Janus PEG-Based Dendrimers for Use in Combination Therapy

Jan 10, 2013 - BA and PPA were attached via two different linkers (carbonate and ester, respectively) to promote differential drug release. The four d...
0 downloads 0 Views 4MB Size
Download PDF
Article pubs.acs.org/Biomac

Janus PEG-Based Dendrimers for Use in Combination Therapy: Controlled Multi-Drug Loading and Sequential Release Aaron L. Acton,† Cristina Fante,† Brian Flatley,† Stefano Burattini,† Ian W. Hamley,† Zuowei Wang,‡ Francesca Greco,*,§ and Wayne Hayes*,† †

Department of Chemistry, §Reading School of Pharmacy, and ‡Department of Mathematics, University of Reading, Whiteknights, Reading, RG6 6AD, United Kingdom S Supporting Information *

ABSTRACT: The increasing use of drug combinations to treat disease states, such as cancer, calls for improved delivery systems that are able to deliver multiple agents. Herein, we report a series of novel Janus dendrimers with potential for use in combination therapy. Different generations (first and second) of PEG-based dendrons containing two different “model drugs”, benzyl alcohol (BA) and 3-phenylpropionic acid (PPA), were synthesized. BA and PPA were attached via two different linkers (carbonate and ester, respectively) to promote differential drug release. The four dendrons were coupled together via (3 + 2) cycloaddition chemistries to afford four Janus dendrimers, which contained varying amounts and different ratios of BA and PPA, namely, (BA)2-G1-G1-(PPA)2, (BA)4-G2-G1-(PPA)2, (BA)2-G1-G2-(PPA)4, and (BA)4-G2-G2-(PPA)4. Release studies in plasma showed that the dendrimers provided sequential release of the two model drugs, with BA being released faster than PPA from all of the dendrons. The different dendrimers allowed delivery of increasing amounts (0.15−0.30 mM) and in exact molecular ratios (1:2; 2:1; 1:2; 2:2) of the two model drug compounds. The dendrimers were noncytotoxic (100% viability at 1 mg/mL) toward human umbilical vein endothelial cells (HUVEC) and nontoxic toward red blood cells, as confirmed by hemolysis studies. These studies demonstrate that these Janus PEG-based dendrimers offer great potential for the delivery of drugs via combination therapy.

■

INTRODUCTION The use of drug combinations is an approach used commonly in the clinic to treat a wide variety of diseases.1−3 For instance, cancer chemotherapy routinely involves administration of different therapeutic agents (e.g., in the case of breast cancer, the combination of cyclophosphamide, doxorubicin (Dox), or fluorouracil).4 Cocktails of drugs are also often employed in the treatment of HIV infection.5 The main advantage of combination therapy over monotherapy is the ability to hit different disease targets simultaneously, which results in increased activity and reduced toxicity.6,7 In recent years, several attempts have been made to deliver combination therapy via traditional drug delivery systems.8−10 Polymer−drug conjugates are drug delivery systems where a drug is covalently attached to a polymeric carrier to improve its therapeutic performance.11 Such systems have shown great promise for the delivery of a single drug, with 16 conjugates having progressed into clinical trials for cancer treatment.12 © 2013 American Chemical Society

Recent studies have applied this technology to the delivery of drug combinations and are showing promise, as the combination polymer conjugates display an increased activity when compared to the parent conjugates bearing a single agent.13−23 These studies on drug delivery systems that bear drug combinations are highlighting that many factors affect the efficacy of the potential treatment. The drug ratio plays a key role as studies carried out in liposomal systems containing drug combinations revealed that drug activity ranged from synergistic to antagonistic depending on this parameter.24 Furthermore, the rate of drug release (both in terms of the order of drug release and the relative dispersal rates) was found to be an important factor for activity.19,25 As the field of polymer−drug conjugates for combination therapy continues to mature, there is a need to develop versatile Received: December 7, 2012 Revised: January 9, 2013 Published: January 10, 2013 564

dx.doi.org/10.1021/bm301881h | Biomacromolecules 2013, 14, 564−574

Biomacromolecules

Article

Scheme 1. Schematic Representation of the Four Dendrons and the Corresponding Dendrimers that Feature Varying Ratios of the Model Drug Compounds BA and PPA

