Silk polymer based adenosine release: therapeutic potential for epilepsy

2.1. Implant Design

Implants designed to deliver the target doses (0, 40, 200 and 1,000 ng per day) were designed based on preliminary experiments to determine drug loading. Implants were also designed to split the target drug load roughly evenly between the three systems chosen for integration in the study, microspheres, macroscale films and nanofilms. The three variables used to impact the final adenosine dose were: (1) the concentration of microspheres in the silk solution that were formed into water-based, porous scaffolds, (2) the concentration of adenosine in the adenosine plus silk solution that the porous scaffolds were coated with to form the macroscale silk films, and (3) the number of layers of nanofilms deposited on the system. The composition of the implants and the theoretical amounts of adenosine loaded via each technique are listed in Table 1.

2.2. Implant Fabrication

The implant fabrication process is represented in Figure 1. Silk for the study was prepared from Bombyx mori cocoons as we have previously described [29]. Adenosine containing microspheres were prepared according to the MeOH based protocol described previously [27]. Briefly, 200 μL of 10 mg/mL adenosine stock solution was mixed with 1 mL of 8% (w/v) silk aqueous solution, which was then added to 200 mg of 1,2-Dioleoyl-sn-Glycero-3-Phosphocholine (DOPC) phospholipid that had been dissolved in 1 mL of chloroform, then dried under N₂ to a film on the interior of a glass test tube. The solution is diluted, then repeatedly freeze-thawed, then lyophilized. Later, the microspheres were treated with methanol (MeOH) to remove the lipids and induce beta sheet silk physical crosslinks to stabilize the structures.

Water-based porous scaffolds were prepared as previously described [31] using the mixture of microspheres and silk solution to imbed the microspheres in the final porous scaffold. Briefly, 4 g of 500-600 um granular NaCl were added to 2 mL of 6% (w/v) silk solution mixed with microspheres in a plastic container and incubated at room temperature for over 24 hours. The scaffolds were then washed for 24 hours to leach out the sodium chloride. To obtain the desired implant geometry (0.6-0.7 mm diameter, 3 mm length), scaffolds were punched out with a 1 mm Miltex biopsy punch and then trimmed with a razor on either end. The porous scaffolds were next coated with macroscale drug-loaded silk films comparable to the films described previously [26]. Implants were soaked in a mixed silk and drug solution for 5 min, then incubated at 60°C for 15-20 minutes, then washed in a 90% MeOH solution.

Finally, nanofilm coatings were applied using our previously described protocol [28] but modified to accommodate coating of a 3D porous scaffold. Scaffolds were first dipped in a 2 mg/mL silk fibroin solution for 2 min, and then washed in 90:10 (v/v) MeOH/water solution for 1 min. Following the methanol wash, the scaffolds were dried at 60°C for 15 minutes. The dried scaffolds were then dipped in a 1 mg/mL adenosine solution for 2 min and dried at 60°C for another 15 minutes. These steps were repeated until the desired number of layers was achieved, ending with a silk layer. After drug loading, all implants were coated with three capping layers (three subsequent silk dips) to delay burst.

2.3. In vitro release studies

For each target dose (40 ng, 200 ng, and 1,000 ng) 3 implants were characterized for release kinetics in vitro (N=3). To evaluate release profiles, implants were immersed in 1 mL of Dulbecco's phosphate buffer, pH 7.2 (PBS) at 37°C. At pre-set time points (1, 2, 3, 4, 6, 10 and 14 days) the PBS was removed and replaced.

Adenosine content in the PBS samples removed from the system was measured using a modified fluorescence assay as previously described [32]. The collected PBS sample was transferred to a 1.5 mL Eppendorf tube and chloroacetaldehyde was added to a final concentration of 220 μM chloroacetaldehyde. These tubes were capped and boiled for 20 min. Boiling of mixed adenosine and chloroacetaldehyde yields the fluorescent derivative 1,N⁶-ethenoadenosine. The fluorescence of the sample was measured with a plate reader (excitation = 310 nm, emission = 410 nm [33]. For each device at each time point, three fluorescence readings were taken and averaged.

