Fragmentation of Injectable Bioadhesive Hydrogels Affords Chemotherapeutic Macromolecules

Abstract

Implantation of drug delivery depots into or proximal to targeted tissue is an effective method to deliver anticancer drugs in a sustained localized manner. Herein, syringe-injectable polydextran aldehyde (PDA)-based bioadhesive gels are prepared that can locally deliver cytotoxins upon their hydrolytic fragmentation. Adhesive gels are formed by mixing doxorubicin (DOX)-functionalized PDA (DOX-PDA) and bovine serum albumin (BSA) using a dual-barrel syringe. Upon mixing and delivery, the DOX-PDA reacts with the cross-linker BSA as well as the extracellular matrix via imine bond formation to define the cohesive and adhesive properties of the gel, respectively. Resulting gels are mechanically rigid (∼10 kPa) and adherent (adhesive stress ∼ 4 kPa). Once formed, the DOX-PDA-BSA gels undergo slow hydrolytic degradation (>2 months) locally releasing free DOX and DOX-PDA as expected. Surprisingly, we found that macromolecules composed of DOX, PDA, and BSA are also released from the bulk material. These DOX-PDA-BSA macromolecules, along with free DOX and DOX-PDA conjugate, are internalized by A549 lung carcinoma cells, resulting in potent cell death.

Graphical Abstract

INTRODUCTION

Local delivery of anticancer drugs to targeted tissues is an effective approach to improve safety, stability, and tumor accumulation of otherwise nonspecific chemotherapeutics.¹ Implantation of delivery depots directly into diseased tissue is one method to accomplish local delivery.^2–7 We had recently reported the design of an inherently antibacterial wound-filling bioadhesive hydrogel that could be delivered to tissue by dual-barrel syringe.⁸ The gel is formed by mixing solutions of polydextran aldehyde (PDA) and branched polyethylenimine (PEI). The aldehyde-containing PDA undergoes imine bond formation with both the amine-rich PEI to initiate gel formation and with the natural amine content of the extracellular matrix to adhere the material to the tissue into which it is injected. Conceptually, the design of this gel material is modular in that nearly any amine-containing macromolecule can be used to cross-link the PDA to form the gel. In addition, small molecule amines can be used to decorate the gel matrix to endow functionality.

Herein, we prepare doxorubicin (DOX)-functionalized PDA bioadhesive gels using bovine serum albumin (BSA) as a cross-linking agent. BSA is a 66 kDa protein commonly used as an excipient in drug formulation that contains 30−35 reactive amines.⁹ DOX is an anthracycline antitumor antibiotic, widely used for cancer therapy due to its wide spectrum of anticancer activity.¹⁰ Importantly, DOX contains a primary amine on its daunosamine ring that can also be used for imine bond formation with PDA.^11–14 DOX-PDA-BSA gels are prepared by first functionalizing PDA with DOX to form a DOX-PDA conjugate. Ligation of drugs to polymers, including dextran, has been reported to reduce toxicity, improve stability and half-life, and enable accumulation in tumor tissues by passive targeting.^11,15–17 Herein, we ligated DOX to PDA to afford a prepolymer that can be cross-linked and is reactive toward tissue. By using substoichiometric amounts of DOX (relative to PDA aldehyde content), the DOX-PDA conjugate is left with residual aldehyde moieties that can be used for reaction with BSA and the extracellular matrix. Gels are formed by simple codelivery of the DOX-PDA conjugate with BSA, as outlined in Figure 1A. As will be shown, DOX-PDA-BSA bioadhesive gels undergo slow hydrolysis to liberate free DOX and DOX-PDA conjugate. Unexpectedly, we found that macromolecules composed of DOX, PDA, and BSA are also released from the bulk gel, presumably by hydrolytic events that lead to material fragmentation, Figure 1B. These DOX-PDA-BSA macromolecules, along with free DOX and the DOX-PDA conjugate, can enter A549 lung carcinoma cells in vitro to induce cell death.

Figure 1.

(A) Schematic illustration of DOX-PDA-BSA bioadhesive gel formation. (B) Hydrolytic degradation of a DOX-PDA-BSA bioadhesive gel.

