This work describes the development of flexible interdigitated electrodes for implementation in 3D brain tumor models, namely, in vitro culture, in ovo model, and in vivo murine model. The proposed method can be used to evaluate the effects of pulsed electric fields on tumors at different levels of complexity.
Glioblastoma is difficult to eradicate with standard oncology therapies due to its high degree of invasiveness. Bioelectric treatments based on pulsed electric fields (PEFs) are promising for the improvement of treatment efficiency. However, they rely on rigid electrodes that cause acute and chronic damage, especially in soft tissues such as the brain. In this work, flexible electronics were used to deliver PEFs to tumors and the biological response was evaluated with fluorescent microscopy. Interdigitated gold electrodes on a thin, transparent parylene-C substrate were coated with the conducting polymer PEDOT:PSS, resulting in a conformable and biocompatible device. The effects of PEFs on tumors and their microenvironment were examined using various biological models. First, monolayers of glioblastoma cells were cultured on top of the electrodes to investigate phenomena in vitro. As an intermediate step, an in ovo model was developed where engineered tumor spheroids were grafted in the embryonic membrane of a quail. Due to the absence of an immune system, this led to highly vascularized tumors. At this early stage of development, embryos have no immune system, and tumors are not recognized as foreign bodies. Thus, they can develop fast while developing their own vessels from the existing embryo vascular system, which represents a valuable 3D cancer model. Finally, flexible electrode delivery of PEFs was evaluated in a complete organism with a functional immune system, using a syngenic, orthograft (intracranial) mouse model. Tumor spheroids were grafted into the brain of transgenic multi-fluorescent mice prior to the implantation of flexible organic electrode devices. A sealed cranial window enabled multiphoton imaging of the tumor and its microenvironment during treatment with PEFs over a period of several weeks.
Glioblastoma multiforme (GBM) is a highly invasive tumor and therefore difficult to eradicate with standard treatments such as resection, radiotherapy, and chemotherapy. Despite multimodal treatments, prognosis remains very poor and most of the patients experience disease progression within 1 year of diagnosis1,2. Recently, the development of bioelectric treatments has shown great potential to improve existing therapies. These therapies use the delivery of pulsed electric fields (PEF), typically in a single treatment session, to disrupt cellular membrane integrity and the microenvironment of tumors. This cell membrane disruption, also known as electroporation, can be reversible or irreversible depending on the electric field intensity and the number of pulses. Irreversible electroporation (IRE) is applied as a non-thermal tissue ablation technique in which electric pulses cause fatal damage to cellular membranes leading to cell death3. Reversible electroporation is applied in electrochemotherapy (ECT), an established technique that consists of the delivery of PEFs in combination with chemotherapy drugs to enhance drug uptake in cancer cells4. Moreover, recent studies demonstrated calcium electroporation as an alternative to ECT with high efficiency for cancer treatment, which is also inexpensive and induces fewer side effects5. Despite these promising advances, PEFs are generally applied using rigid, metallic electrodes which are known to cause damage to soft tissue6. The brain is particularly sensitive to such invasive devices where the mechanical mismatch induces inflammation and astroglial scarring7.
In this context, a flexible PEF delivery system in combination with 3D models of glioblastoma tumors is presented, from microfabrication to a murine model. Conformal electrodes are made with standard thin-film microfabrication processes, including the use of soft and biocompatible materials such as parylene-C, gold, and PEDOT:PSS8,9. An interdigitated electrode design is used to cover a large surface area while maintaining adequate transparency for imaging between the electrode fingers10. For the tumor model, 3D spheroids of glioblastoma cells expressing a genetically encoded fluorescence reporter are produced using a variation of the liquid-overlay 96-well plate method11. The spheroids are grafted into the chorioallantoic membrane of a quail embryo, resulting in an in ovo model that has been extensively used to study angiogenesis or drug toxicology12,13. Tumors can be grafted and vascularized by the embryo's vasculature in the absence of an immune system at this stage of embryonic development12. Flexible electrodes are then placed on top of the vascularized tumor to study the effect of PEF delivery on the spheroid and its vasculature. Finally, these effects are investigated on a complete living organism, including tumor microenvironment and immune system, by implanting engineered spheroids into the brain parenchyma of murine models14. Flexible electrodes are placed on top of the insertion site and the craniotomy is sealed with a glass window, allowing repeated two-photon imaging over several weeks.
