- Email: [email protected]
Temperature-triggered on-demand drug release enabled by hydrogen-bonded multilayers of block copolymer micelles
Journal of Controlled Release 171 (2013) 73–80
Contents lists available at ScienceDirect
Journal of Controlled Release journal homepage: www.elsevier.com/locate/jconrel
Temperature-triggered on-demand drug release enabled by hydrogen-bonded multilayers of block copolymer micelles Zhichen Zhu 1, Ning Gao 1, Hongjun Wang ⁎, Svetlana A. Sukhishvili ⁎⁎ Department of Chemistry, Chemical Biology and Biomedical Engineering, Stevens Institute of Technology, Hoboken 07030, USA
a r t i c l e
i n f o
Article history: Received 10 March 2013 Accepted 20 June 2013 Available online 2 July 2013 Keywords: Temperature responsive Block copolymer micelles Layer-by-layer Drug release Poly(N-isopropylacrylamide)
a b s t r a c t We report on hydrogen-bonded layer-by-layer (LbL) films as a robust, reusable platform for temperaturetriggered “on-demand” release of drugs. Films with high drug loading capacity, temperature-controlled on–off drug release, and stability at physiological conditions were enabled by assembly of tannic acid (TA) with temperature-responsive block copolymer micelles (BCMs), which were pre-formed by heating solutions of a neutral diblock copolymer, poly(N-vinylpyrrolidone)-b-poly(N-isopropylacrylamide) (PVPON-b-PNIPAM), to a temperature above the lower critical solution temperature (LCST) of PNIPAM. The BCM/TA films exhibited temperature-triggered swelling/deswelling transitions at physiological conditions (swelling ratios of 1.75 and 1.2 at 37 °C and 20 °C, respectively). A model drug, doxorubicin (DOX) was incorporated into the film at a high drug-to-matrix ratio (~9.3 wt.% of DOX per film mass), with a total loading capacity controlled by the film thickness. At 37 °C, DOX was efficiently retained within the hydrophobic BCM cores of BCM/TA films, whereas exposure to a lower temperature (20 °C) triggered fast DOX release. While neither bare BCM-containing films nor films loaded with DOX showed cytotoxicity at 37 °C, drug released from films at lower temperature exhibited high potency against breast cancer cells. Repeated on/off drug release was demonstrated with 1.5-μm-thick DOX-loaded films, allowing at least three 30-min cooling cycles with consistent DOX (~12–16% of loaded DOX released for each cycle) released over a 4-day period. Despite significant stress associated with multiple swelling/deswelling cycles, films maintained their structural integrity in PBS, and each film could be repeatedly loaded with drug and used more than 15 times with only ~7% loss in film thickness and no obvious changes in reloading capacity or release profiles. This work presents the first proof-of-concept utility of temperature-responsive BCM-containing films for repeated on-demand release of a drug. © 2013 Elsevier B.V. All rights reserved.
1. Introduction On-demand drug release represents a promising strategy to potentially increase the therapeutic accuracy and minimize over-dosage induced side effects [1,2]. Ideally, the on-demand release system should release little or no drug in the “off” state, has fast stimuli response and reproducible drug release in each “on” state, and exhibits the ability to adjust drug dosage according to patient needs. More importantly, the system needs to maintain stability upon repeated switch to the “on” state without mechanical disruption.
⁎ Correspondence to: H. Wang, Department of Chemistry, Chemical Biology and Biomedical Engineering, Stevens Institute of Technology, McLean Building Room 416, 1 Castle Point on Hudson, Hoboken, NJ 07030, USA. Tel.: +1 201 216 5556; fax: +1 201 216 8240. ⁎⁎ Correspondence to: S.A. Sukhishvili, Department of Chemistry, Chemical Biology and Biomedical Engineering, Stevens Institute of Technology, McLean Building Room 315, 1 Castle Point on Hudson, Hoboken, NJ 07030, USA. Tel.: +1 201 216 5544; fax: +1 201 216 8240. E-mail addresses: [email protected] (H. Wang), [email protected] (S.A. Sukhishvili). 1 Equal contribution. 0168-3659/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jconrel.2013.06.031
Block copolymer micelles (BCMs) have been explored as promising vesicular drug carriers. The use of stimuli-responsive block copolymers for BCM assembly allows BCM carriers to release their load in response to environmental stimuli, such as pH, temperature, sound, or light [3–6]. Compared to other stimuli such as pH and ionic concentration alterations, tissue cells can better tolerate temperature changes especially in the region lower than physiological temperature. In this regard, temperature-triggered drug release is highly desirable. Temperatureresponsive BCMs can retain and release drugs loaded within their micellar cores as a result of temperature-induced hydrophobic/ hydrophilic transitions [7–10]. Often, this is realized through the inclusion of a BCM core of poly(N-isopropylacrylamide) (PNIPAM) — a widely used temperature-responsive polymer with phase transition occurring at close-to-physiological temperatures. Specifically, the lower critical solution temperature (LCST) of PNIPAM is ~ 32–33 °C [11,12], and this value can be further adjusted within a wide temperature range by using NIPAM-based copolymers. However, uncrosslinked micelles with responsive cores often lose their structural stability, and covalent stabilization through crosslinking of micellar cores or shells is necessary to improve structural stability and achieve tunable drug release [13–15].
74
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
Unlike freely diffusible individual micelles that potentially accumulate in the body, film-embedded micelles can be permanently bound within coatings, preventing systemic distribution of micelles and providing localized therapeutic effects. In this study, we stabilize temperature-responsive drug-delivering micelles by binding them onto a surface using layer-by-layer (LbL) assembly with a branched polymer. The LbL technique affords non-line-of-sight deposition of conformal coatings at substrates of virtually any shape, and allows for control of film structure, composition and delivery of multiple therapeutic components from surfaces [16–18]. With non-micellar LbL films, release of functional compounds through diffusion [19] in response to environmental triggers [20–24], or through film biodegradation [25,26] has been previously demonstrated. However, little is known about the functionality of BCM-containing LbL films as drug delivery films, especially for repeated “on-demand” drug release. While demonstrating drug delivery from BCM-containing films through hydrolysis or biodegradation [27,28], serious concerns have been raised regarding the morphologic and structural integrity of responsive micelles upon external stimuli. For example, morphologic changes were observed with films containing pH responsive BCMs even after a single pH stimulation [29,30]. Our team demonstrated pH-regulated release of a model dye from LbL films with temperature-responsive diblock copolymer micelles [31], however, further work showed that the use of diblock copolymers as BCM components compromised film structural stability during long-term exposure to changing environmental conditions [32], and triblock copolymers were required to assure film stability [33,34]. Moreover, electrostatic pairing responsible for assembly of the above-described films is vulnerable to attack by small ions and cannot be used as a universal strategy for constructing functional films under high-salt physiological environments. Here, we report on the use of hydrogen-bonded assembly to construct high-capacity, temperature-responsive, BCM-containing, reusable films for on-demand release of therapeutic compounds to cells. Previously, we demonstrated the ability of tannic acid (TA) to form stable hydrogen-bonded assemblies with a range of neutral homopolymers, including poly(N-vinylpyrrolidone) (PVPON) [35]. The assembled films were highly stable in a wide pH range and with tolerance to salt concentrations [36]. This all-homopolymer assembly was later used as a non-toxic, cytocompatible coating for encapsulation of yeast cells [37]. Following this strategy, here we report on novel films built via hydrogen-bonded assembly of TA with temperature-responsive micelles, formed from a neutral diblock copolymer, poly(N-vinylpyrrolidone)-b-poly(N-isopropylacrylamide) (PVPON-b-PNIPAM). Temperature-responsive micelles enable films with temperature-controlled swelling/deswelling transitions, as well as pulsed drug release, as a result of the LCST behavior of coreforming PNIPAM blocks. Unlike previous films constructed from similar BCMs but with another hydrogen donor, PMAA [31], the films described here remained highly stable at physiological conditions (PBS, 37 °C). At the same time, the films exhibited temperature-triggered swelling/deswelling transitions, which were used to “switch”, on or off, the release of doxorubicin (DOX), a widely used hydrophobic anticancer drug. To our knowledge, this is the first report on pulsed, on-demand release of drugs from biocompatible hydrogen-bonded temperature-responsive LbL films containing BCMs. Due to the unique capacity of TA to effectively bridge micelles within the film via multisite hydrogen bonding with PVPON in the BCM coronas, the films were able to withstand at least 15 temperature-induced swelling/deswelling cycles and could be reused for repeated DOX loading and release. As a proof-of-concept experiment of functionality of the BCM/TA matrix, we show that pulsed release of DOX from these films effectively eradicated human breast cancer (MCF-7) cells. The robust, reversible, externally controlled swelling and drug release modes make these films promising candidates for on-demand drug release from surfaces.
2. Materials and methods 2.1. Materials Benzyl chloride (Aldrich), elemental sulfur (Aldrich, St Louis, MO), and sodium methoxide (Fluka, St Louis, MO) were used as received. N, N-dimethylformamide (DMF) (99%), dioxane and tetrahydrofuran (THF) were received from Aldrich and distilled under reduced pressure prior to use. N-isopropylacrylamide (NIPAM) and azobis(isobutyronitrile) (AIBN) were received from Aldrich and recrystallized from benzene and methanol, respectively. Water with a resistivity of 18.2 MΩ cm−1 was supplied by a Millipore Milli-Q system. Branched polyethyleneimine (BPEI) with a weight-average molecular weight (Mw) of ~25 kDa, and Mw/Mn = 2.5, and PMAA with Mw ~ 150 kDa, were received from Sigma-Aldrich. DOX-HCl, hydrochloric acid, sodium hydroxide, sodium chloride, dibasic and monobasic sodium phosphate, as well as all other reagents were purchased from Sigma-Aldrich and used as received. 2.2. Synthesis of diblock copolymer PVPON165-b-PNIPAM140 block copolymer was synthesized using reversible addition–fragmentation chain transfer (RAFT) polymerization and characterizations are described in our previous work [31]. 2.3. Preparation and characterization of BCMs 2.3.1. Self-assembly of PVPON-b-PNIPAM BCMs 20 mg of block copolymer PVPON-b-PNIPAM was dissolved in 40 mL of 0.01 M phosphate buffer at pH 5.0, and the solution was gradually heated to 40 °C to enable the formation of BCMs. The presence of micelles in solution was confirmed by atomic force microscopy (AFM) and dynamic light scattering (DLS) as described below. 2.3.2. Dynamic light scattering (DLS) Hydrodynamic sizes of block copolymer solutions were measured using Zetasizer Nano-ZS equipment (Malvern Instruments Ltd, Worcestershire, UK). 2.4. Preparation and characterization of BCM-containing multilayer films 2.4.1. Deposition of BCM/TA films Silicon wafers or glass slides were cleaned as described elsewhere [38]. To enhance the adhesion of subsequently deposited multilayers to the surface, substrates were primed with a BPEI/PMAA bilayer as a precursor by alternately incubating with 0.2 mg/mL BPEI and PMAA solutions for 10 min with two 1-min rinsing cycles in buffer solution between polymer deposition. After rinsing with Milli-Q water, the substrates were then modified with neutral BCM/PMAA multilayers. All the deposition and rinsing steps were performed in 0.01 M pH 5.0 phosphate buffer at 40 °C. Hydrogen-bonded BCM/TA films were deposited by alternately exposing the substrate to respective solutions of 0.2 mg/mL BCM and 0.2 mg/mL TA solution for 10 min, with two 1-min intermediate rinsing steps with 0.01 M phosphate buffer. Deposition always started with BCM solution, and continued until the desired number of bilayers n was deposited. The produced [BCM/TA]n films were dried in air at 40 ºC. 2.4.2. AFM AFM measurements were performed in air using NSCRIPTOR Dip Pen Nanolithography system (Nanoink Inc., Skokie, IL). For AFM imaging, micelles were deposited on silicon wafers by drying BCM-containing solution at 40 °C.