MATERIALS AND METHODS

Materials.

T-25 dextran was purchased from Pharmacosmos. Doxorubicin hydrochloride (DOX) was purchased from AvaChem Scientific. Sodium periodate, diethylene glycol, and bovine serum albumin (BSA) were purchased from Sigma.

Preparation of PDA.

PDA (M_n = 15.5 kDa; M_w/M_n = 1.4) was prepared as previously reported.⁸ Briefly, T-25 dextran (5 g, 30.9 mmol glucose monomer) was dissolved in water (150 mL). Then sodium periodate (5.23 g, 24.5 mmol) in water (150 mL) was added to the dextran solution and stirred for 24 h at room temperature. The reaction was quenched with the addition of diethylene glycol (2.8 mL) and stirred for 2 h. Then the reaction mixture was dialyzed (MWCO 12.3 kDa) against water over 3 days and lyophilized, affording a white fluffy powder. The percent oxidation of PDA was determined by both colorimetric analysis (52%) and ¹³C NMR (51%) as previously reported.⁸

Preparation of DOX-PDA Conjugate.

A 20 wt % PDA stock solution was first prepared by dissolving 200 mg of PDA in 800 μL of 2× phosphate buffered saline (20 mM phosphate, 300 mM NaCl, pH 7.4). A portion of the PDA stock solution (500 μL) was mixed with an equal volume (500 μL) of DOX dissolved in water (16 mg/mL, 29.4 mM). The mixture was allowed to react at 37 °C for 24 h. The percent conjugation was determined by ultrafiltration and associated absorbance measurements of unconjugated DOX at 480 nm. The amount of DOX used to prepare the DOX-PDA conjugate can be varied; thus, the final composition of DOX and PDA within the DOX-PDA conjugate can be defined by the final concentrations of DOX and PDA used to prepare the conjugate. For example, the DOX-PDA conjugate prepared above can be defined as DOX(8 mg/mL)-PDA(10 wt %). This nomenclature will be used throughout the manuscript. Solutions of DOX-PDA conjugate prepared by mixing DOX and PDA can be used directly to prepare the final DOX-PDA-BSA bioahdesive gel as described below.

Preparation of DOX-PDA-BSA Gels.

Gels can be prepared using DOX-PDA solutions generated by mixing PDA with DOX without isolating the DOX-PDA conjugate as described above. For example, 200 μL of a DOX(8 mg/mL)-PDA(10 wt %) solution can be mixed with 200 μL of 20 wt % BSA stock solution (PBS, pH 7.4) and allowed to gel at 37 °C for 24 h, resulting in a DOX(4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) hydrogel.

Dynamic Oscillatory Rheology.

Oscillatory rheology experiments were performed on an ARG2 rheometer (TA Instruments) using a 25 mm stainless steel parallel plate geometry. DOX(4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) gels were formed directly on the rheometer as follows. A solution of DOX(8 mg/mL)-PDA(10 wt %) conjugate (200 μL) and an equal volume (200 μL) of 20 wt % BSA stock solution were mixed and then 300 μL of the mixture was transferred to the rheometer plate and the tool lowered to a gap height of 0.5 mm. Standard S6 oil was placed around the tool to prevent evaporation during the measurements. A dynamic time sweep was performed to measure the evolution of storage modulus (G′) and loss modulus (G″) at an angular frequency of 6 rad/s and 0.2% strain at 37 °C for 6 h. After the time sweep, a dynamic frequency sweep (0.1−100 rad/s at constant 0.2% strain) and strain sweep (0.1−1000% strain at constant 6 rad/s) were performed to ensure that the time sweep data was collected in the linear viscoelastic regime, Figure S2. Rheological experiments were conducted in triplicate.

Maximal Adhesive Stress Determination with Porcine Skin.

