Gold-silver alloy nanoshells: a new candidate for nanotherapeutics and diagnostics
© Gheorghe et al; licensee Springer. 2011
Received: 12 July 2011
Accepted: 13 October 2011
Published: 13 October 2011
Skip to main content
© Gheorghe et al; licensee Springer. 2011
Received: 12 July 2011
Accepted: 13 October 2011
Published: 13 October 2011
We have developed novel gold-silver alloy nanoshells as magnetic resonance imaging (MRI) dual T 1 (positive) and T 2 (negative) contrast agents as an alternative to typical gadolinium (Gd)-based contrast agents. Specifically, we have doped iron oxide nanoparticles with Gd ions and sequestered the ions within the core by coating the nanoparticles with an alloy of gold and silver. Thus, these nanoparticles are very innovative and have the potential to overcome toxicities related to renal clearance of contrast agents such as nephrogenic systemic fibrosis. The morphology of the attained nanoparticles was characterized by XRD which demonstrated the successful incorporation of Gd(III) ions into the structure of the magnetite, with no major alterations of the spinel structure, as well as the growth of the gold-silver alloy shells. This was supported by TEM, ICP-AES, and SEM/EDS data. The nanoshells showed a saturation magnetization of 38 emu/g because of the presence of Gd ions within the crystalline structure with r 1 and r 2 values of 0.0119 and 0.9229 mL mg-1 s-1, respectively (Au:Ag alloy = 1:1). T 1- and T 2-weighted images of the nanoshells showed that these agents can both increase the surrounding water proton signals in the T 1-weighted image and reduce the signal in T 2-weighted images. The as-synthesized nanoparticles exhibited strong absorption in the range of 600-800 nm, their optical properties being strongly dependent upon the thickness of the gold-silver alloy shell. Thus, these nanoshells have the potential to be utilized for tumor cell ablation because of their absorption as well as an imaging agent.
The search for new composite systems consisting of a wide range of metal and semiconductor core materials with an outer inert shell has led to the discovery of novel nanoparticles with a broad range of biomedical applications in areas such as tissue engineering, gene and drug delivery, photo-thermal therapy, cell tracking, and storage systems [1–3]. Among this broad area of nanometer-sized systems, iron oxide nanocores have gained special attention because of their unique physical properties in which their size, morphology, composition, and surface chemistry can be tailored to many biological and biomedical applications [4, 5]. Tantamount to their physical versatility, these nanocores possess unique magnetic properties that facilitate proton relaxation within specific tissues, which thus make them suitable as T 2 contrast agents for magnetic resonance imaging (MRI) [6, 7]. Typical iron oxides that have extensively been explored for MRI applications include magnetite, Fe3O4, with alternating lattices of Fe(II) and Fe(III) ions or maghemite, γ-Fe2O3, in which the lattice structure consist of Fe(III) ions only . However, despite their widespread use in the biomedical field, an open challenge for using iron oxide nanoparticles as contrast agents is to improve their magnetic properties, which in turn would lead to higher imaging sensitivity. This can be done by doping them with transition metal ions, in particular lanthanides which have distinctive magnetic and optical properties, associated with their electronic configuration [9, 10].
However, apart from their magnetic properties, these nanocores are unstable both in air and in solution. Iron oxide typically forms aggregates in solution and undergoes oxidation in air [11, 12]. In addition, even though iron oxide nanoparticles are good magnetic materials, the cores are susceptible to corrosion in the presence of water [13, 14] and are more reactive in bulk materials because of their high surface-to-volume ratio . Also, aside from their physical instability and relative biocompatibility [16, 17], iron oxide nanocores have been shown to be toxic . However, these hurdles can be overcome by coating the cores with outer shells, such as inorganic oxides (silica) [19, 20], inert metals (silver, gold, or gold alloys) [11, 21–23], or bioactive macromolecules such as liposomes and micelles, which can encapsulate a numerous amount of nanoparticles within the structures that then impart an inert physical property to the cores in biological media [2, 24–26]. Subsequent surface functionalization of the nanocomposite materials with hydrophilic polymers such as poly(ethylene glycols) (PEGs) [27–30] and dextrans [31, 32] can further increase the physical stability of the nanoparticles particularly in solution. In addition, the coating of the magnetic nanoparticles with inert metals, such as gold, is extremely attractive because the nanocomposite system is resistant to corrosion in biological conditions and can decrease the toxicity of the cores [33–35].
