Unraveling the interaction between doxorubicin and DNA origami nanostructures for customizable chemotherapeutic drug release

Abstract

Doxorubicin (DOX) is a common drug in cancer chemotherapy, and its high DNA-binding affinity can be harnessed in preparing DOX-loaded DNA nanostructures for targeted delivery and therapeutics. Although DOX has been widely studied, the existing literature of DOX-loaded DNA-carriers remains limited and incoherent. Here, based on an in-depth spectroscopic analysis, we characterize and optimize the DOX loading into different 2D and 3D scaffolded DNA origami nanostructures (DONs). In our experimental conditions, all DONs show similar DOX binding capacities (one DOX molecule per two to three base pairs), and the binding equilibrium is reached within seconds, remarkably faster than previously acknowledged. To characterize drug release profiles, DON degradation and DOX release from the complexes upon DNase I digestion was studied. For the employed DONs, the relative doses (DOX molecules released per unit time) may vary by two orders of magnitude depending on the DON superstructure. In addition, we identify DOX aggregation mechanisms and spectral changes linked to pH, magnesium, and DOX concentration. These features have been largely ignored in experimenting with DNA nanostructures, but are probably the major sources of the incoherence of the experimental results so far. Therefore, we believe this work can act as a guide to tailoring the release profiles and developing better drug delivery systems based on DNA-carriers.

Figure 1.

Schematic of the doxorubicin (DOX) loading into a DNA origami nanostructure (DON) and subsequent release upon enzymatic degradation. Here, we 1) study how DOX is loaded into DONs (in seconds), optimize the conditions for the loading by monitoring the spectroscopic features of DOX, and characterize the formed DOX–DON complexes. Through simultaneous real-time detection of the absorbance and fluorescence changes of the DOX-loaded DONs, we then 2) monitor the degradation of DONs into single-stranded DNA fragments by nucleases (DNase I, green) (in minutes to hours under a DNase I concentration of 34 U ml⁻¹) and 3) characterize the subsequent DOX release profiles of different DONs and show that the release profiles of DOX depend on the DNA origami superstructure and the applied DOX content.

Figure 2.

Effects of buffer conditions on the spectral features of DOX in the absence of DNA. (A) Absorption and emission (inset) spectra of 3 μM DOX in 40 mM Tris, 0 mM MgCl₂ at pH 6.0–9.0. The emission spectra were obtained at an excitation wavelength of 494 nm. (B) Spectral features of 3 μM DOX in 40 mM Tris, pH 7.4 buffer at different MgCl₂ concentrations. The inset figure shows a comparison of the emission spectra of the 0 mM and 100 mM samples at 494 nm excitation, with the maximum emission intensity of the 0 mM MgCl₂ sample normalized to 1. The 0 mM MgCl₂ spectrum (black) corresponds to the pH 7.4 spectrum in (A). (C) Absorption (black/gray lines) and emission (orange lines) spectra of 3 μM DOX in the chosen experimental conditions: 40 mM Tris, 10 mM MgCl₂, pH 7.4. The effect of the 10 mM MgCl₂ concentration is shown by comparing the spectra measured at 10 mM MgCl₂ (solid lines) with spectra measured at 0 mM MgCl₂ (dashed lines).

Figure 3.