from Lancaster Synthesis and were used as received without purification. Solvents were purchased from Fisher Scientific except for ethyl acetate and hexane which were purchased from Sigma Aldrich. All solvents were used as supplied with the exception of THF that was distilled under argon from sodium and benzophenone prior to use. Fisher Scientific Silica 60A (particle size 35−70 μm) was used to perform column chromatography. Thin-layer chromatography (TLC) was performed on aluminum sheets coated with Merck 5735 Kieselgel 60 F254. Developed plates were air-dried and stained using a potassium permanganate solution. Characterization. 1H (400 MHz) and 13C (100 MHz) NMR spectra were recorded on a Bruker Nano 400 (9.39 T) or Bruker DPX 400 (9.39 T) instrument with tetramethylsilane (TMS) used as the internal standard. Fourier transform infrared spectroscopy (FTIR) spectra were recorded on a PerkinElmer Spectrum FT-IR in transmission mode, all samples were analyzed directly using a diamond ATR sampling accessory. Matrix-assisted laser desorption-ionization time-of-flight mass spectra (MALDI-TOF) were obtained using a Bruker Daltonics Ultraflex 1 spectrometer operating in reflection mode. The instrument was calibrated using a standard peptide mixture (Bruker Daltonics, Calibration standard II) using 2,5-dihydroxybenzoic acid (DHB) as the matrix. A typical sample preparation is described as follows: 3 μL of a solution of the analyte in acetonitrile (5 mg/mL) was combined with 3 μL of the freshly prepared matrix (20 mg/mL in 1:1 acetonitrile/water with 1% TFA), an aliquot (1 μL) was spotted onto a ground steel MALDI target plate and left to dry in air prior to analysis. Mn, Mw, and PDI were calculated from the MALDI spectra as reported40 by Roeraade et al. Mass spectrometry was conducted using ThermoFisher Scientific Orbitrap XL LCMS. The sample was introduced by liquid chromatography and sample ionization was achieved by electrospray ionization (ESI). High performance liquid chromatography (HPLC) was performed using an Agilent 1100 featuring a variable wavelength UV detector. A Phenomenex Aqua 5μ C18 125A (150 × 4.60 mm 5 μ micrometer) column fitted with a Phenomenex Luna 3μ C18(2) (50 × 4.60 mm 3 μ micrometer) guard column was used. Samples were analyzed using gradient elution consisting of HPLC grade (A) 1% acetic acid in water and (B) 1% acetic acid in acetonitrile following 80% A/20% B 0−5 min, 70% A/30% B 5−10 min, 20% A/80% B 10−20 min. A flow rate of 1 mL/min at ambient temperature was used, and chromatograms were monitored using a UV detector operating at 258 nm. Linker Stability Studies of Model Compounds 1−4. For hydrolysis in buffer, model compounds 1−4 were dissolved in 0.1 M PBS (pH 7.4) or 0.1 M citrate buffer (pH 5) in deuterium oxide at a concentration of 5 mg/mL. The mixtures were vortexed for 2 min and incubated in a water bath at 37 °C. At regular time intervals, aliquots (0.4 mL) were taken and analyzed by 1H NMR spectroscopic analysis. In contrast, for hydrolysis in human plasma (Sigma Aldrich, U.K.), compounds 1−4 were dissolved in plasma (reconstituted with deuterium oxide) at a concentration of 50 mg/mL. The samples were incubated in a

polymeric carriers that are able to exert control over drug ratios and release rates. In recent years, the use of dendrimers in biomedical applications has grown as a result of their monodisperse nature, low toxicity, water solubility, encapsulation ability, and large number of functionalizable peripheral groups.26 Janus dendrimers27−29 (also referred to as “bow-tie” dendrimers30), which are comprised of two heterogeneous segments, have been designed for drug delivery applications.31−33 As a result of their unique chemical structure and the convergent methodology used to prepare them, Janus dendrimers are intrinsically suitable for the delivery of more than one drug. Within this context, we report the design and synthesis of novel Janus PEG-based dendrimers for potential use in combination therapy. In this approach, the degree of drug loading can be tailored by varying the dendron generation and hence the number of drug loading sites (i.e., two or four drugs for first (G1) or second generation (G2), respectively). PEG was used as the polymeric backbone because of its FDA approval and proven biocompatibility (as demonstrated by commercially available polymer−protein conjugates).34 In recent years, dendritic or highly branched PEG-based polymers have been used as effective drug delivery systems.35−37 To generate the desired dendritic branched architectures from linear PEG units, a tertiary amine branching motif (as found in biocompatible PAMAM dendrimers38) was employed. To prove applicability to combination therapy, benzyl alcohol (BA) and 3-phenylpropionic acid (PPA) were selected for use as model drug compounds. Four dendrons were synthesized which comprised either azide or alkyne moieties at the core and carried different amounts of BA and PPA, namely, (BA)2-G1N3, (PPA)2-G1alk, (BA)4-G2N3, and (PPA)4-G2alk. Conjugation of the various dendrons via [3 + 2] cycloaddition chemistries39 afforded four Janus dendrimers (Scheme 1), (BA)2-G1-G1-(PPA)2, (BA)4-G2-G1-(PPA)2, (BA)2-G1-G2-(PPA)4, and (BA)4-G2-G2-(PPA)4, which differ in terms of either model drug loading capacity or ratio. The effectiveness of the dendrimers for use in combination therapy was also probed. In particular, their ability to deliver the BA and PPA payload was tested by release studies under physiological conditions. Finally, the biocompatibility of the dendrimers and their metabolites were assessed against the healthy human cell line (HUVEC) and toxicity toward red blood cells using hemolysis studies.

