Novel flurbiprofen derivatives with improved brain delivery: synthesis, in vitro and in vivo evaluations.

http://informahealthcare.com/drd ISSN: 1071-7544 (print), 1521-0464 (electronic) Drug Deliv, Early Online: 1–10 ! 2014 Informa Healthcare USA, Inc. DOI: 10.3109/10717544.2014.954165

ORIGINAL ARTICLE

Novel flurbiprofen derivatives with improved brain delivery: synthesis, in vitro and in vivo evaluations Dan Zheng1*, Xiao Shuai2*, Yanping Li1, Peng Zhou1, Tao Gong1, Xun Sun1, and Zhirong Zhang1 1

Drug Delivery Downloaded from informahealthcare.com by Universitat de Girona on 10/17/14 For personal use only.

Key Laboratory of Drug Targeting and Drug Delivery Systems, Ministry of Education, West China School of Pharmacy, Sichuan University, Chengdu, Sichuan, People’s Republic of China and 2West China Hospital, Sichuan University, Chengdu, Sichuan, People’s Republic of China Abstract

Keywords

Tarenflurbil (R-flurbiprofen) was acknowledged as a promising candidate in Alzheimer’s disease (AD) therapy. However, the Phase III study of tarenflurbil was extremely restricted by its poor delivery efficiency to the brain. To tackle this problem, the novel carriers for tarenflurbil, racemic flurbiprofen (FLU) derivatives (FLU-D1 and FLU-D2) modified by N,N-dimethylethanolaminerelated structures were synthesized and characterized. These derivatives showed good safety level in vitro and they possessed much higher cellular uptake efficiency in brain endothelial cells than FLU did. More importantly, the uptake experiments suggested that they were internalized via active transport mechanisms. Biodistribution studies in rats also illustrated a remarkably enhanced accumulation of these derivatives in the brain. FLU-D2, the ester linkage form of these derivatives, achieved a higher brain-targeting efficiency. Its Cmax and AUC0–t were enhanced by 12.09-fold and 4.61-fold, respectively compared with those of FLU. Additionally, it could be hydrolyzed by esterase in the brain to release the parent FLU, which might facilitate its therapeutic effect. These in vitro and in vivo results highlighted the improvement of the braintargeted delivery of FLU by making use of N,N-dimethylethanolamine ligand, with which an active transport mechanism was involved.

Active transport, Alzheimer’s disease, brain targeting, flurbiprofen, N,N-dimethylethanolamine

Introduction The prolonged life expectancy has aroused great concerns of the increased rate of neurodegenerative disorders during the last few decades. Azheimer’s disease (AD), characterized by an insidious onset with memory impairment and an inexorably progressive cognitive decline, has become the most common form of dementia in the ageing population (Gasparini et al., 2004). Brain inflammation is one of the biochemical processes contributing to neurodegeneration in AD, which is mainly associated to the accumulated b-amyloid peptide (Ab) and neurofibrillary tangles in the degenerating neurons (Imbimbo, 2009a). Epidemiological studies indicate that long-term use of non-steroidal anti-inflammatory drugs (NSAIDs), such as flurbiprofen (FLU), could reduce the risk of AD and delay its progression (Launer et al., 2002; Etminan et al., 2003). FLU has attracted tremendous attention because of its actions on key hallmarks of AD. By modulating the activity of g-secretase, the pivotal enzyme that generates the Ab peptide from the Ab protein precursor, FLU either in

*These authors contributed equally to this work. Address for correspondence: Prof. Zhirong Zhang, West China School of Pharmacy, Sichuan University, No.17, Block 3, Southern Renmin Road, Chengdu, People’s Republic of China, 610041. Tel: +86-28-85501566. Fax: +86-28-85501615. E-mail: [email protected]

History Received 15 June 2014 Revised 8 August 2014 Accepted 9 August 2014

the racemic form or its R or S enantiomers could selectively reduce Ab42 level and inhibit the aggregation of Ab (Eriksen et al., 2003). The Phase II, 12-month study with tarenflurbil (R-flurbiprofen) revealed apparent positive effects in mildly AD patients (Wilcock et al., 2008). Unfortunately, the consequential Phase III study in AD patients with tarenflurbil has failed due to its poor brain penetration and the high dosage-related (800 mg b.i.d.) gastrointestinal adverse effects (Green et al., 2009; Imbimbo, 2009b). Preclinical studies had shown that tarenflurbil did not penetrate blood–brain barrier (BBB) satisfactorily in rodents, with a 1.3% ratio of cerebrospinal fluid (CSF) to plasma and negligible tarenflurbil level measured in the brain (2 mM) (Peretto et al., 2005; Kukar et al., 2007). In addition, both the high affinity to serum albumin and the active efflux transport restricted the passive diffusion of FLU to the brain (Parepally et al., 2006). The limited BBB penetration of tarenflurbil makes large dosage a necessity thus increased the risk of adverse effects. Therefore, enhancing the transport of FLU into brain becomes crucial to make it an effective neuroprotective agent. In fact, most central nervous system (CNS) therapeutic agents were hampered in their clinical application since they cannot be effectively delivered into the brain. BBB, the tight junction formed by brain endothelial cells, blocks paracellular transfer and forces most agents to enter brain either by lipophilic passive diffusion or by specific carrier- or receptor-


2

D. Zheng et al.

mediated transport mechanisms (Alam et al., 2010). Our previous work has proved ethanolamine-related structures as promising ligands with excellent brain specificity. It was found that drugs conjugated with these ligands could significantly enhance their accumulations in the brain after intravenous (i.v.) injection into rats (Zhang et al., 2012). Ethanolamine-related structures are widely found in either endogenous substances including phosphatidylcholine (e.g. neurotransmitters) or various CNS therapeutic agents, such as chlorpromazine, meclofenoxate and procaine. With the amino group, ethanolamine is positively charged in physiological condition, and might become a substrate of the organic cation or choline transporters located at the BBB (Zhang et al., 2012). Moreover, there are several advantages for drug development with this ligand, including a simple, small and well-defined structure, and a lack of toxicity, which contrasts with the toxic effects sometimes reported with polymeric nanocarriers (Ai et al., 2011). In this study, FLU was conjugated with N,N-dimethylethanolamine-related ligands in order to increase its permeability into brain. FLU derivatives were synthesized via ester bond or amide bond, whose physicochemical properties and braintargeting abilities were evaluated both in vitro and in vivo. Cytotoxicity assay was performed on bEnd.3 cells and L929 cells to determine the safety level of these derivatives. The potential transport mechanism of these derivatives was also explored in vitro using brain endothelial cells.

