Poly(curcumin-co-poly(ethylene glycol)) films provide neuroprotection following reactive oxygen species insult in vitro

Abstract

Objective:

Curcumin is an antioxidant and anti-inflammatory molecule that may provide neuroprotection following central nervous system (CNS) injury. However, curcumin is hydrophobic, limiting its ability to be loaded and then released from biomaterials for neural applications. We previously developed polymers containing curcumin, and these polymers may be applied to neuronal devices or to neural injury to promote neuroprotection. Thus, our objective was to evaluate two curcumin polymers as potential neuroprotective materials for neural applications.

Approach:

For each curcumin polymer, we created three polymer solutions by varying the weight percentage of curcumin polymer in solvent. These solutions were subsequently coated onto glass coverslips, and the thickness of the polymer was assessed using profilometry. Polymer degradation and dissolution was assessed using brightfield microscopy, scanning electron microscopy, and gel permeation chromatography. The ability of the polymers to protect cortical neurons from free radical insult was assessed using an in vitro cortical culture model.

Main Results:

The P50 curcumin polymer (containing greater poly(ethylene glycol) content than the P75 polymer), eroded readily in solution, with erosion dependent on the weight percentage of polymer in solvent. Unlike the P50 polymer, the P75 polymer did not undergo erosion. Since the P50 polymer underwent erosion, we expected that the P50 polymer would more readily protect cortical neurons from free radical insult. Unexpectedly, even though P75 films did not erode, P75 polymers protected neurons from free radical insult, suggesting that erosion is not necessary for these polymers to enable neuroprotection.

Significance:

This study is significant as it provides a framework to evaluate polymers for future neural applications. Additionally, we observed that some curcumin polymers do not require dissolution to enable neuroprotection. Future work will assess the ability of these materials to enable neuroprotection within in vivo models of neural injury.

1. Introduction

Injury to the central nervous system (CNS) occurs after trauma, such as with traumatic brain injury (TBI) or spinal cord injury (SCI) [1], or through the implantation of neural interfacing devices into the brain or spinal cord [2–3]. As a consequence of these insults, neuronal and glial cell membranes are damaged, leading to neuronal and glial cell death, termed as the primary injury [4–5]. Additionally, the rupture of blood vessels leads to the infiltration of immune cells such as neutrophils and macrophages[6,7], the release of harmful reactive oxygen species (ROS) from cellular debris[8–9], and the recruitment of glial cells within the vicinity of the injury.[9–10] Oxidative stress plays a crucial role in promoting secondary injury following primary injury [11–12] and can lead to additional neural tissue death and possible electrode failure.[10] Thus, in order to mitigate this inflammatory response and quench reactive oxygen species, researchers have begun investigating the use of anti-inflammatory biomaterials to circumvent secondary injury following primary injury [13] or as electrode coating materials.[14]

Curcumin is a naturally occurring compound found in plants of the Curcuma longa species and is the principal and most widely studied curcuminoid of turmeric.[15] Over the last two decades, research has shown curcumin to exhibit diverse benefits[15–19] by acting in anti-inflammatory[20,21], antioxidant[22,23], tissue protective[24–30], chemoprotective[31], antibacterial[32], antiviral[33], and immunomodulatory roles. Although curcumin appears promising as a therapeutic candidate, its effectiveness is limited due to its poor stability and low solubility.[18] To overcome its limited bioavailability, current research has focused on curcumin delivery from biomaterials. Biomaterials can be designed as drug-delivery vehicles and tailored to release therapeutics, such as curcumin, for extended periods within the body at sites of interest. Furthermore, biomaterials can be synthesized from poly(pro-drugs), where the drug of interest is incorporated into a polymer backbone, to increase therapeutic solubility.[34] Thus, local delivery of curcumin via poly(prodrugs) limits potential off-target side effects and increases curcumin efficacy as a lesser amount of drug is required to achieve the desired therapeutic effect.[35]

The poly(pro-curcumin) polymer examined within this work is poly(curcumin-co-polyethylene glycol). Within their study, Chen et al. conjugated curcumin to a polyethylene glycol (PEG) backbone to increase hydrophilicity, thus increasing bioavailability and therapeutic efficacy.[36] A poly(pro-curcumin) library was synthesized consisting of four polymers, Px: P25, P50, P75, and P100, where x is the molar percentage of curcumin in the monomer feed. Additionally, Chen et al. demonstrated that the P25 material, which solubilizes instantly due to its high PEG content, exhibits neuroprotective effects. To that end, the aim of this current study is to characterize the P50 and P75 polymers, which contain higher curcumin content. We report an investigation of the P50 and P75 polymer films at several concentrations to explore how polymer solution concentration plays a role in material dissolution and neuroprotection following simulated neural injury. Polymer films were selected as the biomaterial to study as they replicate electrode coatings or films placed on or near sites of neural injury. By investigating P50 and P75 at multiple concentrations, this study will provide insight into the appropriate material selection for future applications as coating materials for neural interfaces or for direct application to sites of neural injury.

2. Methods

2.1. Poly(pro-curcumin) synthesis

Polymers (P50 and P75) were synthesized as published.[36] Briefly, anhydrous pyridine was added to a solution of curcumin and PEG1000 in dry dichloromethane (DCM). Sebacoyl chloride was added dropwise, and the mixture was then stirred at room temperature. Dissolution-precipitation processes were repeated until the final products were obtained, as confirmed by nuclear magnetic resonance (NMR) spectroscopy (Fig. S1).

