HAuNS were synthesized by sacrificial galvanic replacement of cobalt nanoparticles in the presence of chloroauric acid according to the method of Schwartzberg et al.31
The average diameter of the resulting HAuNS was 38.5±1.7 nm, as determined by the dynamic light scattering method. Transmission electron microscopy (TEM) revealed the morphology of the HAuNS () and indicated that they had average diameter of 43±4.6 nm and average Au shell thickness of 4.0 nm. Surface modification by polyethylene glycol (PEG) resulted in a moderate increase in the average diameter of the HAuNS from 38.5±1.7 nm to 48.6±1.3 nm. Compared with HAuNS, PEG-HAuNS had a significantly increased colloidal stability: no aggregation was observed when PEG-HAuNS were stored in water at room temperature over a period of 3 months. The absorption spectra showed that the plasma resonance peaks for both HAuNS and PEG-HAuNS were tuned to the NIR region (~800 nm) (). Thus, PEG modification did not affect the spectrum characteristic of HAuNS but did increase their physical stability.
Fig. 1 Fabrication and characterization of DOX-loaded hollow gold nanospheres (DOX@HAuNS). A, Schematic of HAuNS synthesis and TEM images of plain and DOX-loaded HAuNS (DOX@HAuNS and DOX@PEG-HAuNS). B, Absorption spectra of HAuNS and PEG-HAuNS. C, Absorbance (more ...)
The complexes of DOX@HAuNS and DOX@PEG-HAuNS were readily formed by simple mixing of HAuNS or PEG-HAuNS solutions with DOX for 24 h at room temperature and then repeated washing to remove unbound DOX. TEM images clearly showed a DOX layer with a thickness of about 4-6 nm covering the surface of both DOX@HAuNS and DOX@PEG-HAuNS, which did not exist in HAuNS and PEG-HAuNS prior to DOX coating (). The color of the HAuNS changed from greenish to henna upon absorption of DOX. DOX@HAuNS displayed a UV–Vis absorption peak at 490 nm characteristic of DOX and a broad NIR plasmon absorption peak at ~800 nm characteristic of HAuNS (). At 24 h after mixing, the absorbance peak intensity of DOX in the UV-visible region was significantly reduced compared with the absorbance peak intensity immediately after mixing of DOX and HAuNS (0 h), when DOX was unbound (). In addition, compared to free DOX, which exhibited strong fluorescence emission, the fluorescence signal from DOX in DOX@HAuNS was almost completely quenched (, inset), a phenomenon known to occur when fluorophores are attached to a metal nanoparticle surface with close proximity.32
These results indicate that DOX was tightly bound to HAuNS and PEG-HAuNS after 24 h incubation.
shows Langmuir adsorption isotherms at room temperature for the adsorption of DOX by HAuNS and PEG-HAuNS. PEG modification on HAuNS significantly affected DOX's absorbing efficiency. At a lower PEG coating density (PEG:Au molar ratio = 0.008:1), PEGylation of HAuNS resulted in higher DOX loading. However, the loading efficiency of DOX decreased with increasing PEG coating density. This may be explained by increased surface area as a result of increased colloidal stability and thus reduced population of aggregates formed among HAuNS when a small amount of PEG was introduced to the HAuNS surface. With excessive PEG modification, however, interaction between HAuNS and DOX was impeded because the steric hindrance of PEG prevented DOX molecules from approaching the surface of HAuNS. The saturation of DOX absorption to PEG-HAuNS at higher PEG coating densities (PEG:Au molar ratio = 0.04:1 and 0.125:1) suggests that the surface of HAuNS was completely occupied by DOX and PEG molecules under such conditions, whereas at the lower PEG coating density (PEG:Au molar ratio = 0.008:1), the DOX absorption isotherm did not plateau up to an initial DOX amount of 180 μg. Further increase in DOX payload is possible for this particular formulation. At the DOX amount of 180 μg in the initial DOX-HAuNS mixtures, the drug contents absorbed to HAuNS were 41%, 63%, and 27% for HAuNS, PEG-HAuNS (PEG:Au molar ratio = 0.008:1), and PEG-HAuNS (PEG:Au molar ratio = 0.125:1), respectively, corresponding to weight ratios of 0.69, 1.7, and 0.37 between DOX and gold (). DOX loading to HAuNS and PEG-HAuNS led to an increase in the size of the nanoparticles, suggesting that the presence of DOX reduced the colloidal stability of these nanoparticles. DOX@HAuNS having higher PEG density were more stable than DOX@HAuNS having lower PEG density, which in turn were more stable than HAuNS without PEG coating (). In the following descriptions of our studies, “PEG-HAuNS” refers to PEG-HAuNS with PEG:Au molar ratio of 0.008:1 unless specified otherwise.