Materials and methods Materials and animals FLU (racemic form,498% purity) was purchased from Tokyo Chemical Industry Co. Ltd, Tokyo, Japan, N,N-dimethylethylenediamine and N,N-dimethylethanolamine were supplied by Kelong Chemical Reagent Factory (Chengdu, China) and Sigma-Aldrich Corporation (Cleveland, OH), respectively. Fetal bovine serum (FBS) was purchased from Shanghai FuMeng Gene Bio-technology Co., Ltd (Shanghai, China). The other chemicals and reagents were of analytical grade or better. The culture plates were obtained from Nest Biotechnology Co., Ltd (Wuxi, China). Thin-layer chromatography (silica gel GF254) was employed to detect spots by ultraviolet (UV) radiation. Purification of the desired compounds was carried out by column chromatography on silica gel. 1H- and 13C-NMR (nuclear magnetic resonance) analyses were performed by AMX-400 Bruker Spectrometer (Bruker BioSpin, Karlsruhe, Germany). Chemical shifts were given in parts per million (). Mass spectroscopy was performed by Agilent 1200 series Rapid Resolution LC (RRLC) system. Protein contents of cell lysates were analyzed by the Pierce BCA (bicinchoninic acid) protein assay reagent kit (Pierce Biotechnology Inc., Rockford, IL). Male Sprague-Dawley rats (220 ± 20 g) were maintained in a germ-free environment with free access to food and water. All animal experiments protocols were approved by the Ethics Committee for Animal Experiments for Sichuan University, according to the requirements of the National Act on the use of experimental animals (People’s Republic of China).

Drug Deliv, Early Online: 1–10

Cell line and cell culture bEnd.3 cells, the immortalized mouse brain endothelial cell line, were kindly provided by Fudan University, Shanghai, China. The cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) with high glucose (HyClone, Waltham, MA) supplemented with 10% FBS, 100 IU mL1 penicillin and 100 mg mL1 streptomycin. L929 cells (own reserved) were grown in RPMI-1640 (Roswell Park Memorial Institute) medium (HyClone) supplemented with 10% calf serum (Minhai, Gansu, China), 100 IU mL1 penicillin and 100 mg mL1 streptomycin. Routinely, cells were maintained at 37 C, 5% CO2 and 95% relative humidity. The culture medium was changed every other day. Methods Synthesis of FLU derivatives The synthesis route of FLU amide bond derivative (FLU-D1) and ester bond derivative (FLU-D2) was outlined in Figure 1. Flurbiprofen chloride was first prepared by treating FLU (2.05 mmol, 500 mg) with SOCl2 at reflux for 3 h followed by evaporation under reduced pressure, as reported before (Song et al., 2004). The oily residue was dissolved in 4 ml anhydrous dichloromethane and was added dropwise into 4 ml TEA alkalized dichloromethane containing N,N-dimethylethylenediamine (4.1 mmol, 446.7 ml) or N,N-dimethyl ethanolamine (4.1 mmol, 415.5 ml) in ice-water bath (Halen et al., 2007). The reaction mixture was then stirred at 35 C for 1 h. After that, the solvent was diluted and washed several times with brine to quench the reaction and remove some hydrophilic hybrids. The organic phase was dried over anhydrous sodium sulfate and evaporated under reduced pressure. The final residues were purified by column chromatography on silica gel with CH2Cl2/MeOH, 50:1 (v/v) and 80:1 (v/v) to yield the desired compound FLU-D1 and FLU-D2, respectively. High-performance liquid chromatography analysis A reversed-phase high-performance liquid chromatography (HPLC) procedure was established for quantitative measurement. HPLC analysis was performed using Agilent instruments (Agilent Technologies, Palo Alto, USA) consisting of a 1260 Quaternary pump, a G1314C 1260 UV detector and a G1329B 1260 Autosampler. Separations were carried out using a Kromasil C18 reverse-phase column (250 4.6 mm, 5 mm). The column effluent was monitored at 246 nm with a flow rate of 1 ml/min at 35 C. The mobile phase was composed of acetonitrile and phosphate buffer (pH 3.5), and the ratios of solvents were 60:40, 35:65 and 40:60 (v/v) for analyses of FLU, FLU-D1 and FLU-D2, respectively. In vitro stabilities of FLU-D1 and FLU-D2 The in vitro stabilities of FLU-D1 and FLU-D2 were investigated in phosphate buffers (pH 2.47, 5.02, 6.82, 7.36 and 7.99), freshly prepared rat plasma and brain homogenates (diluted with 0.9% physiological saline), respectively. Initially, 200 ml of stock solution of derivatives (1 mg/ml) in saline was added to 5.0 ml of preheated medium. The mixtures were then kept at 37 ± 0.5 C with continuous shaking at 100 rpm. Samples of 100 ml were taken from the


DOI: 10.3109/10717544.2014.954165

Novel flurbiprofen derivatives with improved brain delivery

3

Figure 1. General procedure for the synthesis of FLU derivatives. Reagents and conditions: (A) SOCl2, DMF, refluxing for 2 h at 60 C; (B and C) CH2Cl2, Et3N, stirring at 35 C for 1 h.

