Production and characterization of no-carrier-added ¹⁶¹Tb as an alternative to the clinically-applied ¹⁷⁷Lu for radionuclide therapy

Abstract

Background

¹⁶¹Tb is an interesting radionuclide for cancer treatment, showing similar decay characteristics and chemical behavior to clinically-employed ¹⁷⁷Lu. The therapeutic effect of ¹⁶¹Tb, however, may be enhanced due to the co-emission of a larger number of conversion and Auger electrons as compared to ¹⁷⁷Lu. The aim of this study was to produce ¹⁶¹Tb from enriched ¹⁶⁰Gd targets in quantity and quality sufficient for first application in patients.

Methods

No-carrier-added ¹⁶¹Tb was produced by neutron irradiation of enriched ¹⁶⁰Gd targets at nuclear research reactors. The ¹⁶¹Tb purification method was developed with the use of cation exchange (Sykam resin) and extraction chromatography (LN3 resin), respectively. The resultant product (¹⁶¹TbCl₃) was characterized and the ¹⁶¹Tb purity compared with commercial ¹⁷⁷LuCl₃. The purity of the final product (¹⁶¹TbCl₃) was analyzed by means of γ-ray spectrometry (radionuclidic purity) and radio TLC (radiochemical purity). The radiolabeling yield of ¹⁶¹Tb-DOTA was assessed over a two-week period post processing in order to observe the quality change of the obtained ¹⁶¹Tb towards future clinical application. To understand how the possible drug products (peptides radiolabeled with ¹⁶¹Tb) vary with time, stability of the clinically-applied somatostatin analogue DOTATOC, radiolabeled with ¹⁶¹Tb, was investigated over a 24-h period. The radiolytic stability experiments were compared to those performed with ¹⁷⁷Lu-DOTATOC in order to investigate the possible influence of conversion and Auger electrons of ¹⁶¹Tb on peptide disintegration.

Results

Irradiations of enriched ¹⁶⁰Gd targets yielded 6–20 GBq ¹⁶¹Tb. The final product was obtained at an activity concentration of 11–21 MBq/μL with ≥99% radionuclidic and radiochemical purity. The DOTA chelator was radiolabeled with ¹⁶¹Tb or ¹⁷⁷Lu at the molar activity deemed useful for clinical application, even at the two-week time point after end of chemical separation. DOTATOC, radiolabeled with either ¹⁶¹Tb or ¹⁷⁷Lu, was stable over 24 h in the presence of a stabilizer.

Conclusions

In this study, it was shown that ¹⁶¹Tb can be produced in high activities using different irradiation facilities. The developed method for ¹⁶¹Tb separation from the target material yielded ¹⁶¹TbCl₃ in quality suitable for high-specific radiolabeling, relevant for future clinical application.

Electronic supplementary material

The online version of this article (10.1186/s41181-019-0063-6) contains supplementary material, which is available to authorized users.

Background

The use of the β⁻–emitter ¹⁷⁷Lu (E_β-av = 134 keV (100%), T_1/2 = 6.7 d) (Solá 2000), in combination with somatostatin analogues (e.g. DOTATOC, DOTATATE), is considered a promising tool for the treatment of neuroendocrine tumors (NET). It has been extensively utilized in clinics, which recently resulted in the approval of Lutathera® (¹⁷⁷Lu-DOTATATE), by the U.S. Food and Drug Administration (FDA) (Kam 2012; CHMP 2016). Treatment with ¹⁷⁷Lu-DOTATATE resulted in longer progression-free survival time (65.2% at Month 20, compared to 10.8% in the control group); nevertheless, partial remission remained at ≤50% in the assessable patients, with the complete response being ≤12% (CHMP 2016; Strosberg 2017; Kwekkeboom 2005; van Essen 2007; Sansovini 2013). The radiolanthanide ¹⁶¹Tb shows similar decay characteristics (E_β-av = 154 keV (100%), T_1/2 = 6.9 d (Solá 2000)) and coordination chemistry to ¹⁷⁷Lu. ¹⁶¹Tb can, therefore, be stably coordinated with a DOTA chelator and be used in combination with a number of small molecules, peptides and antibodies currently employed with ¹⁷⁷Lu. ¹⁶¹Tb may show an increased therapeutic efficacy over ¹⁷⁷Lu, due to the co-emission of a substantially larger number of conversion and Auger electrons at a favorable energy range (~12 e⁻, ~36 keV per decay for ¹⁶¹Tb and ~ 1 e⁻, ~ 1.0 keV per decay for ¹⁷⁷Lu, respectively) (Eckerman 2008; Müller 2014). The possibility of using ¹⁶¹Tb as an alternative to ¹⁷⁷Lu was first proposed by Lehenberger et al. (Lehenberger 2011) and, subsequently, corroborated by Müller et al. by comparison of in vitro and in vivo studies using a DOTA-folate conjugate labeled with ¹⁶¹Tb and ¹⁷⁷Lu (Müller 2014). The enhanced anti-tumor effect, as well as higher average survival time, was found in mice treated with ¹⁶¹Tb-folate over those which received ¹⁷⁷Lu-folate. In a preliminary therapy study using ¹⁶¹Tb-PSMA-617, PSMA-positive PC-3 PIP tumor-bearing mice demonstrated significant tumor-growth delay, as compared to the control group, without causing early side effects (Müller 2018a). Better therapeutic efficacy was also observed for a ¹⁶¹Tb-labeled radioimmunoconjugate in an ovarian cancer model when compared to the ¹⁷⁷Lu-radioimmunoconjugate counterpart (Grünberg 2014). The low-energy conversion and Auger electron emission from ¹⁶¹Tb contribute 26% – 88% to the total absorbed dose (compared to 10% – 34% for ¹⁷⁷Lu), depending on the tumor size, which could be associated to its enhanced therapeutic efficacy over that of ¹⁷⁷Lu (Champion 2016). The doses delivered by ¹⁶¹Tb or ¹⁷⁷Lu to 10 mm-diameter spheres were calculated to be comparable for both radionuclides, however, for 100 μm-diameter and 10 μm-diameter spheres ¹⁶¹Tb could deliver 1.8 and 3.6 times higher dose than ¹⁷⁷Lu, respectively, making ¹⁶¹Tb the more appropriate candidate for treating micrometastases (Hindie 2016). Also, the co-emission of 48.9 keV and 74.6 keV ¹⁶¹Tb γ-rays allows for the acquisition of single photon emission computed tomography (SPECT) images for dosimetry determination before administration of the therapeutic dose, comparable to that performed with ¹⁷⁷Lu (Marin 2018). In addition, ¹⁶¹Tb could be used in combination with diagnostic radioisotopes, namely, ¹⁵²Tb (PET) or ¹⁵⁵Tb (SPECT) as a matched pair towards the concept of theragnostics (Müller 2012; Müller 2018b).

