Electrophoretic Delivery of γ-aminobutyric Acid (GABA) into Epileptic Focus Prevents Seizures in Mice

Summary

The challenge of epilepsy research is to develop novel treatments for patients where classical therapy is inadequate. Using a new protocol—with the help of an implantable drug delivery system—we are able to control seizures in anesthetized mice by the electrophoretic delivery of GABA into the epileptic focus.

Abstract

Epilepsy is a group of neurological disorders which affects millions of people worldwide. Although treatment with medication is helpful in 70% of the cases, serious side effects affect the quality of life of patients. Moreover, a high percentage of epileptic patients are drug resistant; in their case, neurosurgery or neurostimulation are necessary. Therefore, the major goal of epilepsy research is to discover new therapies which are either capable of curing epilepsy without side effects or preventing recurrent seizures in drug-resistant patients. Neuroengineering provides new approaches by using novel strategies and technologies to find better solutions to cure epileptic patients at risk.

As a demonstration of a novel experimental protocol in an acute mouse model of epilepsy, a direct in situ electrophoretic drug delivery system is used. Namely, a neural probe incorporating a microfluidic ion pump (µFIP) for on-demand drug delivery and simultaneous recording of local neural activity is implanted and demonstrated to be capable of controlling 4-aminopyridine-induced (4AP-induced) seizure-like event (SLE) activity. The γ-aminobutyric acid (GABA) concentration is kept in the physiological range by the precise control of GABA delivery to reach an antiepileptic effect in the seizure focus but not to cause overinhibition-induced rebound bursts. The method allows both the detection of pathological activity and intervention to stop seizures by delivering inhibitory neurotransmitters directly to the epileptic focus with precise spatiotemporal control.

As a result of the developments to the experimental method, SLEs can be induced in a highly localized manner that allows seizure control by the precisely tuned GABA delivery at the seizure onset.

Introduction

Epilepsy is the fourth most common neurological disorder: about 1% of the population suffers from epilepsy, and about one-third of the affected have recurrent seizures. In most cases, seizures can be controlled with medication. However, drug treatment needs to be set for every patient individually, where proper dosing can take years to find^1,2. Additionally, most of the medication has serious side effects that reduce the quality of life^3,4,5,6,7. Finally, in 30% of the cases patients are resistant to medication, and in case of a constant single seizure generator locus, only resective neurosurgery can attenuate the occurrence of seizures⁸. Therefore, a major initiative in modern epilepsy research is to discover new strategies which can prevent recurrent seizures in patients at risk, while reducing the necessity of strong drug therapies and invasive resective surgeries.

Epileptic seizures occur when there is an imbalance within excitatory and inhibitory circuits either throughout the brain (generalized epilepsy) or in a localized part of the brain (focal epilepsy), such that neurons discharge in an abnormal fashion^9,10,11. Antiepileptic drugs can act in two different ways in seizure prevention: either decreasing excitation or enhancing inhibition¹². Specifically, they can either modify the electrical activity of neuronal cells by affecting ion channels in the cell membrane¹³ or act on chemical transmission between neurons by affecting the inhibitory neurotransmitter GABA or the excitatory glutamate in the synapses^14,15. For some medications, the mode of action is unknown¹⁸. Also, drug treatments have a continuous effect on patients and cannot adapt to the prevalence dynamics of seizures. Ideally, drugs with specific mechanisms of action would act on the underlying epileptic processes. An optimal treatment would not touch the brain interictally but would act immediately when a seizure starts developing. In contrast to that, in all cases of epilepsy, medication now means a systematic treatment, affecting the whole brain and the whole body of the patient⁹.

Epileptic seizures can appear many years after the initial insult such as brain trauma. The period between the initial insult and the occurrence of the first spontaneous seizures is characterized by considerable molecular and cellular reorganizations, including neuronal death with the disappearance of neuronal network connections and axonal sprouting/neosynaptogenesis with the appearance of new connections^19,20,21. Once seizures become recurrent, their frequency and severity tend to increase, involving more brain regions. It is important to distinguish the sites of seizure onset (epileptogenic regions) from propagation networks, as the rules of seizure genesis and propagation may differ. Research performed on human tissue and experimental models of epilepsy have provided important data regarding the reorganization of circuits and their ability to generate seizures^20,21,22,23. However, it is difficult to determine if these reorganizations are adaptive responses or whether they are causally related to epileptogenesis or seizure genesis and propagation¹².