These methods will be useful for people interested in various domains ranging from microelectronics engineering to oncology applications. The microfabrication protocol can be used and adapted for any application requiring thin-film metal electrodes coated with PEDOT:PSS. Further, the biological models developed for the evaluation of antitumor electrical treatments will be of general interest for the investigation of the differentiation of cellular, vascular, and immune response to implanted materials.
All experimental procedures were performed in accordance with the French legislation and in compliance with the European Community Council Directive of November 24, 1986 (86/609/EEC) for the care and use of laboratory animals. The research on animals was authorized by the Direction Départementale des Services Vétérinaires des Bouches-du-Rhône and approved by the ethical committee of Provence Cote D'Azur (Apafis # 22689-2019100414103054).
1. Flexible device microfabrication (Figure 1)
2. Generation of Glioblastoma GCaMP6f stable cell line
3. 3D models
4. Pulsed Electric Field (PEF) delivery and imaging
This protocol allows the application to two glioblastoma models in which a flexible PEF delivery system is integrated. Following microfabrication and packaging steps, flexible electrodes are characterized in saline solution by electrochemical impedance spectroscopy (EIS) to assess and validate their performance. The PEDOT:PSS-coated electrodes show the typical capacitive and resistive dominated regions separated by a cut-off frequency, whereas the uncoated electrodes display only capacitive behavior (Figure 2).
A variation of the liquid-overlay 96-well plate method is used to grow 3D tumors made of transfected glioblastoma cells stably expressing a fluorescent intracellular calcium reporter. The growth of the spheroids can be observed with a bright-field microscope (Figure 3; ED 0). At least 2 or 3 days are needed to obtain spherical and dense spheroids, depending on the cell line and the number of cells seeded.
In the in ovo model, spheroids are grafted in the chorioallantoic membrane of a quail embryo (Figure 3; ED 6). The success of the graft can be assessed by fluorescence microscopy a few days later, as living cells have intracellular calcium, and are thus fluorescent (Figure 3; ED 9). The vascularization of the tumor can be observed under a fluorescence microscope by injecting a fluorescent dye into the blood vessels (Figure 3; ED9). However, it might not always be possible to visualize the blood vessels inside the tumor as the spheroid is very dense. The flexible interdigitated electrodes are placed on top of the vascularized tumor (Figure 3; ED 9) and connected to a pulse generator. The probe must be gently placed to avoid bleeding of the embryo; otherwise, the fluorescent dye can spread, which obstructs any observation by imaging. Correct delivery of the pulse to the biological environment can be verified by measuring the current going through the circuit. Imaging of these in ovo models allows real-time monitoring of the effect of PEFs on the intracellular calcium in a 3D glioblastoma tumor, as well as the vasoconstriction induced on the tumor's vasculature, avoiding any influence of other cell types including the immune system15.
The study of the PEF effect on glioblastoma can also be performed in a more complete and predictive model. Indeed, the in vivo model described above14 consists of grafting a 3D glioblastoma tumor in the brain parenchyma of a mouse (Figure 4). The injection site of the tumor is plugged by a cross-linked dextran gel hemi-bead, to recapitulate the physiological biophysical constraints during the growth of the tumor. Although described in reference14, it is worth re-emphasizing that it is critically important that the dextran hemi-bead be precisely superglued to the dura mater; otherwise, the tumor can escape through the open dura and completely cover the brain, making the imaging impossible. For any chronic imaging, tissue ingrowth that takes place as the cranial window heals poses a serious barrier, as the new tissue is non-transparent and makes images foggy or non-usable. Therefore, after inserting and gluing the hemi-bead, the sidewalls of the opened cranial window need to be sealed with a thin layer of superglue meticulously placed all around the cavity wall, without letting the superglue slip or flow onto the dura. When the flexible probe is placed on top of the tumor injection site, no bubbles can stay under the probe, for two reasons. Firstly, imaging cannot proceed when bubbles are present. Secondly, bubbles serve as insulators, thus changing the electrical stimulation properties. After taking the precautions described above, the craniotomy is sealed with a glass window cemented to the skull to allow chronic imaging over weeks. As the tumor consists of GCaMP or DsRed expressing cells, the injection can be confirmed with a fluorescence microscope. The electrochemical impedance of the electrodes must be measured to validate the performance after implantation. Compared to the impedance in saline solution, an increase of the impedance is expected in vivo at frequencies above 100 Hz due to the presence of a biological environment (Figure 5). Vascularized neural parenchyma and tumor infiltration can be observed and characterized through the transparent substrate over weeks by two-photon microscopy (Figure 6). The use of transgenic animals expressing fluorescent proteins in cells of interest (immune cells and neurons) can, for example, allow demonstration of the minimal inflammatory process induced by electrode implantation alone (Figure 6A) or show the presence of microglia and monocytes 26 days after implantation of a PEF stimulated electrode implanted on top of a growing GBM tumor (Figure 6B1). In the latter case, both peripheral-monocyte-derived cells and brain-resident microglial cells were found around and inside the tumor (Figure 6B2). On the day of PEF delivery, contact pads of the flexible electrodes can be connected to the pulse generator, directly under the two-photon microscope. Overall, this model can be used to investigate the effect of bioelectric treatments over time using various types of cells involved in brain tumor development, up to a depth of around 500 µm.
