Radiosynthesis of Stable 198Au-Nanoparticles by Neutron Activation of αvβ3-Specific AuNPs for Therapy of Tumor Angiogenesis

This paper reports on the development of stable tumor-specific gold nanoparticles (AuNPs) activated by neutron irradiation as a therapeutic option for the treatment of cancer with high tumor angiogenesis. The AuNPs were designed with different mono- or dithiol-ligands and decorated with different amounts of Arg-Gly-Asp (RGD) peptides as a tumor-targeting vector for αvβ3 integrin, which is overexpressed in tissues with high tumor angiogenesis. The AuNPs were evaluated for avidity in vitro and showed favorable properties with respect to tumor cell accumulation. Furthermore, the therapeutic properties of the [198Au]AuNPs were evaluated in vitro on U87MG cells in terms of cell survival, suggesting that these [198Au]AuNPs are a useful basis for future therapeutic concepts.

The focus of this work was the development of highly stable targeted gold nanoparticles for neutron activation [33].Therefore, AuNPs with mono-and di-thiol linkers with low and high loading of target-specific peptides were synthesized to compare their specific avidity in cell binding assays and their stability during and after neutron irradiation.To achieve target-specific accumulation in tissues with high tumor angiogenesis, the AuNPs were functionalized with a c(RGDfK) derivative [32,34].The Arg-Gly-Asp (RGD) peptide motif is known to bind to the transmembrane α v β 3 integrin [35,36], which is overexpressed in tumor angiogenesis in tumors of various origins, for example, on glioma cells (U87MG) [37][38][39].

Synthesis and Functionalization of Gold Nanoparticles
Integrin α v β 3 , a transmembrane protein expressed on endothelial cells, binds the RGD triple amino acid peptide motif of extracellular matrix proteins.Growing malignant tumors require continuous angiogenesis, and the integrin α v β 3 is overexpressed for this purpose.As a result, α v β 3 is preferentially expressed in tumor angiogenesis and is a potential target for AuNPs decorated with RGD peptides [36].Therefore, ultra-small AuNPs 3 and 6 (3 ± 2 nm) were synthesized by Turcu et al. [40] and Brust and Schiffrin [41], respectively.The AuNPs contained thiol-PEG 3 -OH or a thioctic acid(TA)-PEG 3 -OH derivative 2 used as the stabilizing ligands and to achieve enhanced biocompatibility (Figure 1).The AuNPs were further functionalized by ligand exchange with low and high amounts (4-8 mg) of TA-PEG 4 -c(RGDfK) derivative 5 to obtain mixed AuNP-thio-PEG-dithio-PEG-RGD 7a (high RGD loading), 7b (low RGD loading) and AuNP-dithio-RGD 8a (high RGD loading), and 8b (low RGD loading), respectively.The AuNPs were purified by dialysis and sizeexclusion chromatography.The size and stability of AuNPs 7a,b, and 8a,b were confirmed by UV/Vis spectroscopy and high performance liquid chromatography (HPLC).The organic shell of the AuNPs was characterized by mass loss using thermogravimetric analyses for each functionalization step.After knowing the number of newly attached molecules, a formula by Zhu et al. was used to calculate the total molar mass of the AuNPs [42] (Table 1).A brief description of the synthesis and characterization can be found in Appendix A. All AuNPs were fully characterized by thermogravimetric analysis (TGA) (Table 1), electron microscopy (EM) (Figures A1 and A2), UV/Vis spectroscopy (Figure A9), HPLC (Figure A10), and nuclear magnetic resonance spectroscopy (NMR) (Figures A11-A16).The AuNPs could be stored in lyophilized form at −20 • C for >12 months without loss of integrity.In contrast, when stored in solution at room temperature, aggregation in the form of precipitation could occur within weeks, especially for peptide-decorated particles [43].

