Biodegradable pH-sensitive polyurethane micelles with different polyethylene glycol (PEG) locations for anti-cancer drug carrier applications

Abstract

Biodegradable multi-blocked polyurethane (PU) based micelles with a hydrophilic PEG corona were extensively studied for anti-cancer drug delivery systems. The hydrophilic PEG segment usually incorporated as a soft segment or as an end capping reagent, which has difficulty forming a dense PEG coating with a brush like conformation due to the low mobility of the PEG domains at the soft segment and the low amount of the PEG at the end of PU chains. In the present study, biodegradable pH sensitive polyurethane micelles with a dense brush like coating of PEG were prepared by a new kind of PEG grafted polyurethanes (PEG-g-PU) which were synthesized using PEGylated diethanolamine (MPEG-DEAM) as a chain extender. The high mobility of pendant MPEG in PEG-g-PU results in the formation of a dense and brush like PEG corona on the PEG-g-PU micelles. Meanwhile the MPEG attached on the hard segment will transfer the diethanolamine (DEAM) to the surface of nanoparticles during the self-assembly process and the DEAM render the particles with positive charges which potentially enhances cellular uptake and endosomal escape. DLS, TEM and AFM showed that a dense PEG domain was formed on the surface of PEG-g-PU micelles while no obvious PEG microdomain was observed for the other two kinds of micelles. FTIR and DSC results demonstrated the enhanced microphase separation of PEG-g-PU micelles compared with PEG-g-PU bulk materials and the other two contrast PU micelles, i.e. PEG-b-PU and PEG-c-PU. Paclitaxel (PTX) was chosen as a model hydrophobic drug to evaluate the loading and pH-triggered release of the PU micelles. The enhanced cytotoxicity of PTX-loaded PEG-g-PU-3 micelles against H460 cancer cells reveals that they are more potent for intracellular delivery of PTX as compared to PEG-b-PU-3 and PEG-c-PU-3 micelles.

---

1. Introduction

Various self-assembled micelles based on amphiphilic block copolymers have been developed in recent years for anticancer drug delivery systems.^1–5 The hydrophobic core of polymer micelles improves the solubility of the hydrophobic drugs by physical encapsulation while the hydrophilic corona of polymer micelles stabilizes the colloid in an aqueous medium. Meanwhile, the nanocarriers improve the solubility of hydrophobic drugs, prolong the circulation time of the drugs and accumulate passively on the solid tumor by an Enhanced Permeability Effect (EPR).⁶ After accumulating at the tumor tissue, rapid intracellular drug release is particularly important to ensure a sufficient drug concentration for killing the tumor.⁷ Thus, it is reasonable to construct smart nanocarriers which respond simultaneously to the stimuli of the intracellular micro-environment of tumor, i.e. pH,^8,9 redox^8,10–12 and enzyme.¹³ The intracellular pH value of cancer is much lower than that of the normal tissues due to the high rate of glycolysis in cancer cells.^14,15 Therefore, pH-sensitive nanoparticles that are prone to swelling or dissolution at endosomal/lysosomal pH have been designed for intracellular drug delivery in cancer therapy. Among the different types of nanocarriers based on pH-sensitive polymers, cationic polymers with amine-functional groups incorporating into the main-chain or side-chain have positive surface charge which potentially enhances cellular uptake and endosomal escape.^16,17 However, one significant obstacle to in vivo applications of cationic polymers is their clearance by the reticuloendothelial system, which leads to inefficient targeting to specific sites in the body.¹⁸

Stealth coating on the surface of cationic polymer nanoparticles (NPs) with hydrophilic poly(ethylene glycol) (PEG) is extensively used to prevent the aggregation, reduce clearance and extend circulation time in the blood stream by rendering the particles high resistance to protein adsorption.¹⁹ The PEG layer on NPs surface can also improve drug encapsulation by providing a physico-chemical barrier to drug escape and it could affect drug release pattern.²⁰ In addition to the above mentioned biological properties, it has been recently demonstrated that very dense coating of PEG (about 0.5–1 PEG per nm² for 200 nm diameter particle) can achieve particles transport coefficients in mucus comparable to those in liquid medium which can potentially improve drug delivery efficacy as electrostatic and steric hindrances prevent NPs to cross mucosal barriers²¹ or to penetrate tissues beyond the perivascular region.²² It is reported that the biological properties of the PEGylation polymer particles depended on the chain coverage-density of PEG layer. The resistance to protein binding (the so-called “antifouling effect”) is dependent on PEG chain coverage conformation and is usually achieved at high coverage-density, in the polymeric brushes regime.²³

Multi-blocked polyurethanes (MPUs) are a flexible platform of materials that can be designed to fit the requirements for different applications.^24–26 Biodegradable multi-blocked polyurethane based micelles with a hydrophilic PEG corona were extensively studied for anti-cancer drug delivery systems.^4,27,28 The hydrophilic PEG segment was usually incorporated as soft segment (PEG-b-PU)¹ or as an end capping reagent (PEG-c-PU).^4,28 However, it is difficult to form a dense PEG coating with a brush like conformation due to the low mobility of the PEG domains at the soft segment and the low amount of the PEG at the end of PU chains. In this study, the major objective was to optimize NPs formulation using three PEGylated cationic polyurethanes of different PEG location, one is MPEG grafted on the chain extenders (PEG-g-PU), the second is PEG copolymerized into PU as soft segment (PEG-b-PU), and the third is MPEG attached on the end of PU chain (MPEG-c-PU). We speculated that the biodegradable pH sensitive polyurethane micelles with a dense coating of PEG could be developed by using poly(ε-caprolactone) (PCL) as soft segment, and N-(methoxypolyethylene glycol)-diethanolamine (MPEG-DEAM) as chain extender (PEG-g-PU) due to the relatively higher mobility of grafted PEG than that located at soft segment (PEG-b-PU) and attached on the endcap (MPEG-c-PU). The MPEG attached on the hard segment will transfer the DEAM to the surface of NPs during self-assembly process and the DEAM render the particles with positive charges which potentially enhances cellular uptake and endosomal escape. The MPEG that stretch out and formed a dense brush-like PEG coating on the particles prolong the circulation time of the micelles. Furthermore, the tertiary amine under the PEG stealth protonated in acid environment and disassembled and released the cargos into the cancer cells rapidly. PTX was chosen as mode hydrophobic drug to evaluate the loading and pH triggered release profiles of the PU micelles.

