In vitro attenuation of astrocyte activation and neuroinflammation through ibuprofen-doping of poly(3,4-ethylenedioxypyrrole) formulations

Neuroinﬂammation is often associated with poor functional recovery and may contribute to or initiate the development of severe neurological disorders, such as epilepsy, Parkinson’s disease or Alzheimer’s disease. Ibuprofen (IBU), being one of the most commonly used non-steroidal anti-inﬂammatory drugs, is known to possess neuroprotective activity and serve as a promising therapeutic for the treatment of neuroinﬂammation. In this study, the potential of an IBU-loaded poly(3,4-ethylenedioxypyrrole) (PEDOP) matrix has been assessed as a neural interface material with an aim to control astrocyte activation and suppress neuroinﬂammation in vitro. Three types of drug immobilization protocols were inves- tigated, leading to the fabrication of IBU-loaded PEDOP matrices exhibiting a broad spectrum of electrical characteristics, drug release proﬁles, as well as biological responses. Among all investigated PEDOP for- mulations, PEDOP matrices formed through a three-step immobilization protocol exhibited the highest charge storage capacity (30 ± 1 mC/cm 2 ) as well as a double layer capacitance of 645.0 ± 51.1 m F, associ- ated with a relatively enlarged surface area. Demonstrating a total drug loading capacity of 150 m g/ml and a release rate constant of 0.15 1/h, this coating formulation may be employed as a safe electrical conduct- ing drug eluting system. (cid:1) 2020 The Authors. Published by Elsevier B.V. ThisisanopenaccessarticleundertheCCBYlicense(http:// creativecommons.org/licenses/by/4.0/).


Introduction
Neuroinflammation, defined as an inflammatory response within the central or peripheral nervous system, can be associated with mechanical trauma, infection or neurodegenerative disease [1,2]. The neuroinflammatory response can be described as a composition of overlapping cytotoxic and cytotrophic processes, which in evolved organisms are regulated in order to promote homeogenesis of the nervous system [3]. Regardless of the source of neuroinflammation, this state is often associated with delayed recovery [1], and chronic brain inflammation is a key barrier to regeneration [3]. Critically, prolonged inflammation may contribute to the formation of disease pathogenesis across the central nervous systems and the development of neurological disorders [2]. For instance, disregulation of inflammatory cell activation in the neural tissue injuries is seen to be associated with the development of epilepsy in humans [4]. Recently, inflammation has also acquired relevance as one of the principal mechanisms of neuronal dysfunction in Parkinson's disease [5], b-amyloid pathology and the progression of Alzheimer's disease [6].
Although non-steroidal anti-inflammatory drugs (NSAIDs) are known to exhibit neuroprotective functions, the origins of these effects are not yet well-known [7,8]. One of the most commonly used over-the-counter NSAIDs is iso-butylphenylpropionic acid, also known as ibuprofen (IBU). IBU is a potent free radical scavenger that can reduce free radical generation, and is known to decrease the production of nitric oxide, protect neurons against glutamate toxicity and decrease the production of proinflammatory cytokines in humans [9,10]. IBU has been also shown to down-regulate the inflammatory response induced by traumatic brain injury [11], protect dopaminergic cells and slow down the progression of Parkinson's disease [12], as well as exhibit neuroprotective effects against Alzheimer's [9] and Machado-Joseph diseases [13] in vivo. Because of its neuroprotective activity, relative safety, and long history of use, IBU may serve as a promising therapeutic for the treatment of neuroinflammation [9]. To maximize the efficacy of NSAIDs in central nervous tissue and avoid off site-effects, various strategies for localized drug delivery have been https://doi.org/10.1016/j.bioelechem.2020.107528 1567-5394/Ó 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). investigated in vivo [14]. Of particular interest in neurospecific applications are drug-loaded conducting polymers matrices which may serve as precisely controlled drug delivery platforms [1516] that can be controlled through an electrical signal [17]. Indeed, conducting polymer systems have already been successfully applied for the controlled release of NSAIDs, particularly naproxen [16] and ibuprofen [18,19].
