Abstract
The advanced use of a pH-responsive biomaterial-based injectable liquid implant for effective chemotherapeutic delivery in glioblastoma multiforme brain (GBM) tumour treatment is presented. As an implant, we proposed a water-in-oil-in-water multiple emulsion with encapsulated doxorubicin. The effectiveness of the proposed therapy was evaluated by comparing the cancer cell viability achieved in classical therapy (chemotherapeutic solution). The experimental study included doxorubicin release rates and consumption for two emulsions differing in drop sizes and structures in the presence of GBM-cells (LN229, U87 MG), and a cell viability. The results showed that the multiple emulsion implant was significantly more effective than classical therapy when considering the reduction in cancer cell viability: 85% for the emulsion-implant, and only 43% for the classical therapy. A diffusion-reaction model was adapted to predict doxorubicin release kinetics and elimination by glioblastoma cells. CFD simulations confirmed that the drug release kinetics depends on multiple emulsion structures and drop sizes.
1. INTRODUCTION
Drug delivery in the treatment of the central nervous system diseases - CNS (brain tumours, trauma, infections, neurodegenerative problems, amongst others) requires passing through, or bypassing, the blood-brain barrier (BBB). The methods of drug administration to the CNS can be divided into three main groups: invasive techniques, non-invasive techniques and alternative routes.1,2,3 The non-invasive techniques explore approaches in which pharmaceuticals are re-engineered to cross the BBB via: (i) chemical methods (lipophilic analogues, prodrugs, enzymatic reactions or chemical bonding of drug molecules with transport facilitated molecules), or (ii) biological methods (drug attachment to proteins specific for receptors responsible for transport across the BBB, transport vectors or barrier-crossing peptides). In addition, nanoparticles, dendrimers, liposomes, micelles, micro/nanoemulsions, including targeted drug delivery systems, and stimuli-responsive functional biomaterials in the drug delivery area, are used to cross the BBB.4,5,6 The invasive techniques include: (i) local surgical treatment combined with adjuvant therapy (intracerebral polymer implants or microchips, intraventricular/intrathecal or interstitial drug delivery, biological tissue delivery) or (ii) controlled BBB damage with drug delivery (e.g. convection-enhanced drug delivery, osmotic or ultrasound disruption of the BBB). Surgical treatment in combination with radio and chemotherapy, in the case of brain tumours, plays a fundamental role in neurooncology. If possible, tumours should be removed completely. In most cases, only part of the tumour is surgically removed, and the remainder is irradiated or subjected to chemotherapy for destruction. The alternative methods bypass the cardiovascular system and include transnasal administration of drugs or iontophoretic delivery. Modern medicine responds to the needs of the patient based on various strategies including mathematical modelling for predicting the effect of chemotherapy on cancer cells. The approaches to simulating cancer interaction with therapy can generally range from (i) cancer growth models, then (ii) mass transfer models, which combine drug release, transport, and elimination in tissue, to (iii) cell-based models at the molecular level. The first group includes models of tumour growth/volume change based on the analysis of cell proliferation and cell death, along with the structure of an avascular (solid) tumour (homogeneous or heterogeneous), and a vascular heterogeneous tumour, including tumour-induced angiogenesis.7,8 The second group – reaction-diffusion/convection models are related to transport mechanisms for delivering drugs to the tumour and drug elimination in tissue by chemical reaction. These models provide macroscopic descriptions of the system based on space- and time-dependent variables such as drug concentration distribution, pressure in the tumour environment, and drug release rates.9,10,11 The third group represents discrete cell-based models describing the dynamics of cancer cells via individual cell behaviours within the tumour tissue based on the transformation and transport of substances to the cell at the molecular level. Models that treat cells individually can combine descriptions at the subcellular and cellular levels, with a macroscopic description of the tumour environment. This extended approach via, e.g. the lattice-based method or cellular automata, creates new hybrid