Neutron Irradiation Experiments
First neutron irradiation experiments with thermal neutrons at the TRIGA Mainz reactor were performed with non-tumor specific AuNP-dithio-PEG 3 (3-1-3-5) and AuNPthio-PEG 6 (6-1-6-3) in different weights and concentrations.Samples were frozen and removed from the freezer immediately before irradiation.Irradiation was performed at 100 kW for 1-2 h with a neutron flux of 1.6 × 10 12 cm −2 × s −1 .With the reactor running, the background dose rate (DR) at the measurement position was ~2 µSv/h.Gamma measurements were not possible for probes >500 µg on the day of irradiation due to the high activity.The dead time for samples 3-5 (Table 2) was about 30 min at the end of the bombardment and still 7.5 min at 20 cm distance.Therefore, most of the gamma measurements of [ 198 Au]AuNPs had to be performed one day after irradiation.Sample [ 198 Au]3-5 still had ~3% dead time in 20 cm distance (Table 2).Precipitation was observed for [ 198 Au]6 but not for [ 198 Au]3 in the solution or on the vessel wall in any case (Figure 2).The activated samples were stored in the freezer for transport and further experiments.In addition, the half-life of [ 198 Au]3 was determined experimentally (mean 2.80 ± 0.07 d) by measuring the activity of different concentrations with a gamma counter for 28 d (Figure A8).UV-Vis measurements showed a strong broadening of the plasmon bands for [ 198 Au]6-1 and [ 198 Au]6-2, indicating aggregation (Figure 3).For [ 198 Au]3-1 and [ 198 Au]3-2 a typical absorption for AuNPs at 514 nm was observed, indicating stable AuNPs even 5 months after neutron activation ([ 198 Au]3-3, Figure A9).The production of ~100 MBq [ 198 Au]3 showed stable AuNPs even at high activity concentration for at least 15 d by HPLC measurements (Figure A10).

Cell Experiments 2.3.1. Determination of Target Avidities
Several different IC 50 values for RGD derivatives have been reported in the literature, ranging from 0.1 nM up to 6.7 µM.The main reason for the observed differences is the assay method used to determine the IC 50 values.IC 50 values of 0.1-1 nM can be found for RGD peptides having been determined by ELISA assays [38] and IC 50 values around 20 nM have been reported for solid-phase α v β 3 binding assays for monomeric RGD derivatives [37].Those IC 50 values were derived by non-living experiments.However, cell experiments are closer to in vivo conditions.Therefore, for the AuNPs 7 and 8, the α v β 3 integrinavidities were determined by competitive displacement experiments on α v β 3 -expressing U87MG cells using 125 I-echistatin as the α v β 3 -specific radioligand and competitor.The RGD monomer c(RGDfK) was evaluated as an internal reference.The evaluation of RGD derivatives by displacement experiments yielded IC 50 values comparable to those reported in the literature [34].For the c(RGDfK) monomer, a mean IC 50 value of 0.7 µM was determined (Table 3, Figure A3).The multi-RGD decoration on the surface of AuNPs 7a and 7b resulted in a lower mean IC 50 value of 27.8 and 38.3 nM, respectively (Figures A4  and A5).Mean IC 50 values of 82.4 and 103.6 nM were found for AuNPs 8a and 8b, respectively (Figures A6 and A7).It was observed that the higher the loading with α v β 3specific RGD peptide, the lower the IC 50 values.AuNP-dithio-PEG-RGD low 103.6 ± 3.5

Determination of Cell Survival
Colony formation assays were performed with [ 198 Au]3 with U87MG cells.For this proof-of-concept experiment, 5-10 Gy was chosen as the incubation dose.To achieve this dose, 1-2 MBq [ 198 Au]3 per well in a 24-well plate within a 96 h incubation period was calculated using Formula (1).
Formula ( 1)-Calculation of dose to a cell monolayer at the bottom of a multi-well plate or Eppendorf tube for 198 Au using Geant4-simulation [44].D: energy dose, S: S-value, A: activity, T 1/2 : half-life of the radionuclide, t: irradiation time.
This dose corresponds to concentrations of [ 198 Au]AuNPs of 0.515-0.939µM, which is at least 10 times higher than the IC 50 of AuNP-dithio-RGD 7 and 8.It was observed that the survival fraction (sf) of the cells was significantly reduced for [ 198 Au]3 and that higher doses of 10 Gy (sf = 18.2%) were more effective in damaging the tumor cells than 5 Gy (sf = 33.9%)(Figure 4).