■

EXPERIMENTAL SECTION

Materials. All chemical reagents were purchased from Sigma Aldrich with the exception of 4-dimethylaminopyridine which was purchased 565

dx.doi.org/10.1021/bm301881h | Biomacromolecules 2013, 14, 564−574

Biomacromolecules

Article

water bath at 37 °C and at regular time intervals aliquots (50 μL) taken and diluted with deuterium oxide (400 μL), followed by 1H NMR spectroscopic analysis. On the basis of peak area, the consumption of 1−4 over time was estimated, and the half-life of each compound in different media calculated using linear regression analysis. Analysis was repeated in triplicate for pH studies and duplicate for plasma stability studies. Determination of Free Model Drug Content. The percentage of free benzyl alcohol (BA) and 3-phenylpropionic acid (PPA) present in the four dendrimer samples was evaluated using HPLC, calibrated with BA and PPA (0.001−1 mg/mL). The four Janus dendrimers were dissolved in acetonitrile (1 mg/mL) and the percentage free model drug determined based upon peak integrations. Model Drug Release Studies from the Janus Dendrimers. Release studies were conducted by dissolving each dendrimer in methanol (20 mg/mL), followed by dilution to a concentration of 1 mg/mL in either PBS (pH 7.4) or citrate (pH 5) buffers. The samples were incubated in a water bath at 37 °C and at regular time intervals aliquots (100 μL) taken and analyzed by HPLC. Release studies in plasma involved dissolving each dendrimer in methanol (20 mg/mL), followed by dilution in human plasma to a molar concentration of (0.248 mM). The samples were incubated in a water bath at 37 °C and at regular time intervals aliquots (50 μL) taken and added to acetonitrile (100 μL). The resulting solutions were centrifuged for 2 min and the supernatants filtered through a microfilter with a pore size of 0.45 μm (Millipore PTFE filter) and analyzed by HPLC. Release studies were repeated in triplicate and upon the basis of peak area, the release of benzyl alcohol and 3-phenylpropionic acid from each dendrimer was determined. Cytotoxicity of Dendrimers. Human umbilical vein endothelial cells (HUVEC; Lonza, U.K.) were cultured in EGM-2 (Lonza, U.K.) in standard tissue culture conditions (37 °C, humidified 5% CO2 atmosphere). HUVECs were seeded (6 × 103 cells/mL) in a 96-well plate and incubated for 16 h. Fresh medium containing serial dilutions of the dendrimers (100 μL of (OH)2-G1-G1-(OH)2, (OH)2-G1-G2(OH)4, (OH)4-G2-G1-(OH)2, (OH)4-G2-G2-(OH)4, which had been filtered through a 0.2 μm sterilized filter) was added to give a final concentration in the range of 0−1 mg/mL. HUVECs were incubated with the treatments for 72 h, after which time the treatments were removed, the cells washed once with PBS (100 μL) and incubated with trypsin-EDTA (50 μL; Lonza, U.K.) for 5 min. Fresh medium was added (50 μL), an aliquot of the cell suspension (20 μL) was diluted 1:1 with tryptan blue (Sigma Aldrich, U.K.) and viable HUVECs were counted using a hemocytometer. The experiment was repeated in triplicate and cell viability was expressed as % of control (medium without treatment added) based on the average cell concentration of four wells for each experiment. Evaluation of Red Blood Cell (RBC) Lysis. Blood was obtained from male Wistar rats (p > 24; Harlan, U.K.) after death by severed spine. RBCs were obtained by centrifuging the blood three times in chilled (4 °C) PBS for 10 min. The pellet was resuspended in PBS to give a 6% w/v solution. A 96-well plate was prepared with serial dilutions of the dendrimers ((OH)2-G1-G1-(OH)2, (OH)2-G1-G2(OH)4, (OH)4-G2-G1-(OH)2, (OH)4-G2-G2-(OH)4) 0−2 mg/ mL,100 μL) in PBS. The RBC suspension (100 μL) was added to the dendrimer solutions and the plate was incubated for 24 h at 37 °C. The plate was then centrifuged for 10 min and the supernatants (150 μL) were transferred into a new plate. Cell lysis was assessed by measuring the absorbance of the released hemoglobin at 570 nm using a plate reader (Benchmark Bio-Rad Microplate Reader). The detergent Triton X-100 (1% v/v) and PBS were used as a positive and negative control, to produce 100 and 0% RBC lysis, respectively. RBC lysis was expressed as percentage of the control. General Procedure for Dendrimer Synthesis. Dendrimers (BA)2-G1-G1-(PPA)2, (BA)2-G1-G2-(PPA)4, (BA)4-G2-G1-(PPA)2, and (BA)4-G2-G2-(PPA)4 were synthesized by one of two methods. Method 1: The corresponding azide (1 equiv) and corresponding alkyne (1 equiv) were dissolved in deuterated acetone (1 mL). To this solution was added a catalytic quantity of copper(I) iodide (ca. 10 mg) and DIPEA (1 equiv). The resulting mixture was stirred under argon