Figure 1: Microfabrication of flexible electrodes. (A) Gold electrode patterning and Parylene C substrate. (B) Outline opening. (C) PEDOT:PSS coating. (D) Connections and packaging. Please click here to view a larger version of this figure.
Figure 2: Electrochemical impedance spectroscopy of flexible gold electrodes and PEDOT:PSS coated cold electrodes in a saline solution. Please click here to view a larger version of this figure.
Figure 3: The in ovo model of glioblastoma. ED 0: Spheroid observed with a bright-field microscope. ED 3: Shell less culture of a quail embryo 3 days after opening. ED 6: Tumor implanted in the CAM observed with a bright-field microscope. ED 9: Flexible device placed on the vascularized tumor (tumor in green and blood vessels in red). Please click here to view a larger version of this figure.
Figure 4: The in vivo application. (A) Scheme for in vivo experiments. (B) Probe placement before the application of cover glass and acrylic resin. (C) Completed probe implantation. Please click here to view a larger version of this figure.
Figure 5: Electrochemical impedance spectroscopy of flexible gold electrodes in a saline solution compared to an implanted probe. Please click here to view a larger version of this figure.
Figure 6: Intravital multispectral two-photon imaging through electrodes. (A) Tiled image of the healthy brain surface in a control multifluorescent AMU-Neuroinflam mouse 3 days after electrode implantation. Cyan shows dendritic arborization of layer 5 pyramidal neurons, green shows recruited granulocytes and monocytes, and yellow shows activated microglia and dendritic cells. Pink shows infrared diffusion due to heat accumulation. (B1) Similar image as in A but 26 days after tumor spheroid implantation 200 µm deep inside the cortex immediately followed by electrode implantation. Note the accumulation of green and yellow immune cells. (B2) Similar image as in B1 but 100 µm below the surface of the electrodes. Note the presence of blue neuronal dendritic arborization in the periphery of the red tumor mass itself infiltrated by yellow microglia and dendritic cells. Deep blue shows a second harmonic signal from the peritumoral collagen. (B3) Zoomed view of B2 showing the presence of interneuron somas (indicated by arrows) in the vicinity of the tumor. Please click here to view a larger version of this figure.
The approach described in this work enables brain tumor models with an integrated PEF delivery system to study the effect of PEFs at different levels of biological organization. The microfabrication protocol consists of standard thin-film processes, which provide a large degree of freedom in electrode design that can be adapted to the specific application. Sometimes, an additional thermal annealing step can be useful at the end of the fabrication, to reduce bending of the electrodes that occurred during manufacturing.
The use of a stable glioblastoma cell line expressing a fluorescent calcium indicator avoids all complications linked with dye delivery and retention, especially in 3D tumors that are very dense16. Indeed, a high expression level is observed over a long period compared to standard chemical fluorescent calcium indicators17. This protocol can be applied to various cell lines, as it is commonly used for imaging neural activity11. Here, human and murine cell lines were used (U87 and Gl261 for implantation in immunodeficient or immunocompetent mice, respectively). Indeed, recent studies showed that the U87 cell line is different from that of the original cells as many mutations were acquired over years of cell culture, affecting experimental reproducibiliy18. The method used for the preparation of 3D tumors is high-throughput, reproducible, and allows the generation of spheroids of a specific size depending on the cell line, the number of cells at seeding, and the time of growth19. However, these spheroids are dense, which presents a disadvantage when imaging at the core of the tumor.