In this paper, the electrochemical performance, drug release profile and neuronal response to IBU-loaded poly(3,4ethylenedioxypyrrole) (PEDOP) matrices is assessed. Although the most frequently used conducting drug carriers are poly(3,4-et hylenedioxythiophene) (PEDOT) and polypyrrole (PPy), their structural analogue, PEDOP, exhibits a combination of the advantages of its ''parent" polymers and is devoid of some of their disadvantages. In contrast to EDOT, EDOP is soluble in water and can be electropolymerized under aqueous conditions [20]. Similarly to PEDOT [21], due to the presence of dioxyethylene bridging group, PEDOP is supposed to exhibit higher electrochemical stability than PPy. Furthermore, PEDOP has been described previously as exhibiting neurocytocompatibility [22], and having a general potency to serve as a drug carrier in vitro [18]. The low oxidation potential of EDOP enables to initialize the process of polymerization at lower potential than for EDOT [20], making PEDOP an efficient drug carrier also for biologically active compounds which are prone to oxidative degradation. Herein, three different types of IBU immobilization protocols, drug release profiles, as well as biological response of IBU-loaded PEDOP matrices were investigated and characterized in vitro. Finally, the optimized conditions for the fabrication of an electroactive conducting polymer coating with superior electrochemical characteristics and enhanced neuroprotective functionality in vitro is presented.

Drug immobilization
The process of electrochemical polymerization was performed by means of a PARSTAT 2273 potentiostat in a three-electrode set-up, comprising a Pt-coated Thermanox coverslip (Electron Microscopy Sciences) as a working electrode, Ag/AgCl (3 M KCl) (EDAQ) as a reference electrode and a glassy carbon rod as an auxiliary electrode. 10 mM 3,4-ethylenedioxypyrrole (EDOP, 2% (w/v) in THF, Sigma Aldrich) was polymerized in a course of a cyclic voltammetry (CV) in the aqueous solutions of either 0.1 M sodium p-toluenesulfonate (PTS, 95%, Sigma Aldrich) to form P-PTS matrix, or 0.1 M ibuprofen sodium salt (IBU, !98%, Sigma-Aldrich) to form P-IBU(1) matrix, or 0.1 M PTS and 0.1 M IBU to form P-IBU(2) matrix. In each case, the geometrical surface area of the working electrode was equal to 0.283 cm 2 . CV curves were collected at the scan rate of 0.1 V/s for 50 CV cycles within the potential range between À0.5 V and 0.9 V (vs. Ag/AgCl). P-IBU(3) matrix was formed in the course of a three-step immobilization protocol [23], comprising a) the electrochemical polymerization of EDOP (10 mM) in 0.1 M PTS aqueous solution, followed by b) dedoping of P-PTS matrix through chronoamperometric reduction at À0.5 V (vs. Ag/AgCl) in 0.1 M PTS aqueous solution for 600 s and c) secondary doping in the presence of 0.1 M IBU aqueous solution by chronoamperometric oxidation at 0.5 V (vs. Ag/AgCl) for 600 s. The potentials of chronoamperometric reduction and oxidation were selected basing on the CV curves of PEDOP to allow the matrix to be fully reduced/oxidized. Grade 1 (R > 10 MX/cm) deionised water was employed as a solvent for preparation of all used solutions.

