Mechanism of Brain Targeting by Dexibuprofen Prodrugs Modified with Ethanolamine-Related Structures

Abstract

The first molecular insights into how prodrugs modified with ethanolamine-related structures target the brain were generated using an in vitro BBB model and in situ perfusion technique. Prodrugs were delivered safely and efficiently to the brain through tight interaction with the anionic membrane of brain capillary endothelial cells, observed as a shift in zeta potential, followed by uptake into the cells. Prodrugs III and IV carrying primary and secondary amine modifications appeared to enter the brain via energy-independent passive diffusion. In contrast, besides the passive diffusion, prodrugs I and II carrying tertiary amine modifications also appeared to enter via an active process that was energy and pH dependent but was independent of sodium or membrane potential. This active process involved, at least in part, the pyrilamine-sensitive H⁺/OC antiporter, for which the N,N-diethyl-based compound II showed a much lower affinity than the N,N-dimethyl-based compound I, likely due to steric hindrance. These new insights into brain-targeting mechanisms may help guide efforts to design new prodrugs.

INTRODUCTION

The blood–brain barrier (BBB) poses the greatest challenge for therapeutic agents meant to act in the brain.^1,2 Most drugs cannot enter the brain in pharmacologically relevant amounts, leading investigators to abandon efforts to target them specifically to this organ. ³ One strategy to improve poor drug uptake in the brain is to administer compounds as prodrugs that can be shuttled by transporters specifically expressed at the BBB. ⁴ Using this approach, researchers have succeeded in designing drug-carrier conjugates that can specifically interact with glucose transporters (GluT1), ⁵ amino acid transporters (LAT1), ⁶ and choline transporters, ⁷ thereby targeting the drugs to the brain. For cationic drugs, however, the transporters that may move them across BBB remain unknown.^8,9 Elucidating the mechanism of cationic drugs transport across the barrier may help guide the development of novel brain-targeting prodrugs.

Six transporters that can shuttle cationic drugs across the cell membrane have already been identified in the kidney and liver: OCT1 (SLC22A1), OCT2 (SLC22A2), OCT3 (SLC22A3), OCTN1 (SLC22A4), OCTN2 (SLC22A5), and MATE1 (SLC47A1).^10,11 Whether these transporters are expressed and functional at the BBB remains unknown. Several studies suggest that an antiporter system capable of transporting of organic cations may have a useful role in carrying cationic drugs passed the barrier. Yamazaki et al^12,13 showed that the cationic H1-antagonist pyrilamine is transferred from blood to brain by capillary endothelial cells (BCECs) in a carrier-mediated process. Subsequently, the same transport pathway was found to carry into the brain, other lipophilic basic drugs, such as imipramine, diphenhydramine, propranolol, pramipexole, ¹⁴ and oxycodone. ⁸ In fact, these drugs competed with pyrilamine in this transport process, suggesting the existence of putative pyrilamine cationic transporters in the brain. This process was found to be energy and pH dependent but independent of sodium and membrane potential.

Evidence that these putative pyrilamine cationic transporters may participate in an antiporter system came when the transport of secondary or tertiary amine drugs (e.g., nicotine, ¹⁵ clonidine, ¹⁶ and tramadol ¹⁷ ), which shows similar characteristics as pyrilamine transport, was found to be driven by an adverse proton gradient. Pyrilamine inhibited the tranport process, suggesting the existence of a pyrilamine-senstive H⁺/OC-antiporter. This antiporter is the best current candidate for a transporter capable of moving cationic drugs across the BBB. Little is known about how the transporter moves drugs across the barrier, whether it is the only mechanism that can do so and what are the chemical or steric requirements for effective drug transport.

Previous work from our laboratory has shown that prodrugs modified with ethanolamine-related structures (prodrugs I, II, III, and IV, Figure 1) can target the brain effectively. ¹⁸ They showed brain-to-plasma concentration ratios that were 9.31- to 17.0-fold higher than that of the drug dexibuprofen on which they were based. Subsequently, we found that N,N-dimethyl amino as a modification group significantly enhanced the brain-uptake efficiency of naproxen, ¹⁹ flurbiprofen, ²⁰ and dopamine. ²¹ These prodrugs are lipophilic and positively charged at physiologic pH, raising the possibility that they are carried into the brain by the pyrilamine-senstive H⁺/OC-antiporter, as well as lipid-mediated simple diffusion. Therefore, the present study analyzed the mechanisms of brain targeting of prodrugs I, II, III, and IV and the role of the H⁺/OC-antiporter. The structural and chemical diversity of the four prodrugs allowed us to probe structure–function relationships of the prodrugs and substrate selectivity of the antiporter system. These results may help guide the design of brain-targeting prodrugs, particularly ones with ethanolamine-related modifications.

Figure 1.

Chemical structures of dexibuprofen and prodrugs I, II, III, and IV.