Fig. 2 A, Langmuir adsorption isotherms showing the adsorption of DOX by HAuNS and PEG-HAuNS with various PEG:Au molar ratios. The absorbed DOX was plotted against the initial amount of DOX in the solution. B, Comparison of DOX payload (left) and particle size (more ...)
In addition to colloidal stability, we also examined the stability of absorbed DOX on HAuNS and PEG-HAuNS. After an initial release of 15%-20% over the first 2-day period, no further release of DOX from either DOX@HAuNS or DOX@PEG-HAuNS was observed in water, phosphate-buffered saline (PBS, pH 7.0), or cell culture medium containing 10% fetal bovine serum over the second 2-day period (). These results indicated that DOX was stably absorbed to HAuNS and PEG-HAuNS.
To investigate the mechanism of DOX binding to HAuNS, the amino group in DOX was blocked by an acetyl protecting group. The binding of DOX-acetamide to HAuNS was reduced to nearly zero (), suggesting that the amino group of DOX was key for its high payload onto HAuNS. According to synthesis protocol, HAuNS were stabilized by negatively charged citrate with a zeta potential of -25 mA in PBS solution (pH 7.4). On the other hand, DOX was positively charged at pH 7.4. As a result, DOX was absorbed onto HAuNS via electrostatic interaction. The formation of complexes between drug molecules and nanocarriers is advantageous compared to covalent conjugation approach owing to easy fabrication and scale up, low cost, and predictable release profile.19, 33
DOX release from DOX@HAuNS and DOX@PEG-HAuNS could be readily controlled by using NIR laser. After the first NIR laser irradiation (begun at 1 h) at 4.0 W/cm2 output power for 5 min, the cumulative release, defined as the ratio of released DOX to total loaded DOX expressed as a percentage, increased from 4.1% to 22.2% for DOX@HAuNS and from 4.1% to 31.9% DOX for DOX@PEG-HAuNS (). Release of DOX was significantly reduced or almost completely stopped when the NIR laser was switched off over the next 1 h of incubation. Similar results were observed when the laser treatment protocol was repeated beginning at 2 h. However, there was less DOX released during the second treatment cycle. By the third treatment cycle (beginning at 3 h), almost no DOX was released. These data suggested that DOX release from both DOX@HAuNS and DOX@PEG-HAuNS could be triggered by external NIR laser.
Fig. 3 NIR light triggered release of DOX from DOX@HAuNS and DOX@PEG-HAuNS. A, DOX release profiles in the presence and absence of NIR laser. Irradiation with NIR laser caused rapid DOX release during NIR exposure (5 min, red lines), and the release was turned (more ...)
After NIR laser irradiation of DOX@HAuNS (~1.0 ×1012 particles/mL), the peak absorption intensity at around 490 nm was significantly increased, indicating release of free DOX from the nanocomplexes. The color of DOX@HAuNS changed from brown to green due to detachment of DOX from DOX@HAuNS (). After centrifugal removal of free DOX, the remaining HAuNS solution displayed an absorption spectrum similar to that of DOX@HAuNS acquired prior to NIR laser irradiation but at a lower intensity, and with loss of the characteristic peak absorption of DOX at 490 nm (). The reduced peak intensity at 800 nm after NIR laser irradiation may be attributed to fragmentation of a small fraction of HAuNS particles (data not shown).