2. Experimental

2.1 Materials

Polyethylene glycol (PEG), methoxypolyethylene glycols (MPEG) [M_n = 1000, Sunshine Biotechnology (Nanjing) Co., Ltd. China] and poly(ε-caprolactone)diol (PCL) (M_n = 2000, Dow Chemical, USA) were dehydrated under reduced pressure at 100 °C for 2–3 h before use. N,N-Dimethylacetamide (DMAc) was dried over CaH₂, vacuum distilled before use. Isophoronediisocyanate (IPDI), diethanolamine and p-toluenesulfonyl chloride (p-TsCl, Aladdin Industrial Corporation, China) were used as received. Triethylamine (TEA, Aladdin Industrial Corporation, China) was distilled under vacuum before use. A dialysis tubing with molecular weight cutoff from 8000 to 14 000 was purchased from Sinopharm Chemical Reagent Co., Ltd (SCRC, Shanghai, China). PTX (99.5%) was obtained from AstaTech (Chengdu) Pharmaceutical Co. Ltd. China. Doxorubicin hydrochloride (DOX, –NH₃⁺Cl⁻ salt form, >98%) was obtained from AstaTech (Chengdu) Pharmaceutical Co. Ltd. China.

2.2 Synthesis of the polyurethanes

The three kinds of polyurethanes were synthesized from the same amount of PEG, PCL, IPDI and tertiary amine (DEMA and/or MPEG-DEAM) with different raw materials and synthesis processes, which resulted in different locations of PEG on the polyurethane, i.e. PEG-g-PU, PEG-b-PU and PEG-c-PU. The feed ratios are listed in Table 1. MPEG-g-PU was synthesized from PCL, IPDI, BDO and MPEG-DEMA by the traditional two step solution polymerization process, i.e. pre-polymerization and chain extend step.^12,28 In brief, anhydrous PCL-diol was dissolved in DMAc under a dry nitrogen atmosphere. IPDI was added into DMAc solution of PCL and pre-polymerized in the presence of 0.1% stannous octoate (Sn(Oct)₂) at 90 °C for 4 h under stirring. Then, the prepolymer solution was cooled to 60 °C and the chain extender MPEG-DEMA was added, and the mixture was reacted at 90 °C for 4 h to complete the chain extending reaction. After that, the reaction mixture was cooled to room temperature. To remove the organic solvents and impurities, the crude product was subsequently poured into a water and methanol mixtures with volume ratio of 3 : 1. The precipitant was redissolved in DMAc and precipitated by water and methanol mixtures again, which could remove low M_w residues in PUs. Then the polymer was harvested and dried at 40 °C in atmosphere for 5 h and at 60 °C in reduced pressure for 2 days to get the finally product. PEG-b-PU and PEG-c-PU were synthesized from PCL, PEG, IPDI, BDO and DEMA by the traditional two step solution polymerization process similar to the synthesis process of MPEG-g-PU.

Table 1 Composition and characteristics of pH-sensitive polyurethanes and their micelles

Feed ratio (mmol)				Molecular weights (g mol)
Samples^a	PCL	MPEG or PEG	MPEG-DEAM	IPDI	MDEA	BDO	Mn	Mw	Mn/Mw	Size (nm)	PDI	Zetapotential (mV)
a pH-sensitive polyurethanes are denoted as PEG-x-PU-y, where x is for location PEG (c, PEG attached on the end of PU chain; g, PEG grafted on PU chain; b, PEG copolymerized into PU as soft segment), y is the molar content of PEG in PUs.
PEG-c-PU-1	1.5	0.4		3.5	1	0.8	35 006	54 280	1.55	48.83 ± 1.1	0.17 ± 0.07	−12.9 ± 1.56
PEG-c-PU-2	1.5	0.7		3.5	1	0.65	37 332	50 380	1.34	45.32 ± 3.2	0.09 ± 0.03	−15.9 ± 2.78
PEG-c-PU-3	1.5	1		3.5	1	0.5	24 794	39 007	1.46	52.72 ± 2.8	0.07 ± 0.02	−13.0 ± 1.98
PEG-g-PU-1	1.5		0.4	3.5	0.6	1	24 638	35 225	1.42	54.15 ± 1.2	0.24 ± 0.06	−17.9 ± 2.65
PEG-g-PU-2	1.5		0.7	3.5	0.3	1	28 685	40 594	1.41	121.12 ± 4.1	0.27 ± 0.05	4.88 ± 1.23
PEG-g-PU-3	1.5		1	3.5	0	1	33 415	54 835	1.64	103.34 ± 3.1	0.19 ± 0.18	12.2 ± 2.32
PEG-b-PU-1	1.5	0.4		3.5	1	1	34 495	54 876	1.59	46.65 ± 0.9	0.08 ± 0.05	−16.4 ± 4.23
PEG-b-PU-2	1.5	0.7		3.5	1	0.6	26 326	40 640	1.54	59.94 ± 1.2	0.13 ± 0.09	−15.3 ± 3.65
PEG-b-PU-3	1.5	1		3.5	1	0.3	30 112	45 907	1.52	61.15 ± 2.1	0.09 ± 0.07	−13.7 ± 3.26

---

2.3 Characterization of polyurethanes

¹H NMR was recorded on an Agilent-NMR-VNMRS 400 (400 MHz) spectrometer using deuterated dimethyl sulfoxide (DMSO-d₆) and deuteroxide (D₂O) as the solvent. Infrared data were obtained on KBr slice with the Nicolet-560 spectrophotometer between 4000 and 600 cm⁻¹ in the resolution of 4 cm⁻¹. The molecular weights and molecular weight distributions of synthesized PUs were determined by a Waters-1515 gel permeation chromatograph. Tetrahydrofuran (THF) was used as the mobile phase at a flow rate of 1 mL min⁻¹ at 40 °C and the molecular weights were reported relative to polystyrene (PS) standards. The DSC analysis was performed with a differential scanning calorimeter (TA instrument 2910 thermal analyzer) to measure the heat variation at a heating rate of 10 °C min⁻¹. The measuring range of this analysis was set between −60 °C and 100 °C.

2.4 Preparation of polyurethane micelles

A dialysis method was used to prepare the polyurethane micelles. Typically, polyurethane (PEG-g-PU, PEG-b-PU or PEG-c-PU, 20 mg) was completely dissolved in 4 mL DMAc. Afterwards, the solution was added drop wise into 10 mL distilled water under vigorous stirring. Subsequently, the micelle solution was transferred to the dialysis bag (MWCO, 3.5 kDa) and dialysed against distilled water for 72 h to eliminate the organic solvent at room temperature. Finally, the micelle solution was passed through a 0.45 μm pore-sized syringe filter (Millipore, Carrigtwohill, Co. Cork, Ireland) and stored at 4 °C.

