Carrier-Mediated Prodrug Uptake to Improve the Oral Bioavailability of Polar Drugs: An Application to an Oseltamivir Analogue

Abstract

The goal of this study was to improve the intestinal mucosal cell membrane permeability of the poorly absorbed guanidino analogue of a neuraminidase inhibitor, oseltamivir carboxylate (GOC) using a carrier mediated strategy. Valyl amino acid prodrug of GOC with isopropyl-methylenedioxy linker (GOC-ISP-Val) was evaluated as the potential substrate for intestinal oligopeptide transporter, hPEPT1 in Xenopus laevis oocytes heterologously expressing hPEPT1 and an intestinal mouse perfusion system. The diastereomers of GOC-ISP-Val were assessed for chemical and metabolic stability. Permeability of GOC-ISP-Val was determined in Caco-2 cells and mice. Diastereomer 2 was about two times more stable than diastereomer 1 in simulated intestinal fluid and rapidly hydrolyzed to the parent drug in cell homogenates. The prodrug had a nine times enhanced apparent permeability (P_app) in Caco-2 cells compared to the parent drug. Both diastereomer exhibited high effective permeability (P_eff ) in mice, 6.32±3.12 and 5.20±2.81 x10⁻⁵ cm/s for diastereomer 1 and 2, respectively. GOC-ISP-Val was found to be a substrate of hPEPT1. Overall, this study indicates that the prodrug, GOC-ISP-Val seems to be a promising oral anti-influenza agent that has sufficient stability at physiologically relevant pHs prior to absorption, significantly improved permeability via hPEPT1 and potentially rapid activation in the intestinal cells.

1. Introduction

Influenza, a common infection caused by influenza A and B viruses, results in three to five million cases of severe illness and 250,000 to 500,000 deaths worldwide annually.¹ The recent outbreaks of highly pathogenic avian influenza A/H5N1 virus about a decade ago and the new avian influenza A/H7N9 virus in March 2013 have raised concerns about the potential emergence of another pandemic after the first influenza pandemic of the 21^st century due to influenza A/H1N1 virus in the spring of 2009.^2–5 While vaccination is the primary strategy for the prevention and control of influenza, its effectiveness is limited by the antigenic mismatch between the vaccine and the circulating influenza virus strains, or the inability of the host to mount a sufficient and rapid immune response to the vaccine.⁶ Furthermore when a potential pandemic caused by a new strain of influenza virus occurs, it takes approximately five to six months for the first supplies of approved vaccine to become available.⁷ Thus, the availability of anti-influenza drugs that are effective against circulating influenza virus strains is extremely important for the prevention and treatment of influenza, particularly during the flu pandemic.

There are currently two types of anti-influenza drugs for the treatment and prophylaxis of influenza: M2 ion channel inhibitors and neuraminidase inhibitors. The use of M2 ion channel inhibitors (amantadine and rimantadine) are limited due to the prevalent drug resistant virus strains, side effects and lack of activity against influenza B viruses.⁶ The neuraminidase inhibitors are associated with a lower frequency of side effects and are effective against both influenza A and B viruses.⁶ Currently, there are two neuraminidase inhibitors in the US market: zanamivir and oseltamivir (Figure 1). Zanamivir has a very low oral bioavailability, averaging 2% and is dosed by oral inhalation.⁸ Oseltamivir is the ethyl ester prodrug of the poorly absorbed neuraminidase inhibitor, oseltamivir carboxylate. Due to rapid absorption and effective metabolism of the parent drug, the oral bioavailability of oseltamivir carboxylate is enhanced to about 80% by the prodrug (ethyl ester) approach.⁹ Currently, oseltamivir is the predominant drug used in the US to treat influenza and the only oral drug recommended by the Centers for Disease Control and Prevention (CDC).¹⁰ In addition to zanamivir and oseltamivir, other neuraminidase inhibitors under development include peramivir and laninamivir (Figure 1). Peramivir is an intravenously administered neuraminidase inhibitor approved in Japan and South Korea, and it is in Phase III clinical studies in the U.S.¹¹ During the 2009 H1N1 influenza pandemic, the US Food and Drug Administration (FDA) issued an Emergency Use Authorization (EUA) for intravenous (I.V.) peramivir to treat certain hospitalized patients with confirmed or suspected 2009 H1N1 influenza infection.¹¹ Laninamivir octanoate, a long-acting prodrug of neuraminidase inhibitor laninamivir, is approved for the treatment of influenza in Japan. Compared to the another inhalable anti-influenza agent zanamivir that requires twice a day, 5 day treatment, laninamivir octanoate needs only one single dose, via inhalation, to complete the treatment, which is much more acceptable to the patients.¹²

Figure 1.

Structures of potent neuraminidase inhibitors and their prodrugs. The starred carbon in the GOC-ISP-Val structure indicates a chiral center.

Although oseltamivir is effective for the prevention and treatment of influenza, resistance to oseltamivir drives the medical need for the development of new orally available anti-influenza drugs.¹³ On the other hand, resistance to the other three neuraminidase inhibitors (zanamivir, laninamivir and peramivir) have been rarely reported, however they are very poorly absorbed after oral administration. The low absorption is likely due to the highly polar and positively charged guanidino functionality^{14, 15} which importantly appears to make a significant contribution to the neuraminidase inhibition potency. Compared to the amino group in oseltamivir carboxylate, the guanidino functionality adds more hydrophilic character to the molecule making it less likely to be absorbed after oral administration. The poor oral absorption also prevents many guanidino-containing drug candidates from further development. The more potent guanidino analogue of oseltamivir carboxylate, GS 4116 (GOC), was abandoned due to poor oral bioavailability, and the prodrug approach that was applied to oseltamivir carboxylate (ethyl ester) did not significantly improve the oral bioavailability of GOC.¹⁶

Given the importance of the high binding affinity of guanidino group to the drug target, influenza neuraminidase and the generally poor oral bioavailability when a guanidino group is present in a drug structure, it would be advantageous to develop a strategy to increase the oral absorption of guanidino containing drugs and drug candidates.¹⁴ Various approaches have been adopted to enhance the lipophilicity of these promising anti-influenza agents, aiming to increase the passive diffusion and thus the oral absorption. A common approach is to make a carboxylic ester prodrug, which was successfully applied to oseltamivir. However, this approach was not successful with zanamivir and GOC, presumably due to the free polar guanidino functionality hindering the penetration of these prodrugs through the apical membrane of the epithelial cell.^{16, 17} In addition to the esterification of the carboxyl group, two separate research groups used counterions to shield the positive charge of guanidino functionality to further increase the drug lipophilicity and enhance the oral absorption. Both of the approaches demonstrated encouraging results for zanamivir, but not for GOC.^{18, 19} On the contrary, it has been recently demonstrated that the prodrug strategies based on the bioisosteric replacement of the guanidino group by an acetamidine (zanamivir amidoxime) and lowering the basicity of the guanidino and its bioisosteric amidine by N-hydroxylation fail to improve the oral bioavailability of zanamavir (F ≤ 3.7).¹⁵