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
2.4.3. Measurements of swelling ratios and refractive indices of BCM/TA films Thicknesses of dry and swollen BCM/TA films were measured by a custom-built phase-modulated ellipsometer at an incidence angle of 70°. Optical properties of the substrates and silicon dioxide layer thickness were determined prior to polymer deposition. In the case of dry polymer films, the refractive index was fixed at a value of 1.5. The swelling of neutral BCM/TA films was measured in situ using a custom-made cylindrical flow-through quartz cell. Briefly, to measure the thickness of swollen films, substrates containing [BCM/TA]4 films were inserted in the glass cells, and cells were filled with PBS. Solution temperature (typically 20 °C or 40 °C) was adjusted using an electric thermal controller and allowed for an equilibrium time of 10 min before measurements. In separate experiments, film thickness was also determined by AFM analysis of the depth profiles of razor-cut dry or wet films. 2.4.4. Confocal laser scanning microscopy (CLSM) Confocal images of films were obtained with an LSM 5 PASCAL laser scanning microscope (Carl Zeiss, Thornwood, NY) equipped with C-Apochromat 63 ×/1.2 water immersion objective. Prior to imaging, films were treated with Alexa 488 according to the procedure recommended by the supplier. 2.5. Preparation of DOX-loaded films Briefly, 2.0 mg of DOX-HCl was dissolved in 1 mL of DMSO, and 0.05 mL of triethylamine was then added into the solution to remove hydrochloride. The DOX solution was added dropwise into 20 mL of PBS at 37 °C. To load DOX within the BCM/TA films, substrates were immersed in the DOX solution for 24 h at 37 °C under stirring. The films were then immersed in a fresh PBS for 24 h at 37 °C to remove free DOX and/or organic solvent. The PBS was replaced every 4 h. To determine the DOX amount loaded within BCM/TA multilayers, 2.5 cm-by-2.5 cm silicon wafers covered with DOX-loaded [BCM/TA]40.5 films were immersed in 10-mL DMSO solution to extract the loaded DOX. The fluorescence intensity of DMSO extract was measured with a fluorometer (excitation at 480 nm), and the amount of DOX was then calculated from a calibration curve obtained with DOX/DMSO solutions of known concentrations. Drug loading content (DLC) was calculated according to the following formula: DLCðwt:% Þ ¼ ðweight of loaded drug=weight of polymerÞ 100%:
2.6. DOX release at different temperatures To measure the release of DOX from free BCMs, 2 mL of DOXloaded BCM solution (0.5 mg/mL, in PBS at pH 7.4 at 37 °C) was transferred into a dialysis membrane bag (MWCO:1000) and then incubated in 20 mL of fresh PBS buffer at 37 °C or 20 °C with gentle shaking. At predetermined time intervals, samples (0.5 mL) were harvested, quickly analyzed by fluorometer (excitation at 480 nm) and then injected back to keep the total volume constant. DOX concentration was determined using a standard curve obtained from DOX solutions of known concentrations in pH 7.4 PBS. Release of DOX from [BCM/TA]40.5 films was similarly measured to that of BCMs mentioned above, except 2.5-cm-by-2.5-cm substrates with DOX-loaded BCM films rather than free BCMs were placed in the dialysis bag. Temperature-triggered release of DOX from [BCM/TA]40.5 films was accomplished by incubating DOX-loaded films in a beaker containing 20 mL of pH 7.4 PBS placed in a 37 °C water bath for a period of 24 h, followed by a quick transfer of the beaker into another water bath at 20 °C for a period of 30 min under magnetic stirring
75
and then transferring the beaker back to the 37 °C water bath. The procedure was repeated for several cycles between 37 °C and 20 °C. Samples (0.5 mL) were harvested from the buffer solutions at defined times, quickly analyzed by fluorometer and injected back to the releasing buffer. The concentration of released DOX was calculated based on the standard curve. In all cases, cumulative release (%) expressed as the total percentage of drug released, was calculated as Wt / W × 100, where Wt is the cumulative weight of drug released by time t, and W is the total weight of drug loaded. 2.7. Cell culture Breast cancer cells (MCF-7, ATCC, Manassas, VA) were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 1% penicillin/streptomycin and 10% fetal bovine serum. The culture was maintained at 37 °C with 5% CO2 and 95% humidity and subculture was routinely performed at a 60–70% cell confluence using 0.25% trypsin-EDTA to detach the cells. 2.8. Tetrazolium salt (MTT) assay MTT assay was used to evaluate the biocompatibility of BCMcontaining films, as well as the effect of temperature-triggered release of DOX on cancer cells. To evaluate film biocompatibility, MCF-7 cells (1.5 × 104 cells per film) were seeded onto the surface of various films and cultured for up to 7 days. The same coverslip surfaces without films were used as controls. To assess the cell-killing efficacy of DOX released from BCM films, MCF (2 × 103 per well) were seeded to 96-well plates and cultured overnight. Then, 100 μL of sample media with or without released DOX was added to each well and incubated for 24 h. For MTT assays, the culture was incubated with thiazolyl blue tetrazolium bromide solution (0.5 mg/mL in culture media) at 37 °C in the dark for 2 h. After discarding nonreacted dye solution, the formazan product was extracted with DMSO and 100 μL of the extract was transferred to a 96-well plate for absorbance measurements. The absorbance was measured at a wavelength of 570 nm with a Synergy HT Multi-Detection Microplate Reader (BioTek Instruments, Winooski, VT). Cell viability was expressed as the percentage of control group: cell viability (%) = Atest / Acontrol × 100%. 2.9. Fluorescence staining for cell morphology and viability To better visualize cell morphology, fluorescence staining for F-actin and cell nuclei were performed as reported previously [39,40]. TRITC-conjugated-phalloidin (Sigma) and SlowFade® Gold antifade reagent with DAPI (Invitrogen, Carlsbad, CA) were used. Briefly, cells were fixed in 3.7% paraformaldehyde in PBS for 5 min and permeabilized with 0.5% Triton X-100 in PBS for another 5 min. The cells were stained with a 50 μg/mL TRITC-conjugated-phalloidin solution at room temperature for 30 min and then washed with PBS to remove unbound phalloidin conjugates. Cell nuclei were stained with DAPI. The stained cells were examined under a Nikon 80i epi-fluorescence microscope with an excitation/emission wavelength at ex 360 nm/em 460 nm (DAPI) and ex 540 nm/em 570 nm (TRITC). To assess cell viability, cultured cells were stained with a LIVE/ DEAD cell vitality assay kit (Invitrogen) following the manufacturer's manual. Briefly, cultured cells were incubated with a staining solution containing 0.125 μM C12-resazurin and 0.02 μM SYTOX in Hank's balanced salt solution (HBSS) for 15 min. Afterwards, the cells were washed with HBSS and fixed with 3.7% paraformaldehyde in PBS for 5 min. The cells were then examined under the fluorescent microscope with an excitation/emission wavelength at ex 490–505 nm/em 530 nm for SYTOX (dead cells) and ex 560–580 nm/em 575 nm for C12-resorufin (live cells), respectively.
76
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
2.10. Statistical analysis All quantitative results were obtained from at least three samples. Each experiment was repeated separately at least three times. Data were expressed as the mean ± standard deviation (SD). An unpaired t-test was used in the statistical analysis of experimental data. A value of p b 0.05 was considered to be statistically significant.
3. Results and discussion 3.1. LbL BCM-containing films and their stability at physiological pH Our design of on-demand drug-releasing films is shown in Fig. 1. Temperature-responsive block copolymer, poly(N-vinylpyrrolidone)b-poly(N-isopropylacrylamide) (PVPON165-b-PNIPAM140), was synthesized by reversible addition–fragmentation chain transfer (RAFT) polymerization as described earlier [31]. Heating the block copolymer solution to temperatures above PNIPAM's LCST of 32–34 °C resulted in the formation of BCMs with a hydrodynamic diameter of ~90 nm and PNIPAM block micellar cores [31]. While in our previous work these micelles were used for assembly with a weak polycarboxylic acid, yielding films unstable at neutral pH values, here we explored the binding of PVPON-b-PNIPAM micelles with another hydrogen-bonding partner, tannic acid (TA) (Fig. 1). Our earlier study has shown that TA forms strong multiple hydrogen bonds with linear neutral polymers, and such hydrogen bonding persists at neutral and slightly basic pH values [35]. In this regard, we expected that binding of TA with copolymer block included within micellar coronae could be equally efficient. Scheme 1 suggests that nonlinear molecules of TA effectively bind BCM micelles within LbL films by forming physical crosslinks between PVPON block in micellar coronas. Fig. 1a shows the thickness growth of [BCM/TA]n films as monitored by ellipsometry and AFM. The data obtained from both techniques were consistent within 10% variation. Nearly linear film growth was observed with [BCM/TA]n films at pH 5, indicating the presence of strong binding between BCM shells and TA. The average bilayer thickness of [BCM/TA]n films as calculated from the slopes of growth curves in Fig. 1a was 37.0 ± 3.5 nm. Fig. 1b shows the evolution of surface morphology of [BCM/TA]n films with an increased thickness. As the number of deposited BCM/TA bilayers exceeded three, the surface became fully covered with micelles, while
Scheme 1. Schematic illustration of the layer-by-layer assembly of PVPON-b-PNIPAM BCMs with TA into stable films at a temperature above PNIPAM's LCST. Right: hydrogen bonding formed between TA and BCMs within the LbL films.
the root mean square (rms) roughness remained modest with a small variation as a function of the number of assembled layers. More specifically, the rms roughness was 24.1, 19.5, 17.9 and 21.3 nm for 1-, 3-, 6-, and 12-bilayer [BCM/TA]n films, respectively. Fig. 1c shows an SEM image of a [BCM/TA]40.5 film (with thickness of 1.49 ± 0.15 μm), deposited on a silicon substrate at 40 °C. Closely-packed spherical micellar morphology suggests structural stability of BCMs within LbL assemblies. One critical aspect in prospective applications of BCM-containing films for drug delivery is their stability at physiological conditions. Fig. 2 illustrates pH-triggered disintegration of BCM/TA films. In our earlier study, we demonstrated that dissociation of multilayers assembled from TA with linear PVPON occurred as a result of deprotonation of weakly acidic phenolic hydroxyl groups of TA, when external pH exceeds the 8.5 pKa of TA [35,41]. Fig. 2 illustrates that [BCM/TA]n films remained stable at neutral and slightly basic pH values. Inset in Fig. 2 shows the effect of immersion time on the stability of [BCM/TA]n films in PBS at 20 °C. Less than 5% thickness decrease was observed for hydrogen-bonded [BCM/TA]n films after immersion in PBS for 24 h. The stability of films at physiological conditions revealed the intrinsic low sensitivity of hydrogen bonding to salts in the solution.
1.0
0.5
Fraction retained
Fraction retained
1.0
in PBS buffer, pH 7.4
0.5 1
3
5
7
Immersion time (d)
0.0 4
6
8
10
pH Fig. 1. (a) AFM images of PVPON-b-PNIPAM [BCM/TA]n films with n = 1, 3, 6, and 12 (scanned area 2 μm × 2 μm). (b) Dry thicknesses of PVPON-b-PNIPAM [BCM/TA]n films as a function of bilayer number deposited at 40 °C as monitored by ellipsometry and AFM. In ellipsometic measurements, the film refractive index was fixed at 1.5. (c) SEM image of the surface morphology of PVPON-b-PNIPAM [BCM/TA]40.5 films. Inset: photograph of a [BCM/TA]40.5 film deposited on a 2.5 cm × 2.5 cm glass coverslip. BCM/TA films were deposited from solutions of pH 5.0 at 40 °C and dried in air at 40 °C.
Fig. 2. Thickness loss of dry [BCM/TA]8 films as a function of pH in PBS at 20 °C. 0.5 M NaOH solution was used for pH adjustment. Thickness retained on silicon substrates was measured by ellipsometry after exposure to buffer solutions for 30 min and dried. Fraction retained was calculated as the percentage of the original dry film thickness. Inset shows fraction of a [BCM/TA]8 film as a function of immersion time in PBS buffer at 20 °C. Films immersed in pH 7.4 buffer solutions were removed at defined time intervals, dried for thickness measurements by ellipsometry and then immersed back to pH 7.4 buffer solutions.
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
Fig. 3. (a) In situ ellipsometry measurements of the swelling ratio and refractive index of a [BCM/TA]4 film (dry thickness ~ 115 nm) immersed in PBS solution at pH 7.4 upon temperature cycling between 20 °C and 40 °C. (b) CLSM imaging of the thickness of [BCM/TA]40.5 films at 20 °C (top) and 40 °C (bottom). Films were stained with Alexa 488. The scale bars are 5 μm. (c) AFM images of the surface morphology of [BCM/ TA]40.5 films at 20 °C (top) and 40 °C (bottom). (d) Schematic illustration of temperature-triggered swelling transitions of BCM-containing films.
