Pulmonary Pharmacokinetics of Oseltamivir Carboxylate in Rats after Nebulization or Intravenous Administration of Its Prodrug, Oseltamivir Phosphate

The aim of this study was to investigate the pharmacokinetics of oseltamivir phosphate, a prodrug, and its active moiety in plasma and lung after its nebulization and intravenous administration in rats. Only 2% of prodrug was converted into active moiety presystematically, attesting to a low advantage of oseltamivir phosphate nebulization, suggesting that oseltamivir phosphate nebulization is not a good option to obtain a high exposure of the active moiety at the infection site within lung.

ABSTRACT

The aim of this study was to investigate the pharmacokinetics of oseltamivir phosphate, a prodrug, and its active moiety in plasma and lung after its nebulization and intravenous administration in rats. Only 2% of prodrug was converted into active moiety presystematically, attesting to a low advantage of oseltamivir phosphate nebulization, suggesting that oseltamivir phosphate nebulization is not a good option to obtain a high exposure of the active moiety at the infection site within lung.

TEXT

A recent series of studies conducted in our laboratory (1–6) has shown that membrane permeability is a key parameter to support drug nebulization (NEB) for local efficacy enhancement. From a biopharmaceutical standpoint, antimicrobial agents for parenteral administration, such as aminoglycosides, are much better candidates for NEB due to their limited membrane permeability (4) than those with high membrane permeability, allowing oral administration, such as fluoroquinolones (2). We have also shown that the anti-influenza drug oseltamivir carboxylate (OC) could be a good candidate for nebulization (1), and we hypothesize now that its orally administered prodrug, oseltamivir phosphate (OP), is not. The aim of this study was to confirm this hypothesis.

In vitro permeability experiments were performed in Calu-3 cell line with determination of the apparent permeability (Papp) (7). Briefly, OP and OC (Sigma, Saint Quentin Fallavier, France; AlsaChimand, Illkirch-Graffenstaden, France) solutions were used at 0.250 and 1 μg/ml, respectively, for both apical-basolateral (AP/BL) and basolateral-apical (BL/AP) studies. Samples (150 μl) were taken at various times in each compartment in function of the studied direction. The in vivo study was authorized by the ethics committee COMETHEA and registered by the French Ministry of Higher Education and Research (no. 2015070211159865). Forty male Sprague Dawley rats (300 to 330 g; Charles River, l’Arbresle, France) were divided in two groups (n = 20; intravenous [i.v.] or intratracheal [i.t.]). OP solution was administered at a dose of 6 mg/kg by i.v. bolus (1 ml) through the tail vein or by i.t. administration (100 μl) using a nebulizer (Microsprayer 1A-1B; PennCentury, Inc., Wyndmoor, PA) (8). Bronchoalveolar lavage (BAL) and blood samples were collected until 6 h postadministration (five rats per sampling time). Dichlorvos (4 mg/ml in 0.9% saline solution [5%, vol/vol]; Sigma) was added to samples. BAL and plasma samples were frozen at –80°C until analysis. OP and OC concentrations were determined by liquid chromatography-tandem mass spectrometry according to the method of Gupta et al. (9) using a system that included a Shimadzu HPLC module (Nexera XR; Shimadzu, Marne-la-Vallée, France) coupled with a API 3000 mass spectrometer (Sciex Life Sciences, Les Ulis, France). Standard curves between 0.25 and 200 ng/ml for OP and between 0.625 and 500 ng/ml for OC were determined in plasma and 0.9% saline solution. Quality controls were performed at 0.5, 25, and 150 ng/ml and 1.25, 62.5, and 375 ng/ml for OP and OC, respectively. To a 150-μl aliquot of spiked sample (plasma or saline solution), 750 μl of 1% formic acid in water, 40 μl of mixed internal standard in methanol-water (50:50 [vol/vol]), and 5% (vol/vol) dichlorvos (4 mg/ml in saline solution 0.9%) were added, and the samples were vortexed for 30 s. Samples and calibration points were loaded on SPE IST Evolute ABN 25-mg cartridges (Biotage, Hengoed, UK), washed with 1 ml of 1% formic acid in water, and eluted with 1 ml of methanol. Eluates were evaporated under nitrogen at 45°C and dissolved with 150 μl of a mixture of mobile phase-water (50:50 [vol/vol]) for injection (5 μl). Intra- and interday variabilities were evaluated, and the precision and accuracy were less than 15% for each level of control concentrations. Urea concentrations in plasma and BAL fluid were determined, and concentrations in epithelial lining fluid (ELF) were estimated using the urea dilution method (8). Plasma and ELF log-transformed data obtained after OP administration were simultaneously analyzed using nonlinear mixed effect modeling approach with S-ADAPT (v1.57) (10). To these data were added plasma and ELF concentrations previously obtained by our group after direct OC NEB and i.v. administrations (1). Only unbound drugs were assumed to exchange between plasma and ELF. Protein binding in plasma was fixed at 42% for OP and was considered negligible for OC (11). Protein binding was supposed to be negligible within ELF (12). In accordance with previous observations made using a Penn-Century microsprayer (2, 5), systemic bioavailability after NEB was fixed at 100%. The final model (see Fig. S1 in the supplemental material) corresponds to systemic pharmacokinetics of both compounds was characterized by a bicompartmental model with central (V_cOP and V_cOC) and peripheral compartments (V_pOP and V_pOC) linked together by two-way passive diffusion clearances (Q₁ and Q₂). OP was assumed to be renally excreted (CL_ROP) and converted into OC (CLconv_pl), and OC was only excreted in urine (CL_ROC). Pharmacokinetics in ELF was characterized by one compartment for OP (V_elf1OP) and by two compartments for OC (V_elf1OC and V_elf2OC) connected by a bidirectional clearance (Q₃). The distribution between plasma and ELF compartments was described by a passive two-way diffusion clearance for both molecules (Q_elfOP and Q_elfOC) to which an efflux clearance from plasma to ELF was added only for OP (CL_outOP). Enzymatic conversion of OP into OC within ELF was characterized by a conversion clearance (CLconv_elf). Other details of modeling are presented in the supplemental material. The unbound plasma area under the concentration-time curve (AUC_u,plasma,OP and AUC_u,plasma,OC) and the ELF AUC (AUC_ELF,OP and AUC_ELF,OC) from 0.5 h to infinity (∞) after either i.v. or NEB administration of OP and OC were calculated using the final model (Berkeley Madonna, v8.3.18; University of California). All extrapolated AUC values from t to ∞ were less than 10% of the AUC_0.5–∞.