2.2. Preparation of poly(pro-curcumin) films by drop casting

P50 and P75 polymers were dissolved in chloroform to create three solutions per polymer: P50 (6% w/w), P50 (8% w/w), P50 (10% w/w), P75 (6% w/w), P75 (8% w/w), and P75 (10% w/w), where (w/w) is Px/CHCl₃. All solutions were protected from light and allowed to stir on a magnetic stir plate for at least one hour before use. Films were fabricated by drop casting 40 μL of each solution onto 15-mm diameter round coverslips. Films were covered to provide protection from light and dried over heat at 25 °C for 45 minutes. After 45 minutes, the films were removed from heat, air dried overnight, then dried under vacuum for 8 hours. To determine the film weight, glass coverslips were weighed before film casting and then again after drying.

2.3. Profilometry

Profile thickness of all films was assessed to determine its role in dissolution and potential degradation. Drop casted films were prepared as above, where three separate solutions per group were used to fabricate films in material triplicate. The thickness of the films as cast was measured using a Veeco Dektak 6M Stylus Profiler equipped with Dektak 32. A 4000 μm region containing both coated and glass surfaces was scanned, using a force of 5 mg, resolution of 0.444 μm/sample, duration of 30 sec, and measurement range of 2620 kÅ hills and valleys. The average step height of the samples was determined at the region where the profile thickness leveled out, excluding thick film edges, and was compared to the height of the glass-only baseline. Three samples were assessed per group, each from a different material replicate, with three measurements taken per sample. Statistical significance between groups was determined using ordinary One-Way ANOVA and Tukey’s multiple comparisons test (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) in GraphPad Prism 10.

2.4. Poly(pro-curcumin) film dissolution and delamination

Poly(pro-curcumin) films were incubated in PBS to determine dissolution and potential degradation throughout the experiment. P50 (6%, 8%, and 10%) and P75 (6%, 8%, and 10%) films were prepared following the protocol above, where three separate solutions per group were used to fabricate films in material triplicate. Films were incubated in 500 μL PBS in 24-well plates at 37 °C in an incubator. After 30 minutes and 25 hours, the films were removed and dried under vacuum. After 48 hours under vacuum, the films were removed and weighed. The individual film weight was subtracted from the corresponding post-incubation film weight per coverslip to determine mass loss after incubation. Poly(pro-curcumin) films were incubated in PBS over 25 hours to match the time scale of the in vitro work. Three separate solutions per group (polymer and concentration) were used to fabricate films in a material triplicate. One material replicate from this experiment was used for further analysis under scanning electron microscopy (SEM) and brightfield microscopy. Statistical significance between groups and time points was determined using repeated measures Two-Way ANOVA with the Geisser-Greenhouse correction and Tukey’s multiple comparisons test (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) in GraphPad Prism 10.

To examine the solubility differences of poly(pro-curcumin) films between PBS and water, additional mass loss experiments were performed in deionized water (DI). Drop casted films from P50 and P75 solutions at 8% (w/w) were fabricated and were incubated in 500 μL DI water at 37 °C, similar to protocols above. After 30 minutes and 25 hours, the films were removed and dried under vacuum. The corresponding supernatants were collected and stored at -20 °C until further use. To determine mass loss after incubation, the individual film weight as cast was subtracted from the corresponding post-incubation film weight per coverslip. One solution per group was used to fabricate films for a single material replicate.

2.4.1. Brightfield microscopy. 

Incubated film scaffold images were acquired using a Leica TCS SP8 STED confocal microscope using a 5× dry objective and transmitted light detector. To obtain images of the coverslips in their entirety, Tile Scan acquisition mode was utilized, with a total of 36 tiles obtained per coverslip with 10% overlap between tiles. Images were merged using the Leica Las X software and were processed using ImageJ.

2.4.2. Scanning electron microscopy (SEM). 

Incubated film scaffolds were secured to a square SEM stub with carbon tape such that all edges of the glass coverslip were covered with carbon tape. A ~0.5-nm layer of Au/Pd was sputter coated onto the film scaffolds using a Technics Hummer V Sputter Coater. An FEI Versa 3D Dual Beam SEM was used to acquire one image at 250× and 1000× magnifications at representative locations on the scaffold. The following imaging parameters were used: 2.0-kV accelerating voltage, 5.0-nm spot size, and 10.0-mm working distance.

2.5. Gel permeation chromatography (GPC) to assess polymer degradation

Post-incubation films and supernatants were analyzed via gel permeation chromatography (GPC) to determine the film and supernatant composition and to assess potential degradation following incubation in PBS. Drop casted films from P50 and P75 solutions at 8% (w/w) were prepared using the protocol above. Films were incubated in PBS at 37 °C for three timepoints (t=0m, t=30m, t=25h), following the protocol from Methods 2.4. The films were removed from PBS and dried under vacuum, while the supernatants from the corresponding films were collected and stored at -20 °C until further use. Changes in the molecular weight distribution (MWD) of P50 (8%) and P75 (8%) films were assessed via GPC in THF, relative to PS standards, at 40ºC. The supernatant aqueous media was thawed and lyophilized to a fine powder, dissolved in THF, and analyzed by GPC.