at room temperature overnight. The solvent was removed and the crude product purified by preparative TLC to afford the corresponding dendrimer. Method 2: The corresponding azide (1 equiv) and corresponding alkyne (1 equiv) were dissolved in dry THF (5 mL). To this solution was added a catalytic quantity of copper(I) iodide (ca. 10 mg) and DIPEA (1 equiv). The resulting mixture was stirred under argon at room temperature overnight. The solvent was removed and the crude product purified by column chromatography. Column fractions containing product were combined, diluted with CH2Cl2 (100 mL), washed with water (1 × 20 mL), brine (1 × 20 mL), dried over MgSO4, filtered, and concentrated in vacuo to afford the corresponding dendrimer. Synthesis of (BA)2-G1-G1-(PPA)2. Method 1: (BA)2-G1N3 (0.103 g, 0.094 mmol), (PPA)2-G1alk (0.100 g, 0.094 mmol), copper(I) iodide (catalytic quantity), and DIPEA (0.017 mL, 0.094 mmol) in acetone (1 mL). Purification by preparative TLC (5% MeOH/CH2Cl2) afforded dendrimer (BA)2G1-G1(PPA)2 (0.093 g, 48%, Rf = 0.25) as a colorless oil. 1H NMR (400 MHz, CDCl3) δH 1.55 (2H, quin, J = 7.5 Hz, NHCH2CH2CH2CH2), 1.88−1.96 (4H, m, 1 × NHCH2CH2CH2CH2 + 1 × CH2CH2CH2NH), 2.60−2.77 (14H, m, 2 × NHCH2CH2N, 2 × OCH2CH2N, 2 × ArCH2CH2, 1 × CH2CH2CH2NH), 2.85 (4H, t, J = 6.0 Hz, 2 × NCH2CH2OCOO), 2.93 (4H, t, J = 8.0 Hz, 2 × ArCH2CH2), 3.28−3.36 (8H, m, 1 × CH2CH2CH2NH, 2 × NHCH2CH2N, 1 × NHCH2CH2CH2CH2), 3.46−3.80 (80H, m, 20 × OCH2CH2O), 3.95−3.99 (8H, m, 4 × CH2O), 4.09 (4H, t, J = 6.0 Hz, 2 × OCH2CH2N), 4.17 (4H, t, J = 6.0 Hz, 2 × NCH 2 CH 2 OCOO), 4.35 (2H, t, J = 7.0 Hz, NHCH2CH2CH2CH2), 5.13 (4H, s, 2 × CH2Ar), 7.07−7.37 (24H, m, 20 × ArH, 4 × NH), 7.44 (1H, s, triazoleH) ppm; 13C NMR (100 MHz, CDCl3) δc 22.9, 26.7, 27.6, 29.2, 30.8, 35.7, 36.7, 37.9, 38.0, 49.6, 52.7, 52.8, 53.8, 53.9, 62.5, 65.9, 69. 6, 70.1, 70.2, 70.3, 70.3, 70.4, 70.5, 70.8, 70.9, 70.9, 126.3, 128.2, 128.3, 128.4, 128.5, 128.6, 135.2, 140.4, 155.0, 169.9, 170.1, 172.7; FTIR (ATR) ν 3348 (N−H), 2869 (C−H), 1737 (CO), 1668 (CO), 1532, 1455, 1347, 1249, 1098 cm−1; MALDI-TOF MS = 2016.0 ± n × 44 m/z; Mn = 2034.6, Mw = 2041.6, PDI = 1.003. Synthesis of (BA)4-G2-G2-(PPA)4. Method 2: (BA)4-G2N3 (0.1 g, 0.036 mmol), (PPA)4-G2alk (0.095 g, 0.036 mmol), copper(I) iodide (catalytic quantity), and DIPEA (0.0062 mL, 0.036 mmol) in dry THF (5 mL). Purification by column chromatography (MeCN/MeOH/ NH4OH, 10:2:1) afforded dendrimer (BA)4-G2-G2-(PPA)4 (0.02 g, 10%, Rf = 0.39) as a colorless oil. 1H NMR (400 MHz, CDCl3) δH 1.56 (2H, quin, J = 7.5 Hz, NHCH2CH2CH2CH2), 1.88−1.96 (4H, m, NHCH2CH2CH2CH2 + CH2CH2CH2NH), 2.60−2.77 (36H, m, 10 × NHCH2CH2N, 4 × OCH2CH2N, 4 × ArCH2CH2, 1 × CH2CH2CH2NH), 2.85 (8H, t, J = 6.0 Hz, 4 × NCH2CH2OCOO), 2.93 (8H, t, J = 8.0 Hz, 4 × ArCH2CH2), 3.28−3.37 (24H, m, 1 × CH2CH2CH2NH, 10 × NHCH2CH2N, 1 × NHCH2CH2CH2CH2), 3.45−3.82 (240H, m, 60 × OCH2CH2O), 3.95−3.99 (24H, m, 12 × CH2O), 4.09 (8H, t, J = 6.0 Hz, 4 × OCH2CH2N), 4.17 (8H, t, J = 6.0 Hz, 4 × NCH 2 CH 2 OCOO), 4.35 (2H, t, J = 7.0 Hz, NHCH2CH2CH2CH2), 5.14 (8H, s, 4 × CH2Ar), 7.10−7.38 (52H, m, 40 × ArH, 12 × NH), 7.43 (1H, s, triazoleH) ppm; 13C NMR (100 MHz, CDCl3) δc 22.9, 26.7, 27.6, 29.7, 30.9, 35.8, 36.7, 36.9, 37.9, 38.1, 49.6, 52.7, 52.8, 53.6, 53.8, 53.9, 62.5, 66.0, 69.6, 70.3, 70.4, 70.5 (×3), 70.8 (×2), 77.3, 126.3, 128.3 (×2), 128.5 (×2), 128.6, 135.2, 140.4, 155.1, 169.8, 170.0, 170.1 (×2), 172.8 ppm. FTIR (ATR) ν 3345 (N−H), 2868 (C−H), 1738 (CO), 1669 (CO), 1531, 1455, 1348, 1250, 1097 cm−1; MALDI-TOF MS = 5467.2 ± n × 44 m/z; Mn = 5444.9, Mw = 5454.9, PDI = 1.002. Synthesis of (BA)2-G1-G2-(PPA)4. Method 2: (BA)2-G1N3 (0.039 g, 0.038 mmol), (PPA)4-G2alk (0.1 g, 0.038 mmol), copper(I) iodide (catalytic quantity), and DIPEA (0.0066 mL, 0.038 mmol) in dry THF (5 mL). Purification by column chromatography (MeCN/MeOH/ NH4OH, 10:2:1) afforded dendrimer (BA)2-G1-G2-(PPA)4 (0.053 g, 36%, Rf = 0.11) as a colorless oil. 1H NMR (400 MHz, CDCl3) δH 1.56 (2H, quin, J = 7.5 Hz, NHCH2CH2CH2CH2), 1.88−1.96 (4H, m, NHCH2CH2CH2CH2 + CH2CH2CH2NH), 2.60−2.77 (30H, m, 6 × NHCH2CH2N, 4 × OCH2CH2N, 4 × ArCH2CH2, 566

