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

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.

The radiolanthanide 161 Tb shows similar decay characteristics (E β-av = 154 keV (100%), T 1/2 = 6.9 d (Solá 2000)) and coordination chemistry to 177 Lu. 161 Tb 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 177 Lu. 161 Tb may show an increased therapeutic efficacy over 177 Lu, 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 161 Tb and~1 e − ,~1.0 keV per decay for 177 Lu, respectively) (Eckerman 2008;Müller 2014). The possibility of using 161 Tb as an alternative to 177 Lu 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 161 Tb and 177 Lu (Müller 2014). The enhanced anti-tumor effect, as well as higher average survival time, was found in mice treated with 161 Tb-folate over those which received 177 Lu-folate. In a preliminary therapy study using 161 Tb-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 161 Tb-labeled radioimmunoconjugate in an ovarian cancer model when compared to the 177 Lu-radioimmunoconjugate counterpart (Grünberg 2014). The low-energy conversion and Auger electron emission from 161 Tb contribute 26% -88% to the total absorbed dose (compared to 10% -34% for 177 Lu), depending on the tumor size, which could be associated to its enhanced therapeutic efficacy over that of 177 Lu (Champion 2016). The doses delivered by 161 Tb or 177 Lu to 10 mm-diameter spheres were calculated to be comparable for both radionuclides, however, for 100 μm-diameter and 10 μm-diameter spheres 161 Tb could deliver 1.8 and 3.6 times higher dose than 177 Lu, respectively, making 161 Tb the more appropriate candidate for treating micrometastases (Hindie 2016). Also, the co-emission of 48.9 keV and 74.6 keV 161 Tb 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 177 Lu (Marin 2018). In addition, 161 Tb could be used in combination with diagnostic radioisotopes, namely, 152 Tb (PET) or 155 Tb (SPECT) as a matched pair towards the concept of theragnostics (Müller 2012;Müller 2018b). Gracheva et al. EJNMMI Radiopharmacy and Chemistry (2019)  The 161 Tb production route was proposed by Lehenberger et al. via the 160 Gd(n, γ) 161 Gd → 161 Tb nuclear reaction, which provided no-carrier-added radiolanthanide at high specific activities (~4 TBq/mg) (Lehenberger 2011). Enriched 160 Gd(NO 3 ) 3 targets (ampoules) were prepared by dissolving 160 Gd 2 O 3 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 161 Tb separation method from 160 Gd(NO 3 ) 3 targets was previously developed at Paul Scherrer Institute (Villigen-PSI, Switzerland), but the radiolabeling capability of the 161 Tb product was three times lower than the commercial no-carrier-added 177 Lu (Lehenberger 2011). This implied that the 161 Tb 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 161 Tb production from 160 Gd 2 O 3 target material, suitable for introduction into a process in accordance with Good Manufacturing Practice (GMP) and, thereafter, clinical application. The 161 Tb purification method was improved by optimization of the Tb/Gd separation process, followed by characterization of the final product ( 161 TbCl 3 ). The 161 Tb purity was compared with no-carrier-added 177 Lu (EndolucinBeta), as currently produced by ITG GmbH, Germany, for worldwide clinical application.