3.2. Temperature-triggered swelling of [BCM/TA] films at physiological conditions With successful demonstration of the structural stability of BCM/TA films at physiological pH, we explored temperature-controlled water uptake within assembled micellar PNIPAM cores and the resultant swelling of LbL films. The temperature effect on the film swelling was studied using phase-modulated in situ ellipsometry. Fig. 3a shows variations in thickness and the refractive index of wet [BCM/TA]4 films when cycling the solution temperature between 20 °C and 40 °C. At 40 °C, when PNIPAM cores of BCMs dehydrated, the film swelled from its dry state to a modest swelling ratio of 1.2 ± 0.07, with a refractive index decrease from 1.5 to 1.42. The small degree of swelling at 40 °C reflects the presence of collapsed PNIPAM cores within BCM micelles, and dehydration of BCM/TA film as a result of hydrogenbonding between PVPON coronal blocks and TA. After cooling the solution to 20 °C, the BCM/TA film swelling ratio increased to ~1.75 ± 0.06 (corresponding to a film thickness of 210 nm), and its refractive index decreased to 1.36 ± 0.07, suggesting the substantial hydration of the BCM/TA film below LCST. CLSM confirmed dramatic changes in thickness of swollen [BCM/TA]40.5 films from 2.8 ± 0.2 μm at 20 °C (top) to 1.7 ± 0.2 μm at 40 °C (bottom) (Fig. 3b). Fig. 3c illustrates a corresponding shrinkage of assembled micelles, with a decrease in their average lateral size from 160 ± 15 nm at 20 °C to 105 ± 10 nm at 40 °C, indicating dehydration of PNIPAM micellar cores above LCST. At the same time, hydrogen bonding between TA and PVPON is not affected by temperature variations; as indicated by Fig. 3, binding of BCM coronal chains with TA prevents dissociation of BCMs into unimers even at low temperature, while allowing reversible volumetric expansion and contraction of micellar PNIPAM cores and swelling transitions within the entire BCM-containing films (Fig. 3d).
77
attributed to the entrapment of DOX within hydrophobic PNIPAMforming cores (at temperatures higher than PNIPAM's LCST). Note that a value of 0.4% has been reported in our previous work for inclusion of a hydrophobic fluorescent dye (pyrene) within BCMs assembled within BCM/PMAA films [31], which were unstable at physiological conditions. Here, much higher loading capacity was achieved with BCM/TA films, which were highly stable at neutral and slightly basic pH range due to hydrogen-bonding-assisted bridging of micellar shells by TA molecules. To determine the biocompatibility of drug-loaded films, MCF-7 cancer cells were cultured on the surface of [BCM/TA]n films for up to 7 days. The cell proliferation on the films was evaluated by using MTT assay. Although the MTT assay is not accurate to determine the exact cell number, it measures cell metabolic activity, which is proportional to cell number under a similar culture condition [42]. Therefore, MTT assay can be used to roughly indicate the cell proliferation trend. Considering that films could be constructed with either TA or BCM as the outermost layer, we evaluated both film types in cell culture experiments. Fig. 4b shows viable cell numbers of MCF-7 cells at DOX-loaded [BCM/TA]n films after culturing for 1, 3 and 7 days. When experiments were performed at 37 °C, [BCM/TA]n films with either BCM or TA as a top layer supported cell proliferation similarly to control glass coverslip surfaces (Fig. 4b). After culturing for 1, 3 and 7 days, cell morphology was observed by fluorescent phalloidinTRITC staining of the cytoskeletal protein filament actin. The fluorescent images revealed that morphologies of all the cells cultured on LbL films remained similar to those observed with the glass coverslip control group, with a typical cobblestone-like morphology in a colony (Fig. 4c). In addition, the average colony size and number were also comparable among various groups (data not shown). This observation agrees well with the cell proliferation results, suggesting compatibility of both bare and DOX-loaded [BCM/TA]n films to MCF-7 at 37 °C. While cellular response was independent of the film top layer, we have used [BCM/TA]40.5, i.e. films containing BCMs as the outmost layer, in our further experiments.
3.3. DOX-loaded BCM/TA films and their biocompatibility Because of the known adverse side effects of DOX, entrapment of DOX within micelles (Fig. 4a) can minimize potential damage to healthy tissues. Loading of DOX within micellar PNIPAM cores was achieved by immersing [BCM/TA]40.5 films in the DOX solution and dialyzing against water at 37 °C. For [BCM/TA]40.5 film, a high DOX loading content of 9.3% of the film mass was obtained as determined by fluorometry. The high amount of DOX loaded within the LbL film is
Fig. 4. (a) Schematic representation of a BCM/TA film retaining DOX in PBS solution at pH 7.4 and 37 °C. (b) Proliferation of MCF-7 cells cultured on DOX-loaded [BCM/TA] films with TA top layer ([BCM/TA]40 film) or BCM top layer ([BCM/TA]40.5 film) at a constant temperature of 37 °C. MCF-7 cells at a density of 1.5 × 104 cells/well were cultured on the films for 1, 3 and 7 days prior to MTT assay. Glass coverslip surface was used as a control. (c) Morphologies of MCF-7 cells on the control surface, as well as on [BCM/TA]40 and [BCM/TA]40.5 films. Cytoskeletal F-actin was stained red with TRITC-phalloidin, and cell nuclei were stained blue with DAPI. Scale bar is 50 μm.
78
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
Fig. 5. (a) Cumulative release of DOX from free BCMs (circles) or from [BCM/TA]40.5 films (squares) in PBS buffer (pH 7.4) at 20 °C (open symbols) and 37 °C (filled symbols). DOX release was monitored by fluorometry (excitation at 480 nm), and released amounts normalized to the total amount of DOX loaded within the film. (b) Release profiles of DOX from [BCM/TA]40.5 films immersed in PBS at a constant temperature of 37 °C, and cooled down to 20 °C for 30 min at cumulative time of 1, 2 and 3 days. Inset shows percentage of DOX released in each cooling cycle. (c) Schematic representation of “on-demand” DOX release from BCM-containing films.
3.4. Temperature-triggered release of DOX from BCM/TA films With successful entrapment of DOX into BCM-containing hydrogenbonded films, it becomes essential to understand how temperature regulates DOX on-demand release. Fig. 5a illustrates the rate of accumulation of DOX released from [BCM/TA]40.5 films during continuous exposure to PBS at a constant temperature of 20 °C or 37 °C. It is necessary to mention that although 20 °C was continuously used for DOX release in this study, in order to be consistent with above experiments, any other temperature below PNIPAM's LCST of 32 °C could equally serve the purpose. For comparison, DOX released from individual BCMs in solution is also shown in Fig. 5a. DOX release from free BCMs was very fast at 20 °C, with nearly 85% of incorporated DOX released during the initial 30 min, reflecting dissociation of micelles to unimers when the solution temperature was below PNIPAM's LCST. At temperatures above LCST, hydrophobic cores of micelles efficiently retained hydrophobic DOX, leading to slower DOX release. Indeed, at 37 °C the release of DOX was drastically slower, with only ~25% DOX released within the first day and very slow release at longer times. Binding of BCMs with TA within the LbL film significantly slowed DOX release as compared to its release from free BCMs. At 37 °C and pH 7.4, only ~10% of DOX was released the first day, and a large amount of DOX was retained in the [BCM/TA]40.5 films even after 4-day incubation. Remarkably, when the temperature was lowered to 20 °C, DOX release from [BCM/TA]40.5 films was greatly accelerated, quickly reaching ~18–20% within 30 min. The temperature-triggered DOX release patterns strongly correlated with the results of temperature-controlled swelling transitions of [BCM/TA]n films in Fig. 3, indicating that retention/release of DOX within/from the film was dictated by differential partitioning the hydrophobic drug in collapsed (above PNIPAM's LCST) or hydrated (below PNIPAM's LCST) micellar cores. In addition to demonstrated continuous release of DOX from the assembled films at temperatures below LCST, it would be beneficial to achieve pulsed release of drugs from the films. To that end, a cyclic thermal stimulation at temperatures above and below LCST was applied to the DOX-loaded [BCM/TA]n films. Fig. 5b shows the release kinetics of DOX from [BCM/TA]40.5 films when films were immersed in PBS at 37 °C, and cooled down to 20 °C for 30 min at days 1, 2 and 3, respectively. Pulsed release of DOX was observed with a step-like increase in cumulative
DOX concentration in solution, with every exposure to 20 °C leading to a burst release of 12–16% of loaded DOX. Note that the released dose decreased from ~16% to ~14% and ~12% in the 1st, 2nd, and 3rd temperature lowering cycles. The lower amounts released in the 2nd and 3rd cycles are attributed to the decreased total ‘supply’ of DOX within the film. In practice, it would be an appealing feature for the films to be recharged with drugs for continued use. Here, we explored the possibility of reusing BCM/TA films for repeated temperature-triggered drug release. Specifically, after complete release of DOX from [BCM/TA]40.5 constructs, films were reloaded with DOX and re-evaluated for their drug release profiles. Interestingly, DOX-reloaded films exhibited essentially identical release profiles to those observed with freshly prepared DOX-loaded BCM films (Fig. 6). Importantly, we were able to perform as many as 15 film re-loading cycles, without significant changes in the DOX release profiles (data not shown). Also note that films repeatedly reloaded with DOX demonstrated highly reproducible pulsed release patterns (Fig. 5b). The ability to be replenished with bioactive compounds is an important feature in designing medical devices [43], and our BCM/TA films seem to satisfy this requirement. The fact that drug-hosting micellar cores preserve their capacity to DOX partitioning even after many film swelling/deswelling cycles suggest no significant morphological changes within non-covalently assembled films. The result is encouraging but yet somewhat surprising, considering the non-covalent assembly of film components. In contrast to previous findings that electrostatically assembled films of temperatureresponsive diblock copolymer micelles showed significant mass loss during long-term exposure to low temperature [34], here, we demonstrate that the use of TA – a non-linear hydrogen-bonding molecule – assures strong binding and efficient stabilization of BCMs within a film during temperature variations, and does not require the use of triblock copolymers in order to enhance the stability of the polymer matrix.
3.5. In vitro efficiency of BCM/TA films in eradicating cancer cells In order to demonstrate film functionality, a proof-of-concept experiment was designed to explore whether the temperaturetriggered release of DOX from [BCM/TA]40.5 films was therapeutically efficient in killing breast cancer cells (MCF-7) as a model. Moreover, we were seeking to explore whether each temperature cycle resulted in reproducible and repeatable cell killing patterns. To that end, release media samples were collected from DOX-loaded [BCM/TA]40.5 films for each of three temperature cycles (one cycle included exposure of the film to 37 °C for 1 day, followed by lowering the
Fig. 6. Representative release profiles of DOX from repeatedly used, DOX-reloaded [BCM/TA]40.5 films in PBS solutions (pH 7.4) at 20 °C (open symbols) and 37 °C (filled symbols). Prior to drug reloading, DOX loaded within the film during previous loading cycles was completely removed by immersing the films in PBS buffer for 7 days at 20 °C.
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
79
with those in Fig. 5b, showing that DOX-loaded [BCM/TA]40.5 films exhibited a very slow release of loaded drug at physiological temperature, while temporarily cooling the films below the LCST of PNIPAM remarkably accelerated the release of DOX. The repeatable tunability of biological activity of the [BCM/TA]n films via simple temperature variations renders such films a promising platform for on-demand release devices. Moreover, by tailoring the size and chemical nature of micellar cores for example, by endowing this block high critical solution temperature rather than LCST behavior at the step of polymer synthesis, and by controlling the number of BCM layers within LbL films, as well as the duration of temperature variation cycles, this platform can be programmed to release desired doses of therapeutics at many scheduled time points. Therapeutic dosages of drugs can be delivered on demand within short periods of time at desirable temperatures, and the delivery can then be “switched off” at physiological temperatures in order to minimize undesirable side effects on healthy tissues. 4. Conclusions
Fig. 7. (a) Cytotoxicity of [BCM/TA]40.5 films against MCF-7 cells after multiple temperature cycles between 37 °C and 20 º°C. Films were immersed in culture medium at a constant temperature of 37 °C, and then cooled down to 20 °C for 30 min at cumulative time of 1, 2 and 3 days. After each 37 °C/20 °C cycling, the medium was separately collected and replaced with fresh medium at 37 °C to start a new release cycle. MCF-7 cells cultured overnight were treated with collected medium samples for 24 h and analyzed by MTT assay to determine cell viability. (b) Immunofluorescence images of MCF-7 cells cultured for 24 h. F-actin was stained red with phalloidin-TRITC and cell nuclei stained blue with DAPI. (c) LIVE/DEAD staining of MCF-7 cultured for 24 h. Red = live; green = dead.