MATERIALS AND METHODS

Materials

Dexibuprofen (99.8% purity) was purchased from the Fourth Pharmaceutical Company of Suzhou (Suzhou, China). Prodrugs I, II, III, and IV were synthesized and identified according to our previous work. ¹⁸ 1-(4-(Trimethylamino) phenyl)-6-phenyl-1,3,5-hexatriene (TMA-DPH), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), trypsin, N-methylglucamine, valinomycin, pyrilamine, propranolol, imipramine, hemicholinium-3, choline, L-carnitine, tetraethylammonium chloride (TEA), L-lysine, L-arginine, spermine, and rotenone were purchased from Sigma-Aldrich (St Louis, MO, USA). All other chemicals were of analytical or high performance liquid chromatography grade.

Animals and Cell Culture

All animal experiments were approved by the Institutional Animal Care and Ethic Committee of Sichuan University (Approved No. SYXK2013-185) and performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals. All animals were fed on a light and dark cycle and allowed free access to standard chow and water. Temperature and relative humidity were kept at 25° and 50%, respectively. After experiment, rats were killed by neck dislocation under anesthesia, and all efforts were made to minimize suffering. The report was written in compliance with ARRIVE guidelines. ²²

L929 cells (passages: 3 to 20) that were purchased from the American Type Culture Collection (Rockville, MD, USA) and cultured in DMEM with high glucose (Hyclone, Logan, UT, USA) supplemented with 10% FBS (Hyclone), 100IU/mL penicillin, and 100 μg/mL streptomycin. Cells were maintained at 37° in a humidified atmosphere containing 5% CO₂, and the culture medium was changed every other day.

In Vitro Co-Culture Model of the Blood–Brain Barrier

Blood to brain by capillary endothelial cells and astrocytes (ACs) were isolated from rat brain cortex and used to establish primary cocultures as described in ‘Supplementary Information’. Co-cultures were set up using the ‘contact through feet’ approach, in which endothelial cells are cultured on the upper face of transwell inserts and ACs are seeded on the lower face. In this way, ACs can spread their processes through the membrane pores to contact endothelial cells. Cell culture inserts with a microporous polyethylene terephtalate membrane (1 μm diameter, Millipore, Darmstadt, Germany) were turned upside down, and ACs were seeded on top at a density of 10 ⁴ cells per insert. After 4 hours, the insert was turned over (right side-up) and placed in 24-well plates. Astrocyte medium was added and replaced every other day. After 5 days, BCECs were seeded onto the upper side of the insert at a concentration of 5 × 10 ⁴ cells per insert, BCEC culture medium was added and replaced every other day. After an additional 5 days, transendothelial electrical resisitance (TEER) of the modeled BBB was measured using a TEER instrument (Millipore). When TEER reached 250 Ω cm², the BBB model was considered ready for experiments. The morphologic characteristics and barrier integrity were analyzed by transmission electron microscopy (JEM 100CX, JEOL, Tokyo, Japan) and scanning electron microscopy (S-450, Hitachi, Japan).

Prodrug Effects on Blood–Brain Barrier Integrity In Vitro Prodrug cytotoxicity in blood to brain by capillary endothelial cells. Cell viability was assessed using MTT method. Briefly, BCECs were seeded in 96-well culture plates at a density of 4×10 ⁴ cells/well and incubated for 48 hours at 37°. Then, the cells were treated for 4 hours with different concentrations of dexibuprofen (DIBU) or prodrugs. Next, 20μL MTT was added to the wells at a final concentration of 0.5 mg/mL, and the cells were incubated another 4 hours at 37°. Insoluble formazan was dissolved in DMSO, and absorbance was measured at 570 nm in a plate reader (Varioskan Flash, Waltham, MA, USA).

Prodrug Effects on Transendothelial Electrical Resisitance. The effects of the prodrugs on the TEER values of in vitro BBB model were measured by a Millicell electrical resistance system (Millipore). After 4 hours permeability test of prodrugs across the BBB as described in the ‘Prodrug transendothelial transport in an in vitro BBB model’, the TEER values of the cell monolayers with/without the presence of prodrugs were recorded and expressed in Ω cm².

Prodrug Effects on F-Actin Staining of Blood to Brain by Capillary Endothelial Cells. After 5 days seeded, BCEC monolayers were incubated with DIBU or prodrugs I, II, III, or IV at a final concentration of 80 μmol/L. At predetermined time points, cells was rinsed with cold phosphate-buffered saline (PBS, pH 7.4), fixed with 4% paraformaldehyde, stained with FITC-phalloidin (Sigma-Aldrich) and visualized by confocal laser scanning microscopy (Leica, Wetzlar, Germany).

