Model solution
The model equations were solved numerically using ANSYS Fluent v. 2020R2 software and the finite volume method (FVM). A computational domain was defined, considering the drops and the external fluid in which the glioblastoma cells absorb (consume) the drug (DOX) released from drops of emulsions DOX-E1 and DOX-E2. Based on the analysis of the number of drops per unit volume (drops + the external surrounding fluid), the simulation domain was determined as the relationship between the radius of the drops and the radius of the external fluid into which the drug was released. Due to the spherical geometry and axial symmetry, the computational domain was defined and discretised in cylindrical coordinates.
3. RESULTS AND DISCUSSION
3.1 The controlled anti-cancer drug release from the pH-responsive multiple emulsion and elimination by GBM cells: experiments and simulations
The experimental data and simulations of the release rates of DOX from multiple emulsions DOX-E1 and DOX-E2 in the cancer cell environment (LN229 and U87 MG) are presented in Fig. 2 (c, d, f, g) as a cumulative mass of DOX released vs time for two DOX concentrations in emulsions.
The experimental results showed that the release rates of DOX are influenced by the drop sizes and emulsion structures in the presence of the tumour environment characterised by acidic pH. The diffusional release of DOX from the emulsion with smaller drops (DOX-E2) was faster than those from the emulsion with larger drops (DOX-E1). Faster release from DOX-E2 resulted from a larger interfacial related to the smaller drop sizes of these emulsions. In addition, the internal structure of emulsion DOX-E2 (single droplet in a drop-Fig. 2b) shortened the DOX diffusion path to the drops’ interface and outside, and so thus led to a faster release compared to the emulsion DOX-E1, structured as many internal droplets in a drop (Fig. 2a). Our previous results related to the release rates of DOX from multiple emulsions in pH=6.3 simulating the biological system of cancer cells confirmed a faster release of DOX in comparison with pH=7.4 representing healthy cells. The release of DOX from multiple emulsions is based on a pH-dependent mechanism, shown in Fig. 3a.
This idea was realised by introducing pH-responsive biopolymer (sodium carboxymethylcellulose - CMC-Na) into the external phase of multiple emulsions, which bring forth spatial conformational changes influencing drug release.31 As the biopolymer (CMC-Na) is an adhesive polymer, its molecules are present not only in the external phase but also on the surfaces of drops.32 As shown in Fig. 3a, under lower pH conditions, the chains of CMC-Na are coiled and form aggregates that interact with one another weakly, facilitating drug release. At a higher or neutral pH, the polymer stretches to form long-loose chains, which interact strongly, leading to overlapping and ultimately hindering the drug release (Fig. 3 a). The spatial conformation changes in polymer chains also promote changes in some of its physicochemical properties.31 Our measurements demonstrated differences between viscosities under acidic and slightly base pH levels for emulsions and their external phases (Fig. 3 b, c). Moreover, viscosity changes affect the values of the diffusion coefficient of the drug. Due to the low concentration of Na-CMC (0.2 wt. %) the changes in the viscosity were small but significant enough to observe the difference in the diffusional transport of DOX both within and out of the drops (Table 3). The drug release rates also depend on the amount of its elimination by cancer cells. Drug elimination is determined by its physicochemical properties, formulation, type of the cells, route of administration, and rate of drug transport to the brain. In general, drugs can be eliminated including mechanisms such as degradation, metabolism, permeation, local internalisation or binding, and through blood capillaries.33 The drug (DOX) depletion due to elimination by GBM cells was determined based on the experimental data of the mass fraction of DOX released from emulsions with and without cancer cells (Fig. 2 c, d, e and f, g, h). A more general model of drug binding in the biological tissue was used to determine the elimination rate constant of DOX, according to the kinetics of an irreversible first-order reaction.9 On this basis, the elimination rate constants of DOX (k) by LN229 and U87 MG cells for the emulsions DOX-E1 and DOX-E2 were calculated and compared with those for the DOX in the solution representing classical therapy (DOX in a solution), (Tab. 2). The obtained values of (k) have demonstrated that exposing cancer cells to DOX in emulsions led to higher drug elimination, as compared to classical therapy. Higher values of