Combination of terbium-161 with somatostatin receptor antagonists—a potential paradigm shift for the treatment of neuroendocrine neoplasms

Purpose The β¯-emitting terbium-161 also emits conversion and Auger electrons, which are believed to be effective in killing single cancer cells. Terbium-161 was applied with somatostatin receptor (SSTR) agonists that localize in the cytoplasm (DOTATOC) and cellular nucleus (DOTATOC-NLS) or with a SSTR antagonist that localizes at the cell membrane (DOTA-LM3). The aim was to identify the most favorable peptide/terbium-161 combination for the treatment of neuroendocrine neoplasms (NENs). Methods The capability of the 161Tb- and 177Lu-labeled somatostatin (SST) analogues to reduce viability and survival of SSTR-positive AR42J tumor cells was investigated in vitro. The radiopeptides’ tissue distribution profiles were assessed in tumor-bearing mice. The efficacy of terbium-161 compared to lutetium-177 was investigated in therapy studies in mice using DOTATOC or DOTA-LM3, respectively. Results In vitro, [161Tb]Tb-DOTA-LM3 was 102-fold more potent than [177Lu]Lu-DOTA-LM3; however, 161Tb-labeled DOTATOC and DOTATOC-NLS were only 4- to fivefold more effective inhibiting tumor cell viability than their 177Lu-labeled counterparts. This result was confirmed in vivo and demonstrated that [161Tb]Tb-DOTA-LM3 was significantly more effective in delaying tumor growth than [177Lu]Lu-DOTA-LM3, thereby, prolonging survival of the mice. A therapeutic advantage of terbium-161 over lutetium-177 was also manifest when applied with DOTATOC. Since the nuclear localizing sequence (NLS) compromised the in vivo tissue distribution of DOTATOC-NLS, it was not used for therapy. Conclusion The use of membrane-localizing DOTA-LM3 was beneficial and profited from the short-ranged electrons emitted by terbium-161. Based on these preclinical data, [161Tb]Tb-DOTA-LM3 may outperform the clinically employed [177Lu]Lu-DOTATOC for the treatment of patients with NENs. Supplementary Information The online version contains supplementary material available at 10.1007/s00259-021-05564-0.

More recently, the challenge of tumor radioresistance triggered the concept of targeted α-therapy, which is particularly effective due to the high linear energy transfer (LET) of α-particles (50-230 keV/μm) [13][14][15]. PRRT using [ 213 Bi]Bi-DOTATOC or [ 225 Ac]Ac-DOTATATE in patients, refractory to 90 Y-and/or 177 Lu-based PRRT, resulted in partial remission or stable disease [16,17]. There are, however, a number of uncertainties regarding the application of α-emitters, among those, the unfavorably short half-life of bismuth-213 and the complicated decay scheme of actinium-225, which comprises a considerable risk of undesired side effects. The inability of imaging these α-emitters and, finally, the difficult production scenarios make their application challenging [18][19][20].
In 2006, Ginj et al. introduced the concept of SSTR antagonists for targeting NENs [21]. Preclinical studies demonstrated much higher tumor accumulation of these non-internalizing SST analogues than for SSTR agonists. In a first-in-human clinical application with four patients with metastatic NENs, the antagonist [ 177 Lu]Lu-DOTA-JR11 performed better than [ 177 Lu]Lu-DOTATATE [22], which led to a Phase 1/2 multicenter clinical study (NCT02592707). More recently, [ 177 Lu]Lu-DOTA-LM3, another SSTR antagonist [23], was tested in 51 patients with metastatic NENs [24]. It resulted in higher tumor doses compared to [ 177 Lu]Lu-DOTATOC and was well tolerated in patients [24]. Unfortunately, this promising new class of SSTR antagonists cannot be combined with α-particle emitters with complex decay chain, such as actinium-225, for which effective internalization is an essential prerequisite to avoid toxicity to healthy tissues as a result of released daughter nuclides [18].
A central question in the context of using terbium-161 refers to the type of targeting agent which would ensure the greatest benefit from the effect of the co-emitted conversion and Auger electrons. It is commonly believed that nuclear localization of a peptide would be essential to induce DNA double-strand breaks (DSBs) by short-ranged electrons [36,37]. It was, however, also shown that the cell membrane may be a suitable target for Auger electron emitters [38]. We believe that the non-internalizing SSTR antagonists may, thus, benefit from the co-emitted Auger electrons of terbium-161, which could eventually lead to a paradigm shift in PRRT.
The aim of this study was, therefore, to investigate the impact of the localization of 161 Tb-based SST analogues to benefit from the short-ranged electrons in PRRT of NENs. The therapeutic efficacy of three 161 Tb-labeled SST analogues supposed to localize (i) in the cytoplasm (DOTA-TOC, SSTR agonist [39]), (ii) in the cytoplasm and cell nucleus (DOTATOC-NLS [40,41]) or at the cell membrane (DOTA-LM3, SSTR antagonist [23]) was compared with those of their 177 Lu-labeled counterparts.