computational models for simulating the cancer intercellular adhesion and invasion process.12,13,14,15,16 In the present study, a diffusion-reaction model was adapted to predict the accurate chemotherapeutic concentration after its in vitro release from a liquid implant in the presence of glioblastoma multiforme cells. Glioblastoma multiforme (GBM) is a primary malignant tumour of the central nervous system with one of the worst prognoses. Despite huge progress in the field of oncology, the median survival rate for patients after diagnosis is less than 2 years.2 As the liquid implant, we proposed multiple emulsions W/O/W (water-in-oil-in-water) type with encapsulated doxorubicin (chemotherapeutics) in the internal droplets, Fig. 1. Multiple emulsions are complex dispersed systems with a hierarchical structure of droplets of the first liquid dispersed in larger drops of a second immiscible liquid, which is a dispersion medium for smaller droplets. The larger drops are then dispersed either in the continuous phase of the first liquid (double emulsion) or in other still larger drops of the first or other immiscible liquid and so on (a multiple emulsion). A multiple emulsion can be then a double, triple, quadruple, quintuple, or even more structured system. These dispersed systems offer a wide range of possible applications in separation processes and environmental protection (alternative fuel), pharmaceuticals, and medicine, especially for the encapsulation and controlled release of active ingredients (drugs, living cells, antigen delivery, cosmetics, food).17,18,19,20,21,22,23,24,25Controlling by multiple emulsions is achieved through the size and physicochemical parameters of drops forming liquid-permeable membranes separating the internal droplets from the continuous external phase. The proposed method involves inserting the liquid implant intraoperatively into the cavity after surgical resection of the tumour, Fig. 1. This method is a bypassing of the BBB and belongs to the group of adjuvant treatments. Administering the drug in the form of multiple emulsions is intended to support the treatment, i.e. the cytotoxic effect on the tumour cells remaining after the surgery to prevent the recurrence of the tumour. The chemotherapeutic agent is released from the liquid implant into the GBM environment in a manner controlled by the pH of the tumour environment and also the drop size, structure, or physicochemical parameters of the emulsion (viscosity, density). The external phase of the emulsion contains a biopolymer (sodium carboxymethylcellulose) which, depending on the pH, changes the conformation of the chains and the viscosity, thus affecting the release rates. Such properties of the polymer are exploited in this concept for controlled release under the acidic tumour microenvironment. The use of the implant in a liquid form also reduces the risk of mechanical damage to healthy tissue during the intraoperative insertion in comparison to a solid implant. The paper aims to find optimal brain cancer treatment based on locally controlled chemotherapeutic release from an emulsion-based implant, including numerical simulations using a diffusion-reaction model for drug transport and consumption. To evaluate this model, the paper includes a comparison of these simulations with experimental data on anti-cancer drug release and consumption for two emulsions differing in drop sizes and structures. Also, cancer cell viability was investigated in comparison with classical chemotherapy involving cells treated with a drug in a solution.
2. MATERIALS AND METHODS
2.1 Preparation and characterization of multiple emulsions with anti-cancer drug (drop size, viscosity)
Multiple/double emulsions W1/O/W2 with doxorubicin hydrochloride - DOX (anti-cancer drug) were prepared in a Couette-Taylor flow (CTF) apparatus where liquid phases were intensively mixed due to rotational and axial flows. The emulsions preparation conditions are shown in Table 1. The detailed procedure for the preparation of the emulsion can be found in the previous authors’ works26,27,28. In short, the internal water phase (W1) with DOX and soybean oil as an organic membrane phase (O) were introduced in the inlet cross-section of the CTF apparatus and intensively mixed to create simple emulsions W1/O. Then, after introducing the water phase (W2) to the simple emulsions in the middle section, double emulsions W1/O/W2 were formed. The CTF apparatus provides high mass transfer parameters and uniform shear flow, contributing to the high encapsulation efficiency and formation of stable emulsions.28,29,30 