Therefore, localizing the epileptic focus and applying antiepileptic drugs locally are one of the main challenges in contemporary epilepsy research. Several experiments using animal models of epilepsy and some clinical studies aimed to find the onset of the seizure events and define the underlying mechanisms in the brain^24,25,26,27. To this end, we developed a new experimental protocol using the 4AP-induced epilepsy model^28,29,30,31 in an acute mouse preparation, which allows the precise insertion of three devices into the given area of the hippocampus, where network activity in vivo is manipulated in a highly localized manner. Localized 4AP injection by a glass micropipette helps to induce epileptic SLEs in a localized spot in the hippocampus, while with the help of the novel polymer-based µFIP probe the control of the seizure activity is achieved simultaneously by recording the neuronal electrical activity with the device’s recording sites. Hippocampal local field activity is also monitored with a multichannel silicon probe in a layer-specific manner in the cortex and in the hippocampus simultaneously.

The recently invented µFIP probes work by using an applied electric field to push charged drugs stored in a microfluidic channel across an ion exchange membrane (IEM) and out to the surrounding tissue (Figure 1). The IEM selectively transports only one type of ion (cation or anion) and, thus, works to limit both passive diffusion in the “off” state and  transport of oppositely charged species from the surrounding tissue into the device. The electric field is created on demand by applying a small voltage (<1 V) between the source electrode which is internal to the microfluidic channel and a target electrode which is external to the device (in this case, the head screw on the animal model). The rate of drug delivery is proportional to the applied voltage and the measured current between the source and target electrodes. The precise tunability of drug delivery is one of the primary advantages of the µFIP. Another critical advantage, compared to fluidic or pressure-based drug delivery systems, is that in the µFIP there is only a negligible pressure increase at the drug delivery outlet as drugs are delivered across the IEM without their carrier solution.

There is a small amount of passive leaking of GABA when the µFIP is “off”, but this was found not to effect SLEs. The µFIP are custom-made following conventional microfabrication methods that we reported previously³¹.

Since one way of preventing recurrent seizures is the blockade of network discharges at the very beginning or even before the first seizure event, the presented method for delivering the inhibitory neurotransmitter GABA into the epileptic focus has great therapeutic potential for seizure control in patients with focal epilepsy. Since GABA is an endogenous substrate, it leaves intrinsic neuronal properties unchanged in physiological concentrations. The local application of low levels of GABA will only affect cells naturally responsive to inhibition, and will only cause similar effects to physiological inhibition, contrary to deep brain stimulation (DBS), which has unspecific actions by stimulating all cells of the neuronal network in its environment, causing a mixed response involving both excitation and inhibition. In conclusion, the proposed method provides a more specific approach to seizure control than DBS.

Protocol

All experimental procedures were performed according to the ethical guidelines of the Institut de Neurosciences des Systèmes and approved by the local Ethical Committees and Veterinary Offices.

NOTE: Seventeen adult male OF1 mice were used for the experiments. Mice were entrained to a 12 h light/dark cycle with food and water available ad libitum.

1. Anesthesia

1. Inject intraperitoneally a mixture of ketamine and xylazine (100 mg/kg body weight and 10 mg/kg body weight, respectively) to anesthetize the animal.
2. Check the level of anesthesia by observing the respiratory rate and whisking and by checking the mouse’s response to pain.
   NOTE: When the mouse’s breathing becomes regular, no whisking can be observed, and the animal does not react to tail pinches, the anesthesia is deep enough to continue.
   1. Place the animal on an electrically programmable heating pad. Cover the rectal temperature probe with a petroleum jelly-based product (see Table of Materials) and place it gently into the rectum (1–2 cm deep) of the mouse to monitor its body temperature. Maintain a body temperature between 36.5 and 37.5 °C during the surgical procedures and experimental recordings.
   2. Monitor the anesthesia level by checking the mouse’s reflexes, whisker movements, and its frequency of breathing. Taking into account the level of anesthesia recorded at least every 30 min, give a small dose of a ketamine-xylazine cocktail (20–50 µL, the same concentration used as before) intramuscularly.