Radiopeptide preparation
No-carrier-added (n.c.a.) terbium-161 was produced at Paul Scherrer Institute, Switzerland, as previously reported [42] and n.c.a. lutetium-177 was obtained from ITM Medical Isotopes GmbH, Germany. Radiolabeling and quality control of the SST analogues were performed as previously reported [43]

Study design
Initially, the tumor uptake, internalization, and subcellular localization of the three SST analogues were assessed in SSTR-positive cancer cells. Afterwards, the therapeutic effect of the 177 Lu-and 161 Tb-labeled somatostatin analogues was investigated in vitro using cell viability and survival assays. In vivo, the biodistribution of the radiopeptides was determined under variable conditions, after which the most promising setting was used for an in vivo comparison of the therapeutic effects of the 177 Lu-and 161 Tb-labeled SST analogues.
Since it was previously confirmed that the SST analogues have equal in vitro and in vivo behavior regarding (sub)cellular uptake and tissue distribution, irrespective of whether they were labeled with lutetium-177 or terbium-161 [43], those experiments were performed with only one version of the radiolabeled peptides.

AR42J tumor cell uptake and internalization
AR42J tumor cells, a SSTR-positive exocrine rat pancreatic cancer cell line [44] (ECACC 93,100,618, Health Protection Agency Culture Collections, Salisbury, UK), were cultured as previously reported (Supplementary Material) [43]. Cell uptake and internalization studies were performed to determine the fractions of radiopeptides localized at the membrane or in the cytoplasm, respectively. AR42J tumor cells were grown as monolayers in 12-well plates overnight, followed by incubation of the cells with each of the 177 Lulabeled radiopeptides (~ 15 kBq, ~ 0.375 pmol per well) for 2 h. The total uptake and internalization were determined after rinsing the cells with phosphate-buffered saline (PBS) and acidic glycine buffer, respectively. The uptake after SSTR saturation was evaluated after incubating the cells with increasing molar amounts of the SST analogues (Supplementary Material).

Determination of the cellular localization of the radiopeptides
The nuclear accumulation of activity was determined after a 2 h-incubation period of the AR42J tumor cells with the respective radiopeptide using a nuclei isolation kit (Nuclei EZ Prep Kit, Sigma-Aldrich, St. Louis, U.S.) (Supplementary Material). The nuclei isolation assay was performed in triplicate with each 177 Lu-labeled radiopeptide. The nuclear fraction of the respective radiopeptide was expressed as percentage of total cell uptake determined as described above.