mixtures at scheduled time points and 300 ml of acetonitrile was added. After immediate mixing and centrifugation for 10 min at 13 500 rpm, 20 ml of clear supernatants were analyzed by HPLC for residual FLU-D1/FLU-D2. Three samples were prepared at each time point. Determination of partition coefficients The apparent partition coefficients (PC) of FLU and its derivatives were measured in n-octanol/pH 7.4 buffer at 25 ± 0.5 C (Tammara et al., 1993). Mutually saturated phases were prepared beforehand. The initial aqueous concentrations of all drugs were 1 mg/ml. Aliquots sampled from each phase were diluted with acetonitrile and analyzed via HPLC method mentioned above. The PC were calculated by the equation: PC ¼ Co =Cw where Co and Cw represent the concentrations of solute in the oil phase and water phase, respectively. For each chemical, triplicate samples were tested. Determination of plasma protein binding The ultrafiltration method was employed to determine plasma protein binding ratio of FLU and its derivatives (Tammara et al., 1993). About 50 ml of FLU, FLU-D1 or FLU-D2 saline solution (1.0 mg/ml) was added to 950 ml of fresh rat plasma. The mixture was vortexed for a few seconds, and then incubated at 37 C for 2 h. Consequently, 0.1 ml of the mixture was taken for analysis of the total drug concentration. About 0.5 ml of mixture was then placed in an ultrafiltration tube (10 kDa molecular weight cutoff; Millipore Corporation, Bedford, MA) and centrifuged at 4000 rpm for 10 min. Samples collected from the bottom of the ultrafiltration tube was used to measure the unbounded fraction. Drug concentrations were analyzed using HPLC.

In vitro cytotoxicity on bEnd.3 cells and L929 cells To evaluate the in vitro cytotoxicity of FLU and its derivatives, MTT ((3-4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assays were performed using bEnd.3 and L929 cell lines. Briefly, bEnd.3 cells were seeded onto a 96-well bottom plates at a density of 5 103 per well and then incubated overnight; and L929 cells were seeded onto 96-well bottom plates at a density of 1 104 per well and incubated at 37 C for 2 days. On the day of administration, the culture medium was replaced with 200 ml of fresh medium containing FLU, FLU-D1 or FLU-D2 (20, 40, 80, 160 and 320 nmol/ml). Control wells were treated with an equivalent volume of drugfree medium. The cells were incubated at 37 C for 24 h. After that, all medium was removed and wells were rinsed with phosphate-buffered saline (PBS). About 200 ml of MTT solution (0.5 mg/mL) was added to each well and incubated for another 4 h. Then, the formed dark blue formazan crystals in each well were dissolved in 200 ml of dimethyl sulfoxide (DMSO). The absorbance of each individual well was read on a Varioskan Flash microplate reader (Thermo Fisher Scientific; Waltham, MA, USA) at 570 nm wavelength. Cell viability was calculated using the formula: ½Abssample ½Absblank 100% Cell viability ¼ ½Abscontrol ½Absblank In vitro cellular uptake studies Several different experiments were conducted using bEnd.3 cells. In one experiment designed to explore drug uptake at different time points after administration, bEnd.3 cells were seeded in NEST culture plate (Nest Biotechnology Co., Ltd) at a density of 2 105 cells per dish. On the second day, the culture medium was replaced with new medium containing Cellular uptake on bEnd.3 cells.


4

D. Zheng et al.

FLU-D1, FLU-D2 or FLU at a concentration of 80 nmol/ml, and the cells were incubated for 0.25, 0.5, 1 or 2 h, respectively. In another experiment which aimed to study the effects of different drug concentrations on its uptake, bEnd.3 cells were exposed to different concentrations (20, 40, 80 or 160 nmol/ml, respectively) of the three drugs and were incubated for 0.5 h at 37 C. We then further investigated the effects of temperature and NaN3 on cellular uptake of the drugs mentioned above. bEnd.3 cells were treated with the drugs (80 nmol/ml) and incubated for 0.5 h at 4 and 37 C, respectively. For NaN3 group, cells were pre-incubated with serum-free medium containing 0.1% NaN3 for 1 h at 37 C, and were then exposed to different drugs (80 nmol/ml) together with 0.1% NaN3 for another 0.5 h. After incubation, cells were washed with icecold PBS for three times to remove drugs attached on cell surface. Then, cells were digested and collected after centrifugation at 5000 rpm for 3 min. Subsequently, the cells were resuspended in 150 ml of ultra-pure water and lysed by five freeze/thaw cycles to release the intracellular drugs. About 20 ml of the cell lysate from each sample was taken to determine the total cell protein content using BCA assay reagent kit. 100 ml of the cell lysates were collected as samples for analysis of intracellular drug content. FLU-D2 was preliminarily hydrolyzed to release FLU by adding 10 ml 6 M NaOH to the cell lysates, and after incubation at 30 C for 5 min, 10 mL of 6 M HCl was added in to neutralize the excess NaOH (Chen et al., 2009). Finally, 100 ml of methanol was added to cell samples to precipitate the protein, and the mixtures were centrifuged at 13 500 rpm for 10 min. The resulting supernatants were analyzed using HPLC. The result of cellular uptake was expressed as the amount (nmol) of FLU, FLU-D1 or FLU-D2 associated with a unit weight (1 mg) of cellular protein. The experiment was repeated for three times for each group and the data were presented as mean ± standard deviation (SD). L929 cells were seeded in NEST culture plates (Nest Biotechnology Co., Ltd) at a density of 2 105 cells/dish and incubated for 2 days. Then cells were treated with FLU, FLU-D1 or FLU-D2 at a concentration of 80 nmol/ml for 0.5 h at 4 and 37 C, respectively. After that, cells were washed three times with ice-cold PBS and collected for determination of intracellular drug concentration by HPLC. The methods of sample processing and HPLC assay were mentioned above.

Cellular uptake on a different cell line.

In vivo studies Male Sprague-Dawley rats (220 ± 20 g) were assigned randomly into three groups (n ¼ 5 in each group) and fasted overnight. Then FLU, FLU-D1 or FLU-D2 dissolved in normal saline was intravenously injected at a single-dose equivalent to 10 mg/kg body weight of FLU. In the pharmacokinetic study, at predetermined time intervals (0.083, 0.25, 0.5, 1, 2, 4, 8, 12 and 24 h), blood samples were collected and plasma was obtained through centrifugation at 5000 rpm for 5 min. In distribution study, at Pharmacokinetics and biodistribution studies in rats.