The maximal adhesion stress of a DOX(4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) gel was determined utilizing a G2-RSA (TA Instruments) dynamic mechanical analyzer using a porcine skin adhesion model. Porcine skin was purchased from Wagner’s meats (Mt. Airy, MD), and the fat was removed from the dermal tissue layer. Skin sections were subsequently cut to approximately 2 × 6 cm. The sectioned skin was soaked in PBS at 4 °C overnight and then allowed to warm to room temperature prior to adhesive testing. A 75 μL solution of DOX(4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) was freshly prepared as described above and quickly applied to a distal 2 × 2 cm area of one of the tissue sections. Then, the distal portions (2 × 2 cm) of the two skins were brought into contact and the gel allowed to set at 37 °C for 2 h in an incubator. Using the G2-RSA, a tensile load was applied to the sample at a rate of 0.10 mm/s and the adhesive stress was monitored. The maximum adhesive stress was considered to be the stress at which the two sample tissue sections became completely separated, concomitant with bond failure. The experiments were conducted in triplicate and the results are presented as an average. The adhesion stress value for a control PDA(5 wt %)-BSA(10 wt %) gel was collected similarly. The fibrin glue (Tisseel) control was determined previously in the lab using the same model and is reported in ref 8.

Analysis of Hydrolytic Degradation Products of DOX-PDA-BSA Bioadhesive Gels.

A DOX(4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) gel (100 μL) was prepared in a glass vial by incubation for 24 h at 37 °C. Then PBS (1 mL) was added onto the gel and the vial was shaken at 37 °C with agitation (100 rpm) for 28 or 56 days to allow hydrolysis. The supernatant at day 28 or 56 was then injected to an Agilent 1200 series analytical HPLC (Vydac C18 peptide/protein column) with solvents consisting of solvent A (0.1% TFA in water) and solvent B (0.1% TFA in 90% acetonitrile) monitoring at both 220 and 480 nm. A linear gradient of 0 to 100% solvent B over 100 min was employed at 40 °C. The resulting chromatogram was compared to control chromatograms of pure DOX, DOX-PDA, and BSA, which eluted at 30, 20−40, and 49 min, respectively. Peaks corresponding to DOX and DOX-PDA conjugate were identified along with an unidentified peak at 46−51 min, which was isolated, lyophilized, and determined to contain macromolecules consisting of BSA, PDA, and DOX via the characterization described below.

Primary Characterization of Macromolecules.

First, the particle size of the species isolated as fraction (46−51 min) was determined by dynamic light scattering (DLS) using a Zetasizer Nano Series instrument (Malvern Instruments Ltd.), as was the zeta potential. Lyophilized powder corresponding to fraction (46−51 min) was dissolved in PBS (0.2 mg/mL, pH 7.4) and 1.0 mL loaded into a disposal DLS cuvette. Correlograms were collected (25 °C, scattering angle = 173° and fit using Malvern’s distribution analysis algorithm to yield the number based size distribution. BSA in PBS (1 mg/mL, pH 7.4) was also examined for comparison. The composition of the macromolecules was determined via a combination of absorbance and fluorescence experiments. First, a UV absorbance spectrum was collected of a solution of macromolecules originally isolated from HPLC in PBS (0.2 mg/mL, pH 7.4, pl = 1 cm) and analyzed. Second, a separate gel void of DOX was prepared, PDA(5 wt %)-BSA(10 wt %), using FITC-labeled BSA (Sigma). The hydrogel was hydrolyzed and macromolecules isolated by HPLC as described above. Fluorescence spectra of a macromolecule solution (1 mg/mL, PBS, pH 7.4) were collected (25 °C, pl = 0.2 cm, λ_ex = 450 nm) on a PTI fluorimeter and analyzed.

Release Profile of DOX-Containing Species from DOX-PDA-BSA Bioadhesive Gels.

DOX(1, 2, or 4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) bioadhesive gels were prepared in glass vials. Then PBS (1 mL) was added onto the gels and the vials were shaken at 37 °C with agitation (100 rpm). At each time point, the supernatant above the gel was removed and replaced with fresh PBS (1 mL). The concentration of DOX-containing species in the removed supernatant was determined by absorbance at 480 nm. Release experiments were conducted in triplicate and the results are presented as the average.

MTT In Vitro Cytotoxicity Assay.