Material characterization
A PARSTAT 2273 potentiostat was used for the electrochemical studies performed in a three-electrode set-up as described above. CV curves were collected in a physiologically relevant phosphate buffered saline solution (PBS, 0.01 M Na 2 HPO 4 , 0.0018 KH 2 PO 4 , 0.0027 M KCl and 0.137 M NaCl, pH = 7.4, Sigma Aldrich), within the potential range from À0.7 to 0.7 V (vs. Ag/AgCl) at 0.1 V/s for 5 CV cycles. CV curves were used to determine charge storage capacity (CSC), calculated as the electric charge integrated under a corresponding CV curve during one CV cycle [19], according to the formula: where t 1 is the time of beginning of a CV cycle (s), t 2 is the time of end of a CV cycle (s), and I is the current density (A/cm 2 ). EIS spectra were collected in a PBS solution within a frequency range from 100 mHz to 100 kHz, with an AC amplitude of 40 mV (vs. Ag/AgCl) and a DC potential of 0 V (vs. Ag/AgCl). The results were presented on Bode plots and compared to those of a bare Pt electrode. EIS Spectrum Analyzer 1.0 software [24] and the Powell algorithm were used to fit the experimental data to an equivalent circuit model. Capacitance was calculated basing on the parameters of a constant phase element (CPE) according to the formula: where C is the capacitance (F), R is the film resistance (X), P and n are CPE parameters. The chemical structure of polymer matrices was characterized by means of a Raman spectroscopy using Renishaw InVia confocal microRaman system equipped with a laser operating at 830 nm and a CCD detector. Raman spectra were acquired in the spectral range between 300 and 1800 cm À1 . The surface morphology of polymer matrices was studied by means of a Hitachi S-4700 Scanning Electron Microscope operating at 15 kV. ImageJ (NIH) image analysis software was used to quantify the sizes of polymer grains. Origin software was used to plot histograms with normal distribution curve overlays. Thickness of samples was determined by means of a Profilm3D Profilometer and 3D Optical Profiler (Filmetrics). The results were expressed as a mean of at least 12 measurements ± standard deviation.

Drug release studies
The release experiments were conducted according to the previously developed protocol [25]. Briefly, Pt-coated Thermanox coverslips coated with P-IBU(1), P-IBU(2) and P-IBU(3) were immersed in 1 mL PBS solution and left under gentle shaking (80 rpm, 37°C) for 14 days. 0.2 mL of the supernatant was taken at specific time points (1 h, 3 h, 6 h, 8 h, 24 h, 2 days, 4 days, 6 days, 8 days, 10 days, 12 days and 14 days) which was supplemented with 0.2 mL of a fresh PBS solution to keep a constant elution volume of 1 mL. This procedure allowed the maintenance of a so-called sink condition preventing the saturation of the supernatant with IBU. The concentration of released IBU was determined through UV-Vis spectrophotometry with the use of a Thermo Scientific Evolution 60S Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) and was based on the absorbance at the peak characteristic for IBU (264 nm) (Fig.S1A). IBU concentration in solution was determined through a calibration curve (Fig.S1B) plotted for absorbance versus drug concentration. A linear relationship was observed between 0.001 mM and 1.0 mM IBU satisfying the equation y = 0 .4x + 0.01 (R 2 = 0.999). The measurements were performed for four technical replicas and the results were expressed as a mean ± standard deviation.
Two models were used to analyze the release kinetics: (a) power equation [26]: rearranged to its linear form: where: M t is the amount of drug released at time t, M 1 is the amount of drug at the equilibrium state, k is the release rate constant, and n is the exponent of release related to drug release mechanism.
rearranged to its linear form: where: X is the fraction of a drug released at time t, n is the Avrami parameter and k is the release rate constant. The fitting of the data describing the release kinetics was made with the use of two free parameters, namely n and k. The equilibrium concentration was determined experimentally.

Biological characterization
The cytocompatibility of P-PTS, P-IBU(1), P-IBU(2) and P-IBU(3) formulations was determined with respect to a primary mixed neural population obtained from the mesencephalon of embryonic Sprague-Dawley rats, and cultured for 3, 7 and 14 days, according to a previously developed protocol [28]. The cellular composition on control and PEDOP-coated Pt electrodes was assessed by immunofluorescence markers for astrocytes (GFAP) and neurons (b-tubulin III), and the fluorescent images were collected by an Olympus Fluoview 1000 Confocal Microscope (Olympus, Tokyo, Japan). Cell density and the average number of astrocytes were analyzed by counting the number of nuclei corresponding to neurons and astrocytes in an area of 900 lm 2 in at least 20 random images taken from test and control groups [29]. The same sets of fluorescent images were used to determine the mean area of astrocytes, as carried out with the use of ImageJ (NIH) image analysis software.