the consumption rate constants of DOX in emulsion implied a greater effectiveness of the emulsion form of the drug administration, which was confirmed by cytotoxicity studies of both forms of the drug (DOX in emulsion and DOX in a solution). The experimentally obtained values of the drug elimination rate constants (k) were further used in simulations. The simulations of the drug (DOX) release in the GBM cells environment were based on a diffusion model with a chemical reaction (representing elimination/consumption of the DOX by cancer cells). Model equations (eqs. 1-13) were solved for parameters D32, d32, kLa, φ, De, De,z, k, calculated based on the experimental data (Table 3) to find the spatiotemporal drug concentration distribution and then averaged to a defined volume for simulating the cumulative mass of DOX released vs time in the presence of GBM cells. The results of the release rates simulations in comparison to experiments for both investigated emulsions and cell lines (LN229 and U87 MG cells) are presented in Fig. 2 (c, d, f, g). Simulations predicting release rates of DOX from the emulsion implant showed good agreement with experimental data. These results proved that the developed model may be a vital tool in the planning and evaluating brain oncotherapy. One of the stages is, among others, determining the dose and duration of therapy and whether and when the implant should be replaced with a new portion of the emulsion - a new implant. Then, based on the simulation results, it is possible to determine the time of the complete release of the drug and thus the time after which it is necessary to replace the inserted implant with a new portion of the emulsion, i.e. the time when the consumption of DOX by the cells ceases to increase or increases insufficiently. Moreover, by having a process model, the prediction of the drug release rates in the tumour cellular environment is feasible over a longer time scale, concerning the treatment time, compared to experimental studies. In vitro studies of the DOX release from the emulsion in the presence of cells were carried out for a maximum of 24 h due to the specific nature and limitations of tests with biological material related to the cell-culturing under conditions of anti-cancer therapy. The use of this model avoids long-term and costly experimental studies that require work with highly toxic chemotherapeutic agents and biological material. In addition, having then basic data on cell viability (from cytotoxicity study) connected with drug elimination, the effectiveness of the therapy may be easily evaluated based on the model predictions.
3.2 The modelling of the spatiotemporal drug concentration distribution within drops of the emulsion implant and external environment with glioblastoma cells
The numerical simulations also included the spatial concentration distribution of DOX within drops and the surrounding fluid with cancer cells LN229 at a given time for emulsion DOX-E1 (Fig. 4a) and DOX-E2 (Fig. 4b). As shown in exemplary simulations in Fig 4, the concentration of DOX released outside the drop, in the presence of GBM cells, achieved a relatively fast (within 50 seconds) and practically constant value within an already small distance from the surface of the implant (representative drop of emulsion) and within the bulk phase. This is the advantage of the drug emulsion-based implant, which ensures a constant concentration of the drug in the tumour environment controlled by emulsion structures and drop sizes. Simulations enable the prediction of the required drug concentration, and thus the dose of the anti-cancer drug for any cancer cells based on comparing the concentration distribution of DOX within the emulsion implant and the concentration of DOX in the cancer cells environment (external fluid). As shown in Fig. 4 (a, b), despite the same dose of DOX in both emulsions, the gradient of DOX in emulsions DOX-E2 (containing smaller drops) changes faster, due to faster diffusional release, providing smaller values of the drug concentration within and outside the drop, compared to emulsion DOX-E1 (with larger drops) at the same release time. This feature of the emulsion-based implant as a local drug delivery system may be important for planning individualised and tailored drug dosing, and also for achieving the desired therapeutic efficacy and avoiding undesired effects.
3.3 Cytotoxicity of the drug encapsulated in multiple emulsions for GBM cell lines
The in vitro cytotoxicity of the multiple emulsions with and without encapsulated DOX, and DOX in a solution (classical chemotherapy), were tested in the presence of glioblastoma multiforme (GBM) cell lines: U87 MG and LN229, after different contact times (24h, 48h, 72h) with cells (Fig. 5).