The formation of DOX–DON complexes and DOX aggregates during the loading process. (A). In the applied loading protocol, (i) DOX (c₀ = 2 mM, 200 μM or 20 μM) is mixed with 2.5 nM triangle DONs and incubated at RT. Control samples are prepared without DONs. (ii) After t = 24, 48 or 96 h, centrifugation for 10 min at 14 000 g is used to separate high-MW particles from the solution; either DOX-loaded DONs or DOX aggregates. (iii) The concentration of DOX removed from the solution by precipitation (c_aggregate) is quantified by removing a small volume of the supernatant and determining the DOX concentration in the supernatant (c_free) from DOX absorbance (A₄₈₀); c_aggregate = c₀ – c_free. (B). The concentration of DOX in the pellet (c_aggregate) versus incubation time in the presence and absence of DONs. The amount of sedimentation was determined for 2 mM, 200 μM and 20 μM DOX loading concentration (c₀) in three different buffers: FOB pH 8.0 (12.5 mM MgCl₂), Tris/Mg^{2 +} pH 8.0 (10 mM MgCl₂) and Tris/Mg^{2 +} pH 7.4 (10 mM MgCl₂). For the DOX-DON samples, the DOX/bp ratio additionally indicates the number of DOX molecules in the precipitate per DNA base pair (c_bp = 18 μM for 2.5 nM DONs). The c_aggregate values are expressed as the mean ± standard error, n = 3.

Figure 4.

Titration experiments for determining the DOX-loading properties of DONs. (A) The models and microscopy images of the studied 2D and 3D DONs. The triangle, bowtie, and double-L 2D DONs are shown on the left accompanied by atomic force microscopy (AFM) images. The 3D DONs—the capsule and the 24-helix bundle (24HB) are shown on the right in TEM images. The AFM images are 500 nm × 500 nm in size, and the TEM images are 300 nm × 300 nm. (B) Representative changes in the absorption spectrum and fluorescence emission after 494 nm excitation (inset) of 3 μM DOX when the concentration of DNA base pairs (bp) in the solution is increased. The spectra have been measured for the triangle DON after the system has reached an equilibrium. The DNA concentration at each titration step is expressed as the molar ratio between DNA base pairs and DOX molecules in the sample (bp/DOX), and indicated in the legend. The fluorescence spectra have been corrected for the decrease of the molar extinction coefficient at the excitation wavelength (ε₄₉₄), and represent the quantum yield of the emitting molecules (Φ). (C) The dependency of ε₄₉₄ and Φ on the bp/DOX ratio (left panel) and the loading kinetics (right panel). In the left panel, the measured values for ε₄₉₄ and Φ during a titration with the triangle DON have been fitted with a 2-component binding model (Equation 2). The corresponding spectra and titration isotherms for the other DONs are presented in the Supplementary Figure S4. The kinetics of ε₄₉₄ (empty circles) and Φ (filled circles) in the right panel have been measured by monitoring the absorption and fluorescence spectra of the samples after adding DONs (triangle or 24HB) at the indicated bp/DOX ratio at t = 0. The data sets have been fitted with a 1-component exponential decay model of the form ae^−bx + c to illustrate the observed kinetic trends. (D) Increase of the fraction of bound DOX molecules (f_b) when the DNA base pair concentration in the sample increases, obtained by fitting the fluorescence data.

Figure 5.

DNase I digestion of the DONs and the subsequent DOX release. (A) Representative (n = 1) digestion and DOX release profiles of the studied DONs at different DOX loading concentrations, after introducing 34 U ml⁻¹ DNase I into the sample at t = 0. The structural integrity of the DONs ( intact) has been determined from the increase of the A₂₆₀ signal, and shown with the white, gray, and black markers depending on the DOX concentration (0, 3 or 6 μM). For samples containing 3 or 6 μM DOX, the DOX release ( DOX released) is shown with the light and dark orange markers and represents the percentage of initially bound molecules that have been released due to the digestion. DOX released has been determined from the recovery of the fluorescence quantum yield of DOX. (B) The cross-correlation between % DOX released and % intact for all DONs at 3 and 6 μM DOX loading concentration. (C) DNase I digestion rates of the DONs at 0, 3 or 6 μM DOX loading concentration. All digestion rates are shown relative to the triangle DON at 0 μM DOX concentration, and have been averaged from the digestion rates determined from three individual measurements; expressed as the mean ± standard error. (D) DOX release rates of the DONs at 3 and 6 μM DOX loading concentration. Relative dose stands for the absolute number of DOX molecules released per unit of time. All values are expressed as the mean ± standard error (n = 3).