2. Surgery/Craniotomy

1. Fix the head of the mouse in a stereotaxic frame. Using a 30 G needle, inject local analgesic ropivacaine (5 µL, 7.5 mg/mL, see Table of Materials) subcutaneously at the planned incision site. Allow 5 min for it to take effect.
2. Make a straight cut midline in the skin above the skull with a scalpel. Gently pull the skin toward the sides with fine forceps and clamp it aside with bulldog serrefine clamps to leave the skull exposed for further work.
3. Clean the skull of fascia with a scalpel or any similar tool. In case of superficial bleeding, remove the blood with cotton swabs or small pieces of paper towel.
4. Take a poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS)-coated ground screw (size: #00, diameter: 0.047 in, length: 1/8 in, see Table of Materials) with a soldered wire and connect it with a connector to the amplifier headstage.
5. Moisten the skull at the desired hole site and drill a hole at high speed using a fine, round drill bit (with a 0.4 mm diameter) on the skull above the cerebellum until the dura is visible. Put the ground screw into the hole and screw it in with a precision screwdriver until it reaches the top of the cerebellum.
   NOTE: The head screw was dip-coated with a PEDOT:PSS solution containing 1% 3-glycidyloxypropyl)trimethoxysilane (GOPS) by weight followed by baking at 140 °C for 90 min. PEDOT:PSS is a conjugated polymer with a volumetric capacitance that is known to be biocompatible. GOPS is a cross-linker mixed with PEDOT:PSS to increase stability in aqueous media (Figure 2).
6. With the help of the stereotaxic frame, measure the stereotaxic coordinates for the desired brain region. For example, the region of interest is the hippocampus, anteroposterior (AP) -1.8 mm and mediolateral (ML) 1.8 mm from the Bregma point based on the brain atlas for mice³².
   NOTE: These are coordinates for the right hemisphere (Figure 2).
7. Thin an approximately 1 to 2 mm diameter area of the skull above the target region using a reliable dental drill (see Table of materials) set at a fast speed until a thin, well-polished, transparent bone membrane remains.
8. Then, if the thickness of the bone membrane is thin enough (<200 µm), make a small hole with thin forceps and gently remove the thin layer of the bone³³. Use custom-made hook-tipped needle to remove the dura. Minimize the size of craniotomy and durotomy to prevent the development of edemas and to minimize cardiac and/or respiratory pulsations of the brain.
   NOTE: The craniotomy must be filled with a droplet of saline solution to prevent drying and then regularly refilled during the experiment (Figure 2).

3. Insertion of the Multichannel Silicon Probe

1. Use stereotaxic arms in a slight AP angle (20°) for the silicon probe to leave ample space for the positioning of the other two implants and to have the recording and the injection sites of the electrode, ion pump, and micropipette as close as possible.
   NOTE: Electrodes, syringes, and ion pumps were covered with a drop of DiI stain solution (1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate [DiI]), for the post hoc visualization of the implantation traces (0.5 mg/ml DiI in dimethyl sulfoxide).
2. Place the silicon probe on the stereotaxic arm attached to a magnetic holder and place it next to the stereotaxic frame. Set the AP angle (20°), and then connect the probe to the headstage and to the ground screw.
3. Slowly lower the silicon probe into the hippocampus with the help of the micron-precise stereotaxic arm or a motorized micromanipulator to avoid lateral movements (Figure 2 and Figure 3).
   1. Initiate the recording software and record—with the headstage, the connected amplifier, and a computer—electric neuronal signals while moving the multichannel silicon probe from the top of the cortex until the targeted dorsoventral (DV) position is reached (-1,800 µm from the cortical surface). Record and watch the local field potential signal (LFP) during penetration on the computer screen.
      NOTE: Control the descent of the probe so that it is moving slowly and continuously while recording, to have better visual control for the penetration and for reaching the target zone.
   2. Use the ripple activity in the pyramidal layer of the hippocampal formation in the recorded LFP as a marker of the target zone.
      NOTE: Ripple activity is visible on one or two neighboring channels of the multichannel silicon (Si) probe having a 100 µm distance between recording sites (Figure 4).
   3. Record LFP signals from the layers of the cortex and the hippocampus simultaneously through the multichannel amplifier’s software (see Table of Materials) with the help of the multichannel Si probes (Figure 4).