Cell viability and survival after treatment
The viability of AR42J tumor cells after incubation with 161 Tb-or 177 Lu-labeled DOTATOC, DOTATOC-NLS or DOTA-LM3 (0.001-40 MBq/mL, 0.01-400 pmol/mL) was determined using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay [33,45]. The cell viability of treated cells was presented as average ± standard deviation (SD) of 4-5 independent experiments and expressed as percentage of sham-treated cells (set as 100%). The data were plotted against the applied activity concentration in logarithmic scale and fitted with a dose-response curve to determine the activity concentration necessary to reduce AR42J tumor cell viability to 50% of untreated control cells (EC 50 ). The survival of the treated AR42J tumor cells was investigated using a clonogenic assay [46]. AR42J tumor cells were seeded as single cells in 6-well plates (2000 cells per well) and exposed to variable activity concentrations (0.01-5 MBq/mL, 0.33-170 pmol/mL) of 161 Tb-and 177 Lu-labeled SST analogues. Two hours later, the activity was removed and the cells were rinsed and let to grow into colonies over 2 weeks. The colonies were colored with crystal violet and counted as previously described (Supplementary Material) [46]. The survival of treated cells was expressed as average ± SD of 3 independent experiments performed in triplicates and expressed as percentage of sham-treated cells (set as 100%). The survival data were analyzed with a two-way ANOVA with Sidak's multiple comparisons post-test.

DNA damage evaluation
The number of DNA DSBs induced in AR42J tumor cells after exposure to the radiopeptides was assessed by immunostaining of γH2AX. The cells were treated in Petri dishes for 2 h using 2.5 MBq/mL or 10 MBq/mL of each radiopeptide followed by incubation with fresh culture medium for additional 24 h. Cell pellets obtained after centrifugation were fixed and embedded in paraffin and cut into sections. The immunostaining was performed using a phosphohistone H2A.X (Ser139) rabbit monoclonal antibody (Cell Signaling Techonology, Danvers, Massachusetts, USA) and an anti-rabbit, horseradish peroxidase-derivatized secondary antibody with DAB substrate buffer (Agilent Technologies, Santa Clara, California, USA) (Supplementary Material). The sections were scanned using a digital slide scanner (NanoZoomer-XR C12000; Hamamatsu, Japan) and the positively and negatively stained AR42J tumor cells were quantified with the pathology image analysis software VIS (Visiopharm Integrator System, Version 208 2019.02.2.6239, Visiopharm, Hoersholm, Denmark) (Supplementary Material). The percentage of γH2AX-positive cells in sham-treated samples was in average 1%. Data were analyzed with a one-way ANOVA with Dunnet's multiple comparisons post-test.

In vivo studies
The animal experiments were carried out according to the guidelines of Swiss Regulations for Animal Welfare, ethically approved by the Cantonal Committee of Animal Experimentation and permitted by the responsible cantonal authorities (license N° 75721 and 79692). Five-week-old female, athymic nude mice (CD-1 Foxn-1/nu) were obtained from Charles River Laboratories (Sulzfeld, Germany). Mice were subcutaneously inoculated with AR42J tumor cells (5 × 10 6 cells in 100 µL PBS) for SPECT/CT imaging, biodistribution, and therapy studies.

SPECT/CT imaging and biodistribution studies
The studies were performed 10-14 days after tumor cell inoculation when the tumor size reached a volume of ~ 250 mm 3 .  [43]. In a first series of biodistribution studies, it was subsequently assessed that a molar peptide amount of 0.2 nmol per mouse was the optimum to achieve high tumor-to-background ratios (Supplementary Material). The time-dependent biodistribution studies were, therefore, performed with mice (n = 3 per group) intravenously injected with 0.2 nmol of the radiolabeled DOTATOC or DOTA-LM3 (5 MBq, in 100 μL PBS containing 0.05% bovine serum albumin (BSA)), respectively. Selected tissues and organs were collected, weighed, and counted for activity using a γ-counter (Perkin Elmer). The decay-corrected data were listed as a percentage of the injected activity per gram of tissue mass (% IA/g). The data were analyzed for significance using a two-way ANOVA with Tukey's multiple comparisons post-test.

Therapy study
The therapy study was initiated with mice randomly assigned to five groups (n = 6) when the AR42J tumors reached an average volume of 99 ± 16 mm 3 . At day 0 and day 7 of the study, the mice were intravenously injected with vehicle only ( Table 1). The relative body weight (RBW) and the relative tumor volume (RTV) were defined based on the values at therapy start as previously described (Supplementary Material) [47]. The endpoint criteria were defined according to a scoring system which required euthanasia of mice with a score ≥ 3 (Supplementary Material).