0.083, 0.25, 0.5, 1, 2 and 4 h after i.v. injection, the blood samples were collected and the rats were sacrificed. Tissues including brains, hearts, livers, spleens, lungs and kidneys were harvested followed by quick washing with cold saline and then homogenized with 2-fold volume of 0.9% saline (g/ml). The homogenates and blood samples were stored at 40 C until assays were performed. Before conducting analysis, samples of FLU-D2 were preliminarily hydrolyzed to FLU following the method mentioned earlier. About 100 ml of plasma or tissue homogenates was extracted by 300 ml methanol, and then quantitative measurement was done by HPLC. Data analysis The area under the curve (AUC0–t), the maximal concentration (Cmax) and the clearance (CLZ) were calculated using the Data and Statistics software package (DAS 3.2.5; Shanghai, China). The relative uptake efficiency (RE) and concentration efficiency (CE) were calculated to evaluate the brain-targeting properties of FLU-D1 and FLU-D2. The values of RE and CE are defined as follows: RE ¼ ðAUC0t ÞD =ðAUC0t ÞF CE ¼ ðCmax ÞD =ðCmax ÞF where D and F represent derivatives and FLU, respectively. The statistical analysis was performed using a student’s t-test, and groups were considered significantly different when p value is 50.05.

Results Synthesis of FLU derivatives The derivatives were synthesized as outlined in Figure 1. FLU-D1 was obtained as white solid substance (75.6%) and FLU-D2 as colorless oil (83.5%). 1H-NMR, 13C-NMR and ESI-MS confirmed the assigned structures. The detailed results were shown below. FLU-D1: 1H-NMR (400 MHz, DMSO-d6): 7.54–7.37 (m, 6H, aromatic), 7.26–7.22 (m, 2H, ’-H), 3.71–3.66 (q, 1H, CH–CH3), 3.19–3.08 (m, 2H, NH–CH2), 2.28–2.25 [m, 2H, CH2–N(CH3)2], 2.13 [s, 6H, N(CH3)2], 1.36–1.34 (d, 3H, CH3). 13C-NMR (100 MHz, DMSO-d6): 172.88 (s), 160.26–157.82 (d), 144.56–144.49 (d), 135.25 (s), 130.64– 130.61 (d), 128.92–128.80 (d), 128.90 (s), 127.90 (s), 126.57– 126.44 (d), 124.03 (s), 115.18–114.95 (d), 58.35 (s), 45.37 (s), 44.66 (s), 37.12 (s), 18.59 (s). ESI-MS: m/z [M + H]+: 315.3. FLU-D2: 1H-NMR (400 MHz, DMSO-d6): 7.55–7.38 (m, 6H, aromatic), 7.28–7.22 (m, 2H, ’-H), 4.15–4.12 (m, 2H, OCH2), 3.92–3.87 (q, 1H, CH–CH3), 2.45–2.42 [m, 2H, CH2–N(CH3)2], 2.11 [s, 6H, N(CH3)2], 1.44–1.42 (d, 3H, CH3). 13C-NMR (100 MHz, DMSO-d6): 173.44 (s), 160.35–157.90 (d), 142.70–142.62 (d), 135.05 (s), 130.93– 130.90 (d), 128.92–128.80 (d), 128.90 (s), 128.00 (s), 127.08– 126.95 (d), 124.21–124.18 (d), 115.50–115.27 (d), 62.58 (s), 57.33 (s), 45.43 (s), 44.18 (s), 18.55 (s). ESI-MS: m/z [M + H]+: 316.3. In vitro stabilities of FLU-D1 and FLU-D2 Data concerning the chemical and enzymatic stabilities of the derivatives were presented in Figure 2. For FLU-D1, no

DOI: 10.3109/10717544.2014.954165


substantial hydrolysis (510%) was found in either PBS buffers or enzymatic circumstances after 12-h incubation, suggesting that the amide bond in FLU-D1 is highly stable (Figure 2A). As expected, the ester bond in FLU-D2 was more susceptible to hydrolysis than the amide bond in FLU-D1. The ester bond appeared to be highly stable in pH 2.5 and pH 5.0 buffers as the remained FLU-D2 in the buffered solutions was 95–105% of the original amount at any time points. Nevertheless, both the alkaline condition and enzymatic environment accelerated its hydrolysis. After 12-h incubation, 51% of the intact FLU-D2 remained in the plasma and homogenate (Figure 2B). This implies that FLU-D2 with an ester bond might be a potential substrate of specific esterase in the plasma and the brain (Rautio et al., 2008).

FLU-D2 could not be calculated because of its rapid hydrolysis in plasma.


PC assay It is well-known that BBB drug penetration is strongly influenced by solute lipophilicity, therefore, PC measurement is necessary for this study. The log PC data of FLU, FLU-D1 and FLU-D2 were 2.38, 1.64 and 2.2, respectively, indicating that both the derivatives are less lipophilic than the parent drug FLU (Table 1). However, the ester bond derivative (FLU-D2) is more lipophilic than the amide bond one (FLU-D1).

5

Evaluation of cytotoxicity on bEnd.3 and L929 cells An important aspect of the investigation of novel drug compounds for brain delivery is the evaluation of their tolerability and toxicity upon the interaction with their target cells. The results illustrated that both cells were well-tolerated when exposed to the three drugs in different dosage ranging from 20 to 320 nmol/ml, and the cell viabilities were all above 80% (Figure 3). Statistical analysis showed no significant differences of cell viability among FLU, FLU-D1 and FLU-D2 treatments. These results indicate that the FLU derivatives are safe in vitro, which paves way for further study.

Table 1. PC of FLU and its derivatives represented as log PC values. Parameter Log PC

FLU

FLU-D1

FLU-D2

2.38 ± 0.008

1.64 ± 0.025

2.20 ± 0.019

Data represented the mean ± SD (n ¼ 3).

Table 2. Plasma binding ratios of FLU and its derivatives.

Plasma protein binding assay The plasma protein binding is considered to be a critical factor influencing brain uptake. As shown in Table 2, FLU and FLU-D1 possessed a plasma binding ratio of 98.9% and 98.5%, respectively, the high affinity of which might reduce their free fractions in blood circulation. The binding ratio of

Figure 2. In vitro stabilities of FLU-D1 (A) and FLU-D2 (B) in phosphate buffer solutions of different pH values, plasma and brain homogenates. Data represented as mean ± SD (n ¼ 3).