The biological experiments were conducted to include three biological replicas for all test and control groups. The results were expressed as a mean ± standard deviation or standard error of the mean. One-way Anova and t-test were performed to determine the statistical significance (p < 0.05).

Electrochemical characterization
The immobilization of a drug into a conducting polymer matrix during synthesis is typically employed in order to balance positive charges caused by the oxidation of the growing polymer chain. Therefore, if the drug has an anionic structure, it can act as a primary doping ion. Some strategies, however, involve the use of a mixed electrolyte, namely a primary dopant and a drug co-dopant, to facilitate the process of electropolymerization. Alternatively, drug mole-cules can be incorporated into conducting matrix following polymer synthesis, and this process is based on an ion-exchange mechanism between the primary dopant, which is removed from the matrix during electrochemical reduction, and drug molecules, which are incorporated in the polymer matrix during electrochemical oxidation. These three immobilization strategies result in the formation of polymer matrices possessing different characteristics in terms of electrochemical properties, drug loading capacity and drug release kinetics. Here, drug-loaded PEDOP matrices were formulated via three immobilization protocols as summarized in Scheme 1, leading to the formation of three IBU-loaded PEDOP-based systems, namely P-IBU(1), P-IBU(2) and P-IBU (3).
The corresponding cyclic voltammetric (CV) curves of electrochemical polymerization, presented in Fig.S2A-C, indicate that in all cases the monomer was irreversibly oxidized and a conductive film was electrodeposited onto a Pt electrode. The thickness of asformed polymer films was comparable, achieving the highest value for P-PTS (332 ± 61 nm) and slightly lower values for IBU-loaded PEDOP: 307 ± 58 nm, 282 ± 47 nm and 304 ± 64 nm, for P-IBU (1), P-IBU(2) and P-IBU(3), respectively. The CV curves recorded during the polymerization of EDOP in the presence of IBU exhibited oxidation peaks at a potential of 0.5/0.6 V (vs. Ag/AgCl). The obtained chronoamperometric curves (Fig.S2D) were typical for a three-step immobilization procedure.
After the synthesis, all PEDOP-based matrices were further examined to assess their electrochemical properties. CV curves were obtained from electrodeposited PEDOP and IBU-loaded PEDOP matrices, namely P-PTS, P-IBU(1), P-IBU(2) and P-IBU (3), as well as from bare Pt electrodes (Fig. 1A). It was observed that the area under the CV curves (equivalent to charge passing through the modified electrode) obtained from P-PTS and P-IBU(3) coated Pt electrodes was more developed than the area under CV curves obtained from P-IBU(1) and P-IBU(2) coatings. The beneficial effect of a three-step immobilization procedure on the electrochemical properties of the resulting material was revealed when assessing the charge storage capacity (CSC) [30] of all investigated PEDOP formulations (Fig. 1B). Here, the P-IBU(3) coating was observed to exhibit a CSC value of 30 ± 1 mC/cm 2 , significantly higher than was in the case of P-IBU(1) (22 ± 2 mC/cm 2 ), P-IBU(2) (21 ± 1 mC/cm 2 ), as well as a bare Pt electrodes (7 ± 1 mC/cm 2 ), indicating that P-IBU(3) fims were able to store the highest charge before reaching overpotential and undergoing irreversible faradaic reactions relative to other investigated matrices.