2.5 Characterization of polyurethane micelles

The critical micelle concentration (CMC) of the obtained polyurethanes in distilled water was determined by fluorescence spectroscopy using pyrene as a probe. The polyurethane concentration was varied from 1.0 × 10⁻⁵ to 0.2 mg mL⁻¹ and the final pyrene concentration was fixed to 5.2 × 10⁻⁶ M. The combined solution of pyrene and micelles was sonicated for 4 h in the dark before the fluorescence measurement. Fluorescence excitation spectra was recorded by a fluorometer (Varian Cary Eclipse Fluorescence Spectrophotometer, USA) at a wavelength range of 285 to 355 nm, with the emission wavelength at 372 nm and slits at 5 nm for both excitation and emission. Size and zeta potentials of the nanoparticles in aqueous solution were measured with a Zetasizer analyzer (Malvern Zetasizer Nano, Zen 3690+MPT2, Malvern, UK). The morphologies of PEG-g-PU, PEG-b-PU or PEG-c-PU micelles were observed by transmission electron microscopy (TEM) (Tecnai G² F20 transmission electron microscope, Royal Dutch Philips Electronics Ltd., Holland) at an accelerating voltage of 200 kV. The self-assembled polyurethane micelles were freeze-dried, and the lyophilized powders were subjected to DSC analysis with a differential scanning calorimeter (TA instrument 2910 thermal analyzer) at a heating rate of 10 °C min⁻¹. The measuring range of this analysis was set between −60 °C and 100 °C. Surface morphology and phase imaging of NPs were studied using atomic force microscope (SPM-9600, Shimadzu, Japan). Samples were prepared by deposition of particle suspension in water on freshly cleaved mica followed by air-drying. Topography and phase images of these samples were captured simultaneously using TappingMode™ Au probes (NSG11) with tip radius of 10 nm, spring constant of 5.5–22.5 N m⁻¹ and resonance frequency of 190–325 kHz. Cantilever length was 100 ± 5 μm.

2.6 Preparation of PTX-loaded micelles

A micelle extraction technique was used to load PTX into reduction-sensitive polyurethane micelles. Prior to loading, the given amount of PTX was dissolved in 5 mL of acetone to a concentration of 5 mg mL⁻¹ and placed in a vial. The acetone was allowed to evaporate from the vial under nitrogen atmosphere to form a drug film. Then 10 mL of micelle solution (1 mg mL⁻¹) was transferred into the vial containing the drug film with different amounts of PTX and ultrasonicated for 2 h to achieve maximum loading. The excess PTX was removed by a 0.45 μm filter. The drug concentration of the loaded micelles was evaluated using a high-performance liquid chromatography (HPLC) system (Waters Isocratic HPLC Pump, US) equipped with a reverse-phase C18 column (4.6 × 250 mm, 5 μm). An acetonitrile–water (60/40 v/v) mixture was used as the mobile phase and the flow rate was 1.0 mL min⁻¹. The UV adsorption of PTX outflow from the chromatographic column was recorded by a Waters 2489 UV/Visible Detector at a wavelength of 227 nm. The loading content (%) and encapsulation efficiency (%) were calculated based on the equations below:

Loading content (LC) (%) = mass of drugs in micelles/total mass of loaded micelles × 100%

Encapsulation efficiency (EE) (%) = mass of drugs in micelles/initial amount of feeding drugs × 100%.

2.7 Preparation of DOX-loaded micelles using dialysis method

Water (5 mL) was added drop-wise to a clear solution consisting of the purified, dried PU (10 mg), DOX (2 mg), and Et₃N (3 molar equivalents to DOX) in DMF (2 mL). The resulting dispersion was dialyzed over water (500 mL) for 4 days, yielding DOX-loaded micelles of PU at 1.0 mg mL⁻¹. The excess DOX was removed by a 0.45 μm filter. The drug concentration of the loaded micelles was evaluated using a high-performance liquid chromatography (HPLC) system (Waters Isocratic HPLC Pump, US) equipped with a reverse-phase C18 column (4.6 × 250 mm, 5 μm). An acetonitrile–water (32/68 v/v) mixture was used as the mobile phase and the flow rate was 1.0 mL min⁻¹. The UV adsorption of DOX outflow from the chromatographic column was recorded by a Waters 2489 UV/Visible Detector at a wavelength of 254 nm.

2.8 In vitro release of PTX

In vitro drug release from the drug-loaded micelles was performed using the dialysis method. In brief, 10 mL of drug loaded micelle solution was added to a dialysis bag (MWCO: 3.5 kDa) and immersed in 100 mL of PBS (0.01 M, pH 7.4, pH 6.8 and pH 5.5) containing 0.1 M sodium salicylate. The dialysis system was kept at 37 °C in a thermostatic incubator with a shaking speed of 110 rpm. Samples were removed and replaced with the same volume of fresh medium at the desired time interval. The concentration of PTX released from drug loaded micelles was analyzed by HPLC.

2.9 Cell culture

The human liver cell line H460 (Chinese Academy of Science Cell Bank for Type Culture Collection, China) was cultured in RPMI 1640 media supplemented with 2 mM L-glutamine, 100 U mL⁻¹ penicillin and 10% foetal bovine serum (FBS) (FBS, HyClone, Logan, UT) at 37 °C in a humidified atmosphere containing 5% CO₂ (Sanyo Incubator, MCO-18AIC, Japan).

Human umbilical vein endothelial cells (HUVECs, Chinese Academy of Science Cell Bank for Type Culture Collection, China) were maintained in Dulbecco's modified Eagle's medium (DMEM, Gibco Life, Grand Island, NY, USA) supplemented with 10% (v/v) foetal bovine serum, 2 mM L-glutamine and 1% (v/v) antibiotic mixture (10 000 U of penicillin and 10 mg of streptomycin) (Gibco). The cells were incubated in a humidified atmosphere of 5% CO₂ at 37 °C (Sanyo Incubator, MCO-18AIC, Japan).