The in ovo model is useful as a first approach to study the effect of PEF on 3D tumors and their vasculature, without interactions with other cell types that are present in the brain. This model is inexpensive, fast, high-throughput, and raises fewer ethical issues than animal models. It is important to maintain the integrity of the embryo throughout the entire experiment, as it could affect its survival and the quality of the imaging. Special care must be taken while opening the quail egg, to avoid damages to the embryonic membrane. The graft and the placement of the flexible electrodes must also be performed carefully, to avoid bleeding that could kill the embryo. Injection of fluorescent dye in the blood vessels allows simultaneous visualization of the tumor cells and vascularization with fluorescence microscopy. The intraocular injection must be performed carefully to avoid dye leaking into the embryonic liquid, which could cause a residual fluorescence in the background that degrades the quality of the imaging. This model can also be used for following drug uptake, as it allows access to the circulatory system. However, experiments are limited by the 12-day survival time of the embryo, thus allowing 7 days of observation, which is significantly shorter than the in vivo model21.
The in vivo brain tumor model can be monitored for 4 to 5 weeks before animals reach an ethical experimental endpoint determined by a sudden 20% weight loss. It is well tolerated and remains in place if the connecting tail of the electrode is not too long. Otherwise, animals tend to scratch the flipping connector, which might ultimately be torn, hence preventing subsequent connection to the stimulator. This 4-week period is nevertheless valuable to cover the different stages of glioblastoma development. When comparing the tumor cell densities in the same volume of interest at different time intervals, the evolution of the tumor growth kinetics can be observed. In particular, enhanced tumor growth was observed at the time of the immune switch22. A similar study in the presence of a stimulating electrode would inform on the effect of PEF on tumor proliferation rate and tumor sensitivity to immune elimination. In comparison to in ovo model, the in vivo model can be seen as a valuable preclinical model to study the impact of immune cells on tumor progression and their contribution to the therapeutic effect of PEF. This protocol is adapted from a previous article with the addition of a flexible electrode device on the tumor before placing a cranial window14. Both the acute and chronic bioelectric treatments of tumors can be characterized by direct and subsequent observations with two-photon microscopy given that initial stimulation is expected to induce cell death and to trigger lasting dysregulation of the immune response.
The connections of the flexible probe are easily accessible under the two-photon microscope. Electrical stimulation parameters can thus be adjusted in real-time based on the observed effect on the neural tissue and/or the targeted cells, similar to how a medical doctor would perform interventional procedures while observing MRI or CT images of his patient. A final consideration is the importance of careful sealing of the electrode on the brain with superglue and silicone glue to prevent tissue regrowth.
In conclusion, the protocol described here represents an innovative model to study the effect of PEF therapy with flexible organic polymer electrodes for glioblastoma tumor models. The two models exhibit different levels of complexity such that cellular, vascular, or immune effects can be separated for a better understanding of the mechanisms of action. Conformal, superficial electrodes reduce the iatrogenic damage while enabling disruption of the tumor microenvironment, triggering vasoconstriction or dysregulation of intracellular calcium15.
The authors have nothing to disclose.
The work reported here was supported by the French National Research Agency (ANR-18-CE19-0029). The authors warmly thank S.M. Bardet for her contribution to the generation of a stable GCaMP6f cell line and D. O'Connor for her help with the in ovo model.