Hydrogel cytotoxicity was assessed using human lung adenocarcinoma A549 cells. A549 cells were maintained with RPMI-1640 medium containing 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin. DOX(0.06−4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) bioadhesive gels (50 μL) were prepared in BD Falcon cell culture inserts (8 μm pore size) and incubated at 37 °C for 24 h. Then, gels were incubated with fresh PBS for an additional 24 h, 28 days, or 56 days at 37 °C. After these incubation times, the cell culture inserts were transferred into an A549 cell-seeded 24-well plate (2.5 × 10⁴ cells/500 μL/well, seeded on the previous day) to assess the activity of each of the gels. After 3 days, cell culture media was removed and 0.5 mg/mL MTT in cell culture media (500 μL) was added. After incubation for 2 h at 37 °C, the media was removed and DMSO (500 μL) was added to lyse the cells. Absorbance of the cell lysates was measured at 540 nm, and relative cytotoxicities of the bioadhesive gels compared to a control (nontreated A549 cells) were obtained. The experiments were conducted in triplicate and the results were presented as the average. Similar experimental protocol was used in separate experiments to access the cytotoxicity of the adhesive toward noncancerous human dermal fibroblasts.

Separate comparative cell proliferation studies measuring the activity of gel supernatant versus free DOX were performed by measuring proliferation as a function of time. A549 cells were seeded in a 96-well plate (5 × 10³ cells/100 μL/well) and allowed to incubate for 24 h, after which time, the media was removed and replaced with either 100 μL of free DOX in solution (4 μg/mL in 10% FBS supplemented RPMI-1640) or 100 μL of gel supernatant (which contains a total of 4 μg of DOX-containing species/mL in 10% FBS supplemented RPMI-1640). Cell viability was measured via MTT at days 0, 1, 2, and 3. The experiments were conducted in triplicate and the results are presented as the average.

Live-Cell Imaging Monitoring Uptake and Localization of DOX and DOX-Containing Species.

For these experiments, DOX(4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) gels were first prepared using FITC-labeled BSA. Gels (50 μL) for each experiment were prepared in glass vials and incubated for 24 h at 37 °C. After incubation, PBS (500 μL) was added onto the gels. The vials were shaken at 37 °C with agitation (100 rpm) for 2 weeks. Either gel supernatant (10 μL) or a free DOX solution (1.25 mg/mL, 10 μL) was added into A549 cell-seeded 96-well plate (5 × 10³ cells/100 μL/well, seeded on the previous day). After a 24 h incubation, 11 μL of a Hoechst 33342 solution in PBS (0.1 mg/mL) was added and incubated for 1 h at 37 °C. Then the cells were rinsed by media two times, and visualized using an EVOS FL Auto Cell Imaging System (Thermo Fisher Scientific Inc.).

Live-Cell Imaging Monitoring Reactive Oxygen Species (ROS) Generation.

A549 cells were seeded in a 96-well plate (5 × 10³ cells/100 μL/well) and allowed to incubate for 24 h. After which time, the media was removed and replaced with either 100 μL of free DOX in solution (4 μg/mL in 10% FBS supplemented RPMI-1640) or 100 μL of gel supernatant (which contains a total of 4 μg of DOX-containing species/mL in 10% FBS supplemented RPMI-1640). After 2 days culture, ROS generation was visualized using an Image-iT LIVE Green Reactive Oxygen Species Detection Kit (Thermo Fisher Scientific Inc.) following the given instructions.

RESULTS AND DISCUSSION

Synthesis of DOX-PDA-BSA Bioadhesive Gels.

Bioadhesive gels can be prepared by mixing solutions of DOX-PDA conjugate and bovine serum albumin (BSA) either via delivery by a dual barrel syringe or by simply mixing in a 1:1 volumetric ratio, Figure 1A. The prerequisite DOX-PDA conjugate is prepared by reacting polydextran aldehyde (PDA) with DOX in aqueous buffer using substoichiometric amounts of DOX relative to the aldehyde content of the PDA. This ensures that sufficient free aldehyde is present in the DOX-PDA conjugate for subsequent reaction with BSA and the extracellular matrix when the bioadhesive is eventually formed. The imine forming reaction of DOX and PDA is nearly quantitative as assessed by ultrafiltration experiments. In these experiments, free DOX is removed from DOX-PDA in the crude solution and quantitatively measured by absorbance. Loading concentrations of 2, 4, and 8 mg/mL of DOX in the coupling reaction afford loading efficiencies of 99, 99, and 98%, respectively (Figure S1). Thus, the resulting solution of DOX-PDA conjugate from the reaction can be used directly in the next reaction with BSA without isolating the conjugate.