Recently, a targeted carrier-mediated prodrug approach to increase the oral absorption of polar drugs²⁰ especially guanidino containing analogues have been developed.¹⁴ This strategy was based on the design of amino acid esters of guanidino containing molecules targeting the intestinal oligopeptide transporter, hPEPT1 and human valacyclovirase (hVACVase) to enhance oral absorption. To evaluate this targeted prodrug strategy, acyloxy ester prodrugs of GOC and zanamivir conjugated with amino acids have been synthesized.^{21, 22} The prodrugs, especially the valine esters exhibited high jejunal membrane permeability in the in situ rat perfusion study, indicating that this strategy is effective to significantly increase the intestinal uptake permeability of polar influenza neuraminidase inhibitors.^{21, 22}

The purpose of the present study was to investigate the stability, metabolism and transport of the valine GOC prodrug with the isopropyl-methylene-dioxy linker (GOC-ISP-Val) in Caco-2 cells and mice. The isopropyl-methylene-dioxy group has been used as linker for this prodrug strategy to increase the chemical stability of the prodrug prior absorption while maintaining the high epithelial cell permeability and rapid prodrug activation following its absorption. The prodrug GOC-ISP-Val has been evaluated for chemical and enzymatic stabilities, activation using mice and human VACVase, as well as hPEPT1-mediated uptake and transport in hPEPT1-expressing Xenopus laevis oocytes and mice, respectively. The hPEPT1-expressing oocytes has been previously demonstrated to be a suitable experimental system to investigate the role of PEPT1 in the transport of amino acid ester prodrugs such as valganciclovir.²³ Intestinal permeability of the prodrug has been also investigated across Caco-2 cell monolayers and in mice single-pass intestinal perfusion (SPIP) model.

2. Material and Methods

2.1. Materials

Diastereomers of prodrug GOC-ISP-Val were synthesized at TSRL, Inc. (Ann Arbor, MI). The ethyl ester of GOC and valacyclovir (VACV) were gifts from TSRL, Inc. (Ann Arbor, MI) and GlaxoSmithKline, Inc. (Research Triangle Park, NC), respectively. Potassium chloride, sodium chloride and HPLC and LC/MS grade acetonitrile, trifluoroacetic acid (TFA) and formic acid were obtained from Fisher Scientific Inc. (Pittsburgh, PA). Physiological saline solution was purchased from Hospira Inc. (Lake Forest, IL). Glycyl-l-proline (Gly-Pro), propranolol, metoprolol, phenol red, calcium chloride, magnesium chloride, sodium dihydrogen phosphate, potassium dihydrogen phosphate, sodium acetate, sodium hydroxide, D-glucose, 2-morpholinoethanesulfonic acid (MES), 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), pepsin, pancreatin and all other reagents and solvents were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO). Cell culture reagents were obtained from Gibco® Life Technologies Inc. (Carlsbad, CA), and cell culture supplies were from Corning Costar Co. (Corning, NY). All chemicals were either analytical or HPLC and LC/MS grade.

2.2. Methods

2.2.1. Cell Culture

Human epithelial colorectal adenocarcinoma (Caco-2) cells (passage 53–56 ) and human liver hepatocellular carcinoma (HepG2) cells (passage 91) from American Type Culture Collection (Rockville, MD) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS), 1% nonessential amino acids, 1 mM sodium pyruvate and 1% L-glutamine. Cells were grown in an atmosphere of 5% CO₂ and 90% relative humidity at 37°C.

2.2.2. Chemical Stability

The chemical stabilities of GOC and the diastereomers of prodrug GOC-ISP-Val were determined in pH 1.2 hydrochloric acid buffer, 50 mM sodium acetate buffer (pH 4.5 and 5.5), 50 mM MES buffer (pH 6.0), pH 6.8 simulated gastric fluid (SIF) and 10 mM potassium phosphate buffer (pH 7.4) at 37 °C. Stock solutions (200 mM in DMSO) of the test compound were diluted to a final concentration of 0.2 mM in the respective buffers and 100 μL aliquots were taken at 0., 5., 10., 30., 60. and 120. min and quenched with 100 μL of 1% (v/v) TFA in water. The samples were analyzed by HPLC.

2.2.3. Enzymatic Stability

2.2.3.1. Hydrolysis in Buffers Containing Pepsin and Pancreatin

Hydrolysis of GOC and the diastereomers of prodrug GOC-ISP-Val was determined in pH 1.2 simulated gastric fluid (SGF) with pepsin, pH 6.8 SIF with pancreatin. Hydrolysis of the prodrug was also carried out in the presence of pancreatin in 50 mM sodium acetate buffer (pH 4.5 and 5.5) and 50 mM MES buffer (pH 6.0). Hydrolysis experiments were performed as described for chemical stability. pH 1.2 SGF with pepsin and pH 6.8 SIF with pancreatin were prepared as described in United States Pharmacopeia 33- National Formulary 28.

2.2.3.2. Hydrolysis in Caco-2 and HepG2 Cell Homogenates

To prepare cellular homogenates, confluent cultures of Caco-2 or HepG2 cells were washed with 0.15 M sodium chloride solution and collected in 50 mM MES buffer (pH 6.0). The cell suspension was sonicated in ice bath and clarified by centrifugation at 9,700 g for 10 min at 4°C. The protein concentration of the homogenate was determined with the Bio-Rad DC protein assay (Hercules, CA) using bovine serum albumin (BSA) as a standard, and adjusted to 500 μg/mL. Hydrolysis of the diastereomers of GOC-ISP-Val in cell homogenates was determined at 37 °C. The hydrolysis reaction was initiated by adding the stock solution of the test compound to the cell homogenates, resulting in a final concentration of 0.2 mM. 100 μL of the reaction mixture was removed at 0., 5., 10., 30., 60. and 120. min and quenched with 100 μL of ice-cold TFA solution (10% (v/v)). The samples were centrifuged at 390 g for 10 min at 4 °C and filtered. The filtrates were analyzed for the prodrug and parent drug by HPLC.