2. 4. Animals and surgery

All animal procedures were conducted in a facility accredited by the Association for Assessment and Accreditation of Laboratory Animal Care in accordance with protocols approved by the Institutional Animal Care and Use Committee and the principles outlined in the NIH Guide for the Care and Use of Laboratory Animals. Adult male Sprague-Dawley rats were used at a body weight of 280 to 300 g. All rats were acclimatized for one week before being used in the experiments. The rats were housed under a 12 h light / dark cycle (lights on from 8:00 A.M) with food and water provided ad libitum.

Anesthesia was induced with 3% isoflurane, 67% N₂O, 30% O₂ and maintained with 1.5% isoflurane, 68.5% N₂O, 30% O₂, while rats (N=22) were placed in a Kopf stereotactic frame. First, polymers with a target release rate of 0, 40, 200, and 1000 ng adenosine per day (N=5-6 per dose) were implanted using a stereotactic implantation device (internal diameter 0.7 mm, external diameter 1 mm) as described [34]. The polymer-loaded device was stereotactically inserted into the brain using a drill hole above the left hemisphere 2 mm rostral to bregma and 1.6 mm lateral to the midline. Using this drill hole the loaded device was inserted into the brain using an angle of 45⁰ from vertical and an angle of 45⁰ from midline. Thus, a diagonal injection tract was created aiming at a coordinate of 5.5 mm caudal to bregma, 5.5 mm to the right of the midline and 7.5 mm below the dura. Upon reaching the target site the 3 mm long polymer was released and deposited within the infrahippocampal cleft by slowly retracting the outer tube of the device. Finally, the device was fully retracted as described previously [34]. Thus, the implanted polymers were deposited within a formed cavity of 3 mm length within the right infrahippocampal cleft and adjacent to the electrode implantation site (Figure 5). The diagonal implantation approach pursued here and previously [11, 14] is characterized by a number of important advantages: (i) Coverage of 3 mm of the dorso-ventral extent of the hippocampus by placement of the implants into the infra-hippocampal fissure; (ii) Minimization of damage to the ipsilateral hippocampus; (iii) Compatibility with the electrode-containing head-set of the animals. Next, a bipolar, coated, stainless steel electrode (0.20 mm in diameter, Plastics One, Roanoke, VA) was implanted into the right hippocampus and fixed with a head-set of dental acrylate. Coordinates for the hippocampal electrodes were (tooth bar at 0): 5.5 mm caudal to bregma, 5.5 mm lateral to midline, and 7.5 mm ventral to dura.

2.5. Kindling

Four days after surgery, the animals were stimulated unilaterally 6 times every second or third day with a Grass S-88 stimulator (1-ms square-wave pulses of 5 V at 50-Hz frequency for 10 s, 30-min interval between stimulations). Behavioral seizures were scored according to the scale of Racine [35]. Each animal received a total of 48 stimulations, which was equivalent to 8 test days spread out between day 4 and day 20 following polymer implantation. The electroencephalogram (EEG) was recorded for periods of 1 min before and 5 min after application of each stimulating pulse using a Nervus EEG monitoring system.

2.6. Histology

After completion of the experiments the rats were transcardially perfused with 4% paraformaldehyde in phosphate buffer (0.15M, pH 7.4). Brains were then post fixed in the same fixative for 6 hours and cryoprotected in 10% DMSO in PBS (v/v) before being cut into a total of 72 coronal sections (40 μm thickness) covering the full extent of the lateral hippocampus (3.5 mm to 6.5 mm caudal to bregma). To characterize gross anatomy of the brain in particular in the implant region, and to verify the precise location of the kindling electrodes, a Cresyl violet stain was performed on every 6^th section.

2.7. Statistics

In the in vitro studies standard deviations were calculated for the three repetitions per dose. The values for each of the three repetition implants loaded with the same dose were obtained by averaging three fluorescence readings per implant. Standard deviations are listed in Table 2. In vivo seizure data are based on N=5 rats for recipients of control polymers (0 ng adenosine) and for recipients of the 40 ng adenosine per day polymers, and N=6 for recipients of the 200 and 1,000 ng adenosine per day polymers. Individual seizure scores and afterdischarge durations were pooled and averaged for each of the 8 test days in each experimental group. Errors are given as ± SD and data were analyzed using one-way ANOVA with Student-Newman-Keuls Test. * P<0.05, ** P<0.01, *** P<0.001.