temperature to 20 °C for 30 min), and used to treat MCF-7 cells for MTT assay. The cell viability assay showed that media samples from DOX-loaded [BCM/TA]40.5 films at 37 °C had negligible cytotoxicity against MCF-7 cells, with only a small reduction (~7%) in cell viability for 24 h of incubation (Fig. 7a). In contrast, the release samples from 20 °C for 30-min cycles significantly (p b 0.001) reduced cell viability to ~ 7% (1st), 9% (2nd), and 10% (3rd) compared to untreated controls, respectively. Immunofluorescent staining for cell morphology with phalloidin-TRITC and cell viability with LIVE/DEAD assay kit further confirmed the MTT results, showing vast differences between 1-day at 37 °C and 30-min at 20 °C samples. MCF-7 cells treated with 37 °C media samples retained their cobblestone morphology with all the cells alive, similar to that of control groups, while only cell nuclei were seen in the culture treated with 20 °C samples and no living cells were identified (Fig. 7b and 7c). These results indicated that DOX-loaded [BCM/TA]40.5 films maintained high bioactivity even after multiple temperature cycles. The results also showed that at each 20 °C incubation cycle, the amount of released DOX was sufficiently high to eradicate breast cancer cells (MCF-7), while the DOX amount released at 37 °C was well below the therapeutic dose, resulting in minimal cell killing even after several days of contact with MCF-7 cells. The data in Fig. 7 are in excellent agreement
We have constructed a new type of hydrogen-bonded, non-toxic, BCM-containing LbL films, which demonstrated high stability at physiological pH and ionic concentrations, and exhibited highly sensitive swelling/deswelling response to temperature variations. The films were highly efficient in entrapping high amounts of a hydrophobic drug, DOX, within PNIPAM cores of assembled BCMs and supported on demand or pulsated temperature-controlled drug release. While both drug-free and DOX-loaded BCM/TA films were not cytotoxic at 37 °C, lowering the temperature to 20 °C for 30 min triggered release of DOX doses sufficient to eradicate breast cancer cells with N90% killing efficacy. The large-amplitude, reversible, temperature-induced swelling transitions also enabled repeatable, temperature-regulated on/off release of DOX from BCM-containing films, with similar cell killing capacity of the drug in each cycle. Importantly, films could be reloaded with the drug at least 15 times, giving reproducible drug release profiles in multiple release cycles. Our results demonstrate strong potential of novel hydrogen-bonded, BCM-containing LbL films as high-efficiency, reusable matrices for temperature-regulated release of hydrophobic drugs. While this study presents a proof-of-concept demonstration of the application of these films for repeatable on-demand release of bioactive compounds from surfaces, these matrices can be integrated with a broad range of biomedical devices (e.g. wound dressing) for on-demand release of antibiotics and growth factors to stimulate tissue regeneration. Acknowledgments The authors thank the National Science Foundation (S. Sukhishvili, Award DMR-0906474) and the National Institute of Arthritis and Musculoskeletal and Skin Diseases (H. Wang, Award 1R21 AR056416) for financial support of this work. The image resources used in this study were partially funded by the National Science Foundation through NSF Grant DMR-0922522. References [1] B.P. Timko, T. Dvir, D.S. Kohane, Remotely triggerable drug delivery systems, Adv. Mater. 22 (44) (2010) 4925–4943. [2] Y. Shao, W. Huang, C. Shi, S.T. Atkinson, J. Luo, Reversibly crosslinked nanocarriers for on-demand drug delivery in cancer treatment, Ther. Deliv. 3 (12) (2012) 1409–1427. [3] G. Gaucher, M.H. Dufresne, V.P. Sant, N. Kang, D. Maysinger, J.C. Leroux, Block copolymer micelles: preparation, characterization and application in drug delivery, J. Control. Release 109 (1–3) (2005) 169–188. [4] V.P. Torchilin, Micellar nanocarriers: pharmaceutical perspectives, Pharm. Res. 24 (1) (2007) 1–16. [5] N. Rapoport, Physical stimuli-responsive polymeric micelles for anti-cancer drug delivery, Prog. Polym. Sci. 32 (8–9) (2007) 962–990. [6] A.W. York, S.E. Kirkland, C.L. McCormick, Advances in the synthesis of amphiphilic block copolymers via RAFT polymerization: stimuli-responsive drug and gene delivery, Adv. Drug Deliv. Rev. 60 (9) (2008) 1018–1036.
80
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
[7] D. Neradovic, C.F. Van Nostrum, W.E. Hennink, Thermoresponsive polymeric micelles with controlled instability based on hydrolytically sensitive N-isopropylacrylamide copolymers, Macromolecules 34 (2001) 7589–7591. [8] C.J. Rijcken, O. Soga, W.E. Hennink, C.F. van Nostrum, Triggered destabilization of polymeric micelles and vesicles by changing polymers polarity: an attractive tool for drug delivery, J. Control. Release 120 (3) (2007 Jul 31) 131–148. [9] H. Wei, S.X. Cheng, X.Z. Zhang, R.X. Zhuo, Thermo-sensitive polymeric micelles based on poly(N-isopropylacrylamide) as drug carriers, Prog. Polym. Sci. 34 (9) (2009) 893–910. [10] X. He, X. Wu, C. Gao, K. Wang, S. Lin, W. Huang, M. Xie, D. Yan, Synthesis and self-assembly of a hydrophilic, thermo-responsive poly(ethylene oxide) monomethyl ether-block-poly(acrylic acid)-block-poly(N-isopropylacrylamide) copolymer to form micelles for drug delivery, React. Funct. Polym. 71 (2011) 544–552. [11] S. Fujishige, K. Kubota, I. Ando, Phase transition of aqueous solutions of poly(Nisopropylacrylamide) and poly(N-isopropylmethacrylamide), J. Phys. Chem. 93 (8) (1989) 3311–3313. [12] F.M. Winnik, Fluorescence studies of aqueous solutions of poly(N-isopropylacrylamide) below and above their LCST, Macromolecules 23 (1990) 233–242. [13] X. Jiang, Z. Ge, J. Xu, H. Liu, S. Liu, Fabrication of multiresponsive shell cross-linked micelles possessing pH-controllable core swellability and thermo-tunable corona permeability, Biomacromolecules 8 (10) (2007) 3184–3192. [14] Y. Li, B.S. Lokitz, S.P. Armes, C.L. McCormick, Synthesis of reversible shell cross-linked micelles for controlled release of bioactive agents, Macromolecules 39 (8) (2006) 2726–2728. [15] Y. Zeng, W.G. Pitt, Poly(ethylene oxide)-b-poly(N-isopropylacrylamide) nanoparticles with crosslinked cores as drug carriers, J. Biomater. Sci. Polym. Ed. 16 (3) (2005) 371–380. [16] S. Pavlukhina, S. Sukhishvili, Polymer assemblies for controlled delivery of bioactive molecules from surfaces, Adv. Drug Deliv. Rev. 63 (9) (2011) 822–836. [17] A.J. Karlsson, R.M. Flessner, S.H. Gellman, D.M. Lynn, S.P. Palecek, Polyelectrolyte multilayers fabricated from antifungal β-peptides: design of surfaces that exhibit antifungal activity Candida albicans, Biomacromolecules 11 (9) (2010) 2321–2328. [18] B.-S. Kim, R.C. Smith, Z. Poon, P.T. Hammond, Mad (multiagent delivery) nanolayer: delivering multiple therapeutics from hierarchically assembled surface coatings, Langmuir 25 (2009) 14086–14092. [19] M.C. Berg, L. Zhai, R.E. Cohen, M.F. Rubner, Controlled drug release from porous polyelectrolyte multilayers, Biomacromolecules 7 (1) (2006) 357–364. [20] B.M. Wohl, J.F. Engbersen, Responsive layer-by-layer materials for drug delivery, J Control. Release 158 (1) (2012) 2–14. [21] E. Kharlampieva, S.A. Sukhishvili, Release of a dye from hydrogen-bonded and electrostatically assembled polymer films triggered by adsorption of a polyelectrolyte, Langmuir 20 (2004) 9677–9685. [22] S. Pavlukhina, Y. Lu, A. Patimetha, M. Libera, S. Sukhishvili, Polymer multilayers with pH-triggered release of antibacterial agents, Biomacromolecules 11 (12) (2010) 3448–3456. [23] X. Wang, L. Zhang, L. Wang, J. Sun, J. Shen, Layer-by-layer assembled polyampholyte microgel films for simultaneous release of anionic and cationic molecules, Langmuir 26 (11) (2010) 8187–8194. [24] M.J. Serpe, C.D. Jones, L.A. Lyon, Layer-by-layer deposition of thermoresponsive microgel thin films, Langmuir 19 (2003) 8759–8764.
[25] E. Vázquez, D.M. Dewitt, P.T. Hammond, D.M. Lynn, Construction of hydrolyticallydegradable thin films via layer-by-layer deposition of degradable polyelectrolytes, J. Am. Chem. Soc. 124 (47) (2002) 13992–13993. [26] J. Zhang, L.S. Chua, D.M. Lynn, Multilayered thin films that sustain the release of functional DNA under physiological conditions, Langmuir 20 (19) (2004) 8015–8021. [27] B.S. Kim, H.I. Lee, Y. Min, Z. Poon, P.T. Hammond, Hydrogen-bonded multilayer of pH-responsive polymeric micelles with tannic acid for surface drug delivery, Chem. Commun. 28 (2009) 4194–4196. [28] B.S. Kim, S.W. Park, P.T. Hammond, Hydrogen-bonding layer-by-layer-assembled biodegradable polymeric micelles as drug delivery vehicles from surfaces, ACS Nano 2 (2) (2008) 386–392. [29] T. Addison, O.J. Cayre, S. Biggs, S.P. Armes, D. York, Incorporation of block copolymer micelles into multilayer films for use as nanodelivery systems, Langmuir 24 (23) (2008) 13328–13333. [30] I. Erel, Z. Zhu, A. Zhuk, S.A. Sukhishvili, Hydrogen-bonded layer-by-layer films of block copolymer micelles with pH-responsive cores, J. Colloid Interface Sci. 355 (1) (2011) 61–69. [31] Z. Zhu, S.A. Sukhishvili, Temperature-induced swelling and small molecule release with hydrogen-bonded multilayers of block copolymer micelles, ACS Nano 3 (11) (2009) 3595–3605. [32] L. Xu, Z. Zhu, S.A. Sukhishvili, Polyelectrolyte multilayers of diblock copolymer micelles with temperature-responsive cores, Langmuir 27 (1) (2011) 409–415. [33] W.S. Tan, R.E. Cohen, M.F. Rubner, S.A. Sukhishvili, Temperature-induced, reversible swelling transitions in multilayers of a cationic triblock copolymer and a polyacid, Macromolecules 43 (4) (2010) 1950–1957. [34] W.S. Tan, Z. Zhu, S.A. Sukhishvili, M.F. Rubner, R.E. Cohen, Effect of block copolymer architecture on the thermally induced swelling of micelle-containing multilayer thin films, Macromolecules 44 (19) (2011) 7767–7774. [35] I. Erel-Unal, S.A. Sukhishvili, Hydrogen-bonded multilayers of a neutral polymer and a polyphenol, Macromolecules 41 (11) (2008) 3962–3970. [36] E. Kharlampieva, S.A. Sukhishvili, Hydrogen-bonded layer-by-layer films, Journal J. Macromol. Sci. C Polym. Rev. J. 46 (2006) 377–395. [37] V. Kozlovskaya, S. Harbaugh, I. Drachuk, O. Shchepelina, N. Kelley-Loughnane, M. Stone, V.V. Tsukruk, Hydrogen-bonded LbL shells for living cell surface engineering, Soft Matter 7 (2011) 2364–2372. [38] V. Kozlovskaya, S. Yakovlev, M. Libera, S.A. Sukhishvili, Surface priming and the self-assembly of hydrogen-bonded multilayer capsules and films, Macromolecules 38 (11) (2005) 4828–4836. [39] T. Wieland, Phallotoxins, Springer-Verlag, New York, 1986. [40] M. Kubista, B. Akerman, B. Nordén, Characterization of interaction between DNA and 4′,6-diamidino-2-phenylindole by optical spectroscopy, Biochemistry 26 (14) (1987) 4545–4553. [41] W. Bors, C. Michel, K. Stettmaier, Electron paramagnetic resonance studies of radical species of proanthocyanidins and gallate esters, Arch. Biochem. Biophys. 374 (2) (2000) 347–355. [42] E. Vega-Avila, M.K. Pugsley, An overview of colorimetric assay methods used to assess survival or proliferation of mammalian cells, Proc. West. Pharmacol. Soc. 54 (2011) 10–14. [43] R. Lo, P.Y. Li, S. Saati, R. Agrawal, M.S. Humayun, E. Meng, A refillable microfabricated drug delivery device for treatment of ocular diseases, Lab Chip 8 (7) (2008) 1027–1030.