2.2.3.3. Hydrolysis in Mice Intestinal Tissue Homogenates

Hydrolysis of GOC and diastereomers of the prodrug GOC-ISP-Val was determined in the intestinal tissue homogenates of wild-type and BPHL (biphenyl hydrolase-like protein, valacyclovirase) knock-out male mice (Charles River, IN). The intestinal segments taken from the wild-type and BPHL knock-out mice were rinsed with saline. Intestinal tissue homogenates were collected in pH 6.0 MES buffer and lysed by homogenizer (Tissue tearor, Model 985370, Biospec Products, Inc., Bartlesville, OK). The mixture was centrifuged at 9,700 g for 5 min. The amount of protein in supernatant was quantified with Bio-Rad DC Protein assay using bovine serum albumin (BSA) as a standard. Protein amount was adjusted to 500 μg/mL. Hydrolysis reaction was carried out with the initial concentration of 0.2 mM of GOC and GOC-ISP-Val at 37 °C in triplicate. 100 μL samples were collected at 0., 5., 10., 30., 60. and 120. min and quenched with 100 μL acetonitrile containing 0.1% (v/v) TFA. The samples were centrifuged at 9,700 g for 10 min at 4 °C. The supernatants were analyzed for prodrug and parent drug by HPLC. VACV was used as a standard substrate of VACVase.

2.2.3.4. Human Valacylovirase (hVACVase)-Mediated Hydrolysis

Potential activation mechanism of the prodrug was evaluated using hVACVase. Recombinant hVACVase was overexpressed and purified from Escherichia coli as previously described.²⁴ The purified hVACVase was concentrated and stored at −80 °C until use. The protein amount of the concentrated enzyme was determined with the Bio-Rad DC protein assay (Hercules, CA) using BSA as a standard. hVACVase-mediated hydrolysis was performed in 50 mM MES buffer (pH 6.0) at 37°C. The reaction buffer was initially warmed to 37°C for 5 min, and enzyme (final concentration of 700 ng/mL) and test compound (final concentration of 1 mM) were added to the buffer and incubated at 37°C. Aliquots were removed over a predetermined times and quenched with an equal volume of 10% (v/v) TFA in water. The samples were analyzed for prodrug and parent drug by HPLC. The specific activity of VACV was routinely monitored to normalize the active protein concentration.

2.2.4. Uptake Studies in Xenopus Laevis Oocytes

Several potential intestinal membrane transporters including organic cation/carnitine transporters, OCTN1 and OCTN2, the organic cation transporters, OCT1 and OCT3, and the organic anion transporting polypeptides, OATP2B1 in addition to the targeted hPEPT1 were evaluated for GOC-ISP-Val uptake in Xenopus laevis oocytes. Only hPEPT1 and OATP2B1 (data not shown) demonstrated enhanced uptake in the oocyte over expression system.

hPEPT1-mediated transport of the prodrug GOC-ISP-Val was further evaluated in Xenopus laevis oocytes heterologously expressing hPEPT1. Xenopus laevis oocytes were prepared and injected with in vitro-copied RNA (cRNA) of hPEPT1, and subjected to the uptake study for GOC-ISP-Val as described previously. ²⁵ In brief, uptake of GOC-ISP-Val was initiated by placing water-injected (controls) or cRNA-injected oocytes into modified Barth’s solution (MBS; 96 mM sodium chloride, 1 mM potassium chloride, 2.4 mM sodium bicarbonate, 0.82 mM magnesium sulphate, 0.33 mM calcium nitrate, 0.41 mM calcium chloride and 10 mM HEPES) containing the prodrug (50–5000 μM) for the indicated times at room temperature. At the end of the uptake study, the oocytes were washed with ice-cold MBS and homogenated with a sonicator (M8808; Wakenyaku Co., Ltd., Kyoto, Japan) in acetonitrile and 20 mM hydrochloride solution (4:6, v:v). The resultant supernatant was dried under a vacuum and reconstituted in 15% (v/v) acetonitrile and 85% (v/v) formic acid solution (0.1% (v/v)). The samples were subjected to LC-MS-MS analysis to quantify the intracellular amount of prodrug and parent drug. Intracellular accumulation of the prodrug was evaluated by measuring the intact form and parent drug, and expressed as the sum of the diastereomer of GOC-ISP-Val and GOC.

2.2.5. Intracellular Accumulation and Transepithelial Transport Across Caco-2 Cells

The apparent permeabilities of GOC, the prodrug GOC-ISP-Val and metoprolol were determined in Caco-2 cell monolayers. Caco-2 cells were seeded on collagen treated polytetrafluoroethylene membrane inserts with 0.4 μM pore size and 24 mm diameter (6-well Transwell plate, Corning Costar Co., Corning, NY). The cells on the inserts were cultured for 21 days at 37 °C in a humidified incubator containing 5% CO₂ in air. The differentiation status of the formed monolayer was evaluated by measuring the transepithelial electrical resistance (TEER) (Millicell- ERS epithelial Voltohmmeter, Millipore Co., Bedford, MA). Permeability studies were conducted with the monolayers that developed TEER values > 280 Ωcm² following 21 days in cell culture. 5 mM MES buffer (pH 6.0) and 5 mM HEPES buffer (pH 7.4 ) were used in the apical (AP) and basolateral (BL) side in Caco-2 permeability studies, respectively.

On the day of the experiment, DMEM was removed and the monolayer were rinsed and incubated with a blank transport buffer for 15 min. Following 15 min incubation, blank transport buffer was removed from the AP side and replaced by 1.5 mL of 0.2 mM test compound in pH 6.0 MES buffer, and 2.5 mL of uptake buffer (HEPES, pH 7.4) was added to the receiver compartment. The mixture of two diastereomers of prodrug GOC-ISP-Val was used in Caco-2 transport studies. Throughout the experiment, the transport plates were kept in an incubator at 37 °C. Samples (100 μL) were taken from the receiver side at 15., 30, 45., 60., 75., 90. and 120. min, and the volume withdrawn was replaced with fresh transport buffer. Samples were also taken from the donor side at 30., 60., 90. and 120. min to determine the levels of parent drug and the diastereomers of the prodrug in the apical side. The appearance rates of the test compounds on the receiver side were obtained under sink conditions. All samples were immediately acidified with 100 μL acetonitrile containing 0.1% (v/v) TFA. Drug and prodrug concentrations in the samples were immediately analyzed by HPLC.