Discussion
c(RGDfK) is a highly potent and selective integrin α v β 3 antagonist and therefore could disrupt cell viability by inhibiting angiogenesis [45].Radiolabeled RGD derivatives can be used as tracers for tumor angiogenesis [46].Multimerization leads to better tumor accumulation [47,48].Therefore, AuNPs decorated with a multitude of c(RGDfK) motifs could lead to better tumor accumulation, which is important for therapy.
Methods for the preparation of [ 198 Au]AuNPs are already known in the literature [11,29,32].However, the synthesis starts with neutron activation of gold foil, which is then dissolved in aqua regia, followed by nanoparticle synthesis and further functionalizations with targetspecific ligands.All these steps are performed with radioactive 198 Au, resulting in higher dose accumulation for the personnel and more radioactive waste as the consequence.In this work, it was decided to first complete the synthesis of tumor-specific AuNPs with a high target avidity and high stability, and to perform neutron activation as the last step in order to reduce the personnel dose and enable a highly efficient synthesis pathway, which is mandatory for high clinical relevance.The challenge was to synthesize AuNPs that withstand neutron activation without aggregation and loss of the ligand shell.
Stable α v β 3 -specific AuNPs 7 and 8 were successfully synthesized with a better avidity compared to the monomeric peptide ligand c(RGDfK).During the irradiation experiments, it was observed that AuNPs containing monothiol ligands were unstable against neutron activation.However, all AuNP derivatives containing only dithiol ligands were stable against neutron activation even at the highest concentrations and irradiation times (~7.5 mg/mL within 2 h).It is known that sulfur can also be activated by neutrons via the 32 S(n,p) 32 P reaction [49,50].Presumably, once a sulfur atom is activated to 32 P, it loses its covalent bond to the AuNP surface and a monothiol ligand is lost to the environment.In contrast, a dithiol ligand could remain bound to the surface even if a binding interaction is lost by activation of one of the sulfur atoms.
To determine the therapeutic influence of [ 198 Au]AuNPs, cell survival was addressed by a colony formation assay.The activity and incubation time to reach relevant doses between 5 and 10 Gy were calculated for monolayer cell culture in 24-well plates (Formula (1)).To reach these doses of 5-10 Gy concentrations of a factor >10 times higher than the IC 50 for AuNP-RGD 8a and 8b had to be used within 96 h of incubation.Therefore, cell viability should be considered to be very low when using such high concentrations of [ 198 Au]8 in cell survival experiments, as the antagonist RGD may interfere with angiogenesis and thus cell viability [45].To circumvent this problem, future experiments should use higher activity concentrations (due to longer activation of AuNPs) or longer cell incubation with lower doses >5 Gy, when evaluating cell survival.However, in the proof-of-concept cell survival experiments, non-specific [ 198 Au]3 showed a significant effect on U87MG cells with a survival fraction as low as 18.2% at 10 Gy.Therefore, the combination of β − -emission from 198 Au and the antagonistic effect of RGD could dramatically reduce the therapeutically relevant dose of applied [ 198 Au]AuNP-RGDs.