The average absorptive (AP/BL) and secretory (BL/AP) Papp values for OP and OC are presented Fig. 1. Ratios between the mean secretory Papp (BL/AP) and the corresponding mean absorptive Papp (AP/BL) were estimated to be 4.4 and 0.9 for OP and OC, respectively. The pharmacokinetic profiles and parameter estimates for OC and OP after administration of the prodrug OP are presented in Fig. 2 and Table S1, respectively. After the OP i.v. bolus (Fig. 2A), the OP plasma concentrations decline rapidly due to both renal excretion of prodrug and a fast systemic conversion into OC, with the rapid apparition of OC in plasma at the first sampling time (t = 0.5 h) (Fig. 2B). After i.v. administration, OP ELF and plasma concentrations decline in parallel with a maximal concentration in ELF obtained at an early time (t = 0.5 h) after i.v. administration, attesting to the rapid diffusion of OP from plasma to ELF (Fig. 2A). Moreover, OP ELF concentrations were much higher than the unbound plasma concentrations (Fig. 2A), with a AUC_ELF/AUC_u,plasma ratio close to 90 (Table 1). After the i.v. administration of OP, the OC concentrations in ELF and plasma decrease in parallel (Fig. 2B), with an AUC_ELF/AUC_u,plasma ratio close to 5 (Table 1). The route of administration had no major effect on OP and OC ELF and plasma concentration-versus-time profiles (Fig. 2) attesting to the fast equilibrium after NEB. The AUC_ELF/AUC_u,plasma ratios for OP and OC after NEB were 90 and 10, respectively (Table 1). Interestingly, the OC ELF concentrations are higher than the plasma concentrations after OP administration regardless of the route of administration (the AUC_ELF/AUC_plasma = 5 after OP i.v. administration, and the AUC_ELF/AUC_plasma = 10 after OP NEB).

FIG 1.