Similarly, the behaviour of an electric charge passing through the electrode and the electrode coating was determined by means of electrochemical impedance spectroscopic (EIS) studies as shown in Fig. 2A (impedance modulus vs. frequency) and Fig. 2B (phase angle vs. frequency). Analysis of the Bode impedance and phase plots for pristine PEDOP and IBU-loaded PEDOP matrices indicated that P-PTS and P-IBU(3) demonstrated similar mechanism of charge transfer. Likewise P-IBU(1) and P-IBU (2), which also demonstrated similar mechanism of charge transfer. It can be inferred that this effect is due to the presence of a co-dopant, IBU, within the whole volume of PEDOP matrix, which is not the case when IBU is immobilized after the process of electrochemical polymerization (as with a P-IBU(3) matrix).
A detailed investigation of the charge transfer was possible through the simulation of EIS data with an equivalent circuit model. Basing on the visual investigations of collected spectra, particularly Bode phase plots, two equivalent circuits were proposed for the description of charge transfer occuring within investigated samples. Since in the case of P-IBU(3) drug molecules are expected to be immobilized at the near surface region of the polymer matrix rather than in bulk, the mechanism of charge transfer for P-IBU (3) is supposed to resemble more the mechanism of charge transfer for P-PTS than for P-IBU(1) and P-IBU(2). Consequently, for P-PTS and P-IBU(3), a modified Randles circuit was employed (Fig. 2C), comprising a parallel combination of a solution resistance (R S ), charge transfer resistance (R CT ), double layer capacitance (C DL ) and constant phase element (CPE P ) associated with the capacitance (C P ) of the polymer matrix. The model resulted in the good fit between the experimental and simulated data (2.7% and 2.2% deviation for P-PTS and P-IBU(3), respectively). To properly model P-IBU(1) and P-IBU(2), it was necessary to add one more circuit element (red square in Fig. 2C) responsible for the diffusion of drug molecules (W D ) and associated resistance (R D ), forming the equivalent circuit model similar to the one developed prevously for PEDOP [20]. The additional capacitive element, giving rise to a capacitive peak at approx. 100 Hz in the phase plot of P-IBU(1) and P-IBU (2) (Fig. 2B), should be related to the diffusion of drug molecules within polymer matrix as the result of spontaneous release of IBU occuring during the measurement. Also here, the model resulted in a good fit between the experimental and simulated data (0.9% and 1.9% deviation for P-IBU(1) and P-IBU(2), respectively). The EIS fitting parameters for PEDOP and IBU-loaded PEDOP matrices are summarized in Table S1.
Since the charge transfer resistance of a bare Pt electrode was calculated to equal 17.2 ± 1.5 kO, all PEDOP formulations are  shown to significantly contribute to a decrease in the overall impedance moduli of the electrode. The charge transfer resistance varied between 43 ± 5 O for P-IBU(3) matrices and 10 ± 1 O for P-IBU(1) matrices, but it should be noted that although all polymer formulations exhibited low R CT valuses, additional resistive factor associated with the diffusion of IBU should be considered. The depletion of the IBU dopant through spontaneous release leads to diffusional limitations increasing matrix resistance. The P-IBU(3) matrix was also found to exhibut the highest double layer capacitance (645.0 ± 51.1 mF), which could be correlated to its developed surface area (see Section 3.2).

