Exploring the Potential of High-Molar-Activity Samarium-153 for Targeted Radionuclide Therapy with [153Sm]Sm-DOTA-TATE

Samarium-153 is a promising theranostic radionuclide, but low molar activities (Am) resulting from its current production route render it unsuitable for targeted radionuclide therapy (TRNT). Recent efforts combining neutron activation of 152Sm in the SCK CEN BR2 reactor with mass separation at CERN/MEDICIS yielded high-Am 153Sm. In this proof-of-concept study, we further evaluated the potential of high-Am 153Sm for TRNT by radiolabeling to DOTA-TATE, a well-established carrier molecule binding the somatostatin receptor 2 (SSTR2) that is highly expressed in gastroenteropancreatic neuroendocrine tumors. DOTA-TATE was labeled with 153Sm and remained stable up to 7 days in relevant media. The binding specificity and high internalization rate were validated on SSTR2-expressing CA20948 cells. In vitro biological evaluation showed that [153Sm]Sm-DOTA-TATE was able to reduce CA20948 cell viability and clonogenic potential in an activity-dependent manner. Biodistribution studies in healthy and CA20948 xenografted mice revealed that [153Sm]Sm-DOTA-TATE was rapidly cleared and profound tumor uptake and retention was observed whilst these were limited in normal tissues. This proof-of-concept study showed the potential of mass-separated 153Sm for TRNT and could open doors towards wider applications of mass separation in medical isotope production.


Introduction
Samarium-153 is a radioisotope that has been used in clinical practice for many years because of its favorable decay properties. 153 Sm has a half-life of 1.93 days and decays into a stable daughter nuclide ( 153 Eu). Upon its decay, several β-particles (E = 808 keV (18%), 705 keV (50%), 635 keV (32%)) and Auger-Meitner and conversion electrons are emitted which are responsible for the therapeutic potential of 153 Sm. The average β-energy for 153 Sm is 91 keV/particle higher than the current benchmark 177 Lu with almost six times more Auger-Meitner and conversion electrons per nuclear transformation (Table 1) [1]. In parallel, 153 Sm also co-emits γ-rays (103 keV, 28%) that allow single photon emission computed tomography (SPECT) imaging using clinically available collimators [2]. These properties put 153 Sm forward as a promising theranostic radionuclide for targeted radionuclide therapy (TRNT).

Quantification of Radioactivity in Biological Samples
Quantification of activity was performed with an automated gamma counter equipped with a 3-in NaI(Tl) detector coupled to a multichannel analyzer (2480 Wizard 2 , Perkin Elmer, Zaventem, Belgium). The results were corrected for background signal, physical decay during counting and detector dead time.

Cell Culture
Rat pancreatic cancer cell line CA20948 was used given the somatostatin receptor 2 (SSTR 2 ) overexpression as it is found in many human neuroendocrine tumors [9]. Cells were grown in DMEM High Glucose and Pyruvate medium (Gibco 41966029) supplemented with 10% fetal bovine serum (Gibco 10270106) and 100 U/mL penicillin-streptomycin (Sigma-Aldrich P4333) and were maintained in a humidified incubator at 37 • C with 5% CO 2 supply. All in vitro results were obtained from a minimum of three replicates per condition.

Mice
All animal experiments were performed in compliance with the Ethical Committee Animal Studies of Medanex Clinic (EC MxCl 2020-163 and EC MxCl 2021-173), the Belgian laboratory animal legislation and the European Communities Council Directive of 22 September 2010 (2010/63/EU). Male BALB/c and NMRI nude mice were purchased from Janvier (Bio Services, Uden, The Netherlands) and housed in ventilated cages under standard laboratory conditions (12 h light/dark cycle) at the animal facility of SCK CEN. All animals had access to food and water ad libitum.

Generation of Tumor Bearing Xenograft Mice
Male NMRI nude mice (between 8 and 15 weeks old) were anesthetized with 2.5% isoflurane in O 2 at a flow rate of 0.2 L min −1 and subcutaneously injected on the right shoulder with 1 × 10 6 CA20948 cells resuspended in PBS with 30% matrigel ® . Tumor dimensions were measured two or three times per week and volume was calculated as length × width × height. Mice were sacrificed if the tumor volume exceeded 600 mm 3 or if more than 10% of the original body weight was lost.

Samarium-153 Production
Samarium-153 was produced as described in our previous work [6]. In short, 153 Sm was produced by neutron activation of highly enriched 152 Sm in its nitrate form (enrichment grade 98.7%). About 350 µg 152 Sm was sealed in an ampoule and irradiated in the BR2 reactor for 2 or 3 days in thermal neutron fluxes of 2.0 to 2.5 × 10 14 neutrons/cm 2 /s. Up to three ampoules were irradiated in each production. After cooling, the target material was dissolved and shipped to CERN/MEDICIS for mass separation to yield high-A m 153 Sm. After mass separation, the 153 Sm was radiochemically purified to remove any 153 Eu and other impurities arising from the mass separation process (e.g., Zn from the implantation layer), resulting in high-purity [ 153 Sm]SmCl 3 (in 0.05 M HCl) ready for radiolabeling.

Radiolabeling
Radiolabeling was performed by adding up to 200 MBq of the 153 Sm radiolabeling solution to 20 nmol of the DOTA-TATE precursor dissolved in 0.15 M NaOAc buffer (pH 4.7) in a total reaction volume of 1 mL. The reaction vial was placed in a Thermomixer C (Eppendorf) to stir (500 rpm) and the mixture was heated to 95 • C for 15 min. After incubation, the radiolabeling yield was evaluated using thin-layer chromatography. An amount of 1-2 µL of the reaction mixture was spotted on a glass microfiber chromatography paper strip impregnated with silica gel (iTLC-SG, Agilent Technologies, Diegem, Belgium) that was eluted with a 0.5 M citrate solution (pH 5.5). After elution, the TLC papers were cut in half and the activity of the bottom and top part of the TLC paper was counted for 2 min each using an automated gamma counter.