2.10 Cellular uptake and intracellular release of payloads

Confocal laser scanning microscopy (CLSM) was employed to examine the cellular uptake and intracellular release behaviour of polyurethane micelles. The DOX-loaded micelles were incubated with H460 cells for 4 h at 37 °C. After removal of the medium, the cells were washed three times with cold PBS, fixed with 1 mL of 4% paraformaldehyde for 30 min at 4 °C, and stained with 2-(4-amidinophenyl)-6-indolecarbamidine dihydrochloride (DAPI, Roche) for 10 min. Finally, the slides were mounted with a 10% glycerol solution and observed by a Leica TCS SP8 (Leica Microscopy Systems Ltd., Germany).

2.11 Cell viability assay

To evaluate the antitumor activity of PTX-loaded polyurethane micelles and the cytocompatibility of drug-free micelles, H460 cells and HUVEC were seeded in 96-well plates at 4 × 10³ cells per well and incubated for 24 h. The culture medium was removed and replaced with 100 μL of medium containing various concentrations of micelle solutions and incubated for another 24 h. Then, 10 μL of Cell Counting Kit-8 (CCK-8) solution (Shanghai Qcbio Science & Technologies Co., Ltd.) was added to each well. After incubating the cells for 4 h, the absorbance was measured at a wavelength of 450 nm. The cell viability was normalized to that of cells cultured in the full culture media. The dose-effect curves were plotted and the median inhibitory concentration (IC₅₀) was determined using IBM SPSS Statistics software (SPPS, Inc., USA).

3. Results and discussion

3.1 Synthesis and characterization of PUs

Synthesis of PEGylated diethanolamine (MPEG-DEAM) is shown in Scheme 1S,† MPEG-OTs was prepared by MPEG and p-TsCl at the presence of NaOH without any solvent. MPEG-OTs was further reacted with diethanolamine to obtain MPEG-DEAM. In order to prepare PU micelles with positive charges with dense PEG brush-like coating on the outmost surfaces, PEG-g-PUs with different PEG content were prepared as the feed ratios listed in Table 1. In addition, PEG-b-PUs and PEG-c-PUs with the same amount of PEG were prepared as a comparison to clarify the effect of PEG position on the surface properties of PU micelles. The similar synthesis processes were conducted as listed in Scheme 2. The unimodal GPC curves of purified PEG-g-PUs, PEG-b-PUs and PEG-c-PUs (ESI: Fig. S1†) confirmed successful polymerization. The average molecular weight and polydispersity of PEG-g-PUs, PEG-b-PUs and PEG-c-PUs are listed in Table 1. The representative ¹H NMR spectra of polyurethanes and the assignment of the peaks are presented in Fig. 1. The ¹H NMR spectra of PEG-g-PU-3, PEG-b-PU-3 and PEG-c-PU-3 are nearly identical arising from their identical composition. The resonance peak centred at 3.49 ppm (peak f) is assigned to the protons on the PEG units while the peaks at 3.97 ppm (peak a), 2.26 ppm (peak e), 1.52 ppm (peak b and d) and 1.29 ppm (peak c) are attributed to the protons of the PCL units. The resonances at 0.75–0.95 ppm (peak IPDI) are ascribed to the methylene and methyl protons of the IPDI units in the products. The peaks at 3.95 ppm (peak h) and 2.25 ppm (peak g) are attributed to the methylene and the methylene protons of the DEAM units.

Scheme 1 Schematic illustration of the pH sensitive polyurethane with different PEG location (A), the change of size (B) and the zeta potentials (C) of PEG-g-PU, PEG-b-PU, PEG-c-PU micelles with the increase of the PEG content.

Scheme 2 Synthesis of pH-sensitive biodegradable multi-block polyurethanes.

Fig. 1 ¹H NMR spectra of PEG-c-PU-3, PEG-g-PU-3 and PEG-b-PU-3 in DMSO-d₆.

Microphase separation is characteristic for multiblocked polyurethanes due to the incompatibility between the flexible soft segments and the rigid hard segments. Hydrogen bonding between the amide groups (as the donor) in hard segments and the urethane carbonyl (as acceptor) in hard segments (denoted as H–H bonding) can increase the microphase separation, while that between the amide groups and the ether oxygen (in PEG) or the carbonyl (in PCL) in soft segments (denoted as H–S bonding) decrease the microphase separation.²⁹ Thus FTIR was used to investigate the hydrogen bonding of the prepared PU for deep understanding of the phase separation induced by PEG location. The FTIR spectra of PEG-g-PU-3, PEG-b-PU-3 and PEG-c-PU-3 bulk materials are shown in Fig. 2. The disappearance of absorption at ∼2275–2250 cm⁻¹ indicates complete consumption of the isocyanate. All the three kinds of polyurethanes show absorption at ∼1740 cm⁻¹ that assigned to the stretching of carbonyl of PCL.²⁷ In addition to the absorption from carbonyl of PCL, the peak at ∼1680 cm⁻¹ caused by H–H bonding predominates from PEG-g-PU-3, which indicates the highly microphase separation for PEG-g-PU-3.²⁷ However, PEG-c-PU-3 and PEG-b-PU-3 include predominate peak at 1720 cm⁻¹ from free C=O, indicating low microphase separation for these two kinds of polyurethanes. The result seems to indicate that PEG-g-PU micelles will also show highly microphase separation due to its special structure.

Fig. 2 FTIR spectra of PEG-c-PU-3, PEG-g-PU-3 and PEG-b-PU-3 in the urethane and ester C=O stretch region.

3.2 Characterization of PU micelles

The amphiphilic multi-blocked polyurethanes self-assembled in aqueous solution into micelles having a hydrophobic PCL core and a hydrophilic PEG shell. The core–shell structure of the polyurethane micelles was confirmed by the ¹H NMR resonance difference of the hydrophilic PEG and hydrophobic PCL signals when DMSO-d₆ and D₂O were used as solvents (ESI: Fig. S2†). The critical micellar concentration (CMC) was determined by fluorescence spectroscopy with a pyrene probe.³⁰ The ratio of the intensity of the peak at 337 nm to that at 333.5 nm is plotted against the log of polymer concentration and the concentration corresponding to the intersection of the two tangential lines is the CMC value (ESI: Fig. S3†). The CMCs of PEG-c-PU-3, PEG-b-PU-3 and PEG-g-PU-3 were determined to be 8.61 × 10⁻⁴ mg mL⁻¹, 8.85 × 10⁻⁴ mg mL⁻¹ and 7.74 × 10⁻⁴ mg mL⁻¹, respectively.