The ¹⁶¹Tb production route was proposed by Lehenberger et al. via the ¹⁶⁰Gd(n,γ)¹⁶¹Gd → ¹⁶¹Tb nuclear reaction, which provided no-carrier-added radiolanthanide at high specific activities (~ 4 TBq/mg) (Lehenberger 2011). Enriched ¹⁶⁰Gd(NO₃)₃ targets (ampoules) were prepared by dissolving ¹⁶⁰Gd₂O₃ in nitric acid and evaporating to dryness. Lanthanide nitrates are hygroscopic materials, however, and the heating of the ampoule in the nuclear reactor (due to γ-rays from the reactor and β -rays generated in the sample) can create water vapor within the ampoule. The vapor could create overpressure, resulting in ampoule breakage. The ¹⁶¹Tb separation method from ¹⁶⁰Gd(NO₃)₃ targets was previously developed at Paul Scherrer Institute (Villigen-PSI, Switzerland), but the radiolabeling capability of the ¹⁶¹Tb product was three times lower than the commercial no-carrier-added ¹⁷⁷Lu (Lehenberger 2011). This implied that the ¹⁶¹Tb product contained undesired environmental impurities at the end of separation (EOS), thereby, compromising the capability of reproducible routine production.

Herein, we report on the large-scale ¹⁶¹Tb production from ¹⁶⁰Gd₂O₃ target material, suitable for introduction into a process in accordance with Good Manufacturing Practice (GMP) and, thereafter, clinical application. The ¹⁶¹Tb purification method was improved by optimization of the Tb/Gd separation process, followed by characterization of the final product (¹⁶¹TbCl₃). The ¹⁶¹Tb purity was compared with no-carrier-added ¹⁷⁷Lu (EndolucinBeta), as currently produced by ITG GmbH, Germany, for worldwide clinical application.

Methods

Target preparation for the production of ¹⁶¹Tb

Gadolinium oxide (¹⁶⁰Gd₂O₃, 98.2% enrichment, Isoflex, USA) was used as target material for the production of no-carrier-added ¹⁶¹Tb, as previously reported (Lehenberger 2011). The elemental composition of the target material in question is provided in the Supplementary Material (Table S1). To prepare the targets for irradiation at the spallation-induced neutron source (SINQ, Paul Scherrer Institute, 4.10¹³ n.cm^− 2.s^− 1), 80–95 mg ¹⁶⁰Gd(NO₃)₃ were placed in a quartz glass ampoule (Suprasil, Heraeus, Germany) and sealed. Ampoules containing 7–33 mg ¹⁶⁰Gd₂O₃ were prepared in a similar manner and sent for irradiation to two research nuclear reactors (SAFARI-1, South African Nuclear Energy Corporation, 2.10¹⁴ n.cm^− 2.s^− 1; and RHF ILL, Institut Laue–Langevin, 1.10¹⁵ n.cm^− 2.s^− 1). The mass of the target material, required for the irradiation at the chosen facility, was calculated using the ChainSolver code (Romanov 2005).

Determination of the neutron fluxes of the irradiation facilities with ⁵⁹Co monitors

In order to monitor neutron fluxes at ILL, SAFARI-1 and SINQ, quartz ampoules containing ⁵⁹Co as standard (⁵⁹Co in 2% w/w HNO₃, Sigma-Aldrich, USA) with 33 ng – 2 μg ⁵⁹Co (mass determined based on the volume of the standard solution pipetted) were prepared. Ampoules were dried at 80 °C, to ensure water evaporation, and sealed. Ampoules with ⁵⁹Co standard were placed and sealed in the same Al container as the ampoules containing ¹⁶⁰Gd target material, along with empty ampoules (used as references) for the irradiation process. ⁵⁹Co masses were calculated to produce 50–100 kBq ⁶⁰Co activity via ⁵⁹Co(n,γ)⁶⁰Co nuclear reaction, depending on the reactor neutron flux. The ⁶⁰Co activities in the ampoules were measured after irradiations using a high-purity germanium (HPGe) detector (Canberra, France), in combination with the InterWinner software package (version 7.1, Itech Instruments, France), with a statistical uncertainty less than 5%. After irradiation at the facility in question and gamma spectrometry of the ampoules, the activities of ⁶⁰Co produced by the added ⁵⁹Co standards were determined by subtraction of the ⁶⁰Co activities of the reference (empty) ampoules which stem from traces of cobalt impurities in the used quartz. Based on these ⁶⁰Co activity values, the average neutron flux (ϕ_th) of each irradiation was calculated (Table 1) using the following equation:

$${\varphi }_{\mathit{th}}=\frac{A_0}{{\sigma }_n\bullet N\bullet \left(1-e^{-\lambda \bullet t_B}\right)}$$ 1

where A₀ is the ⁶⁰Co activity at the end of bombardment, σ_n the thermal neutron capture cross section (37.18 ± 0.06 b (Mughabghab 2018)), N the number of target atoms, λ the radioactive decay constant and t_B the irradiation time.

Table 1.

Measured neutron fluxes of the irradiation facilities used for ¹⁶¹Tb production

Facility	Irradiation time, d	60Co activity, kBq	Measured perturbed neutron flux, n.cm− 2.s− 1	Nominal unperturbed neutron flux, n.cm− 2.s− 1
ILL	10	33.5	7.4·1014	1.0·1015
SAFARI-1	13–16	41.0–54.0	1.8·1014	2.0·1014
SINQ	21	92.3	1.8·1013 a	4.0·1013

^aMeasurement performed while spallation target was operated at lower than optimum beam current

Development of the procedure for the ¹⁶¹Tb purification process

A chromatographic column (10 mm × 170 mm) was prepared using Sykam macroporous cation exchange resin (Sykam Chromatographie Vertriebs GmbH, Germany; particle size 12–22 μm, NH₄⁺ form). The separation parameters were first optimized by means of bench experiments with the use of radioactive tracers (Additional file 1: Table S2) and subsequently applied towards the separation of an aliquot of reactor-produced ¹⁶¹Tb (230 MBq). These experiments resulted in the design and construction of a chemical separation module, such that high activities (GBq) of the radionuclide can be processed in the hot cell. The quartz glass ampoule with the ¹⁶⁰Gd₂O₃ target material, delivered from the irradiation facility, was placed in a plastic target tube, crushed and attached to the module inside the hot cell with the aid of manipulators. The target material from the ampoule was dissolved in 2.0 mL 7.0 M nitric acid (HNO₃, Suprapur, Merck, Germany), followed by evaporation at 80 °C under nitrogen flow. The residue was taken up in 0.1 M ammonium nitrate (prepared from 25% Suprapur NH₃ and 65% Suprapur HNO₃, Merck, Germany) and loaded onto the cation exchange resin column. The ¹⁶¹Tb separation from the target material and impurities was performed with the use of 0.13 M (pH 4.5) α-hydroxy-isobutyric acid (α-HIBA, Sigma-Aldrich GmbH, Germany) as eluent. Concentration of ¹⁶¹Tb was performed using the bis(2,4,4-trimethyl-1-pentyl) phosphinic acid extraction resin (LN3, Triskem International, France; 6 mm × 5 mm), followed by the elution of the final product (¹⁶¹TbCl₃) in 500 μL 0.05 M hydrochloric acid (HCl, Suprapur, Merck, Germany). The pH of the final product was determined using pH indicator strips (Merck, Germany).

Characterization of the ¹⁶¹Tb product after purification

Radionuclidic purity

The identification and radionuclidic purity of the ¹⁶¹Tb were examined by γ-ray spectrometry using the HPGe detector mentioned above. The aliquot of the final product, containing 5–10 MBq of ¹⁶¹Tb, was measured with the HPGe detector until the 3σ uncertainty was below 5%.

Radiochemical purity

The radiochemical purity of the final product was determined by means of radio thin layer chromatography (radio TLC) using a procedure established for ¹⁷⁷Lu (Oliveira 2011). The aliquot of ¹⁶¹TbCl₃ (2 μL, ~ 100 kBq) was deposited on TLC silica gel 60 F₂₅₄ plates (Merck, Italy) and placed in the chamber with 0.1 M sodium citrate solution (pH 5.5; Merck, Germany) mobile phase. After elution, the plate was dried and analyzed using a radio TLC scanner instrument (Raytest Isotopenmessgeräte GmbH, Germany). The results were interpreted with the MiniGita Control software package (version 1.14, Raytest Isotopenmessgeräte GmbH, Germany).

Radiolabeling yield

Radiolabeling of DOTANOC (ABX GmbH, Germany) at a molar activity of 180 MBq/nmol (1-to-4 nuclide-to-peptide molar ratio) was performed in order to evaluate the success of the purification process. Sodium acetate (Alfa Aesar, Germany; 0.5 M, pH 8) was added to ¹⁶¹TbCl₃ solution (~ 200 MBq) to adjust pH to ~ 4.5. The relevant quantity of DOTANOC was subsequently added from a 1 mM stock solution. The reaction solution was incubated for 10 min at 95 °C. The radiolabeling yield was determined by reverse-phase high performance liquid chromatography (HPLC, Merck Hitachi LaChrom) with a radioactivity detector (LB 506, Berthold, Germany) and a C-18 reverse-phase column (150 mm × 4.6 mm; Xterra™ MS, C18; Waters). Trifluoroacetic acid 0.1% (Sigma-Aldrich, USA) in MilliQ water (A) and acetonitrile (VWR Chemicals, USA; HPLC grade) (B) were used as mobile phase with a linear gradient of solvent A (95–5% over 15 min) in solvent B at a flow rate of 1 mL/min. The sample for the analysis was prepared by diluting ~ 0.3 MBq aliquot of the radiolabeling solution in 100 μL MilliQ water containing sodium diethylenetriamine pentaacetic acid (Na-DTPA, 50 μM). The radiolabeling yield of ¹⁶¹Tb-DOTANOC was determined by integration of the product peak from the obtained HPLC chromatogram in relation to the sum of all radioactive peaks (the radiolabeled product, potentially released ¹⁶¹Tb as well as degradation products of unknown structure), which were set to 100%.

