Fabrication of redox-responsive doxorubicin and paclitaxel prodrug nanoparticles with microfluidics for selective cancer therapy

Xiaodong Ma ab, Ezgi Özliseli b, Yuezhou Zhang *b, Guoqing Pan c, Dongqing Wang *a and Hongbo Zhang *abd
aDepartment of Radiology affiliated Hospital of Jiangsu University Jiangsu University, 212001 Zhenjiang, P.R. China
bPharmaceutical Sciences Laboratory, Faculty of Science and Engineering, Åbo Akademi University FI-00520, Finland. E-mail: yuezhou.zhang@abo.fi; hongbo.zhang@abo.fi
cJiangsu University, Institute for Advanced Materials, School of Material Science and Engineering, 301Xuefu Rd Zhenjiang, Jiangsu, CN212013, China
dTurku Centre for Biotechnology, University of Turku and Åbo Akademi University, Turku, Finland

Received 21st October 2018 , Accepted 28th November 2018

First published on 3rd December 2018


Cancer is an exceptionally confounding disease that demands the development of powerful drug/drugs, without inducing heavy adverse side effects. Thus, different approaches have been applied to improve the targeted delivery of cancer drugs: for example by using nanocarriers. However, nanocarriers are foreign materials, which need further validation for their biocompatibility and biodegradability. In this study, we have chemically conjugated the hydrophilic anticancer drug doxorubicin (DOX) with the hydrophobic drug paclitaxel (PTX) through a redox-sensitive disulfide bond, abbreviated to DOX-S-S-PTX. Subsequently, due to its amphiphilic characterization, the prodrug can self-assemble into nanoparticles under microfluidic nanoprecipitation. These novel prodrug nanoparticles have a super-high drug loading degree of 89%, which is impossible to achieve by any nanocarrier systems, and can be tailored to 180 nm to deliver themselves to the target, and release DOX and PTX under redox conditions, which are often found in cancer cells. By evaluating cell viability in MDA-MB-231, MDA-MB-231/ADR and MEF cell lines, we observed that the prodrug nanoparticles effectively killed the cancer cells, and selectively conquered the MDA-MB-231/ADR. Meanwhile, MEF cells were spared due to their lack of a redox condition. The cell interaction results show that the reduced intermediate of the prodrug can also bind to parent drug biological targets. The hemolysis results show that the nanoparticles are biocompatible in blood. Computer modelling suggested that the prodrug is unlikely to bind to biological targets that parent drugs still strongly interact with. Finally, we confirm that the prodrug nanoparticles have no therapeutic effect in blood or healthy cells, but can selectively eliminate the cancer cells that meet the redox conditions to cleave the disulfide bond and release the drugs DOX and PTX.


1. Introduction

Cancer is globally the second leading cause of mortality,1,2 with nearly 1 in 6 deaths being ascribed to cancer. It can occur at any stage of life without regard to gender or race, and even occurs in animals.3 Hence, the battle against cancer is always on. Among versatile treatments, chemotherapy might be the most important option for most cancers because of its high efficiency. However, unlike surgery or radiation which target specific areas, conventional chemotherapy works throughout the human body. Such undiscriminating administration processes will inevitably attack healthy cells unless the drug itself is target-selective.4 Breast cancer is the second leading cause of cancer-related mortality among women worldwide.5

Prodrugs are compounds with little or no pharmacological activity which, after administration, are metabolized to the active daughter drug through an enzymatic/chemical process. Prodrugs have been discovered by chance or have been designed on purpose. Such efforts are trying confront drug development hurdles that limit formulation options or result in unwanted performance, or side effects.6 It is also an approach which can be used to address the bioavailability of a drug with low aqueous solubility, in which the native hydrophobic drug is masked into a hydrophilic form that can be converted by endogenous enzymes7,8 to the native drug, and has been utilized to “redeem” water-insoluble drug candidates or to improve the availability of existing drugs. Over the past decade, more than 30 prodrugs have been marketed.6

“Nanotech” is a technology for understanding or manipulating matter at the scale of about 1 to 100 nanometers, where unique properties offer novel applications.9 Drug delivery and release through nanotechnology have emerged as a promising approach for a number of drugs or drug candidates to give satisfactory outcomes.10 Drug nanoparticles can be prepared through a variety of protocols, exemplified by emulsion-solvent evaporation,11 double emulsion and evaporation,12 emulsion–diffusion,13 salting out14 and the solvent displacement method.15,16 Solvent displacement involves the diffusion of the organic solvent in an aqueous medium in the presence of a surfactant. The drug and the surfactant are dissolved in a water-miscible mixture with an aqueous solution under magnetic stirring. As a result, nanoparticles are formed instantaneously by rapid solvent diffusion. It was observed that the mixing of the organic phase into the aqueous phase heavily affects the size and polydispersity of particles. For the bulk method, the organic solution is often poured into an aqueous medium or added stepwise without precise control over the organic phase. Therefore, it gives rise to physicochemically unsatisfactory particles. However, microfluidics-based technology is more controllable and has been widely used for drug loading and delivery-related nanoassemblies.17,18 The application of this approach to the fabrication of drug nanocrystals has been shown not only in a proof-of-concept,19 but holds great potential for scaling-up, especially for the nanoprecipitation of the poorly water-soluble anticancer drugs Paclitaxel (PTX) and Sulforaphane (SFN).20 However, due to the instability of these drug crystals, the obtained drug nanocrystals were further encapsulated with a polymer to give core/shell structures and give 42.6% (PTX) and 45.2% (SFN) drug loading by using the superfast sequential nanoprecipitation method.20

The combination of nanoprecipitation with a prodrug seems to be an effective approach for tackling unmet therapeutic need. For instance, “squalenoylation” of the antiretroviral compounds dideoxycytidine and didanosine that spontaneously self-organize in water as stable nanoassemblies has already been applied to intravenous administration.21 Gaudin et al. also conjugated adenosine to the lipid squalene and subsequently formatted a squalenoyl adenosine prodrug into nanoassemblies offering prolonged circulation of the nucleoside, providing neuroprotection in mouse stroke and rat spinal cord injury models in both in vitro and in vivo studies.22 In this prodrug conjugation, squalenoyl functioned as a lipid nanocarrier to balance the hydrophilicity of adenosine but had no therapeutic contributions, resulting in 37% drug loading. The drug loading, in theory, could be as high as 100% if the two therapeutic drugs could be covalently conjugated together without using any linker between them. For example, the hydrophilic anticancer drug irinotecan (Ir) and the hydrophobic anticancer drug chlorambucil (Cb) were joined together via an ester bond which formed between the hydroxyl group of Ir and the carboxylic acid group of Cb.23 It is clear that the selection of drugs for the conjugate will never be arbitrary.