Moreover, coating of these doped cores with an outer inert shell can sequester toxic metals such as gadolinium within the core of the nanoparticles . Gd(III) ions have to be combined with carrier molecules because of their extreme toxicities, which are strongly associated with a systemic fibrosing disorder that is referred to as nephrogenic systemic fibrosis (NSF) in patients with kidney diseases [36, 37]. The Gd ions are toxic because their ionic radius is almost equal to that of divalent calcium ions, which enable Gd to compete with biological systems including enzymes with a higher binding affinity to alter the kinetics of their systems . In addition, Gd ions are inorganic blockers of voltage-gated calcium channels . Hence, the development of a contrast agent that could completely sequester Gd within the core of the nanoparticle, thereby preventing its release into the body would highly be beneficial as a safe and effective alternative to typical Gd-based contrast agents.
Currently, there is an extensive development of monometallic nanoshells with minor attention given to nanosystems consisting of bimetallic and trimetallic alloy shells [40–43]. Recent research in this area has shown that gold-silver alloy nanoshells might have additional biomedical applications (e.g., cancer screening), because of their distinct optical properties ranging from visible to the near-IR wavelength region . Our exploratory research aimed at preparing new biocompatible nanoshells for MR imaging that could overcome NSF and can also be utilized for laser ablation of tumor cells has led to the synthesis of novel Gd-doped magnetite cores, Gd:Fe3O4, covered with a gold-silver alloy shell.
Powder X-ray diffraction analysis was used for the analysis of the synthesized samples. Data were collected at room temperature using a PANanalytical X'Pert PRO diffractometer equipped with Cu Kα radiation (λ = 1.5406 Å). The samples were ground using an agate mortar and pestle to a suitable powder sample. The powder was then placed on a zero background sample holder. The 2θ range was 5-120° with a scan rate of 6.4°/min. Data compilation and analysis were processed using the Diffractometer Management System software package which included a JCPDS powder diffraction database . The composition of the samples was analyzed using a sequential Thermo Jarrell Ash Atomscan 25 inductively coupled plasma (ICP) emission spectrometer. Measurements were made using the following selected lines: 230.562 for Fe, 328.068 for Ag, 242.795 for Au, and 342.347 for Gd. The instrument drift was corrected by running a quality control solution every two to three analyses. The nanoparticles were dissolved in aqua regia and evaporated to dryness. The resulted residues were dissolved in 10% nitric acid and diluted to concentrations in the range of 1-50 ppm for analysis. Standard solutions with concentrations ranging from 0 to 50 ppm were used for calibration curves. Chemical analyses of the samples by means of SEM/EDS were performed on a Jeol JSM 8330 F scanning electron microscope equipped with an energy-dispersive spectrometer (EDS) by means of 15 kV accelerating voltage and 12 μA emission current from a cold-field emission gun. Small volumes of water-dispersed nanoparticles were deposited on a copper disk and allowed to evaporate at room temperature. The copper disk was subsequently attached to a brass sample holder using carbon tape. All the analyses were carried out at approximately 4 × 10-4 Pa. Elemental analysis was used for all our samples. These results complemented the results from XRD and ICP-AES analyses. The size and morphology of the nanoparticles were determined with transmission electron microscopy (TEM) on a JEOL JEM-2000 FX instrument. Statistical analysis and average size distribution of the particles were determined by evaluating a minimum of 100 nanoparticles per sample. The mean standard deviation σ, defined as σ = ((1/(N - 1))*Σ i (x i - x)2)1/2, was also determined for each size distribution. The nanoparticles were dispersed in water and cast on a carbon-coated grid (Pelco® No. 160). The sample was allowed to settle on the grid for approximately 1 min in a humidified atmosphere. The carbon-coated grid was then washed with water, followed by another addition of nanoparticles onto the carbon-coated grid. The grids were then left for another 1 min before gently removing excess liquid with a filter paper. The absorption of the nanoparticles in the ultraviolet-visible-near infrared region (UV-Vis-NIR) was determined with a Beckman Coulter DU 800 spectrophotometer using a 0.1-mm quartz cuvette. The spectra were collected in the range of 190-1100 nm at a resolution of 0.5 nm, and the concentration of the nanoparticles was adjusted so that the maximum absorbance did not exceed 1 absorbance unit. The magnetic susceptibility of the synthesized nanoparticles was measured