Mixing buffered solutions of DOX-PDA conjugate and BSA at pH 7.4 initiates imine bond formation between the lysine side chains of the protein and the aldehydes of the conjugate, with concomitant cross-linking and the onset of gelation. If the mixing is performed in the presence of tissue, the DOX-PDA conjugate also reacts with the ECM, which adheres the gel. The relative amounts of DOX, PDA, and BSA in the adhesive can be varied during the synthesis to alter the material’s mechanical and anticancer properties. Herein, we define the final composition of a given gel by stating the amount of DOX in mg/mL, and PDA and BSA as final wt %, for example, DOX(4 mg/mL)-PDA(5 wt %)-BSA(10 wt %).

Rheological and Adhesive Properties of DOX-PDA-BSA Bioadhesives.

Figure 2A shows a time-sweep oscillatory rheology experiment that follows the change in the storage modulus (G′) and loss modulus (G″) as a function of time after which solutions of DOX-PDA conjugate and BSA have been mixed directly in the rheometer, ultimately forming a DOX(4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) hydrogel. The gel begins to form within about 30 min and continues to stiffen, reaching a modulus of ∼10 kPa after 6 h. The gel will continue to slowly stiffen with longer time suggesting that covalent bond formation defining the gel is still occurring at long times. The value of G′ is over an order of magnitude greater than G″ indicating that a relatively stiff viscoelastic gel had formed. Independent experiments showed that both G′ and G″ were invariant, with a frequency from 0.1 to 100 rad/s, Figure S2A. Strain-sweep experiments show that the gel yields at about 200% strain, Figure S2B. Further, additional frequency-sweep measurements of preformed equilibrated gels showed an equilibrium G′ value of 23.4 kPa, Figure S2C. The rate of gelation for this particular formulation is moderate when compared to other PDA-based gels that use PEI,⁸ ε-polylysine,¹⁸ or multiarm polyethylene glycol (PEG) amine¹⁹ as cross-linker. However, adjusting the amount of either PDA or BSA can significantly increase the rate of gelation for the DOX-PDA-BSA system. Figure 2B shows that by slightly increasing either component by 5 wt %, the gelation time can be decreased to a few minutes. For the characterization experiments performed herein, we found the slower gelation time to be quite convenient.

Figure 2.

(A) DOX(4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) gel formation was monitored as a function of time via time-sweep oscillatory rheology. (B) Time-sweep rheology of different PDA-BSA formulations void of DOX. (C) Adhesive stress measurements of a DOX-PDA-BSA gel, PDA-BSA gel, and fibrin glue as assessed by uniaxial lap-shear measurements employing porcine tissue.

The adhesive ability of the gel was next investigated using a porcine skin model, Figure 2C. Here, a lap-shear analysis was performed during uniaxial loading of a DOX(4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) gel applied between two sections of porcine epidermis, as defined by the American Society for Testing and Materials (ASTM) standard protocol F2255. The maximum adhesive stress is about 4 kPa, which is similar to clinically used fibrin glue. Data for an adhesive void of DOX is also shown, which indicates that DOX loading does not significantly influence the adhesive properties of the material. As reported for other PDA-based materials,⁸ bond failure was largely cohesive in nature as opposed to adhesive failure. Thus, the material is capable of making sufficient interactions with the ECM to ensure its localized placement after injection. Although the material is designed to adhere to ECM, imine bond formation could also take place between PDA and proteins, for example, cytokines, native to the local environment. We currently have no evidence for this mode of reactivity, but it is possible.

Hydrolytic Degradation of the DOX-PDA-BSA Gels.

