Core–shell structured gold nanoparticles as carrier for ¹⁶⁶Dy/¹⁶⁶Ho in vivo generator

Abstract

Background

Radionuclide therapy (RNT) has become a very important treatment modality for cancer nowadays. Comparing with other cancer treatment options, sufficient efficacy could be achieved in RNT with lower toxicity. β⁻ emitters are frequently used in RNT due to the long tissue penetration depth of the β⁻ particles. The dysprosium-166/holmium-166 (¹⁶⁶Dy/¹⁶⁶Ho) in vivo generator shows great potential for treating large malignancies due to the long half-life time of the mother nuclide ¹⁶⁶Dy and the emission of high energy β⁻ from the daughter nuclide ¹⁶⁶Ho. However, the internal conversion occurring after β⁻ decay from ¹⁶⁶Dy to ¹⁶⁶Ho could cause the release of about 72% of ¹⁶⁶Ho when ¹⁶⁶Dy is bound to conventional chelators. The aim of this study is to develop a nanoparticle based carrier for ¹⁶⁶Dy/¹⁶⁶Ho in vivo generator such that the loss of the daughter nuclide ¹⁶⁶Ho induced by internal conversion is prevented. To achieve this goal, we radiolabelled platinum-gold bimetallic nanoparticles (PtAuNPs) and core–shell structured gold nanoparticles (AuNPs) with ¹⁶⁶Dy and studied the retention of both ¹⁶⁶Dy and ¹⁶⁶Ho under various conditions.

Results

The ¹⁶⁶Dy was co-reduced with gold and platinum precursor to form the ¹⁶⁶DyAu@AuNPs and ¹⁶⁶DyPtAuNPs. The ¹⁶⁶Dy radiolabelling efficiency was determined to be 60% and 70% for the two types of nanoparticles respectively. The retention of ¹⁶⁶Dy and ¹⁶⁶Ho were tested in MiliQ water or 2.5 mM DTPA for a period of 72 h. In both cases, more than 90% of both ¹⁶⁶Dy and ¹⁶⁶Ho was retained. The results show that the incorporation of ¹⁶⁶Dy in AuNPs can prevent the escape of ¹⁶⁶Ho released due to internal conversion.

Conclusion

We developed a chelator-free radiolabelling method for ¹⁶⁶Dy with good radiolabelling efficiency and very high stability and retention of the daughter nuclide ¹⁶⁶Ho. The results from this study indicate that to avoid the loss of the daughter radionuclides by internal conversion, carriers composed of electron-rich materials should be used.

Supplementary Information

The online version contains supplementary material available at 10.1186/s41181-022-00170-3.

Introduction

Cancer is one of the leading causes of death in the world (Sung et al. 2021; Bray et al. 2018). Nowadays, surgery and external beam radiation therapy (EBRT) are still the most common treatment modalities for localized tumors. In the case of metastases, systemic treatments such as radionuclide therapy (RNT) are preferred. RNT has been proved to be able to significantly prolong the life expectancy of terminal patients without affecting quality of life (Pool et al. 2010; Humm et al. 2015). In RNT, the therapeutic radionuclides are usually linked to chelators conjugated to tumor targeting vectors such as peptides, nucleotides and antibodies. Once distributed to the tumor site, the ionizing radiation emitted by the radionuclides can damage the DNA of the cancer cells and lead to apoptosis (Gudkov et al. 2016; Dash et al. 2013; Sgouros et al. 2020; Tafreshi et al. 2019; Kostelnik and Orvig 2019).

Over the past decades, many radiopharmaceuticals have been developed and some of them have been already applied in the clinic (Suman et al. 2021). Radionuclides that emit β⁻ particles are more commonly applied in the clinic but the interest in α emitters is also growing (Nelson et al. 2021; Tafreshi et al. 2019). Since β⁻ particles have relatively long tissue penetration depth, they are suitable for treating larger metastases (Pouget et al. 2011; Marcu et al., 2018). Moreover, additional benefits can be achieved with β⁻ emitters by the so called “cross-fire” effect, i.e. due to the long range of β⁻ particles, it is not essential to target every single tumor cell to efficiently irradiate the whole tumor (Pouget et al. 2011).