Contents lists available at ScienceDirect
Journal of Controlled Release journal homepage: www.elsevier.com/locate/jconrel
Temperature-triggered on-demand drug release enabled by hydrogen-bonded multilayers of block copolymer micelles Zhichen Zhu 1, Ning Gao 1, Hongjun Wang ⁎, Svetlana A. Sukhishvili ⁎⁎ Department of Chemistry, Chemical Biology and Biomedical Engineering, Stevens Institute of Technology, Hoboken 07030, USA
a r t i c l e
i n f o
Article history: Received 10 March 2013 Accepted 20 June 2013 Available online 2 July 2013 Keywords: Temperature responsive Block copolymer micelles Layer-by-layer Drug release Poly(N-isopropylacrylamide)
a b s t r a c t We report on hydrogen-bonded layer-by-layer (LbL) films as a robust, reusable platform for temperaturetriggered “on-demand” release of drugs. Films with high drug loading capacity, temperature-controlled on–off drug release, and stability at physiological conditions were enabled by assembly of tannic acid (TA) with temperature-responsive block copolymer micelles (BCMs), which were pre-formed by heating solutions of a neutral diblock copolymer, poly(N-vinylpyrrolidone)-b-poly(N-isopropylacrylamide) (PVPON-b-PNIPAM), to a temperature above the lower critical solution temperature (LCST) of PNIPAM. The BCM/TA films exhibited temperature-triggered swelling/deswelling transitions at physiological conditions (swelling ratios of 1.75 and 1.2 at 37 °C and 20 °C, respectively). A model drug, doxorubicin (DOX) was incorporated into the film at a high drug-to-matrix ratio (~9.3 wt.% of DOX per film mass), with a total loading capacity controlled by the film thickness. At 37 °C, DOX was efficiently retained within the hydrophobic BCM cores of BCM/TA films, whereas exposure to a lower temperature (20 °C) triggered fast DOX release. While neither bare BCM-containing films nor films loaded with DOX showed cytotoxicity at 37 °C, drug released from films at lower temperature exhibited high potency against breast cancer cells. Repeated on/off drug release was demonstrated with 1.5-μm-thick DOX-loaded films, allowing at least three 30-min cooling cycles with consistent DOX (~12–16% of loaded DOX released for each cycle) released over a 4-day period. Despite significant stress associated with multiple swelling/deswelling cycles, films maintained their structural integrity in PBS, and each film could be repeatedly loaded with drug and used more than 15 times with only ~7% loss in film thickness and no obvious changes in reloading capacity or release profiles. This work presents the first proof-of-concept utility of temperature-responsive BCM-containing films for repeated on-demand release of a drug. © 2013 Elsevier B.V. All rights reserved.
1. Introduction On-demand drug release represents a promising strategy to potentially increase the therapeutic accuracy and minimize over-dosage induced side effects [1,2]. Ideally, the on-demand release system should release little or no drug in the “off” state, has fast stimuli response and reproducible drug release in each “on” state, and exhibits the ability to adjust drug dosage according to patient needs. More importantly, the system needs to maintain stability upon repeated switch to the “on” state without mechanical disruption.
⁎ Correspondence to: H. Wang, Department of Chemistry, Chemical Biology and Biomedical Engineering, Stevens Institute of Technology, McLean Building Room 416, 1 Castle Point on Hudson, Hoboken, NJ 07030, USA. Tel.: +1 201 216 5556; fax: +1 201 216 8240. ⁎⁎ Correspondence to: S.A. Sukhishvili, Department of Chemistry, Chemical Biology and Biomedical Engineering, Stevens Institute of Technology, McLean Building Room 315, 1 Castle Point on Hudson, Hoboken, NJ 07030, USA. Tel.: +1 201 216 5544; fax: +1 201 216 8240. E-mail addresses: [email protected] (H. Wang), [email protected] (S.A. Sukhishvili). 1 Equal contribution. 0168-3659/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jconrel.2013.06.031
Block copolymer micelles (BCMs) have been explored as promising vesicular drug carriers. The use of stimuli-responsive block copolymers for BCM assembly allows BCM carriers to release their load in response to environmental stimuli, such as pH, temperature, sound, or light [3–6]. Compared to other stimuli such as pH and ionic concentration alterations, tissue cells can better tolerate temperature changes especially in the region lower than physiological temperature. In this regard, temperature-triggered drug release is highly desirable. Temperatureresponsive BCMs can retain and release drugs loaded within their micellar cores as a result of temperature-induced hydrophobic/ hydrophilic transitions [7–10]. Often, this is realized through the inclusion of a BCM core of poly(N-isopropylacrylamide) (PNIPAM) — a widely used temperature-responsive polymer with phase transition occurring at close-to-physiological temperatures. Specifically, the lower critical solution temperature (LCST) of PNIPAM is ~ 32–33 °C [11,12], and this value can be further adjusted within a wide temperature range by using NIPAM-based copolymers. However, uncrosslinked micelles with responsive cores often lose their structural stability, and covalent stabilization through crosslinking of micellar cores or shells is necessary to improve structural stability and achieve tunable drug release [13–15].
74
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
Unlike freely diffusible individual micelles that potentially accumulate in the body, film-embedded micelles can be permanently bound within coatings, preventing systemic distribution of micelles and providing localized therapeutic effects. In this study, we stabilize temperature-responsive drug-delivering micelles by binding them onto a surface using layer-by-layer (LbL) assembly with a branched polymer. The LbL technique affords non-line-of-sight deposition of conformal coatings at substrates of virtually any shape, and allows for control of film structure, composition and delivery of multiple therapeutic components from surfaces [16–18]. With non-micellar LbL films, release of functional compounds through diffusion [19] in response to environmental triggers [20–24], or through film biodegradation [25,26] has been previously demonstrated. However, little is known about the functionality of BCM-containing LbL films as drug delivery films, especially for repeated “on-demand” drug release. While demonstrating drug delivery from BCM-containing films through hydrolysis or biodegradation [27,28], serious concerns have been raised regarding the morphologic and structural integrity of responsive micelles upon external stimuli. For example, morphologic changes were observed with films containing pH responsive BCMs even after a single pH stimulation [29,30]. Our team demonstrated pH-regulated release of a model dye from LbL films with temperature-responsive diblock copolymer micelles [31], however, further work showed that the use of diblock copolymers as BCM components compromised film structural stability during long-term exposure to changing environmental conditions [32], and triblock copolymers were required to assure film stability [33,34]. Moreover, electrostatic pairing responsible for assembly of the above-described films is vulnerable to attack by small ions and cannot be used as a universal strategy for constructing functional films under high-salt physiological environments. Here, we report on the use of hydrogen-bonded assembly to construct high-capacity, temperature-responsive, BCM-containing, reusable films for on-demand release of therapeutic compounds to cells. Previously, we demonstrated the ability of tannic acid (TA) to form stable hydrogen-bonded assemblies with a range of neutral homopolymers, including poly(N-vinylpyrrolidone) (PVPON) [35]. The assembled films were highly stable in a wide pH range and with tolerance to salt concentrations [36]. This all-homopolymer assembly was later used as a non-toxic, cytocompatible coating for encapsulation of yeast cells [37]. Following this strategy, here we report on novel films built via hydrogen-bonded assembly of TA with temperature-responsive micelles, formed from a neutral diblock copolymer, poly(N-vinylpyrrolidone)-b-poly(N-isopropylacrylamide) (PVPON-b-PNIPAM). Temperature-responsive micelles enable films with temperature-controlled swelling/deswelling transitions, as well as pulsed drug release, as a result of the LCST behavior of coreforming PNIPAM blocks. Unlike previous films constructed from similar BCMs but with another hydrogen donor, PMAA [31], the films described here remained highly stable at physiological conditions (PBS, 37 °C). At the same time, the films exhibited temperature-triggered swelling/deswelling transitions, which were used to “switch”, on or off, the release of doxorubicin (DOX), a widely used hydrophobic anticancer drug. To our knowledge, this is the first report on pulsed, on-demand release of drugs from biocompatible hydrogen-bonded temperature-responsive LbL films containing BCMs. Due to the unique capacity of TA to effectively bridge micelles within the film via multisite hydrogen bonding with PVPON in the BCM coronas, the films were able to withstand at least 15 temperature-induced swelling/deswelling cycles and could be reused for repeated DOX loading and release. As a proof-of-concept experiment of functionality of the BCM/TA matrix, we show that pulsed release of DOX from these films effectively eradicated human breast cancer (MCF-7) cells. The robust, reversible, externally controlled swelling and drug release modes make these films promising candidates for on-demand drug release from surfaces.
2. Materials and methods 2.1. Materials Benzyl chloride (Aldrich), elemental sulfur (Aldrich, St Louis, MO), and sodium methoxide (Fluka, St Louis, MO) were used as received. N, N-dimethylformamide (DMF) (99%), dioxane and tetrahydrofuran (THF) were received from Aldrich and distilled under reduced pressure prior to use. N-isopropylacrylamide (NIPAM) and azobis(isobutyronitrile) (AIBN) were received from Aldrich and recrystallized from benzene and methanol, respectively. Water with a resistivity of 18.2 MΩ cm−1 was supplied by a Millipore Milli-Q system. Branched polyethyleneimine (BPEI) with a weight-average molecular weight (Mw) of ~25 kDa, and Mw/Mn = 2.5, and PMAA with Mw ~ 150 kDa, were received from Sigma-Aldrich. DOX-HCl, hydrochloric acid, sodium hydroxide, sodium chloride, dibasic and monobasic sodium phosphate, as well as all other reagents were purchased from Sigma-Aldrich and used as received. 2.2. Synthesis of diblock copolymer PVPON165-b-PNIPAM140 block copolymer was synthesized using reversible addition–fragmentation chain transfer (RAFT) polymerization and characterizations are described in our previous work [31]. 2.3. Preparation and characterization of BCMs 2.3.1. Self-assembly of PVPON-b-PNIPAM BCMs 20 mg of block copolymer PVPON-b-PNIPAM was dissolved in 40 mL of 0.01 M phosphate buffer at pH 5.0, and the solution was gradually heated to 40 °C to enable the formation of BCMs. The presence of micelles in solution was confirmed by atomic force microscopy (AFM) and dynamic light scattering (DLS) as described below. 2.3.2. Dynamic light scattering (DLS) Hydrodynamic sizes of block copolymer solutions were measured using Zetasizer Nano-ZS equipment (Malvern Instruments Ltd, Worcestershire, UK). 2.4. Preparation and characterization of BCM-containing multilayer films 2.4.1. Deposition of BCM/TA films Silicon wafers or glass slides were cleaned as described elsewhere [38]. To enhance the adhesion of subsequently deposited multilayers to the surface, substrates were primed with a BPEI/PMAA bilayer as a precursor by alternately incubating with 0.2 mg/mL BPEI and PMAA solutions for 10 min with two 1-min rinsing cycles in buffer solution between polymer deposition. After rinsing with Milli-Q water, the substrates were then modified with neutral BCM/PMAA multilayers. All the deposition and rinsing steps were performed in 0.01 M pH 5.0 phosphate buffer at 40 °C. Hydrogen-bonded BCM/TA films were deposited by alternately exposing the substrate to respective solutions of 0.2 mg/mL BCM and 0.2 mg/mL TA solution for 10 min, with two 1-min intermediate rinsing steps with 0.01 M phosphate buffer. Deposition always started with BCM solution, and continued until the desired number of bilayers n was deposited. The produced [BCM/TA]n films were dried in air at 40 ºC. 2.4.2. AFM AFM measurements were performed in air using NSCRIPTOR Dip Pen Nanolithography system (Nanoink Inc., Skokie, IL). For AFM imaging, micelles were deposited on silicon wafers by drying BCM-containing solution at 40 °C.