Imine bond formation is an equilibrium reaction where water can add across the bond to regenerate starting aldehyde and amine components. In these cross-linked gels, these aldehydes and amines can again react to form the original or new imine bonds, or the gels can undergo hydrolytic degradation over time as the imine bonds reversibly sever affording a number of possible components. HPLC was used to investigate the products of hydrolysis. Figure 3A shows a chromatogram of supernatant obtained from a DOX(4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) gel that was allowed to degrade in buffer for two months at 37 °C. The intensely absorbing peak eluting at ∼30 min corresponds to free DOX that had been released from the gel via the hydrolysis of the daunosimine bond appending the small molecule to the material network. A broad peak centered at ∼25 min corresponds to fragments of DOX-PDA conjugate that are released, again from imine bond hydrolysis. Comparative chromatograms of DOX and DOX-PDA verify these assignments, Figure S3. Interestingly, PDA-DOX conjugates have been studied previously. Ueda et al. showed that DOX exhibited higher antitumor activity and lower toxicity in rats when conjugated to PDA.¹¹

Figure 3.

(A) Hydrolytic degradation products of a DOX-PDA-BSA bioadhesive gel assessed by HPLC (absorbance at 480 nm) after 8 weeks. (B) Dynamic light scattering of isolated DOX-PDA-BSA macromolecules and control BSA. (C) UV−visible spectra of isolated DOX-PDA-BSA macromolecules and control BSA. (D) UV−visible spectrum and emission spectrum of isolated PDA-BSA (FITC-labeled) macromolecules.

Finally, a late eluting broad peak was observed at ∼48 min. We initially assigned this peak to free BSA, the protein cross-linker, which has a very similar retention time in the HPLC experiment (Figure S3). However, subsequent DLS experiments show that the material corresponding to this degradation product was comprised of particles characterized by an average size of ∼11 nm in diameter, significantly larger than BSA (∼6 nm; Figure 3B). Further, UV−vis indicated that these macromolecules had a spectral profile distinct from BSA (Figure 3C) that contained an intense absorption at ∼500 nm, reminiscent of the DOX chromophore. Thus, it is likely that the macromolecules contain DOX and PDA but are distinct from the pure DOX-PDA conjugate used in the synthesis of the adhesive. Although the data in Figure 3A resulted from two months of degradation, macromolecules were also observed after one month of degradation as well (Figure S4). The degradation profiles are similar.

We next determined if BSA, in addition to PDA and DOX, could be a component of the macromolecules released from the adhesive via a fluorescence experiment. Here, a separate gel was prepared using FITC-labeled BSA as the cross-linker, but importantly excluded DOX, which emits at a similar wave-length. The resulting gel (PDA(5 wt %)-BSA^FITC(10 wt %)) was allowed to degrade in buffer, and the released macromolecules were isolated. Figure 3D shows that these macromolecules display the characteristic absorbance and fluorescence of the FITC-labeled BSA. Further, the zeta potential of the macromolecule fraction was determined and compared to free BSA. The macromolecules are characterized by an average −21 mV potential, which is very similar to BSA (−24 mV), the cross-linking protein of the bioadhesive and its main contributor to charge since both PDA and DOX are largely neutral (data not shown). Thus, the characterization data in Figure 3 indicate that the DOX-PDA-BSA adhesives degrade via hydrolysis and release free DOX, DOX-PDA conjugate, and macromolecules comprised of DOX, PDA, and BSA. Of note is the narrow size distribution of the macromolecules (Figure 3B) released from the gel, hydrolysis of the material could initially lead to a more broad size distribution of fragments, the larger of which could continue to hydrolyze to a minimally sized particle that is unable to hydrolyze further.

Release Profile of DOX-Containing Moieties from the DOX-PDA-BSA Bioadhesive.