The size change of PEG-g-PU, PEG-b-PU, PEG-c-PU micelles with the increase of the PEG content are illustrated in Scheme 1 and Table 1. The hydrodynamic diameter of PEG-b-PU and PEG-c-PU micelles remain nearly constant with the increase of the PEG content. However, the hydrodynamic diameter of PEG-g-PU micelles increases with the increase of the PEG content. The size of PEG-g-PU series micelles is much larger than the other two kinds of PU micelles with the same amount of PEG, which might suggest that there were much more PEG chains stretched out from the surface of PEG-g-PU micelles than that from the other two micelles. To testify this conclusion, the self-assembled micelles of PEG-b-PU-3, PEG-c-PU-3 and PEG-g-PU-3 in aqueous were visually observed by TEM microphotographs. An obscure and uneven shell was observed for PEG-g-PU-3 micelles in Fig. 3D. However, the shells were relatively smoother than PEG-b-PU-3 and PEG-c-PU-3 micelles. This should be an evidence that much more PEG chains stretching out from the surface of PEG-g-PU micelles than that from the other two micelles. Meanwhile, TEM images indicate that PEG-b-PU-3, PEG-c-PU-3 and PEG-g-PU-3 micelles have an average diameter of 40–50 nm (Fig. 3A, D and G), which is smaller than the size determined by DLS. The difference in micelle sizes between DLS and TEM can be attributed to the dehydrated state of the micelles.¹² Note that the hydrodynamic size of PEG-g-PU-3 micelles was significantly larger than the size observed by TEM. This is because the well-solvated PEG coronas of PEG-g-PU-3 micelles are not directly visible by TEM, but their presence and spatial extent is indicated by DLS.^31,32 This is a further evidence that much more PEG chains stretching out from the surface of PEG-g-PU micelles than that from the other two micelles, which may result in the longer circulation time of PEG-g-PU micelles than PEG-b-PU and PEG-c-PU micelles.

Fig. 3 TEM micrograph of PEG-c-PU-3, PEG-g-PU-3 and PEG-b-PU-3 micelles under pH = 7.4 and pH = 5.5 (A: pH = 7.4, B: pH = 5.5, PEG-c-PU-3; D: pH = 7.4, E: pH = 5.5, PEG-g-PU-3; G: pH = 7.4, H: pH = 5.5, PEG-b-PU-3). Size distribution of pH sensitive polyurethanes micelles determined by DLS (C: PEG-c-PU-3, F: PEG-g-PU-3, I: PEG-b-PU-3).

The PEG-b-PU and PEG-c-PU micelles show negative zeta potentials as shown in Table 1 and Scheme 1C. In addition, the zeta potentials of PEG-b-PU and PEG-c-PU micelles remain nearly constant with the increase of the PEG content. However, the zeta potential of PEG-g-PU micelles increases with the increase of the PEG content. For example, PEG-g-PU-1 shows a negative zeta potential with −17.9 mV, PEG-g-PU-2 and PEG-g-PU-3 with +4.88 mV and +12.2 mV respectively. It has been demonstrated that the PEGlyation surface without other ionic groups usually presented negative zeta potentials as a result of the polarization of water molecules under the effect of PEG.^12,33 Meanwhile, the tertiary amine groups of DMEA in the hard segment will render the particles with positive charges when the DEMA located at the surface of the micelles.³⁴ Based on this, we can conclude that none or few DEMA was migrated to the surface of PEG-b-PU and PEG-c-PU micelles due to the relatively lower mobility of PEG chains. However, the positive charge of PEG-g-PU-2 and PEG-g-PU-3 micelles suggested that many tertiary amine groups (DEMA) were transported to the surface of PEG-g-PU micelles by MPEG chains during self-assembling. Therefore, it can conclude that PU micelles with positive charges covered with a dense PEG chains could be achieved by PEG-g-PU-3.

3.3 Surface morphology and phase analysis of PU micelles

It is reported²⁰ that the biological properties of the PEGylated polymer particles depended on the chain coverage-density and conformation of PEG corona. The high antifouling effect (long circulation time) and good mucus traffic properties of polymer nanoparticles is usually achieved at high coverage-density and in the polymeric brushes regime. In order to investigate surface properties of PEGylated nanoparticles, tapping mode atomic force microscopy (TM-AFM) was used to probe the PU NPs. Fig. 4 shows TM-AFM topography and their corresponding phase images of PEG-b-PU-3, PEG-g-PU-3, and PEG-c-PU-3 respectively. TM-AFM surface analysis revealed that all NPs were in a spherical shape without obvious aggregating tendency. The particles size was in the range of 70–100 nm (Fig. 4) which is larger than TEM results. Phase imaging is based on the use of changes in the phase angle of cantilever probe. This image shows more contrast than the topographic one as well as more sensitivity to material surface properties such as stiffness, viscoelasticity, and chemical composition. Hildgen et al.³⁵ visually observed the PEG domains and their coverage density at the surface of PEGylated polymer micelles using phase image analysis of AFM. Due to the mechanical and hydrophilicity differences between PCL segment, hard segment and PEG segment in multiblocked polyurethanes, a phase contrast would be observed in phase images of PU nanoparticles by AFM. PEG molecule is expected to be softer than PCL since PEG molecules of lower M_w 1000 (used in our study) have smaller Young's modulus than PCL and hard segment of polyurethanes. Thus, it was expected that PEG molecule will appear as darker regions in the phase images due to negative phase shift. Xie et al.³⁶ has already shown that softer PEG segments appeared as darker regions embedded in lighter polystyrene domain for poly(styrene-b-ethylene oxide) polymer films. As is shown in Fig. 4A, D and G, PEG-c-PU-3 and PEG-b-PU-3 particles had nearly homogenous surface without any clear phase separation (Fig. 4A and G). Therefore, no contrast was observed in phase images of PEG-c-PU and PEG-b-PU particles. This indicate that there is no PEG microdomain formed on the surface of PEG-c-PU and PEG-b-PU particles due to the high phase miscibility of PCL and PEG segment. On the other hand, PEG-g-PU-3 NPs showed the presence of an observable phase contrast at the surface of NPs (Fig. 4D). This might be due to mechanical differences between soft segment and hard segment (on which pendant PEG chains attached) that result in microphase separation of soft segment (PCL) and hard segment (including PEG). In the case of PEG-g-PU, highly intense dark regions located at centre of the micelles and which were coated by brighter hydrophobic PCL chains that represent the core. The immiscibility of soft segment (PCL) and hard segment (including PEG) should result in separation of both components during NPs formation. Thus, grafted copolymer NPs will be predominantly consisting of hydrophobic PCL cores surrounded by hydrophilic PEG chains stretching out the surface. In the case of the PEG-c-PU-3 and PEG-b-PU-3 particles, no obvious dark regions are found at the surface of NPs. This might be due to the peculiar architecture of the polymers themselves where PEG segment was covalently linked with PCL segment which enhanced miscibility of PEG and PCL blocks compared to the grafted ones and restrained the aggregation of PEG chains towards the aqueous phase during NPs formation.