Because DOX was attached to HAuNS via electrostatic interaction, it was anticipated that the release of DOX from DOX@HAuNS and DOX@PEG-HAuNS would be pH-dependent. Indeed, while there was no DOX released from DOX@PEG-HAuNS in PBS at pH 10 and only 11% DOX released in PBS at pH 7.4 at room temperature after 2 days of incubation, the DOX release reached 35% at pH 5.0 and 57% at pH 3.0 (). The observed pH dependency is attributed to the increased hydrophilicity and higher solubility of DOX at lower pH caused by increased protonation of -NH2 groups on DOX, which reduce the interaction between DOX and HAuNS. When the pH was decreased, the COO- groups on the surface of HAuNS also became protonated. As a result, the electrostatic interaction between DOX and HAuNS was reduced. The pH-dependent drug release from HAuNS or PEG-HAuNS could be exploited for drug delivery applications: the microenvironments of extracellular tissues of tumors and intracellular lysosomes and endosomes are acidic, and this acidity could facilitate active drug release from HAuNS-based delivery vehicles.
NIR laser irradiation increased the amount of DOX released from DOX@HAuNS or DOX@PEG-HAuNS incubated in PBS at different pH levels; the lower the pH of the medium, the less DOX released. For example, after 5 min of continuous NIR laser irradiation at 4.0 W/cm2
, 14.6%, 16.7%, and 5.1% more DOX was released from DOX@HAuNS when the nanoparticles were incubated in PBS at pH 7.4, pH 5.0, and pH 3.0, respectively (). The ability to achieve higher DOX release with NIR light at pH ~5.0 is favorable because the intracellular lysosome environment of tumor cells has a pH of approximately 5.0.34
To explain the observed high loading capacity of DOX to HAuNS and further evaluate the advantages of HAuNS as drug carriers, we compared the DOX loading efficiency and drug release behavior between HAuNS and AuNPs having similar size (~40 nm) and surface charge (zeta potential ~ -25 mA). HAuNS or AuNPs with the same equivalent Au concentration were incubated with DOX (final concentration of 1.0 mM). On the basis of 1.0 μg Au, the DOX payload increased from ~0.2 μg in AuNPs to 0.7 μg in HAuNS, a 3.5-fold increase (). This increase can be explained by the greater surface area of HAuNS compared to AuNPs. Assuming that the shell thickness of HAuNS is 4.0 nm and the diameter of both HAuNS and AuNP is 40 nm, it is estimated that each solid AuNP is equivalent to two hollow HAuNS on the basis of weight. This means that HAuNS should have twice the DOX payload of AuNPs if all DOX molecules are coated on the outer surface of HAuNS. The fact that 3.5-fold higher DOX was found bound to HAuNS suggests that the inner surface of HAuNS was also coated with DOX. In fact, it is estimated that HAuNS possess 3.2-fold greater surface area than AuNPs when the inner and outer surface areas are counted. TEM showed that the shell of HAuNS was porous (), which makes it possible for DOX molecules to diffuse into the core and bind to the inner surface of HAuNS.
Fig. 4 Comparison between hollow gold nanospheres (HAuNS) and solid gold nanoparticles (AuNP). A, Comparison of DOX payload between HAuNS and solid AuNP. B, Release of DOX from DOX@HAuNS (blue line) and DOX@AuNP (pink line) under repeated NIR laser exposure. (more ...)
In addition to significant difference in DOX loading capacity between HAuNS and AuNP, DOX@HAuNS also displayed distinct characteristic of NIR light-triggered DOX release. In contrast, no DOX release was observed when DOX-coated AuNP was irradiated with NIR light (). This is because unlike DOX@HAuNS, there was no plasmon absorption in the NIR region for DOX@AuNP (). When aqueous solutions of HAuNS, DOX@HAuNS, and DOX@PEG-HAuNS having the same nanoparticle concentration of 0.7×1011 particles/mL were exposed to NIR light (5.0 W/cm2 for 10 min), the temperature was increased 39°C, 27°C, and 30°C, respectively. In comparison, no significant temperature change was observed when PBS or solution of AuNP was exposed to the NIR laser (). Thus, the temperature elevation mediated by HAuNS in the presence of NIR light may be responsible for NIR laser–triggered release of DOX from DOX@HAuNS and DOX@PEG-HAuNS.