Following the transport study, each Transwell membrane with cells was washed with ice-cold transport buffer (pH 6.0). The filter was then cut and placed in 0.5 mL of methanol:water (50:50, v:v) for 1 h incubation in order to lyse the cells. The cell suspension was collected after 1 h incubation and centrifuged at 9,700 g at 4 °C for 10 min. The supernatant was analyzed for prodrug and parent drug by HPLC. Protein amount was determined in lysed cells with the Bio-Rad (Hercules, CA) DC Protein assay using BSA as a standard. Drug accumulation into Caco-2 cells was expressed as nM drug or prodrug per mg protein.

2.2.6. Single-Pass Intestinal Perfusion (SPIP) Studies

All animal experiments were conducted using protocols approved by the University of Michigan Committee of Use and Care of Animals (UCUCA). Male BALB/c mice (Charles River, IN) weighing 20–25 g were used for all perfusion studies. The animals were housed and handled according to the University of Michigan Unit for Laboratory Animal Medicine guidelines. Prior to each experiment, the mice were fasted overnight (12–18 h) with free access to water. Animals were randomly assigned to the different experimental groups.

The procedure for the in situ SPIP was performed on BALB/c mice according to previously published report.²⁶ In brief, mice were anesthetized with an im injection of 5 mg/kg xylazine and 80 mg/kg ketamine. After they were placed on a heated surface maintained at 37 °C (Harvard Apparatus Inc., Holliston, MA), the abdomen was opened by a midline incision of 1.5 cm. The proximal jejunal segment of approximately 8 cm was exposed and glass cannulas (2.0 mm o.d.) were inserted at each end of the jejunal segment and cannulated with flexible PVC tubing (2.06 mm i.d., Fisher Scientific Inc., Pittsburgh, PA). Care was taken to avoid disturbing the circulatory system, and the exposed segment was kept moist with 37°C normal saline solution. All solutions were incubated in a 37°C water bath. The isolated jejunum segment was rinsed with normal saline solution in order to clean out any residual debris.

5 mM MES perfusion buffer (pH 6.0) was prepared with 1 mM calcium chloride, 0.5 mM magnesium chloride, 145 mM sodium chloride, 1 mM sodium dihydrogen phosphate, 3 mM potassium chloride and 5 mM D-glucose. At the start of the study, the perfusion solution containing the test prodrug (0.1 mM), the high permeability reference standard metoprolol (0.1 mM) and phenol red (5 μg/mL) was perfused through the intestinal segment using a peristaltic multichannel pump (Watson Marlow, Model 323S, Watson-Marlow Bredel Inc., Wilmington, MA) at a flow rate of 0.1 mL/min. Phenol red was used as a nonabsorbable marker for water flux measurements.

The perfusion solution containing all components was perfused for 0.5 h to ensure steady state conditions. After reaching steady state, perfusion samples were taken in 10 min intervals for 1.5 h. The experiments were repeated with coperfusion of the oligopeptide transporter inhibitor, Gly-Pro (10 mM) to evaluate the hPEPT1-mediated transport of prodrug GOC-ISP-Val. The concentrations of prodrugs, GOC and metoprolol in all samples including perfusion samples, original drug solution and inlet solution taken at the exit of the tubing were determined by LC-MS.

The concentration of phenol red in perfusion samples was determined by HPLC except in the experiments containing Gly-Pro, where the absorbance of phenol red was measured using an E_max Precision Microplate Reader (Molecular Devices Corporation, Sunnyvale, CA) at 450 nm. At the end of each perfusion experiment, the length of each perfused intestinal segment was accurately measured.

2.2.7. Analytical Methods

2.2.7.1. HPLC Analysis

Two different HPLC-UV systems were put to use for the analysis of the stability and cell culture samples. Either an Agilent HPLC system (Agilent Technologies, Santa Clara, CA) consisted of Agilent pumps (1100 series), an Agilent autosampler (1200 series) and an Agilent UV-Vis detector (1100 series) controlled by Chemstation® 32 software (Version B.01.03) or a Waters HPLC system (Waters Inc., Milford, MA) consisted of two Waters pumps (Model 515), a Waters autosampler (WISP model 712) and a Waters UV detector (996 Photodiode Array Detector) controlled by Waters Millennium® 32 software (Version 3.0.1). The flow rate was 1 mL/min at room temperature. All samples were resolved in an Agilent ZORBAX Eclipse XDB-C₁₈ column (3.5μm, 4.6×150 mm) equipped with a guard column. The mobile phase consisted of 0.1% (v/v) TFA in milli-Q water and 0.1% (v/v) TFA in acetonitrile with acetonitrile gradient changing from 2 to 90% over 11 min. The retention times for GOC, GOC-ISP-Val isomer 1 and isomer 2 were 7.3, 8.0 and 8.5 min, respectively. The detection wavelength was 254 nm. The injection volume was 50 μL. The retention time was 7.6 min, and detection wavelength was 275 nm for metoprolol. For the determination of phenol red in the SPIP studies, the samples were resolved in a Waters Xterra C₁₈ reverse-phase column (5 μm, 4.6 × 250 mm) equipped with a guard column. The mobile phase consisted of 0.03% (v/v) TFA in milli-Q water and 0.03% (v/v) TFA in acetonitrile with the organic phase gradient changing from 2 to 65% over 10 min. The retention time was 11.3 min, and the detection wavelength was 265 nm for phenol red.

2.2.7.2. LC-MS Analysis

The concentrations of metoprolol, GOC, GOC ethyl ester and the diastereomers of GOC-ISP-Val in the mice intestinal perfusion samples were determined using an LC-MS method. Perfusion samples were first treated with acetonitrile to precipitate any proteins and salts in the perfusate samples, then centrifuged at 22,000 g for 15 min. The supernatant was diluted by milli-Q water containing 0.1% (v/v) formic acid prior to analysis.

The LC-MS unit (Shimadzu Scientific Instrument, Kyoto, Japan) used for the analysis consisted of two pumps (LC-20AD), in-line vacuum degasser units (DGU-20A), autosampler (SIL 20AHT) and an LCMS 2010A detector. The system was controlled by LCMS Solution software (Version 3). The LCMS detector conditions for monitoring the prodrugs and drugs had the CDL temperature at 250°C and the heat block at 200°C. The detector voltage was maintained at 1.5 kV, with a nebulizing gas flow of 1.2 mL/min. Separation was performed on a Waters XTerra C₁₈ reverse-phase column (5 μm, 2.1 x 50 mm) in positive-electrospray ionization mode. The injection volume was 10 μL. Mobile phase A contained 0.1% (v/v) formic acid in milli-Q water and mobile phase B contained 0.1% (v/v) formic acid in acetonitrile under organic phase gradient changing from 2 to 90% over 7 min. The flow rate was 0.2 mL/min at room temperature. Propranolol was used as an internal standard. Quantification for metoprolol, propranolol, GOC, GOC ethyl ester and GOC-ISP-Val were performed at m/z 268.35, 260.35, 326.9, 355.4 and 498.4, respectively. Separate standard curves were used for each experiment.