2.6. Cortical neuron isolation and seeding

The day before dissociation, 24-well plates were coated with poly-L-lysine (PLL) diluted 1:1 in deionized (DI) water (250 μL per well) for 1.5 hours at 5% CO₂ and 37 °C. Wells were then washed with DI water 3 times. Each well received 500 μL plating medium (MEM with 0.6% glucose and 5% fetal bovine serum) and was incubated overnight. Cortical neurons were isolated from E18 Sprague Dawley and cultured according to previously published protocols.[36,37] In brief, cortices were dissected in Hank’s balanced salt solution (HBSS) and incubated for 15 minutes at 37 °C with 0.25% trypsin and 0.2 mg/mL DNase. After 3 HBSS washes, neurons were dissociated by repeated trituration with regular and narrowed glass Pasteur pipettes. After dissociation, the cortical cell suspension was plated at 400,000 cells per well (~210,000 cells/cm²) and incubated for 4 hours at 5% CO₂ and 37 °C. After 4 hours, the medium was replaced with 500 μL neurobasal medium (NBM) stock solution: neurobasal medium containing B27 supplement, 1% penicillin-streptomycin, and 0.5 mM GlutaMAX^™. The cells were then incubated at 5% CO₂ and 37 °C for 7 days without media changes prior to treatment with poly(pro-curcumin) films. All procedures were approved by the Rensselaer Polytechnic Institute Institutional Animal Care and Use Committee (IACUC).

For immunocytochemistry experiments, 15-mm diameter glass coverslips in 24-well plates were coated with 250 μL of the 1:1 PLL:DI water mixture for 1.5 hours at 5% CO₂ and 37°C. After 1.5 hours, the wells containing the coverslips were washed with DI water 3 times, replaced with 500 μL plating medium, and incubated overnight. After dissociation of the cortices, the cortical cell suspension was plated at 400,000 cells per well in 24-well plates onto the PLL-coated coverslips and incubated for 4 hours at 5% CO₂ and 37 °C. After 4 hours, the medium was replaced with 500 μL NBM stock solution (neurobasal medium containing B27 supplement, 1% penicillin-streptomycin, and 0.5 mM GlutaMAX^™) and incubated for 7 days without media changes prior to treatment with poly(pro-curcumin) films.

2.7. Polymer film preparation for cell culture

Poly(pro-curcumin) films were prepared using the protocol above. Once films were removed from vacuum and weighed to obtain a post-cast weight, they were sterilized via ethylene oxide exposure in an Anprolene An74i tabletop sterilizer (Andersen Products, Haw River, NC). After sterilization, films were degassed for at least 3 days in a sterile cell culture hood. Drop casted films were prepared as previously mentioned, where three separate solutions per group were used to fabricate films in material triplicate.

2.8. Hydrogen peroxide insult assay

The hydrogen peroxide (H₂O₂) insult assay, derived from a previously published protocol, was used to determine if the poly(pro-curcumin) (Px, where x is either 50 or 75) films could rescue cortical neurons from an insult of reactive oxygen species.[36] Eight groups were used in the P50 experiment: 1) control NBM, 2) 100 μM H₂O₂ insult, 3) P50 6% film alone, 4) P50 8% film alone, 5) P50 10% film alone, 6) both H₂O₂ insult with P50 6% present, 7) both H₂O₂ insult with P50 8% present, and 8) both H₂O₂ insult with P50 10% present. Similar groups were used in the P75 experiment: 1) control NBM, 2) 100 μM H₂O₂ insult, 3) P75 6% film alone, 4) P75 8% film alone, 5) P75 10% film alone, 6) both H₂O₂ insult with P75 6% present, 7) both H₂O₂ insult with P75 8% present, and 8) both H₂O₂ insult with P75 10% present. In each well, 5 μL of media was first removed to account for the addition of the following treatments: H₂O₂ to the insulted groups and plain deionized water to the non-insulted groups. The H₂O₂ was diluted in sterile deionized water and prewarmed in an incubator for 30 minutes, and 5 μL was added to each well to generate a final concentration of 100 μM in each well. 5 μL of plain deionized water was prewarmed in an incubator for 30 minutes and added to the control and Px film groups. Films were added to each well by floating the coverslip on top of the media with the polymer film side facing into the media. Plain glass coverslips were added to the control and 100 μM H₂O₂ insult groups to control for the altered gas exchange due to the diminished area of the liquid/air interface. For the combined group, H₂O₂ and the Px film were added simultaneously. Wells for each group but with media only were generated for background subtraction. Thirty minutes after addition of treatments, 1/10^th volume of PrestoBlue^® reagent was directly added to wells containing culture medium, including the background wells, and incubated for 24 hours. After 24 hours, 200 μL of media was removed from each well in duplicate and placed in a black 96-well plate. Fluorescence measurements were analyzed using a Tecan Infinite M200 plate reader (λ_ex560/ λs_em590) per the manufacturer’s protocols. Four wells were used per group per biological replicate, with each biological replicate using one material replicate (3 biological and material replicates). A simplified schematic of the hydrogen peroxide insult assay is shown in Fig. 1. Upon analysis of the fluorescence data, all measurements had background fluorescence subtracted using the corresponding media-only wells. Statistical significance between groups was assessed using Kruskal-Wallis ANOVA and Dunn’s post hoc test (*p < 0.05, **p < 0.01, ***p < 0.001) in GraphPad Prism 10.