dx.doi.org/10.1021/bm301881h | Biomacromolecules 2013, 14, 564−574

Biomacromolecules

Article

1 × CH2CH2CH2NH), 2.85 (4H, t, J = 6.0 Hz, 2 × NCH2CH2OC OO), 2.93 (8H, t, J = 8.0 Hz, 4 × ArCH2CH2), 3.27−3.38 (16H, m, 1 × CH2CH2CH2NH, 6 × NHCH2CH2N, 1 × NHCH2CH2CH2CH2), 3.45−3.82 (160H, m, 40 × OCH2CH2O), 3.95−3.99 (16H, m, 8 × CH2O), 4.09 (8H, t, J = 6.0 Hz, 4 × OCH2CH2N), 4.17 (4H, t, J = 6.0 Hz, 2 × NCH2CH2OCOO), 4.35 (2H, t, J = 7.0 Hz, NHCH2CH2CH2CH2), 5.14 (4H, s, 2 × CH2Ar), 7.08−7.37 (38H, m, 30 × ArH, 8 × NH), 7.43 (1H, s, triazoleH) ppm; 13C NMR (100 MHz, CDCl3) δc 22.9, 26.7, 27.6, 29.2, 29.7, 30.9, 35.8, 36.7, 36.9, 37.9, 38.0, 49.6, 52.7, 52.8, 53.6, 53.9 (×2), 62.5, 66.0, 69.6, 70.3, 70.8, 70.9 (×3), 121.2, 126.3, 128.3 (×2), 128.5 (×2), 128.6, 135.2, 140.4, 147.1, 155.1, 169.8, 170.0, 170.1 (×2), 172.8 ppm; FTIR (ATR) ν 3346 (N−H), 2867 (C−H), 1733 (CO), 1669 (CO), 1531, 1454, 1348, 1250, 1098 cm−1; MALDI-TOF MS = 3829.7 ± n × 44 m/z; Mn = 3841.3, Mw = 3848.9, PDI = 1.002. Synthesis of (BA)4-G2-G1-(PPA)2. Method 2: (BA)4-G2N3 (0.15 g, 0.054 mmol), (PPA)2-G1alk (0.054 g, 0.054 mmol), copper(I) iodide (catalytic quantity), and DIPEA (0.0094 mL, 0.054 mmol) in dry THF (5 mL). Purification by column chromatography (MeCN/ MeOH/NH4OH, 10:2:0.5 → 10:2:1) afforded dendrimer (BA)4-G2G1-(PPA)2 (0.140 g, 69%, Rf = 0.125 (10:2:1)) as a clear yellow oil. 1 H NMR (400 MHz, CDCl3) δH 1.55 (2H, quin, J = 7.5 Hz, NHCH2CH2CH2CH2), 1.88−1.96 (4H, m, NHCH2CH2CH2CH2 + CH2CH2CH2NH), 2.60−2.77 (22H, m, 6 × NHCH2CH2N, 2 × OCH2CH2N, 2 × ArCH2CH2, 1 × CH2CH2CH2NH), 2.85 (8H, t, J = 6.0 Hz, 4 × NCH2CH2OCOO), 2.93 (4H, t, J = 8.0 Hz, 2 × ArCH2CH2), 3.28−3.37 (16H, m, 1 × CH2CH2CH2NH, 6 × NHCH2CH2N, 1 × NHCH2CH2CH2CH2), 3.45−3.82 (160H, m, 40 × OCH2CH2O), 3.95−3.99 (16H, m, 8 × CH2O), 4.09 (4H, t, J = 6.0 Hz, 2 × OCH2CH2N), 4.17 (8H, t, J = 6.0 Hz, 4 × NCH2CH2OCOO), 4.35 (2H, t, J = 7.0 Hz, NHCH2CH2CH2CH2), 5.14 (8H, s, 4 × CH2Ar), 7.00−7.37 (38H, m, 30 × ArH, 8 × NH), 7.41 (1H, s, triazoleH) ppm; 13C NMR (100 MHz, CDCl3) δc 22.9, 26.7, 27.6, 29.3, 30.9, 35.9, 36.7, 36.9, 37.9, 38.1, 52.8 (×2), 53.7, 53.9, 54.0, 62.6, 66.0, 69.6, 70.3, 70.6, 70.9 (×2), 126.3, 128.3 (×2), 128.5, 128.6 (×2), 135.2, 140.4, 147.4, 155.1, 169.8, 169.9, 170.0, 170.1, 172.7 ppm; FTIR (ATR) ν 3348 (N−H), 2869 (C−H), 1742 (CO), 1663 (CO), 1533, 1455, 1347, 1249, 1095 cm−1; MALDI-TOF MS = 3742.8 ± n × 44 m/z; Mn = 3714.1, Mw = 3721.1, PDI = 1.002.

compounds were designed to simulate the chemical environment of different linkages that can be formed between PEG and a tertiary amine branching motif. The stability of model compounds 1−4 was tested under various physiological conditions (pH 5, 7.4, and plasma at 37 °C in vitro) and monitored by 1 H NMR spectroscopic analysis (see Table 1 and SI) Table 1. Degradation Half-Lives of Model Compounds compound

linker

pH 7.4 t1/2 (h)

pH 5 t1/2 (h)

plasma t1/2 (h)

1 2 3 4

amide ester carbamate carbonate