The total release of DOX in the form of free small molecule, DOX-PDA conjugate, and macromolecules from the adhesive was assessed as a function of time, as shown in Figure 4. In this experiment, total DOX release from adhesives formed in cylindrical glass vials into an infinite sink of buffer is measured. Figure 4A shows that when 4 mg/mL of DOX is initially loaded in a DOX-PDA(5 wt %)-BSA(10 wt %) gel, a slow, sustained, linear accumulated release of DOX is realized over several months. Further, the total amount and the rate of DOX release can be modulated by simply adjusting the amount of DOX used in formulating the adhesive. The three different concentrations of DOX-loaded bioadhesives (1, 2, and 4 mg/mL DOX) showed concentration-dependent release rates, 0.6, 1.3, and 2.2 μg/day, respectively. Figure 4B shows the percent of initially loaded DOX that is released as a function of time. Independent of the absolute amount initially loaded, the percent of DOX released remains nearly constant with only about 40% being released after 56 days. Extrapolating this data suggests that the adhesive can deliver DOX over a time course of 4−5 months before exhausting its payload. Figure 4B also shows that the rate of material degradation is constant for each of the formulations and independent of DOX concentration. Thus, although similar amounts of each material have degraded on any given day, the amount of DOX within the released fraction is dependent on the initial DOX loading as shown in Figure 4A. Collectively, the data nicely show that the bioadhesives release DOX in a nearly linear fashion. This is in contrast to many materials that release DOX via a diffusion controlled burst mechanism where the majority of encapsulated DOX is released early. Lastly, HPLC analysis suggests that the majority of DOX is released in the form of both the macromolecule (40%) and the PDA-DOX conjugate (36%). Comparatively, a smaller fraction (24%) of DOX is released in its free form.

Figure 4.

Total DOX released from DOX-PDA-BSA bioadhesive gels as a function of time and initial DOX loading. Total DOX defined as free DOX, DOX-PDA conjugates, and DOX-PDA-BSA macromolecules. (A) Cumulative release measured in mg of total DOX. (B) % total DOX released.

Cytotoxic Activity of the DOX-PDA-BSA Bioadhesive Gels.

The cytotoxic activity of the DOX-PDA-BSA bioadhesives was evaluated using A549 lung carcinoma cells, Figure 5. In this experiment, DOX-PDA(5 wt %)-BSA(10 wt %) gels containing different loading concentrations of DOX were assessed at different stages of their degradation. Specifically, gels were prepared and allowed to undergo hydrolysis for 1, 28, and 56 days in buffer at 37 °C. After which, the gels were transferred to culture plates containing cells where viability was then measured after 3 days. This challenging experiment tests the ability of the adhesive to remain active and capable of delivering DOX during the time course of its degradation. After 1 day of degradation, the bioadhesive kills cells in manner dependent on the initial loading concentration of the drug with the 4 mg/mL formulation being most active and on par with the DMSO positive control. After 28 days of hydrolytic degradation, gels originally containing 4 mg/mL of DOX showed activity nearly identical to freshly prepared gels. After 2 months (56 days), the DOX(4 mg/mL)-PDA(5 wt %)-BSA(10 wt %) gel was slightly less active, but still able to kill over 60% of the cancer cells. Given, the linear release of DOX species (Figure 4), the gel should be as active at 56 days as it is at 28 days. The decreased activity may be due to different forms of DOX having varying activities. While the total DOX release is linear, the proportion of the different forms within the total DOX release may vary with time. Lastly, the adhesive void of DOX was found to be cytocompatible, demonstrating behavior similar to the negative control at 1, 28, and 56 days. Similar experiments were performed using noncancerous human dermal fibroblasts to access whether the adhesives are selective in their cytotoxic behavior. In agreement with the nonselective behavior of free DOX, the adhesives were also nonselective in their action (Figure S5).

Figure 5.

In vitro cytotoxicity of DOX-PDA-BSA bioadhesive gels. The bioadhesive gels were prepared in cell culture inserts that were preincubated with PBS for 1, 28, or 56 days, washed, and then adhesives were transferred to A549-seeded wells. After 3 days, cell viability was measured by MTT assay.