Figure 1:

Schematic depicting the cortical neuron hydrogen peroxide insult assay and analysis with poly(pro-curcumin) films. Cortical neurons were isolated, dissociated, and cultured for 7 days in vitro (DIV), allowing them to reach stage 5 of neuronal development. After 7 DIV, treatments were added to wells containing neurons, with additional background wells (BG) to account for the fluorescence from media and curcumin films. Finally, 24 hours after treatment addition, cell viability was assessed via PrestoBlue, or cells were stained utilizing immunocytochemistry methods.

2.9. Immunocytochemistry

To visualize how the H₂O₂ and Px films affected cortical neuron morphology, cells were seeded onto PLL-coated coverslips, as previously mentioned, and subjected to the hydrogen peroxide insult assay with the same groups and treatment methods as those in the previous experiment. After 24 hours of exposure to treatments, cells were fixed in prewarmed 4% paraformaldehyde (PFA)/4% sucrose solution diluted in phosphate buffered saline (PBS) for 20 min at 37 °C. After fixing, the cells were washed with PBS 3 times (4 minutes), and membrane permeabilization was initiated through application of 0.1% Triton X-100 in PBS for 11 minutes at room temperature (RT). Cells were then washed with PBS 3 times (4 minutes) and were blocked with 5% (w/v) bovine serum albumin (BSA) in PBS for 1 hour at RT. Next, cells were incubated with the primary antibody solution consisting of NF200 (1:500) in 5% BSA in PBS for 1 hour at RT, followed by 3 PBS washes (4 minutes). The cells were incubated with the secondary antibody solution consisting of Alexa Fluor 594 goat anti-rabbit (1:1000) in 5% BSA in PBS for 1 hour at RT. Cells were quickly washed once with PBS, then incubated in a solution of 1 μg/mL DAPI in PBS for 10 minutes at RT to label all nuclei. Finally, the cells were washed with PBS 3 times (4 minutes). A Leica TCS SP8 STED confocal microscope equipped with Las X was used to acquire Z-series images (step size = 7 μM) of the cells using a 10× objective. Texas Red (TXRD 594) and DAPI fluorescence channels were used for visualization of neurofilament and nuclei, respectively. Images were processed using ImageJ; the DAPI channel had contrast adjusted across all images to increase visualization.

3. Results

3.1. Poly(pro-curcumin) film casting

P50 and P75 films were fabricated using 6%, 8%, and 10% (w/w) (Px/CHCl₃) solutions. Three different weight percentages were selected to determine how film thickness and increased curcumin delivery affected dissolution and neuroprotection. Solution concentrations lower than 6% (w/w) were not chosen due to difficulty with film casting and handling. Furthermore, solution concentrations higher than 10% (w/w) were not utilized due to having potentially cytotoxic effects [38].

3.2. Poly(pro-curcumin) dissolution and delamination

To characterize poly(pro-curcumin) films for experimental applications, we first sought to investigate the dissolution and delamination of films in PBS on the same timescale as the in vitro cortical neuron hydrogen peroxide insult assay. Films were first weighed to determine the average film weight for each concentration group (Fig. S2). Across 3 replicates, P50 6% films weighed 3.51 ± 0.383 mg, P50 8% films weighed 4.53 ± 0.289 mg, and P50 10% films weighed 5.64 ± 0.431 mg (Fig. S2 A). A control group, time = 0m (t=0m), was utilized to ensure that non-experimental films did not break down outside experimental conditions. It is of interest that in all of the groups at t=0m, the control films appear to have gained mass during the time course of the experiment. However, these fluctuations are likely due to scale calibration differences. The final weight of the polymer films, obtained post-incubation, represents the weight of the intact film on the glass coverslip, where the supernatant contained both dissolved P50 and intact, delaminated polymer. P50, which theoretically contains equal amounts of curcumin and polyethylene glycol (PEG), is relatively soluble due to the interactions between water and PEG [39]. This is supported by the mass loss data (Fig. 2A), where P50 6% lost approximately 98% of its mass within 30 minutes of incubation, with no additional mass loss over 25 hours. Additionally, P50 8% lost approximately 70% of its mass within 30 minutes of incubation and 84% of its mass within 25 hours, while P50 10% lost less of its mass than the lower concentrations: approximately 50% within 30 minutes and after 25 hours.

Figure 2:

(A) P50 and (B) P75 film mass loss (%) following incubation in PBS over 30 minutes and 25 hours. t=0m timepoint indicates non-incubated films, where the fluctuations in weight display scale variability. Statistical significance between groups and time points was determined using repeated measures Two-Way ANOVA with the Geisser-Greenhouse correction and Tukey’s multiple comparisons test (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) in GraphPad Prism 10.

P75 films were similarly weighed to determine the average film weight for each group of concentrations (Fig. S2 B). On average, across 3 replicates, P75 6% films weighed 3.59 ± 0.272 mg, P75 8% films weighed 4.50 ± 0.371 mg, and P75 10% films weighed 5.60 ± 0.293 mg. As with P50 film controls at t=0, P75 control films display a slight increase in mass throughout the experiment. P75 contains a higher amount of curcumin than PEG; although water interacts slightly with the PEG molecules within the polymer, it is highly hydrophobic and insoluble due to the chemical structure of curcumin. This is notable across all concentrations, as P75 6%, P75 8%, and P75 10% all had a slight mass increase, likely due to scale variability or the lack of a rinse-step following film removal from PBS, or no loss in mass over 25 hours (Fig. 2B).