2.2.7.3. LC/MS/MS Analysis

Samples from the Xenopus uptake studies were analyzed using an LC-MS-MS system, consisting of an LC-20AD (Shimadzu, Kyoto, Japan) HPLC linked to an API 3200 triple quad mass spectrometer (Applied Biosystems, Foster City, Japan). The mobile phase consisted of acetonitrile:water with 0.1% (v/v) formic acid (85:15, v:v). The flow rate was 0.2 mL/min. Data acquisition was achieved using multiple reaction monitoring in electrospray positive mode. Daughter ion transitions were 327.2 to 80.2 and 498.4 to 327.2 for GOC and GOC-ISP-Val, respectively.

2.2.8. Data Analysis

2.2.8.1. Stability Studies

The apparent first-order degradation rate constants of the test compounds were determined by plotting the natural logarithm of test compound remaining as a function of time. The slope of the plot is equal to the negative rate constant (k, min⁻¹). The degradation half-lives (t_1/2, min) were calculated by Eq. 1:

$${\mathrm{t}}_{1/2}=0.693/\mathrm{k}$$ Eq. (1)

2.2.8.2. Caco-2 Transport Study

The apparent permeability (P_app, cm/s) across Caco-2 cell monolayers was calculated from the linear plot of drug accumulated in the receiver side versus time using Eq. 2:

$$P_{\mathit{app}}=\frac{1}{C_{o}A}\times \frac{dQ}{dt}$$ Eq. (2)

where dQ/dt is the steady-state appearance rate of the test compound on the receiver side, C_o is the initial concentration of the test compound in the donor side, and A is the monolayer growth surface area (4.67 cm²). Linear regression was carried out to obtain steady-state appearance rate of the test compound on the receiver. The steady-state appearance rate was calculated using the sum of the amount of parent drug and the diastereomers of the prodrug in the receiver compartment.

2.2.8.3. Single-Pass Intestinal Perfusion (SPIP) Studies

The effective permeability (P_eff, cm/s) through the mouse gut wall in the SPIP studies was determined assuming the “plug flow” model expressed in Eq. 3.²⁷

$${\mathrm{P}}_{\text{eff}}(\text{cm}/\mathrm{s})=\frac{-\mathrm{Q}ln\frac{{{\mathrm{C}}^{\prime }}_{\text{out}}}{{{\mathrm{C}}^{\prime }}_{\text{in}}}}{2\mathrm{\pi }\text{RL}}$$ Eq. (3)

where Q is the perfusion buffer flow rate (0.1 mL/min), C′_out/ C′_in is the ratio of the outlet and inlet concentrations of the tested drug that has been adjusted for water transport, R and L are the radius (0.1 cm) and length of the intestinal segment, respectively.

The net water flux in the SPIP studies was determined from the perfusate concentrations of phenol red, a non-absorbed and non-metabolized marker. The measured C_out/C_in ratio was corrected for water transport according to Eq. 4.

$$\frac{{{\mathrm{C}}^{\prime }}_{\text{out}}}{{{\mathrm{C}}^{\prime }}_{\text{in}}}=\frac{{\mathrm{C}}_{\text{out}}}{{\mathrm{C}}_{\text{in}}}\times \frac{{\mathrm{C}}_{\text{in}\text{phenol}\text{red}}}{{\mathrm{C}}_{\text{out}\text{phenol}\text{red}}}$$ Eq. (4)

where C_{in phenol red} and C_{out phenol red} are equal to the concentrations of phenol red in the inlet and outlet samples, respectively. After perfusion of the prodrug, some portion of the prodrug was hydrolyzed or metabolized in the intestine. To account for this hydrolysis and metabolism, Eq. 5 was used:

$$\text{Prodrug}(\text{metabolized}/\text{hydrolyzed})=\text{Parent}{\text{drug}}_{\text{out}}-\text{Parent}{\text{drug}}_{\text{in}}$$ Eq. (5)

The corrected concentration for the prodrug (C_out-prodrug) was then calculated using Eq. 6:

$${\mathrm{C}}_{\text{out-prodrug}}=\text{Intact}{\text{Prodrug}}_{\text{out}}+\text{Prodrug}(\text{metabolized}/\text{hydrolyzed})$$ Eq. (6)

2.2.9. Statistical Analysis

All hydrolysis and cell culture experiments were performed in triplicate and SPIP experiments were performed in quadruplicate. The data are presented as mean ± standard deviation (SD). The data of uptake studies in Xenopus laevis oocytes are presented as mean ± standard error of mean (SEM), and the number of consecutive experiments performed was 8–13. The independent t-test was used to assess differences between means of two groups. Differences were considered statistically significant at p<0.05.

3. Results

3.1. Separation of Diastereomers

Due to the chiral center on the linker, the synthetic procedure provided a 1:1 mixture of two diastereomers of GOC-ISP-Val. The two diastereomers were separated by semi-preparative HPLC. GOC-ISP-Val isomer 1 had a shorter retention time than GOC-ISP-Val isomer 2. The separated diastereomers were used individually for the hydrolysis, oocyte uptake and SPIP studies, while the transport experiment in Caco-2 cell monolayers was performed using the mixture of diastereomers of GOC-ISP-Val.

3.2. Chemical and Enzymatic Hydrolysis

The estimated half-life (t_1/2) values of GOC and the diastereomers of prodrug GOC-ISP-Val in various buffers at pH 1.2 with/without pepsin, pH 4.5, 5.5, 6.0, 6.8 with/without pancreatin and pH 7.4, as well as in cell homogenates and mice intestinal tissue homogenates are presented in Table 1. For the diastereomers of the prodrug, the stability decreased with the increase in pH, which was consistent with the previous findings for many amino acid ester prodrugs.²² Both diastereomers were stable at pH 1.2, 4.5 and 5.5 and relatively stable at pH 6.0 and 6.8. At pH 7.4, the chemical hydrolysis increased dramatically, suggesting that the prodrug is chemically activated to the parent drug in the cell cytosol following absorption. Interestingly, GOC-ISP-Val isomer 2 was hydrolyzed about 2.0 times faster than isomer 1 at pH 6.0, 6.8 and 7.4, respectively, indicating that the stereochemistry around the acyloxy center influences the chemical hydrolysis rates. Both diastereomers were stable in pH 1.2 SGF with pepsin. The hydrolysis of GOC-ISP-Val isomer 1 was two and three times faster in pH 6.0 MES buffer and pH 6.8 SIF with pancreatin than in the corresponding buffers without pancreatin, respectively. Whereas pancreatin had no significant effect on the hydrolysis rates of GOC-ISP-Val isomer 2 at pH 6.0 and 6.8. There were no significant differences between the half-lives of GOC-ISP-Val isomer 2 at pH 6.0 with and without pancreatin (p = 0.707) and at pH 6.8 with and without pancreatin (p = 0.886). Both diastereomers were also stable in pH 4.5 and 5.5 acetate buffers with pancreatin.