Methods
Target preparation for the production of 161 Tb Gadolinium oxide ( 160 Gd 2 O 3 , 98.2% enrichment, Isoflex, USA) was used as target material for the production of no-carrier-added 161 Tb, 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 spallationinduced neutron source (SINQ, Paul Scherrer Institute, 4.10 13 n.cm − 2 .s − 1 ), 80-95 mg 160 Gd(NO 3 ) 3 were placed in a quartz glass ampoule (Suprasil, Heraeus, Germany) and sealed. Ampoules containing 7-33 mg 160 Gd 2 O 3 were prepared in a similar manner and sent for irradiation to two research nuclear reactors (SAFARI-1, South African Nuclear Energy Corporation, 2.10 14 n.cm − 2 .s − 1 ; and RHF ILL, Institut Laue-Langevin, 1.10 15 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 59 Co monitors
In order to monitor neutron fluxes at ILL, SAFARI-1 and SINQ, quartz ampoules containing 59 Co as standard ( 59 Co in 2% w/w HNO 3 , Sigma-Aldrich, USA) with 33 ng -2 μg 59 Co (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 59 Co standard were placed and sealed in the same Al container as the ampoules containing 160 Gd target material, along with empty ampoules (used as references) for the irradiation process. 59 Co masses were calculated to produce 50-100 kBq 60 Co activity via 59 Co(n,γ) 60 Co nuclear reaction, depending on the reactor neutron flux.
The 60 Co activities in the ampoules were measured after irradiations using a high- Gracheva et al. EJNMMI Radiopharmacy and Chemistry (2019)  purity germanium (HPGe) detector (Canberra, France), in combination with the Inter-Winner 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 60 Co produced by the added 59 Co standards were determined by subtraction of the 60 Co activities of the reference (empty) ampoules which stem from traces of cobalt impurities in the used quartz. Based on these 60 Co activity values, the average neutron flux (ϕ th ) of each irradiation was calculated (Table 1) using the following equation: where A 0 is the 60 Co 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.
Development of the procedure for the 161 Tb 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, NH 4 + 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 161 Tb (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 160 Gd 2 O 3 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 (HNO 3 , 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 NH 3 and 65% Suprapur HNO 3 , Merck, Germany) and loaded onto the cation exchange resin column. The 161 Tb 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 161 Tb 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 ( 161 TbCl 3 ) 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 161 Tb product after purification

Radionuclidic purity
The identification and radionuclidic purity of the 161 Tb were examined by γ-ray spectrometry using the HPGe detector mentioned above. The aliquot of the final product, containing 5-10 MBq of 161 Tb, 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 177 Lu (Oliveira 2011). The aliquot of 161 TbCl 3 (2 μL,~100 kBq) was deposited on TLC silica gel 60 F 254 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 161 TbCl 3 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 reversephase 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 161 Tb-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 161 Tb as well as degradation products of unknown structure), which were set to 100%.
Determination of the radiolabeling yield of 161 Tb-and 177 Lu-DOTA over a two-week period In order to assess the change of the molar activity of 161 Tb-DOTA at different DOTA-tonuclide molar ratios over time (2 weeks after EOS) and to compare it with the molar activity of 177 Lu-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 161 Tb (corresponding to 3.1-5 pmol), Lu (no-carrier-added, ITG GmbH, Germany) or 177 Lu (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 F 254 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 161 Tb or 177 Lu (R f = 0) and to the 161 Tb-or 177 Lu-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 161 Tb and 177 Lu (no-carrier-added) and once for 177 Lu (carrier-added). The average DOTA-to-nuclide molar ratios, corresponding to 50% labeling efficiency of DOTA with 161 Tb (no-carrier-added) and 177 Lu (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).

Tb/ 177 Lu-DOTATOC stability studies
The radiolabeling of DOTATOC with no-carrier-added 161 Tb or no-carrier-added 177 Lu at 50 MBq/nmol molar activity (300 MBq 161 Tb 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 161 Tb/ 177 Lu-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.