by a superconducting quantum interference device magnetometer (Quantum Design MPMS). The saturation magnetization measurements were performed at 300 and 5 K. Once the desired measurement temperature was reached at zero field, field cycling from 5 to -5 T back and forth was applied followed by data collection for each M(H) measurement curve. In addition, the zero-field-cooled (ZFC) and field-cooled (FC) measurements were performed by cooling the sample from 300 to 5 K at zero field and then applying 200 Oe field for the warming up scan. Longitudinal and traverse relaxation times were measured in a 1.5-T magnet on a Siemens Avanto, full body clinical scanner (Siemens AG Healthcare) at 37°C. The coil used was standard 12 channel head coil and the samples prepared were dispersed in 1% agarose gel at the following nanoparticle concentrations: 0.005, 0.01, 0.02, 0.03, and 0.04 mg/mL. In brief, solutions were solidified in test tubes of 10 mm diameter, placed in a sample holder, and simultaneously imaged together with calibration samples of copper sulfate (CuSO4 H2O) prepared in a similar fashion for the following concentrations: 6.54, 5.872, 5.538, 5.204, 4.87, 4.536, 3.868, 3.2, 2.08, 0.96, and 0.52 mg/mL. Fast spin echo images (echo train length 16, acquisition matrix = 256 × 128, field of view (FOV) = 220 × 110 mm resolution in an isotropic in-plane resolution of 0.86 mm, slice thickness 5 mm) were acquired in two series. In series 1, the echo time was varied from 10 to 160 ms (in increments of 10 ms) with fixed repetition time (TR = 3000 ms). From the signal intensity curves of these images, the spin-spin relaxation time T 2 was determined for each concentration. In series 2, images were acquired for the following TRs: 170, 350, 700, 1000, 2000, 4000, and 5000 ms while keeping the echo time constant at minimum (TE = 9.9 ms). Gray scale values from the MRI images acquired at different TR and TE times were analyzed as follows: signal intensities were determined as average over regions of interest within the test tube (using Image J, NIH) avoiding image voxel at the test tube wall to minimize partial volume averages. The nonlinear Levenberg-Marquardt fitting method was then used to the fit standard equations for MRI signal intensity for spin echo sequences  to both the signal intensities at different TE (series 1) to determine T 2 and to the signal intensities at different TR to determine T 1. These fits were performed for the different concentrations of the Au/Ag alloy-coated nanoparticles, the commercial Omniscan®, and the CuSO4 H2O standards. Relaxivities, r 1 and r 2, were then determined using linear regression analysis assuming a linear relationship between relaxation times T 1 and T 2 and concentration. For Omniscan® and CuSO4 H2O, relaxivities were compared to literature values.
Compositions obtained by ICP-AES for the analysis of Fe3O4, Gd:Fe3O4, Au-seeded functionalized Gd:Fe3O4, and gold-silver alloy nanoshells on gold-seeded amine-functionalized Gd:Fe3O4
Au-seeded Gd: Fe3O4
Au-Ag alloy = 1:1 nanoshells on Au-seeded amine-functionalized Gd: Fe3O4
Au-Ag alloy = 5:1 nanoshells on Au-seeded amine-functionalized Gd: Fe3O4
There was no significant change in the X-ray powder pattern of the Gd:Fe3O4 nanoparticles compared to the parent compound, Fe3O4, which was prepared as a reference by a coprecipitation reaction of Fe(III) and Fe(II) in NaOH at 65°C (see Figure 2A, B). These results are consistent with nanoparticles previously synthesized by the coprecipitation of the Fe(II), Fe(III), and Gd(III) salts in the same basic solution, which did not result in a mixed oxide system, Gd2O3-Fe3O4, but a doped Gd:Fe3O4. The powder pattern of the as-synthesized nanoparticles indicated a smaller unit cell for the doped nanocores that was confirmed by the shift of the diffraction peaks characterized by Miller indices shown in Figure 2A (220, 311, 222, 400, 422, 511, and 440) toward bigger Bragg angles. Indexing of the powder patterns presented in Figure 2A, B indicated for Gd:Fe3O4 a cubic unit cell with lattice parameter a = 8.308 Å, which is approximately 0.8% smaller than the unit cell of the parent compound, Fe3O4, a = 8.373 Å, respectively (reported data for Fe3O4: cubic unit cell, space group Fd-3m, a = 8.384 Å) . It was also suggested that in such doped system, the Gd(III) ions might occupy the octahedral sites . The average particle size of the nanocores was estimated by applying the Debye-Scherrer model, written as d = 0.9λ/ß d cosθ, where d is the crystallite diameter, λ = 1.540562 Å for Cu Kα line, ß d is the full width at half maximum of the strongest reflection peak, and θ is the corresponding Bragg angle . This formula was applied for the (311) reflection in the powder pattern shown in Figure 2B in which an average crystallite size of 3.5 nm for the nanocores was calculated. The gold seeding step of the Gd:Fe3O4 