Table 1.

Estimated half-lives of GOC and the diastereomers of GOC-ISP-Val determined in SGF with/without pepsin, SIF and buffers with/without pancreatin, as well as in Caco-2 and HepG2 cell homogenates, hVACVase and mice intestinal tissue homogenates.

	Half-life (t1/2, min)
Medium	GOC	GOC-ISP-Val isomer 1	GOC-ISP-Val isomer 2
pH 1.2 SGFa without pepsin	stableb	stableb	stableb
pH 1.2 SGFa with pepsin	stableb	stableb	stableb
pH 4.5 Acetate buffer	stableb	stableb	stableb
pH 4.5 Acetate buffer with pancreatin	stableb	stableb	stableb
pH 5.5 Acetate buffer	stableb	stableb	stableb
pH 5.5 Acetate buffer with pancreatin	stableb	stableb	stableb
pH 6.0 MES buffer	stableb	1100 ± 95	773 ± 176
pH 6.0 MES buffer with pancreatin	stableb	570 ± 184	698 ± 166
pH 6.8 SIFc without pancreatin	N/Ad	263 ± 34	136 ± 13
pH 6.8 SIFc with pancreatin	stableb	83.3 ± 4.1	138 ± 21
pH 7.4 Phosphate buffer	stableb	128 ± 5	65.8 ± 1.3
Caco-2 Cell homogenates (pH 6.0)	stableb	158 ± 1	17.9 ± 0.4
HepG2 Cell homogenates (pH 6.0)	stableb	223 ± 11	14.9 ± 0.4
hVACVasee (pH 6.0)	N/Ad	Not hydrolyzedf	Not hydrolyzedf
Wild-type mice intestinal tissue homogenatesg	stableb	144 ± 7	100 ± 4
Knock-out mice intestinal tissue homogenatesg	stableb	124 ± 13	105 ± 15

All values are expressed as mean ± SD; n=3;

^asimulated gastric fluid;

^bno decrease in area was detected by HPLC after 2 h;

^csimulated intestinal fluid;

^dnot available;

^ehuman valacyclovirase;

^fhydrolysis in hVACVase (700 ng/mL) was not significantly faster than that in buffer;

^gvalacyclovir was used as the standard substrate of valacylovirase to compare the enzyme activities in the intestinal tissue. Half-lives of valacyclovir were 95 and 187 min in the intestinal tissue homogenates of wild-type and BPHL knock-out mice, respectively.

Hydrolysis in Caco-2 and HepG2 cell homogenates and mice intestinal tissue homogenates were carried out at pH 6.0 to minimize the effect of chemical hydrolysis in buffer. The prodrug diastereomers were hydrolyzed 5–50 times more rapidly in the cell homogenates than in the comparable buffered solution, indicating likely enzyme-mediated hydrolysis in addition to the chemical hydrolysis at intracellular pH. The differences in the enzymatic hydrolysis between two diastereomers were profound. GOC-ISP-Val isomer 2 hydrolyzed nine and 15 times faster than GOC-ISP-Val isomer 1 in Caco-2 and HepG2 cell homogenates, respectively suggesting significant enzymatic hydrolysis. Enzymatic stability in wild-type mice intestinal tissue homogenates demonstrated that hydrolysis of GOC-ISP-Val isomer 2 was 1.4 times faster than GOC-ISP-Val isomer 1. Both diastereomers were hydrolyzed approximately eight times more rapidly in wild-type mice intestinal tissue homogenates than in the comparable buffered solution (pH 6.0). No significant differences were found for the half-lives of individual diastereomers between wild-type and BPHL knock-out mice intestinal tissue homogenates (p = 0.085 and 0.559 for GOC-ISP-Val isomer 1 and 2, respectively), indicating that GOC-ISP-Val is not the substrate of VACVase. To further examine whether the well-characterized hVACVase²⁴ is responsible for the enzymatic breakdown of the prodrug, the hydrolysis of both diastereomers was examined in the presence of purified hVACVase. However, no differences were observed in hydrolysis over buffer control, indicating that recombinant hVACVase is not responsible for the metabolism of the prodrug, GOC-ISP-Val.

3.3. hPEPT1-mediated Uptake by Xenopus Laevis Oocytes

The uptake of GOC-ISP-Val isomer 1 and 2 and parent drug GOC by Xenopus laevis oocytes injected with hPEPT cRNA at pH 6.0 for 30 min is presented in Figure 2A. The uptake of both diastereomers was approximately two times higher in hPEPT cRNA- injected oocytes expressing hPEPT1 than water-injected control oocytes (Figure 2A). In contrast, no increase in uptake was observed for the GOC parent drug. Since there was no significant difference in hPEPT1-mediated uptake of the diastereomers, the uptake of GOC-ISP-Val isomer 1 was further characterized. The uptake of GOC-ISP-Val isomer 1 increased in a time-dependent manner for the initial 3 h (Figure 2B). The saturation kinetics estimated a K_m (substrate concentration at which the reaction rate is half of the maximum rate (V_max)) value of 7.3 mM for the diastereomer 1 (Figure 2C). A non- metabolized hPEPT1 substrate, glycylsarcosine (Gly-Sar) significantly reduced hPEPT1- mediated uptake of the diastereomer 1 in a concentration dependent manner (Figure 2D).

Figure 2.