To investigate differences in curcumin solubility between solutions, mass loss for P50 8% and P75 8% films was also assessed in deionized (DI) water. Films were first weighed to determine the average film weight for each concentration group (Fig. S3). Dissimilar to the high rate of dissolution and delamination seen with PBS, P50 8% films incubated in DI water showed a mass loss of about 40% after 30 minutes and 25 hours (Figure S4 A). Furthermore, P75 8% films lost about 5% mass across time points, confirming their highly insoluble properties (Figure S4 B). It is well-known that the solubility of curcumin is solvent dependent, in that curcumin is extremely insoluble in water and is frequently reported to have a solubility of <8 μg/mL, while it has a higher solubility in PBS, pH 7.4 (3.0 ±0.56 mg/mL).[40–41] This difference in solubility may contribute to differences seen between mass loss data in PBS and DI water.

3.2.1. Brightfield microscopy. 

Whole coverslips were assessed via brightfield (BF) imaging following the completion of the degradation time course to supplement mass loss data. P50 BF images (Figure S5) showcase films as cast at t=0m for P50 6% (S5 A), P50 8% (S5 D), and P50 10% (S5 G), where no breakdown has occurred. Due to the lack of visual difference between the P50 6% films at t=30m (S5 B) and t=25h (S5 C), it can be inferred that the polymer is almost entirely solubilized within 30m. At t=30m, P50 8% (S5 E) had already undergone a noticeable loss of the polymer film, with nearly all of the polymer dissolved into PBS solution within 25 hours (S5 F). Lastly, P50 10% films after 25 hours (S5 I) show increased speckling compared to films at t=30m (S5 H), suggesting that this polymer may be undergoing slight bulk dissolution, although not to the same extent as P50 6% or 8%, as well as surface erosion.

P75 coverslips over the degradation time course were also assessed under BF (Fig. S6). Unlike P50, P75, across all concentrations and time points, displays little differences in the films. One feature of note is as the films incubate over time (S6 F, H, I), cracking occurs within the surface.

3.2.2. Scanning electron microscopy. 

To gain insight into the dissolution mechanism of the P50 films, incubated films were visualized via SEM (Figure 3). In non-degraded control films (t=0), P50 6% (Fig. 3A) and 10% (Fig. 3G) appeared to have textured, pitted surfaces; however, P50 8% (Fig. 3D) contained globular-like structures. Additionally, after 30 minutes, the spherical structures appeared across all P50 concentrations (Fig. 3B, 3E, 3H). As hinted at by the mass loss data, an extremely low particle density in both P50 6% (t=25h, Fig. 3C) and P50 8% (t=25h, Fig. 3F) appear to undergo bulk dissolution. Due to the lack of a wash step following incubation in PBS, salt artifacts are present within SEM images (Fig. 3C, E, F). P50 10% likely undergoes a slightly different dissolution mechanism than the other two concentrations. At both the t=30m (Fig. 3H) and t=25h (Fig. 3I) time points, gaps between the particle structures indicate that PBS may be diffusing throughout the film; however, the particle structures are highly present, dense, and contain several layers. Thus, P50 10% may undergo a combination of bulk and surface erosion. Changes in surface characteristics were observed quickly following exposure to PBS (t = 30 min) and these changes remained intact even after 24 hours of exposure. P75 films were visualized at the same magnification used to assess the P50 films via SEM to assess degradation (Fig. 4). P75 films, across all concentrations and time points, reveal very little to no differences in surface characteristics following incubation in PBS over 25 hours.

Figure 3:

Dissolution of P50 generates spherical particles as visualized by SEM at 250x and 1000x (inset). (A-C): P50 6% at t=0m, t=30m, t=25h. (D-F): P50 8% at t=0m, t=30m, t=25h. (G-I): P50 10% at t=0m, t=30m, t=25h. Bulk dissolution generates individual or small particle clumps, while combinatorial dissolution via bulk and surface erosion leads to interruption in film surface. Scale bars = 150μm and in inset = 50μm.

Figure 4:

Dissolution of P75 visualized by SEM at 250x and 1000x magnification (inset). (A-C): P75 6% at t=0m, t=30m, t=25h. (D -F): P75 8% at t=0m, t=30m, t=25h. (G-I): P75 10% at t=0m, t=30m, t=25h. Scale bars = 150μm and in inset = 50 μm.

3.3. Profilometry

To determine if film thickness contributed to polymer dissolution, the profile thickness of the P50 (Fig. 5) and P75 (Fig. 6) films was determined in their as-cast state. The films cast from the P50 6% solution exhibited a thickness of 45.1 ± 3.00 μm (mean ± standard deviation), while the films cast from the P50 10% solution (77.7 ± 15.7 μm) were significantly thicker than both P50 6% (****p< 0.0001) and P50 8% films (**p< 0.01). Films from the P50 8% solution, which displayed an intermediate thickness at 61.4 ± 4.39 μm, were significantly thicker than P50 6% films (**p< 0.01) (Fig. 5).

Figure 5:

Profilometry results as determined by average step height calculations with glass baseline. Error bars indicate mean ± standard deviation; a single data point represents one reading. Statistical significance between groups was determined using ordinary One-Way ANOVA and Tukey’s multiple comparisons test (*p < 0.05, **p < 0.01, ***p < 0.001, **** p < 0.0001).