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
2.4.3. Measurements of swelling ratios and refractive indices of BCM/TA films Thicknesses of dry and swollen BCM/TA films were measured by a custom-built phase-modulated ellipsometer at an incidence angle of 70°. Optical properties of the substrates and silicon dioxide layer thickness were determined prior to polymer deposition. In the case of dry polymer films, the refractive index was fixed at a value of 1.5. The swelling of neutral BCM/TA films was measured in situ using a custom-made cylindrical flow-through quartz cell. Briefly, to measure the thickness of swollen films, substrates containing [BCM/TA]4 films were inserted in the glass cells, and cells were filled with PBS. Solution temperature (typically 20 °C or 40 °C) was adjusted using an electric thermal controller and allowed for an equilibrium time of 10 min before measurements. In separate experiments, film thickness was also determined by AFM analysis of the depth profiles of razor-cut dry or wet films. 2.4.4. Confocal laser scanning microscopy (CLSM) Confocal images of films were obtained with an LSM 5 PASCAL laser scanning microscope (Carl Zeiss, Thornwood, NY) equipped with C-Apochromat 63 ×/1.2 water immersion objective. Prior to imaging, films were treated with Alexa 488 according to the procedure recommended by the supplier. 2.5. Preparation of DOX-loaded films Briefly, 2.0 mg of DOX-HCl was dissolved in 1 mL of DMSO, and 0.05 mL of triethylamine was then added into the solution to remove hydrochloride. The DOX solution was added dropwise into 20 mL of PBS at 37 °C. To load DOX within the BCM/TA films, substrates were immersed in the DOX solution for 24 h at 37 °C under stirring. The films were then immersed in a fresh PBS for 24 h at 37 °C to remove free DOX and/or organic solvent. The PBS was replaced every 4 h. To determine the DOX amount loaded within BCM/TA multilayers, 2.5 cm-by-2.5 cm silicon wafers covered with DOX-loaded [BCM/TA]40.5 films were immersed in 10-mL DMSO solution to extract the loaded DOX. The fluorescence intensity of DMSO extract was measured with a fluorometer (excitation at 480 nm), and the amount of DOX was then calculated from a calibration curve obtained with DOX/DMSO solutions of known concentrations. Drug loading content (DLC) was calculated according to the following formula: DLCðwt:% Þ ¼ ðweight of loaded drug=weight of polymerÞ 100%:
2.6. DOX release at different temperatures To measure the release of DOX from free BCMs, 2 mL of DOXloaded BCM solution (0.5 mg/mL, in PBS at pH 7.4 at 37 °C) was transferred into a dialysis membrane bag (MWCO:1000) and then incubated in 20 mL of fresh PBS buffer at 37 °C or 20 °C with gentle shaking. At predetermined time intervals, samples (0.5 mL) were harvested, quickly analyzed by fluorometer (excitation at 480 nm) and then injected back to keep the total volume constant. DOX concentration was determined using a standard curve obtained from DOX solutions of known concentrations in pH 7.4 PBS. Release of DOX from [BCM/TA]40.5 films was similarly measured to that of BCMs mentioned above, except 2.5-cm-by-2.5-cm substrates with DOX-loaded BCM films rather than free BCMs were placed in the dialysis bag. Temperature-triggered release of DOX from [BCM/TA]40.5 films was accomplished by incubating DOX-loaded films in a beaker containing 20 mL of pH 7.4 PBS placed in a 37 °C water bath for a period of 24 h, followed by a quick transfer of the beaker into another water bath at 20 °C for a period of 30 min under magnetic stirring
75
and then transferring the beaker back to the 37 °C water bath. The procedure was repeated for several cycles between 37 °C and 20 °C. Samples (0.5 mL) were harvested from the buffer solutions at defined times, quickly analyzed by fluorometer and injected back to the releasing buffer. The concentration of released DOX was calculated based on the standard curve. In all cases, cumulative release (%) expressed as the total percentage of drug released, was calculated as Wt / W × 100, where Wt is the cumulative weight of drug released by time t, and W is the total weight of drug loaded. 2.7. Cell culture Breast cancer cells (MCF-7, ATCC, Manassas, VA) were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 1% penicillin/streptomycin and 10% fetal bovine serum. The culture was maintained at 37 °C with 5% CO2 and 95% humidity and subculture was routinely performed at a 60–70% cell confluence using 0.25% trypsin-EDTA to detach the cells. 2.8. Tetrazolium salt (MTT) assay MTT assay was used to evaluate the biocompatibility of BCMcontaining films, as well as the effect of temperature-triggered release of DOX on cancer cells. To evaluate film biocompatibility, MCF-7 cells (1.5 × 104 cells per film) were seeded onto the surface of various films and cultured for up to 7 days. The same coverslip surfaces without films were used as controls. To assess the cell-killing efficacy of DOX released from BCM films, MCF (2 × 103 per well) were seeded to 96-well plates and cultured overnight. Then, 100 μL of sample media with or without released DOX was added to each well and incubated for 24 h. For MTT assays, the culture was incubated with thiazolyl blue tetrazolium bromide solution (0.5 mg/mL in culture media) at 37 °C in the dark for 2 h. After discarding nonreacted dye solution, the formazan product was extracted with DMSO and 100 μL of the extract was transferred to a 96-well plate for absorbance measurements. The absorbance was measured at a wavelength of 570 nm with a Synergy HT Multi-Detection Microplate Reader (BioTek Instruments, Winooski, VT). Cell viability was expressed as the percentage of control group: cell viability (%) = Atest / Acontrol × 100%. 2.9. Fluorescence staining for cell morphology and viability To better visualize cell morphology, fluorescence staining for F-actin and cell nuclei were performed as reported previously [39,40]. TRITC-conjugated-phalloidin (Sigma) and SlowFade® Gold antifade reagent with DAPI (Invitrogen, Carlsbad, CA) were used. Briefly, cells were fixed in 3.7% paraformaldehyde in PBS for 5 min and permeabilized with 0.5% Triton X-100 in PBS for another 5 min. The cells were stained with a 50 μg/mL TRITC-conjugated-phalloidin solution at room temperature for 30 min and then washed with PBS to remove unbound phalloidin conjugates. Cell nuclei were stained with DAPI. The stained cells were examined under a Nikon 80i epi-fluorescence microscope with an excitation/emission wavelength at ex 360 nm/em 460 nm (DAPI) and ex 540 nm/em 570 nm (TRITC). To assess cell viability, cultured cells were stained with a LIVE/ DEAD cell vitality assay kit (Invitrogen) following the manufacturer's manual. Briefly, cultured cells were incubated with a staining solution containing 0.125 μM C12-resazurin and 0.02 μM SYTOX in Hank's balanced salt solution (HBSS) for 15 min. Afterwards, the cells were washed with HBSS and fixed with 3.7% paraformaldehyde in PBS for 5 min. The cells were then examined under the fluorescent microscope with an excitation/emission wavelength at ex 490–505 nm/em 530 nm for SYTOX (dead cells) and ex 560–580 nm/em 575 nm for C12-resorufin (live cells), respectively.
76
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
2.10. Statistical analysis All quantitative results were obtained from at least three samples. Each experiment was repeated separately at least three times. Data were expressed as the mean ± standard deviation (SD). An unpaired t-test was used in the statistical analysis of experimental data. A value of p b 0.05 was considered to be statistically significant.
3. Results and discussion 3.1. LbL BCM-containing films and their stability at physiological pH Our design of on-demand drug-releasing films is shown in Fig. 1. Temperature-responsive block copolymer, poly(N-vinylpyrrolidone)b-poly(N-isopropylacrylamide) (PVPON165-b-PNIPAM140), was synthesized by reversible addition–fragmentation chain transfer (RAFT) polymerization as described earlier [31]. Heating the block copolymer solution to temperatures above PNIPAM's LCST of 32–34 °C resulted in the formation of BCMs with a hydrodynamic diameter of ~90 nm and PNIPAM block micellar cores [31]. While in our previous work these micelles were used for assembly with a weak polycarboxylic acid, yielding films unstable at neutral pH values, here we explored the binding of PVPON-b-PNIPAM micelles with another hydrogen-bonding partner, tannic acid (TA) (Fig. 1). Our earlier study has shown that TA forms strong multiple hydrogen bonds with linear neutral polymers, and such hydrogen bonding persists at neutral and slightly basic pH values [35]. In this regard, we expected that binding of TA with copolymer block included within micellar coronae could be equally efficient. Scheme 1 suggests that nonlinear molecules of TA effectively bind BCM micelles within LbL films by forming physical crosslinks between PVPON block in micellar coronas. Fig. 1a shows the thickness growth of [BCM/TA]n films as monitored by ellipsometry and AFM. The data obtained from both techniques were consistent within 10% variation. Nearly linear film growth was observed with [BCM/TA]n films at pH 5, indicating the presence of strong binding between BCM shells and TA. The average bilayer thickness of [BCM/TA]n films as calculated from the slopes of growth curves in Fig. 1a was 37.0 ± 3.5 nm. Fig. 1b shows the evolution of surface morphology of [BCM/TA]n films with an increased thickness. As the number of deposited BCM/TA bilayers exceeded three, the surface became fully covered with micelles, while
Scheme 1. Schematic illustration of the layer-by-layer assembly of PVPON-b-PNIPAM BCMs with TA into stable films at a temperature above PNIPAM's LCST. Right: hydrogen bonding formed between TA and BCMs within the LbL films.
the root mean square (rms) roughness remained modest with a small variation as a function of the number of assembled layers. More specifically, the rms roughness was 24.1, 19.5, 17.9 and 21.3 nm for 1-, 3-, 6-, and 12-bilayer [BCM/TA]n films, respectively. Fig. 1c shows an SEM image of a [BCM/TA]40.5 film (with thickness of 1.49 ± 0.15 μm), deposited on a silicon substrate at 40 °C. Closely-packed spherical micellar morphology suggests structural stability of BCMs within LbL assemblies. One critical aspect in prospective applications of BCM-containing films for drug delivery is their stability at physiological conditions. Fig. 2 illustrates pH-triggered disintegration of BCM/TA films. In our earlier study, we demonstrated that dissociation of multilayers assembled from TA with linear PVPON occurred as a result of deprotonation of weakly acidic phenolic hydroxyl groups of TA, when external pH exceeds the 8.5 pKa of TA [35,41]. Fig. 2 illustrates that [BCM/TA]n films remained stable at neutral and slightly basic pH values. Inset in Fig. 2 shows the effect of immersion time on the stability of [BCM/TA]n films in PBS at 20 °C. Less than 5% thickness decrease was observed for hydrogen-bonded [BCM/TA]n films after immersion in PBS for 24 h. The stability of films at physiological conditions revealed the intrinsic low sensitivity of hydrogen bonding to salts in the solution.
1.0
0.5
Fraction retained
Fraction retained
1.0
in PBS buffer, pH 7.4
0.5 1
3
5
7
Immersion time (d)
0.0 4
6
8
10
pH Fig. 1. (a) AFM images of PVPON-b-PNIPAM [BCM/TA]n films with n = 1, 3, 6, and 12 (scanned area 2 μm × 2 μm). (b) Dry thicknesses of PVPON-b-PNIPAM [BCM/TA]n films as a function of bilayer number deposited at 40 °C as monitored by ellipsometry and AFM. In ellipsometic measurements, the film refractive index was fixed at 1.5. (c) SEM image of the surface morphology of PVPON-b-PNIPAM [BCM/TA]40.5 films. Inset: photograph of a [BCM/TA]40.5 film deposited on a 2.5 cm × 2.5 cm glass coverslip. BCM/TA films were deposited from solutions of pH 5.0 at 40 °C and dried in air at 40 °C.
Fig. 2. Thickness loss of dry [BCM/TA]8 films as a function of pH in PBS at 20 °C. 0.5 M NaOH solution was used for pH adjustment. Thickness retained on silicon substrates was measured by ellipsometry after exposure to buffer solutions for 30 min and dried. Fraction retained was calculated as the percentage of the original dry film thickness. Inset shows fraction of a [BCM/TA]8 film as a function of immersion time in PBS buffer at 20 °C. Films immersed in pH 7.4 buffer solutions were removed at defined time intervals, dried for thickness measurements by ellipsometry and then immersed back to pH 7.4 buffer solutions.
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
Fig. 3. (a) In situ ellipsometry measurements of the swelling ratio and refractive index of a [BCM/TA]4 film (dry thickness ~ 115 nm) immersed in PBS solution at pH 7.4 upon temperature cycling between 20 °C and 40 °C. (b) CLSM imaging of the thickness of [BCM/TA]40.5 films at 20 °C (top) and 40 °C (bottom). Films were stained with Alexa 488. The scale bars are 5 μm. (c) AFM images of the surface morphology of [BCM/ TA]40.5 films at 20 °C (top) and 40 °C (bottom). (d) Schematic illustration of temperature-triggered swelling transitions of BCM-containing films.
