appears to be organic anion transporter protein 1A2 (OATP1A2) (13). Other possible influx transporters, such as the organic cation transport protein (OCT), have been suggested (24). With respect to efflux, despite some evidence to the contrary (13), it is unlikely that P-glycoprotein (P-gp) is involved in the efflux of ciprofloxacin (25). Rather, ciprofloxacin efflux may be more closely aligned to breast cancer resistance protein (BCRP) (14). If this corresponds to protein expression, one might expect substantial amounts of efflux transporter in the terminal ileum. The relative abundance of the various influx and efflux transporters within the human small intestine is discussed elsewhere (26). In humans, BCRP messenger RNA (mRNA) expression was highest in the duodenum and terminal ileum, with smaller but still substantial expression through the sigmoidal colon (27). Based upon the work of Haller et al. (28), mRNA expression of BcrpP in the canine GI tract is highest in the duodenum and upper jejunum. However, Haller et al. also observed that influx transporters such as the peptide transporter 1 (PepT1) and Oct1 are likewise highest in the upper portions of the small intestine (although very low levels of Oct1 mRNA were observed within any of the canine intestinal segments). Thus, there is the possibility that with respect to the duodenum, there is a concomitant presence of BCRP efflux protein plus some influx transporter such as Oatp1A2. Unfortunately, although there does exist some information on canine intestinal transporters (29), much more work is needed to more fully appreciate the segmental distribution and activity of these influx and efflux proteins in dogs and thereby appreciate their potential influence on canine oral drug absorption.

Fig. 7. Percent fraction of ciprofloxacin absorbed as a function of formulation and species. It should be noted that due to the construct of our model whereby CL was defined on the basis of empirical compartmental PK models, the fraction absorbed was synonymous with the bioavailable fraction. The subject-specific data are stacked within vertical columns. Symbols reflect speciesspecific data generated with the solution, fast, or slow tablets

Exploring Canine-Human Differences in Product Performance

Fig. 8. Relationship between mean predicted in vivo dissolution and absorption of fast and slow dissolving tablets in dogs and humans. Plots are presented either as milligrams dissolved as a function of time or as a plot of milligrams absorbed or dissolved in dogs and humans at identical times after drug administration. For the dog versus human plots, the black hatched line reflects the line of unity to facilitate a visual comparison

Despite the potential presence of BCRP in the human ileum, the greater amount of ciprofloxacin absorption occurring in the ileum of the human volunteers as compared to that predicted based upon the canine data (Fig. 5) is consistent with the difference in model physiological parameters. In dogs, the short ileum compartment was associated with very brief residence time (Supplemental Figure 1, default canine physiologic parameter values of 0.02 h for ileum 1, 2, and 3). In contrast, transit time durations of 0.58, 0.42, and 0.29 h were the default physiologic parameter transit time values associated with the human ileum 1, 2, and 3, respectively. Consistent with the results of our study, Papich (21) likewise observed substantial intersubject variability in canine ciprofloxacin oral bioavailability when human generic 250 mg ciprofloxacin tablets were administered with a 12 mL water flush to six fasted, normal healthy beagle dogs (average weight of 11.2 kg). Using numerical deconvolution methods, the dogs in the Papich study were found to absorb from 32% to 80% of the administered dose (mean = 58.3%, %CV = 45%). However, far less variability was seen when four of these six dogs were administered an oral solution (mean = 70.5%, %CV = 7.3%). The authors speculated that if there were transporter-associated drug absorption (in the upper portion of the small intestine), it would lead to far greater variability when the dogs were administered oral tablets (where variability in dissolution would lead to variability in fraction absorbed) as compared to that of the oral solution. This conclusion is consistent with the evaluations generated with the current data using

mechanistic absorption models where the vast majority of drug was predicted to be absorbed in the duodenum, irrespective of whether the formulation was administered as fast or slow dissolving tablets or as the oral solution. However, an additional point raised by our model was that the fundamental challenge to oral tablet bioavailability was that of achieving adequate dissolution. While the focus of this study was to examine the impact of canine-human differences in GI tract physiology on in vivo product performance, we also wish to suggest that these results may have implications with regard to potential alterations in formulation performance encountered with some human gastrointestinal pathologies (e.g., achlorhydria, hypermotility) (30). In that regard, the use of mechanistic models as tools for interpreting canine oral bioavailability data may provide important insights that need to be considered during formulation development for human use (31). Furthermore, the current discussion does not cover canine-human differences in in vivo product performance in the presence of food. Given the recognized interspecies differences in GI responses to a meal (10,32), it would be beneficial to explore future opportunities for evaluating the extent to which mechanistic models can support our prediction of dog-human differences in the influence of food on in vivo product performance. CONCLUDING THOUGHTS Although the current study design employed mechanistic models that were clearly fit for the purpose, the models

Martinez et al.