Determination of the radiolabeling yield of ¹⁶¹Tb- and ¹⁷⁷Lu-DOTA over a two-week period

In order to assess the change of the molar activity of ¹⁶¹Tb-DOTA at different DOTA-to-nuclide molar ratios over time (2 weeks after EOS) and to compare it with the molar activity of ¹⁷⁷Lu-DOTA, thin layer chromatography (TLC) analysis was performed. The required DOTA solutions (1–500 pmol DOTA) were obtained by dilution of the initial 1 mM DOTA solution (CheMatech, France) with 0.5 M sodium acetate (pH 4.5). The prepared DOTA dilutions were mixed with 2.5–4 MBq ¹⁶¹Tb (corresponding to 3.1–5 pmol), ¹⁷⁷Lu (no-carrier-added, ITG GmbH, Germany) or ¹⁷⁷Lu (carrier-added, IDB Holland bv, the Netherlands) at different DOTA-to-nuclide molar ratios (160:1 to 1:1). The reaction solutions were incubated for 20 min at 95 °C and 2 μL of each solution were deposited on TLC silica gel 60 F₂₅₄ plates, which served as a stationary phase. The mixture of 10% ammonium acetate (Sigma-Aldrich, USA) and methanol (Merck, Germany) was used as mobile phase (ratio 1:1 (v/v), pH 5.5). After elution, the phosphor screen (Multisensitive, Perkin Elmer Inc., USA) was illuminated with the TLC plate and analyzed with a Cyclon Phosphor Imager (Perkin Elmer Inc., USA). The peaks corresponding to the unlabeled ¹⁶¹Tb or ¹⁷⁷Lu (R_f = 0) and to the ¹⁶¹Tb- or ¹⁷⁷Lu-DOTA compounds (R_f = 0.4) were integrated with the OptiQuant image analysis software (version 5.0, Perkin Elmer Inc., USA) and the radiolabeling yield determined. Based on the data obtained, the DOTA-to-nuclide molar ratios were plotted against the radiolabeling yield using Origin software, fitted with a Boltzmann’s sigmoidal modified equation. Experiments were repeated three times for ¹⁶¹Tb and ¹⁷⁷Lu (no-carrier-added) and once for ¹⁷⁷Lu (carrier-added). The average DOTA-to-nuclide molar ratios, corresponding to 50% labeling efficiency of DOTA with ¹⁶¹Tb (no-carrier-added) and ¹⁷⁷Lu (no-carrier-added or carrier-added) at different time points (Day 3 to Day 14 after EOS), were determined and compared with each other for statistical significance by an unpaired t test using Graph Pad Prism (version 7.00).

¹⁶¹Tb/¹⁷⁷Lu-DOTATOC stability studies

The radiolabeling of DOTATOC with no-carrier-added ¹⁶¹Tb or no-carrier-added ¹⁷⁷Lu at 50 MBq/nmol molar activity (300 MBq ¹⁶¹Tb activity in total) was performed as described above, in the absence or in the presence of L-ascorbic acid (2.9 mg, Sigma-Aldrich, USA). The radiolabeling yield was determined by means of HPLC (as described above) immediately after the preparation of ¹⁶¹Tb/¹⁷⁷Lu-DOTATOC. The radioactivity concentration of the labeling solutions was adjusted to 250 MBq/500 μL with saline and radiolytic stability of the radioligand was determined over time (1 h, 4 h and 24 h) by means of HPLC.

Results

¹⁶¹Tb production yield (theoretical versus experimental)