Figure 6:

Profilometry results as determined by average step height calculations with glass baseline. Error bars indicate mean ± standard deviation; a single data point represents one reading. Statistical significance between groups was determined using ordinary One-Way ANOVA and Tukey’s multiple comparisons test (*p < 0.05, **p < 0.01, ***p < 0.001, **** p < 0.0001).

The films cast from the P75 solutions displayed similar trends to those of the P50 films, with the exception of P75 10%. P75 6% films exhibited a thickness of 40.0 ± 4.23 μm (mean ± standard deviation). The films cast from the P75 8% solution (59.8 ± 4.78 μm) and P75 10% solution (64.4 ± 3.46 μm) were significantly thicker than the P75 6% films (****p< 0.0001) (Fig. 6).

3.4. Gel permeation chromatography

Following P50 8% film incubation in PBS, substantial changes in the GPC chromatogram were observed (Figure 7A). The as-cast P50 film (t=0m) showed a peak MW of 17,000 g/mol, with a broad multimodal distribution ranging from ~10⁵ to 10³ g/mol. After 30 minutes of incubation (t=30m), in PBS the intact polymer film showed a similarly broad GPC trace; however, the peak MW shifted substantially to a lower MW of 5,000 g/mol. A very similar peak MW is seen after 25 hours relative to the t=30m peak. Hydrolysis of polyesters typically proceeds on timescales of weeks to months.[42] Thus, the changes in the GPC trace at t=30m, are almost certainly not due to chain scission via hydrolysis. It was previously shown that P50 is partially soluble in PBS buffer.[36] Hence, we suspect that the changes in the GPC trace are due to solubilization of a fraction of the total P50 sample. It is important to note that these materials are statistical copolymers of curcumin and PEG, which inherently must possess dispersity, not only in terms of chain length, but also in terms of their copolymer composition (i.e., P50 contains 50 mol% curcumin on average, but the distribution of compositions is putatively quite broad).

Figure 7:

GPC traces of solvent-cast thin films of P50 8% (A) after incubation in PBS at 37°C for 30 min and 25 hrs, compared to that of the as-cast films (t=0m), and GPC traces of the supernatant aqueous media in which P50 films were incubated (B). The data represent normalized refractive index detector response, with the traces offset vertically for clarity.

In that light, it is reasonable to speculate that the subpopulation of P50 which dissolves in water is the more PEG-rich subpopulation within the broad distribution. That notion is further supported by the GPC trace of the supernatant media: after 30 minutes of film incubation in PBS, the supernatant contains a broad MWD with a peak MW of 7,000 g/mol (Figure 7B). Thus, the more curcumin-rich fraction of the whole sample remains in the intact polymer film sample.

On the contrary, the P75 8% films displayed very little change in the GPC data after incubation in PBS for 30 minutes and 25 hours (Figure 8A). Whereas P75 is also a broad distribution of copolymer compositions, it skews much more heavily toward the more curcumin-rich compositions. Thus, it is unsurprising that these materials are almost completely water-insoluble and therefore the MWD remains unchanged upon short times of incubation. Interestingly, a small amount of polymer was detected in the supernatant media with a peak MW of ~1,000 g/mol (Figure 8B).

Figure 8:

GPC traces of solvent-cast thin films of P75 8% (A) after incubation in PBS at 37ºC for 30 min and 25 hrs, compared to that of the as-cast films (t=0m), and GPC traces of the supernatant aqueous media in which P75 films were incubated (B). The data represent normalized refractive index detector response, with the traces offset vertically for clarity.

3.5. Cytotoxicity and hydrogen peroxide rescue assay

The toxicity of the P50 and P75 films across all concentrations, as well as their ability to quench hydrogen peroxide reactive oxygen species to protect neurons, was investigated in 7 DIV primary rat cortical neuron cultures. Neurons were cultured for a period of seven days to create dense cultures which recapitulate the dense and functional neuronal network observed in the cortex.[43] Futhermore, our culture system is devoid of glia since glia in culture can protect neurons from free radical damage.[44] Thus, these initial cellular assays are meant to examine the neuroprotective benefit of neurons only. To ensure film-casting consistency across all experiments, P50 (Fig. S7 A and Fig. S8 A) and P75 (Fig. S7 B and Fig. S8 B) coverslips for all concentrations were weighed prior to film casting and after being removed from vacuum following casting. With the presence of P50 6%, 8%, and 10% films compared with the control, there was little to no loss of fluorescence (Figure 9B), indicating that the polymer is nontoxic at these concentrations. Additionally, the 100 μM hydrogen peroxide insult significantly decreased neuronal viability (****p< 0.0001), while the application of P50 films across all concentrations with the insult led to a considerable increase in cell viability, indicating neuronal rescue (Figure 9B). Addition of P50 to cultures did not appreciably affect the fluorescence intensity of the Prestoblue assay with 6% treatment being slightly above the control (2% increase in fluorescence intensity) and 8% and 10% treatment being slightly below the control sample (2% and 7% decrease in fluorescence intensity respectively). H₂O₂ insult reduced fluorescence intensity by 41% compared to the untreated control sample. Treatment with 6%, 8%, or 10% P50 in the presence of H₂O₂ increased fluorescence intensity by 33%, 17%, and 9% compared to cultures treated by H₂O₂ alone. However, P50 treatment in the presence of H₂O₂ did not restore fluorescence intensity to levels observed in the untreated control (no H₂O₂), and the fluorescence intensity remained 22%, 32%, and 36% lower when 6%, 8%, or 10% P50 films were used to circumvent H₂O₂-induced damage respectively.