>120 5.8 >120 88.9

>120 7.9 >120 103.4

>120 2.7 >120 11.0

Stability studies (Table 1) showed the order of stability of the four linkers to be amide/carbamate > carbonate > ester, with amide and carbamate linkers stable to degradation up to 120 h under any of the conditions tested. For polymer−drug conjugates to act as effective delivery systems, the polymeric architecture must be stable during circulation in vivo.42 Consequently, amide or carbamate linkages offer the required stability for use as branching linkages between PEG and the tertiary amine unit. Ester and carbonate linkages exhibited the fastest hydrolysis in plasma (2 t1/2 = 2.7 h and 4 t1/2 = 11.0 h), and was attributed to the action of hydrolayses and proteases present within the plasma.43,44 Based on these preliminary results, ester and carbonate linkages were selected as the linkers between the model agents and dendritic structure, as such bonds displayed different hydrolysis profiles in plasma, which could be exploited to effect the desired controlled release and/or sequential release. Synthesis and Characterization of Dendrons Loaded with Model Drugs. Based on the linker stability studies, dendrons (BA) 2-G1 N3, (PPA) 2 -G1 alk , (BA) 4 -G2 N3 , and (PPA)4-G2alk were synthesized, which incorporate stable amide branching units and the two different model drugs (BA and PPA) attached to the periphery either through carbonate or ester linkages (Schemes 2 and 3). Synthesis of (BA)2-G1N3, (PPA)2-G1alk. Commercially available dicarboxy PEG-600 (5) was converted initially to the corresponding dimethyl ester (6), followed by addition of a single equivalent of base and the heterobifunctional PEG 7 was then isolated via column chromatography. The free acid group was coupled to either an azide-amine45 or alkyne-amine46 substrate (prepared as reported in the literature, see SI) via activation of the carbonyl center in the form of the corresponding acid chloride to obtain azide (8) or alkyne (9) functionalized PEG. The methyl ester terminus was coupled to a branched amine diol that was prepared using phthalimide chemistries47 (see SI) to yield the desired first generation dendrons (OH)2-G1N3 and (OH)2-G1alk containing two terminal hydroxyl groups. Conjugation of benzyl alcohol to the free hydroxyl groups of (OH)2-G1N3 was achieved through a carbonate intermediate48 (see SI) to afford the first generation dendron (BA)2-G1N3 loaded with two benzyl alcohol groups attached through carbonate linkages. A total of 2 equiv of 3-phenylpropionic acid were attached to (OH)2-G1alk through ester linkages using the Steglich esterification procedure49 to generate the first generation dendron (PPA)2-G1alk that features an alkyne focal point. Synthesis of (BA)4-G2N3 and (PPA)4-G2alk. To generate second generation dendrons that feature either alkyne or azide focal points in conjunction with BA or PPA loadings, the following strategy was employed (see Scheme 3). A total of