Prodrug Effects on Surface Membrane Potential of Blood to Brain by Capillary Endothelial Cells

Confluent BCEC monolayers were washed three times with PBS, trypsinized, and diluted with HEPES buffer to a final concentration of 4×10 ⁶ cells/mL. Then, TMA-DPH, IBU, or prodrugs were added to cell suspensions to final concentrations of 54 μmol/L (TMA-DPH) or 80 μmol/L (DIBU and prodrugs) and incubated for 30 minutes. Surface membrane potential (zeta-potential, Z) was analyzed as described ²³ using a Zetasizer Nano ZS (Malvern, UK) equipped with a He-Ne laser (λ = 632.8 nm). Measurements were taken at 25° using disposable zeta cells with platinum gold-coated electrodes. ²⁴

Prodrug Uptake by Blood to Brain by Capillary Endothelial Cells Blood to brain by capillary endothelial cells were seeded in 6-well culture plates (5×10 ⁵ cells/well), incubated for 48 hours and checked under the microscope for confluence and morphology. Then BCECs were washed twice with 1 mL of PBS and preincubated with HBSS buffer for 20 minutes at 37°. Thereafter, buffer containing different concentrations of DIBU or prodrug was added to initiate uptake. Cells were incubated at 37° for a predetermined time period, after which uptake was terminated by washing three times with 1 mL of ice-cold PBS. Cultures were trypsinized and centrifuged at 1,000 g for 5 minutes; pellets were washed twice with PBS, then resuspended in ultrapure water and lysed by repetitive freeze-thawing to release internalized prodrug.

Sample containing prodrugs was subjected to hydrolysis. Lysates containing prodrugs (100 μL) were mixed with 20 μL 1N NaOH and allowed to stand for 10 minutes at 25°. Acetonitrile (200 μL) was added and mixtures were vortexed for 5 minutes, then centrifuged at 14,000 g for 10 minutes; supernatants (1 μL) were analyzed by high-performance liquid chromatography (Waters, Tokyo, Japan). ¹⁸ Uptake was expressed as nmol of DIBU per mg of total cellular protein, which was assayed in parallel wells using the BCA kit (Pierce, Rockford, IL, USA).

This uptake procedure was performed at both 4° and 37° to examine whether passive diffusion is a driving force for prodrug uptake. The procedure was also performed after 30-minute pretreatment with metabolic energy inhibitors (0.1% NaN₃ or 25 μmol/L rotenone) to examine the energy requirements of prodrug transport.

To begin to examine whether the mechanisms of prodrug uptake depend on cell line, we incubated cultures of BCECs and L929 cells for 1 hour at 37° with DIBU or prodrug (80 μmol/L). Cells were washed three times with ice-cold PBS and assayed for internalized prodrug as described above.

Prodrug Uptake Mechanisms in Blood to Brain by Capillary Endothelial Cells

To examine the mechanisms of prodrug uptake in the assay described above, cells were preincubated in the presence of different additives. To test the sodium ion dependency of uptake, cells were preincubated in medium containing N-methylglucamine or potassium chloride instead of sodium choride. To test the effect of reducing membrane potential, cells were preincubated for 10 minutes with the potassium ionophore valinomycin (10 μmol/L). The uptake assay was performed under both acidic conditions (pH 6.0) and alkaline conditions (pH 8.4). The assay was also performed by pretreating cultures with 30 mmol/L NH4Cl and then removing it to examine the effects of intracellular acidification, or by incubating cultures continuously with 30 mmol/L NH₄Cl to examine the effects of intracellular alkalization. During measurement of prodrug uptake at acidic intracellular pH_i, extracellular NH₄Cl was removed after the NH₄Cl preincubation because intracellular NH₃ rapidly diffuses out of the cells, resulting in accumulation of protons released from NH₄⁺. ¹⁷

The ability of cationic inhibitors to affect prodrug uptake was assessed by pretreating cultures with the inhibitors (including 200 μmol/L of pyrilamine, propranolol, and imipramine, 500 μmol/L astrocytes (ACs) of hemicholinium-3, choline, L-carnitine, TEA, L-lysine, L-arginine, and spermine) for 30 minutes at 37°, then adding prodrug I or II (80 μmol/L) and incubating another 1 hour. Cultures were rinsed three times with ice-cold PBS and processed as described above. Lineweaver-Burk plot describing uptake of prodrugs I and II was generated by performing the uptake assay at concentrations of 20 to 120 μmol/L in the absence or presence of 200 μmol/L pyrilamine.

Prodrug Transendothelial Transport in an In Vitro Blood–Brain Barrier Model

Transendothelial transport was studied as described. ²⁵ Briefly, the medium in the lower chamber of the BBB model was replaced with 1 mL HBSS, and 0.2 mL HBSS containing DIBU or prodrug (80 μM) was added to the upper chamber. At 15, 30, 60, 90, and 120 minutes, 100 μL of HBSS was collected from the lower chamber, and the same volume of fresh HBSS was added to maintain constant volume. The 100-μL sample was diluted with acetonitrile for HPLC analysis. Transendothelial drug permeability was measured in terms of the accumulated cleared volume (μL): accumulated cleared volume (μL) = drug accumulated in the lower chamber/drug initially added to the upper chamber. An apparent permeability coefficient (P_app) was calculated for DIBU and prodrugs using the equation:

$p_{\text{app}}=(dQ/dt)/\mathrm{A}-C_0$

where the dQ/dt (μmol/s) is the drug permeation rate; A, the surface area of the membrane; C₀ (μmol/L), the initial drug concentration.