Parameter Binding ratio (%)

FLU

FLU-D1

FLU-D2

98.9 ± 0.07

98.5 ± 0.14

–

The value for FLU-D2 could not be calculated as a result of the rapid hydrolysis in plasma. Data represented as mean ± SD (n ¼ 3).

Figure 3. In vitro stabilities of FLU-D1 (A) and FLU-D2 (B) in phosphate buffer solutions of different pH values, plasma and brain homogenates. Data represented as mean ± SD (n ¼ 3).

D. Zheng et al.



6

Figure 4. bEnd.3 uptake (A) 80 nmol/ml FLU, FLU-D1and FLU-D2 at 37 C incubation for different times; (B) 20–160 nmol/ml FLU, FLU-D1and FLU-D2 at 37 C incubation for 0.5 h; (C) 80 nmol/ml FLU, FLU-D1 and FLU-D2 at 4 C, 37 C and NaN3 incubation for 0.5 h, respectively; L929 uptake (D) 80 nmol/ml FLU, FLU-D1 and FLU-D2 at 37 C and 4 C incubation for 0.5 h, respectively; (E) Cellular uptake of 80 nmol/ml FLU, FLU-D1and FLU-D2 on bEnd.3 cells and L929 cells at 37 C incubation for 0.5 h, respectively. Data represented as mean ± SD (n ¼ 3), **p50.01, ***p50.001.

In vitro cellular uptake studies Cellular uptake on bEnd.3 cells The uptake of FLU-D1 and FLU-D2 by bEnd.3 cells are found to be dependent on the incubation time within 2 h (Figure 4A). At each time point, the uptake amount of the derivatives was significantly higher than that of FLU. At 0.5 h, both FLU-D1 and FLU-D2 reached a peak concentration of 9.27 ± 0.60 and 13.49 ± 0.43 nmol/mg protein,

respectively. These concentrations are 60.2-fold and 87.6fold of FLU (0.15 ± 0.12 nmol/mg protein). As shown in Figure 4(B), the uptake of both FLU-D1 and FLU-D2 are also concentration-dependent. The intracellular drug concentrations in FLU-D1 and FLU-D2 groups showed a decline trend after an initial ascent, and they reached the peak concentration when the drug dosage was 80 nmol/ml. In contrast, FLU showed continued growth with no saturation tendency, and the uptake amount was significantly lower than

DOI: 10.3109/10717544.2014.954165


7

those of the other two groups. The uptake of FLU-D1 and FLU-D2 also displayed a significant difference between 37 and 4 C. To be specific, the intracellular concentration of FLU-D1 and FLU-D2 incubated at 37 C were 2.3 times and 1.6 times of those at 4 C (p50.01), which indicated the potential dependence of temperature (Figure 4C). Conversely, the uptake of FLU was slightly enhanced at 4 C, and the figure showed no substantial difference between 37 and 4 C. When cells were treated with NaN3, the uptake of FLU-D1 and FLU-D2 were 0.5 times and 0.4 times of those in control groups, which implied that the transport is potentially energy consuming (Figure 4C). The time-, concentration-, temperature- and energy-dependences suggested a process of active transport of FLU-D1 and FLU-D2 mediated by certain carriers on bEnd.3 cells.


Cellular uptake on L929 cells To further investigate the delivery efficiency of FLU-D1 and FLU-D2 in different cell lines, the uptake experiment was also performed using L929 cells at 37 and 4 C. Results are shown in Figure 4(D and E). On one hand, the uptake of FLU-D1, FLU-D2 and FLU showed no significant difference at 37 and 4 C, which might be explained that all of them were transported in a non-active pattern in L929 cell line. On the other hand, the uptake amount of FLU-D1 and FLU-D2 in bEnd.3 cells were 5.2 times and 1.6 times of those in L929 cells. However, no significant difference of the intracellular concentration of FLU was found between L929 and bEnd.3 cells. These results suggest a brain specific targeting property of FLU-D1 and FLU-D2 in vitro. Plasma and brain bioavailability in rats To determine the fate of our FLU derivatives in blood circulation and brain in vivo, plasma and brain concentrations of each drug were monitored by HPLC after drug administration in rats. Hydrolysis of FLU-D2 in both plasma and brain has been detected. As indicated in Figure 5(A) and Table 3, FLU-D2 showed a similar pharmacokinetic pattern with the parent FLU in plasma as no significant difference was found in their pharmacokinetic parameters. In contrast, the concentration of FLU-D1 in plasma was significantly lower than that of FLU throughout the whole time course, which might be caused by their different distribution patterns in vivo. In the case of brain tissue, both FLU-D2 and FLU-D1 showed significantly higher concentrations than FLU (Figure 5B). The maximal concentration (Cmax) of FLU-D2 was 12.1 times and 0.6 times higher than that of FLU and FLU-D1. The area under the curve (AUC0–t) of FLU-D2 was 4.6 times higher than that of FLU (Table 4). These results indicate that the modification of FLU-D2 can greatly improve drug bioavailability in the brain tissue, but has less such effect in the plasma. Tissue distribution study in rats To get more information about the brain-targeting efficiency of FLU-D1 and FLU-D2 in vivo, drugs concentrations in main organs were measured separately. We set the measurement time point at 4 h after i.v. administration, since the drug

Figure 5. Blood and brain concentration–time profiles of FLU, FLUD1and FLU-D2 following i.v. injection at a FLU-equivalent dose of 10 mg/kg in rats. The concentrations of FLU-D1 and FLU-D2 were converted to FLU equivalent. Data represented as mean ± SD (n ¼ 5).

Table 3. Pharmacokinetic parameters of FLU and its derivatives in plasma after i.v. administration to rats. Parameters

FLU

FLU-D1

FLU-D2

AUC0–t (mg/g h) 189.636 ± 16.139 1.567 ± 0.171 171.89 ± 21.974 Cmax (mg/g) 86.868 ± 3.697 2.049 ± 0.382 41.365 ± 2.647 Tmax (h) 0.083 0.083 0.05 T1/2 (h) 3.61 ± 0.422 0.549 ± 0.04 3.939 ± 0.58 CLZ (L/h/kg) 52.701 ± 4.42 6406 ± 661.327 58.103 ± 7.506 The concentrations of FLU-D1and FLU-D2 were converted to FLU equivalent. Data represented as mean ± SD (n ¼ 5).