Uptake of prodrug GOC-ISP-Val by Xenopus laevis oocytes expressing hPEPT1. (A) Uptake of GOC-ISP-Val diastereomers and GOC (50 μM) in water-injected (open column) and hPEPT1 cRNA-injected oocytes (closed column) at pH 6.0 for 30 min at room temperature (B) Time course of uptake of GOC-ISP-Val isomer 1 (50 μM) in water-injected (open circle) and hPEPT1 cRNA-injected oocytes (closed circle) at pH 6.0 for up to 3 h at room temperature. The inset shows the hPEPT1-mediated uptake of GOC-ISP-Val isomer 1 (gray circle) expressed as the difference between the uptakes of the prodrug by hPEPT1-injected and water-injected oocytes. (C) Concentration-dependent uptake of GOC-ISP-Val isomer 1 (50–5000 μM) in hPEPT1 cRNA-injected oocytes at pH 6.0 for 3 h at room temperature. A maximum concentration of the prodrug was set at 5000 μM for the concentration-dependent uptake study due to limited availability of the prodrug. (D) Competitive inhibition assay performed in the presence of 1 and 20 mM Gly-Sar. Effect of Gly-Sar is expressed as the percentage of control (hPEPT1-mediated uptake without Gly-Sar) at room temperature (pH 6.0). All data are presented as the mean ± SEM, n = 8–13 in each experimental group; * Differences between the uptakes of prodrug by water-injected and hPEPT1 cRNA-injected oocytes are considered statistically significant at p < 0.05.

3.4. Intracellular Accumulation and Transepithelial Transport Across Caco-2 Cell

The result of the transport study for GOC and the prodrug GOC-ISP-Val across Caco-2 cell monolayers in the absorptive (AP-BL) direction is presented in Figure 3A. The parent drug, GOC demonstrated negligible apparent permeability (P_app) ( 7.98±4.61 x 10⁻⁷ cm/s), while the prodrug demonstrated a nine times increase in permeability. However P_app of metoprolol was 1.7 times higher than that of the prodrug. The stability of the prodrug was determined in both donor and receiver sides. The amounts of the parent drug GOC and diastereomers of GOC-ISP-Val over time in the donor and receiver sides are presented in Figure 3B and 3C, respectively. The hydrolysis of the prodrug was below 10% in the donor side during 2 h transport period. The amounts of the parent drug and diastereomers of GOC-ISP-Val in the basolateral side were gradually increased with time up to 90 min. Equal amounts of GOC, GOC-ISP-Val isomer 1 and 2 were present in the receiver side over time. The conversion rate of GOC-ISP-Val to the parent drug, GOC was found to be 0.31±0.02 on the receiver side. Permeability was determined by adding the amount of parent drug converted from the prodrug to the sum of intact diastereomers in the receiver side.

Figure 3.

Transport of GOC-ISP-Val across Caco-2 cell monolayers. (A) Apparent permeability (P_app, cm/s) of GOC, GOC-ISP-Val (0.2 mM) and metoprolol (0.2 mM) in the absorptive (AP- BL) direction across Caco-2 monolayers. (B) Apical concentrations of GOC and the diastereomers of GOC-ISP-Val versus time in Caco-2 transport study. (C) Cumulative amount of GOC and the diastereomers of GOC-ISP-Val versus time in the basolateral compartment in Caco-2 transport study. Data are presented as the mean ± SD, (n = 3)

Drug/prodrug intracellular accumulation into Caco-2 cells following the transport study is presented in Figure 4. Prodrug intracellular accumulation into Caco-2 cell was four times higher than that of the parent drug. The amounts of GOC, GOC-ISP-Val isomer 1 and isomer 2 were 49.2, 28.5 and 22.3% in cells, respectively. The data indicated that approximately half of the prodrug (49±12%) is hydrolyzed to the parent drug by intracellular hydrolysis following the apical membrane transport.

Figure 4.

Comparison of GOC and the prodrug GOC-ISP-Val accumulation into Caco-2 cells after 2 h transport study. The accumulated drug/prodrug amount in cells was measured as nmol of GOC and the diastereomers of GOC-ISP-Val per mg of protein in the cell lysates collected from the membrane at the end of the transport study. Data are presented as the mean ± SD, (n = 3)

3.5. In Situ Permeability in Mice

Effective permeability (P_eff) values of GOC, GOC ethyl ester and GOC-ISP-Val diastereomers from the in situ mouse SPIP in proximal jejunum are presented in Figure 5. P_eff was measured as a function of disappearance of compounds after corrections for water flux, hydrolysis and possible metabolism. GOC and the simple ethyl ester of GOC exhibited approximately zero permeability in the jejunum of mouse, which was consistent with the known very low absorption.²¹ However, the P_eff values of both GOC-ISP-Val diastereomers (P_eff = 6.32±3.12 x 10⁻⁵ and 5.20 ±2.81 x10⁻⁵ cm/s for GOC-ISP-Val isomer 1 and 2, respectively) were significantly higher than the P_eff values of GOC and its ethyl ester prodrug. In fact, there were no significant differences between the P_eff values of GOC-ISP-Val diastereomers and metoprolol, an FDA standard for the low/high P_eff class boundary (P_eff = 5.38±1.68 x 10⁻⁵ cm/s for metoprolol and p = 0.620 and 0.810 for GOC-ISP-Val isomer 1 and 2 versus metoprolol, respectively). Also, there was no significant difference between the P_eff values of two diastereomers (p = 0.612). The fractions of prodrug metabolized following SPIP in mice were 4.9±3.2 and 3.6±2.2% for GOC- ISP-Val isomer 1 and 2, respectively. Significantly, when the GOC-ISP-Val prodrugs were co- perfused with the oligopeptide transporter inhibitor, Gly-Pro (10 mM), the P_eff values of GOC- ISP-Val isomers 1 and 2 were reduced by 8.5 and 6.4 times, respectively (Figure 5), strongly suggesting hPEPT1-mediated transport. There was no significant difference between the P_eff values of two diastereomers (p = 0.780) when coperfused with Gly-Pro.

Figure 5.

Effective permeability (P_eff; cm/s) values obtained for the diastereomers of GOC-ISP-Val (0.1 mM) in the presence or the absence of the oligopeptide transporter inhibitor, Gly-Pro (10 mM), as well as for GOC (0.1 mM) and GOC ethyl ester (0.1 mM) in comparison to metoprolol (0.1 mM) following in situ single-pass intestinal perfusion (SPIP) to the mouse jejunum at pH 6.0. Data are presented as the mean ± SD; n = 24 for metoprolol and n = 4 for the others.

4. Discussion

Since the most common route of drug delivery is the oral route, the prodrug strategy to achieve the optimal physicochemical properties to allow high transcellular absorption following oral administration of therapeutic agents is a challenge in modern drug development and prodrug innovations. Ideally, a prodrug should be chemically stable prior to absorption and be quickly metabolized (activated) to systemically active parent drug in the intestinal mucosal cell or hepatic cells following the absorption in order to exert its maximum therapeutic effect.