Firstly, the cytotoxicity of the emulsion without DOX (emulsions E1 and E2) was verified. Reduction in the viability of cells after 72 h contact with the emulsion was observed (for 87 MG max. to 55%, Fig. 5 a-d; for LN229 max to 45% Fig. 5 e-h), but this effect disappeared within a maximum of 7 days cell culturing after removal of the emulsion. Secondly, the cytotoxic effect of DOX introduced as an emulsion implant was compared with DOX in a solution (classical therapy) for GBM cells of both lines, and two doses of DOX (0.1 µM and 0.2 µM). At the lower dose of DOX (0.1 µM), a reduction in cell viability within 24-72 h for both routes of the drug administration was observed in the range: (i) U87 MG to 15-69% for DOX in emulsions, and to 58-77% for DOX in solution (Fig. 5 a, c), (ii) LN229 to 39-69% for DOX in emulsion, and to 54-67% for DOX in solution (Fig. 5 e, g). Whereas at the higher dose of DOX (0.2 µM), a reduction in cell viability was in the range: (i) U87 MG to 25-71% for DOX in emulsions and to 56-70% for DOX in solution (Fig. 5 b, d), (ii) LN229 to 36-60% for DOX in emulsion and to 27-54% for DOX in solution (Fig. 5 f, h). For the tested doses of DOX, a greater cytotoxic effect of DOX in emulsions (max. reduction by 85%) was observed compared to DOX in a solution (max. reduction by 43%) for 0.1 µM DOX in emulsion DOX-E1, and U87 MG cell line. These results revealed that DOX delivered as an emulsion provides a greater cytotoxic effect, and therefore increases the efficacy of the therapy compared to classical chemotherapy. Also, these results confirmed the significance of the routes of drug administration for the final effect of this therapy. Our previous studies with a wider range of DOX concentration (0.01-1 µM) for U87 MG cells also demonstrated greater effectiveness of DOX therapy based on a drug in an emulsion compared with the classical drug administration.26 In addition, DOX in emulsion-based therapy was shown to be effective for the lowest doses of DOX in the range studied, which were ineffective with the classical therapy (DOX solution). Next, the cytotoxic effect between the emulsion form of DOX (DOX-E1 and DOX-E2) was compared. In case of line U87 MG cells, cell viability after DOX administration in emulsion were (i) for DOX-E1: 15-57% for the 0.1 µM DOX dose and 25-63% for the 0.2 µM DOX dose (Fig. 5 a-b), and (ii) for DOX-E2: 50-69% for the 0.1 µM DOX dose and 46-71% for the 0.2 µM DOX dose (Fig. 5 c-d). For the LN229 cell line, cell viability after DOX administration in emulsion achieved (i) for DOX-E1: 36-46% for the 0.1 µM DOX dose and 36-49% for the 0.2 µM DOX dose (Fig. 5 e-f), and (ii) for DOX-E2: 53-69% for the 0.1 µM DOX dose and 48-60% for the 0.2 µM DOX dose (Fig. 5 g-h). Emulsion DOX-E1 demonstrated greater efficacy of reduction in viability of the tested GBM cell lines compared to DOX-E2. The DOX-E1 and DOX-E2 systems differed in the drop sizes and internal structures, as well as in the composition of the external phases, and the volume fraction of the dispersed phases. Differences between emulsions DOX-E1 and DOX-E2 affected the release rates of DOX, which contributed to the different efficacy in reducing cancer cell viability. More cytotoxic and thus effective was emulsion DOX-E1 with larger membrane phase drops with a structure of many small internal droplets, from which DOX release is slower than from emulsion DOX-E2. This proved that the drug release rate from a delivery system has a direct impact on cancer cell viability and thus on the final therapeutic effect.