■

RESULTS AND DISCUSSION This study has afforded a new family of dendrimers for the delivery of two agents in exact ratios and with varied rates of release. Therefore, we initially explored the stability of different chemical linkages that would form the dendritic delivery systems. Even though the stability of esters, amides, carbonates, and carbamates are widely described, bond stability is strongly affected by the chemical environment surrounding the bond (e.g., Greenwald and co-workers have demonstrated that increasing steric hindrance around an ester, carbonate, and carbamate bond improves pro-drug stability).41 Consequently, a comprehensive study was first undertaken to identify chemical linkers that could (i) form a stable branching linker between the tertiary amine unit and PEG and (ii) degrade at varying rates enabling differential release of two model drugs. Linker Stability Studies under Physiological Conditions. Four model compounds 1−4 (Figure 1), featuring four different chemical linkages, amide, ester, carbonate, and carbamate, were synthesized (see Supporting Information (SI)). The model

Figure 1. Chemical structures of model compounds 1−4. 567

dx.doi.org/10.1021/bm301881h | Biomacromolecules 2013, 14, 564−574

Biomacromolecules

Article

Scheme 2. Synthesis of First Generation Dendrons with Model Drugs Attached through Carbonate (BA)2-G1N3 or Ester Linkages (PPA)2-G1alk

Scheme 3. Synthesis of Second Generation Dendrons with Four Model Drugs Attached Either through Carbonate (BA)4-G2N3 or Ester Linkages (PPA)4-G2alk

568

dx.doi.org/10.1021/bm301881h | Biomacromolecules 2013, 14, 564−574

Biomacromolecules

Article

2 equiv of heterofunctional PEG 7 were first reacted with mono Boc-protected TREN50 to afford 10. Subsequent deprotection afforded the amine 11, which was then coupled to either the carboxy groups of 12 or 13, respectively, in order to generate PEG-dendrons, each featuring two PEG methyl-ester end groups and either the desired azide (14) or alkyne (15) focal point moieties. Transesterification of the branched amine diol with the methyl ester residues of 14 or 15 using zirconium tertbutoxide as the catalyst51 afforded second generation dendrons, (OH)4-G2N3 and (OH)4-G2alk, each comprised of four hydroxyl groups at the periphery and either azide or alkyne moieties at the focal point. The four benzyl alcohol groups were then conjugated to (OH)4-G2N3 through carbonate linkages to produce (BA)4-G2N3, and four 3-phenylpropionic acid groups conjugated to (OH)4-G2alk through ester linkages to afford (PPA)4-G2alk. Overall, a total of four dendrons containing 2 or 4 molecules of BA or PPA were successfully synthesized: (BA)2-G1N3, (PPA)2-G1alk, (BA)4-G2N3, and (PPA)4-G2alk. All of the dendrons were fully characterized by 1H and 13C NMR and IR spectroscopy and MALDI-TOF mass spectrometry. The concept of these dendritic delivery systems is to enable combined delivery of two different model drugs in precise ratios. It is, therefore, essential that 100% drug loading is achieved during synthesis of the four dendrons. Drug loading was confirmed by 1H NMR spectroscopic analysis as exemplified in Figure 2 for (BA)2-G1N3 and (PPA)4-G2alk (see SI for (PPA)2G1alk and (BA)4-G2N3). In particular, we observed distinct downfield shifts of the triplet resonance at 3.58 ppm assigned to CH2OH (highlighted B in Figure 2a) to 4.15 and 4.09 ppm for conjugation of BA and PPA, respectively. In addition, in the case of (BA)2-G1N3, a singlet resonance with an integral of four protons was observed at 5.12 ppm and assigned to the benzylic protons of BA, revealing that the desired carbonate conjugation was complete (see resonance E in Figure 2b). In the case of (PPA)4-G2, the resonances observed at 2.63 and 2.93 ppm, respectively, corresponding to methylene protons of PPA, both possessed integrals of eight protons (see highlighted resonances E and F in Figure 2d), which demonstrate effective ester formation for this dendron. Synthesis and Characterization of Janus Dendrimers. Different combinations of first and second generation dendrons were coupled together via (3 + 2) cycloaddition reactions. In total, four dendrimers were produced: (BA)2-G1-G1-(PPA)2, (BA)4-G2-G1-(PPA)2, (BA)2-G1-G2-(PPA)4, and (BA)4-G2G2-(PPA)4, which differ in terms of model drug loading capacity (i.e., total loading 4, 6, or 8 model drug molecules per polymer) or model drug ratio (i.e., the BA to PPA ratio was 1:1, 1:2, or 2:1). Characterization data detailing the successful synthesis of the dendrimers is exemplified below for coupling of (BA)2-G1 and (PPA)4-G2 to produce (BA)2-G1-G2-(PPA)4 (please refer to SI for characterization of other dendrimers). 1 H NMR spectroscopy (Scheme 4) demonstrated that coupling of (BA)2-G1N3 and (PPA)4-G2alk with retained model drug loading was successful. Triazole formation was confirmed by the appearance of the characteristic singlet at 7.45 ppm,52,53 which correlated with consumption of the alkyne of (PPA)4-G2alk, as indicated by the disappearance of the terminal alkyne proton resonance at 2.01 ppm and downfield shift of the CH2 alkyne proton resonances at 2.25 ppm to a triplet at 4.40 ppm. Furthermore, consumption of the azide of (BA)2-G1N3 was evident by the downfield shift of the resonance at 1.6 ppm, which corresponded to CH2N3 (Scheme 4b) to 3.3 ppm upon