Table 4. Pharmacokinetic parameters of FLU and its derivatives in brain after i.v. administration to rats. Parameters AUC0–t (mg/g h) Cmax (mg/g) RE CE

FLU 3.168 ± 0.134 2.43 ± 0.043 – –

FLU-D1

FLU-D2 a

25.73 ± 1.115 20.844 ± 0.672a 8.12 8.37

17.76 ± 0.883a 31.817 ± 2.308a 5.61 13.09

The concentrations of FLU-D1and FLU-D2 were converted to FLU equivalent, ap50.001 versus FLU. Data represented the mean ± SD (n ¼ 5).

concentration in the brain declined to the level which could hardly be measured after 4 h. In Figure 6(A), it is clear that FLU is mainly retained in the plasma, with a tiny proportion evenly distributed throughout main organs, implying a poor tissue-specific localization property in vivo. According to Figure 6(B), FLU-D1 underwent a rapid partition into the main organs after administration and was progressively


8

D. Zheng et al.


Figure 6. Biodistribution profiles of FLU (A), FLU-D1 (B) and FLU-D2 (C) following i.v. injection at a FLU-equivalent dose of 10 mg/kg in rats. The concentrations of FLU-D1 and FLU-D2 were converted to FLU equivalent. Data represented as mean ± SD (n ¼ 5).

eliminated over time. FLU-D1 appeared to be mainly accumulated in the lung, and maintained a prolonged high drug levels over the time course. The plasma concentration remarkably decreased due to its extensive distribution into other organs. All these findings indicated a significantly enhanced tissue penetration ability of FLU-D1, especially to the lung. In the case of FLU-D2, a considerable proportion of released parent drug (FLU) was found in both plasma and other organs. The drug level in plasma was shown significantly higher than that of FLU-D1 (Figure 6C). Their highly different distribution patterns might result from the hydrolysis of FLU-D2. In addition, after 0.5 h injection, the distribution pattern of FLU-D2 was much similar to that of FLU. The drug concentrations of the conjugates and parent drug in brain tissues at different time points were illustrated in Figure 7. The results showed that the concentrations of FLU-D2 and FLU-D1 in the brain were much higher than that of FLU at each time point, which indicated that these derivatives possessed much greater brain penetration ability than FLU did. The CE values, a parameter to evaluate the targeting efficiency of different organs, were calculated and displayed in Table 5. It is interesting that the highest CE value of FLU-D2 group was found in the brain (13.09), while for FLU-D1, the highest CE value appeared in the lung (13.66). All these results suggested that FLU-D2 has a much better

Figure 7. Brain concentrations of FLU, FLU-D1 and FLU-D2 at different times after i.v. injection in rats. The concentrations of FLU-D1 and FLU-D2 were converted to FLU equivalent. The brain concentrations of FLUD1 and FLU-D2 were significantly higher than that of FLU at each time point, and significant difference was marked only at 0.083 h. Data represented as mean ± SD (n ¼ 5), ***p50.001.

Table 5. CE of FLU-D1 and FLU-D2 in different organs after i.v. administration to rats.

CEFLU-D1 CEFLU-D2

Plasma

Brain

Lung

Kidney

Heart

Liver

Spleen

0.04 0.71

8.37 13.09

13.66 3.84

4.13 1.28

1.78 1.36

0.29 0.54

2.51 1.24

DOI: 10.3109/10717544.2014.954165


brain specificity than FLU or FLU-D1, which is consistent with the results of cellular uptake on bEnd.3 cells in vitro.

brought toxic effects on the cells since cell viability was always above 80%. Statistical analysis showed no significant differences of cell viability among the FLU-D1-, FLUD2- and FLU-treated groups, which suggested that the N,N-dimethylethanolamine structure has a good biocompatibility as a ligand. Nevertheless, long-term toxicity of these derivatives deserves to be further studied. Cellular uptake studies in bEnd.3 cells were performed to evaluate the in vitro delivery efficacy of the derivatives. A dramatic improvement of cellular uptake was detected. Indeed, comparing with FLU, the intracellular drug concentrations were enhanced by 59.2 times and 86.6 times with FLU-D1 and FLU-D2, respectively. More importantly, the time-, concentration-, temperatureand energy-dependent uptake of FLU-D1 and FLU-D2 implied a process of active transport in bEnd.3 cells. This change is probably owed to the introduction of the N,N-dimethylethanolamine ligand to FLU molecule. Comparisons of the uptake of FLU derivatives have also been made using L929 cells. The intracellular concentrations of drugs in L929 cells were much lower than that in bEnd.3 cells, which implies that the derivatives could selectively target to brain cells. As no substantial difference was found between 37 and 4 C conditions during the uptake of FLU-D1 or FLU-D2 in L929 cells, a passive diffusion transport process of the derivatives might be involved in this cell line. Here, for the first time, the N,N-dimethylethanolamine conjugates are demonstrated to mediate an active transport of FLU into brain endothelial cells. This provides a completely novel strategy for designing brain drug delivery systems. Our experiments also displayed that FLU-D2 was more efficient and more specific than FLU-D1, considering the delivery ability towards the brain cells, which suggested that the ester linkage was probably more suitable for brain delivery of FLU. Our studies in vivo showed that both conjugates could remarkably enhance brain delivery efficiencies of FLU. The highest brain concentrations of the derivatives appeared at the first sampling time point and then slowly decreased (Figure 5B). The Cmax of FLU-D2 in the brain was 13.09-fold and 1.5-fold of FLU and FLU-D1, respectively, suggesting an extremely high efficiency of brain transport for FLU-D2. The CE value of the brain was much higher than other tissues in FLU-D2 treated rats (Table 5), suggesting that FLU-D2 specifically accumulated in the brain. Furthermore, the in vivo metabolism of FLU-D2 in plasma and other tissues was much similar to that of FLU due to its easy hydrolysis, which implies a similar safety level as FLU. Nonetheless, detailed pharmacological profile needs to be further investigated concerning these derivatives. Despite the less lipophilic and high plasma-binding nature of FLU-D1 and FLU-D2 compared with FLU (Tables 1 and 2), the derivatives showed an efficient brain delivery efficacy in vivo. This further supports the hypothesis that an active transport mechanism is involved, which most probably mediated by the N,N-dimethylethanolamine ligand. Further detailed competitive inhibition assays and transporters searches have been planned as the next stage of this project. It is speculated that the linkage form plays an important role