To fulfill the advantages of the prodrug and to facilitate subsequent nanoprecipitation, several principles should be taken into account: (1) the feasibility of covalent conjugation, which means the components need to bear a suitable functional group to enable the chemical reaction to occur; (2) synergism of the candidate with a specific biological targets is always pursued, or at least it should have no side effects, while antagonism must be avoided; (3) there must be a balance of hydrophobicity and hydrophilicity, as a bias on either side will impair the formation of the desired nanoparticle; (4) the necessary fragility of the prodrug in a biological microenvironment, requiring the input prodrug to be fragmented; (5) the possibility of monitoring the pharmacokinetics of the processed prodrug particles. Following the above-mentioned know-how, we chose DOX and PTX as monomers to synthesize a drug conjugate for the following reasons: (1) DOX and PTX are synergistic in breast cancer cells;24 (2) both the amine group of DOX and the hydroxyl group of PTX can react with carboxylic acid, indicating that a linker with a bi-carboxylic acid functional group is needed to join them together; (3) DOX is hydrophilic while PTX is hydrophobic; so the coupling of the two drugs results in an amphiphilic molecule; (4) 3,3′-dithiodipropionic acid contains a disulfide bond which is breakable by a reducing reagent. The reduced glutathione (L-γ-glutamyl-L-cysteinyl-glycine; GSH) level is 100 times higher in cancer cells than in normal ones;25 therefore, it is a biomarker of diseased cells and prodrug destructor. (5) DOX is known to be fluorescent; hence it can be monitored by using a spectroscopy instrument. As illustrated in Fig. 1, we have conjugated DOX with PTX with a redox-sensitive linker and subsequently precipitated the prodrug to nanoparticles (NPs) with microfluidics. The effect and selectivity of those prodrug nanoparticles on cancer cells are evaluated with the breast cancer cell lines MDA-MB-231, MDA-MB-231/ADR and mouse embryonic fibroblasts (MEF). As a proof-of-concept, we hypothesize that the prepared NPs will be selectively internalized and be accumulated into cancer tissues, attributed to EPR effects.26


image file: c8bm01333k-f1.tif
Fig. 1 Illustration of DOX-S-S-PTX nanoparticle for breast cancer cell.

2. Experimental

2.1. Materials

Doxorubicin (DOX) and paclitaxel (PTX) were purchased from Arisun ChemPharm Co., Ltd (China). 3,3′-Dithiodipropionic acid (DTDP), 4-dimethylaminopyridine (DMAP) and N-(3-dimethylaminopropyl)-N-ethylcarbodiimide hydrochloride (EDC·HCl) were purchased from Alfa Aesar (Finland). DL-Dithiothreitol (DTT), acetyl chloride, N,N-dimethylformamide (DMF), triethylamine (TEA), dichloromethane (DCM), and menthol (MeOH) were purchased from Sigma-Aldrich (Finland). Cellulose ester membranes (dialysis bag) with a molecular weight cut-off value (MWCO) of 3500 were purchased from Solarbio.com, Beijing, China.

2.2. Synthesis and characterization of DOX-S-S-PTX

As shown in Fig. 2, the amphiphilic DOX-S-S-PTX prodrug conjugate was synthesized through esterification and an amide coupling reaction.
image file: c8bm01333k-f2.tif
Fig. 2 Synthesis route of DOX-S-S-PTX.
2.2.1. Synthesis of dithiodipropionic anhydride (DTDPA). DTDPA was obtained by acylation of 3,3′-dithiodipropionic acid (DTPA) with acetyl chloride, according to the previous literature.27 Briefly, DTPA (3.0 g, 48 mmol) was dissolved in acetyl chloride (30 mL) and refluxed at 70 °C for 12 h. After the solvent was removed, the residue was precipitated into excess ethyl ether to afford DTDPA and vacuum-dried (2.7 g, 90%) as a white solid. 1H NMR (400 MHz, DMSO-d6): δ1H NMR (500 MHz, DMSO-d6) δ 2.87 (t, J = 6.9 Hz, 4H), 2.61 (t, J = 6.9 Hz, 4H).
2.2.2. Synthesis of PTX-S-S-COOH. The PTX derivative PTX-S-S-COOH was synthesized through esterification.28 In brief, PTX (1.0 g, 1.17 mmol), DTDPA (270.2 mg, 1.41 mmol), TEA (142.2 mg, 191.7 μl) and DMAP (28.61 mg, 0.23 mmol) were dissolved in 20 ml of methylene chloride. The mixture was stirred at room temperature overnight. The completion of the reaction was monitored by LCMS until the starting material vanished. Then, the crude product was purified by silica gel column chromatography to obtain the pure intermediate (yield = 88%). 1H NMR (400 MHz, DMSO-d6): δ1H NMR (500 MHz, DMSO-d6) δ1H NMR (500 MHz, DMSO-d6) δ 9.48 (d, J = 7.1 Hz, 1H), 7.98 (d, J = 7.2 Hz, 2H), 7.88 (d, J = 7.1 Hz, 2H), 7.73 (d, J = 6.9 Hz, 1H), 7.67 (d, J = 7.3 Hz, 2H), 7.54 (d, J = 7.0 Hz, 1H), 7.45 (d, J = 7.3 Hz, 4H), 7.19 (d, J = 6.3 Hz, 1H), 6.30 (s, 1H), 5.81 (t, J = 8.1 Hz, 1H), 5.56 (t, J = 8.6 Hz, 1H), 5.41 (t, J = 7.7 Hz, 2H), 4.91 (d, J = 9.0 Hz, 1H), 4.66 (s, 1H), 4.14–4.09 (m, 1H), 4.01 (s, 2H), 3.61–3.56 (m, 2H), 2.95–2.74 (m, 8H), 2.35 (s, 1H), 2.25 (s, 3H), 2.10 (s, 3H), 1.79 (s, 3H), 1.72–1.57 (m, 2H), 1.50 (s, 3H), 1.23 (s, 1H), 1.01 (d, J = 15.1 Hz, 5H). HRMS (ES−) for C53H58NO17S2 [M − H] calculated 1044.3152, found 1044.3259 (Fig. S3). HRMS [2M] calculated 2090.6449, found 2090.6547.
2.2.3. Synthesis of PTX-S-S-DOX. PTX-S-S-DOX was synthesized through amide coupling. Briefly, DOX·HCl (77.93 mg, 0.14 mmol), PTX-S-S-COOH (100 mg, 0.1 mmol) EDC·HCl (27.49 mg, 0.14 mg) and TEA (14.51 ml, 19.55 μl) were dissolved in DMSO and protected from light with aluminium foil. The reaction mixture was stirred at room temperature overnight; then DMSO was removed by lyophilisation. The residues were purified by silica gel column chromatography (MeOH–CH2Cl2) to obtain the pure products to obtain a red powder (yield 56%). δ1H NMR (500 MHz, chloroform-d) δ 13.90 (s, 1H), 13.16 (s, 1H), 8.10 (d, J = 7.2 Hz, 2H), 7.96 (d, J = 7.4 Hz, 1H), 7.74 (d, J = 8.2 Hz, 1H), 7.72 (d, J = 7.2 Hz, 2H), 7.60 (t, J = 7.4 Hz, 1H), 7.50 (t, J = 7.7 Hz, 2H), 7.43 (t, J = 7.4 Hz, 1H), 7.37–7.33 (m, 8H), 6.31 (d, J = 8.4 Hz, 1H), 6.28 (s, 1H), 6.15 (t, J = 8.6 Hz, 1H), 5.90 Hz, (1H), 5.43 (s, 1H), 5.21 (s, 1H), 4.70 (s, 2H), 4.39 (dd, J = 10.7, 6.8 Hz, 1H), 4.28 (d, J = 8.4 Hz, 1H), 4.17 (d, J = 8.5 Hz, 1H), 4.11 (d, J = 6.5 Hz, 2H), 4.02 (s, 3H), 3.76 (d, J = 7.0 Hz, 1H), 3.58 (s, 1H), 3.19 (d, J = 18.6 Hz, 1H), 2.93 (d, J = 18.7 Hz, 1H), 2.88–2.69 (m, 8H), 2.54–2.44 (m, 2H), 2.39 (s, 3H), 2.33–2.23 (m, 3H), 2.20 (s, 3H), 2.13–2.04 (m, 3H), 1.88 (s, 3H), 1.77 (d, J = 7.7 Hz, 2H), 1.65 (s, 3H), 1.24 (d, J = 6.3 Hz, 4H), 1.19 (s, 3H), 1.12 (s, 3H). HRMS (ES+) for C80H86N2O27S2 [M + Na]+ calculated 1593.4757, found 1593.4638 (Fig. S3).
2.2.4. Characterization of the prodrug. The 1H NMR spectra of DTDP, DTDPA, DOX, PTX, PTX-S-S-COOH and DOX-S-S-PTX were recorded on Bruker 500 NMR spectrometers (Bruker, Billerica, MA, USA) and chemical shifts (δ, ppm) are quoted relative to the residual solvent peak (Chemical shifts (δ, ppm) are reported relative to the solvent peak (CDCl3, 7.26 [1H]; DMSO-d6, 2.50 [1H]). Mass spectra were recorded on a Bruker Daltonics microTOF-Q mass spectrometer (Bruker, Billerica, MA, USA). In addition, the FTIR spectra of DOX, PTX, PTX-S-S-COOH and DOX-S-S-PTX were recorded on a Thermo Scientific Nicolet iS50 Fourier transform infrared spectrometer in the wavenumber range 400–4000 cm−1.