Both DOX@HAuNS and DOX@PEG-HAuNS were internalized into MDA-MB-231 cells and were retained in the endolysosomal compartments. After 1 h incubation, DOX@HAuNS showed strong red fluorescence signal from DOX despite quenching effect with DOX bound to HAuNS. The fluorescence signal was limited to spots scattered throughout the cytoplasm. The white bright dots obtained from dark-field imaging indicated the presence of HAuNS, which to a large extent colocalized with DOX (). These results suggested that HAuNS together with DOX were phagocytosed by the cancer cells and distributed to the endolysosomal vehicles. After 48 h, DOX and HAuNS remained trapped in the endolysosomal vehicles (). This observation contrasts with our earlier findings that DOX could be released from DOX@HAuNS at ~pH 5. It may be that in the microenvironment of endolysosomes, the detached DOX re-attached to HAuNS before the drug molecules had a chance to diffuse out of the vehicles. In contrast to DOX@HAuNS, free DOX was taken up by the tumor cells and distributed to cell nuclei 1 h after incubation (). NIR laser irradiation (1.0 W/cm2 for 3 min per treatment, 4 treatments over a 2-h period) caused release of DOX from DOX@HAuNS, and DOX was distributed to cell nuclei (). Thus, it is possible to control intracellular DOX release from DOX@HAuNS and DOX@PEG-HAuNS by NIR laser irradiation.
Fig. 5 Uptake of DOX@HAuNS and free DOX in MDA-MB-231 breast cancer cells. A, Cell uptake of DOX@HAuNS after 1 h and 48 h of incubation. The reflectance of HAuNS was visualized using a dark-field condenser. The red color from the fluorescence signal of DOX@HAuNS (more ...)
Both DOX@HAuNS and DOX@PEG-HAuNS were cytotoxic against MDA-MB-231 cells in a dose-dependent manner. About 33.5% and 39.5% of cells were killed by DOX@HAuNS and DOX@PEG-HAuNS, respectively at an equivalent DOX concentration of 10 μg/mL (). However, free DOX exhibited higher toxicity, with 77.4% cell killed at the same drug concentration. The lower cell killing potency with DOX@HAuNS and DOX@PEG-HAuNS could be attributed to relatively stable complexes formed between DOX and HAuNS and delayed DOX release inside cells. After NIR laser irradiation (2 W/cm2
for 3 min per treatment, 4 treatments over a 2-h period), both DOX@HAuNS and DOX@PEG-HAuNS showed significant enhanced cell-killing effect toward cancer cells, with about 83.5% and 86.4% cell killed, respectively, at an equivalent DOX concentration of 10 μg/mL (). HAuNS was not cytotoxic with Au concentrations ranging from 0.04 μg/mL to 160 μg/mL (). This is consistent with literature findings that in general, Au-based nanoparticles are well tolerated.14, 15
Plain unloaded HAuNS exhibited significant cell-killing effect (>69% cells killed), when cells incubated with HAuNS were treated with NIR laser at Au concentrations greater than 16 μg Au/mL (). These results indicate that HAuNS produced significant photothermal ablation at concentrations greater than 16 μg Au/mL. At an Au concentration of 5.9 μg/mL (10 μg/mL equivalent DOX), 86.4% of cells were killed with combined DOX@PEG-HAuNS and NIR laser treatments. In comparison, treatments with combined HAuNS and NIR laser killed only 40.6% of the cancer cells (). The enhanced cytotoxicity of DOX@HAuNS and DOX@PEG-HAuNS at lower concentrations primarily resulted from DOX released upon NIR laser irradiation. Thus, at higher concentrations of DOX@HAuNS and DOX@PEG-HAuNS, both photothermal ablation and cytotoxic activity of DOX contributed to the killing of cancer cells.
Fig. 6 Cell survival as a function of DOX concentration (left) and Au concentration (right). MDA-MB-231 cells were either not exposed to NIR light or irradiated with NIR light (2 W/cm2 for 3 min per treatment, 4 treatments over 2 h). The viability of cells was (more ...)