4. CONCLUSIONS
This paper presents results of a promising strategy for the efficient local delivery of an anti-cancer drug (doxorubicin-DOX) to brain tumours based on an injectable three-phase liquid implant in the form of W1/O/W2 multiple emulsion. The multiple emulsions have structures of droplets in drops, therefore protecting healthy cells by encapsulating an aggressive anti-cancer drug within their internal droplets surrounded by larger drops of patient-friendly oil. The drug is gradually released by diffusion at predetermined rates from the internal droplets of the emulsions-based implant containing biopolymer (sodium carboxymethylcellulose) in response to the acidic tumour microenvironment, and is then transported to the tumour. The implant was designed to sustainably deliver therapeutics for up to 100 hours or longer, depending on the parameters of the emulsion: drop sizes, structure, viscosity, and the encapsulated drug dose. The comprehensive experimental study included drug release, in the presence and absence of glioblastoma multiforme (GBM) cell lines (LN229, U87 MG), and cell viability to evaluate the effectiveness of the proposed therapy. Also, a diffusion–reaction model has been adapted to analyse and predict doxorubicin release kinetics and drug elimination by glioblastoma cells to evaluate the proposed therapy. The model equations include parameters that take into account the structure of the emulsion (drop size and packing volume fraction), drug absorption (elimination rate constant) by cancer cells, and drug diffusion coefficients inside the emulsion drops and in the tumour environment. Drug elimination was modelled assuming first-order reaction kinetics. The numerical simulations of the drug concentration distribution in time and space were performed for the release process from emulsions DOX- E1 and DOX-E2 in the presence of glioblastoma cells (U87 MG and LN229) according to model parameters based on the experimental data. The CFD numerical simulation confirmed that the drug release process is controlled by the parameters of the emulsion structures. The obtained fractional release of the chemotherapeutic drug showed a faster release rate, from the emulsion DOX-E2, of smaller drops (higher interfacial area) compared to the emulsion DOX-E1 (bigger drops - smaller interfacial area) in the presence of the tested cancer cell lines. In addition, simulations of the spatiotemporal drug concentration distribution outside the drops of the emulsion implant confirmed a constant drug concentration, close to the implant surface and in the bulk phase, essential for the effectiveness of the therapy. Maintaining a constant concentration of the drug in the cancer cells environment confirmed the advantage of multiple emulsion as an implant delivering the drug. The validation of this model by comparison with experimental data showed good agreement under a variety of conditions (emulsions drop sizes and structures: DOX-E1 and DOX-E2, DOX concentration and types of cancer cells). The best effectiveness of the therapy was experimentally confirmed by a significant reduction in the GBM cell viability by 85% for emulsion DOX-E1, whereas for DOX in a solution (classical chemotherapy) by 43% depending on the dose of the drug. This was also confirmed by the higher drug elimination rate constants by cancer cells treated with a chemotherapeutic in a multiple emulsion compared to classical chemotherapy. Moreover, emulsion-based therapy, even with the lowest dose of DOX responded with reduced cell viability, whereas for the traditionally delivered drug this dose was ineffective. The obtained results showed some considerable promise and proved that emulsion-based implant anti-cancer drugs delivery might succeed in the unequal fight against brain tumours. The adapted diffusion-reaction model has several potential uses, especially in early pre-clinical development in the planning of optimal oncotherapy, including drug dose and release kinetics and treatment duration. This model, supported by experimental data, also provides insights into the relevant mechanisms and parameters, which quantitatively describe the complex processes accomplished by drug release in a biological system.
ACKNOWLEDGMENT
Research was funded by (POB Biotechnology and Biomedical Engineering) of Warsaw University of Technology within the Excellence Initiative: Research University (IDUB) programme (project BIOTECHMED-2) and by the National Science Centre – Poland (grant number: 2014/13/B/ST8/04274).
REFERENCES
  1. Lu CT, Zhao YZ, Wong HL, Cai J, Peng L, Tian XQ. Current approaches to enhance CNS delivery of drugs across the brain barriers. Int J Nanomedicine. 2014;9:2241-2257.
  2. Tan AC, Ashley DM, López GY, Malinzak M, Friedman HS, Khasraw M. Management of Glioblastoma: State of the Art and Future Directions. CA Cancer J Clin. 2020;70:299–312.