2.3. Fabrication of three-dimensional (3D) microfluidic devices

A 3D microfluidic co-flow focusing device was fabricated by assembling two (inner and outer) borosilicate glass capillaries on a glass slide.29 One end of the cylindrical capillary (outer diameter of around 1000 μm; World Precision Instruments Ltd, UK) was tapered using a magnetic glass microelectrode horizontal needle puller (P-31, Narishige Co., Ltd, Japan). The inner tapered capillary was polished using sandpaper (Indasa Rhynowet, Portugal) until the cross-section of the shape and end became flattened. The inner tapered capillary was inserted into another cylindrical capillary with an inner dimension of around 1100 μm (World Precision Instruments Ltd, UK), and coaxially aligned. A transparent epoxy resin (5 Minute® Epoxy, Devcon) was used to seal the capillaries when required.

2.4. Preparation of nanoparticles

We firstly prepared the prodrug particles using a bulk method. In detail the ethanolic prodrug solution was added dropwise into water with a surfactant, followed by centrifugation to remove ethanol and resuspend the particles. By comparison, the DOX-S-S-PTX nanoparticles were also prepared by our in-house microfluidics devices.30,31 Two miscible liquids (aqueous and methanol) were injected separately into the microfluidic device through polyethylene tubes connected to syringes by a needle at constant flow rates. The flow rate of the different liquids was controlled by pumps (PHD 2000, Harvard Apparatus, USA). The DOX-S-S-PTX in methanol (5 mg mL−1) served as the inner dispersed phase; meanwhile, a Pluronic® F-127 (Sigma-Aldrich, Finland), 0.1% aqueous solution was selected as the outer continuous fluid. The inner (2 mL h−1) and outer (40 mL h−1) fluids were separately pumped into the microfluidic device, in which the inner fluid was focused by the outer continuous fluid. In this procedure, water-amphiphilic DOX-S-S-PTX self-assembled into nanoparticles during diffusion from the ethanol solution into water, and thus, bare DOX-S-S-PTX nanoparticles were obtained. In order to optimize the physicochemical properties of the prepared nanoparticles, including particle size, polydispersity index (PDI) and zeta (z)-potential, several process variables and formulation parameters were evaluated, such as the flow ratio between the inner and outer fluids and the concentration of DOX-S-S-PTX.

2.5. Characterization of the nanoparticles

Particle sizing was performed using dynamic light scattering with a Zetasizer Nano ZS (Malvern Instruments Ltd, UK). For each measurement, the sample (1.0 mL) was put in a disposable polystyrene cuvette (SARSTEDT AG & Co., Germany). The nanocarrier surface ζ-potential was measured with a Zetasizer Nano ZS by using disposable folded capillary cells (DTS1070, Malvern, UK). Both the size and ζ-potential were recorded as the average of three measurements. The structure of the fabricated nanoparticle was evaluated by a transmission electron microscope (TEM; JEOL 1400 Plus, JEOL, USA) at an acceleration voltage of 120 kV. The TEM samples were prepared by depositing 10 μL of the nanoparticle suspensions (1.0 mg mL−1) onto carbon coated copper grids (200 mesh; Ted Pella, Inc., USA). Samples were blotted away after 5 min of incubation, then air-dried prior to imaging.

2.6. The in vitro release of DOX and PTX from the nanoparticles

The cumulative levels of DOX and PTX released from the DOX-S-S-PTX nanoparticles were characterized using the dialysis method.32 In brief, 5 mL of PTX-loaded micellar solutions (1.0 mg mL−1) in PBS (1×, pH 7.4, used as before except for additional statement) were transferred into a dialysis membrane bag (MWCO 3500, Fisher Scientific), which was then immersed in 50 mL of PBS with or without DTT (10 mM), and suspended in a water bath at a constant temperature of 37 °C with horizontal shaking. Herein, DTT was used as a reducing agent to mimic the role of GSH, which in the microenvironment of cancer cells provides a reducing environment. At each predetermined time interval, 1 mL of incubated solution was taken out and replenished with an equal volume of corresponding PBS. The amounts of PTX and DOX were all quantified by HPLC using an Agilent 1100 system (Agilent Technologies, USA). PTX and DOX were simultaneously determined with a mobile phase composed of water and acetonitrile. The wavelengths used for PTX and DOX were 254 nm. The flow rate of the mobile phase was 1.0 mL min−1, the temperature was set at 30 °C, using a Waters Symmetry Shield RP18 Column (4.6 × 250 mm, 5 mm, Waters Corporation, USA) as the stationary phase, and the sample injection volume was 20 μL. A binary solvent system was used (solvent A, 0.1% aqueous TFA; solvent B, 0.1% TFA ACN), with UV detection by a detector (UV-975, Jasco) at 254 nm. A gradient of 5–95% of solvent B over 20 min for a 25 min run time was used to first identify the retention time of the parent drug (Fig. S1); then build-up standard curves (Fig. S2) were established from known concentrations of PTX in ethanol solution and DOX in Milli-Q water.

2.7. Cell studies

2.7.1. Cell culture and maintenance. Triple negative breast cancer cell lines MDA-MB-231, MDA-MB-231/ADR (MDA-MB-231 cell line with induced drug resistance by doxorubicin) and healthy cells of mouse embryonic fibroblasts (MEF) were used for in vitro studies. The p-glycoprotein expression level of the MDA-MB-231/ADR cells was confirmed (data not shown). The cells were cultured in high-glucose Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% FBS, 1% penicillin–streptomycin and 2 mM L-glutamine at 37 °C, in a humidified incubator with 5% CO2. Cells were passaged 2–3 times a week once they reached 90–100% confluency.

Triple negative breast cancer cell lines MDAMB-231 (ATCC), MDAMB-231/ADR (MDAMB-231 cell line with induced drug resistance) (a generous gift from F. Chen's group, Åbo Akademi, Finland) and mouse embryonic fibroblasts (MEF)(ATCC) as healthy cells were used for in vitro studies. The cell source of MDAMD-231 and MEF is ATCC and the drug resistant cell line is a gift from Chen's group.