nanocores had as a result--a decrease in the intensities characteristic to the doped spinel structure (Figure 2C). The X-ray powder diffraction was also used as a tool to verify the formation of the gold-silver alloy nanoshells on the gold-seeded Gd:Fe3O4 nanocores. As can be seen from Figure 2D, apart from the weak peaks which corresponded to the (311), (422), (511), and (440) reflections of the Gd:Fe3O4, three extra peaks positioned at 2θ values of 38.4°, 44.6°, and 64.7° that corresponded to the gold-silver alloy (molar ratio Au:Ag = 1:1) were observed. In Figure 2E, peaks corresponding to the gold-silver alloy (Au:Ag = 5:1) only are observed. These data indicated the formation of a layer of gold-silver alloy on the nanocores and not the formation of discrete gold and silver nanoparticles, which would have shown six individual peaks: 38.2°, 44.4°, and 64.6° corresponding to (111), (200), and (220) planes for gold  and 38.4°, 44.3°, and 64.7° corresponding to (111), (200), and (220) planes for silver , respectively. Moreover, the XRD patterns of gold and silver are completely overlapped in Figure 1D, E; therefore, the gold-silver alloy cannot be distinguished from the powder patterns of gold and silver as monometallic phases . In addition, the absence of any diffraction peaks for Gd-doped magnetite cores in Figure 2E is because of the heavy atom effect from gold; it was observed that gold shells dominate the powder pattern because of its high electron density, which is consistent with previously reported data . Our observation is consistent with data reported for gold nanoshells deposited on magnetite cores by others [62–64], as well as with ICP-AES results (Table 1) which indicated a molar ratio of Au:Ag = 0.4 (theoretical ratio 1:1) and 5.3, respectively (theoretical ratio 5:1). In addition, Liu et al. showed that the powder pattern of gold-silver alloy shells deposited on silica cores exhibited only the peaks characteristic to the gold or silver monometallic phase, whereas the peaks corresponding to the silica support were found to be absent . Thus, our findings from the TEM, ICP-AES, and SEM/EDS data together with previously published one suggested that the Gd:Fe3O4 nanocores were completely covered with the gold-silver alloy and hence the formation of the nanoshells. Consistent with these findings was the optical absorption data for the nanoshells, which exhibited only one plasmon band that further supported the suggested nanoshell structure (discussed later). Therefore, we have shown that the reduction of gold and silver salts in the presence of gold-seeded Gd:Fe3O4 cores successfully led to the formation of gold-silver alloy nanoshells. However, the gold-silver shell is very thin when Au:Ag = 1:1 since the diffraction from gold and silver does not dominate the powder pattern, but a decrease of the intensity of the reflection peaks for the Gd:Fe3O4 nanocores was observed (Figure 1D) [1, 64]. As the ratio of Au:Ag was increased to 5:1, the reflections of gold and silver started to dominate the powder pattern and most of the characteristic peaks of the nanocore disappeared (Figure 2E).
Based on previous investigations [44, 64, 78] as well as our experimental observations, we anticipate that the optical properties of the nanoshells can be tailored in a controlled manner by adjusting the volume of the gold seed solution as well as the amounts of gold and silver salts used as precursors in the formation of the nanoshells. Studies regarding this matter, the optimization of the synthesis conditions of the nanoshells as well as the investigation of their biocompatibility are currently under way. Thus, our findings have shown the successful synthesis of a gold-silver alloy shell on composite magnetite nanocores. In addition, these nanosystems hold great promise for medical applications since based on their composition, the nanoparticles should be able to overcome corrosion and toxicity issues as well as aggregation. Furthermore, the magnetic properties of the nanocores and the thickness of the alloy shell may be systematically investigated and exploited. Complete structural characterizations and measurements of these relevant physical properties of these novel nanoshells are currently underway as well as the evaluation of the nanoshells in biological assays including in-depth cytotoxicity studies and ablation studies in various cell lines.
More details of the sample characterization including results from TEM, UV-Vis-NIR, and M(T) measurements are given in "Figure S4 in Additional file 1."
We thank Dr. Irene Rusakova for assistance with TEM analysis. We also gratefully acknowledge the financial support of this study that was made available from the Department of Defense (DOD) (OC073093) and Texas Higher Education Coordinating Board-ARP Award (003652-0217-2007).
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.