Materials and Methods
General procedures.All reagents and solvents were purchased from commercial suppliers (Sigma, Merck) and were used without further purification.NMR spectra were recorded on a 300 MHz Mercury Plus and a 500 MHz NMR System spectrometer (Varian, Palo Alto, CA, USA).Chemical shifts (δ) are given in ppm and are referenced to the residual solvent resonance signals relative to (CH 3 ) 4 Si ( 1 H, 13 C).Mass spectra were obtained on a microflex MALDI-TOF mass spectrometer (Bruker Daltonics, Bremen, Germany) and HR-ESI-MS spectra on a LTQ FT Ultra Fourier Transform Ion Cyclotron Resonance spectrometer (Thermo Finnigan, Dreieich, Germany).When applicable, purity was determined by HPLC.The purity of all final compounds was 95% or higher.HPLC was performed on a Dionex UltiMate 3000 HPLC system (Thermo Scientific, Dreieich, Germany), equipped with a reverse phase column (Analytical: Merck Chromolith RP-18e; 100 × 4.6 mm plus a guard column 5 × 4.6 mm; semipreparative: Chromolith RP-18e; 100 × 10 mm plus a guard column 10 × 4.6 mm), and a UV-diode array detector (210 nm, 254 nm).The solvent system used was a gradient of acetonitrile:water (containing 0.1% TFA) (0-5 min: 0-100% MeCN) at a flow rate of 4 mL/min, unless otherwise stated.The purity and stability of AuNPs/[ 198 Au]AuNPs were investigated by size exclusion HPLC using a PolySep™-SEC GFC-P 4000, LC column 300 × 7.8 mm, and a 35 mm PolySep guard column (Phenomenex, Aschaffenburg, Germany) with water (0.8 mL/min) as eluent (Figure A10).Purification of AuNPs was performed by dialysis (tubes with molecular weight cut-off of 14,000 g/mol, Visking, Roth, Karlsruhe, Germany) against distilled water and by sizeexclusion chromatography using Sephadex G25 PD10 columns (Fisher Scientific, Schwerte, Germany) and distilled water as eluent.
A brief description of the AuNP syntheses can be found in Appendix A. Determination of the number of ligands on the surface of the AuNPs.The thermogravimetric analyses were performed using a Mettler Toledo TGA 2 STAR e system.AuNPs (1-2 mg) were weighed into 70-µL-aluminum oxide crucibles (Mettler Toledo, Gießen, Germany) and heated from 25-750 • C (10 K/min) in a stream of N 2 or CO 2 (30 mL/min).The loading of the different AuNPs is shown in Table 1 and was calculated from the different mass losses, which increase as the AuNPs are functionalized.Therefore, the amount of different ligands per particle can be calculated according to the formula of Zhu et al. [42].Since the nanoparticles have an average diameter of ~3 nm, the calculated amount of gold atoms is ~834 Au atoms per nanoparticle.This gives a molecular weight of an AuNP of 164,298 g/mol.Using TGA, the following ligand numbers were determined:
Avidity experiments.The α v β 3 -binding affinities of the respective RGD peptides and AuNPs were determined using in vitro competitive displacement experiments on U87MG tumor cells (HTB-14, ATCC ® , Manassas, VA, USA).U87MG cells were harvested and resuspended in the binding buffer at a cell concentration of 2 × 10 6 /mL to reach 10 5 cells per well.
Neutron irradiation experiments.Production of [ 198 Au]AuNPs by neutron activation of 0.05-15.5 mg AuNPs was performed in pneumatic transfer tube one for 1-2 h at 100 kW with a thermal neutron flux of 1.6 × 10 12 cm −2 × s −1 in the TRIGA research reactor (Mainz, Germany).For calibration of the dose calibrator ISOMED 2010 (NUVIA Instruments, Dresden, Germany) 12.7 mg solid Au was irradiated for 1 h to produce 87 MBq (calculated) [ 198 Au]Au with a measured dose rate of 57 µSv/h.26 h later, the activity was measured with the dose calibrator, and 60 MBq was obtained (using the 137 Cs-channel, 66 MBq calculated).In addition, the solid [ 198 Au]Au (40 MBq) was carefully dissolved in 2 mL aqua regia at 50 • C within 15 min in order to find the correct calibration factors of the dose calibrator for different volumes in vials and syringes.
Colony formation assay.Three days before the experiments, 150,000 cells were seeded in 24-well plates.U87MG cells were incubated for 96 h in the presence of the α v β 3 -specific or non-radioactive AuNPs or 1-2 MBq [ 198 Au]AuNPs to achieve the calculated doses of 5-10 Gy.After incubation, the cell medium was removed, the cells were washed and harvested, and a colony formation assay was performed in triplicate for each irradiation point with 1000 cells per well in a 6-well plate.Colonies were cultured in cell medium for 28 days, then washed with 1 mL PBS, fixed with 2 mL 4% formaldehyde in PBS for 15 min, and incubated with 2 mL 0.5% crystal violet dye solution for 30 min.Afterward, colonies were washed with distilled water, dried, and counted by light microscopy.Colonies of more than 50 cells were considered viable, and the plating efficiency for each sample was estimated based on the initial number of cells seeded.Clonogenic cell survival was calculated as the relative plating efficiency of treated versus untreated samples.Triplicate samples were prepared for each treatment and experimental condition.

Conclusions
α v β 3 -specific RGD-containing AuNPs with a higher target avidity compared to α v β 3specific RGD were successfully synthesized.This proof-of-concept work should demonstrate, that activation of AuNPs with a ligand shell is possible without losing their organic shell and integrity.Irradiation experiments demonstrated the stability and consistency of [ 198 Au]AuNPs with dithiol ligands compared to [ 198 Au]AuNPs with monothiol ligands, which always aggregated at each applied concentration after neutron activation.In vitro experiments determine the therapeutic effect of [ 198 Au]AuNPs by addressing the survival fraction of U87MG cells proved a significant influence on cell death.Therefore, the [ 198 Au]AuNPs could serve as a tool for endoradiotherapy.