Assessment of the therapeutic efficacy
The efficacy of the treatment was assessed by comparison of the RTVs, measured every second day, of mice in each group using a two-way ANOVA with Sidak's multiple comparisons post-test The average tumor growth delay, herein defined as the time during which the tumors did not grow or even decreased in size, was determined for mice of each group. For the subsequent phase, in which the tumors started to regrow, the doubling time of the tumor volume was calculated based on the fitted exponential tumor growth curve. The average ± SD of tumor growth delay and of the doubling time of the tumor volume in single mice, respectively, were compared among groups with a one-way ANOVA with Tukey's multiple comparisons post-test. The survival times of mice were presented by Kaplan-Meier curves and analyzed using a log-rank test (Mantel-Cox).

Assessment of early side effects
Early side effects were assessed based on the RBW of the mice. When an endpoint was reached or at the end of the study (at day 49), blood plasma parameters (albumin (ALB), creatinine (CRE), blood urea nitrogen (BUN), alkaline phosphatase (ALP), and total bilirubin (TBIL)) were determined (Supplementary Material). At the same time, relevant organs and tissues were collected, and the masses were put into relation to the brain mass of the respective mouse in order to allow the comparison of organ-to-brain and organ-to-body weight ratios among the single treatment groups (Supplementary Material). Data were analyzed for significance using a one-way ANOVA test with a Tukey's multiple comparisons post-test.

Statistical analysis and graphs preparation
The GraphPad Prism software (version 8) was used for preparation of the graphs, for the analysis of the data and to perform statistical analysis. A p value < 0.05 was considered statistically significant. SPECT/CT images were prepared using the software CorelDRAW (version X7).

In vitro tumor cell viability
It was observed that, in all cases, the 161 Tb-labeled SST analogue was more potent in reducing AR42J tumor cell viability than the 177 Lu-labeled analogue (Fig. 2a-

In vitro tumor cell survival
Colony-forming assays confirmed that the 161 Tb-labeled peptides reduced AR42J tumor cell survival more effectively than the respective 177 Lu-labeled analogues (Fig. 2d/

SPECT/CT imaging studies
As previously reported [43], [ 161 Tb]Tb-DOTATOC accumulated in the tumors, but retention in the kidneys was also visible on the 2 h p.i.-SPECT/CT scan (Fig. 4a). At the same timepoint after injection of [ 161 Tb]Tb-DOTA-TOC-NLS, the tumor uptake was slightly increased, but retention in the liver and kidneys was exceedingly high. As a result, this radiopeptide would not be applicable for therapeutic purposes (Fig. 4b) (Supplementary Material,  Fig. S4). [ 161 Tb]Tb-DOTA-LM3 showed the most favorable tissue distribution, with an increased tumor uptake, resulting in the highest tumor-to-kidney ratios (Fig. 4c) [43]. The tumor uptake was SSTR-specific for all radiopeptides, as previously reported (Supplementary Material) [43].
The tumor uptake and retention of [ 161 Tb]Tb-DOTA-LM3 was significantly higher (35 ± 7% IA/g; 4 h p.i.; 21 ± 4%IA/g at 48 h p.i.) than for [ 161 Tb]Tb-DOTATOC at all investigated timepoints (p < 0.05). In the lungs, stomach, and pancreas, activity was detected during the first 4 h but effectively cleared over the hours that followed. The accumulated activity of [ 161 Tb]Tb-DOTA-LM3 in the pancreas and stomach was significantly higher than in the case of [ 161 Tb]Tb-DOTATOC (p < 0.05) at all investigated timepoints, whereas renal uptake was similar as observed for [ 161 Tb]Tb-DOTATOC (p > 0.05) (Fig. 5b and Supplementary Material Table S5). As previously demonstrated [43], the in vivo distribution of 161 Tb-and 177 Lu-labeled SSTanalogues is equal; hence, the results hold true irrespective of which radionuclide was employed.