This assay was repeated under different conditions to probe the mechanisms involved. It was performed at 4° to allow comparison with 37° results and thereby assess the role of lipid-mediated simple diffusion. It was performed in the presence of 0.1% NaN₃ to assess the energy dependence of the process. To assess the possible role of pyrilamine-sensitive H^+/OC antiporter, medium in the upper chamber was supplemented with cationic inhibitors at 37° for 15 minutes before addition of prodrugs I and II.

In situ Rat Brain Perfusion with Prodrugs

Brain perfusion was performed as described. ²⁶ Male Sprague-Dawley rats weighing 220 ±20 g (Dashuo Biotechnology, Chengdu, China) were anesthetized using intraperitoneal chloral hydrate (400 mg/kg), and the right carotid artery was perfused with Krebs-Henseleit buffer (118 mmol/L NaCl, 4.7 mmol/L KCI, 25 mmol/L NaHCO₃, 1.2 mmol/L KH₂PO₄, 2.5 mmol/L CaCl₂, 1.2 mmol/L MgSO₄, 10mmol/L D-glucose, 10mmol/L HEPES, and 20IU/mL heparin, pH 7.4) containing DIBU or prodrug (80 μmol/L). Perfusion was performed for 30 seconds at a flow rate of 4.9mL/min. After perfusing the brain with drug-free Krebs-Henseleit buffer for another 30 seconds, the right brain hemisphere was collected, rinsed with ice-cold saline solution, wiped dry, weighed, and stored at − 20° until analysis.

To examine the role of passive diffusion in prodrug uptake by the brain, brain capillaries were first perfused for 30 seconds with drug-free perfusion medium cooled to 4°. In additional experiments, medium containing prodrug was precooled to 4° before perfusion, or it was supplemented with 0.1% NaN₃. To examine the role of putative OC transporters at the BBB, medium containing prodrug I or II (80 μmol/L) was supplemented with 1 mmol/L pyrilamine before perfusion.

Statistical Analysis

Statistical comparisons were performed by one-way analysis of variance (ANOVA) and Student's t-test, P < 0.05 and P < 0.01 were considered indications of statistical differences and statistically significant different, respectively.

RESULTS

In the first set of experiments, BCEC monolayers were used to mimic the structure and function of the BBB. These monolayers were incubated with DIBU or each of the four prodrugs to examine cytotoxicity, effects on membrane potential, and mechanisms of prodrug uptake. In the second set of experiments, transendothelial transport of prodrugs in BCEC-AC cocultures was analyzed as a model of the BBB. This two-step experimental approach mirrors the overall process of drug uptake by the brain, which is thought to comprise uptake by endotheliocytes, followed by passage across the BBB into the brain parenchyma.

Characterization of the In Vitro Model of the Blood-Brain Barrier The barrier was created and then analyzed by electron microscopy to confirm the expected morphology. Scanning electron microscopy of BCEC-AC primary cocultures showed that each cell type presented the expected morphology (Figures 2A and 2B). Cells in the confluent BCEC monolayer were spindle-shaped, tightly packed, and nonoverlapping (Figure 2C). Tight junctions between the BCECs were visible by transmission electron microscopy (Figure 2D), indicating the successful establishment of an in vitro model of the BBB. Transendothelial electrical resisitance of cocultures measured 279±12 Ω cm².

Figure 2.

Morphologic analysis of in vitro models of blood-brain barrier. Scanning electron micrographs of (A) blood to brain by capillary endothelial cells (BCECs) and (B) ACs; (C) Light micrographs of BCECs. (D) Transmission electron micrograph of the in vitro BCEC-AC coculture model.

Prodrug Cytotoxicity in Blood to Brain by Capillary Endothelial Cells

Before beginning any other culture experiments, the MTT assay was used to measure the cytotoxicity of DIBU or prodrugs with ethanolamine-related structures. Even at the highest prodrug concentrations tested, more than 80% of cells remained viable after 4-hour incubation with prodrugs (Figure 3A). This suggests that modifying DIBU with ethanolamine-related structures does not increase its cytotoxicity.

Figure 3.

Effects of prodrugs on the tight junctions of the in vitro model of blood-brain barrier (BBB). (A) Cytotoxicity of dexibuprofen (DIBU) and of prodrug I, II, III, or IV on blood to brain by capillary endothelial cells (BCECs) after 4-hour incubation. Results were presented as average cell viability %± s.d. (n = 5). (B) Transendothelial electrical resisitance (TEER) changes in the BBB model before and after exposure to DIBU or one of the four prodrugs. Data represent mean ±s.d. (n = 3). (C) F-actin filaments in a BCEC monolayer were incubated for 30 minutes with DIBU or prodrug, then observed by laser confocal scanning microscopy. Cells without any drug treatment served as controls. Scale bar = 20 μm. (D) Effects of DIBU and prodrugs on the zeta potential of BCECs. Cultures were treated for 30 minutes at 37° with 80 μmol/L DIBU or prodrug. Cultures were treated with TMA-DPH served as a positive control, while cultures incubated with phosphate-buffered saline (PBS) served as a negative control. Data represent mean ±s.d. (n = 3). **P < 0.01 versus control group. TMA-DPH, 1-(4-(trimethylamino) phenyl)-6-phenyl-1,3,5-hexatriene.