Material characterization
The chemical structure of all experimental PEDOP chemistries was confirmed through Raman analysis (Fig. 3). which indicated characteristic PEDOP peaks attributable to the asymmetric stretching mode of C@C bonds of pyrrole (1613 cm À1 ), the symmetric stretching mode of C@CAO groups of PEDOP (1396 cm À1 ), the CAC stretching mode (low intensity peaks in the range of 1300-1350 cm À1 ), the CAN stretching of the pyrrole rings (1186 cm À1 ), the NAH vibrations (1020 cm À1 ) and the NAH out-of-plane bending mode (706 cm À1 ) [18,31,32]. The presence of PTS dopant in P-PTS polymer formulations was confirmed by the appearance of a peak assigned to symmetric SO 3 À bending at 634 cm À1 [33]. The disappearance of this signal in the Raman spectrum of P-IBU (3) proves that the conditions of dedoping were sufficient to entirely remove PTS from the near surface region of the polymer matrix. The presence of IBU moiety in all IBU-loaded matrices: P-IBU(1), P-IBU(2) and P-IBU(3), was confirmed by the appearance of bands characteristed for the IBU molecule (Fig.S3), namely the out of plane CAC ring deformation (484 cm À1 ), in plane CAH ring bending (1062 and 1342 cm À1 ), CAC ring stretching (1264 cm À1 ), as well as CAH deformations (1502 cm À1 ) [34]. Therefore, all three drug immobilization protocols were shown to be efficient for entrapment of IBU moieties inside PEDOP matrix. Chemical doping is known to have a strong effect on the morphology of conducting polymer formulations [20]. As shown in Fig. 4A, the surface of electrodeposited P-PTS films was composed of polymer grains with an average diameter of~0.6 mm, and which  were regularly dispersed over the Pt electrode. When IBU was incorporated as a primary dopant (P-IBU(1)), the polymer grains were grouped into agglomerates with an average diameter of 1.4 mm (Fig. 4B). Conversely, the concurrent presence of both doping ions, namely PTS and IBU, enhanced the formation of a compact surface layer, in which the polymer grains were no longer individual structures (Fig. 4C). The electrochemical reduction and reoxidation of the initial P-PTS matrix, through a three-step procedure of IBU immobilization, led to noticeable surface erosion and matrix stress during the ion exchange process. Interestingly, the initial globular structure (particles with an average diameter of 0.5 mm) was maintained in P-IBU(3) films (Fig. 4D) which could partially contribute to the enhanced electrochemical performance of this matrix as previously observed in PEDOT formulations [35]. Similar morphological reversibility during doping and dedoping processes has been already observed by Schoetz et al. [36], who investigated charging and discharging of PEDOT by atomic force microscopy. Noteworthly, the distribution of average diameters of polymer grains of P-IBU(3) films (Fig.S4D) resembled the standard normal size distribution, in contrast to other PEDOP formulations ( Fig.S4A-C). Due to the presence of polymer grain agglomerates, P-IBU(2) was found to diverge significantly from a normal size distribution.

In vitro IBU release studies
The suitability of PEDOP formulations for drug delivery was verified by IBU elution studies under spontaneous release conditions over a period of 14 days. The drug release profiles of IBU-loaded PEDOP matrices, namely P-IBU(1), P-IBU(2) and P-IBU(3) are presented in Fig. 5, indicating both the experimental values as well as the corresponding fitted curves calculated by means of power and Avrami's models. Power equation is a well-established semiempirical model frequently used to describe drug release from polymeric systems [37,38]. Although Avrami's equation was originally used to describe crystallization mechanisms, it has been also used many times for the description of release kinetics [39][40][41], especially for conducting polymer based drug delivery systems [27,[42][43][44][45]. In our case, the use of power law resulted in higher values of the correlation coefficient (R 2 = 0.99 for all three IBU-loaded PEDOP matrices) than the use of Avrami's model (R 2 between 0.96 and 0.98) ( Table 1). Also the visual investigation of the fitted curves (solid and dash lines in Fig. 5 representing fitted curves calculated by means of power or Avrami's equations, respectively), demonstrated that the power law is describing the kinetics of release in a better way than Avrami's model. Particularly, fitting for Avrami's model led to the overestimation of the concentration of released drug in the first 7 days of release, and the underestimation of IBU concentration in longer periods. Both models gave the values of the exponent of release (n) lower than 1. Since in both models n is related to the drug release mechanism, its value indicates that  the release of IBU from IBU-loaded PEDOP matrices has a diffusive character [38,42]. This is consistent with previous literature reports on polypyrrole-based drug delivery systems, for which a diffusive mechanism was found at open circuit conditions (spontaneous release) [27,42,45]. According to Papadopoulou et al. [46], such low values of n may occur in highly disordered matrices, which differ much from the percolation cluster.