3.1. Adenosine release kinetics in vitro

The average release of adenosine from the implants at each time point and the cumulative release over 14 days are summarized in Table 2. Over 14 days, the 8 ng target implant released a total of 100.9 ng (average release rate = 7.2 ng/day), the 40 ng target implant released 467.4 ng (average release rate = 33.4 ng/day), the 200 ng target implant released 2387.3 ng (average release rate = 170.5 ng/day) and the 1,000 ng target implant released 11,465.3 ng (average release rate = 819 ng/day). Release rates for each time point are summarized in Table 3.

3.2. Dose dependent suppression of kindling epileptogenesis by infrahippocampal adenosine releasing implants

Infrahippocampal implantation of silk-based polymers with different target release rates of adenosine ranging from 40 to 1000 ng adenosine per day should suppress kindling epileptogenesis in a dose-dependent way. To test this hypothesis, four groups of rats received infrahippocampal implants with target release rates of 0 ng (control, N=5), 40 ng (N=5), 200 ng (N=6), and 1000 ng (N=6) adenosine per day. These target doses were based on previous dose assessments suggesting an effective dose to be in the range of 200 ng adenosine released per day [15, 16]. Four days after polymer implantation hippocampal kindling was initiated with each rat receiving a total of 8 kindling sessions every second or third day, with each session comprised of 6 stimulations delivered every 30 min. Thus, during a time span ranging from 4 to 20 days after stimulation, each rat received a total of 48 stimulations (Figure 2A).

Seizure scores were determined after each stimulation according to the scale of Racine [35]. Stages 1 to 3 correspond to partial seizures of progressing intensity, whereas stages 4 and 5 represent generalized seizures with tonic and clonic components. Compared to control (0 ng adenosine) kindling development in recipients of target doses of 40, 200, and 1000 ng was initially delayed in a dose-dependent relation during the first 24 sessions or until the end of session 4 that corresponds to 11 days after polymer implantation. Most notably, during this time recipients of polymers releasing a target dose of 1000 ng per day were completely protected from any seizure activity (Figure 2A,B). This dose-dependent therapeutic effect was also reflected by the number of stimulations, needed to elicit the first partial (stage 1 or 2) seizure (Figure 3). While in recipients of control implants partial seizures first occurred during session 2 (i.e. around stimulation 10), the first emergence of partial seizures was markedly delayed with increasing doses of released adenosine with recipients of a target dose of 1000 ng adenosine per day needing 2.5 times as many stimulations to reach the same kindling state as the control group (Figure 3).

From stimulation 25 onwards (i.e. session 5 at day 13 after implantation) the therapeutic effects of the implants declined (Figure 2), which is consistent with the reduced rate of adenosine release from the polymers (Table 3). While recipients of control implants reached their kindling criteria (generalized stage 4/5 seizures) from session 5 onwards, all recipients of adenosine releasing polymers showed a delay in the acquisition of generalized seizure responses (Figure 2). Interestingly, the kindling rates (i.e. the progressive increase in seizure stages with time) were similar in recipients of target doses of 0, 200, and 1000 ng adenosine per day, however, with a dose dependent delay in kindling onset, indicating polymer-based adenosine-release suppressed kindling. Most importantly, recipients of a target dose of 1000 ng adenosine per day showed a delay of kindling acquisition of about three sessions or one week, indicating that doses of around 400 ng adenosine per day (Table 3) are sufficient to suppress kindling.

3.3. Infrahippocampal adenosine releasing implants reduce afterdischarge duration in EEG recordings

The therapeutic effect of the adenosine-releasing infrahippocampal implants was further quantified by determination of the duration of electroencephalographic afterdischarges (ADD) elicited by each test stimulus during the kindling acquisition phase (Figure 4).

The ADD was averaged according to session number (N=6 stimulations, each) and treatment target dose (0, 40, 200, and 1000 ng adenosine per day). During the first three sessions a dose-dependent reduction of the ADD became evident. From session 4 onwards ADDs from recipients of target doses of 40 and 200 ng adenosine per day were similar to those from sham controls (0 ng adenosine per day), being consistent with the decline in adenosine released from these polymers with time (Table 3). Remarkably, afterdischarges remained significantly reduced throughout the whole experiment in recipients of polymers releasing a target dose of 1000 ng adenosine per day. At that time the ADD of recipients of target doses of 0, 40, and 200 ng adenosine per day had reached a plateau with average values exceeding 120s, whereas the ADD of recipients of target doses of 1000 ng adenosine per day remained reduced with values of <100s. Thus, the behavioral seizure response as observed in Figure 2 was closely paralleled by the duration of afterdischarges in respective EEGs. These results also indicate that a target dose of 1,000 ng adenosine per day has the potential for prolonged seizure suppression, even if the rates of adenosine release decline with time. After completion of seizure monitoring all animals were sacrificed at day 20 after polymer implantation and subjected to gross histological analysis. In coronal brain sections we detected a cavity resulting from the polymer deposited within the infrahippocampal fissure and traces of the electrode insertion were found in correct locations (Figure 5). No obvious since for inflammation or bleeding were found.