Fig. 4 Tapping mode AFM images of NPs, left panel shows topography (T), middle panel shows corresponding phase images (P), and right panel shows corresponding space pattern, all images are acquired in air (A–C: PEG-c-PU-3; D–F: PEG-g-PU-3; G–I: PEG-b-PU-3).

DSC was used to detect the effect of molecular structure of the used PU on the thermal properties of NPs,³⁷ which provide information on transition temperatures such as crystallization melting endotherms (T_m) of PEG and PCL segment. T_m of the PU bulk materials and PU micelles are shown in Fig. 5. For the bulk polyurethanes, a T_m at ∼43 °C corresponding to the melting of crystalline PCL was observed for all kinds of polyurethanes. The T_m of the PCL segment was quite lower, which may be a consequence of mixing with the PEG segments.³⁸ Notably, a melting endotherm at ∼31 °C corresponding to the melting of crystalline phase PEG was observed only for PEG-g-PU,³⁹ indicating the presence of orderly arranged the PEG chains in PEG-g-PU due to the greater phase separation between the soft and hard segments of PEG-g-PU as has proven by FTIR results. Furthermore, the three kinds of PU micelles were lyophilized and subjected to DSC study to investigate the package of PEG and PCL chains in the self-assembled micelles. The results are shown in Fig. 5B. The absence of melting endotherm of crystalline phase PEG for both PEG-b-PU and PEG-c-PU micelles suggested the random packaged or loop like PEG chains located at the surface of these two kinds of micelles. In addition, the melting temperature of crystalline phase PCL decreased to ∼40 °C and ∼37 °C for PEG-b-PU and PEG-c-PU micelles respectively, indicating the less ordered PCL chain package in the core of the micelles than that in the bulk materials due to the stretch of PEG chains during self-assembling of the amphiphilic polyurethanes. Interestingly, an obvious melting endotherm of crystalline phase PEG at ∼32 °C was observed for PEG-g-PU micelles, indicating the ordered chain package of the PEG chains in lyophilized PEG-g-PU micelles. In other words, the PEG was in a brush like conformation in PEG-g-PU micelles as we expected. In addition, the melting endotherm of crystalline phase PCL was disappeared for PEG-g-PU micelles, indicating the loosely package of the PCL chains in the core of PEG-g-PU micelles due to the higher mobility of the PEG chains grafted on the hard segment than that blocked at soft segment or encaped at the end of polyurethanes. This provides further evidence that the grafted PEG (PEG-g-PU) have relatively higher mobility than that located at soft segment (PEG-b-PU) or attached on the endcap (PEG-c-PU).

Fig. 5 DSC curves of bulk PEG-c-PU-3, PEG-g-PU-3, PEG-b-PU-3 (A) and PEG-c-PU-3 micelles, PEG-g-PU-3 micelles, PEG-b-PU-3 micelles (B). Temperature ranges from 10 °C to 60 °C.

3.4 Stimuli-responsiveness of polyurethane micelles

We incorporated pH active tertiary amine into the polyurethane main chain by using EDMA or PEGylated EDMA as chain extenders. The tertiary amine will be protonated at low pH value which result in the swell of the polyurethane micelles and the release of the encapsulated cargos.⁴ To determine the pH responsiveness of the PU micelles, the size change of micelles at acidic environment (pH = 5.5) was carried out. As is shown in Fig. 3B, E and H, all the PU micelles are swollen in the acidic environment, suggesting the protonation and the repulsion of the positively charged tertiary amine groups. Notably, the morphology of all kinds PU micelles incubated at pH 5.5 still kept intact in a spherical shape without disintegration, which may encapsulate the drug in the core and protect the drug from degradation by enzyme in the escape process.¹

3.5 Drug loading and release of the polyurethane micelles

PTX was used as a model hydrophobic drug to investigate the loading capacity of the PEG-b-PU-3, PEG-c-PU-3, PEG-g-PU-3 micelles. A micelle extraction technique was used to load the PTX into polyurethane micelles. The excess PTX was removed by filtration through a 0.45 μm filter. The loading level of PTX for PTX-loaded micelles was determined using HPLC. The drug loading content (LC) and encapsulation efficiency for PTX loaded micelles are in the range of 1.23–5.77% and 2.5–10.24%, respectively. The loading content (LC) and encapsulation efficiency for DOX loaded micelles are in the range of 1.22–1.61% and 6.20–8.20%, respectively. It should be noted that the size and size distribution of the PTX and DOX loaded micelles are in the same range (ESI: Fig. S4†).

The drug release behaviour of the PTX-loaded PEG-b-PU-3, PEG-c-PU-3 and PEG-g-PU-3 micelles was investigated at neutral (pH 7.4) and acidic (pH 5.5 and pH 6.8) buffer solutions. Fig. 6 shows the cumulative drug release profiles as a function of time. As shown in Fig. 6, the PTX release of all the three kinds of polyurethane micelles were correlated markedly with the pH value of the solution.

Fig. 6 Time dependent cumulative release of PTX from pH sensitive polyurethane micelles in PBS buffer solutions (pH 7.4, 10 mM), (pH 6.8, 10 mM) and (pH 5.5, 10 mM).

In addition, the effect of PEG position in PU chains on drug encapsulation and release could be ignored due to no obvious difference was observed for the drug release from PEG-b-PU-3, PEG-c-PU-3 and PEG-g-PU-3 micelles. In neutral environment, PTX release was suppressed significantly due to the good stability and compacted structure of the polyurethane micelles, with less than 20% released within 48 h. However, the PTX release was enhanced dramatically at acidic environment, i.e. pH 6.8 and pH 5.5, due to the protonation of the DMEA and the swollen of the PU micelles. In detail, at pH 5.5, nearly 70% of PTX is released from the three kinds of PU micelles within 24 h. In contrast, the release rate is ∼50% at pH 6.8 within 24 h.