The highest IBU release rate constant (0.45 1/h and 0.39 1/h according to the Avrami's and power models, respectively) was noted in P-IBU(2) formulations, corresponding to approx. 225 mg/ ml IBU released over a period of 14 days. Further simulation with power kinetic parameters indicated that 95% of the total amount of immobilized IBU (250 mg/ml) will be eluted within 22 days. The release of IBU from both P-IBU(1) and P-IBU(3) formulations was reduced relative to P-IBU(2) formulation which was supposed to arise from IBU acting as a main dopant, necessary for maintaining the electroneutrality of the polymer. Here, over a period of 14 days only 160 mg/ml and 130 mg/ml of IBU was eluted for P-IBU(1) and P-IBU(3), respectively, and the release of 95% of the total amount of immobilized IBU (200 mg/ml and 150 mg/ml, respectively) was accomplished in circa 20 days (according to power equation).
Although a high drug loading capacity would seem to be an indicator of an efficient drug carrier, achieving a gradual, constant release of the optimal portion of drug over a prolonged time should be considered the ultimate goal of an ideal drug release system. According to recent studies, high doses of IBU could be hazardous in specific applications. In particular, IBU is used to close the ductus arteriosus in cases of neonate hyperbilirubinemia. Berns et al. [47] investigated bilirubin-associated brain damage in premature infants, showing that IBU delivery above a concentration of 125 lg/ml could interfere with bilirubin-albumin binding, decreasing neuron viability. Therefore, even though exhibiting the lowest drug loading capacity in this study (150 mg/ml) and the slowest release rate constant (approx. 0.15 1/h), P-IBU(3) could be considered as the most beneficial and the safest drug eluting system among all investigated IBU-loaded PEDOP matrices for neural applications.

Biological characterization
In order to evaluate the cytocompatibility of IBU doped PEDOP formulations, the analysis of cell density of astrocyte and neuron populations, and astrocyte activation, observed as the increase in astrocyte cell area [48], was assessed in vitro. Consequently, the analysis of the density of astrocyte and neuron populations on bare Pt and PEDOP-coated Pt electrodes (Fig. 6) revealed a significant decrease in the presence of astrocytes and an increase in the presence of neurons for all PEDOP-coated Pt electrodes with respect to bare Pt substrates. The lowest astrocyte presence was observed on P-IBU(3) formulations after 3 days in culture (4.8 ± 0.6%), significantly lower than on bare Pt substrates (55.0 ± 2.9%) and all other PEDOP formulations. Since the high percentage of a specific type of cells at an early time point should be associated with enhanced adhesion of cells to the surface, it could be stated that the surfaces of P-IBU(3), P-PTS and P-IBU(1) exhibit specific interactions with neurons. Interestingly, the surface of P-IBU(2) seems to increase the adhesion of neurons less efficiently than other PEDOP formulations, and this could be associated with its surface morphology, which was shown to be the most compact among all investigated PEDOP formulations. On the surface of a bare Pt electrode, the ratio between neurons and astrocytes is close to 1:1, showing no specific interactions of this substrate with a particular type of cells.
The next two time points (7 and 14 days) represent the growth and the development of cell population, in which the density of neurons and astrocytes should be balanced to form a healthy neural environment. The surface of a bare Pt electrode was shown to initially favour the development of neurons, but after 14 days the ratio between astrocytes and neurons was again equal to approx. 1:1. Apart from P-IBU(1), for which there was no significant change in cell density in time, other PEDOP formulations showed the increase in the average percentage of astrocytes. Although astrocyte presence was observed to increases with time, reaching 15.5 ± 1.5% after 14 days in culture, it remained significantly lower on P-IBU(3) relative to Pt substrates, indicating that these PEDOP materials may promote the development of a healthy neural environment. The similar cell density of astrocyte and neuron populations was observed on the surface of PEDOT doped with dexamethasone and decorated with Au particles [49]. Although dexamethasone is a potent anti-inflammatory corticosteroid commonly used in the field of neural devices, when administrated systemically it is found to cause serious side effects, including myopathy and diabetes [50]. Ibuprofen, as a non-steroidal anti- Table 1 Release kinetic parameters calculated basing on power and Avrami's equations for the spontaneous release of IBU from IBU-loaded PEDOP matrices.: k -release rate constant, nrelease exponent, t 0.95 -time necessary for the release of 95% of immobilized drug, R 2 -correlation coefficient.