No-carrier-added ¹⁶¹Tb was produced by neutron irradiation of enriched ¹⁶⁰Gd₂O₃ (98.2% enrichment) targets via the ¹⁶⁰Gd(n,γ)¹⁶¹Gd → ¹⁶¹Tb nuclear reaction. The mass of the target material had to be calculated precisely in order to ensure that the ¹⁶¹Tb activity allowed for international transportation was not exceeded. ¹⁶¹Tb is not explicitly listed in the dangerous goods tables of the ADR (European Agreement concerning the International Carriage of Dangerous Goods by Road) and International Air Transport Agency (IATA) regulations which are, in turn, based on IAEA (International Atomic Energy Agency) recommendations. As a result, the generic A2 value (activity limit of radioactive material) according to the “Basic Radionuclide Values for Unknown Radionuclides or Mixtures” of 0.02 TBq has to be applied (IAEA 2012). Due to this restriction, a maximum of 20 GBq ¹⁶¹Tb could be transported internationally (in this case, shipping from ILL and SAFARI-1 research reactors to PSI) (IAEA 2012). Masses of the ¹⁶⁰Gd target material, required to produce 20 GBq ¹⁶¹Tb after bombardment at the irradiation facilities, were calculated with the ChainSolver code using the recommended cross-section for thermal neutron capture on ¹⁶⁰Gd of 1.4(3) b (Mughabghab 2018). Two-week irradiations at ILL (6.5 mg ¹⁶⁰Gd₂O₃, 1.10¹⁵ n.cm^− 2.s^− 1, 1 day cooling) and at SAFARI-1 (32 mg ¹⁶⁰Gd₂O₃, 2.10¹⁴ n.cm^− 2.s^− 1, 1 day cooling) would theoretically result in 20 GBq ¹⁶¹Tb. At PSI’s neutron source facility (SINQ, 4·10¹³ n.cm^− 2.s^− 1) each irradiation cycle is 3 weeks, which was calculated to provide 17.2 GBq ¹⁶¹Tb after the bombardment of 100 mg ¹⁶⁰Gd(NO₃)₃. The masses of the target material could be adapted to operator/user requirements based on the ChainSolver code calculations and neutron fluxes, calculated from the measured ⁶⁰Co activity values of the ⁵⁹Co monitors (Table 1). Three ampoules with ⁵⁹Co were bombarded at the SAFARI-1 nuclear reactor and one ampoule each at ILL and at SINQ irradiation facilities (together with the ¹⁶⁰Gd ampoules), respectively. The measured values of the perturbed neutron fluxes in the samples irradiated at the ILL and SAFARI-1 nuclear reactors scaled as expected with the unperturbed neutron flux values reported by the facility in question.

In practice, one-to-two-week irradiations of ¹⁶⁰Gd target ampoules using SAFARI-1 (22–33 mg ¹⁶⁰Gd₂O₃) and ILL (7–13 mg ¹⁶⁰Gd₂O₃) research reactors resulted in production of 10–20 GBq of ¹⁶¹Tb (Table 2). The calculated neutron flux at the PNA irradiation position of the SINQ facility was determined experimentally to be only ~ 50% of its originally reported value of 4 10¹³ n.cm^− 2.s^− 1 (Table 1), which resulted in 6–9 GBq ¹⁶¹Tb after 3 weeks irradiation of the enriched ¹⁶⁰Gd(NO₃)₃ target material (Table 2). This was due to the fact that the spallation target had recently been replaced and was being operated at a lower beam current (1.3 mA protons instead of 2.3 mA) than normally specified (Table 1).

Table 2.

Production yields of several ¹⁶¹Tb batches, obtained from the irradiation facilities

Facility	Irradiation time, d	Target material	Mass of the target material, mg	Measured 161Tb activity (EOB), GBq
ILL	5	160Gd2O3	12.5	11.6
ILL	10	160Gd2O3	7.3	16.7
SAFARI-1	14	160Gd2O3	33.3	19.6
SAFARI-1	7	160Gd2O3	32.5	11.9
SINQ	21	160Gd(NO3)3	94.9	8.8
SINQ	21	160Gd(NO3)3	86.4	6.0

Radiochemical separation of ¹⁶¹Tb from the target material and accumulated impurities

By performing bench experiments using Sykam cation exchange resin column (10 mm × 170 mm) and long-lived radioactive tracers, conditions for the efficient separation of Tb from up to 140 mg of Gd₂O₃ and the presence of various impurities were established (Additional file 1: Figure S1 and S2). Subsequently, the established experimental conditions (Sykam resin; 0.13 M α-HIBA, pH 4.5; 0.6 mL/min eluent flow rate) were applied towards the purification of the reactor-produced ¹⁶¹Tb (230 MBq). During the irradiation, radioactive side products were co-produced from the impurities of the target material (⁴⁶Sc, ¹²⁴Sb, ¹⁴¹Ce, ¹⁴⁷Nd, ¹⁵³Gd, ¹⁵³Sm, ^{152/154/155/156}Eu, ¹⁶⁹Yb) and the ampoule material (⁶⁵Zn, ⁶⁰Co, ¹⁹²Ir). Despite this impurity formation, the established method demonstrated effective ¹⁶¹Tb separation from Gd target material and the impurities using the Sykam resin column (Fig. 1). ¹⁶¹Tb was eluted from the Sykam resin with 20 mL α-HIBA, followed by the concentration of the radionuclide on the LN3 resin column (Additional file 1: Figure S3). LN3 extraction resin was reported to have low affinity for Tb ions in low concentrated acids (McAlister 2007), thereby, allowing ¹⁶¹Tb to be eluted in a small volume of 0.05 M HCl.

Fig. 1.

Elution profile of ¹⁶¹Tb separation from the irradiated target material and side products (10 mm × 170 mm Sykam resin column, 8 mg ¹⁶⁰Gd₂O₃, 0.6 mL/min eluent flow rate)

Based on the developed two-column purification method (combination of Sykam and LN3 resin columns), a ¹⁶¹Tb purification module was designed (Fig. 2). The module was constructed and placed inside the hot cell, making it possible to perform separations with higher activities (up to 20 GBq) of the reactor-produced ¹⁶¹Tb.

Fig. 2.

Schematic diagram of the ¹⁶¹Tb chemical separation system

The established procedure for the ¹⁶¹Tb purification process using the designed module resulted in the elution of the final product (¹⁶¹TbCl₃) in a small volume (500 μL) of 0.05 M HCl, with an activity concentration of 11–21 MBq/μL. A separation yield of 80–90% was achieved at EOS. Losses of 10–20% of ¹⁶¹Tb activity were observed in the target dissolving, column loading and final elution steps.