Fig. 9. Intersubject variability comparison in predicted amount of drug absorbed versus dissolved in humans and dogs over the first 4 h after drug administration: a oral solution, b fast tablets, and c slow tablets

nevertheless succeeded in providing important insights into the differences in in vivo absorption and dissolution behavior of two tablet formulations in dogs versus humans and identifying questions for future study. Within the framework of this investigation, the following are points to ponder when considering interspecies differences in in vivo product performance: 1. The inherent permeability of the drug and the location of drug absorption within the GI tract: Based upon the results of our evaluation, this is a critical variable to consider when determining whether or not interspecies extrapolations may result in biased conclusions. The presence of an absorption window in the upper small intestine will lead to greater sensitivity to formulation effects in dogs as compared to humans. In turn, variability in in vivo dissolution will be magnified, resulting in greater within- and between-subject variability in dogs as compared to humans. 2. The more rapid GI transit time in dogs: With more rapid transit through the GI tract, there is a greater likelihood of variability in vivo dissolution in dogs as compared to humans. This can be of particular concern when there are factors that can lead to slow drug solubilization, be it a function of a poorly soluble compound or a slow dissolving formulation. 3. GI fluid content: While there are any number of variables that can differ between the GI fluid content

in dogs and humans (see previous discussion), a focus in this study was gastric pH because of the zwitterion nature of ciprofloxacin. The potential for the precipitation of solubilized drug is also a challenge as the drug leaves the stomach and enters the duodenum. Particularly in dogs where there appears to be random surges in gastric pH, this would lead to the risk for even greater variability in oral drug absorption as compared to that in humans. As discussed by Papich and Martinez (4), there are drugs for which dogs exhibit higher oral bioavailability than humans, others where oral bioavailability is lower, and yet others where it is similar. So what makes the difference? That is an important question that still requires further research. To facilitate the prediction of oral drug product performance across species, it is important to distinguish situations where interspecies differences are attributable to presystemic drug metabolism, the activity of influx or efflux transporters, or where the fundamental issue involves such in vivo factors as membrane permeability (or absorptive surface area), GI residence time, or in vivo product dissolution and drug solubilization. Clearly, as we acquire more examples of research where human-canine differences are explored (across a range of drug physicochemical characteristics and, optimally, across a range of formulations), there will be an improvement in the ability to extrapolate data from dogs to humans and vice versa. Such extrapolations can be invaluable

Exploring Canine-Human Differences in Product Performance in supporting the development of oral drug products that support human and/or veterinary health. The hope is that by showcasing the use of in silico models for exploring potential reasons for observed human-canine differences in in vivo oral product performance, this investigation will encourage future use of a mechanistic approach for exploring interspecies differences in in vivo formulation behavior. As with the dataset from study 1, the full concentration versus time data generated in this dog study is available upon request. If you wish to obtain a copy of these data, please contact Marilyn Martinez ([email protected]).

13.

14.

15. 16. 17. 18.

REFERENCES

1.

Akimoto M, Nagahata N, Furuya A, Fukushima K, Higuchi S, Suwa T. Gastric pH profiles of beagle dogs and their use as an alternative to human testing. Eur J Pharm Biopharm. 2000;49:99–102. 2. Lui CY, Amidon GL, Berardi RR, Fleisher D, Youngberg C, Dressman JB. Comparison of gastrointestinal pH in dogs and humans: implications on the use of the beagle dog as a model for oral absorption in humans. J Pharm Sci. 1986;75:271–4. 3. Martinez MN, Papich MG. Factors influencing the gastric residence of dosage forms in dogs. J Pharm Sci. 2009;98:844–60. 4. Papich MG, Martinez MN. Applying Biopharmaceutical Classification System (BCS) criteria to predict oral absorption of drugs in dogs: challenges and pitfalls. AAPS J. 2015;17:948–64. 5. Martinez M, Amidon G, Clarke L, Jones WW, Mitra A, Riviere J. Applying the biopharmaceutics classification system to veterinary pharmaceutical products. Part II. Physiological considerations. Adv Drug Deliv Rev. 2002;54:825–50. 6. Sutton SC. Companion animal physiology and dosage form performance. Adv Drug Deliv Rev. 2004;56:1383–98. 7. Arndt M, Chokshi H, Tang K, Parrott NJ, Reppas C, Dressman JB. Dissolution media simulating the proximal canine gastrointestinal tract in the fasted state. Eur J Pharm Biopharm. 2013;84:633–41. 8. Zhang H, Xia B, Sheng J, Heimbach T, Lin TH, He H, et al. Application of physiologically based absorption modeling to formulation development of a low solubility, low permeability weak base: mechanistic investigation of food effect. AAPS PharmSciTech. 2014;15:400–6. 9. Jamei M, Turner D, Yang J, Neuhoff S, Polak S, RostamiHodjegan A, et al. Population-based mechanistic prediction of oral drug absorption. AAPS J. 2009;11:225–37. 10. Sjögren E, Abrahamsson B, Augustijns P, Becker D, Bolger MB, Brewster M, et al. In vivo methods for drug absorption—comparative physiologies, model selection, correlations with in vitro methods (IVIVC), and applications for formulation/API/excipient characterization including food effects. Eur J Pharm Sci. 2014;57:99–151. 11. Martinez M, Mistry B, Lukacova V, Polli J, Hoag S, Dowling T, et al. Use of modeling and simulation tools for understanding the impact of formulation on the absorption of a low solubility compound: ciprofloxacin. AAPS J. 2016;18:886–97. 12. USP29 monographs, ciprofloxacin www.pharmacopeia.cn/ v29240/usp29nf24s0_m17865.html.