Therapy study in AR42J tumor-bearing mice
Sham-treated mice of Group A showed an exponential tumor growth so that the endpoint was reached within the first 14 days in all cases. The tumor growth was delayed in treated mice of Groups B-E, resulting in significantly prolonged median survival times compared to the 9 days in the control group (Fig. 6, Table 2). Mice treated with [ 161 Tb]Tb-DOTATOC (Group B) showed a slightly slower tumor growth compared to mice treated with [ 177 Lu]Lu-DOTATOC (Group C). After 12 days from the therapy start, the RTV of mice of these two groups were significantly different (2.0 ± 0.7 vs 4.0 ± 3.3, p < 0.05). The tumor growth delay and doubling time were higher in Group B (9.0 ± 5.5 days and 3.4 ± 3.6 days, respectively) compared to those of Group C (6.0 ± 4.4 days and 3.4 ± 0.8 days, respectively, p > 0.05). Mice of Group B were, thus, euthanized at a later stage (day 20-26) compared to Group C (day [12][13][14][15][16][17][18][19][20][21][22]. The median survival (21 vs 19.5 days) was, however, comparable between the groups (p > 0.05).
The tumor growth delay for mice treated with [ 161 Tb] Tb-DOTA-LM3 was 44 ± 5 days, but only 35 ± 7 days in mice treated with [ 177 Lu]Lu-DOTA-LM3 (p < 0.05). Afterwards, the tumors started to regrow exponentially in 5 out of 6 mice of both groups; however, the doubling time was considerably longer for mice treated with

Assessment of potential early side effects
No signs of early side effects were detected in any of the treated groups (Fig. 7, Supplementary Material, Tables S6-S9). The mice gained weight over the course of the study and the body weights did not differ between the groups at any timepoint (p > 0.05) (Fig. 7a/b). ALB plasma levels, an indicator of general health status of mice, were comparable among the groups (Fig. 7c). No signs of kidney toxicity were observed at the time of euthanasia, as kidneyto-brain mass ratios and plasma CRE levels showed no significant difference among the groups with the exception of one outlier (Fig. 7d/e). The BUN levels were elevated in mice treated with [ 161 Tb]Tb-DOTA-LM3 (9.9 ± 1.5 mmol/L) and [ 177 Lu]Lu-DOTA-LM3 (8.2 ± 1.9 mmol/L) compared to control mice (6.2 ± 0.7 mmol/L) and mice treated with [ 161 Tb] Tb-or [ 177 Lu]Lu-DOTATOC, but in the physiological range reported for this particular mouse strain by the breeding company (Charles River, Germany) (Fig. 7f). No liver toxicity was observed based on liver-to-brain mass ratios, plasma ALP, and TBIL which were comparable among all groups (p > 0.05) (Fig. 7g/h/i).