Discussion As introduced, FLU, an effective neuroprotective agent in AD therapy, was greatly limited in clinical use due to its poor brain penetration ability (Imbimbo, 2009b). Various strategies have been experimented to overcome the low brain permeability of CNS therapeutic agents (Kasinathan et al., 2014; Mittal et al., 2014). Prodrug strategy, one of the medicinal chemistry drug delivery systems, has been investigated extensively in order to increase the targeting efficiency of drugs to the brain (Anderson, 1996). Previous work of our research group demonstrated that ethanolamine-related structure could effectively improve drug availability in the brain (Zhang et al., 2012). In this study, we introduced N,N-dimethylethanolamine ligand, with active transport function, to FLU and developed novel FLU conjugates. We showed that these derivatives targeted to the brain efficiently both in vitro and in vivo, and also were considerably safe in vitro. Both racemic FLU and its S- or R-isomer have been shown to be equipotent in lowering Ab42 levels in mouse brain and share the same mechanism of interaction with the g-secretase complex (Eriksen et al., 2003; Pignatello et al., 2007). Thus, we chose to perform the chemical derivatization using the racemic form of FLU in this study. The type of linkage between model drug and the ligand is crucial for creating potent prodrugs for brain delivery. Among the well-studied linkage types, ester and amide linkages are the most commonly used (Halmos et al., 1996). In this study, FLU-D1 and FLU-D2 were successfully synthesized by conjugating FLU with N,N-dimethylethylenediamine and N,N-dimethylethanolamine via amide bond and ester bond, respectively. The structures of final products were confirmed by 1H-NMR, 13C-NMR as well as HPLC-MS spectroscopy. A series of experiments were designed to investigate the different fates of the derivatives in vitro and in vivo. Our stability studies in PBS buffers and biological samples indicated that the amide linkage form FLU-D1 is quite stable under all given conditions (490% remaining percentage). This implies that no significant amounts of FLU would be released in the blood stream and would work in its intact form. In contrast, the ester bond in FLU-D2 was more susceptible to hydrolysis than FLU-D1, which suggested that FLU-D2 might undergo hydrolysis at its target sites and release FLU to exert therapeutic effect. bEnd.3 cells are an immortalized mouse brain endothelial cell line exhibiting endothelial properties. It is an ideal candidate as a model of the BBB because of its rapid growth, maintenance of blood brain barrier characteristics over repeated passages, formation of functional barriers and amenability to numerous molecular interventions (Brown et al., 2007). With these advantages, it was selected as the BBB model in this study to investigate the brain delivery properties of FLU derivatives in vitro. Cytotoxicity studies were performed using bEnd.3 cells and L929 cells to evaluate the safety of FLU-D1and FLU-D2. As illustrated in Figure 3, neither FLU-D1 nor FLU-D2

9

10

D. Zheng et al.

in brain uptake, and the ester bond might have higher affinity to the unknown transporters on the BBB than amide linkage (Halmos et al., 1996). This point is also supported by our FLU-D2 data. Besides the vastly improved brain availability of FLU-D2, the released parent FLU in target sites could also facilitate its potential use in AD therapy.


Conclusion Two N,N-dimethylethanolamine conjugated FLU derivatives (FLU-D1 and FLU-D2) were synthesized as a novel approach to deliver FLU into brain. In vitro cell viability experiments displayed a good biocompatibility of these novel conjugates. The significantly increased uptake of the FLU derivatives by bEnd.3 cells compared with that of FLU could be attributed to an active transport mediated by the introduction of the N,N-dimethylethanolamine ligand. This hypothesis is supported by the time-, temperature-, concentration- and energy- dependent uptake patterns of these derivatives. A remarkable enhanced brain bioavailability of these derivatives has also been detected after an intravenous injection to rats. Moreover, FLU-D2 could be easily hydrolyzed to the parent FLU when being distributed to the target sites. All these findings indicate that FLU-D2 is promising to be developed into a safe and effective therapeutic agent for Alzheimer’s disease.

Declaration of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article. This work was supported by National Basic Research Program of China (No. 2013CB932504) and the National Natural Science Foundation of China (No. 81130060).

References Ai J, Biazar E, Jafarpour M, et al. (2011). Nanotoxicology and nanoparticle safety in biomedical designs. Int J Nanomed 6:1117– 1127. Alam MI, Beg S, Samad A, et al. (2010). Strategy for effective brain drug delivery. Eur J Pharm Sci 40:385–403. Anderson BD. (1996). Prodrugs for improved CNS delivery. Adv Drug Deliv Rev 19:171–202. Brown RC, Morris AP, O’Neil RG. (2007). Tight junction protein expression and barrier properties of immortalized mouse brain microvessel endothelial cells. Brain Res 1130:17–30. Chen Q, Gong T, Liu J, et al. (2009). Synthesis, in vitro and in vivo characterization of glycosyl derivatives of ibuprofen as novel prodrugs for brain drug delivery. J Drug Target 17:318–28.