(i) Effective dose assessment

It is important to note that first signs of kindling development (e.g. behavioral seizure scores) typically appear with a delay of 1 session comprised of 6 stimulations. Thus, recipients of control implants display first signs of kindling during session 2 (Figure 2); this implies that kindling epileptogenesis is initiated in the session (i.e. at the relevant dose of adenosine released at that time) preceding the first signs of kindling. Consequently, first signs of kindling appear with a dose dependent delay of about 1 session in the control group (stimulation #6), 2 sessions in the 40 ng group (stimulation #10), 3 sessions in the 200 ng group (stimulation #16), and 4 sessions in the 1000 ng group (stimulation #24) (Figure 2A). This close dose-dependency is also reflected in the average number of stimulations needed to elicit the first stage 1 or 2 seizure (Figure 3) and suggests a linear dose-response relation. Since our experiments were designed to prolong kindling after cessation of adenosine release from the polymers, parallel in vitro adenosine release data (Table 3) allowed us to determine the minimal effective dose needed to delay the onset of kindling. This becomes most evident in recipients of a target dose of 1000 ng adenosine per day. These animals continued to be protected from any seizure activity during the first 23 stimulations (i.e. during session 1 to 4) (Figure 2). First seizures occurred during session 5; taking in consideration the one-session-delay of detecting behavioral seizures, we conclude that kindling epileptogenesis was initiated during session 4, corresponding to day 11 after polymer implantation. At that time the polymers released approximately 200 ng adenosine per day (Table 3). Similar calculations performed in recipients of target doses of 40 and 200 ng adenosine per day indicated that kindling epileptogenesis was initiated as soon as daily release rates of adenosine fell below 50 to 100 ng adenosine (Figure 2, Table 3). These calculations allow us to estimate the minimal effective dose to be in the range of 50 to 200 ng adenosine per day. These data are in agreement with previous cell transplantation experiments, in which adenosine release rates of about 20 to 40 ng adenosine released from 10⁵ cells was sufficient to suppress kindled seizures [15, 17]. It is also important to note that high doses of adenosine (> 2000 ng adenosine per day) appeared to be well tolerated, without any gross behavioral changes of the animals.

(ii) Possible antiepileptogenic effect

An important finding of this study is the observation that kindling acquisition occurs with a dose-dependent delay but at the same kindling rate (i.e. progression in seizure severity). This observation suggests that implant-derived adenosine might prevent epileptogenesis. Our unique experimental design, limiting adenosine release to 14 days (corresponding to session 1 to 5), but extending kindling to 20 days (corresponding to session 8) allowed us to study resuming of kindling after expiration of the polymers. Reduced seizure stages (Figure 2) combined with the presence but reduced duration of ADDs (Figure 4) in recipients of adenosine releasing implants during the early stages of kindling could have two alternative explanations: suppression of seizures, however seizure suppression may mask ongoing epileptogenesis, or prevention of epileptogenesis. In the first case, if seizures were simply suppressed by graft-derived adenosine and epileptogenesis was masked by seizure suppression, cessation of adenosine release from the polymers would imply a sudden, rapid rise in seizure severity. However, this is clearly not the case (Figure 2). Instead, the rate of kindling acquisition was not changed in recipients of adenosine releasing polymers. Thus, kindling development in recipients of a 1,000 ng per day target dose proceeded at the same rate as kindling development in control animals (Figure 2), however, with a delay in kindling onset. These data suggest that focal polymer-based delivery of adenosine might prevent epileptogenesis. It is important to note that afterdischarges that are considered as the “epileptogenic” stimulus are not suppressed by implant-derived adenosine. The interpretation of our data is consistent with previous studies attributing antiepileptogenic effects to increased brain adenosine in adenosine kinase mutant mice [11] or to stem cell derived brain implants that release adenosine [11, 14].