Holmium-166 (¹⁶⁶Ho) is a β⁻ emitter that decays to ¹⁶⁶Er with a half-life time of 26.8 h and emits β⁻ particles with maximum energy of 1.85 MeV (Fig. 1a). The high energy of the β⁻ particles results in a maximum tissue penetration depth of 8.7 mm which makes ¹⁶⁶Ho a promising radionuclide for treating larger malignancies (Klaassen et al. 2019). In addition, ¹⁶⁶Ho can also be imaged by single-photon emission computed tomography (SPECT) due to its gamma emission at 80.57 keV (Elschot et al. 2011).¹⁶⁶Ho is generally produced by the neutron activation of ¹⁶⁵Ho following the (n, γ) reaction. An alternative route for ¹⁶⁶Ho production is the ¹⁶⁶Dy/¹⁶⁶Ho generator (Klaassen et al. 2019). Dysprosium-166 (¹⁶⁶Dy) has a half-life time of 81.6 h, decays to ¹⁶⁶Ho via β⁻ decay and can be produced by a double neutron capture reaction from ¹⁶⁴Dy (Fig. 1b). ¹⁶⁶Dy/¹⁶⁶Ho can also serve as in vivo generator which is capable of delivering higher radiation dose per administrated activity due to the three times longer half-life time of ¹⁶⁶Dy than ¹⁶⁶Ho (Poty et al. 2018; Baidoo et al. 2013; Edem et al. 2016). Therefore, better treatment outcome could be expected by using ¹⁶⁶Dy/¹⁶⁶Ho in vivo generator instead of the direct administration of ¹⁶⁶Ho.

Fig. 1.

a Decay scheme of ¹⁶⁶Dy and ¹⁶⁶Ho including the major transitions. b The double-neutron capture nuclear reaction of ¹⁶⁴Dy to produce ¹⁶⁶Dy and the corresponding cross-sections

However, Zeevaart et al. reported the radiolabelling of ¹⁶⁶Dy on dodecane tetraacetic acid (DOTA) and surprisingly found that about 72% of the daughter ¹⁶⁶Ho was released from the ¹⁶⁶Ho-DOTA complex (Zeevaart et al. 2012). The ¹⁶⁶Ho loss was attributed to the de-excitation of ¹⁶⁶Ho^* via internal conversion instead of γ emission. Internal conversion is a process where the excited daughter nucleus electromagnetically interacts with inner orbital electrons and results in the emission of an inner electron from K shell or L shell along with the creation of electron vacancies. The electrons from the outer shells will be reorganized to fill in the vacancies while emitting auger electrons as well as characteristic X-ray. As the result of the emission of auger electrons, the de-excited ¹⁶⁶Ho ions become highly charged and will extract electrons from the surrounding environment (i.e. DOTA). Due to the electron transfer to ¹⁶⁶Ho, the DOTA component also becomes positively charged while the ¹⁶⁶Ho ion acquires its original oxidation state (+ 3). The repulsion force between the two components having the same charge results in the rupture of the bonds between ¹⁶⁶Ho and DOTA. Thus, ¹⁶⁶Ho is released as free ion. The theoretical calculation predicts 73.6% ¹⁶⁶Ho release which matches well with the published experimental results (Zeevaart et al. 2012). Being an isotope of a lanthanide element, free ¹⁶⁶Ho tends to accumulate in liver, kidney, spleen and bone and may cause severe side effect to the patient (Suzuki et al. 1998). Therefore, to implement the ¹⁶⁶Dy/¹⁶⁶Ho in vivo generator in the clinic, a carrier that can prevent the loss of the internally converted ¹⁶⁶Ho has to be developed.

Nowadays, the medical application of different types of nanoparticles has been extensively reported for diagnostics and the treatment of cancer and other diseases (Mitchell et al. 2021; Pelaz et al. 2017; Thomas and Weber 2019; Shi et al. 2017; Wong et al. 2020). Gold nanoparticles (AuNP) have shown great potential as carriers for anti-cancer agents due to their unique properties such as biocompatibility, precisely controlled size and the possibility of easy surface modification (Singh et al. 2018; Boisselier and Astruc 2009). Besides using AuNP as carriers for conventional payloads, multiple reports on the chelator-free labelling of medical radionuclides on AuNP have been published (Ge et al. 2020; Silva et al. 2021). In these studies, radionuclides in the form of metallic ions or halogen ions are either co-reduced into the lattice of AuNP (e.g. ⁶⁴Cu (Frellsen et al. 2016; Sun et al. 2014; Pretze et al. 2019; Zhao et al. 2014), ¹¹¹In (Zheng et al. 2021) and ⁶⁸ Ga (Zheng et al. 2021)) or chemically absorbed on the surface of AuNP (¹²⁵I, ¹²⁴I (Lee et al. 2016; Lee et al. 2017) and ²¹¹At (Dziawer et al. 2017)). In most cases, the radiolabelling stability and the tumor uptake of the radionuclides appear to be improved after being loaded on AuNPs when compared to the common chelator approaches (Pretze et al. 2019). The improved tumor uptake is likely from the prolonged circulation time of AuNPs comparing with small molecules. However, the toxicity of AuNP itself have to be considered even gold is considered to be biocompatible (Ranjbar Bahadori et al. 2021).