The structures of the obtained emulsions were analyzed using an Olympus BX60 optical microscope, with Olympus SC50 digital camera (Olympus, Japan), and image analysis software, Image Pro Plus 4.5 (Media Cybernetics, USA). For each of the double emulsions, at least 800 drops of the membrane phase and 1000 drops of the internal phase were measured. Then drop sizes distributions were determined and the average drop sizes: the Sauter mean diameter of the internal (d32) and membrane phase (D32) drops. Also, the volume fractions of the internal phase drops in the membrane phase drops were determined (packing volume fraction). The fluorescence spectrofluorometer FLUOstar Optima (BMG LABTECH, Germany) was used to measure the concentration of the chemotherapeutic agent in the emulsion continuous phase to determine the encapsulation efficiency of the DOX in the internal phase droplets. The emulsions rheological tests were performed with a RheolabQC rotational rheometer (the measuring system of concentric cylinder geometry - gap size: 1.64mm, length 60mm, cone angel: 120°, ratio of radii: 1.08, range of shear rate: 1-1500 s−1, 37°C, Anton Paar, Austria). Two stable soybean oil-based emulsions, differing in the internal structure of the drops, with pH-responsive biopolymer (sodium carboxymethylcellulose - CMC-Na) in the external phase were selected for further experiments (Table 1). The emulsions were characterized by a high encapsulation efficiency of (DOX) (>95%), calculated based on the difference in DOX concentration introduced to the CTF apparatus and in the external phase of created multiple emulsions according to the procedure26. All compounds used to prepare emulsions were supplied by Sigma Aldrich. Detailed data on the composition and preparation conditions of the emulsions can be found in Table 1.
2.2 Glioblastoma model cell-lines culturing
The in vitro studies on cell viability and release of the anti-cancer/cytostatic agent from multiple emulsions were conducted for selected tumour cell lines of glioblastoma multiforme: U87 MG, LN229. The cell lines were procured from the Institute of Biochemistry and Biophysics PAS (Poland). The cells were cultured on 10 cm dishes (BD-Falcon, USA) to 80-90% confluence in DMEM medium with high glucose and L-glutamine (HyClone, Poland), with 10% fetal bovine serum (FBS; Gibco, Poland) and 1% penicillin/streptomycin antibiotics (Life Technologies, Poland) in an incubator (37°C; 5% CO2). The cells were passaged 24 hours prior to release and cytotoxicity experiments (cells were first rinsed with phosphate-buffered saline-PBS buffer (Lab Empire, Poland), then trypsinized by 0.25% trypsin and 0.1% EDTA (HyClone, Poland).
2.3 Release of anti-cancer drug from multiple emulsions
The in vitro release experiments of the anti-cancer/cytostatic drug (DOX) from multiple emulsions were carried out in the systems with and without glioblastoma cells (U87 MG, LN229) to determine the drug consumption by cells. Standard 12-well plates were used, 4000 cells/wells were seeded 24h before the release experiment and placed in the incubator (temperature: 37°C, 5% CO2, release volume: 1 cm3). The release process was analyzed for a specific concentration of DOX encapsulated in the emulsion (0.1 µM, 0.2 µM). The target concentration of DOX was obtained by diluting the emulsions with PBS buffer of pH = 7.4 (1:100, 1:50 - volume of emulsion to PBS buffer). At certain time intervals during the 24h experiment with cells and 96 h without cells, the entire volume of the diluted emulsion was taken from each well (1 cm3) and filtered through a hydrophilic syringe filter (nylon filter membrane, 0.2 µm). The DOX concentration in the samples was determined using a fluorescence spectrofluorometer - FLUOstar Optima (BMG LABTECH, Germany), measured at a wavelength of Ex 488 nm/ Em 593 nm (excitation/emission). Prior to the release studies, the effect of PBS buffer at pH 7.4 on cell viability was monitored. It was assumed that the relative viability of cells was not lower than 80%. The measurement of cell viability was performed in the presence of cells in PBS buffer for 3h, 6h, 9h, 24h. The measurements were carried out 24h after re-positioning the cells in the full culture medium, according to the procedure described in the cytotoxicity studies. These results determined the limitation of the experiment time, when the cells were surely in good condition, to the maximum time of 7h. Measurements of concentration after 24 hours in a system with cells were excluded due to cell viability below the minimum required (U87 MG 58±12%.; LN229 65±11%).