3.2. Temperature-triggered swelling of [BCM/TA] films at physiological conditions With successful demonstration of the structural stability of BCM/TA films at physiological pH, we explored temperature-controlled water uptake within assembled micellar PNIPAM cores and the resultant swelling of LbL films. The temperature effect on the film swelling was studied using phase-modulated in situ ellipsometry. Fig. 3a shows variations in thickness and the refractive index of wet [BCM/TA]4 films when cycling the solution temperature between 20 °C and 40 °C. At 40 °C, when PNIPAM cores of BCMs dehydrated, the film swelled from its dry state to a modest swelling ratio of 1.2 ± 0.07, with a refractive index decrease from 1.5 to 1.42. The small degree of swelling at 40 °C reflects the presence of collapsed PNIPAM cores within BCM micelles, and dehydration of BCM/TA film as a result of hydrogenbonding between PVPON coronal blocks and TA. After cooling the solution to 20 °C, the BCM/TA film swelling ratio increased to ~1.75 ± 0.06 (corresponding to a film thickness of 210 nm), and its refractive index decreased to 1.36 ± 0.07, suggesting the substantial hydration of the BCM/TA film below LCST. CLSM confirmed dramatic changes in thickness of swollen [BCM/TA]40.5 films from 2.8 ± 0.2 μm at 20 °C (top) to 1.7 ± 0.2 μm at 40 °C (bottom) (Fig. 3b). Fig. 3c illustrates a corresponding shrinkage of assembled micelles, with a decrease in their average lateral size from 160 ± 15 nm at 20 °C to 105 ± 10 nm at 40 °C, indicating dehydration of PNIPAM micellar cores above LCST. At the same time, hydrogen bonding between TA and PVPON is not affected by temperature variations; as indicated by Fig. 3, binding of BCM coronal chains with TA prevents dissociation of BCMs into unimers even at low temperature, while allowing reversible volumetric expansion and contraction of micellar PNIPAM cores and swelling transitions within the entire BCM-containing films (Fig. 3d).
77
attributed to the entrapment of DOX within hydrophobic PNIPAMforming cores (at temperatures higher than PNIPAM's LCST). Note that a value of 0.4% has been reported in our previous work for inclusion of a hydrophobic fluorescent dye (pyrene) within BCMs assembled within BCM/PMAA films [31], which were unstable at physiological conditions. Here, much higher loading capacity was achieved with BCM/TA films, which were highly stable at neutral and slightly basic pH range due to hydrogen-bonding-assisted bridging of micellar shells by TA molecules. To determine the biocompatibility of drug-loaded films, MCF-7 cancer cells were cultured on the surface of [BCM/TA]n films for up to 7 days. The cell proliferation on the films was evaluated by using MTT assay. Although the MTT assay is not accurate to determine the exact cell number, it measures cell metabolic activity, which is proportional to cell number under a similar culture condition [42]. Therefore, MTT assay can be used to roughly indicate the cell proliferation trend. Considering that films could be constructed with either TA or BCM as the outermost layer, we evaluated both film types in cell culture experiments. Fig. 4b shows viable cell numbers of MCF-7 cells at DOX-loaded [BCM/TA]n films after culturing for 1, 3 and 7 days. When experiments were performed at 37 °C, [BCM/TA]n films with either BCM or TA as a top layer supported cell proliferation similarly to control glass coverslip surfaces (Fig. 4b). After culturing for 1, 3 and 7 days, cell morphology was observed by fluorescent phalloidinTRITC staining of the cytoskeletal protein filament actin. The fluorescent images revealed that morphologies of all the cells cultured on LbL films remained similar to those observed with the glass coverslip control group, with a typical cobblestone-like morphology in a colony (Fig. 4c). In addition, the average colony size and number were also comparable among various groups (data not shown). This observation agrees well with the cell proliferation results, suggesting compatibility of both bare and DOX-loaded [BCM/TA]n films to MCF-7 at 37 °C. While cellular response was independent of the film top layer, we have used [BCM/TA]40.5, i.e. films containing BCMs as the outmost layer, in our further experiments.
3.3. DOX-loaded BCM/TA films and their biocompatibility Because of the known adverse side effects of DOX, entrapment of DOX within micelles (Fig. 4a) can minimize potential damage to healthy tissues. Loading of DOX within micellar PNIPAM cores was achieved by immersing [BCM/TA]40.5 films in the DOX solution and dialyzing against water at 37 °C. For [BCM/TA]40.5 film, a high DOX loading content of 9.3% of the film mass was obtained as determined by fluorometry. The high amount of DOX loaded within the LbL film is
Fig. 4. (a) Schematic representation of a BCM/TA film retaining DOX in PBS solution at pH 7.4 and 37 °C. (b) Proliferation of MCF-7 cells cultured on DOX-loaded [BCM/TA] films with TA top layer ([BCM/TA]40 film) or BCM top layer ([BCM/TA]40.5 film) at a constant temperature of 37 °C. MCF-7 cells at a density of 1.5 × 104 cells/well were cultured on the films for 1, 3 and 7 days prior to MTT assay. Glass coverslip surface was used as a control. (c) Morphologies of MCF-7 cells on the control surface, as well as on [BCM/TA]40 and [BCM/TA]40.5 films. Cytoskeletal F-actin was stained red with TRITC-phalloidin, and cell nuclei were stained blue with DAPI. Scale bar is 50 μm.
78
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
Fig. 5. (a) Cumulative release of DOX from free BCMs (circles) or from [BCM/TA]40.5 films (squares) in PBS buffer (pH 7.4) at 20 °C (open symbols) and 37 °C (filled symbols). DOX release was monitored by fluorometry (excitation at 480 nm), and released amounts normalized to the total amount of DOX loaded within the film. (b) Release profiles of DOX from [BCM/TA]40.5 films immersed in PBS at a constant temperature of 37 °C, and cooled down to 20 °C for 30 min at cumulative time of 1, 2 and 3 days. Inset shows percentage of DOX released in each cooling cycle. (c) Schematic representation of “on-demand” DOX release from BCM-containing films.
3.4. Temperature-triggered release of DOX from BCM/TA films With successful entrapment of DOX into BCM-containing hydrogenbonded films, it becomes essential to understand how temperature regulates DOX on-demand release. Fig. 5a illustrates the rate of accumulation of DOX released from [BCM/TA]40.5 films during continuous exposure to PBS at a constant temperature of 20 °C or 37 °C. It is necessary to mention that although 20 °C was continuously used for DOX release in this study, in order to be consistent with above experiments, any other temperature below PNIPAM's LCST of 32 °C could equally serve the purpose. For comparison, DOX released from individual BCMs in solution is also shown in Fig. 5a. DOX release from free BCMs was very fast at 20 °C, with nearly 85% of incorporated DOX released during the initial 30 min, reflecting dissociation of micelles to unimers when the solution temperature was below PNIPAM's LCST. At temperatures above LCST, hydrophobic cores of micelles efficiently retained hydrophobic DOX, leading to slower DOX release. Indeed, at 37 °C the release of DOX was drastically slower, with only ~25% DOX released within the first day and very slow release at longer times. Binding of BCMs with TA within the LbL film significantly slowed DOX release as compared to its release from free BCMs. At 37 °C and pH 7.4, only ~10% of DOX was released the first day, and a large amount of DOX was retained in the [BCM/TA]40.5 films even after 4-day incubation. Remarkably, when the temperature was lowered to 20 °C, DOX release from [BCM/TA]40.5 films was greatly accelerated, quickly reaching ~18–20% within 30 min. The temperature-triggered DOX release patterns strongly correlated with the results of temperature-controlled swelling transitions of [BCM/TA]n films in Fig. 3, indicating that retention/release of DOX within/from the film was dictated by differential partitioning the hydrophobic drug in collapsed (above PNIPAM's LCST) or hydrated (below PNIPAM's LCST) micellar cores. In addition to demonstrated continuous release of DOX from the assembled films at temperatures below LCST, it would be beneficial to achieve pulsed release of drugs from the films. To that end, a cyclic thermal stimulation at temperatures above and below LCST was applied to the DOX-loaded [BCM/TA]n films. Fig. 5b shows the release kinetics of DOX from [BCM/TA]40.5 films when films were immersed in PBS at 37 °C, and cooled down to 20 °C for 30 min at days 1, 2 and 3, respectively. Pulsed release of DOX was observed with a step-like increase in cumulative
DOX concentration in solution, with every exposure to 20 °C leading to a burst release of 12–16% of loaded DOX. Note that the released dose decreased from ~16% to ~14% and ~12% in the 1st, 2nd, and 3rd temperature lowering cycles. The lower amounts released in the 2nd and 3rd cycles are attributed to the decreased total ‘supply’ of DOX within the film. In practice, it would be an appealing feature for the films to be recharged with drugs for continued use. Here, we explored the possibility of reusing BCM/TA films for repeated temperature-triggered drug release. Specifically, after complete release of DOX from [BCM/TA]40.5 constructs, films were reloaded with DOX and re-evaluated for their drug release profiles. Interestingly, DOX-reloaded films exhibited essentially identical release profiles to those observed with freshly prepared DOX-loaded BCM films (Fig. 6). Importantly, we were able to perform as many as 15 film re-loading cycles, without significant changes in the DOX release profiles (data not shown). Also note that films repeatedly reloaded with DOX demonstrated highly reproducible pulsed release patterns (Fig. 5b). The ability to be replenished with bioactive compounds is an important feature in designing medical devices [43], and our BCM/TA films seem to satisfy this requirement. The fact that drug-hosting micellar cores preserve their capacity to DOX partitioning even after many film swelling/deswelling cycles suggest no significant morphological changes within non-covalently assembled films. The result is encouraging but yet somewhat surprising, considering the non-covalent assembly of film components. In contrast to previous findings that electrostatically assembled films of temperatureresponsive diblock copolymer micelles showed significant mass loss during long-term exposure to low temperature [34], here, we demonstrate that the use of TA – a non-linear hydrogen-bonding molecule – assures strong binding and efficient stabilization of BCMs within a film during temperature variations, and does not require the use of triblock copolymers in order to enhance the stability of the polymer matrix.
3.5. In vitro efficiency of BCM/TA films in eradicating cancer cells In order to demonstrate film functionality, a proof-of-concept experiment was designed to explore whether the temperaturetriggered release of DOX from [BCM/TA]40.5 films was therapeutically efficient in killing breast cancer cells (MCF-7) as a model. Moreover, we were seeking to explore whether each temperature cycle resulted in reproducible and repeatable cell killing patterns. To that end, release media samples were collected from DOX-loaded [BCM/TA]40.5 films for each of three temperature cycles (one cycle included exposure of the film to 37 °C for 1 day, followed by lowering the
Fig. 6. Representative release profiles of DOX from repeatedly used, DOX-reloaded [BCM/TA]40.5 films in PBS solutions (pH 7.4) at 20 °C (open symbols) and 37 °C (filled symbols). Prior to drug reloading, DOX loaded within the film during previous loading cycles was completely removed by immersing the films in PBS buffer for 7 days at 20 °C.
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
79
with those in Fig. 5b, showing that DOX-loaded [BCM/TA]40.5 films exhibited a very slow release of loaded drug at physiological temperature, while temporarily cooling the films below the LCST of PNIPAM remarkably accelerated the release of DOX. The repeatable tunability of biological activity of the [BCM/TA]n films via simple temperature variations renders such films a promising platform for on-demand release devices. Moreover, by tailoring the size and chemical nature of micellar cores for example, by endowing this block high critical solution temperature rather than LCST behavior at the step of polymer synthesis, and by controlling the number of BCM layers within LbL films, as well as the duration of temperature variation cycles, this platform can be programmed to release desired doses of therapeutics at many scheduled time points. Therapeutic dosages of drugs can be delivered on demand within short periods of time at desirable temperatures, and the delivery can then be “switched off” at physiological temperatures in order to minimize undesirable side effects on healthy tissues. 4. Conclusions
Fig. 7. (a) Cytotoxicity of [BCM/TA]40.5 films against MCF-7 cells after multiple temperature cycles between 37 °C and 20 º°C. Films were immersed in culture medium at a constant temperature of 37 °C, and then cooled down to 20 °C for 30 min at cumulative time of 1, 2 and 3 days. After each 37 °C/20 °C cycling, the medium was separately collected and replaced with fresh medium at 37 °C to start a new release cycle. MCF-7 cells cultured overnight were treated with collected medium samples for 24 h and analyzed by MTT assay to determine cell viability. (b) Immunofluorescence images of MCF-7 cells cultured for 24 h. F-actin was stained red with phalloidin-TRITC and cell nuclei stained blue with DAPI. (c) LIVE/DEAD staining of MCF-7 cultured for 24 h. Red = live; green = dead.