Figure 9:

P50 films rescue cortical neuron viability following hydrogen peroxide insult. (A) Representative images of 7 DIV primary rat cortical neuron cultures after receiving treatments: water (control), 100 μM H₂O₂, a P50 film (6%, 8%, or 10%), or both 100 μM H₂O₂ and a P50 film (6%, 8%, or 10%) simultaneously. Cultures were stained for neurofilament (green) and nuclei (blue). Scale bars = 250 μm. (B) Graph of mean arbitrary fluorescence units after 24 hours of incubation with PrestoBlue viability reagent. Error bars indicate ± standard deviation; a single data point represents fluorescence from one culture well. Statistical analysis was assessed with Kruskal-Wallis ANOVA and Dunn’s post hoc test (*p< 0.05, **p< 0.01, ***p< 0.001, ****p< 0.0001).

Although the cultures quantified with PrestoBlue were unable to be stained with immunocytochemistry, cultures that underwent the same experimental conditions were visualized. The cultures depict a loss of neuronal structure and integrity in the hydrogen peroxide insult group, demonstrated by a lack of neurofilament and nuclei, compared to the control, and an increase in neuronal integrity following application of P50 films to insult groups (Figure 9A).

Similar to the P50 films, there was no statistically significant change in cell viability with the presence of P75 6%, 8%, and 10% films compared to the control (Figure 10), which indicates that the polymer is nontoxic at these concentrations. The hydrogen peroxide insult decreased neuronal viability from the control value; however, unlike the P50 experiments, this change was not statistically significant. Like P50 films, P75 films at all concentrations were able to rescue the neurons from the hydrogen peroxide insult to some extent, as there was an increase in fluorescence with the film compared to the group that only received hydrogen peroxide Addition of P75 to cultures did not appreciably affect the fluorescence intensity of the Prestoblue assay with 6%, 8%, and 10% treatment being slightly above the control (4%, 5%, and 1% respectively). H₂O₂ insult reduced fluorescence intensity by 33% compared to the untreated control sample. Treatment with 6%, 8%, or 10% P75 in the presence of H₂O₂ increased fluorescence intensity by 36%, 22%, and 16% compared to cultures treated by H₂O₂ alone. However, P75 treatment in the presence of H₂O₂ did not restore fluorescence intensity to levels observed in the untreated control (no H₂O₂), and the fluorescence intensity remained 8%, 18%, and 22% lower when 6%, 8%, or 10% P75 films were used to circumvent H₂O₂-induced damage respectively (Figure 10B).

Figure 10:

P75 films rescue cortical neuron viability following hydrogen peroxide insult. (A) Representative images of 7 DIV primary rat cortical neuron cultures after receiving treatments: water (control), 100 μM H₂O₂, a P75 film (6%, 8%, or 10%), or both 100 μM H₂O₂ and a P75 film (6%, 8%, or 10%) simultaneously. Cultures were stained for neurofilament (green) and nuclei (blue). Scale bars = 250 μm. (B) Graph of mean arbitrary fluorescence units after 24 hours of incubation with PrestoBlue viability reagent. Error bars indicate ± standard deviation; a single data point represents fluorescence from one culture well. Statistical analysis was assessed with Kruskal-Wallis ANOVA and Dunn’s post hoc test (*p< 0.05, **p< 0.01, ***p< 0.001, ****p< 0.0001).

Immunostained cultures show increased neuronal integrity via increased neurofilament fluorescence with film application to the insulted groups compared to the group that only received hydrogen peroxide (Figure 10A).

4. Discussion

Overall, the work presented in this manuscript reveals that although the P50 and P75 polymers exhibit similar neuroprotective effects, they do so via distinct mechanisms. When incubated in PBS, P50, which theoretically contains equal amounts of curcumin and PEG, is relatively soluble due to the interactions between PBS and PEG (Fig. 2A). The loss of solubility of the films as the casting solution concentration increased aligns with current literature. [45–46] It is well known that at a higher polymer concentration, more polymer molecules are present leading to an increase in chain overlap and entanglement within the solution.[45] Greater chain entanglement causes an increase in viscosity hindering solvent evaporation, resulting in greater layer thickness (Fig. 5). Furthermore, it has been noted that as a polymer film becomes thicker, there is a transition from bulk erosion to surface degradation due to water molecules having a reduced ability to penetrate the entangled polymer chains.[46] Although there were differences in mass loss data between solvents, PBS and DI water, we elected to highlight experiments done within PBS, as it more closely reflects in vitro experiments. Within the cell culture experiments, neurobasal medium (NBM) is utilized, which contains a myriad of inorganic salts, including calcium chloride and potassium chloride. PBS is a buffered solution, which also contains calcium chloride and potassium chloride. Thus, the results from experiments with PBS more closely replicate the in vitro film dissolution and delamination.