2.7.2. Cytotoxicity assay. A WST-1 cell viability assay was used to determine drug efficacy in cancerous and healthy cells. MDA-MB-231, MDA-MB-231/ADR cancer cells and MEF cells were incubated overnight in a 96-well-plate (7000 cells per well) in cell growth media at 37 °C with 5% CO2. The following day, the cell growth media were replaced with fresh media containing 0.05, 0.1, 0.5, 1, 5 or 10 μM free drug or nanoparticle concentrations and incubated for 24 h or 48 h. Free drug stock solutions (DOX, PTX, DOX + PTX) were prepared in DMSO and nanoparticles (DOX-S-S-PTX) were suspended in water. All the dilutions for the cell viability assay, including the control, were prepared in cell growth media with DMSO concentration of 1%. After incubation with free drug or nanoparticles, 10 μL of WST-1 reagent was added to each well and the cells were incubated for a determined time (2 h for cancer cells and 3 h for fibroblasts) at 37 °C with 5% CO2. After incubation, the absorbance was measured by a Varioskan Flash Multimode Reader (Thermo Scientific Inc., Waltham, MA, USA) at 440 nm. Duplicates or triplicates were used for the experiment and averaged absorbance readings were plotted. To eliminate the background due to doxorubicin, absorbance values of cells without WST-1 reagent were measured and subtracted prior to the plotting.
2.7.3. Cellular uptake study. Cells were incubated overnight for attachment in 12-well plates (15 × 104 cells per well). DOX, PTX, DOX + PTX and DOX-S-S-PTX NPs were incubated with cells to keep the final concentration of DOX at 0.4 μM in cell media for 6 h and 24 h timepoints at 37 °C. Cells were collected by trypsin, washed twice with PBS, and acquisition of cellular uptake was determined by a flow cytometer BD LSRFortessa (BD Biosciences) by using the PE channel (Exmax 496 nm/Emmax 578 nm). All measurements were carried out in triplicate, and the results were analysed by Flowing Software 2.0. The gate was defined for live cells only; 20[thin space (1/6-em)]000 cells were recorded per sample. The fluorescence intensities of stocks (drug and nanoparticles) were measured by a Thermofisher Varioskan plate reader for the same Ex/Em values and the results were normalized for comparison of cellular uptake. The main fluorescence intensity of the control cells was proportioned with the results to achieve uptake efficiency.
2.7.4. Doxorubicin localization by confocal microscopy. Evaluations of drug and nanoparticle localization in cells were determined by confocal microscopy. Cells were grown on coverslips (15 × 104 cells per sample) overnight in 6-well plates. The medium was replaced by solutions of DOX, PTX, DOX + PTX and nanoparticles (0.4 μM), respectively. After 1 h, 6 h and 24 h timepoints, the cells were rinsed with PBS, fixed with 4% PFA and the sample was mounted using VECTASHIELD® with DAPI for microscopy. A Zeiss LSM780 confocal microscope (Plan-Apochromat 100×/1.40 Oil DIC), oil objective and Zen 2010 software setup was used for imaging. Detection of DAPI was performed with 405 nm laser excitation and 450–500 nm emission. Argon laser 488 excitation was utilized for doxorubicin and emission was collected at 530–600 nm.

2.8. Ex vivo red blood cell hemolysis assay

Intravenous injection compatibility was investigated by ex vivo hemolysis assay, as described by Evans et al.33 Briefly, 5 ml of blood was obtained from a healthy anonymous human donor according to ethical requirements and drawn directly into Na2EDTA coated tubes (1.6 mg ml−1) to prevent coagulation. Full blood was centrifuged at 500g for 5 min and haematocrit and plasma levels were marked on the tube. Red blood cells were washed by replacing plasma with 150 mM NaCl solution to the original plasma level. Cells were mixed gently and centrifuged at 500g for 5 min. Red blood cells were washed twice with NaCl solution, and once with PBS and thereafter diluted 50 times in PBS. Stock solutions of DOX, PTX, DOX + PTX and DOX-S-S-PTX were prepared in DMSO at ×1000 the desired concentration to be tested, to eliminate the toxic effect of DMSO. 1 μL of stock solution was added to 1 ml of red blood cell suspension and incubated at 37 °C for 24 h. For a positive control, 1% final concentration of Triton-X100 and for a negative control 0.1% final concentration of DMSO was used. After incubation, the cells were centrifuged at 500g for 5 min, 200 μL of supernatant was collected from each sample and the absorbance of haemoglobin was measured by a Varioskan plate reader at 500 nm. The experiment was done in triplicate and %hemolysis was calculated according to:
image file: c8bm01333k-t1.tif

2.9. Modelling

To demonstrate that the prodrug itself has no or negligible therapeutic effects until it is broken down into daughter drugs PTX-SH and DOX-SH, a docking simulation was also carried out. The protein structures (PDB code: 1JFF and 1D12) were prepared through the Protein Preparation Wizard panel34 to assign atom type and side chain protonation states before use. The molecular modelling suite Glide_XP35 was used to predict the binding affinity of PTX-SH and DOX-SH by defining a 10 Å empirical box localized at the centroid of PTX or DOX in complex structures.

3. Results and discussion

3.1. Characterization of DOX-S-S-PTX

The FTIR spectra of DOX, PTX, PTX-S-S-COOH and DOX-S-S-PTX are presented in Fig. 3. As can be seen in the spectra of DOX·HCl (Fig. 3A), the peak at 1617 cm−1 is assigned to the –NH2 group bending vibration of DOX.36,37 The characteristic –OH stretching vibration peaks of PTX (Fig. 3B) are at 3398 cm−1.38 An intense peak was observed at 1715 cm−1 (Fig. 3C), which is due to the absorption of the C[double bond, length as m-dash]O of carboxylic acid. In theory, one equivalent DTPA bearing two carboxylic acid groups can possibly react with two stoichiometric equivalents of PTX to give the PTX dimmer PTX-S-S-PTX. During our experiments, a maximum yield of 10% PTX dimmer was detected while at least an 80% yield of the desired PTX-S-S-COOH was recovered by following the suggested synthetic route of Yin et al.28 To completely avoid unwanted dimmer product PTX-S-S-PTX, DTPA was dehydrated by refluxing in methylene chloride at 70 °C to get cyclic anhydride DTDPA at first, then the DTDPA reacted with PTX to obtain only monomeric PTX-S-S-COOH as the main product. This result is consistent with the report of cyclic succinic anhydride which reacted with propargyl alcohol39 and β-cyclodextrin40 to give carboxylic acid terminated products. After the amide coupling reaction between PTX-S-S-COOH and DOX·HCl, the –NH2 bending vibration peak disappears, while new peaks at 1721 cm−1 and at 1579 cm−1 are presented in Fig. 3D, which are attributed to the C[double bond, length as m-dash]O stretching and –NH– bending vibration, respectively. The above results indicate the successful synthesis of DOX-S-S-PTX.
image file: c8bm01333k-f3.tif
Fig. 3 FTIR spectra of (A) DOX·HCl; (B) PTX; (C) PTX-S-S-COOH; (D) DOX-S-S-PTX.

1H NMR spectra of DOX-S-S-PTX and its intermediate products are shown in Fig. 4. As can be seen in Fig. 4A, the –COOH peak of DTDP is at d 12.35 ppm. After refluxing in acetyl chloride, the –COOH peak of DTDP disappeared in Fig. 4B, suggesting the formation of DTDPA.


image file: c8bm01333k-f4.tif
Fig. 4 1H NMR spectra of (A) DTDP in DMSO-d6; (B) DTDPA in CDCl3; (C) PTX in CDCl3; (D) PTX-S-S-COOH in DMSO-d6; (E) DOX in DMSO-d6; (F) DOX-S-S-PTX in DMSO-d6.