Transendothelial electrical resisitance of the in vitro BBB was approximately 250 Ω cm² before and after incubation with prodrug, suggesting that none of the prodrugs significantly affected the integrity of the in vitro model. F-actin staining for 1 hour revealed continuous actin filaments with a fluorescent intensity similar to that of DIBU-treated cells. This further suggested that prodrugs did not compromise the tight junctions of the BCEC monolayer.

Prodrug Effects on the Blood to Brain by Capillary Endothelial Cell Membrane Potential

As reported previously, ²² membrane surface charge (zeta potential) of BCEC monolayers was highly negative, ζ = − 21.50± 1.37 mV (Figure 3D). TMA-DPH, a cationic probe of membrane dynamics in living cells, served as a positive control. This probe intercalates among the hydrophilic head groups of membrane phospholipids and neutralizes the anionic charges of lipids in the outer layer of the cytoplasmic membrane. As expected, TMA-DPH, increased the zeta value (ζ = − 15.80± 0.66 mV), bringing it closer to the uncharged state. While incubating the monolayers with DIBU did not significantly change the zeta potential from the value in the control group (no treatment, P >0.05), incubating them with any of the prodrugs brought the potential significantly closer to an uncharged state (P < 0.001). In contrast, no significant differences were observed between DIBU and the control group, indicating that DIBU did not interact with negatively charged cell membrane. These results suggest that prodrug modification with ethanolamine-related structures increased affinity with the anionic surface of endothelial cells, which may help to explain why the prodrugs were taken up by the brain to a greater extent than unmodified DIBU.

Prodrug Uptake by Blood to Brain by Capillary Endothelial Cells All prodrugs showed significantly greater uptake than DIBU with increasing incubation time and concentrations (Figure 4). Relative uptake under these experimental conditions followed the sequence: I > II > III > IV> DIBU. This is consistent with our previous study of brain targeting by these prodrugs. ¹⁸

Figure 4.

Uptake of dexibuprofen (DIBU) and prodrugs I, II, III, and IV by blood to brain by capillary endothelial cell (BCEC) monolayers under various conditions. (A) Cultures were exposed for 1 hour at 37° to various DIBU or prodrug doses (20, 40, 60, 80, and 120 μmol/L). (B) Cultures were exposed to 80 μmol/L DIBU or prodrug and analyzed at 10, 20, 30, 60, and 120 minutes. (C) Cultures were exposed for 1 hour to 80 μmol/L DIBU or prodrug at either 37 ° or 4°. (D) Cultures were exposed for 1 hour at 37 ° to 80 μmol/L prodrug in the presence of metabolic energy inhibitors. (E) Cultures of either BCECs or L929 cells were exposed at 37° for 1 hour to 80 μmol/L DIBU or prodrug. Data represent mean ± s.d. (n = 3). **P < 0.01.

Uptake characteristics were different for the prodrugs. Uptake of DIBU and of prodrugs III and IV increased linearly with incubation time and concentrations, it did not differ significantly between 37° and 4°, and it was not inhibited by sodium azide or rotenone. These results suggest that DIBU and prodrugs III and IV enter the brain primarily via passive membrane diffusion. Similar results were obtained in BCEC and L929 cultures, suggesting that the uptake mechanism is cell line independent.

Unlike the uptake of these two prodrugs, the uptake of prodrugs I and II by BCECs was time, concentration, temperature, and energy dependent (Figure 4). Uptake was greater at 37° than at 4° for both prodrug I (11.55 ± 1.51 versus 6.36±0.45 nmol/mg) and prodrug II (7.35 ±0.83 versus 3.95 ±0.41 nmol/mg). Uptake of I and II was significantly inhibited by pretreatment with sodium azide and rotenone in BCECs (Figure 4D). These results suggest an active transport process. Consistent with this interpretation, higher doses of prodrug I and II showed saturation behavior (Figure 4B). Prodrug II showed a much lower affinity for the putative transporter system, reflected in a lower maximum uptake (6.36 ±0.42 versus 9.83 ±0.19 nmol/mg). Uptake of prodrugs I and II was much greater in BCECs than in L929 cells (Figure 4E), suggesting that the functional characteristics of the transport system that brings these prodrugs into the brain may differ across cell lines.

In Vitro Mechanistic Studies of Prodrug Uptake by the Brain To define the functional characteristics of the putative OC transporter system in BCECs, the uptake experiments were repeated using culture medium at different pH values. Uptake of I and II depended on extracellular pH, with uptake decreasing significantly as the pH decreased from 8.4 to 6.0 (Figure 5A). If uptake only involves passive diffusion, then the pH partition theory predicts that uptake should decrease as the proportion of the unchanged species decrease. ²⁷ In moving from pH 7.4 to 6.0, the proportion of unchanged prodrug II is expected to decrease from 15.4% to 0.72% based on a pKa of 8.14 (see Supplementary information, Table 1), corresponding to a change of only 15% in the proportion of unchanged species. Nevertheless, uptake was 3.5-fold lower at pH 6.0 than at pH 7.4, a difference much greater than one would predict for a mechanism of simple passive diffusion. Similar results were obtained for prodrug I. This provides further evidence in favor of a mechanism involving active transport.