4. Insertion of µFIP

1. Connect tubes (see Table of Materials) to the inlet of the µFIP and fill the probe with 0.05 M GABA solution. Remove the tubes and close the inlet with paraffin film wrapping. Connect electrical leads to the source measurement unit.
2. Insert the µFIP with the help of the stereotaxic arm at a mediolateral (MP) angle (20°). The Si probe remains inserted during the whole process.
   NOTE: µFIP is very flexible and may benefit from the support of a small and clean paintbrush to keep it straight until it reaches the brain surface. After that step, µFIP can be lowered gently with axial movements.
3. Lower the µFIP slowly with axial movements and never let it bend during the trajectory until it reaches the dorsoventral (DV) coordinate (-1,200 µm from the cortical surface).
   NOTE: Try to put the two devices (µFIP and silicon probe) as close to each other as possible, considering the 300 μm distance of the outlet from the µFIP tip.
   CAUTION: Avoid any mechanical issues among the devices and their connectors during insertion (Figure 2B and Figure 3B).

5. Preparation of devices for Seizure Induction

1. Change the metal needle of the syringe (10 µL) (see Table of Materials). Remove the needle-holding metal part, place and fix the micropipette (outer diameter [OD]: 1.2 mm, inner diameter [ID]: 0.75 mm, tip diameter: 20–50 µm with ± 0.5 cm of tapering of the shank), and then replace the needle-holding element.
2. Position the syringe and the attached borosilicate micropipette at a 20° lateromedial (LM) angle for the injection of 4AP (50 mM in artificial cerebrospinal fluid [ACSF]).
   CAUTION: Do not use the metal needle of the syringe or a micropipette with a tip bigger than 50 µm.
3. Draw 500 nL–1 µL of 50 mM 4AP with the help of an automated microinjection pump.

6. Insertion of the Glass Pipette Attached to a Syringe for 4AP Injection

1. Lower the glass micropipette attached to the syringe to the aimed DV position (-1,500 µm), and then inject 250 nL of the 4AP solution (Figure 2 and Figure 3). Start recording with the recording software. Watch the screen and wait for the first interictal spike to appear.
2. Start the GABA delivery by µFIP immediately with the appearance of the first interictal spike. Deliver GABA by applying 1 V between source and target for 100 s followed by 1 s off, for 30 cycles. With the help of the recording software, record for a minimum of 2 h.
   NOTE: The total mass of the delivered GABA is around 1 nmol (Figure 5).
3. At the end of the experiment, gently remove the inserted probes and the ground screw, and remove the animal from the stereotaxic equipment. Animals were euthanized using an overdose of drug (i.p.100mg/kg pentobarbital). Death was confirmed by cessation of breath and circulation.

7. Evaluation of the Placement of the Implants

1. After euthanizing the animal, perfuse it transcardially, first with 50 mL of saline and then with 150 mL of an ice-cold fixative solution containing 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB)³⁴.
   CAUTION: PFA is hazardous and must be handled with care.
2. Decapitate the animal, and then remove the skin and the muscle from the top and sides of the skull. Starting from the foramen magnum, make lateral incisions in the skull toward the ears and a sagittal midline incision, taking great care not to damage the brain. Gently remove the skull with a bone trimmer. Remove the brain, and then cut a tissue block from the region of interest (from the Bregma point, -1 to -3 mm AP) with the help of a brain matrix (see Table of Materials).
3. Glue the tissue block to the specimen holder of a vibratome, put the stand into it, and set the vibratome to 40 µm thickness in a PB bath make 40 µm coronal sections.
4. Wash extensively with 0.1 M PB. Follow the histological protocol for glial fibrillary acidic protein (GFAP) staining³¹.
5. Mount sections on slides and cover them with a mounting medium containing 2-(4-amidinophenyl)-1H-indole-6-carboxamidine (DAPI) (see Table of Materials).