Eriksen JL, Sagi SA, Smith TE, et al. (2003). NSAIDs and enantiomers of flurbiprofen target g-secretase and lower Ab42 in vivo. J Clin Invest 112:440–9. Etminan M, Gill S, Samii A. (2003). Effect of non-steroidal antiinflammatory drugs on risk of Alzheimer’s disease: systematic review and meta-analysis of observational studies. BMJ 327:128. Gasparini L, Ongini E, Wenk G. (2004). Non-steroidal anti-inflammatory drugs (NSAIDs) in Alzheimer’s disease: old and new mechanisms of action. J Neurochem 91:521–36. Green RC, Schneider LS, Amato DA, et al. (2009). Effect of tarenflurbil on cognitive decline and activities of daily living in patients with mild Alzheimer disease. JAMA 302:2557–64. Halen PK, Chagti KK, Giridhar R, Yadav MR. (2007). Combining anticholinergic and anti-inflammatory activities into a single moiety: a novel approach to reduce gastrointestinal toxicity of ibuprofen and ketoprofen. Chem Biol Drug Des 70:450–5. Halmos T, Santarromana M, Antonakis K, Scherman D. (1996). Synthesis of glucose-chlorambucil derivatives and their recognition by the human GLUT1 glucose transporter. Eur J Pharm 318:477–84. Imbimbo BP. (2009a). An update on the efficacy of non-steroidal antiinflammatory drugs in Alzheimer’s disease. Exp Opin Investig Drugs 18:1147–68. Imbimbo BP. (2009b). Why did tarenflurbil fail in Alzheimer’s disease? J Alzheimer’s Dis 17:757–60. Kasinathan N, Jagani HV, Alex AT, et al. (2014). Strategies for drug delivery to the central nervous system by systemic route. Drug Deliv [Epub ahead of print] 1–15. Kukar T, Prescott S, Eriksen JL, et al. (2007). Chronic administration of R-flurbiprofen attenuates learning impairments in transgenic amyloid precursor protein mice. BMC Neurosci 8:54. Launer L, Breteler M, Hofman A, Stricker BC. (2002). Pharmacologic agents associated with a preventive effect on Alzheimer’s disease: a review of the epidemiologic evidence. Epidemiol Rev 24:248–68. Mittal D, Ali A, Md S, et al. (2014). Insights into direct nose to brain delivery: current status and future perspective. Drug Deliv 21:75–86. Parepally JMR, Mandula H, Smith QR. (2006). Brain uptake of nonsteroidal anti-inflammatory drugs: ibuprofen, flurbiprofen, and indomethacin. Pharm Res 23:873–81. Peretto I, Radaelli S, Parini C, et al. (2005). Synthesis and biological activity of flurbiprofen analogues as selective inhibitors of b-amyloid1-42 secretion. J Med Chem 48:5705–20. Pignatello R, Pant V, Salmaso S, et al. (2007). Flurbiprofen derivatives in Alzheimer’s disease: synthesis, pharmacokinetic and biological assessment of lipoamino acid prodrugs. Bioconjug Chem 19:349–57. Rautio J, Kumpulainen H, Heimbach T, et al. (2008). Prodrugs: design and clinical applications. Nat Rev Drug Discov 7:255–70. Song N, Li Y, Sun X, Qu F. (2004). Synthesis of (+/) ibuprofen sugar derivatives. Acta Pharmacol Sinica 39:105–109. Tammara VK, Narurkar MM, Crider AM, Khan MA. (1993). Synthesis and evaluation of morpholinoalkyl ester prodrugs of indomethacin and naproxen. Pharm Res 10:1191–9. Wilcock GK, Black SE, Hendrix SB, et al. (2008). Efficacy and safety of tarenflurbil in mild to moderate Alzheimer’s disease: a randomised phase II trial. Lancet Neurol 7:483–93. Zhang X, Liu X, Gong T, et al. (2012). In vitro and in vivo investigation of dexibuprofen derivatives for CNS delivery. Acta Pharmacol Sinica 33:279–88.

Synthesis and Antiproliferative and Metabolic Evaluations of Novel Securinine Derivatives.

Synthesis, in vitro and in vivo antitumor and antiviral activity of novel 1-substituted benzimidazole derivatives.

Improved Delivery Through Biological Membranes. XXIV. Synthesis, in Vitro Studies, and in Vivo Characterization of Brain-Specific and Sustained Progestin Delivery Systems.

Synthesis of novel 1,2,3-triazole derivatives of isoniazid and their in vitro and in vivo antimycobacterial activity evaluation.

Novel 4-aminoquinoline-pyrimidine based hybrids with improved in vitro and in vivo antimalarial activity.

Design, synthesis, and evaluations of antifungal activity of novel phenyl(2H-tetrazol-5-yl)methanamine derivatives.

Colloidal nanocarriers for the enhanced cutaneous delivery of naftifine: characterization studies and in vitro and in vivo evaluations.

Development of drug-in-adhesive patch formulations for transdermal delivery of fluoxetine: In vitro and in vivo evaluations.

Correlating In Vitro Splice Switching Activity With Systemic In Vivo Delivery Using Novel ZEN-modified Oligonucleotides.

Synthesis, in vitro and in vivo antitumor activity of symmetrical bis-Schiff base derivatives of isatin.

Identifying lipidic emulsomes for improved oxcarbazepine brain targeting: In vitro and rat in vivo studies.

Anti-CD20 Immunoglobulin G Radiolabeling with a 99mTc-Tricarbonyl Core: In Vitro and In Vivo Evaluations.

Synthesis and in-vivo hypolipidemic activity of some novel substituted phenyl isoxazol phenoxy acetic acid derivatives.

Design, synthesis and in vitro trypanocidal and leishmanicidal activities of novel semicarbazone derivatives.

Nanomiemgel--a novel drug delivery system for topical application--in vitro and in vivo evaluation.

Synthesis and in vitro antiproliferative evaluation of novel N-alkylated 6-isobutyl- and propyl pyrimidine derivatives.

Correction: Nanomiemgel--a novel drug delivery system for topical application--in vitro and in vivo evaluation.

Novel hydrophobin-coated docetaxel nanoparticles for intravenous delivery: in vitro characteristics and in vivo performance.

A novel peptide for efficient siRNA delivery in vitro and therapeutics in vivo.

Preparation of novel pirfenidone microspheres for lung-targeted delivery: in vitro and in vivo study.

Design, synthesis and in vitro evaluation of novel ursolic acid derivatives as potential anticancer agents.

Enkephalin analogues with N-phenyl-N-(piperidin-2-ylmethyl)propionamide derivatives: Synthesis and biological evaluations.

Controlled delivery of zoledronate improved bone formation locally in vivo.

Disparate effects of the prostaglandin synthesis inhibitors, meclofenamate, and flurbiprofen on monkey luteal tissue in vitro.