In addition, the thermogravimetric analysis shows that the melting point of DTDP is in the range 153–155 °C, and that for DTDPA is 65–7 °C. A dramatic change in melting points between DTDP and DTDPA also demonstrates the successful synthesis of DTDPA. The reaction of DTDPA and PTX is confirmed by the formation of an ester bond. The methylene (–CH2–CH2–) peaks of DTDPA at δ 2.87 ppm and 2.61 ppm appear in both DTDPA and PTX-S-S-COOH spectra (Fig. 4B and D).

The aromatic proton peaks of PTX are in the range 7.25–8.25 ppm, and 1.0–2.5 ppm for acetyl and methyl protons (Fig. 4C). Among all the hydroxyl groups in PTX, activated DTPA (DTDPA) prefers to react with the hydroxyl group of PTX linked with C-2′ to form an ester bond. A comparison of the 1H NMR spectra of PTX (Fig. 4C) and COOH-S-S-PTX (Fig. 4D) shows that the –OH peak at δ 4.71 ppm in PTX disappears in Fig. 4D, but the –COOH peak of COOH-S-S-PTX appears at 9.49 ppm; furthermore, proton peaks from 2.80 to δ 2.90 ppm show the methylene (–CH2–CH2–S–S) peak of DTPA. The HRMS spectra of the product give a mass of 1044.3259 ([M − H] in Fig. S3A), which together suggest that the desired molecules have been obtained.

Fig. 4E and F are the 1H NMR spectra of DOX and DOX-S-S-PTX, respectively. Among all the functional groups in DOX, the –NH2 of daunosamine is in the most suitable position for modification of the structure. As can be seen in Fig. 4E, the –NH2 peak of DOX is at 8.03 ppm. After the reaction of DOX with COOH-S-S-PTX, the –NH2 peak disappeared in Fig. 4F. Furthermore, the –COOH of COOH-S-S-PTX is at 9.49 ppm, which also disappeared in Fig. 4F. In addition, the observed mass of 1593.4638 [M + Na]+ from the HRMS spectra also suggested the successful conjugation of COOH-S-S-PTX with DOX into the prodrug DOX-S-S-PTX.

3.2. The morphology, size and size distribution of the DOX-S-S-PTX nanoparticles

The intrinsic amphiphilicity of the DOX-S-S-PTX prodrug favours its self-assembly into nanoparticles in the aqueous environment, which was demonstrated by both the bulk method and the microfluidics platform. However, the bulk approach fails to give ideal results. As shown in Table S1, the smallest particle obtained is about 350.1 nm which is beyond the empirical particle size suitable for cellular uptake of no larger than 200 nm. In general, both the size and polydispersity properties of particles made by bulk methods are significantly poorer than those from microfluidic devices. Using microfluidic devices, the effects of concentration of prodrug and inner[thin space (1/6-em)]:[thin space (1/6-em)]outer (I[thin space (1/6-em)]:[thin space (1/6-em)]O) fluid flow on the characteristics of DOX-S-S-PTX nanoparticles were investigated to optimize the formulation. The morphology and size of DOX-S-S-PTX nanoparticles are shown in Fig. 5. The size, size distribution and ζ-potential of the nanoparticles were measured by DLS. At a fixed I[thin space (1/6-em)]:[thin space (1/6-em)]O fluid flow of 2[thin space (1/6-em)]:[thin space (1/6-em)]40, changing the concentration of prodrug in the inner flow led to variation in particle size and PDI (Fig. 5A and B). At the lowest DOX-S-S-PTX of 1 mg ml−1, an average particle size of 99.04 nm (Fig. 5A) and a polydispersity index (PDI) of 0.106 (Fig. 5B) were achieved. The increased prodrug concentration barely affects the average size of the nanoparticles but weakens their uniformity (Fig. 5B). When DOX-S-S-PTX is increased from 2 to 3 mg ml−1, the particle size enlarged to 181.9 nm but the polydispersity improved (PDI of 0.084). A more concentrated inner fluid payload of up to 4 mg ml−1 worsened the characteristics of the as-prepared nanoparticles, especially the PDI. Similarly, the physicochemical characteristics of the prodrug particles are also dependent on I[thin space (1/6-em)]:[thin space (1/6-em)]O fluid flow. The average size of the DOX-S-S-PTX nanoparticles at I[thin space (1/6-em)]:[thin space (1/6-em)]O fluid flow = 1[thin space (1/6-em)]:[thin space (1/6-em)]40 is 147.2 nm (Fig. 5C) with a PDI of 0.187 (Fig. 5D) at a fixed DOX-S-S-PTX prodrug concentration of 3 mg ml−1, while the average size is 181.9 nm (Fig. 5C) with a PDI of 0.084 (Fig. 5D), and is therefore monodispersed when I[thin space (1/6-em)]:[thin space (1/6-em)]O = 2[thin space (1/6-em)]:[thin space (1/6-em)]40. The increased PDI at the lowest I[thin space (1/6-em)]:[thin space (1/6-em)]O fluid flow (1[thin space (1/6-em)]:[thin space (1/6-em)]40) indicates possible aggregation of already formed small nanoparticles due to the slow mixing of inner and outer fluid.41 At I[thin space (1/6-em)]:[thin space (1/6-em)]O = 4[thin space (1/6-em)]:[thin space (1/6-em)]40, both particle average size and size distribution increase to 220.9 nm and 0.174. A further increase of inner fluid flow worsens the particle size and PDI up to 274.6 nm and 0.22 compared to the 2[thin space (1/6-em)]:[thin space (1/6-em)]40 ratio. This is because the increased inner fluid flow leads to the formation of microvortices which enhance the average mass transfer rates between inner and outer fluid flow, and therefore accelerate the prodrug precipitation, making the formation of big sized nanoparticles more likely. The ζ-potential of prepared prodrug nanoparticles also varies (Fig. 5E) with fluctuation in I[thin space (1/6-em)]:[thin space (1/6-em)]O fluid flow. At I[thin space (1/6-em)]:[thin space (1/6-em)]O fluid flow = 2[thin space (1/6-em)]:[thin space (1/6-em)]40 the ζ-potential of the nanoparticles is about −18.5 mV, demonstrating the most stable product against aggregation in the current experiment.
image file: c8bm01333k-f5.tif
Fig. 5 The morphology and size of DOX-S-S-PTX nanoparticles. (A) Hydrodynamic size and (B) PDI at different DOX-S-S-PTX prodrug concentrations with inner[thin space (1/6-em)]:[thin space (1/6-em)]outer (I[thin space (1/6-em)]:[thin space (1/6-em)]O) flow fixed at 2[thin space (1/6-em)]:[thin space (1/6-em)]40 ml h−1; (C) hydrodynamic size, (D) PDI at different inner[thin space (1/6-em)]:[thin space (1/6-em)]outer (I[thin space (1/6-em)]:[thin space (1/6-em)]O) flows and (E) ζ-potential at 3 mg ml−1 with DOX-S-S-PTX concentration of 3 mg ml−1; (F) TEM of DOX-S-S-PTX nanoparticles with inner[thin space (1/6-em)]:[thin space (1/6-em)]outer (I[thin space (1/6-em)]:[thin space (1/6-em)]O) flow fixed at 2[thin space (1/6-em)]:[thin space (1/6-em)]40 ml h−1 at 3 mg ml−1 with DOX-S-S-PTX of 3 mg ml−1.