Discussion
Terbium-161 and lutetium-177 share similar physical decay properties in terms of half-life and β¯-energy; however, the co-emission of short-ranged electrons is a specific characteristic of terbium-161. A direct comparison of the therapeutic effect of terbium-161 and lutetium-177 is feasible due to the similar chemical properties of these radiolanthanides and, thus, equal tumor cell uptake and biodistribution profiles of the investigated pairs of radiopeptides herein [43]. In this study, we investigated the impact of the subcellular localization of SST analogues on the therapeutic efficacy of terbium-161 and lutetium-177, respectively. Both DOTATOC and DOTATOC-NLS internalized to a large degree into the cytoplasm (80% and 84%) whereof, in the case of DOTATOC-NLS, ~ 15% of the internalized fraction were shuttled to the nucleus. In contrast, the largest fraction of the SSTR antagonist DOTA-LM3 remained at the cell membrane (> 90%).
Indeed, terbium-161 was more effective in reducing AR42J tumor cell viability and survival than lutetium-177, irrespective of the SST analogue employed. This indicated clearly that the conversion and Auger electrons emitted by terbium-161 contributed positively to its therapeutic efficacy. The difference in therapeutic efficacy between 161 Tb-and 177 Lu-labeled peptides was, however, not the same among the three types of SST analogues.
In line with the common belief that Auger electron emitters should be shuttled in close proximity to the cell DNA [36,37] . 7 Parameters of potential side effects in therapy mice. a/b/c Indicators for the general health status of the mice: RBW during the therapy (a) shown until the first mouse of the group reached an endpoint, RBW at the endpoint (b), and plasma ALB levels (c). d/e/f Indicators for kidney toxicity: kidney-to-brain mass ratios (d), plasma CRE (e), plasma BUN levels (f). g/h/i Indicators for liver toxicity: liver-tobrain mass ratios (g), plasma ALP levels (h), and plasma TBIL levels (i). Organ-to-brain mass ratios and plasma parameters refer to mice at the endpoint Interestingly, the non-internalizing [ 161 Tb]Tb-DOTA-LM3 revealed a 102-fold increased efficacy to reduce cell viability in vitro compared to [ 177 Lu]Lu-DOTA-LM3 (Fig. 2). These findings were in line with a previous observation made with internalizing and non-internalizing antibodies, wherof the latter were more effective in killing tumor cells when labeled with the Auger electron-emitting iodine-125 [38,48,49]. Obviously, the cell membrane is a better target to exploit the effect of short-ranged electrons than the cytoplasm. Current endeavors in our laboratories focus, thus, on a further radiobiological investigation of the effect of [ 161 Tb]Tb-DOTA-LM3 to better understand the "membrane effect" of shortranged electrons.
Preclinical therapy studies with AR42J tumor-bearing mice confirmed the superior efficacy of 161 Tb-labeled SST analogues over those labeled with lutetium-177. The obvious positive effect of Auger electrons emitted by terbium-161 was particularly pronounced when using DOTA-LM3. The overall efficacy of DOTA-LM3-based PRRT was much better than the results obtained with DOTATOC, irrespective of which radionuclide was applied. It can be ascribed to the higher tumor accumulation of SSTR antagonists compared to agonists. [ 161 Tb]Tb-DOTA-LM3 induced a tumor growth delay of over 44 days and was, thus, the most powerful candidate, with a significant advantage over the clinically applied [ 177 Lu] Lu-DOTATOC, which induced a tumor growth delay of only 6 ± 4 days. No significant early side effects nor signs of liver or kidney toxicity were observed in any of the treated mice, yet further investigations are necessary in order to assess potentially negative effects to SSTR-expressing normal tissues or bone marrow toxicity.
Even though AR42J tumor cells show neuroendocrine properties [50] and have, thus, been extensively employed for the evaluation of SST analogues, it can be considered a limitation of this study to have only used rat tumor cells. It will, thus, be crucial to confirm the reported findings using human endocrine pancreatic tumor cells. The investigation of further "agonist/antagonist" pairs that target SSTR or other receptors will provide clarification about the general validity of the proposed advantage of terbium-161 when applied in combination with non-internalizing tumor-targeting agents. Both topics are in the focus of our currently-ongoing research activities.
Our previous work already showed that terbium-161 provides an advantage over the use of lutetium-177 [33][34][35]. If the concept presented herein will be confirmed in follow-up studies, terbium-161 may be used preferentially with noninternalizing tumor-targeting agents, thereby, complementing the use of α-particle emitters, which should be combined with fast-internalizing molecules.
Finally, our results, together with the recently reported data of the successful application of [ 177 Lu]Lu-DOTA-LM3 in patients [24,51], encourage a clinical translation of [ 161 Tb]Tb-DOTA-LM3 for the treatment of NEN's. Current efforts at PSI are, therefore, dedicated to the scale-up of the terbium-161 production, which is performed in analogy to the production of n.c.a. lutetium-177 (Supplementary Material, Fig. S6) [25,42] to enable the preparation of therapeutic quantities of [ 161 Tb]Tb-DOTA-LM3 under good manufacturing practice (GMP) for clinical application.

Conclusion
This study showed that the cellular localization of 161 Tblabeled SST analogues is relevant to exploit the therapeutic potential of short-ranged electrons. [ 161 Tb]Tb-DOTA-LM3 revealed to be powerful for the treatment of NENs, as it combines the favorable properties of SSTR antagonists with the advantages of terbium-161. Should our results be confirmed with other tumor models and targeting concepts, these results may initiate a paradigm shift towards the application of Auger electron emitters with membrane-targeting biomolecules to enable more effective PRRT in NENs.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.