Figure 5.

Mechanistic studies of the uptake of prodrugs I and II by blood to brain by capillary endothelial cell (BCEC) monolayers. (A) Effect of extracellular pH on uptake of 80 μmol/L prodrug after 1-hour incubation at 37°. (B) Effect of intracellular pH on uptake of 80 μmol/L prodrug after 1-hour incubation at 37°. (C) Effects of replacing sodium ion and altering membrane potential on uptake of 80 μmol/L prodrug after 1-hour incubation at 37°. Cultures were incubated in sodium-containing buffer (control) or sodium-free buffer in which sodium had been replaced with N-methylglucamine or potassium. Cultures were also pretreated with 10 μmol/L valinomycin, and then uptake was measured in buffer containing sodium and valinomycin. Potential competitive inhibitors (pyrilamine, propranolol, imipramine, hemicholinium-3, choline, l-carnitine, tetraethylammonium chloride (TEA), L-lysine, L-arginine, and spermine) were incubated with cultures for 1 hour at 37°, and uptake of (D) prodrug I and (E) prodrug II was examined. Lineweaver-Burk plot of the inhibitory effect of pyrilamine on the uptake of prodrug I (F) and prodrug II (G). Cellular uptake of I and II (20 to 120 μmol/L, for 20 minutes) was measured in the absence (-) or presence of 200 μmol/L pyrilamine (•). Data represent mean ± s.d. (n = 3). *P < 0.05; **P < 0.01.

Table 1.

Structures of the reported substrates of the H⁺/OC antiporter

Substrate	Structure	Substrate	Structure
Pyrilamine		Tramadol
Pramipexole		Nicotine
Oxycodone		Propranolol
Diphenhydramine		Codeine
Apomorphine		Clonidine
Clonidine

To examine the effect of intracellular pH on cell uptake, cells were treated with NH₄Cl. Acute treatment with NH₄Cl, which causes intracellular alkalinization, significantly reduced the uptake of prodrugs I and II to ~ 50% of the control. Pretreatment with NH4Cl, which causes intracellular acidification, significantly stimulated uptake of prodrugs I and II (Figure 5B).

Replacing extracellular Na⁺ with N-methylglucamine or K⁺ did not significantly affect uptake of prodrug I or II (Figure 5C), nor did treatment with the potassium ionopore valinomycin. These results suggest that the prodrug uptake is insensitive to sodium and membrane potential.

To begin to define the molecular characteristics of the putative transporters involved in prodrug uptake in BCECs, the uptake assay was repeated in the presence of potential inhibitors. The cationic drugs such as pyrilamine, propranolol, and imipramine significantly inhibited uptake of I and II (Figures 5D and 6E), but the choline transport inhibitor hemicholinium-3 and the OCTN2-specific substrate L-carnitine did not. Choline, the substrate of the choline transport system, and tetraethylammonium, a classical inhibitor of OCTs, only slightly inhibited uptake. Similarly, alkaline amino acids such as L-lysine, L-arginine, and spermine did not significantly affect uptake.

Figure 6.

Mechanism of transendothelial transport of DIBU and prodrugs in an in vitro BCEC-AC coculture model of the blood–brain barrier. (A) Accumulated cleared volume (μL) of DIBU and prodrugs across the barrier. (B) P_app value of DIBU and prodrugs at 37° or 4° or in the presence of NaN₃. *P < 0.05; **P < 0.01, compared with P_app values obtained at 37°. Transport efficiency of (C) prodrugs I and (D) II in the presence of the inhibitors pyrilamine, propranolol, or imipramine. The amount of I or II transported at 37° served as the control. *P < 0.05; **P < 0.01, compared with the control group. Data represent mean ± s.d. (n = 3). BCEC, blood to brain by capillary endothelial cell.

Lineweaver-Burk analysis of prodrug I and II uptake in the presence or absence of pyrilamine showed that this cationic H1 antagonist competitively inhibited uptake of both prodrugs (Figures 5F and 5G); the observed K_i values were 102.61 umol/L for I and 123.35 μmol/L for II.