temperature to 20 °C for 30 min), and used to treat MCF-7 cells for MTT assay. The cell viability assay showed that media samples from DOX-loaded [BCM/TA]40.5 films at 37 °C had negligible cytotoxicity against MCF-7 cells, with only a small reduction (~7%) in cell viability for 24 h of incubation (Fig. 7a). In contrast, the release samples from 20 °C for 30-min cycles significantly (p b 0.001) reduced cell viability to ~ 7% (1st), 9% (2nd), and 10% (3rd) compared to untreated controls, respectively. Immunofluorescent staining for cell morphology with phalloidin-TRITC and cell viability with LIVE/DEAD assay kit further confirmed the MTT results, showing vast differences between 1-day at 37 °C and 30-min at 20 °C samples. MCF-7 cells treated with 37 °C media samples retained their cobblestone morphology with all the cells alive, similar to that of control groups, while only cell nuclei were seen in the culture treated with 20 °C samples and no living cells were identified (Fig. 7b and 7c). These results indicated that DOX-loaded [BCM/TA]40.5 films maintained high bioactivity even after multiple temperature cycles. The results also showed that at each 20 °C incubation cycle, the amount of released DOX was sufficiently high to eradicate breast cancer cells (MCF-7), while the DOX amount released at 37 °C was well below the therapeutic dose, resulting in minimal cell killing even after several days of contact with MCF-7 cells. The data in Fig. 7 are in excellent agreement
We have constructed a new type of hydrogen-bonded, non-toxic, BCM-containing LbL films, which demonstrated high stability at physiological pH and ionic concentrations, and exhibited highly sensitive swelling/deswelling response to temperature variations. The films were highly efficient in entrapping high amounts of a hydrophobic drug, DOX, within PNIPAM cores of assembled BCMs and supported on demand or pulsated temperature-controlled drug release. While both drug-free and DOX-loaded BCM/TA films were not cytotoxic at 37 °C, lowering the temperature to 20 °C for 30 min triggered release of DOX doses sufficient to eradicate breast cancer cells with N90% killing efficacy. The large-amplitude, reversible, temperature-induced swelling transitions also enabled repeatable, temperature-regulated on/off release of DOX from BCM-containing films, with similar cell killing capacity of the drug in each cycle. Importantly, films could be reloaded with the drug at least 15 times, giving reproducible drug release profiles in multiple release cycles. Our results demonstrate strong potential of novel hydrogen-bonded, BCM-containing LbL films as high-efficiency, reusable matrices for temperature-regulated release of hydrophobic drugs. While this study presents a proof-of-concept demonstration of the application of these films for repeatable on-demand release of bioactive compounds from surfaces, these matrices can be integrated with a broad range of biomedical devices (e.g. wound dressing) for on-demand release of antibiotics and growth factors to stimulate tissue regeneration. Acknowledgments The authors thank the National Science Foundation (S. Sukhishvili, Award DMR-0906474) and the National Institute of Arthritis and Musculoskeletal and Skin Diseases (H. Wang, Award 1R21 AR056416) for financial support of this work. The image resources used in this study were partially funded by the National Science Foundation through NSF Grant DMR-0922522. References [1] B.P. Timko, T. Dvir, D.S. Kohane, Remotely triggerable drug delivery systems, Adv. Mater. 22 (44) (2010) 4925–4943. [2] Y. Shao, W. Huang, C. Shi, S.T. Atkinson, J. Luo, Reversibly crosslinked nanocarriers for on-demand drug delivery in cancer treatment, Ther. Deliv. 3 (12) (2012) 1409–1427. [3] G. Gaucher, M.H. Dufresne, V.P. Sant, N. Kang, D. Maysinger, J.C. Leroux, Block copolymer micelles: preparation, characterization and application in drug delivery, J. Control. Release 109 (1–3) (2005) 169–188. [4] V.P. Torchilin, Micellar nanocarriers: pharmaceutical perspectives, Pharm. Res. 24 (1) (2007) 1–16. [5] N. Rapoport, Physical stimuli-responsive polymeric micelles for anti-cancer drug delivery, Prog. Polym. Sci. 32 (8–9) (2007) 962–990. [6] A.W. York, S.E. Kirkland, C.L. McCormick, Advances in the synthesis of amphiphilic block copolymers via RAFT polymerization: stimuli-responsive drug and gene delivery, Adv. Drug Deliv. Rev. 60 (9) (2008) 1018–1036.
80
Z. Zhu et al. / Journal of Controlled Release 171 (2013) 73–80
[7] D. Neradovic, C.F. Van Nostrum, W.E. Hennink, Thermoresponsive polymeric micelles with controlled instability based on hydrolytically sensitive N-isopropylacrylamide copolymers, Macromolecules 34 (2001) 7589–7591. [8] C.J. Rijcken, O. Soga, W.E. Hennink, C.F. van Nostrum, Triggered destabilization of polymeric micelles and vesicles by changing polymers polarity: an attractive tool for drug delivery, J. Control. Release 120 (3) (2007 Jul 31) 131–148. [9] H. Wei, S.X. Cheng, X.Z. Zhang, R.X. Zhuo, Thermo-sensitive polymeric micelles based on poly(N-isopropylacrylamide) as drug carriers, Prog. Polym. Sci. 34 (9) (2009) 893–910. [10] X. He, X. Wu, C. Gao, K. Wang, S. Lin, W. Huang, M. Xie, D. Yan, Synthesis and self-assembly of a hydrophilic, thermo-responsive poly(ethylene oxide) monomethyl ether-block-poly(acrylic acid)-block-poly(N-isopropylacrylamide) copolymer to form micelles for drug delivery, React. Funct. Polym. 71 (2011) 544–552. [11] S. Fujishige, K. Kubota, I. Ando, Phase transition of aqueous solutions of poly(Nisopropylacrylamide) and poly(N-isopropylmethacrylamide), J. Phys. Chem. 93 (8) (1989) 3311–3313. [12] F.M. Winnik, Fluorescence studies of aqueous solutions of poly(N-isopropylacrylamide) below and above their LCST, Macromolecules 23 (1990) 233–242. [13] X. Jiang, Z. Ge, J. Xu, H. Liu, S. Liu, Fabrication of multiresponsive shell cross-linked micelles possessing pH-controllable core swellability and thermo-tunable corona permeability, Biomacromolecules 8 (10) (2007) 3184–3192. [14] Y. Li, B.S. Lokitz, S.P. Armes, C.L. McCormick, Synthesis of reversible shell cross-linked micelles for controlled release of bioactive agents, Macromolecules 39 (8) (2006) 2726–2728. [15] Y. Zeng, W.G. Pitt, Poly(ethylene oxide)-b-poly(N-isopropylacrylamide) nanoparticles with crosslinked cores as drug carriers, J. Biomater. Sci. Polym. Ed. 16 (3) (2005) 371–380. [16] S. Pavlukhina, S. Sukhishvili, Polymer assemblies for controlled delivery of bioactive molecules from surfaces, Adv. Drug Deliv. Rev. 63 (9) (2011) 822–836. [17] A.J. Karlsson, R.M. Flessner, S.H. Gellman, D.M. Lynn, S.P. Palecek, Polyelectrolyte multilayers fabricated from antifungal β-peptides: design of surfaces that exhibit antifungal activity Candida albicans, Biomacromolecules 11 (9) (2010) 2321–2328. [18] B.-S. Kim, R.C. Smith, Z. Poon, P.T. Hammond, Mad (multiagent delivery) nanolayer: delivering multiple therapeutics from hierarchically assembled surface coatings, Langmuir 25 (2009) 14086–14092. [19] M.C. Berg, L. Zhai, R.E. Cohen, M.F. Rubner, Controlled drug release from porous polyelectrolyte multilayers, Biomacromolecules 7 (1) (2006) 357–364. [20] B.M. Wohl, J.F. Engbersen, Responsive layer-by-layer materials for drug delivery, J Control. Release 158 (1) (2012) 2–14. [21] E. Kharlampieva, S.A. Sukhishvili, Release of a dye from hydrogen-bonded and electrostatically assembled polymer films triggered by adsorption of a polyelectrolyte, Langmuir 20 (2004) 9677–9685. [22] S. Pavlukhina, Y. Lu, A. Patimetha, M. Libera, S. Sukhishvili, Polymer multilayers with pH-triggered release of antibacterial agents, Biomacromolecules 11 (12) (2010) 3448–3456. [23] X. Wang, L. Zhang, L. Wang, J. Sun, J. Shen, Layer-by-layer assembled polyampholyte microgel films for simultaneous release of anionic and cationic molecules, Langmuir 26 (11) (2010) 8187–8194. [24] M.J. Serpe, C.D. Jones, L.A. Lyon, Layer-by-layer deposition of thermoresponsive microgel thin films, Langmuir 19 (2003) 8759–8764.
[25] E. Vázquez, D.M. Dewitt, P.T. Hammond, D.M. Lynn, Construction of hydrolyticallydegradable thin films via layer-by-layer deposition of degradable polyelectrolytes, J. Am. Chem. Soc. 124 (47) (2002) 13992–13993. [26] J. Zhang, L.S. Chua, D.M. Lynn, Multilayered thin films that sustain the release of functional DNA under physiological conditions, Langmuir 20 (19) (2004) 8015–8021. [27] B.S. Kim, H.I. Lee, Y. Min, Z. Poon, P.T. Hammond, Hydrogen-bonded multilayer of pH-responsive polymeric micelles with tannic acid for surface drug delivery, Chem. Commun. 28 (2009) 4194–4196. [28] B.S. Kim, S.W. Park, P.T. Hammond, Hydrogen-bonding layer-by-layer-assembled biodegradable polymeric micelles as drug delivery vehicles from surfaces, ACS Nano 2 (2) (2008) 386–392. [29] T. Addison, O.J. Cayre, S. Biggs, S.P. Armes, D. York, Incorporation of block copolymer micelles into multilayer films for use as nanodelivery systems, Langmuir 24 (23) (2008) 13328–13333. [30] I. Erel, Z. Zhu, A. Zhuk, S.A. Sukhishvili, Hydrogen-bonded layer-by-layer films of block copolymer micelles with pH-responsive cores, J. Colloid Interface Sci. 355 (1) (2011) 61–69. [31] Z. Zhu, S.A. Sukhishvili, Temperature-induced swelling and small molecule release with hydrogen-bonded multilayers of block copolymer micelles, ACS Nano 3 (11) (2009) 3595–3605. [32] L. Xu, Z. Zhu, S.A. Sukhishvili, Polyelectrolyte multilayers of diblock copolymer micelles with temperature-responsive cores, Langmuir 27 (1) (2011) 409–415. [33] W.S. Tan, R.E. Cohen, M.F. Rubner, S.A. Sukhishvili, Temperature-induced, reversible swelling transitions in multilayers of a cationic triblock copolymer and a polyacid, Macromolecules 43 (4) (2010) 1950–1957. [34] W.S. Tan, Z. Zhu, S.A. Sukhishvili, M.F. Rubner, R.E. Cohen, Effect of block copolymer architecture on the thermally induced swelling of micelle-containing multilayer thin films, Macromolecules 44 (19) (2011) 7767–7774. [35] I. Erel-Unal, S.A. Sukhishvili, Hydrogen-bonded multilayers of a neutral polymer and a polyphenol, Macromolecules 41 (11) (2008) 3962–3970. [36] E. Kharlampieva, S.A. Sukhishvili, Hydrogen-bonded layer-by-layer films, Journal J. Macromol. Sci. C Polym. Rev. J. 46 (2006) 377–395. [37] V. Kozlovskaya, S. Harbaugh, I. Drachuk, O. Shchepelina, N. Kelley-Loughnane, M. Stone, V.V. Tsukruk, Hydrogen-bonded LbL shells for living cell surface engineering, Soft Matter 7 (2011) 2364–2372. [38] V. Kozlovskaya, S. Yakovlev, M. Libera, S.A. Sukhishvili, Surface priming and the self-assembly of hydrogen-bonded multilayer capsules and films, Macromolecules 38 (11) (2005) 4828–4836. [39] T. Wieland, Phallotoxins, Springer-Verlag, New York, 1986. [40] M. Kubista, B. Akerman, B. Nordén, Characterization of interaction between DNA and 4′,6-diamidino-2-phenylindole by optical spectroscopy, Biochemistry 26 (14) (1987) 4545–4553. [41] W. Bors, C. Michel, K. Stettmaier, Electron paramagnetic resonance studies of radical species of proanthocyanidins and gallate esters, Arch. Biochem. Biophys. 374 (2) (2000) 347–355. [42] E. Vega-Avila, M.K. Pugsley, An overview of colorimetric assay methods used to assess survival or proliferation of mammalian cells, Proc. West. Pharmacol. Soc. 54 (2011) 10–14. [43] R. Lo, P.Y. Li, S. Saati, R. Agrawal, M.S. Humayun, E. Meng, A refillable microfabricated drug delivery device for treatment of ocular diseases, Lab Chip 8 (7) (2008) 1027–1030.