In this study, we developed a chelator-free radiolabelling method to incorporate ¹⁶⁶Dy in AuNP. In this radiolabelling method, we co-reduced ¹⁶⁶Dy³⁺ ions with gold and platinum precursors to form either a bimetallic (¹⁶⁶DyAuNP) or trimetallic (¹⁶⁶DyPtAuNP) nanoparticle. In addition, an extra gold layer was added to the ¹⁶⁶DyAuNP to form a core–shell structured ¹⁶⁶DyAu@AuNP. We first characterized the physical properties of the DyAu@AuNP and DyPtAuNP with non-radioactive Dy. Then the radiolabelling of ¹⁶⁶Dy was performed and the retention of ¹⁶⁶Ho on ¹⁶⁶DyAu@AuNP and ¹⁶⁶DyPtAuNP was evaluated.

Methods and materials

Materials

Gold(III) chloride trihydrate (≥ 99.9%, HAuCl₄ · 3H₂O), chloroplatinic acid hexahydrate (≥ 37.50% Pt, H₂PtCl₆ · 6H₂O), sodium borohydride (≥ 98.0%, NaBH₄), cetyltrimethylammonium bromide (≥ 98%, CTAB), cetyltrimethylammonium chloride solution (25 wt.% in water, CTAC), L-Ascorbic acid (≥ 99%, AA), sodium hydroxide (NaOH) and dysprosium(III) chloride hexahydrate (≥ 99.9%) were purchased from Sigma-Aldrich (Zwijndrecht, the Netherlands). 90% enriched dysprosium-164 oxide powder (¹⁶⁴Dy₂O₃) was obtained from Oak Ridge National Laboratory (sample number 122502, ORNL, Tennessee, USA). Ethylenediaminetetraacetic acid disodium salt dihydrate (Na₂EDTA · 2H₂O), Diethylenetriamine pentaacetate (DTPA), hydrochloric acid (HCl, 30%, Suprapur®) and nitric acid (HNO₃, 69%, Supelco®) was supplied by Merck. All chemicals were used as received without further purification. MiliQ water was obtained from an in-house MiliQ system (Millipore) and used throughout this study.

Production of ¹⁶⁶Dy

¹⁶⁶Dy was produced by the double neutron capture reaction of ¹⁶⁴Dy. 3 mg 90% enriched ¹⁶⁴Dy₂O₃ powder was irradiated in the reactor facilities of the SCK•CEN—BR2 Reactor (Mol, Belgium), the Institute of Energy Security and Environmental Safety Centre for Energy Research (Budapest, Hungary) or the nuclear reactor research facility (HOR, Hoger Onderwijs Reactor) at the Department of Radiation Science and Technology of the Delft University of Technology (Delft, the Netherlands). The obtained ¹⁶⁶Dy₂O₃ powder was dissolved in 5 ml 1 M HCl under mild heating to prepare a stock solution of ¹⁶⁶DyCl₃. 2.5 ml of the stock solution was transferred to a 20-ml glass vial and the pH of the stock solution was adjusted to ~ 5.5 by adding 2.35 ml of 1 M NaOH solution (checked by pH test paper). The activity of ¹⁶⁶Dy and ¹⁶⁶Ho in the stock solution was measured on a calibrated well-type HPGe detector (Canberra).

Synthesis of AuNP seed

The synthesis was adapted from a published protocol with some changes (Zheng et al. 2014) and is schematically illustrated in Fig. 2. The AuNP seeds were synthesized by the reduction of HAuCl₄ by NaBH₄ using CTAB as capping agent. 0.1 ml 25 mM HAuCl₄, 4 ml 250 mM CTAB and 5.9 ml MiliQ water was added to a glass vial and mixed for 10 min. 0.6 ml freshly prepared, ice-cold 10 mM NaBH₄ solution was added to the mixture dropwise under vigorous stirring. The color of the solution changed from yellow to dark brown rapidly. The obtained AuNP seeds were left undisturbed at 27 °C for 1.5 h before further usage.

Fig. 2.

Schematic illustration of the synthesis of ¹⁶⁶DyAu@AuNP and ¹⁶⁶DyPtAuNP

Growth of AuNP seed to 5 nm AuNP

2 ml 200 mM CTAC, 1.5 ml 100 mM AA and 1 ml AuNP seed dispersion were added to a glass vial and mixed for 5 min at 27 °C. 2 ml 0.5 mM HAuCl₄ was then added in one-shot by a pipet. The reaction was continued at 27 °C for another 15 min.