Transendothelial Transport of Prodrugs in an In Vitro Blood–Brain Barrier Model

In an effort to model how our prodrugs modified with ethanolamine-related groups may move across the blood–brain barrier into the brain parenchyma, we supplemented the BCEC monolayer studies with studies in an in vitro model of the barrier. In the BCEC-AC coculture BBB model, accumulated clearance volume (μL) varied in the order I > II > III > IV> DIBU (Figure 6A). This suggests that modifying DIBU with ethanolamine-related structures substantially increased the efficiency of its transport across the barrier. P_app (%) relative to DIBU was 981.28% for prodrug I, 769.61% for II, 329.93% for III, and 164.03% for IV (Figure 6B). For DIBU and prodrugs III and IV, transendothelial transport, like the uptake process, was energy independent, suggesting a passive process. For prodrugs I and II, in contrast, P_app was significantly higher at 37° than at 4°, and significantly higher in the absence than in presence of NaN₃. Pyrilamine, propranolol, and imipramine significantly decreased P_app of prodrugs I and II (**P < 0.01, Figures 6C and 6D).

Prodrug Transport from Blood to Brain in Rats

After in vivo brain perfusion at 37°, prodrugs concentrations (nmol/g) in the right brain hemisphere were 9.24 ±1.50 for DIBU, and 117.69±20.66, 99.59±9.88, 66.51 ±16.85, and 50.03± 11.76 for the four prodrugs (Figure 7A). Modifying DIBU with ethanolamine-related structures significantly enhanced the blood-to-brain transport. Perfusing at 4° or in the presence of NaN₃ substantially decreased blood-to-brain transport of prodrug I to 63.65 ±14.19 and 70.47 ± 16.56 nmol/g, respectively, and decreased transport of prodrug II to 68.06 ± 16.05 nmol/g and 71.87 ±11.24 nmol/g. Transport of DIBU or prodrugs III or IV did not differ significantly between 37° and 4° or between the presence and absence of NaN3. These results, consistent with those of the uptake studies in BCEC monolayers, suggest that uptake of I and II primarily involves an energy-dependent process, while uptake of DIBU, III, and IV primarily involves passive diffusion.

Figure 7.

In vivo prodrug transport from blood to brain in rats, measured using the in situ brain perfusion method. (A) Brain uptake of DIBU or prodrug after 30-second perfusion at 37° or 4° or in the presence of NaN₃. *P < 0.05 versus 37° **P < 0.01 versus 37°. (B) Uptake of I and II by the brain after perfusion with or without pyrilamine (1 mmol/L). *P < 0.05 versus the control; **P < 0.01 versus the control. Data represent mean ± s.d. (n = 5).

To verify the functional characteristics of the putative transport system observed in vivo, the perfusion medium was supplemented with cationic drug inhibitors. Pyrilamine significantly decreased brain uptake of I by 29.27% and of II by 18.64% (Figure 7B). These results confirm the in vitro findings suggesting that a putative transporter-mediated system is involved in the brain uptake of I and II.

DISCUSSION

In previous work as part of our ongoing efforts to synthesize a new generation of drugs capable of reaching the brain in pharmacologically effective doses, we designed and synthesized a series of prodrugs modified with ethanolamine-related structures and showed that they delivered DIBU safely and efficiently to the brain. Nevertheless, despite sharing a similar structure, the four prodrugs exhibited dramatically different targeting efficiencies. Here, we extend that work by providing the first molecular insights into brain targeting of cationic drugs. Prodrugs III and IV appear to be taken up by an energy-independent, passive process. In contrast, prodrugs I and II appeared to be taken up by an active transport process depend on time, concentration, temperature, and energy. These results not only provide the basis for future, more detailed mechanistic studies; they also indicate that even prodrugs with globally similar structure can cross the BBB by quite different mechanisms.

Cytotoxicity of cationic materials or drugs has become a major concern in drug development. Modification with ethanolamine-related structures did not increase the cytotoxicity of prodrugs compared with DIBU (Figure 3A). Nicotine is reported to increase the BBB permeability via the modulation of BBB tight junctions. ²⁸ Ethanolamine-related structures share the same structure unit with nicotine, and thus, we studied the effect of prodrugs on the tight junction of in vitro BBB model. Transendothelial electrical resisitance measurements and F-actin staining showed that they did not compromise barrier integrity (Figures 3B and 3C). This suggests that the observed entry of prodrugs into the brain reflected passive or active transport instead of intercellular leakage.

Endothelial cell membranes are negatively charged, consisting mainly of phosphatidylcholine with smaller amounts of anionic phosphatidylserine. ²⁹ The brain appears to absorb cationic peptides primarily as a result of their interaction with the anionic membrane, ³⁰ and the membranes of brain cells are quite anionic, with their lipids relatively exposed. This may explain why lipophilic cationic compounds are more likely than other molecules to cross the BBB. ²³ To analyze interactions between our prodrugs and the membrane of BCECs, we measured the surface charge (zeta potential) in the presence of prodrugs. The results suggest that modifying DIBU with ethanolamine-related structures increased affinity for the anionic surface of endothelial cells (Figure 3D), which may help enhance brain uptake. At the same time, the different ethanolamine modifications lead to the same extent of partial membrane neutralization, in contrast to their substantially different uptake efficiencies. We speculate that while the modifications globally increased interaction with BCECs, chemical and structural differences also strongly contribute to the overall transport efficiency and perhaps even to the underlying transport mechanism.