3.6 Internalization and intracellular release of the polyurethane micelle payload

The cellular uptake of the micelles and the intracellular location of the encapsulated payloads were monitored by CLSM in H460 cells. Since PTX molecules are not fluorescent, doxorubicin (DOX, red fluorescence) was encapsulated in the hydrophobic micellar cores by dialysis to label the nanocarriers and to track the internalization and intracellular localization of DOX in H460 cells.⁴⁰ The nuclei of H460 cells were stained by DAPI, which emits blue fluorescence to distinguish from the red fluorescence of DOX. Fig. 7 shows CLSM images of H460 cells incubated with DOX-loaded polyurethane micelles for 4 h. As shown in Fig. 7D, DOX is diffusible molecule that can directly pass through the cell membranes and rapidly enter the nucleus.¹ Meanwhile, the images of DOX fluorescence suggest that DOX-loaded polyurethane micelles were internalized and DOX was released to reach cell nuclei.^14,41 These results may demonstrate that pH-sensitive PU can effectively escape from endo-lysosome to cytoplasm with loaded drugs.¹

Fig. 7 CLSM images of H460 cells incubated with DOX-loaded PEG-c-PU-3, PEG-g-PU-3 and PEG-b-PU-3 micelles for 4 h. Scale bar = 20 μm (A: PEG-c-PU-3, B: PEG-g-PU-3, C: PEG-b-PU-3, D: free DOX).

3.7 In vitro cytotoxicities of polyurethane micelles and PTX-loaded polyurethane micelles

The in vitro cytotoxicity of drug-free and drug-loaded polyurethane micelles was evaluated by the CCK8 assay. Fig. 8 shows the effect of PEG-b-PU-3, PEG-c-PU-3 and PEG-g-PU-3 concentration on the viability of HUVEC and H460 cells. The results demonstrate that empty PEG-b-PU-3, PEG-c-PU-3 and PEG-g-PU-3 micelles show very low cytotoxicity (greater than 90% cell viability) in two different cell lines even at micelle concentrations up to 0.5 mg L⁻¹, suggesting the non-toxic nature of polyurethane micelles to HUVEC and H460 cells. Further, the in vitro cytotoxicity of PTX-loaded PEG-g-PU-3 micelles was compared to that of PTX-loaded PEG-b-PU-3 and PEG-c-PU-3 and micelles. Fig. 9A shows the cytotoxicity results given as a function of PTX concentration from 0.01 to 10 000 ng mL⁻¹. As the test concentration was increased to 100 ng mL⁻¹, PEG-g-PU-3 encapsulated PTX exhibited lower cell viability (30%) than PEG-b-PU-3 and PEG-c-PU-3 encapsulated PTX (40% for PEG-c-PU-3 and 50% for PEG-b-PU-3). The IC₅₀ (i.e., inhibitory concentration to produce 50% cell death) values of PEG-g-PU-3, PEG-b-PU-3 and PEG-c-PU-3 encapsulated PTX were determined to be ∼700, ∼2300 and ∼1700 ng mL⁻¹ respectively for H460 cells (Fig. 9B) (p < 0.05). PEG-g-PU-3 micelles provide more efficient intracellular delivery of PTX as compared to PEG-b-PU-3 and PEG-c-PU-3 micelles, which might be attributed to the grafted PEG with high mobility transfer the diethanolamine (DEAM) to the surface of nanoparticles during self-assembly process and the DEAM render the particles with positive charges that potentially enhances cellular uptake and endosomal escape.^1,42

Fig. 8 Viability of H460 cells (A) and HUVECs (B) after 48 h of incubation with various concentrations of empty pH sensitive polyurethane micelles determined by the CCK8 assay.

Fig. 9 Cytotoxicity (A) and IC₅₀ values (B) of PTX-loaded polyurethane micelles against H460 cells after incubation for 24 h.

4. Conclusions

We have developed three kinds of pH responsive multi-blocked polyurethanes with PEG incorporated as soft segment (PEG-b-PU), PEG positioned at the endcap (PEG-c-PU), and PEG grafted on the hard segment (PEG-g-PU), as antitumor drug delivery systems. The PEG content has little effect on particles size and zeta potentials for PEG-b-PU and PEG-c-PU due to relatively low mobility of PEG. However, the particles size and zeta potential increased with the increasing of the PEG content for PEG-g-PU micelles. The results suggested that the PEG with relatively higher mobility transported DEAM to the surface of micelles during self-assembly in aqueous for PEG-g-PUs. Furthermore, TEM, AFM, and DSC together showed that a dense PEG domain was formed on the surface of PEG-g-PU micelles while no obvious PEG domain was observed for the other two kinds of micelles. In response to acidic environments, the protonated DEAM will repel each other, which caused swell of the micelles and controlled release of the anticancer drugs inside the cancer cells. Notably, the intact sphere morphology of PU micelles at acidic solutions may protect the drug from degradation by enzyme during lysosomal escape. Internalization into H460 cancer cells and intracellular release of anticancer drugs was shown by CLSM. The in vitro cytotoxicity of drug-free and drug-loaded polyurethane micelles was evaluated by the CCK8 assay. The enhanced cytotoxicity of PTX-loaded PEG-g-PU-3 micelles against H460 cancer cells reveals that the PEG-g-PU-3 micelles are more potent for intracellular delivery of PTX as compared to PEG-b-PU-3 and PEG-c-PU-3 micelles. Meanwhile, the three empty polyurethane micelles exhibited no cytotoxicity against cancer cells or healthy cells at concentrations as high as 0.5 mg mL⁻¹. These results demonstrate that multi-blocked polyurethanes are a promising drug delivery platform and that the surface properties of the PU micelles drugs can easily be tuned by the location and content of PEG segment in the polyurethanes.

Acknowledgements

This work was financially supported by the Applied Basic Research Programs Foundation of Sichuan Province (No. 2015JY0126) and by the Fundamental Research Funds for the Central Universities (Southwest University for Nationalities No. 2016NGJPY04).