It is generally suggested that nanoparticles with a size of 10–200 nm can passively accumulate in tumour cells via the EPR effect.42,43 Therefore, a fluid flow I[thin space (1/6-em)]:[thin space (1/6-em)]O = 2[thin space (1/6-em)]:[thin space (1/6-em)]40 and inner fluid with 3 mg ml−1 DOX-S-S-PTX prodrug concentration were applied to the inner fluid to fabricate as-prepared DOX-S-S-PTX for passive delivery to tumour cells. As shown in Fig. 5F, the transmission electron microscopy (TEM) images of the nanoparticles demonstrate the uniform size of the prodrug nanoparticles which are spherical and porous with centred large pores surrounded by multiple smaller ones. The regular shape and hollow structure of the particles may indicate the mechanical properties of the particles, which are made of DOX-S-S-PTX prodrug.

3.3. In vitro release of DOX from the DOX-S-S-PTX nanoparticles

The in vitro release profiles of the DOX-S-S-PTX nanoparticles in the absence and presence of DTT are shown in Fig. 6A. The accumulative release level of DOX from the nano-preparation of the DOX-S-S-PTX prodrug shows no difference in the first 10 h with or without DTT. A possible explanation is that both amide bond hydrolysis and disulfide bond breakage are involved in this process. In the beginning, it is not easy for the DTT to penetrate the prodrug nanoparticles, since hydrophilic DOX forms the outer layer of the nanoparticles. Therefore, the DOX predominantly gets released by hydrolysis of the amide bond in the nanoparticles of the prodrug by up to 40.9% in the PBS solution. After 10 h, the DTT can easily approach the disulfide bond in the prodrug nanoparticles due to the swelling effect. DOX can now be released by breaking the disulfide bond by DTT. Nevertheless, the release profiles of samples with and without DTT scientifically differ after 10 h. When DTT is not charged, the concentration of DOX remains unchanged, indicating that no further DOX molecules are released.
image file: c8bm01333k-f6.tif
Fig. 6 (A) The in vitro release of DOX from DOX-S-S-PTX; (B) the haemolytic ratio of DOX-S-S-PTX nanoparticles.

In contrast, the sample with DTT DOX molecules continues to release until 72 h, and gives an 89.7% accumulative level, suggesting that the disulfide bond damage contributes about 50% of the DOX release from DOX-S-S-PTX nanoparticles. The DOX release profile agrees with the report of the behaviour of mPEG-S-S-PTX micelles under similar experimental conditions.44 However, the in vitro release profile of PTX (Fig. 6A) is not coincident with DOX, since only 1.5% and not a corresponding amount of PTX was detected. This is because the poor solubility prevents PTX being in aqueous PBS, while the liquid chromatographic method cannot monitor undissolved PTX. It is possible to introduce a surfactant such as Tween 80 to the experimental buffer to generate a homogeneous PTX suspension. But concern that the surfactant may also increase the dissociation of prodrug molecules from the prodrug nanoparticles warns us not to do so.

3.4. The haemolytic test of the DOX-S-S-PTX nanoparticles

A haemolytic test was carried out to investigate the safety and suitability of the fabricated DOX-S-S-PTX prodrug nanoparticles for intravenous injection. The results of the haemolytic test of NPs, DOX, PTX, and a mixture of DOX and PTX are shown in Fig. 6B. The haemolytic ratio of DOX-S-S-PTX nanoparticles at concentrations of 0.5 and 5 μM are all lower than 1%, suggesting that the NPs cannot cause red blood cell lysis.45 In particular, at a higher concentration of 5 μM, NPs are safer than DOX alone or their combination, since they triggered a two to three fold higher haemolytic ratio. Therefore, the results show that our as-prepared prodrug NPs are biocompatible for intravenous injection.

3.5. In vitro cytotoxicity and cell apoptosis

The cell viability of free drugs (DOX, PTX, DOX + PTX) and prodrug NPs was investigated with the WST-1 assay for MDA-MB-231, MDA-MB-231/ADR and MEF cells. Time points of 24 h and 48 h were chosen for this study. Fig. 7 demonstrates that both free drugs reduce the viability of MDA-MB-231, MDA-MB-231/ADR and MEF cells. However, prodrug NPs show selective inhibition with regard to cancer cells, while they do not alter the cell viability of healthy cells (Fig. 7E and F). In terms of free DOX, PTX, a DOX + PTX mixture and prodrug NPs, a dose-dependent manner was coherently observed in the inhibition of cell proliferation. In addition, the 48 h inhibition ratio was higher than that for 24 h. These observations indicate that free DOX, PTX, and NPs have time-dependent and concentration-dependent cytotoxic effects on cancer cells, but no effects were exerted on MEF cells by NPs. Fig. 7C and D suggest that the inhibition ratios of prodrug NPs to drug-resistant cancer cells were higher than for free DOX, PTX or a physical mixture of DOX and PTX, implying that the faster NP uptake and synergism of DOX and PTX may play a role.
image file: c8bm01333k-f7.tif
Fig. 7 Cytotoxicity of DOX, PTX, DOX + PTX mixture and NPs. MDA-MB-231 cells incubated for (A) 24 h, (B) 48 h, MDA-MB-231/ADR cells incubated for (C) 24 h, (D) 48 h, MEF cells incubated for (E) 24 h, (F) 48 h.

3.6. Cellular uptake

The cellular uptakes of DOX, DOX + PTX and prodrug NPs with MDA-MB-231, MDA-MB-231/ADR and MEF cells were investigated by flow cytometry for 6 h and 24 h time points. As illustrated in Fig. 8A and B, NPs exhibited poor uptake in healthy cells, whereas higher uptake was observed for free drugs; therefore NPs have the potential to eliminate possible side effects. PTX incorporation does not affect cellular uptake of DOX. In all tested cell lines, the uptake of NPs is lower compared to free drugs, but faster since no drastic difference was observed between 6 h and 24 h, implying the cellular uptake of NPs can quickly equilibrate. The lower NP uptake is consistent with Cho's report,46 and could be a result of the negativity of the NP surface ζ-potential since the cell membrane is negatively charged47 due to its phospholipid bilayer structure; hence isoelectric charge repulsion prevents NPs from cell internalization. However, it was also reported that negatively charged particles had more cellular uptake than cationic particles. So it is suggested that the surface charge of nanoparticles is not the only affecting factor for their cellular uptake, but hydrophobicity is equally important.48,49 In our as-prepared prodrug NPs, hydrophilic DOX tends to deploy on the surface of particles while hydrophobic PTX embedded centered in the particles; the arrangement as such therefore may hinder the internalization of NPs into the hydrophobic cell membrane.50
image file: c8bm01333k-f8.tif
Fig. 8 The fluorescence microscopy images and flow cytometry of cellular uptake of DOX, PTX, DOX + PTX and prodrug NP. (A) Mean fluorescence intensity of drug; (B) flow cytometry of uptake in MDAMA-231-DOX and MEF cells and (C) fluorescence microscopy images with scale bar 10 μm.

DOX is a fluorescent drug which localizes in the nucleus as a mechanism of anticancer action,51 which enables the microscopy imaging of nuclear localization. The result of the nuclear internalization of DOX was investigated by confocal fluorescence microscopy and images of the prodrug NPs and free drugs with MDA-MB-231, MDA-MB-231/ADR and MEF cells for 6 h and 24 h are shown in Fig. 8B and Fig. S4. For MDA-MB-231 and MDA-MB-231/ADR cells, DOX can already be detected in the nucleus after 1 h of incubation for free drug and prodrug NPs. The fluorescence can still be detected at 6 h and 24 h timepoints, which proves the efficacy of the prodrug NPs. It can clearly be seen that the intensity of the NPs is less than that of free drug, and this suggests that the fluorescence intensity loss could be due to prodrug conjugation. Regardless of the incubation timepoints (Fig. 8A), the fluorescence intensity of DOX in MEF cells treated with NPs showed no obvious change, whereas free DOX and DOX + PTX internalization was observed, which suggests the NPs mainly accumulate in MDAMA-231/ADR cells, hence eliminating possible unwanted side effects.