19. 20. 21. 22.

23.

24. 25. 26.

27.

28.

29. 30. 31.

32.

Arakawa H, Shirasaka Y, Haga M, Nakanishi T, Tamai I. Active intestinal absorption of fluoroquinolone antibacterial agent ciprofloxacin by organic anion transporting polypeptide, Oatp1a5. Biopharm Drug Dispos. 2012;33:332–41. Haslam IS, Wright JA, O’Reilly DA, Sherlock DJ, Coleman T, Simmons NL. Intestinal ciprofloxacin efflux: the role of breast cancer resistance protein (ABCG2). Drug Metab Dispos. 2011;39:2321–8. Agoram B, Woltosz WS, Bolger MB. Predicting the impact of physiological and biochemical processes on oral drug bioavailability. Adv Drug Deliv Rev. 2001;50 Suppl 1:S41–67. Hendriksen BA, Felix MV, Bolger MB. The composite solubility versus pH profile and its role in intestinal absorption prediction. AAPS PharmSci. 2003;5(1):E4. Lu ATK, Frisella ME, Johnson KC. Dissolution modeling: factors affecting the dissolution rates of polydisperse powders. Pharm Res. 1993;10:1308–14. Mithani SD, Bakatselou V, TenHoor CN, Dressman JB. Estimation of the increase in solubility of drugs as a function of bile salt concentration. Pharm Res. 1996;13:163–7. Kararli TT. Comparison of the gastrointestinal anatomy, physiology, and biochemistry of humans and commonly used laboratory animals. Biopharm Drug Dispos. 1995;16:351–80. Mahar KM, Portelli S, Coatney R, Chen EP. Gastric pH and gastric residence time in fasted and fed conscious beagle dogs using the Bravo pH system. J Pharm Sci. 2012;101:2439–48. Papich MG. Ciprofloxacin pharmacokinetics and oral absorption of generic ciprofloxacin tablets in dogs. Am J Vet Res. 2012;73:1085–91. Abadía AR, Aramayona JJ, Muñoz MJ, Pla Delfina JM, Saez MP, Bregante MA. Disposition of ciprofloxacin following intravenous administration in dogs. J Vet Pharmacol Ther. 1994;17:384–8. Abadia AR, Aramayona JJ, Muñoz MJ, Pla Delfina JM, Bregante MA. Ciprofloxacin pharmacokinetics in dogs following oral administration. Zentralbl Veterinarmed A. 1995;42:505– 11. Vance-Bryan K, Guay DR, Rotschafer JC. Clinical pharmacokinetics of ciprofloxacin. Clin Pharmacokinet. 1990;19:434–61. Park MS, Okochi H, Benet LZ. Is ciprofloxacin a substrate of Pglycoprotein? Arch Drug Inf. 2011;4:1–9. Drozdzik M, Gröer C, Penski J, Lapczuk J, Ostrowski M, Lai Y, et al. Protein abundance of clinically relevant multidrug transporters along the entire length of the human intestine. Mol Pharm. 2014;11:3547–55. Gutmann H, Hruz P, Zimmermann C, Beglinger C, Drewe J. Distribution of breast cancer resistance protein (BCRP/ABCG2) mRNA expression along the human GI tract. Biochem Pharmacol. 2005;70:695–9. Haller S, Schuler F, Lazic SE, Bachir-Cherif D, Krämer SD, Parrott NJ, et al. Expression profiles of metabolic enzymes and drug transporters in the liver and along the intestine of beagle dogs. Drug Metab Dispos. 2012;40:1603–10. Cho SM, Park SW, Kim NH, Park JA, Yi H, Cho HJ, et al. Expression of intestinal transporter genes in beagle dogs. Exp Ther Med. 2013;5:308–14. Zhou H. Pharmacokinetic strategies in deciphering atypical drug absorption profiles. J Clin Pharmacol. 2003;43:211–27. Kuentz M, Nick S, Parrott N, Röthlisberger D. A strategy for preclinical formulation development using GastroPlus as pharmacokinetic simulation tool and a statistical screening design applied to a dog study. Eur J Pharm Sci. 2006;27:91–9. Kalantzi L, Persson E, Polentarutti B, Abrahamsson B, Goumas K, Dressman JB, et al. Canine intestinal contents vs. simulated media for the assessment of solubility of two weak bases in the human small intestinal contents. Pharm Res. 2006;23:1373–81.