Synthesis of 5 nm non-radioactive DyAu@AuNP

33 μl, 20 μl or 10 μl 25 mM DyCl₃ solution (pH 5.5) was mixed with 0.1 ml 25 mM HAuCl₄ and 4 ml 250 mM CTAB in a glass vial to achieve the Dy:Au (n:n) feeding ratio of 1:3, 1:5 or 1:10. The total volume was adjusted to 10 ml by MiliQ water. 0.6 ml ice-cold 10 mM NaBH₄ solution was then added to the mixture dropwi under vigorous stirring. The growth of the DyAuNP seed to 5 nm core–shell structured DyAu@AuNP was performed in the same way as the growth of AuNP seed to 5 nm AuNP after aging the DyAuNP seed at 27 °C for 1.5 h.

Synthesis of non-radioactive DyPtAuNP

10 μl 25 mM H₂PtCl₆, 90 μl 25 mM HAuCl₄ and 4 ml 250 mM CTAB were mixed in a glass vial. 33 μl or 10 μl 25 mM DyCl₃ solution (pH 5.5) was then added to achieve the Dy:(Pt + Au) (n/n) feeding ration of 1:3 or 1:10. MiliQ water was added to adjust the total volume to be 10 ml and stirred for 10 min. 0.6 ml freshly prepared, ice-cold 10 mM NaBH₄ solution was added to the mixture dropwise under vigorous stirring. The colour of the solution changed from yellow to dark brown rapidly. The obtained DyPtAuNP was left undisturbed at 27 °C for 1.5 h before purification.

Synthesis of 5 nm ¹⁶⁶DyAu@AuNP

88.2 μl stock solution of ¹⁶⁶DyCl₃ containing approximately 0.134 MBq ¹⁶⁶Dy and 0.2 MBq ¹⁶⁶Ho was mixed with 0.1 ml 25 mM HAuCl₄ and 4 ml 250 mM CTAB in a glass vial. 30.1 μl, 17.1 μl or 7.1 μl of 25 mM DyCl₃ solution (pH 5.5) was then added to ensure the Dy:Au (n:n) feeding ratio to be 1:3, 1:5 or 1:10. The synthesis of ¹⁶⁶DyAuNP seed and growth to ¹⁶⁶DyAu@AuNP was performed in the same way as non-radioactive DyAu@AuNP which was described above.

Synthesis of ¹⁶⁶DyPtAuNP

88.2 μl stock solution of ¹⁶⁶DyCl₃ containing approximately 0.134 MBq ¹⁶⁶Dy and 0.2 MBq ¹⁶⁶Ho was mixed with 10 μl 25 mM H₂PtCl₆, 90 μl 25 mM HAuCl₄ and 4 ml 250 mM CTAB in a glass vial. 30.1 μl or 7.1 μl of 25 mM DyCl₃ solution (pH 5.5) was then added to ensure Dy:Au (n:n) feeding ratio of 1:3 or 1:10. The final volume was then adjusted to 10 ml by MiliQ water and stirred for 10 min. 0.6 ml freshly prepared, ice-cold 10 mM NaBH4 solution was added to the mixture dropwise under vigorous stirring. The obtained ¹⁶⁶DyPtAuNP was left undisturbed at 27 °C for 1.5 h before purification.

Characterization of non-radioactive nanoparticles

The morphology and size of the AuNP, DyAu@AuNP and DyAuNP was determined with a JEM-1400 Plus transmission electron microscope (TEM, JEOL) at the acceleration voltage of 120 kV. The UV–vis absorption spectra of AuNPs were measured by a UV–VIS-NIR spectrophotometer (UV-6300PC, VWR). The hydrodynamic radius of the samples was determined by dynamic light scattering (DLS) which consisted of a JDS uniphase 633 nm 35 mW laser source, an ALV sp 125 s/w 93 goniometer, a fibre detector and a Perkin Elmer photo counter. The data was fitted using the CONTIN method and the Stokes–Einstein equation (Eq. 1) was used to determine the hydrodynamic radius of the nanoparticles.

$$R_H=\frac{\mathit{kT}}{6\pi \eta D}$$ 1

Determination of ¹⁶⁶Dy radiolabelling efficiency

100 μl 100 mM EDTA or 100 mM DTPA was added to the ¹⁶⁶DyAu@AuNP and ¹⁶⁶DyPtAuNP samples and incubated at 27 °C for 30 min to bind with free ¹⁶⁶Dy³⁺ ions. Then the samples were centrifuged (4000 rpm, 10 min) and washed three times using spin filters (MWCO 10 KDa, Amicon). The final volume of the washed samples was adjusted to 4 ml by MiliQ water and stored at 37 °C. The counts of the nanoparticles and filtrates (¹⁶⁶Dy-EDTA) of all samples were measured by an automatic gamma counter (Wallac Wizard² 2480, Perkin Elmer) or a low energy Ge-detector (GL2020R, Canberra). ¹⁶⁶Dy was measured using its gamma emission at 425.99 keV. The radiolabelling efficiency of ¹⁶⁶Dy was calculated by the following formula: Counts(NPs)/[Counts(NPs) + ∑Counts(filtrate)] × 100%.