Characteristics of the ¹⁶¹TbCl₃ product

¹⁶¹Tb, obtained after the purification process, was characterized to provide a product specification (Table 3). The identification of the product was confirmed by the ¹⁶¹Tb-characteristic γ-lines (Fig. 3). The content of long-lived ¹⁶⁰Tb (T_1/2 = 72.3 d), produced by the ¹⁵⁹Tb(n,γ)¹⁶⁰Tb nuclear reaction due to the presence of ¹⁵⁹Tb impurity in the target material (as sold by the vendor), was determined after the decay of ¹⁶¹Tb and did not exceed 0.007% of the total ¹⁶¹Tb activity at EOS (Additional file 1: Figure S4). The radiochemical purity of the ¹⁶¹TbCl₃, determined using radio-TLC, was > 99% (Additional file 1: Figure S5). The radiolabeling yield of ¹⁶¹Tb-DOTANOC showed ≥99% efficiency at 180 MBq/nmol molar activity, which corresponds to 1-to-4 nuclide-to-peptide molar ratio (Fig. 4).

Table 3.

Product data specification of ¹⁶¹TbCl₃, developed in this work, compared to the commercially-available ¹⁷⁷LuCl₃

Test	Specification of 161TbCl3	Specification of 177LuCl3 (EndolucinBeta)a
Radioactivity concentration	11–21 MBq/μL	36–44 MBq/μL
Appearance	Clear and colorless solution	Clear and colorless solution
pH	1–2	1–2
Radiolabeling yield HPLC based on radiolabeling with161Tb of DOTANOC, molar ratio 1:4 (180 MBq/nmol)	> 99%	> 99%
Identity 161Tb (γ-ray spectrometry)	48.9 keV γ-line 74.6 keV γ-line	113 keV γ-line 208 keV γ-line
Radionuclidic purity (γ-ray spectrometry)	160Tb ≤ 0.007%	175Yb ≤ 0.01%
Radiochemical purity (radio-TLC)	> 99%	> 99%

^aSpecification from the EndolucinBeta certificate of analysis (ITG)

Fig. 3.

Gamma spectrum of ¹⁶¹Tb, obtained after the purification process. The radionuclidic impurity ¹⁶⁰Tb is not visible due to the negligible activity as compared to that of ¹⁶¹Tb at end of separation (EOS)

Fig. 4.

HPLC chromatogram of ¹⁶¹Tb-DOTANOC (2.3 min retention time would indicate “free” or unlabeled ¹⁶¹Tb, while 8.2 min indicates ¹⁶¹Tb-DOTANOC)

Comparison of the ¹⁶¹Tb and ¹⁷⁷Lu quality, based on the ¹⁶¹Tb- and ¹⁷⁷Lu-DOTA molar activities over a two-week period

Radiolabeling of DOTA with ¹⁶¹Tb (no-carrier-added) and ¹⁷⁷Lu (either carrier-added or no-carrier-added) was performed at different DOTA-to-nuclide molar ratios in order to monitor the quality change of the radiolanthanides of interest over a two-week period after EOS. DOTA could be complexed with ¹⁶¹Tb and ¹⁷⁷Lu (no-carrier-added) at 15:1 and 13:1 DOTA-to-nuclide molar ratios, respectively, with > 90% radiolabeling yield at Day 14 after EOS (Fig. 5 a, b). This indicates the possibility of using the prospective drug product (e.g. DOTA peptides radiolabeled with ¹⁶¹Tb) for up to 2 weeks after the chemical separation. With carrier-added ¹⁷⁷Lu, 90% radiolabeling yield was only achieved when using a much higher DOTA-to-nuclide radio (32:1) at the two-week time point (Fig. 5c).

Fig. 5.

Comparison of the radiolabeling yield of (a) no-carrier-added ¹⁶¹Tb (SAFARI-1); b no-carrier-added ¹⁷⁷Lu (ITG) and (c) carrier-added ¹⁷⁷Lu (IDB) in combination with DOTA over time at different DOTA-to-nuclide molar ratios

The average DOTA-to-nuclide molar ratios corresponding to 50% labeling efficiency of DOTA with ¹⁶¹Tb and ¹⁷⁷Lu were determined at specific time points (Day 3, Day 7, Day 10 and Day 14 – Table 4) (Additional file 1: Figure S7). The values allow an estimation of the possible radiolabeling yield of different biomolecules conjugated with a DOTA chelator, labeled with the radionuclide of interest, over a certain decay period, as well as comparison of the radiolabeling capability of the radionuclides of interest. When DOTA was radiolabeled with carrier-added ¹⁷⁷Lu, 50% labeling efficiency was obtained at higher DOTA-to-nuclide molar ratios at each time point as compared to no-carrier-added ¹⁶¹Tb and ¹⁷⁷Lu, respectively. The 50% labeling efficiency of ¹⁶¹Tb-DOTA was found to be comparable with that of no-carrier-added ¹⁷⁷Lu at Day 3 (p = 0.13), while a slight increase of the values was observed for Day 7, Day 10 and Day 14 (p > 0.05), respectively.

Table 4.