Previously, the chemical and enzymatic stabilities of methoxy, ethoxy and propylene glycol linker containing amino acid prodrugs were evaluated to find a suitable linker for prodrugs with amino acids.²⁸ It was found that the propylene glycol linker is an optimal linker for amino acid prodrugs, since it has good chemical stability and is enzymatically hydrolyzed to parent drug.²⁸ The acyloxy (alkyl) ester based amino acid linked prodrugs of polar neuraminidase inhibitors, GOC and zanamavir were further evaluated to target the intestinal membrane transporter, hPEPT1.^{21, 22} The corresponding valyl amino acid prodrugs of GOC (GOC-L-Val) and zanamivir (Zan-L-Val) presented significantly higher intestinal permeabilities than their parent drugs.^{21, 22} Although the methyl-methylene-dioxy linker for these amino acid ester prodrugs are relatively stable at physiologically relevant intestinal pHs, the half-lives of 2.5 and 3.5 h for valyl prodrugs of zanamivir and GOC at pH 6.0 suggested that a significant amount of these prodrugs may be hydrolyzed prior to recognition by the intestinal transporters.^{21, 22}

In the present study, the more bulky isopropyl-methylene-dioxy linker was used to make the valyl amino acid prodrug of GOC (GOC-ISP-Val) to achieve higher chemical stability compared to the previously evaluated prodrugs with methyl-methylene-dioxy linker. Both diastereomers of GOC- ISP-Val were highly stable at low pHs and exhibited significantly longer half-lives (t_1/2 = 1100 and 773 min for diastereomer 1 and 2, respectively, at pH = 6.0) than the previously developed GOC prodrug with methyl-methylene-dioxy linker (GOC-L-Val, t_1/2 = 212 min at pH 6.0)²¹ resulting in sufficient stability in the gastrointestinal lumen to exhibit potentially good oral absorption. In pH 7.4 buffer and Caco-2 and HepG2 cell homogenates, as well as in mice intestinal tissue homogenates, the diastereomers of GOC-ISP-Val were efficiently hydrolyzed to the parent drug GOC, indicating its potential for rapid activation following mucosal cell absorption in vivo. The differences between the chemical hydrolysis rates of the two diastereomers at high pHs may likely be explained by the steric hindrance. Although the chemical structure of prodrug GOC-ISP-Val is relatively flexible, the bulky isopropyl group may allow one of the diastereomers to be sterically more favorable for either acid or base catalyzed hydrolysis (general or specific). Furthermore, the chemical instability with increasing pH suggests the possible involvement of the amino group from the valine in an intramolecular catalyzed cleavage at the acyloxy center. It can be also influenced by the stereochemistry around the acycloxy center carbon. The selectivity between the two diastereomers was even more pronounced in both Caco-2 and HepG2 cell homogenate hydrolysis. Diastereomer 2 was approximately two times much more stable than diastereomer 1 in pH 6.8 simulated intestinal fluid with pancreatin, indicating that diastereomer 2 is more likely to remain relatively stable at the upper and mid small intestine and may release the parent drug in the intracellular compartment. Therefore, the better stability in the intestine and the more rapid in vivo activation suggest that diastreomer 2 may be a better candidate for oral administration compared to diastereomer 1. Since the clinical potency of the diastereomers is significantly dependent on the intestinal stability and in vivo activation of the diastereomers to the parent drug GOC, the more potent analogue of oseltamivir carboxylate. Furthermore, it is clear that the key determinant for the choice of appropriate isomers of drugs/prodrugs for clinical use is contingent on their relative potencies. Hydrolysis studies in wild-type and BPHL knock-out mice intestinal tissues demonstrated that GOC-ISP-Val is not a substrate of VACVase in mice. In separate studies, neither GOC-ISP-Val isomer 1 nor isomer 2 were found to be the substrate of previously identified amino acid ester prodrug activating enzyme, hVACVase, presumably due to the bulky isopropyl group on the prodrug side chain hindering the binding to hVACVase enzyme.^{24, 29} Since GOC-ISP-Val is not a substrate of hVACVase, another stereoselective enzyme(s) must be responsible for its intestinal and hepatic activation.

The uptake study in Xenopus oocytes as a heterologous expression system demonstrated that GOC-ISP-Val is recognized by the oligopeptide transporter hPEPT-1, indicating the improved intestinal permeability of the prodrug is, at least in part, mediated by hPEPT1. The increase in hPEPT1-mediated uptake of each diastereomer was similar to the previous result for the prodrug VACV, which is transported via hPEPT1 in the intestine.³⁰ Since the estimated affinity, K_m of diastereomer 1 was found to be 7.3 mM which was very close to that of VACV (5.9 mM).³⁰ Gly-Sar inhibition of the hPEPT1 transport confirmed that the uptake of the prodrug is mediated by hPEPT1. In addition to being a substrate of hPEPT1, the accumulation of the diastereomers of GOC-ISP-Val into water-injected oocytes was several times higher than the parent drug, suggesting that hydrophobicity of the prodrug could contribute to its improved bioavailability as well. Furthermore, there may be some possible contributions of other transporters such as the organic anion transporting polypeptide, OATP2B1 in the intestinal absorption of GOC-ISP-Val required to be clarified.

The transport of GOC-ISP-Val was further examined using Caco-2 cell culture and SPIP in mice. The prodrug demonstrated significantly enhanced uptake and permeability compared to the parent drug in Caco-2 cells. However, Caco-2 permeability of the prodrug was approximately two times less than that of the high permeability reference standard, metoprolol. The transport activity of peptide like drugs by hPEPT1 is reported to be lower in Caco-2 cells than in the small intestinal epithelial cells.^{31, 32} Despite the low transepithelial permeability across Caco-2 cells, both diastereomers of GOC-ISP-Val exhibited high intestinal permeability in mice, comparable to that of metoprolol. The result of the inhibition study with Gly-Pro in mice intestinal perfusion was consistent with the uptake study in oocytes. The oligopeptide transporter inhibitor, Gly-Pro significantly reduced the intestinal permeability of both diastereomer, suggesting that hPEPT1 is responsible for the dramatic enhancement of intestinal permeability of this prodrug.

In conclusion, the present study clearly demonstrated that the carrier-mediated prodrug approach can be successfully applied to very polar GOC. The stability and intestinal mucosal cell uptake of this polar and poorly absorbed neuraminidase inhibitor were optimized, thus increasing the potential for human oral administration.