Determination of ¹⁶⁶Ho and ¹⁶⁶Dy retention

To assess the stability of ¹⁶⁶Ho and ¹⁶⁶Dy on nanoparticles, the samples were dispersed in 4 ml MiliQ water or 2.5 mM DTPA (pH 7.5) and incubated at 37 °C for 24, 48 and 72 h. At each time point, the samples were collected and washed by MiliQ water using spin filters under centrifugation (4000 rpm, 10 min). The counts of the nanoparticles and the filtrate was measured to calculate the retention of both ¹⁶⁶Ho and ¹⁶⁶Dy.

Determination of gold and dysprosium content in ¹⁶⁶DyAu@AuNP and ¹⁶⁶DyPtAuNP

1 ml of each completely decayed sample was completely destructed in 1 ml aqua regia (HCl/HNO₃ = 3:1) and diluted by MiliQ water to a final volume of 10 ml. The concentration of Au and Dy were then measured by ICP-OES (Optima 8000, Perkin Elmer).

Results

Synthesis and characterization of non-radioactive DyAu@AuNP

In this study, we designed a core–shell structured AuNP to function as the carrier for ¹⁶⁶Dy /¹⁶⁶Ho in vivo generator. The gold precursor was first co-reduced with ¹⁶⁶Dy³⁺ ions to form the ¹⁶⁶DyAuNPs. Subsequently, an extra gold shell was grown by reducing gold precursor with ascorbic acid to prevent the possible escape of free ¹⁶⁶Ho³⁺ ions. Besides assisting to retain free ¹⁶⁶Ho³⁺ ions, the growth of an extra gold layer can also improve the colloidal stability of the DyAuNPs (Zheng et al. 2014). It is important that the original physiochemical properties of the AuNPs are not altered upon ¹⁶⁶Dy encapsulation. Thus, we first performed a pilot study with non-radioactive DyCl₃ to explore the influence of Dy content on the physical properties of the DyAu@AuNPs. The non-radioactive DyCl₃ was co-reduced with HAuCl₄ by a strong reducing agent NaBH₄ to form the bimetallic DyAuNPs. Three samples with different Dy:Au feeding ratios of 1:3, 1:5 and 1:10 were prepared. An extra gold shell was then grown on the seed particles via the reduction of HAuCl₄ at lower concentration using ascorbic acid and resulting in the core–shell structured DyAu@AuNPs. The non-incorporated Dy³⁺ ions were removed by incubating DyAu@AuNPs with EDTA or DTPA, followed by multiple cycles of washing with MiliQ water. Au@AuNP without Dy content was also prepared with the same method and used as the control group.

The size and shape of the DyAu@AuNPs were characterized by transmission electron microscope (TEM). As shown in Fig. 3a–d, DyAu@AuNPs with varying Dy:Au feeding ratios as well as the Au@AuNP all showed a diameter of 4.9 nm. The hydrodynamic radius (R_H) of the DyAu@AuNPs and Au@AuNP was measured by dynamic light scattering (DLS). As shown in Fig. 3e, the intensity weighted R_H was determined to be within the range of 12 ~ 14 nm for both the DyAu@AuNPs and the Au@AuNP. The hydrodynamic radius of the DyAu@AuNPs was found to be larger than the radius measured by TEM, since DLS measures the hydration layer formed around CTAB/CTAC on the surface of the nanoparticles. Due to the surface plasmon resonance (SPR) effect of AuNPs, the characteristic UV–vis spectrum can be used as an indication of the size of AuNPs (Barbosa et al. 2010). The UV–vis spectrum of the DyAu@AuNPs and Au@AuNP is shown in Fig. 3f. The wavelength of the SPR peak (λ_SPR) of all samples were detected near 520 nm, indicating that the DyAu@AuNPs all had comparable size to the Au@AuNP. All results of the characterization of the DyAu@AuNPs and Au@AuNP are summarized in Table 1.

Fig. 3.

Characterization of DyAu@AuNPs. a–d Representative TEM image of samples with different Dy:Au feeding ratios: Dy:Au = 1:3 (a), 1:5 (b), 1:10 (c), no Dy addition (d). Scale bar is 20 nm. See supporting information for size distribution histograms (Additional file 1: Fig. S1). e Hydrodynamic radius (R_H) of the samples measured by DLS. f) UV–vis spectrum of the DyAu@AuNPs

Table 1.

Summary of the physical properties of DyAu@AuNPs with different Dy:Au feeding ratios

	No Dy	1:3	1:5	1:10
d (nm)	4.9 ± 0.8	4.9 ± 0.6	4.9 ± 0.7	4.9 ± 0.7
RH (nm)	12.3 ± 0.3	12.9 ± 0.3	14.1 ± 0.5	11.8 ± 0.3
λSPR (nm)	518	522	519	520

Based on the TEM, UV–vis and DLS measurements, we conclude that incorporating different amounts of Dy into the gold nanoparticle had no influence on the final size and shape of the core–shell structured DyAu@AuNPs.