DOTA-to-nuclide molar ratios, corresponding to 50% labeling efficiency of ¹⁶¹Tb-DOTA and both carrier-added and no-carrier-added ¹⁷⁷Lu-DOTA over a two-week decay period

Nuclide	Day 3	Day 7	Day 10	Day 14
161Tb (PSI) No-carrier-added	4.4 ± 0.3	5.7 ± 0.4	7.6 ± 0.5	9.4 ± 0.3
177Lu (ITG) No-carrier-added	3.8 ± 0.1	4.2 ± 0.1	5.7 ± 0.1	7.1 ± 0.1
177Lu (IDB) Carrier-added	5.8*	7.5*	9.2*	12.2*

* Statistically different to no-carrier-added ¹⁶¹Tb and ¹⁷⁷Lu (p < 0.05)

Radiolytic stability of ¹⁶¹Tb-DOTATOC in comparison to ¹⁷⁷Lu-DOTATOC

In order to determine whether the conversion and Auger electrons from ¹⁶¹Tb may cause additional radiolytic degradation of radiolabeled somatostatin analogues, stability tests were performed using DOTATOC. The preparation of the any clinically-applied radiopharmaceutical (e.g. ¹⁷⁷Lu-DOTATOC, ¹⁷⁷Lu-DOTATATE), requires the use of a stabilizer (e.g. L-ascorbic acid, gentisic acid) in order to inhibit peptide autoradiolysis (LUTATHERA n.d.; Mukherjee et al. 2015; Liu et al. 2003; Dash 2015; Esser 2006; Schuchardt 2013). The stability studies of ¹⁶¹Tb-DOTATOC and ¹⁷⁷Lu-DOTATOC were performed with and without addition of L-ascorbic acid. Both ¹⁶¹Tb- and ¹⁷⁷Lu-DOTATOC were stable over 24 h incubation in the presence of a stabilizer (L-ascorbic acid) and showed > 98% radiolabeling yield at the 24-h time point. ¹⁶¹Tb-DOTATOC and ¹⁷⁷Lu-DOTATOC were stable over 1 h (> 99% intact product) in the absence of L-ascorbic acid, but both showed radiolytic degradation after an incubation period of 24 h. After 4 h, slight degradation of both ¹⁶¹Tb-DOTATOC (78% intact product) and ¹⁷⁷Lu-DOTATOC (85% intact product) was observed (Fig. 6). No significant influence of conversion and Auger electrons from ¹⁶¹Tb on radioligand stability was observed, when compared to ¹⁷⁷Lu.

Fig. 6.

Stability of the radiopeptide over time, in the presence and absence of ascorbic acid: the graph shows the % intact ¹⁶¹Tb/¹⁷⁷Lu-DOTATOC over a period of 24 h. In the presence of ascorbic acid ¹⁷⁷Lu and ¹⁶¹Tb-DOTATOC curves overlap, making it impossible to visualize both simultaneously

Discussion

In the present study, the development of a reproducible chemical separation to produce no-carrier-added ¹⁶¹Tb from enriched ¹⁶⁰Gd targets and the characterization of the final product (¹⁶¹TbCl₃) is reported. Gd (NO₃)₃, previously used as target material (Lehenberger 2011), is not suitable for large-scale ¹⁶¹Tb production as lanthanide nitrates are hygroscopic materials, which begin to decompose at 70 °C (Fukuda 2018; Kalekar 2017). The pressure change due to the water evaporation and release of the gases at higher temperatures inside the ampoule may result in the ampoule cracking. The use of oxide targets (melting point 2420 °C) eliminates the potential issues described and will potentially give rise to large-scale ¹⁶¹Tb production in future (TBq activity). The ChainSolver code allows for quick calculations of the required masses of ¹⁶⁰Gd targets for the production of the desired ¹⁶¹Tb activity, based on irradiation times and neutron fluxes, obtained with ⁵⁹Co monitors.

Another big advantage of the newly-devised method over that of the previously-developed one is that the ¹⁶¹Tb product is eluted from LN3 in a small volume and can be used directly for labeling, while the previous method required the use of evaporation after elution from AG50W-X8, thereby, increasing the risk of introduction of environmental contamination and reducing the quality of the final product. The use of LN3 also ensures elution of elements in reverse order when compared to the Sykam/α-HIBA system, therefore, any impurities that may have eluted with ¹⁶¹Tb before introduction to the second column will be eluted differently from the LN3 resin column.

Recently, Brezovcsik et al. reported a Tb separation procedure from massive Gd targets (> 100 mg) using a 20 cm long analytical HPLC column (GE), indicating no mass influence of Gd on the separation process (Brezovcsik 2018). The efficiency of Tb separation from Gd was only 85%, however, showing significant overlapping of Tb and Gd peaks, which would result in the presence of “cold” Gd in the Tb final product and significantly decrease radiolabeling efficiency. The purification method described in this work provides an effective ¹⁶¹Tb separation from the Gd₂O₃ target material of masses up to 140 mg. This is a valuable result for possible future commercial application of the developed method for ¹⁶¹Tb separation from massive Gd₂O₃ targets (> 100 mg). For example, 2 weeks irradiation of 160 mg ¹⁶⁰Gd₂O₃ target in ILL's nuclear reactor could theoretically result in the production of 0.5 TBq ¹⁶¹Tb (ChainSolver Code calculations (Romanov 2005)) which can be efficiently purified with the investigated method. ¹⁶⁰Gd₂O₃ target material contains various trace elements (Additional file 1: Table S2), which may show similar chemical behavior to ¹⁶¹Tb on the resin in question and result in the elution of Tb and the impurities in the same fraction. It was demonstrated that, despite the impurities that could be produced via activation of trace elements in the target material or in the quartz ampoule, the established purification method effectively separated the ¹⁶¹Tb from potential impurities present in the system, based on the combination of Sykam and LN3 resin columns.