  3. Kasinathan N, Jagani HV, Alex AT, Volety SM, Rao JV. Strategies for drug delivery to the central nervous system by systemic route. Drug Deliv. 2015;22:243-257.
  4. Chakroun RW, Zhang P, Lin R, Schiapparelli P, Quinones‐Hinojosa A, Cui H. Nanotherapeutic systems for local treatment of brain tumors. WIREs Nanomed Nanobiotechnol. 2018;10(1):e1479.
  5. Haumann R, Videira JC, Kaspers GJL, van Vuurden DG, Hulleman E. Overview of Current Drug Delivery Methods Across the Blood–Brain Barrier for the Treatment of Primary Brain Tumors. CNS Drugs. 2020;34:1121–1131.
  6. Mueller E, Himbert S, Simpson MJ, Bleuel M, Rheinstadter MC, Hoare T. Cationic, Anionic, and Amphoteric Dual pH/Temperature-Responsive Degradable Microgels via Self-Assembly of Functionalized Oligomeric Precursor Polymers. Macromolecules. 2021;54(1):351-363.
  7. Gerlee P. The model muddle: in search of tumor growth laws. Cancer Res. 2013;73(8):2407-2411.
  8. Preziosi L. Cancer modelling and simulation (1st edition). Chapman & Hall, CRC, 2003.
  9. Weiser JR, Saltzman WM. Controlled release for local delivery of drugs: barriers and models. J Control Release. 2014;190:664-673.
  10. Siepmann J, Siepmann F. Modeling of diffusion controlled drug delivery. J Control Release. 2012;161(2):351-362.
  11. Kim M, Gillies RJ, Rejniak KA. Current advances in mathematical modeling of anti-cancer drug penetration into tumor tissues. FRONT ONCOL. 2013;3:278.
  12. Montigny JD, Iosif A, Breitwieser L, Manca M, Bauer R, Vavourakis V. An in silico hybrid continuum-/agent-based procedure to modelling cancer development: interrogating the interplay amongst glioma invasion, vascularity and necrosis. Methods. 2020;185(4):94-104.
  13. Chen J, Weihs D, Vermolen FJ. A Cellular Automata Model of Oncolytic Virotherapy in Pancreatic Cancer. Bull Math Biol. 2020;82(103):1-25.
  14. Metzcar J, Wang Y, Heiland R, Macklin P. A Review of Cell-Based Computational Modeling in Cancer Biology. JCO Clin. Cancer Inform. 2019;3(3):1-13.
  15. Trucu D, Lin P, Chaplain MA, Wang Y. A multiscale moving boundary model arising in cancer invasion. Multiscale Model. Simul. 2013;11(1):309-335.
  16. Ramis-Conde I, Chaplain MAJ, Anderson ARA. Mathematical modelling of cancer cell invasion of tissue. Math. Comput. Modelling. 2008;47:533–545.
  17. Dluska E, Hubacz R, Wronski S, Kamienski J, Dylag M, Wojtowicz R. The influence of helical flow on water fuel emulsion preparation. Chem. Eng. Commun. 2007;194(10): 1271–1286.
  18. Dluska E, Markowska-Radomska A, Metera A, Kosicki K. Hierarchically structured emulsions for brain therapy. Colloid Surf. A Physicochem. Eng. Asp. 2019;575:205-211.
  19. Eugenia Carlotti M, Gallarate M, Sapino S, Ugazio E, Morel S. W/O/W Multiple Emulsions for Dermatological and Cosmetic Use, Obtained with Ethylene Oxide Free Emulsifiers. J Dispers Sci Technol. 2005;26(2):183-192.
  20. Loya-Castro MF, Sánchez-Mejía M, Sánchez-Ramírez DR, Domínguez-Ríos R, Escareño N, Oceguera-Basurto PE, Figueroa-Ochoa ÉB, Quintero A, del Toro-Arreola A, Topete A, Daneri-Navarro A. Preparation of PLGA/Rose Bengal colloidal particles by double emulsion and layer-by-layer for breast cancer treatment. J. Colloid Interface Sci. 2018;518:122-129.