Power equation
Avrami's equation  inflammatory over-the-counter drug, possessing similar ability to reduce tissue reactions following electrode implantation, can be considered as a safe alternative to dexamethasone. Immunofluorescent analysis (Fig. 7A-E) was subsequently performed to quantify the mean neuron number (Fig. 7F) and the mean area of the astrocyte cell body (Fig. 7G) in mixed neural populations cultured on all experimental and control PEDOP formulations. Here, the number of neurons was observed to be the highest on P-IBU(3) for-mulations relative to other experimental PEDOP formulations and a control material, indicating that P-IBU(3) formulations promoted neural adhesion in vitro. Indeed, after 14 days in culture a~6x increase in neuron presence was observed on the surface of P-IBU(3) formulations relative to bare Pt electrodes, and a~2x increase in neuron presence relative to other experimental PEDOP formulations. Therefore, P-IBU(3) was found to outperform other electrode materials intended for neural applications, e.g. polyhy- droxyalkanoate/carbon nanotube nanocomposites, for which a~3x increase in neuron presence was observed relative to bare Pt electrodes [29]. An increase in the size of the astrocyte soma is associated with astrocyte activation [48] and can be employed to assess neuroinflammation in vitro. Here, the analysis of the mean astrocyte cell area was performed to assess the anti-inflammatory function of IBU doped PEDOP films (Fig. 7G). It was observed that astrocytes cultured on P-IBU(3) formulations possessed a mean cell area of 666 ± 74 lm 2 , significantly less that than astrocytes cultured on bare Pt electrodes (1483 ± 179 lm 2 ), and similarly to astrocytes cultured on self-supporting CNT films (637 ± 54 lm 2 ) introduced previously by our group as a material able to prevent reactive gliosis due to its mechanical biomimicry [51]. Also P-IBU(1) formulations were observed to reduce the mean astrocyte cell area (819 ± 95 lm 2 ), indicating that this PEDOP formulation also possesed an anti-inflammatory functionality. Interestingly, even though the mean area of astrocytes cultured on P-PTS and P-IBU (2) was moderately reduced relative to neural populations cultured on bare Pt substrates, these changes were not statisticaly significant.

Conclusions
In this study, the electrochemical and biological effects of ibuprofen doping on poly(3,4-ethylenedioxypyrrole) thin film substrates was assessed in vitro. It was observed that PEDOP formulations formed via a three-step IBU doping/immobilization protocol, P-IBU(3), exhibited the highest CSC (30 ± 1 mC/cm 2 ) relative to all other experimental PEDOP formulations and control substrates. Although all PEDOP formulations were shown to possess a significantly decreased electrochemical impedance modulus relative to Pt substrates, P-IBU(3) formulations exhibited a significant increase in a double layer capacitance (645.0 ± 51.1 mF) which was associated with its developed surface topography. Characterisation of the total drug loading capacity indicated that different PEDOP formulations were able to immobilize from 150 mg/ml (P-IBU(3)) to 250 mg/ml (P-IBU(2)) of IBU, and elute 95% of loaded drug within circa 20 days. With the release rate constant of 0.15 1/h, P-IBU(3) was found to provide the optimal amount of drug necessary to promote the development of a healthy neural environment and to prevent the activation of astrocytes. Consequently, P-IBU(3) formulations formed via a three-step drug immobilization protocol could be considered as the most beneficial and the safest drug eluting system among all investigated IBU-loaded PEDOP matrices for applications in neural drug delivery.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.