Absorptive (apical to basolateral [AP/BL]) and secretory (basal to apical [BL/AP]) permeability (Papp) of OC (1 μg/ml) and OP (0.250 μg/ml) through Calu-3 cells (means ± the standard deviations, n = 3; *, P < 0.05, t test [Prism 5.02; GraphPad, La Jolla, CA]).

FIG 2.

Experimental (means ± the standard deviations) and predicted concentration-time profiles of OP (A and C, left panel) and OC (B and D, right panel) after OP i.v administration (A and B, top panel) or NEB (C and D, bottom panel). Closed and open symbols represent concentrations measured in plasma and ELF, respectively. Solid lines represent predicted total plasma concentrations, dotted lines represent unbound plasma concentrations, and dashed lines represent predicted ELF concentrations.

TABLE 1.

OP and OC AUC in plasma and ELF after NEB or i.v. administration of OP or OC^a

Treatmentb	AUC (μg/ml/h)
	AUCu,plasma,OP	AUCELF,OP	AUCu,plasma,OC	AUCELF,OC
i.v. OP	198	17,316	1,551	8,275
NEB OP	200	17,532	1,639	16,747
i.v. OC			2,810	5,626
NEB OC			2,842	3.77 × 106

^aAUC_u,plasma,OP values were calculated by multiplying the experimental data by the unbound fraction of OP (i.e., 0.58) (14).

^bFor i.v. OC and NEB OC, the AUC values were calculated by using the final model and data from Bedor et al. (1).

After i.v. administration, the OP AUC_ELF/AUC_u,plasma of 90 suggests that OP is a substrate of efflux transporters at the bronchoalveolar barrier, which is consistent with its 4.4-fold-higher secretory Papp (BL/AP) than its absorptive Papp (AP/BL) through Calu-3 cells (Fig. 1) and in agreement with previous studies presenting OP as a P-glycoprotein substrate (11, 12). Antimicrobial agents with low membrane permeability are much better candidates for NEB (4), but OP presents a relatively high absorptive permeability (Papp = 1.07 ± 0.11 × 10⁻⁶ cm/s) through the Calu-3 cell line, a level intermediate between those for norfloxacin (Papp = 0.6 ± 0.05 × 10⁻⁶ cm/s) and pefloxacin (Papp = 7.1 ± 0.28 × 10⁻⁶ cm/s) in the same model (13). Consequently, OP does not appear to be a good candidate for NEB, as confirmed by the results of the in vivo experiments showing that the route of administration for OP has no effect on both plasma and ELF concentrations versus time profiles (Fig. 2A and C). However, OP is a prodrug, and ELF exposure of the active moiety, OC, is actually of greater interest but also more complex to assess. OC ELF exposure after OP NEB should be as high as possible, which depends upon two major parameters: (i) the relative rates of OP conversion and absorption and (ii) the relative rates of OC formation and absorption (8). In the present study, the OP AUC_u,plasma was similar after NEB and i.v. administration (Table 1), indicating that virtually 100% of the OP nebulized dose was directly absorbed and therefore that the fraction converted presystematically into the active OC was extremely limited. This is due to the fact that OP systemic absorption is 45 times faster than its conversion into OC (Q_elfOP = 0.018 liters/h/kg versus CLconv_elf = 0.00042 liters/h/kg, Table S1), meaning that roughly only 2% of the OP nebulized dose is converted presystematically into OC. On the top of that, OC appeared in ELF rapidly disappears since its presystemic formation is three times slower than its absorption (CLconv_elf = 0.00042 liters/h/kg versus Q_elfOC = 0.0011 liters/h/kg; see Table S1 in the supplemental material). In a previous study with direct administrations of OC, ELF concentrations after NEB were 840 times higher than after i.v. administration (1), and OC exposure within ELF was 200 times higher after direct OC NEB than after OP NEB (Fig. S2). Noticeably, the relatively high contribution of OP efflux transport Cl_outOP (1.58 liters/h/kg) compared to its passive diffusion Q_elfOP (0.018 liters/h/kg) 90 times higher leads to much higher OP ELF than plasma concentrations (Fig. 1A) but has limited consequences on OC ELF concentrations (Fig. S2). The impacts of OP efflux transport and CLconv_elf on OC ELF concentrations are presented in the supplemental material.

In conclusion, this study suggests that OP NEB instead of i.v. administration is not a good option to increase exposure of the active OC at the infection site within lung but that direct NEB of OC remains the best strategy to reach this objective.