3.7. In silico simulation

PTX with the brand name Taxol is a cytoskeletal drug that targets β tubulin to interfere with the cell microtubule assembly-disassembly process to lead to an arrested cell in the G2/M-phase cycle, ultimately leading to cell apoptosis.52 It has been proposed that the release of PTX from mPEG-SS-PTX conjugated micelles can be achieved through both esterase-induced hydrolysis and GSH-related disulfide breakdown.44 The ester bond hydrolysis leads to native drug PTX, which we believe should go on to target macromolecules to enable the function. The reduction of the disulfide bond by GSH gave PTX-CH2-CH2-SH, which we believe should bind to PTX's biological target, since only a short spacer (–CH2–CH2–SH) was introduced. To confirm this, molecular docking was performed. Fig. 9A represents the binding of PTX in β tubulin. In parallel, the docked PTX–CH2–CH2–SH to β tubulin with the highest docking score is shown in Fig. 9B, demonstrating a very similar binding pose to PTX. The framework of PTX–CH2–CH2–SH localizes into the binding pocket of PTX while the spacer (–CH2–CH2–SH) moiety points to the solvent (highlighted by the dashed circle). Another half of the prodrug DOX–CH2–CH2–SH was also docked to the therapeutic target of DOX DNA topoisomerase II.54 A similar binding pose between DOX and the biological target was also obtained (the data is not shown). There is no doubt that the reduced products PTX–CH2–CH2–SH and DOX–CH2–CH2–SH can be enzymatically hydrolyzed into PTX and DOX by esterase and proteases, respectively.
image file: c8bm01333k-f9.tif
Fig. 9 The complex of β-tubulin with (A) PTX (PDB file: 1JFF53) and (B) predicted binding modes of PTX-SH. The dashed circle highlights the thiol moiety.

4. Conclusions

In this work, we propose a new drug delivery system based on a PTX-S-S-DOX prodrug by the conjugation of PTX and DOX through a disulfide bond. Due to the hydrophobicity balance of carefully selected components, the obtained DOX-S-S-PTX amphiphilic prodrug self-assembled into NPs through microfluidics-tailored nanoprecipitation. The size of the as-prepared NPs can be tuned by adjusting the inner[thin space (1/6-em)]:[thin space (1/6-em)]outer flow and concentration of prodrug in the inner solvent channel. The obtained NPs are stable and safe. The disulfide bond in the DOX-S-S-PTX conjugate can be broken in the reducing environment of tumour cells, leading to the controlled simultaneous release of two daughter drugs. The prepared NPs inhibit the cancer cells in a dose- and time-dependent manner but not healthy MEF cells. Although less NPs were taken up, their uptake is quicker compared to free drugs. NPs internalize into diseased cells in a large quantity but in a lower quantity in healthy ones, implying this prodrug nanoprecipitation formulation has advantages over free drug administration in terms of minimizing drug side effects.

Conflicts of interest

Authors declare no conflicts of interests.

Acknowledgements

This work was supported by Distinguished Clinical Investigator Grant of Jiangsu Province, China (Grant No. JSTP201701), Academy of Finland (Grant No. 297580) and Sigrid Jusélius Foundation (decision no. 28001830K1). Authors are thankful to Peiru Yang and Fang Chen for donating the DOX resistant cell line, Dhayakumar Rajan Prakash to reshape the Fig. 1 of this manuscript. Gunilla Henriksson is kindly acknowledged for helping hemolysis study.