Synthesis and characterization of non-radioactive DyPtAuNP

To better understand the behavior of ¹⁶⁶Ho and ¹⁶⁶Dy on nanoparticles and check if internally converted ¹⁶⁶Ho can be retained even without the extra gold layer, we attempted to directly use the DyAuNPs as the carrier for ¹⁶⁶Dy. However, the DyAuNPs were not stable and aggregated to larger AuNPs within 24 h (Additional file 1: Fig. S3). To improve the colloidal stability of DyAuNPs, we hereby prepared trimetallic DyPtAuNPs by replacing 10% of Au with Pt while the Dy:(Au + Pt) feeding ratios was still set to be 1:3 and 1:10. PtAuNP with no Dy content was also prepared and used as control. The size of the DyPtAuNPs as determined by TEM are shown in Fig. 4a–c. The diameter of the DyPtAuNP with Dy feeding ratio of 1:3 was measured to be 4.4 ± 1.1 nm which is comparable to the PtAuNP (4.0 ± 1.3 nm). However, larger particles (d = 6.6 ± 1.8 nm) were measured for the DyPtAuNP with Dy feeding ratio of 1:10. The UV–vis spectrum of the DyPtAuNPs and PtAuNP is given in Fig. 4d. The absence of SPR peak near 500 nm further confirmed the small size of the nanoparticles (Alric et al. 2013).

Fig. 4.

Characterization of DyPtAuNPs. a–c Representative TEM image of samples with different Dy:Au feeding ratios: Dy:Au = 1:3 (a), 1:10 (b) and no Dy addition (c). Scale bar is 50 nm. See supporting information for size distribution histograms (Additional file 1: Fig. S2). d UV–vis spectrum of the DyPtAuNPs

Radiolabelling of ¹⁶⁶Dy on DyAu@AuNP and DyPtAuNP

The radiolabelling of ¹⁶⁶Dy was carried out by a similar method as used for the preparation of non-radioactive DyAu@AuNPs and DyPtAuNPs. The Dy source was changed to a mixture of non-radioactive DyCl₃ and ¹⁶⁶DyCl₃ stock solution containing 0.134 MBq ¹⁶⁶Dy. Due to the decay of ¹⁶⁶Dy, ¹⁶⁶Ho was also present in the stock solution of ¹⁶⁶DyCl₃. Considering the trace amount of ¹⁶⁶Ho³⁺ ions, we expect that this to have negligible influence on the formation of the NPs. Three independent samples of ¹⁶⁶DyAu@AuNP and ¹⁶⁶DyPtAuNP with different Dy:Au feeding ratios were prepared and washed thoroughly by EDTA/DTPA and MiliQ water to remove all unbounded ¹⁶⁶Dy. The ¹⁶⁶Dy radiolabelling efficiency was calculated by comparing the counts of nanoparticles and the washing solution at 425.99 keV. The calculated results are shown in Fig. 5. Radiolabelling efficiency of 60% and 70% was achieved for ¹⁶⁶DyAu@AuNPs and ¹⁶⁶DyPtAuNPs respectively. No significant difference of the radiolabelling efficiency was found among the groups with different Dy:Au feeding ratios.

Fig. 5.

¹⁶⁶Dy radiolabelling efficiency of ¹⁶⁶DyAu@AuNP (a) and ¹⁶⁶DyPtAuNP (b) with different Dy:Au feeding ratios. The error bars represent the standard deviations of three independent experiments ( n.s. indicates non-significant difference, 2way ANOVA test)

Due to the big lattice mismatch (11.9%) and the large difference of reduction potential between Dy (III, − 2.29 V) and Au (III, [AuCl₄]⁻, + 0.93 V), not all initially added Dy was reduced in the AuNP core which resulted in ¹⁶⁶Dy radiolabelling efficiency of around 60%. As the same activity of ¹⁶⁶DyCl₃ was used during the synthesis of ¹⁶⁶DyAu@AuNPs and ¹⁶⁶DyPtAuNPs, the activity of radiolabelled ¹⁶⁶Dy was the same for all samples with different Dy:Au feeding ratios. No improvement of the ¹⁶⁶Dy radiolabelling efficiency was achieved by lowering the initial amount of Dy³⁺.