Notes and references

1. N. Song, L. Zhou, J. Li, Z. Pan, X. He, H. Tan, X. Wan, J. Li, R. Ran and Q. Fu, Nanoscale, 2016, 8, 7711–7722.
2. N. Chan, S. Y. An and J. K. Oh, Polym. Chem., 2014, 5, 1637–1649.
3. A. E. Dennis and E. Discher, Science, 2002, 297, 967–972.
4. S. Yu, C. He, J. Ding, Y. Cheng, W. Song, X. Zhuang and X. Chen, Soft Matter, 2013, 9, 2637–2645.
5. S. Zhang, Y. Wu, B. He, K. Luo and Z. Gu, Sci. China: Chem., 2014, 57, 461–475.
6. J. W. H. Maedaa, T. Sawaa, Y. Matsumurab and K. Horic, J. Controlled Release, 2000, 65, 271–284.
7. W. Shao, K. Miao, H. Liu, C. Ye, J. Du and Y. Zhao, Polym. Chem., 2013, 4, 3398–3410.
8. S. Zhang and Y. Zhao, Bioconjugate Chem., 2011, 22, 523–528.
9. G. H. Hwang, K. H. Min, H. J. Lee, H. Y. Nam, G. H. Choi, B. J. Kim, S. Y. Jeong and S. C. Lee, Chem. Commun., 2014, 50, 4351–4353.
10. J. Yang, Y. Duan, X. Zhang, Y. Wang and A. Yu, J. Mater. Chem. B, 2016, 4, 3868–3873.
11. H. Kim, S.-M. Jeong and J.-W. Park, J. Am. Chem. Soc., 2011, 133, 5206–5209.
12. Y. Yao, H. Xu, C. Liu, Y. Guan, D. Xu, J. Zhang, Y. Su, L. Zhao and J. Luo, RSC Adv., 2016, 6, 9082–9089.
13. M. Khuphe, A. Kazlauciunas, M. Huscroft and P. D. Thornton, Chem. Commun., 2015, 51, 1520–1523.
14. Z. Ge and S. Liu, Chem. Soc. Rev., 2013, 42, 7289–7325.
15. M. Karimi, A. Ghasemi, P. S. Zangabad, R. Rahighi, S. M. M. Basri, H. Mirshekari, M. Amiri, Z. S. Pishabad, A. Aslani, M. Bozorgomid, D. Ghosh, A. Beyzavi, A. Vaseghi, A. R. Aref, L. Haghani, S. Bahrami and M. R. Hamblin, Chem. Soc. Rev., 2016, 45, 1457–1501.
16. J. Lai, Z. Xu, R. Tang, W. Ji, R. Wang, J. Wang and C. Wang, Polymer, 2014, 55, 2761–2771.
17. T. Kavitha, S. I. Abdi and S. Y. Park, Phys. Chem. Chem. Phys., 2013, 15, 5176–5185.
18. R. K. Kainthan, M. Gnanamani, M. Ganguli, T. Ghosh, D. E. Brooks, S. Maiti and J. N. Kizhakkedathu, Biomaterials, 2006, 27, 5377–5390.
19. Z. Amoozgar and Y. Yeo, Wiley Interdiscip. Rev.: Nanomed. Nanobiotechnol., 2012, 4, 219–233.
20. J. M. Rabanel, P. Hildgen and X. Banquy, J. Controlled Release, 2014, 185, 71–87.
21. B. C. Tang, M. Dawson, S. K. Lai, Y. Y. Wang, J. S. Suk, M. Yang, P. Zeitlin, M. P. Boyle, J. Fu and J. Hanes, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 19268–19273.
22. T. Stylianopoulos, B. Diop-Frimpong, L. L. Munn and R. K. Jain, Biophys. J., 2010, 99, 3119–3128.
23. A. Vonarbourg, C. Passirani, P. Saulnier and J. P. Benoit, Biomaterials, 2006, 27, 4356–4373.
24. Z. Wang, L. Yu, M. Ding, H. Tan, J. Li and Q. Fu, Polym. Chem., 2011, 2, 601–607.
25. M. Ding, Z. Qian, J. Wang, J. Li, H. Tan, Q. Gu and Q. Fu, Polym. Chem., 2011, 2, 885–891.
26. H. S. Lee, N. Tomczyk, J. Kandel, R. J. Composto and D. M. Eckmann, J. Mater. Chem. B, 2013, 1, 6382–6391.
27. Z. Pan, L. Yu, N. Song, L. Zhou, J. Li, M. Ding, H. Tan and Q. Fu, Polym. Chem., 2014, 5, 2901–2910.
28. X. He, M. Ding, J. Li, H. Tan, Q. Fu and L. Li, RSC Adv., 2014, 4, 24736–24746.
29. Y. He, D. Xie and X. Zhang, J. Mater. Sci., 2014, 49, 7339–7352.
30. S. Zhang and Y. Zhao, J. Am. Chem. Soc., 2010, 132, 10642–10644.
31. Z. Li, E. Kesselman, Y. Talmon, M. A. Hillmyer and T. P. Lodge, Science, 2004, 306, 98–101.
32. T. P. Lodge, A. Rasdal, Z. Li and M. A. Hillmyer, J. Am. Chem. Soc., 2005, 127, 17608–17609.
33. Y. Hu, X. Jiang, Y. Ding, L. Zhang, C. Yang, J. Zhang, J. Chen and Y. Yang, Biomaterials, 2003, 24, 2395–2404.
34. J. Hu, G. Zhang, Z. Ge and S. Liu, Prog. Polym. Sci., 2014, 39, 1096–1143.
35. S. Essa, J. M. Rabanel and P. Hildgen, Eur. J. Pharm. Biopharm., 2010, 75, 96–106.
36. H. Wang, A. B. Djurišić, W. K. Chan and M. H. Xie, Appl. Surf. Sci., 2005, 252, 1092–1100.
37. G. Wu, Y. Fan, X. He and Y. Yan, RSC Adv., 2015, 5, 80893–80900.
38. H. M. Jeonga, B. K. Kimb and Y. J. Choib, Polymer, 2000, 41, 1849–1855.
39. F. Martini, S. Borsacchi, M. Geppi, F. Pilati and M. Toselli, Polymer, 2011, 52, 4536–4544.
40. P. Zhao, J. Zhang, Y. Zhu, X. Yang, X. Jiang, Y. Yuan, C. Liu and C. Li, J. Mater. Chem. B, 2014, 2, 8372–8377.
41. P. S. Pramod, R. Shah and M. Jayakannan, Nanoscale, 2015, 7, 6636–6652.
42. J. Lu, H. Jia, L. Guo, G. Zhang, Y. Cao, H. Yan and K. Liu, Eur. Polym. J., 2015, 66, 376–385.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra20613a

‡ Yongchao Yao and Deqiu Xu contributed equally to the writing of this article.

---