Figure 2. Example of 1H NMR spectroscopic analysis that confirmed complete model drug loading onto the PEG dendrons: (a) (OH)2G1N3, (b) (BA)2-G1N3, (c) (OH)4-G2alk, (d) (PPA)4-G2alk.

triazole formation. Finally, the model drug loading of the resultant dendrimers was determined by assessment of the triplet resonances at 4.09 and 4.15 ppm, respectively; proton integrals of 8 and 4 revealed the presence of PPA and BA in the designed 2:1 ratio for (BA)2-G1-G2-(PPA)4. Comparable analyses were conducted for the other dendrimers within this series (see SI) FTIR spectroscopic analysis of (BA) 2-G1-G2-(PPA) 4 (Figure 3) further supported successful conjugation of the dendrons via complete disappearance of the azide absorption band (ca. 2096 cm−1) observed in (BA)2-G1N3. MALDI-TOF mass spectrometry also provided evidence that supported the desired coupling of (BA)2-G1N3 and (PPA)4G2alk to form (BA)2-G1-G2-(PPA)4, as highlighted in Figure 4. Addition of (BA)2-G1N3 (1069.5 m/z) to (PPA)4-G2alk (2758.4 m/z) resulted in a theoretical mass of 3827.9 m/z for (BA)2-G1-G2-(PPA)4, which is in exact agreement with the experimental mass obtained 3829.7 m/z (with an isotope peak evident at 3827.8 m/z). The characterization data reported demonstrate successful synthesis of four Janus dendrimers conjugates ((BA)2-G1-G1(PPA)2, (BA)4-G2-G1-(PPA)2, (BA)2-G1-G2-(PPA)4, and (BA)4-G2-G2-(PPA)4), which differed in terms of drug loading capacity or loading ratio. The structural characteristics of each dendrimer are summarized in Table 2. 569

dx.doi.org/10.1021/bm301881h | Biomacromolecules 2013, 14, 564−574

Biomacromolecules

Article

Scheme 4. Synthesis of Dendrimers Using (3 + 2) Cycloadditiona

a

Example shown is for coupling of (BA)2-G1N3 and (PPA)4G2alk to form the Janus dendrimer (BA)2-G1-G2-(PPA)4.

conjugation.54 However, in this case, dendritic polymer backbones are used and, hence, facilitate higher drug loadings as the generation number of the dendrimer increases. Furthermore, the dendritic nature of the delivery systems results in a very low polydispersity (PDI≤ 1.003), which in turn affords a more consistent body distribution compared to polydispersed systems.55 Furthermore, batch to batch reproducibility is precisely controlled, which is not possible in current polymer−drug conjugates carrying two agents, which employ a random drug conjugation process resulting in an uncontrolled and nondirectional point of drug attachment.16,21 In Vitro Model Drug Release from Dendrimers. Model drug (BA and PPA) loading was confirmed at 100% of maximum theoretical loading (see Table 2) by 1H NMR spectroscopy and MALDI-TOF mass spectrometry. Purification of all conjugates from free, unreacted “drug” was a key part of the preparation of a polymer−drug conjugate as the free drug will have a different biological profile from the conjugated−drug.56 Removal of BA and PPA from our conjugates was successful and in all cases the amount of free model drug was quantified by HPLC analysis and was determined to be