The completely decayed ¹⁶⁶DyAu@AuNP and ¹⁶⁶DyPtAuNP samples were also destructed and further analysed by ICP-OES to measure the concentration of Au and Dy. Comparing with the Au concentration of the Au@AuNP and PtAuNP samples, little difference of the Au concentration was found from the ¹⁶⁶DyAu@AuNP and ¹⁶⁶DyPtAuNP samples (Table S1). Taking ICP-OES measurements together with other characterizations, we further confirmed that the reduction of gold precursor by NaBH₄ as well as the formation of nanoparticles was not affected by the addition of Dy³⁺. The ¹⁶⁶Dy radiolabelling efficiency was also calculated using the total concentration of Dy (including both radioactive and non-radioactive Dy) measured by ICP-OES (Additional file 1: Fig. S4). Similar to the results from Ge-detector measurement, the radiolabelling efficiency was not influenced by the Dy:Au feeding ratios. However, we found that the ¹⁶⁶Dy radiolabelling efficiency calculated from ICP-OES data was approximately 10% lower than that from the Ge-detector data. Further studies will be carried out to explain this phenomenon.

Retention of ¹⁶⁶Ho and ¹⁶⁶Dy

In vivo generator of therapeutic radionuclides can generally increase the delivered dose per administrated activity because of the longer half-life time of the mother nuclides (Edem et al. 2016). To make sure the radiation dose is mainly delivered to the tumor while sparing the normal tissues, both the mother and the daughter nuclides should be kept within the carrier. Therefore, we radiolabelled core–shell structured gold nanoparticles, i.e. the ¹⁶⁶DyAu@AuNPs with ¹⁶⁶Dy. An outer layer of gold was added to prevent the diffusion of free ¹⁶⁶Ho if it escapes from the core nanoparticle. On the other hand, nanoparticles without the shell structure, i.e. the ¹⁶⁶DyPtAuNPs were also radiolabelled with ¹⁶⁶Dy for comparison. ¹⁶⁶DyAuNP seeds were not studied because of the low colloidal stability (Additional file 1: Fig. S3).

To measure the retention of the internally converted ¹⁶⁶Ho as well as the retention of ¹⁶⁶Dy, ¹⁶⁶DyAu@AuNPs and ¹⁶⁶DyPtAuNPs were incubated in MiliQ water or 2.5 mM DTPA (pH 7.5) at 37 °C for 72 h. Every 24 h, the samples were centrifuged to separate NPs from free ¹⁶⁶Dy³⁺ and ¹⁶⁶Ho³⁺. The counts of the nanoparticles and the washing solution was measured at 65 ~ 90 keV and 340 ~ 460 keV for ¹⁶⁶Ho and ¹⁶⁶Dy respectively. As the NPs were still capped by CTAB/CTAC, the nanoparticles would form aggregation upon interaction with high concentration salt solution or protein (Zhang and Lin 2014). Thus, the in vitro stability tests were not performed in PBS or serum to avoid the interference of nanoparticle aggregation. As shown in Fig. 6, more than 95% of ¹⁶⁶Ho was found to be retained in both ¹⁶⁶DyAu@AuNPs and ¹⁶⁶DyPtAuNPs for at least 72 h in MiliQ water (Fig. 6a, b). The retention of ¹⁶⁶Dy was also found to be more than 95% for both ¹⁶⁶DyAu@AuNPs and ¹⁶⁶DyPtAuNPs during the 72 h incubation in MiliQ water (Fig. 6c, d). For all the samples challenged by DTPA, about 90% of both ¹⁶⁶Ho and ¹⁶⁶Dy was still bounded to the nanoparticles even after 72 h incubation (Fig. 7). These results indicate that very high ¹⁶⁶Ho and ¹⁶⁶Dy retention was achieved independent from the Dy:Au feeding ratio and the extra shell of coating.

Fig. 6.

¹⁶⁶Ho and ¹⁶⁶Dy retention of ¹⁶⁶DyAu@AuNPs (a, c) and ¹⁶⁶DyPtAuNPs (b, d) with different Dy:Au feeding ratios in MiliQ water at 37 °C as function of time. The error bars represent the standard deviation of three independent experiments

Fig. 7.

¹⁶⁶Ho and ¹⁶⁶Dy retention of ¹⁶⁶DyAu@AuNPs (a, c) and ¹⁶⁶DyPtAuNPs (b, d) with different Dy:Au feeding ratios in 2.5 mM DTPA at 37 °C as function of time. The error bars represent the standard deviation of three independent experiments