2.4 Elimination rate constant of anti-cancer drug in the presence of cancer cells
The doxorubicin (DOX) elimination/consumption rate constants by GBM cells (U87 MG, LN229) were determined based on the DOX depletion in the system. The mass fraction of DOX absorbed by cells was calculated as the difference between the concentration of DOX released from the multiple emulsion in a cell-free environment and the presence of cancer cells. The mass fraction of the drug available was determined by subtracting from the value 1 (the mass fraction corresponding to the total drug availability at time t=0) the mass fraction of DOX absorbed (consumed) by cells for a given time. The obtained values were presented as a function: ln(1-mass fraction of consumed DOX) vs time with a linear approximation following the kinetics of the first-order reaction. The drug elimination rate constant was then found as a slope of this function. The drug consumption rate constants and the linear function fit coefficients were determined for each glioblastoma cell line and for the tested DOX concentrations in emulsions DOX-E1 and DOX-E2, which are summarised in Table 2.
2.5 Cytotoxicity study of multiple emulsions with DOX
In vitro cytotoxicity tests were performed using the REDOX test based on Alamar Blue reagent (Invitrogen, USA) according to the manufacturers’ instructions. Cells (U87 MG, LN229) were seeded on standard 96-well plates with 400 cells/wells to maintain a proportional number of cells in relation to vessel geometry for release and cytotoxicity studies. After 24 hours from seeding, cells were exposed to a potentially cytotoxic agent (DOX). Cytotoxic effects were determined for the DOX concentrations in the emulsions at 0.1 µM and 0.2 µM. These dosages of DOX were obtained by the dilution of emulsion in a culture medium with the proportions 1:100 and 1:50, respectively. In cytotoxicity studies, the emulsion was diluted with a full culture medium, with a composition identical to that used for the cell culture. In earlier studies, the effect of changes in DOX concentration on U87MG cell viability was checked26. The cell viability was determined for cells treated with emulsion, without DOX (E1 and E2 -negative control), emulsion with encapsulated DOX (DOX-E1 and DOX-E2), and DOX solution in full culture medium. The measurements were carried out for the three different times of cancer cell contact with a potentially cytotoxic agent: 24h, 48h, 72h. After the established contact time, the culture medium was exchanged for a fresh culture medium with Alamar Blue reagent (1:10 v/v). After 24h and 168h the RFUs (relative fluorescence units) were measured at a wavelength of Ex 560 nm/Em 590 nm (excitation/emission) using a multimode detector (DTX880, Beckman Coulter, Canada). The relative cell viability was calculated as a ratio to that of the untreated control cells. All obtained values are the result of three independent experiments. Each of the parameters was determined in triplicate.
2.6 Emulsion viscosity measurements
A RheolabQC rotational viscometer with a double gap measuring cylinder (DG42) (Anton Paar, Austria) was used to study the rheological properties of the emulsions DOX-E1 and DOX-E2, and the external phases of both emulsions. To study multiple emulsions as systems responding to changes in a pH environment, emulsions and their external phases were diluted before measurement with a phosphate buffer (PBS buffer) at pH 7.4 (alkaline) and pH 6.3 (acidic). Samples were diluted in a volume dilution ratio of emulsion or external phase in PBS buffer at 1:10 (volume of tested sample to PBS buffer).
Measurements were performed for 3 independently prepared samples in the shear rate range 50-2500 s-1 during testing. Samples of emulsions for rheological tests were placed in the measuring cylinder immediately after the emulsification process. The tests were performed after establishing the temperature of the sample at 37°C. During the measurement, the emulsion structures were microscopically observed to ensure the stability of the emulsions.
2.7 Statistical analysis
All data in this study were expressed as a mean±standard error (±SD) of at least three independent experiments. Each of the parameters/values was determined in triplicate.
2.8 Mathematical model development of drug release from the emulsion-based implant and drug elimination by cancer cells