Tb production yield (theoretical versus experimental)
No-carrier-added 161 Tb was produced by neutron irradiation of enriched 160 Gd 2 O 3 (98.2% enrichment) targets via the 160 Gd(n,γ) 161 Gd → 161 Tb nuclear reaction. The mass of the target material had to be calculated precisely in order to ensure that the 161 Tb activity allowed for international transportation was not exceeded. 161 Tb 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 161 Tb could be transported internationally (in this case, shipping from ILL and SAFARI-1 research reactors to PSI) (IAEA 2012). Masses of the 160 Gd target material, required to produce 20 GBq 161 Tb after bombardment at the irradiation facilities, were calculated with the ChainSolver code using the recommended cross-section for thermal neutron capture on 160 Gd of 1.4(3) b (Mughabghab 2018). Two-week irradiations at ILL (6.5 mg 160 Gd 2 O 3 , 1.10 15 n.cm − 2 .s − 1 , 1 day cooling) and at SAFARI-1 (32 mg 160 Gd 2 O 3 , 2.10 14 n.cm − 2 .s − 1 , 1 day cooling) would theoretically result in 20 GBq 161 Tb. At PSI's neutron source facility (SINQ, 4·10 13 n.cm − 2 .s − 1 ) each irradiation cycle is 3 weeks, which was calculated to provide 17.2 GBq 161 Tb after the bombardment of 100 mg 160 Gd(NO 3 ) 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 60 Co activity values of the 59 Co monitors (Table 1). Three ampoules with 59 Co were bombarded at the SAFARI-1 nuclear reactor and one ampoule each at ILL and at SINQ irradiation facilities (together with the 160 Gd 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 160 Gd target ampoules using SAFARI-1 (22-33 mg 160 Gd 2 O 3 ) and ILL (7-13 mg 160 Gd 2 O 3 ) research reactors resulted in production of 10-20 GBq of 161 Tb ( 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 10 13 n.cm − 2 .s − 1 (Table 1), which resulted in 6-9 GBq 161 Tb after 3 weeks irradiation of the enriched 160 Gd(NO 3 ) 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).

Radiochemical separation of 161 Tb 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 Gd 2 O 3 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 161 Tb (230 MBq). During the irradiation, radioactive side products were co-produced from the impurities of the target material ( 46 Sc, 124 Sb, 141 Ce, 147 Nd, 153 Gd, 153 Sm, 152/154/155/156 Eu, 169 Yb) and the ampoule material  161 Tb separation from Gd target material and the impurities using the Sykam resin column (Fig. 1). 161 Tb 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 161 Tb to be eluted in a small volume of 0.05 M HCl.
Based on the developed two-column purification method (combination of Sykam and LN3 resin columns), a 161 Tb 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 161 Tb.
The established procedure for the 161 Tb purification process using the designed module resulted in the elution of the final product ( 161 TbCl 3 ) 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 161 Tb activity were observed in the target dissolving, column loading and final elution steps.
Characteristics of the 161 TbCl 3 product 161 Tb, obtained after the purification process, was characterized to provide a product specification (Table 3). The identification of the product was confirmed by the 161 Tb-characteristic γ-lines (Fig. 3). The content of long-lived 160 Tb (T 1/2 = 72.3 d), produced by the 159 Tb(n,γ) 160 Tb nuclear reaction due to the presence of 159 Tb impurity in the target material (as sold by the vendor), was determined after the decay of 161 Tb and did not exceed 0.007% of the total 161 Tb activity at EOS (Additional file 1: Figure S4). The radiochemical purity of the 161 TbCl 3 , determined using radio-TLC, was > 99% (Additional file 1: Figure  S5). The radiolabeling yield of 161 Tb-DOTANOC showed ≥99% efficiency at 180 MBq/ nmol molar activity, which corresponds to 1-to-4 nuclide-to-peptide molar ratio (Fig. 4).
Comparison of the 161 Tb and 177 Lu quality, based on the 161 Tb-and 177 Lu-DOTA molar activities over a two-week period Radiolabeling of DOTA with 161 Tb (no-carrier-added) and 177 Lu (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 161 Tb and 177 Lu (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 161 Tb) for up to 2 weeks after the chemical separation. With carrier-added 177 Lu, 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).
The average DOTA-to-nuclide molar ratios corresponding to 50% labeling efficiency of DOTA with 161 Tb and 177 Lu 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 177 Lu, 50% labeling efficiency was obtained at higher DOTA-to-nuclide molar ratios at each time point as compared to no-carrier-added 161 Tb and 177 Lu, respectively. The 50% labeling efficiency of 161 Tb-DOTA was found to be comparable with that of no-carrier-added 177 Lu 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.