TA-PEG 4 -c(RGDfK) 5
TA-NH-PEG 4 -COOH dissolved in 0.5 mL of DMF (1.1 eq., 0.055 mmol, 22.4 mg) was added to a solution of PyBOP (1.9 eq., 0.094 mmol, 49.1 g) in 0.5 mL of DMF.DIPEA (3 eq., 0.149 mmol, 26 µL) was added and the mixture was stirred at ambient temperature until the reaction was completed.DMF was then removed under reduced pressure, and the crude product was purified by semi-preparative HPLC.After lyophilization the product TA-NH-PEG 4 -c(RGDfK) was obtained as colorless solid (yield: 28.91%, 14.AuNP-dithio-PEG 3 -OH 3 [40] Hydrogen tetrachloroaurate (1 eq., 0.525 mmol, 207 mg) was dissolved in 205 mL MeOH to give a bright yellow solution and under stirring a solution of TA-NH-PEG 3 -OH dissolved in 205 mL of MeOH was added and stirred for 2 h until the reaction color became nearly colorless.A solution of sodium borohydride dissolved in 20.5 mL of water was quickly added to the reaction.The solution immediately turned black.The reaction mixture was stirred overnight, MeOH was removed under reduced pressure, and the residue was redissolved in 9 mL of tracepure water and dialyzed in distilled water for 4 days.After lyophilisation TA-AuNP was obtained as black powder.(39.47%, 180.7 mg).
AuNP-thio-PEG 6 [6] General procedure for the preparation of PEGylated AuNPs: Briefly, hydrogen tetrachloroaurate(III) trihydrate (1 eq., 0.4 mmol, 157.5 mg, ≥99.9% trace metal basis) was dissolved in 12.5 mL of trace pure water resulting in a bright yellow solution, and then extracted by mixing with 125 mL of a tetraoctylammonium bromide (TOABr, 1.2 eq., 0.48 mmol, 263 mg) toluene solution.The contents were stirred vigorously for 20 min at room temperature to facilitate the phase transfer of the Au(III) into the toluene layer, which resulted in the organic layer turning to a dark orange color and the aqueous layer becoming clear colorless.After the phase transfer was complete, the aqueous layer was removed.The organic layer was dried with MgSO 4 and filtered to remove excess of water.The solution was cooled to 0 • C in an ice bath.Freshly-prepared HO-PEG 3 -thiol (3 eq., 1.2 mmol, 199 mg) in 6.3 mL of dichloromethane was added and stirred until the orange solution turned to colorless (~1 h).A fresh solution of tetrabutylammonium borohydride (TBABH) (10 eq., 4.0 mmol, 1.03 g) in 6.3 mL dichloromethane was then added to the rapidly stirring toluene solution over 5 s.The solution immediately turned black.The PEG-AuNP began to precipitate from the toluene after 1 h.After stirring the mixture for 16 h from 0 • C to 20 • C, 6.3 mL of trace pure water was added under slow stirring to extract the PEGylated AuNPs for 120 min.The organic layer was decanted and the aqueous layer was washed alternately with 3 × 13 mL toluene/1.3mL MeCN and 3 × 13 mL toluene/1.3mL isopropanol.The black aqueous layer was transferred to a Visking cellulose dialysis tube (molecular cut-off 14,000 Da) with 3 × 6.3 mL trace pure water and dialyzed in 3 × 10 L of distilled water for 1 h, 2.5 h and 16 h.The AuNPs were lyophilized to yield 69.5 mg (25.10%) of black powder.

AuNP-PEG-RGDs by ligand exchange
General procedure for the preparation of RGD-decorated AuNPs: Briefly, the functionalization of AuNPs 3 and 6 was performed by a place-exchange reaction with TA-PEG 4c(RGDfK) 5. TA-PEG 4 -c(RGDfK) 5 was dissolved in a mixture trace pure H 2 O:MeOH (1:1) and was added to AuNPs 3 or 6 in 2 mL trace pure H 2 O and stirred overnight.Purification of AuNPs was performed in two steps: First, the AuNP solution was transferred into a Visking cellulose dialysis tube (molecular cut-off 14,000 Da) with 3 × 1 mL trace pure water and dialyzed in distilled water for 4 days, and second, the AuNP solution was eluted by size-exclusion chromatography using Sephadex G25 PD10 columns and distilled water.The AuNPs were then lyophilized and obtained as a black powder.
Appendix A.2. Electron Microscopy AuNP samples were diluted at will in deionized water (fade red solution), particles were adsorbed onto glow-discharged carbon-coated EM grids and directly observed by TEM (Zeiss EM912, Carl Zeiss Oberkochen, Germany).Images were digitally captured with a CCD camera (Sharp eye, TRS, Moorenweiss, Germany).Particle number and size were measured using the FIJI software (v1.50e).

Figure A1 .
Figure A1.Transmission electron microscope image of AuNP 6 and corresponding histogram of AuNP diameter distribution.

Figure A2 .
Figure A2.Transmission electron microscope image of AuNP 3 and corresponding histogram of AuNP diameter distribution.

Figure A9 .
Figure A9.Representative absorption spectrum of [ 198 Au]3 at different time points after neutron activation with a typical absorption maximum for ultrasmall AuNPs at 514 nm indicating no aggregation of the particles.

Table 1 .
Calculated number of ligands and resulting molecular mass of the AuNPs.

Table 2 .
Summary of neutron activation of various AuNPs, mass, and calculated half-life.