Notes and references

  1. H. K. Weir, R. N. Anderson, S. M. C. King, A. Soman, T. D. Thompson, Y. Hong, B. Moller and S. Leadbetter, Prev. Chronic Dis., 2016, 13, 1160211 Search PubMed.
  2. K. Curtius, N. A. Wright and T. A. Graham, Nat. Rev. Cancer, 2018, 18, 19 CrossRef CAS PubMed.
  3. P. A. Pesavento, D. Agnew, M. K. Keel and K. D. Woolard, Nat. Rev. Cancer, 2018, 1 Search PubMed.
  4. J. Dubach, C. Vinegoni, R. Mazitschek, P. F. Feruglio, L. Cameron and R. Weissleder, Nat. Commun., 2014, 5, 3946 CrossRef CAS PubMed.
  5. F. Kamangar, G. M. Dores and W. F. Anderson, J. Clin. Oncol., 2006, 24, 2137–2150 CrossRef PubMed.
  6. J. Rautio, N. A. Meanwell, L. Di and M. J. Hageman, Nat. Rev. Drug Discovery, 2018, 559–587 CrossRef CAS PubMed.
  7. D. Platel, S. Bonoron-Adele, R. Dix and J. Robert, Br. J. Cancer, 1999, 81, 24 CrossRef CAS PubMed.
  8. M. Kapoor and R. A. Siegel, Mol. Pharm., 2013, 10, 3519–3524 CrossRef CAS PubMed.
  9. J. Hulla, S. Sahu and A. Hayes, Hum. Exp. Toxicol., 2015, 34, 1318–1321 CrossRef CAS PubMed.
  10. Y. Ma, Q. Mou, X. Zhu and D. Yan, Mater. Today Chem., 2017, 4, 26–39 CrossRef.
  11. N. T. Chi, N. M. Triet and D. M. Chien, J. Phys.: Conf. Ser., 2009, 187, 012047 Search PubMed.
  12. M. Iqbal, N. Zafar, H. Fessi and A. Elaissari, Int. J. Pharm., 2015, 496, 173–190 CrossRef CAS PubMed.
  13. D. Quintanar-Guerrero, M. de la Luz Zambrano-Zaragoza, E. Gutiérrez-Cortez and N. Mendoza-Muñoz, Recent Pat. Drug Delivery Formulation, 2012, 6, 184–194 CrossRef CAS.
  14. B. Nagavarma, H. K. Yadav, A. Ayaz, L. Vasudha and H. Shivakumar, Asian J. Pharm. Clin. Res., 2012, 5, 16–23 CAS.
  15. M. Beck-Broichsitter, E. Rytting, T. Lebhardt, X. Wang and T. Kissel, Eur. J. Pharm. Sci., 2010, 41, 244–253 CrossRef CAS PubMed.
  16. K. S. Yadav and K. K. Sawant, AAPS PharmSciTech, 2010, 11, 1456–1465 CrossRef CAS PubMed.
  17. J. Xu, S. Zhang, A. Machado, S. Lecommandoux, O. Sandre, F. Gu and A. Colin, Sci. Rep., 2017, 7, 4794 CrossRef PubMed.
  18. E. Chiesa, R. Dorati, T. Modena, B. Conti and I. Genta, Int. J. Pharm., 2018, 536, 165–177 CrossRef CAS PubMed.
  19. D. Liu, C. R. Bernuz, J. Fan, W. Li, A. Correia, J. Hirvonen and H. A. Santos, Adv. Funct. Mater., 2017, 27, 1604508 CrossRef.
  20. D. Liu, H. Zhang, S. Cito, J. Fan, E. Mäkilä, J. Salonen, J. Hirvonen, T. M. Sikanen, D. A. Weitz and H. l. A. Santos, Nano Lett., 2017, 17, 606–614 CrossRef CAS PubMed.
  21. H. Hillaireau, N. Dereuddre-Bosquet, R. Skanji, F. Bekkara-Aounallah, J. Caron, S. Lepêtre, S. Argote, L. Bauduin, R. Yousfi and C. Rogez-Kreuz, Biomaterials, 2013, 34, 4831–4838 CrossRef CAS PubMed.
  22. A. Gaudin, M. Yemisci, H. Eroglu, S. Lepetre-Mouelhi, O. F. Turkoglu, B. Dönmez-Demir, S. Caban, M. F. Sargon, S. Garcia-Argote and G. Pieters, Nat. Nanotechnol., 2014, 9, 1054 CrossRef CAS PubMed.
  23. P. Huang, D. Wang, Y. Su, W. Huang, Y. Zhou, D. Cui, X. Zhu and D. Yan, J. Am. Chem. Soc., 2014, 136, 11748–11756 CrossRef CAS PubMed.
  24. S. Lou, Z. Zhao, M. Dezort, T. Lohneis and C. Zhang, ACS Omega, 2018, 3, 9210–9219 CrossRef CAS PubMed.
  25. M. P. Gamcsik, M. S. Kasibhatla, S. D. Teeter and O. M. Colvin, Biomarkers, 2012, 17, 671–691 Search PubMed.
  26. R. Ngoune, A. Peters, D. von Elverfeldt, K. Winkler and G. Pütz, J. Controlled Release, 2016, 238, 58–70 CrossRef CAS PubMed.
  27. X. Zheng, Z. Li, L. Chen, Z. Xie and X. Jing, Chem. – Asian J., 2016, 11, 1780–1784 CrossRef CAS PubMed.
  28. T. Yin, Q. Wu, L. Wang, L. Yin, J. Zhou and M. Huo, Mol. Pharm., 2015, 12, 3020–3031 CrossRef CAS PubMed.
  29. D. Liu, H. Zhang, E. Mäkilä, J. Fan, B. Herranz-Blanco, C.-F. Wang, R. Rosa, A. J. Ribeiro, J. Salonen and J. Hirvonen, Biomaterials, 2015, 39, 249–259 CrossRef CAS PubMed.
  30. D. Liu, S. Cito, Y. Zhang, C. F. Wang, T. M. Sikanen and H. A. Santos, Adv. Mater., 2015, 27, 2298–2304 CrossRef CAS PubMed.
  31. B. Herranz-Blanco, E. Ginestar, H. Zhang, J. Hirvonen and H. A. Santos, Int. J. Pharm., 2017, 516, 100–105 CrossRef CAS PubMed.
  32. W. Zhu, Y. Wang, X. Cai, G. Zha, Q. Luo, R. Sun, X. Li and Z. Shen, J. Mater. Chem. B, 2015, 3, 3024–3031 RSC.
  33. B. C. Evans, C. E. Nelson, S. Y. Shann, K. R. Beavers, A. J. Kim, H. Li, H. M. Nelson, T. D. Giorgio and C. L. Duvall, J. Visualized Exp., 2013 Search PubMed , e50166.
  34. G. M. Sastry, M. Adzhigirey, T. Day, R. Annabhimoju and W. Sherman, J. Comput.-Aided Mol. Des., 2013, 27, 221–234 CrossRef PubMed.
  35. R. A. Friesner, R. B. Murphy, M. P. Repasky, L. L. Frye, J. R. Greenwood, T. A. Halgren, P. C. Sanschagrin and D. T. Mainz, J. Med. Chem., 2006, 49, 6177–6196 CrossRef CAS PubMed.
  36. L. Chen, Y. Xue, X. Xia, M. Song, J. Huang, H. Zhang, B. Yu, S. Long, Y. Liu and L. Liu, J. Mater. Chem. B, 2015, 3, 8949–8962 RSC.
  37. D. Zhao, J. Wu, C. Li, H. Zhang, Z. Li and Y. Luan, Colloids Surf., B, 2017, 155, 51–60 CrossRef CAS PubMed.
  38. G. Kathiravan, S. V. Raman, B. Rajangam and A. Rajasekar, J. Anal. Bioanal. Tech., 2014, 5, 1 Search PubMed.
  39. J.-H. Wang, C.-C. Cheng, Y.-C. Yen, C.-C. Miao and F.-C. Chang, Soft Matter, 2012, 8, 3747–3750 RSC.
  40. Y. Xi, Z. Luo, X. Lu and X. Peng, J. Agric. Food Chem., 2017, 66, 228–237 CrossRef PubMed.
  41. Y. Kim, B. Lee Chung, M. Ma, W. J. Mulder, Z. A. Fayad, O. C. Farokhzad and R. Langer, Nano Lett., 2012, 12, 3587–3591 CrossRef CAS PubMed.
  42. N. Hoshyar, S. Gray, H. Han and G. Bao, Nanomedicine, 2016, 11, 673–692 CrossRef CAS PubMed.
  43. E. Blanco, H. Shen and M. Ferrari, Nat. Biotechnol., 2015, 33, 941 CrossRef CAS PubMed.
  44. D. Zhao, H. Zhang, S. Yang, W. He and Y. Luan, Int. J. Pharm., 2016, 515, 281–292 CrossRef CAS PubMed.
  45. H. Wang, Y. Zhao, Y. Wu, Y.-l. Hu, K. Nan, G. Nie and H. Chen, Biomaterials, 2011, 32, 8281–8290 CrossRef CAS PubMed.
  46. E. C. Cho, J. Xie, P. A. Wurm and Y. Xia, Nano Lett., 2009, 9, 1080–1084 CrossRef CAS PubMed.
  47. B. Chen, W. Le, Y. Wang, Z. Li, D. Wang, L. Ren, L. Lin, S. Cui, J. J. Hu and Y. Hu, Theranostics, 2016, 6, 1887 CrossRef CAS PubMed.
  48. Z. Zhang, K. Van Steendam, S. Maji, L. Balcaen, Y. Anoshkina, Q. Zhang, G. Vanluchene, R. De Rycke, F. Van Haecke and D. Deforce, Adv. Funct. Mater., 2015, 25, 3433–3439 CrossRef CAS.
  49. A. Valsesia, C. Desmet, I. Ojea-Jiménez, A. Oddo, R. Capomaccio, F. Rossi and P. Colpo, ChemComm, 2018, 1, 53 CrossRef.
  50. L. Lacerda, H. Ali-Boucetta, S. Kraszewski, M. Tarek, M. Prato, C. Ramseyer, K. Kostarelos and A. Bianco, Nanoscale, 2013, 5, 10242–10250 RSC.
  51. N. S. H. Motlagh, P. Parvin, F. Ghasemi and F. Atyabi, Biomed. Opt. Express, 2016, 7, 2400–2406 CrossRef CAS PubMed.
  52. N. C. Kampan, M. T. Madondo, O. M. McNally, M. Quinn and M. Plebanski, BioMed Res. Int., 2015, 2015, 413076 Search PubMed.
  53. J. Löwe, H. Li, K. Downing and E. Nogales, J. Mol. Biol., 2001, 313, 1045–1057 CrossRef PubMed.
  54. J. L. Nitiss, Nat. Rev. Cancer, 2009, 9, 338 CrossRef CAS PubMed.

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/c8bm01333k

This journal is © The Royal Society of Chemistry 2019