The two-column ¹⁶¹Tb purification process proposed by Lehenberger et al. (Lehenberger 2011) was adopted as the baseline for this study towards further development. The cation exchange resin of the first column, column dimension and pump flow rate were changed, while the concentration and pH of the eluent remained the same (0.13 M α-HIBA pH 4.5). The resin used for the second column was changed from AG 50 W-X8 (cation exchange resin) to LN3 (extraction resin). These modifications mentioned above played a vital role in obtaining the final product (¹⁶¹TbCl₃) in purity comparable to that of the commercially-available no-carrier-added ¹⁷⁷LuCl₃ (EndolucinBeta). The radiolabeling capability of ¹⁶¹Tb in this work was similar to no-carrier-added ¹⁷⁷Lu and three times higher than that of ¹⁶¹Tb obtained by Lehenberger et al. (Lehenberger 2011). Somatostatin analogues could be labeled with no-carrier-added ¹⁶¹Tb (this work) at 1-to-4 nuclide-to-peptide molar ratio (Fig. 4) immediately after the purification process with > 99% radiolabeling yield. Quantitative formation of ¹⁶¹Tb-DOTATATE was possible only at 1-to-12 nuclide-to-peptide molar ratio (Lehenberger 2011).

As expected, the radiolabeling capability of the produced, no-carrier-added ¹⁶¹Tb was higher than the radiolabeling capability of carrier-added ¹⁷⁷Lu at each time point over the two-week decay period after the chemical separation process. The faster drop of ¹⁶¹Tb radiolabeling capability as compared to no-carrier-added ¹⁷⁷Lu at Day 7, Day 10 and Day 14 after purification (Table 4) could be explained by the lower radioactivity concentration of the final product obtained (11–21 MBq/μL for ¹⁶¹TbCl₃ vs 36–44 MBq/μL for ¹⁷⁷LuCl₃). This implies that the mass ratio between ¹⁶¹Tb and the impurities (Zn, Fe, Co, Pb etc.), which can be introduced during product analysis and post-processing, is lower compared to ¹⁷⁷Lu. This results in the potentially stronger interference from environmental impurities during radiolabeling of DOTA with ¹⁶¹Tb than with ¹⁷⁷Lu (Asti 2012). Nevertheless, complexation of ¹⁶¹Tb with DOTA was possible at Day 14 (after radiochemical separation) at 1-to-15 nuclide-to-peptide molar ratio (corresponding to 48 MBq/nmol molar activity), with > 90% radiolabeling yield (Fig. 5). This ratio would be appropriate when using DOTA-functionalized targeting agents, such as peptides for peptide receptor radionuclide therapy (Dash 2015). These excellent achievements indicate the possible clinical use of ¹⁶¹TbCl₃ for a period of up to 2 weeks after EOS. Moreover, clinically-applied DOTATOC radiolabeled with ¹⁶¹Tb was stable over 24 h at high radioactivity concentration, indicating that conversion and Auger electrons had no negative influence on the stability, hence, storage and transportation of ¹⁶¹Tb-labeled somatostatin analogues would be feasible, as is the case for their ¹⁷⁷Lu-labeled counterparts.

Conclusions

A new method to separate ¹⁶¹Tb from its enriched ¹⁶⁰Gd₂O₃ target material and co-produced impurities was developed with the use of cation exchange and extraction chromatography, respectively. The method resulted in radionuclidically and radiochemically pure product (¹⁶¹TbCl₃), comparable to commercially available, no-carrier-added ¹⁷⁷Lu. The quantity and quality of ¹⁶¹Tb is suitable for high-specific radiolabeling, potentially useful for the GMP production of radioligands towards future clinical application.

Abbreviations

EOI

End of Irradiation

EOS

End of Separation

GMP

Good Manufacturing Practice

HPGe

High-purity Germanium

HPLC

High Performance Liquid Chromatography

IATA

International Air Transport Agency

PET

Positron Emission Tomography

SPECT

Single Photon Emission Computed Tomography

TLC

Thin Layer Chromatography

α-HIBA

α-hydroxy-isobutyric acid

Authors’ contributions

NG developed the production and separation process, performed the separation experiments of ¹⁶¹Tb, did the radiolabeling and the stability experiments, analyzed the data and drafted the manuscript. CM was responsible for the quality assessment of the final product and reviewed/revised the manuscript. ZT and SH supported the chemical separation process and reviewed the manuscript. UK and JRZ were responsible for the irradiation of ¹⁶⁰Gd targets at ILL and Necsa, respectively. AV was responsible for the irradiation of ¹⁶⁰Gd targets at PSI, as well as the logistics of irradiated ampoules. RS reviewed the manuscript. NvdM was responsible for the development of the production and separation process of ¹⁶¹Tb, organized and supervised the whole study and finalized the manuscript. All authors read and approved the final manuscript.

Funding

The research was funded by the Swiss National Science Foundation (IZLIZ3_156800). The authors acknowledge the Neuroendocrine Tumor Research Foundation for the support of the project through receiving the Petersen Investigator Award 2018.

Availability of data and materials

The dataset(s) supporting the conclusions of this article is (are) included within the article (and its additional file(s)).

Ethics approval and consent to participate

This article does not contain any studies with human participants performed by any of the authors.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.