Scanning electron microscopy results reveal particular structures within the P50 films across all concentrations (Fig. 3.2). These structures have been previously reported with the P50 polymer and are due to the self-assembling nature of the copolymer.[36] The starting compounds in the monomer feed, curcumin and polyethylene glycol (PEG), have two different reactivities in the presence of sebacoyl chloride.[36] Although PEG contains an alcohol group which is generally considered a better nucleophile than the phenolic group in curcumin, the reaction was carried out with undissolved, waxy PEG, causing the curcumin to be preferentially consumed in the primary stages of the copolymerization reaction. This led to the formation of a gradient polymer sequence, as the curcumin was consumed until the PEG had melted and was able to partake in the reaction. Previous work has detailed the ability of gradient copolymers to self-assemble in selective solvents, where insoluble blocks drive the formation of different morphologies to stabilize the aggregates.[47]

Alternatively, P75, which contains more curcumin than PEG, is highly hydrophobic and insoluble due to the structure of curcumin. As expected, there was little to no mass loss following incubation in PBS (Fig. 2B). SEM images further revealed little differences in the P75 films, across all time points and concentrations (Fig. 4). One feature of note is cracking occurs within the surface as films incubate over time (Fig. S6). The cracking features may be attributed to the di-block copolymer structure. Previous literature has revealed that the combination of a high amphiphilic co-polymer concentration, as with P75 8% and P75 10%, and a high hydrophobic polymer concentration promotes the formation of disrupting effects such as cracks.[48] Thus, although P75 film thickness (Fig. 6) was similar to that of P50, different dissolution characteristics are present due to the nature of the polymers themselves. Interestingly, although P75 appears to have very little to no dissolution over 25 hours, it has slightly greater neuroprotective effects across all concentrations than P50 (Fig 10 and 9, respectively). This is likely due to the higher amount of curcumin present in the polymer; although the film is unable to break down to release curcumin, it quenches the hydrogen peroxide molecules at the film’s surface.

These polymers can be explored in a myriad of applications. Previous studies have examined the benefits of implementing curcumin into hydrogel biomaterials for spinal cord injury applications. Luo et al. demonstrated that their FC-FI-Cur (Fmoc-grafted chitosan and Fmoc peptide with curcumin) hydrogel modulated inflammatory responses as well as promoted remyelination 14 days after SCI in vivo.[49] Additionally, using a peptide hydrogel containing polypyrrole fibers, human-induced neural progenitor cells, and nanocurcumin, Elkhenany et al. showed decreased scar tissue formation and neuropreservation in an in vivo SCI model.[50] Finally, Requejo-Aguilar developed a polyacetal-curcumin conjugate, which led to a decrease in the inflammatory response and enhanced functional recovery in vivo.[51] The P75 polymer films within this study are free-standing and easy to handle. Due to its ease of application and promising neuroprotective effects, P75 would be an ideal candidate for SCI applications.

To our knowledge, two studies have investigated curcumin directly in cortical electrode applications. Potter et al. coated probes with a 3% (w/w) curcumin/poly(vinyl alcohol) (PVA) composite, which aided in maintaining high neuron populations following implantation.[52] Although a greater neuron density, as well as a more stable blood-brain barrier, were observed in vivo following implantation compared to PVA-only probes, in vitro studies reveal most of the curcumin was released from the construct within 10 hours. Furthermore, the nanocomposites released only ~25% of the curcumin content into the medium. To mitigate a burse release of curcumin from its carrier system, Ziemba et. al developed a poly(curcumin-PEG carbamate) (PCPC) microelectrode coating.[53] Within this work, the PCPC coating reduced the mechanical mismatch between the probe and brain tissue, although there was insufficient curcumin release, as noted by the lack of tissue preservation upon implantation.

The current study provides a framework for characterizing drug containing polymers for neural applications. While the neuron study presented here clearly demonstrates the ability of the polymers to provide neuroprotection following free radical insult, the culture is devoid of glia and immune cells and analyzes neuroprotection at a specific time point using a defined concentration of free radicals. As such, application of the polymers to appropriate in vivo models of neural injury would provide more insight into the neuroprotective potential of these polymers. Despite these limitations, the P50 and P75 constructs examined within this work highlight the following: both assist in neuron preservation upon ROS insult and provide neuroprotection either by polymer dissolution (P50) or without dissolution (P75). From analyzing the results of this work, one may propose applying the P50 polymer immediately following neural injury, such that the fast dissolution characteristics enable neuroprotection. Conversely, the P75 polymer demonstrates resilience to dissolution while still enabling neuroprotection, and as such may be appropriate for implants residing in the nervous system over chronic timeframes.

5. Conclusion

In this work, we characterized and assessed the dissolution and neuroprotective effects of two previously synthesized poly(pro-curcumin) polymers at several concentrations. The film incubation studies within this work demonstrate that hydrophilicity plays a major role in dissolution kinetics, while film thickness plays a more minor role. The more hydrophilic polymer dissolves at a much higher rate than the more hydrophobic polymer; additionally, as the polymer concentration increases, increasing the film thickness, the rate of dissolution slows. Both P50 and P75 polymer films, across all concentrations, demonstrated no cytotoxicity and increased neuronal viability following free-radical insult. Furthermore, representative images stained against neurofilament revealed the maintenance of neuronal integrity in the hydrogen peroxide insulted wells with polymer films. P50 was hypothesized to have greater neuroprotective effects due to its higher solubility; unexpectedly, P75 outperformed P50 in our neuroprotective studies, suggesting that the higher curcumin presence is sufficient without high solubility. Thus, although the films seemingly do not release curcumin from the films, they likely exhibit neuroprotective effects through surface quenching. The library of P50 and P75 concentration variants illustrates the ability to tune the poly(pro-curcumin) materials to the desired biomedical application; by characterizing the dissolution and neuroprotective effects at each concentration, future researchers will be able to select the material and concentration that best suits their application’s needs.