Electrically Triggered Release of a Small Molecule Drug from a Polyelectrolyte Multilayer Coating

Abstract

Electrically triggered drug delivery represents an attractive option for actively and remotely controlling the release of a therapeutic from an implantable device (e.g., a "pharmacy-on-a-chip"). Here we report the fabrication of nanoscale thin films that can release precise quantities of a small molecule drug in response to application of a small, anodic electric potential of at least +0.5 V versus Ag/AgCl. Films containing negatively charged Prussian Blue (PB) nanoparticles and positively charged gentamicin, a small hydrophilic antibiotic, were fabricated using layer-by-layer (LbL) assembly. When oxidized, the PB nanoparticles shift from negatively charged to neutral, inducing dissolution of the film. Films with thicknesses in the range 100-500 nm corresponding to drug loadings of 1-4 μg/cm(2) were characterized. We demonstrate control over the drug dosage by tuning the film thickness as well as the magnitude of the applied voltage. Drug release kinetics ranging from triggered burst release to on/off, or pulsatile release, were achieved by applying different electric potential profiles. Finally, the in vitro efficacy of the released drug was confirmed against Staphylococcus aureus bacteria. Given the versatility of an external electrical stimulus and the ability of LbL assembly to conformally coat a variety of substrates regardless of size, shape, or chemical composition, we maintain that electrically controlled release of a drug from an LbL-coated surface could have applications in both implantable medical devices and transdermal drug delivery systems.

Figure 1

A) Growth curve, determined via profilometry, for Chi(PB/Chi)₅(PB/GS)_n and Chi(PB/Chi)_n film architectures, revealing accelerated film growth when adhesion layers are deposited. The lines are best fit lines for exponential and linear growth models for films with and without adhesion layers, respectively. Error bars represent ± one standard deviation in measured thickness values at n = 5–7 locations on each film. B) Photographs of Chi(PB/Chi)₅(PB/GS)_n films for n = 0, 25, 50, and 75.

Figure 2

Atomic force microscopy 3D height images of A) n = 25, B) n = 50, and C) n = 75 bilayer films and D) an optical micrograph of an n = 75 film. Film surface roughness increases and clusters form on the film surface with the deposition of an increasing number of bilayers.

Figure 3

Cyclic voltammograms of a Chi(PB/Chi)₅(PB/GS)₂₅ film subjected to multiple cycles at a scan rate of 50 mV/s in a PBS, pH 7.4 electrolyte. A decrease in peak height (and peak area) with subsequent scans reveals a loss of the electroactive Prussian Blue from the film.

Figure 4

A) Absolute and B) normalized thickness of Chi(PB/Chi)₅(PB/GS)_n films over time at an applied potential of +1.25 V vs. Ag/AgCl in a PBS, pH 7.4 electrolyte. The thickest films dissolve more slowly, and all film thicknesses plateau to approximately 25–45% of initial thickness. The lines represent the best fit to a first order exponential decay model. Error bars represent ± one standard deviation in measured thickness values at n = 5–10 locations on each film.

Figure 5

(A) Drug release profiles from Chi(PB/Chi)₅(PB/GS)_n films at an applied potential of +1.25 V vs. Ag/AgCl and at the open circuit potential (OCP). The total amount of released drug, or the drug dosage, can be set by tuning the number of deposited layers, n. Error bars represent ± one standard deviation in measured values from n = 3 films. (B) Linear regression best fits for a pseudo-second order drug release kinetics model for n = 25, 50, and 75 films.

Figure 5

(A) Drug release profiles from Chi(PB/Chi)₅(PB/GS)_n films at an applied potential of +1.25 V vs. Ag/AgCl and at the open circuit potential (OCP). The total amount of released drug, or the drug dosage, can be set by tuning the number of deposited layers, n. Error bars represent ± one standard deviation in measured values from n = 3 films. (B) Linear regression best fits for a pseudo-second order drug release kinetics model for n = 25, 50, and 75 films.

Figure 6

A) Total amount of gentamicin released from a Chi(PB/Chi)₅(PB/GS)₅₀ film over time at different applied potentials. B) Total amount of gentamicin released from a Chi(PB/Chi)₅(PB/GS)₅₀ film in 1 hr at different applied potentials. The smallest amount of drug is released at the open circuit potential of +0.25 V, while increasing amounts of drug are released at both anodic and cathodic potentials, with the greatest amount released during oxidation of the PB. Error bars represent ± one standard deviation in measured values from n = 3 films. All means are statistically different from each other with p < 0.05 except for those at +1.00 and +1.25 V, for which p = 0.15.

Figure 6

A) Total amount of gentamicin released from a Chi(PB/Chi)₅(PB/GS)₅₀ film over time at different applied potentials. B) Total amount of gentamicin released from a Chi(PB/Chi)₅(PB/GS)₅₀ film in 1 hr at different applied potentials. The smallest amount of drug is released at the open circuit potential of +0.25 V, while increasing amounts of drug are released at both anodic and cathodic potentials, with the greatest amount released during oxidation of the PB. Error bars represent ± one standard deviation in measured values from n = 3 films. All means are statistically different from each other with p < 0.05 except for those at +1.00 and +1.25 V, for which p = 0.15.

Figure 7

Drug release profile from a Chi(PB/Chi)₅(PB/GS)₇₅ film with 2 sec pulses of +1.25 V to turn drug release ‘on’, followed by 30 sec pulses at +0.25 V to turn drug release ‘off’. The films are sufficiently stable to allow for on/off, or pulsatile, drug release controlled by the applied potential. Error bars represent ± one standard deviation in measured values from n = 3 films.

Figure 8

Results of a microdilution assay of gentamicin released from a Chi(PB/Chi)₅(PB/GS)₇₅ film against S. aureus bacteria. The MIC of the drug released from the film corresponds well with that of the free drug. Error bars represent ± one standard deviation in measured values from n = 3 samples.