Given that the majority of DOX is released from the adhesive in the form of the macromolecule and DOX-PDA conjugate, it is likely that these species contribute to the material’s anticancer activity. Free DOX exhibits its activity by partitioning to the nucleus and intercalating DNA, inhibiting topoisomerase II-mediated replication as well as generating reactive oxygen species (ROS).^10,20 Live cell imaging was performed to investigate the localization of the DOX-containing species release from the adhesive. The localization potential of free DOX was also evaluated for comparison. Figure 6 shows that free DOX partitions as expected to the nucleus within 24 h after being added to the cell culture media as evident by the red channel fluorescence (DOX) colocalized with the nuclei imaged in the blue channel (see merged data). Next, the supernatant from a hydrolyzed DOX(4 mg/mL)-PDA(5 wt %)-BSA^FITC(10 wt %) gel was added to cells. The supernatant contains free DOX, DOX-PDA conjugate, and macromolecules comprised of DOX, PDA, and FITC-labeled BSA. Interestingly, very little of the red channel fluorescence colocalizes to the nucleus as evident in the merged data. This suggests that the supernatant contains a comparatively low concentration of free DOX, which is in agreement with the HPLC data described earlier. The green channel measures the fluorescence of the macromolecules, which contain BSA^FITC along with DOX and PDA. The macromolecules partition to the cytoplasm of cells. Also evident is that the red fluorescence colocalizes with the green fluorescence in the cytoplasm, as seen in the merged data (yellow). The merged data also displays pure red fluorescence in the cytoplasm, most likely due to DOX-PDA conjugates. Collectively, the imaging data show that the macromolecules along with the DOX-PDA conjugate partition to the cytoplasm where they exert their action. The mechanism of cellular internalization is not yet known, but most likely involves endocytosis given the punctate nature of the fluorescent species within the cells. In fact, uptake of drug−polymer complexes into cancer cells is generally mediated by endocytosis.¹⁷ Given the fact that these species localize to the cytoplasm, their mechanism of action most likely do not involve the direct binding of nuclear DNA. More likely, DOX contained in the macromolecule and in the PDA-conjugate is generating ROS that kills the cells.²⁰ We tested this possibility by assessing the formation of ROS in A549 cells via live-cell imaging. Figure 7A shows control cells in the absence of any DOX-containing species. Their nuclei fluoresce blue (Hoechst). Figure 7B shows cells after the addition of free DOX. The generation of ROS is seen as green fluorescence via the oxidation of the chemical probe, carboxy-DCFH. Finally, Figure 7C shows ROS generation from cells that have been exposed to the supernatant collected from hydrolyzed bioadhesive. This experiment shows that ROS is being generated and is at least, in part, responsible for the cytotoxic action of the material. Lastly, time-course cell proliferation studies were also performed (Figure S6) that show that the species contained in the adhesive’s supernatant were just as active as free DOX. This suggests that nuclear localization is not necessary for potent activity.

Figure 6.

Live-cell imaging monitoring uptake and localization of free DOX and DOX-containing moieties released from DOX-PDA-BSA^FITC bioadhesives in A549 cells. DOX (red), BSA^FITC (green), and nuclei (blue) were visualized after 24 h incubation. Scale bar = 50 μm. All panels at same scale.

Figure 7.

Live-cell imaging monitoring ROS generation. A549 cells were cultured in the absence (A) or presence of free DOX (B) or supernatant of a DOX-PDA-BSA bioadhesive (C). ROS (green) and nuclei (blue) were visualized after 48 h incubation. Scale bar = 100 μm. All panels at same scale.

CONCLUSION

In this study, we evaluate the mechanical and anticancer properties of a class of injectable adhesives prepared from mixing DOX-functionalized PDA conjugate with the cross-linking protein BSA. Resulting gels are moderately mechanically stiff and display adhesive properties similar to fibrin glue. Hydrolysis mediated degradation releases free DOX, DOX-PDA conjugate as expected. However, we unexpectedly discovered that 11 nm macromolecules that are comprised of DOX, PDA, and BSA were also released as a result of hydrolytic fragmentation of the gel network. Adhesive gels release all of these hydrolysis products with a slow sustained, linear release profile, which is capable of killing A549 lung carcinoma cells for at least 2 months. Hydrolysis products released from the gels enter cells within 24 h, localize to the cytoplasm and kill the cell most likely via ROS-mediated mechanisms. Previous in vivo studies in mice on a PDA/PEI-based adhesive developed by our lab showed no evident necrosis and minimal inflammatory response as compared to a sham injection.⁸ Thus, the DOX-PDA-BSA bioadhesive gels reported herein have potential as an injectable therapy for the slow sustained release of DOX-containing cytotoxins.