Radiolytic stability of 161 Tb-DOTATOC in comparison to 177 Lu-DOTATOC
In order to determine whether the conversion and Auger electrons from 161 Tb 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. 177 Lu-DOTATOC, 177 Lu-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 161 Tb-DOTATOC and 177 Lu-DOTATOC were performed with and without addition of L-ascorbic acid. Both 161 Tb-and 177 Lu-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. 161 Tb-DOTATOC and 177 Lu-DOTATOC were stable over 1 h (> 99% intact product) in the absence of Lascorbic acid, but both showed radiolytic degradation after an incubation period of 24 h. After 4 h, slight degradation of both 161 Tb-DOTATOC (78% intact product) and 177 Lu-DOTATOC (85% intact product) was observed (Fig. 6). No significant influence of conversion and Auger electrons from 161 Tb on radioligand stability was observed, when compared to 177 Lu.

Discussion
In the present study, the development of a reproducible chemical separation to produce no-carrier-added 161 Tb from enriched 160 Gd targets and the characterization of the final product ( 161 TbCl 3 ) is reported. Gd (NO 3 ) 3 , previously used as target material (Lehenberger 2011), is not suitable for large-scale 161 Tb 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 161 Tb production in future (TBq activity). The ChainSolver code allows for quick calculations of the required masses of 160 Gd targets for the production of the desired 161 Tb activity, based on irradiation times and neutron fluxes, obtained with 59 Co monitors.
Another big advantage of the newly-devised method over that of the previouslydeveloped one is that the 161 Tb 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 161 Tb 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 161 Tb separation from the Gd 2 O 3 target material of masses up to 140 mg. This is a valuable result for possible future commercial application of the developed method for 161 Tb separation from massive Gd 2 O 3 targets (> 100 mg). For example, 2 weeks irradiation of 160 mg 160 Gd 2 O 3 target in ILL's nuclear reactor could theoretically result in the production of 0.5 TBq 161 Tb (ChainSolver Code calculations (Romanov 2005)) which can be efficiently purified with the investigated method. 160 Gd 2 O 3 target material contains various trace elements (Additional file 1: Table S2), which may show similar chemical behavior to 161 Tb 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 161 Tb from potential impurities present in the system, based on the combination of Sykam and LN3 resin columns. The two-column 161 Tb 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 ( 161 TbCl 3 ) in purity comparable to that of the commercially-available no-carrier-added 177 LuCl 3 (EndolucinBeta). The radiolabeling capability of 161 Tb in this work was similar to no-carrier-added 177 Lu and three times higher than that of 161 Tb obtained by Lehenberger et al. (Lehenberger 2011). Somatostatin analogues could be labeled with no-carrier-added 161 Tb (this work) at 1-to-4 nuclide-topeptide molar ratio (Fig. 4) immediately after the purification process with > 99% radiolabeling yield. Quantitative formation of 161 Tb-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 161 Tb was higher than the radiolabeling capability of carrier-added 177 Lu at each time point over the two-week decay period after the chemical separation process. The faster drop of 161 Tb radiolabeling capability as compared to no-carrier-added 177 Lu 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 161 TbCl 3 vs 36-44 MBq/μL for 177 LuCl 3 ). This implies that the mass ratio between 161 Tb and the impurities (Zn, Fe, Co, Pb etc.), which can be introduced during product analysis and post-processing, is lower compared to 177 Lu. This results in the potentially stronger interference from environmental impurities during radiolabeling of DOTA with 161 Tb than with 177 Lu (Asti 2012). Nevertheless, complexation of 161 Tb 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 161 TbCl 3 for a period of up to 2 weeks after EOS. Moreover, clinicallyapplied DOTATOC radiolabeled with 161 Tb 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 161 Tb-labeled somatostatin analogues would be feasible, as is the case for their 177 Lu-labeled counterparts.

Conclusions
A new method to separate 161 Tb from its enriched 160 Gd 2 O 3 target material and coproduced impurities was developed with the use of cation exchange and extraction chromatography, respectively. The method resulted in radionuclidically and radiochemically pure product ( 161 TbCl 3 ), comparable to commercially available, no-carrier-added 177 Lu. The quantity and quality of 161 Tb is suitable for high-specific radiolabeling, potentially useful for the GMP production of radioligands towards future clinical application.