  21. Dluska E, Metera A, Markowska-Radomska A, Tudek B. Effective cryopreservation and recovery of living cells encapsulated in multiple emulsions. Biopreserv Biobank. 2019;17(5):468-476.
  22. Mutaliyeva B, Grigoriev D, Madybekova G, Sharipova A, Aidarova S, Saparbekova A, Millere R. Microencapsulation of insulin and its release using w/o/w double emulsion method. Colloids Surf. A Physicochem. Eng. Asp. 2017;521:147–152.
  23. Bozkir A, Hayta G. Preparation and evaluation of multiple emulsions water-in-oil-in-water (w/o/w) as delivery system for influenza virus antigens. J. Drug. Target. 2004;12:157-164.
  24. Soriano-Ruiz JL, Suner-Carbo J, Calpena-Capmany AC, Bozal-de Febrer N, Halbaut-Bellowa L, Boix-Montanes A, Souto EB, Clares-Naveros B. Clotrimazole multiple W/O/W emulsion as anticandidal agent: Characterization and evaluation on skin and mucosae. Colloids Surf. B Biointerfaces. 2018;175:166-174.
  25. McClements DJ, Decker EA, Weiss J. Emulsion-based delivery systems for lipophilic bioactive components. J Food Sci. 2007;72(8):109-124.
  26. Dluska E, Markowska-Radomska A, Metera A, Tudek B, Kosicki K. Multiple emulsions as effective platforms for controlled anti-cancer drug delivery. Nanomedicine. 2017;12(18):2183-2197.
  27. Dluska E, Cui Z, Markowska-Radomska A, Metera A, Kosicki K. Cryoprotection and banking of living cells in a 3D multiple emulsion-based carrier, Biotechnology Journal. 2017;12(8):1–7.
  28. Dluska E, Markowska-Radomska A. Regimes of multiple emulsions of W1/O/W2 and O1/W/O2 type in the continuous Couette-Taylor flow contactor. Chem. Eng. Technol. 2010;33(1):113–120.
  29. Markowska-Radomska A, Dluska E. An evaluation of a mass transfer rate at the boundary of different release mechanisms in complex liquid dispersion. Chem Eng Process. Process Intensification. 2016;101:56-71.
  30. Markowska-Radomska A, Dluska E. The Multiple Emulsion Entrapping Active Agent Produced via One-Step Preparation Method in the Liquid-Liquid Helical Flow for Drug Release Study and Modeling. Progr Colloid Polymer Sci. 2012;139:29-34.
  31. Khaled B, Abdelbaki B. Rheological and electrokinetic properties of carboxymethylcellulose-water dispersions in the presence of salts. Int. J. Phys. Sci. 2012;7(11):1790-1798.
  32. Friedman D, Schwartz J, Amselem S. Bioadhesive emulsion preparations for enhance drug delivery. US Patent Number: USOO5744155A, 1998.
  33. Weiser JR, Saltzman WM. Controlled release for local delivery of drugs: barriers and models. J Control Release. 2014;190:664-673.
  34. Li J, Carr PW. Accuracy of empirical correlations for estimating diffusion coefficients in aqueous organic mixtures. Anal. Chem. 1997;69:2530-2536.
  35. Bartzatt R. Identification of doxorubicin and an imine derivative from liquid and solid samples utilizing liquid chromatography. J Liq Chromatogr Relat Technol. 2006;29:2303-2312.
  36. Yan N. A mass transfer model for type-II-facilitated transport in liquid membranes. Chem. Eng. Sci. 1993;48:3835-3843.
  37. Lian G, Malone ME, Homan JE, Norton IT. A mathematical model of volatile release in the mouth from the dispersion of gelled emulsion particles. J. Control. Release. 2004;98:139–155.
  38. Happel J. Viscosity of suspensions of uniform spheres. J. Appl. Phys. 1957;28:1288-1292.