(3-Glycidyloxypropyl)trimethoxysilane | Sigma | 440167 | GOPS |
0.25% Trypsin-EDTA (1X) | Gibco | 25200-056 | |
4-Dodecylbenzenesulfonic acid | Sigma | 44198 | DBSA |
96-well plate | Falcon | 353075 | |
Acetone | Technic | 530 | |
Acrylic resin | Fischer scientific | NC1455685 | |
agarose | Sigma | A9539 | |
autoclave | Tuttnauer | 3150 EL | |
AZ 10XT | Microchemicals | Positive photoresist | |
AZ 826 MIF Developer | Merck | 10056124960 | Metal-ion-free developer for the negative photoresist |
AZ Developer | Merck | 10054224960 | Metal-ion-free developer for the positive photoresist |
AZ nLof 2070 | Microchemicals | Negative photoresist | |
Buprenorphine | Axience | ||
Carprofen | Rimadyl | ||
Centrifuge Sorvall Legend X1R | Thermo Scientific | 75004260 | |
CMOS camera Prime 95B | Photometrics | ||
CO2 incubator HERAcell 150i | Thermo scientific | ||
DAC board | National Instruments | USB 6259 | |
Déco spray Pébéo | Cultura | 3167860937307 | Black acrylic paint |
Dextran Texas Red 70.000 | Thermofisher | D1830 | |
Die bonding paste "Epinal" | Hitachi | EN-4900GC | Silver paste |
Dimethyl sulfoxide | Sigma | D2438 | |
Dispensing machine | Tianhao | TH-2004C | |
Dulbecco’s Modified Eagle’s Medium + GlutaMAX™-I | Gibco | 10567-014 | |
Dulbecco's Modified Eagle's Medium | Sigma | D6429 | |
Egg incubator COUVAD'OR 160 | lafermedemanon.com | ||
Ethylene glycol | Carl Roth | 6881.1 | |
Fertilized eggs of Japanese quail | Japocaille | ||
Fetal Bovine Serum | VWR | S181BH | |
Flask | Greiner | 658170 | |
Fluorescence macroscope | Leica MZFLIII | ||
Gl261 | DSMZ | ACC 802 | |
Gold pellets – Dia 3 mm x 6 mm th | Neyco | ||
Handheld automated cell counter | Millipore | PHCC00000 | |
Heating and drying oven | Memmert | UF110 | |
Hexadimethrine Bromide Sequa-brene | Sigma | S2667 | |
hot plate Delta 6 HP 350 | Süss Microtec | ||
Illumination system pE-4000 | CoolLed | ||
Infrared tunable femtosecond laser (Maï-Taï) | Spectra Physics (USA) | ||
Ionomycin calcium salt | Sigma | I3909 | |
Kapton tape SCOTCH 92 33×19 | 3M | Polyimide protection tape | |
Lab made pulse generator | |||
Labcoter 2 Parylene Deposition system PDS 2010 | SCS | ||
Lenti-X 293 T cell line | Takara Bio | 63218 | HEK 293T-derived cell line optimized for lentivirus production |
Lenti-X GoStix Plus | Takara Bio | 631280 | Quantitative lentiviral titer test |
Mask aligner MJB4 | Süss Microtec | ||
Micro-90 Concentrated cleaning solution | International Products | M9050-12 | |
Microscope slides 76 x 52 x 1 mm | Marienfeld | 1100420 | |
Needles 30G | BD Microlance 3 | 304000 | |
PalmSens4 potentiostat | PalmSens | ||
parylene-c : dichloro-p-cyclophane | SCS | 300073 | |
PCB Processing Tanks | Mega Electronics | PA104 | |
PEDOT:PSS Clevios PH 1000 | Heraeus | ||
penicillin / streptomycin | Gibco | 15140-122 | |
Petri dish | Falcon | 351029 | |
pGP-CMV-GCaMP6f | Addgene | 40755 | plasmid |
Phosphate Buffer Saline solution | Thermofisher | D8537 | |
Plasma treatment system PE-100 | Plasma Etch | ||
PlasmaLab 80 Reactive Ion Etcher | Oxford Instruments | ||
Plastic mask | Selba | ||
Plastic weigh boat 64 x 51 x 19 mm | VWR | 10770-454 | |
Poly-dimethylsiloxane: SYLGARD 184 Silicone Elastomer Kit | Dow chemicals | 1673921 | |
Polyimide copper film 60 µm (Kapton) | Goodfellow | IM301522 | |
Propan-2-ol | Technic | 574 | |
Protolaser S | LPKF | ||
puromycin | Gibco | A11103 | |
Round cover glass 5 mm diameter | Fischer scientific | 50-949-439 | |
Scepter Sensors – 60 µm | Millipore | PHCC60050 | |
Silicone adhesive Kwik-Sil | World Precision Instruments | ||
spin coater | Süss Microtec | ||
Spin Coater | Laurell | WS-650 | |
Super glue | Office depot | ||
tetracycline-free fœtal bovine Serum | Takara Bio | 631105 | |
Thermal evaporator Auto 500 | Boc Edwards | ||
Two-photon microscope | Zeiss LSM 7MP | ||
U87-MG | ATCC | HTB-14 | Human glioblastoma cells |
Ultrasonic cleaner | VWR | ||
Vortex VTX-3000L | LMS | VTX100323410 | |
Xfect single shots reagent | Takara Bio | 631447 | Transfection reagent |