Discussion

Surprisingly, the ¹⁶⁶DyPtAuNPs were found to be able to retain the same percentage of ¹⁶⁶Ho as the ¹⁶⁶DyAu@AuNPs. This result suggests that high ¹⁶⁶Ho retention could still be achieved even without the addition of an extra gold layer. This finding made us think about a the possible mechanism responsible for the high ¹⁶⁶Ho retention on AuNPs. The internal conversion of ¹⁶⁶Dy results in highly charged ¹⁶⁶Ho ions which tend to seek electrons from the surrounding environment, i.e. the carrier. In the case of ¹⁶⁶Dy coupled to a simple chelator composed of low Z elements such as dodecane tetraacetic acid (DOTA), the number of free electrons in the system is low. Therefore, DOTA molecule could be easily altered to be positively charged after the electron migration to the ¹⁶⁶Ho ions. Due to the repulsion between the entities having the same charge, i.e. ¹⁶⁶Ho³⁺ and [DOTA]ⁿ⁺, the ¹⁶⁶Ho-DOTA complex is ruptured. When a high Z material is used as the carrier for ¹⁶⁶Dy, such as AuNP, many more free electrons are available. When the highly positive ¹⁶⁶Ho ion extracts electrons from its neighbouring Au atoms, electrons can be quickly redistributed to fill in the new vacancies. The redistribution of electrons might cause a transient change of the surface charge of AuNP, but then electrons from the solvent (i.e. water) will be attracted to the AuNP due to the ultra-high affinity of Au to solvated electrons (Ghandi et al. 2015). Therefore, the colloidal stability of AuNP is preserved while the release of ¹⁶⁶Ho is avoided. A similar method was reported to improve the retention of ⁸⁰Br which was internally converted from ^80mBr (49 and 37 keV, α = 1.6 and 300 respectively) by Adamson et al. (Adamson and Grunland 1951; Wexler and Anderson 1960). The authors found that 100% and 86% of ⁸⁰Br was released from [Co(NH₃)₅Br]²⁺ (aq) and solidified [Co(NH₃)₅Br](NO₃)₂ (s) while 47% and even 0% of ⁸⁰Br was released from [PtBr₆]²⁻ (aq) and solidified (NH₄)₂PtBr₆ (s). This result supported our hypothesis on the function of AuNP as electron source for the internally converted ¹⁶⁶Ho. Besides, the results from these studies also suggest that in our case the reduction of ¹⁶⁶Dy³⁺ into solid state (Dy⁰) might also contribute to the high retention of ¹⁶⁶Ho.

Besides the high retention of the internally converted ¹⁶⁶Ho, our radiolabelling method is also simple and quick. The whole procedure can be finished within 8 h without the need of separating ¹⁶⁶Dy from ¹⁶⁶Ho. The interaction between the β⁻ particle emitted by ¹⁶⁶Ho and gold atoms is also favourable for a more efficient dose delivery due to the formation of secondary electrons and free radicals such as · OH radicals (Haume et al. 2016). To make the ¹⁶⁶DyPtAuNPs and ¹⁶⁶DyAu@AuNPs more applicable for clinical application, the current capping ligand, CTAB/CTAC, has to be exchanged with biocompatible ligands such as PEG. In previous studies it has been shown that small AuNPs not conjugated with targeting agents have tumour uptake around 4–5% ID/g depending on the morphology and surface properties of the nanoparticles (Sun et al. 2014; Zhang et al. 2022). In comparison small molecules such as PSMA can achieve much higher tumour uptake (Banerjee et al. 2014). Therefore, it will be very interesting to determine whether the addition of such targeting moieties will increase tumour accumulation.

Conclusion

In summary, we developed a chelator-free radiolabelling method to obtain a ¹⁶⁶Dy/¹⁶⁶Ho in vivo generator and prevented the loss of ¹⁶⁶Ho that is caused by internal conversion. The explanation for the high ¹⁶⁶Ho retention was not experimentally proven but might be related to the high electron density of the gold nanoparticles. To further understand the mechanism of ¹⁶⁶Ho retention on gold nanoparticles, the structure of the nanoparticles should be studied by both experiments as well as theoretical simulations. Besides the further research on ¹⁶⁶Ho retention mechanism, the capping ligands of the nanoparticles should be replaced to increase the biocompatibility of the nanoparticles and make them suitable for medical applications.

Abbreviations

AA

Ascorbic acid

AuNP

Gold nanoparticle

CTAB

Cetyltrimethylammonium bromide

CTAC

Cetyltrimethylammonium chloride

DLS

Dynamic light scattering

DOTA

Dodecane tetraacetic acid

DTPA

Diethylenetriamine pentaacetate

Dy

Dysprosium

EBRT

External beam radiation therapy

EDTA

Ethylenediaminetetraacetic acid

Ho

Holmium

RNT

Radionuclide therapy

SPECT

Single photon emission computed tomography

SPR

Surface plasmon resonance

TEM

Transmission electron microscope

UV–vis

Ultraviolet–visible

Author contributions

AD and HW contributed to the design of the study and oversaw the research project. RW designed and carried out the experiments and analysed the results. BP was involved in the radiochemistry experiments of this work. All authors read and approved the final manuscript.

Funding

Not applicable.

Availability of data and materials

The data associated to this research work are available in this manuscript or in the online supplementary file.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.