Primary BCECs express several transporters, including GluT1, LAT1, and OCTN2. Thus, BCECs appear to be a suitable cell line for investigating the BBB transport of prodrugs. In the cellular uptake study, conjugation of ethanolamine-related structures significantly improved the cellular uptake of dexibuprofen in BCECs (Figure 4). The uptake of III and IV was proven to be energy independent, indicating a passive uptake process. On the contrary, the uptake of I and II by BCECs was time, concentration, temperature, and energy dependent, indicating an active transport process is involved in their uptake. Competition experiments showed that cationic compounds (hemicholinium-3, TEA, and L-carnitine), choline, and its analogs did not inhibit uptake of I or II, nor did alkaline amino acids (Figures 5C and 5D). These results suggest that uptake of I and II does not involve OC transporters (OCTs and OCTNs), choline transporters, or the electrostatic interactions observed in the transport of cationic peptides. Instead, the lipophilic cationic drugs pyrilamine, imipramine, and propranolol significantly inhibited the uptake of I and II; in fact, pyrilamine competitively inhibited uptake of I with Ki value of 102.61 μmol/L and II with Ki value of 123.35 μmol/L. These results, together with the observation that transport was dependent on pH but independent of sodium and membrane potential (Figures 5A and 5B), suggest that I and II share a common H^+/OC transporter with pyrilamine. ³¹

To further explore the fate of prodrugs after they are taken up into BCECs, we examined their transendothelial transport in a BCEC-AC coculture model of the BBB. This model relies on transwell culture inserts to create ‘contact through feet’ interface for interactions between the two cell types. In this coculture model, specific functional proteins such as enzymes, transporters, and tight junctional proteins are expressed, which likely help the system achieve a higher TEER than other BBB models. ³² Similar to the uptake process, the transendothelial transport of DIBU and of prodrugs III and IV was energy independent, suggesting a passive transport process. In contrast, P_app of I and II was time, concentration, and energy dependent and was significantly lower in the presence of pyrilamine, propranolol, and imipramine (Figure 6). These findings suggest that prodrugs III and IV are taken up by the brain primarily via passive diffusion, while prodrugs I and II are taken up via passive diffusion as well as via the same cation transporters such as pyrilamine, propranolol, and imipramine.

The substrate specificity of the H⁺/OC antiporter is controversial. While some speculate that it preferentially interacts with OCs with secondary or tertiary amine moieties, ¹⁶ others suggested that it recognizes only tertiary amines with N-dimethyl or N-diethyl moieties, but not ethanolamine or quaternary amines. ³³ Indeed, a review of the published substrates of the H⁺/OC antiporter (Table 1) suggests that the majority are tertiary amines. Our results support this conclusion, since the antiporter seems to be involved in the brain transport only of tertiary amine-based prodrugs I and II, but not of primary or secondary amine-based prodrugs III and IV.

The lipophilicity of the amine modification may also be important. Kubo ³⁴ conducted an analysis of the relationship between the inhibitory effect and the lipophilicity of inhibitors, revealed a lipophilicity-dependent inhibitory effect of amines on the H⁺/OC antiporter. Work from our laboratory supports the idea that lipophilicity of the N,N-dimethyl amino group-conjugated drug derivatives is essential for brain targeting. ²¹ Differences in lipophilicity may help explain why our four prodrugs were transported into the brain by different mechanisms. At the same time, structural factors, such as steric hindrance, may also have a role: despite their structural similarity, prodrug II with a N,N-diethyl group showed much lower affinity for the putative transporter system than prodrug I with a N,N-dimethyl group. This may attribute to greater steric hindrance with II.

The in situ rat brain perfusion technique is versatile and sensitive for studying the functional characteristics and mechanism of drug uptake across the BBB.^35,36 Consistent with the results in the BCEC monolayers and in the BCEC-AC coculture system, uptake efficiency was significantly higher for the four prodrugs than for DIBU, and the uptake mechanism was primarily passive diffusion for III and IV and primarily active transport for I and II. Also as in the in vitro experiments, uptake of I and II in the in situ perfusion experiments was significantly inhibited by pyrilamine, confirming the involvement of the putative transporter-mediated system. Based on our in vitro and in vivo findings, we propose the following sequence of steps in the transport of prodrugs I and II into the brain. The cationic prodrug interacts with the cell membrane due to its lipophilicity and positive charge, and it is specifically recognized by transporters expressed in the brain microvascular endothelial cells, which then transport it across the BBB. Due to the difference in steric hindrance, N,N-dimethyl-based prodrug I showed a much higher affinity for the putative transporter system compared with N,N-diethyl-based prodrug II, resulting in higher brain targeting efficiency.

The present work provides the first mechanistic details about brain targeting by dexibuprofen prodrugs modified with ethanolamine-related structures. In vitro and in vivo studies showed that prodrugs were taken up into the brain in a structure-dependent manner. These findings lay the foundation for more systematic structure–function analysis of cationic drug transport across the BBB, which may aid the development of novel brain-targeted prodrugs. Future work should focus on the discovery of prodrug efflux processes from brain and the identification of H⁺/OC antiporter at the molecular level.