8. Confocal Microscopy

1. Place the slides with the stained coronal sections under a 20x objective of a confocal microscope. Select the target region.
2. Choose the optimal excitation and emission (exc/ems) filter sets for the dyes as follows: DAPI = 358/461 nm, DiI = 551/569 nm, and fluorescein (see Table of Materials) = 490/525 nm.
   NOTE: Since staining varies per section, a proper range of minimal and maximal excitation and detection needs to be determined for each section, where the least dense and most dense regions both show emission.
3. Choose the least dense region and set the laser intensity and detection to high values, and then verify at the densest regions whether these values cause oversaturation of detected emission. If so, lower the values and recheck them with the least dense region. Iterate these steps until arriving at the highest possible detection at low staining levels and proper, not oversaturated levels at highly stained areas. Repeat this process for all dyes.
4. Use the tile scan function of the microscope with 512 x 512 pixels per tile to obtain a large overview of the probe insertion sites with an adequate resolution for post hoc processing.

Results

Using the procedure presented here with a 4AP epilepsy model in anesthetized mice, control of epileptic seizures can be achieved in the epileptic focus. The precise localization of the implants (Figure 2) helped to record hippocampal local field potentials (LFPs, Figure 4), to induce small hippocampal seizures and to deliver GABA at the seizure onset. The localization of the implants was verified after each experiment by post hoc...

Discussion

By developing a new experimental protocol in an acute mouse model of epilepsy, SLEs could be successfully controlled with the help of a µFIP implanted in the epileptic focus. Thanks to its capability to deliver GABA with temporal and spatial precision, 4AP-induced SLEs were controlled at the onset of the seizures. Treatment of epilepsy is theoretically possible if the control of the neural network discharges is achieved at the place of the seizure start. The presented protocol proved this possible if the localizatio...

Materials

Name	Company	Catalog Number	Comments
4AP	Sigma	275875
Alexa Fluor 488	Abcam	ab15007
Amplifier	Neuralynx, Montana, USA	Digital Lynx 4SX
Amplifier	Ampliplex	KJE-1001
Atlas Stereotaxique	Allen Atlas	978-0470054086
Borosilica glass pipette	Sutter	BF120-69-15
Brain Matrix	WPI	RBMA-200C
Bone trimmer	FST	16109-14
Confocal microscope	Zeiss	LSM 510
Connector	INSTECH	SC20/15
Coton tige	Monoprix	EMD 6107OD
Cover slip	Menzel-Glass	15747592
DiI Stain	Thermo Fisher	D282
DMSO	Sigma	11412-11
Drill	FOREDOM	K1070
Forceps	F.S.T.	11412-11
GABA	Sigma	A2129
GFAP Monoclonal Antibody	Thermofisher	53-9892-80
GOPS	Sigma	440167-100M
Hamilton seringe	Hamilton	80330
Headscrew	Component Supply	TX00-2FH
Heating pad	Harvard apparatus	341446
Injection Pump	WPI	UMP3-3
Keithley	Tektoronix	216A
Ketamine	Renaudin	5787419
Magnetic holder	Narishige	GJ-1
Mice	Charles River	612
Motoric manipulator	Scientifica, UK	IVM
Na2HPO4	Sigma	255793
NaH2PO4	Sigma	7558807
NeuroTrace DiI	Thermofisher	N22880
Paper towel	KIMBERLY CLARK	7552000
PB	Sigma	P4417
PEDOT:PSS	CLEVIOS	81076212
PFA	Acros Organic	30525-89-4
Rectal temperature probe	Harvard apparatus	521591
Ropivacaine	KABI	1260216
Saline	Sigma	7982
Scalpel	F.S.T	AUST R195806
Seringue	BD Medical	324826
Serrefine clamp	F.S.T	18050-28	4 is recommended
Silicon probe	NeuroNexus, Michigan, USA	A2x16-10mm-50-500-177 or A1x16-5mm-150-703
Stereotoxic frame	Stoelting	51733U
Superfrost Slide	ThermoScientific	J38000AMNZ
Tubing	INSTECH	LS20
Vaseline	Laboratoire Gilbert	3518646126611
Vectashield DAPI	Vector Laboratories, California, USA	H-1200-10
Vibratome, Leica VT1200S	Leica Microsystems	1491200S001
Xylazine	Bayer	4007221032311