Stability Test
The stability of [ 153 Sm]Sm-DOTA-TATE was evaluated by incubating the radiopharmaceutical in several relevant media. About 5 MBq of [ 153 Sm]Sm-DOTA-TATE (250 µL) was added to 4.75 mL of human serum or 4.75 mL of PBS and incubated at 37 • C up to 7 days. After 1 h, 3 days, 5 days and 7 days, a 5 µL sample was spotted on the TLC system described above to identify the intact fraction. To get an idea of the shelf life of [ 153 Sm]Sm-DOTA-TATE, the stability of the radiopharmaceutical product was evaluated in radiolabeling buffer at room temperature using the same protocol.

Cell Binding Assay
To study membrane association and internalization of [ 153 Sm]Sm-DOTA-TATE, 50,000 cells were seeded per well in a 24-well plate (26,316 cells/cm 2 ). After 24 h, cells were incubated with 2.5 nM [ 153 Sm]Sm-DOTA-TATE (2.7 kBq) for 1 h. Non-specific binding was assessed by co-incubation of [ 153 Sm]Sm-DOTA-TATE with a 4-fold and 4000-fold molar excess of unlabeled DOTA-TATE in parallel. Hereafter, fractions were collected. First, the membrane-bound fraction was collected following 10 min incubation at room temperature in strip buffer (50 mM glycine, 100 mM NaCl, pH 2.8). Then, the internalized fraction was collected by 30 min of cell lysis (1 M NaOH) at room temperature. The amount of [ 153 Sm]Sm-DOTA-TATE associated with the cell membrane and internalized was measured using an automated gamma counter. In addition, total cell number per well was obtained using a MOXI™ automated cell counter (VWR, Leuven, Belgium).

MTS Viability Assay
Cell viability after exposure to [ 153 Sm]Sm-DOTA-TATE was tested using the CellTiter 96 ® Aqueous One Solution Cell Proliferation Assay (Promega G3582). A total of 10,000 cells were seeded per well into a 96-well plate (31,250 cells/cm 2 ). After 24 h, cells were treated with [ 153 Sm]Sm-DOTA-TATE (0-10 MBq/mL) in fresh cell culture medium. After 4 h incubation at 37 • C in 5% CO 2 , cells were washed and new medium was added. Then, after 4 days, MTS reagent was prepared according to the manufacturer's instructions and added to the cells. Absorbance was measured at 490 nm using the BioTek Synergy H1 reader (Agilent). Cell viability was obtained as the ratio of the absorbance of a sample well to the average absorbance of untreated (0 MBq/mL) wells, after subtraction of the background signal of wells containing MTS reagent without cells. Statistical significance (p < 0.05) was determined using linear models.

Clonogenic Survival Assay
Cell suspensions (0.1 × 10 6 cells) were prepared using 0.25% trypsin/EDTA, resuspended in cell culture medium and treated for 4 h with radiolabeled [ 153 Sm]Sm-DOTA-TATE (0-10 MBq/mL). An aliquot of 300 cells for each treatment condition was plated in 6-well plates with 2 mL of growth medium and incubated at 37 • C in 5% CO 2 . After 14 days, colonies were fixed and stained with 6% glutaraldehyde and 0.5% crystal violet. The number of colonies (blindly coded) was counted to determine the clonogenic survival fraction per treatment condition. Importantly, colonies were only considered if they contained at least 50 cells. Statistical significance (p < 0.05) was determined using linear models.

Biodistribution
The pharmacokinetic behavior of [ 153 Sm]Sm-DOTA-TATE was evaluated by performing biodistribution studies in healthy BALB/c mice as well as in CA20948 tumor bearing xenograft mice [9]. Additionally, similar biodistribution studies were performed with free (unlabeled) [ 153 Sm]SmCl 3 to identify the organs with 153 Sm accumulation in case of release from the complex. All animals were anesthetized with 2.5% isoflurane in O 2 at a flow rate of 1 L min −1 and injected with ∼2-3 MBq of [ 153 Sm]SmCl 3 or [ 153 Sm]Sm-DOTA-TATE via a tail or penile vein. The mice were sacrificed by an overdose pentobarbital (200 µL of 60 mg/mL) 1, 4, 24 and 72 h post-injection (p.i.) (n = 3-4 per time point). Blood was collected via cardiac puncture. Organs of interest were collected in tared tubes and weighed. The activity in the different organs was measured in the gamma counter. For the calculation of the total activity in the blood, bone and muscle, the masses were assumed to be, respectively, 7, 12 and 40% of the total body mass [10,11]. Data were expressed as a percentage of injected activity (%IA) and standardized uptake value (SUV). %IA was calculated as (counts per min (cpm) in organ)/(IA converted to cpm) × 100. SUV was calculated as (activity in cpm in organ/weight of organ in g)/(IA/total body weight in g).

Dosimetry
The mean absorbed doses per amount of injected activity (mGy/mBq) to the kidneys and tumor were calculated based on the MIRD formalism. The time-activity curve (TAC) was created based on the fraction of injected activity per gram of dissected tissue (FIA/g) as a function of time. The FIA was decay-corrected to the time of sacrifice of the mouse. For both tissues, it was assumed that FIA/g = 0 at t = 0 h (injection time) and to increase linearly thereafter until t = 1 h (earliest time point). The time-activity curves for kidneys and tumor were fitted to bi-exponential curves (R 2 > 0.99 and R 2 = 0.97, respectively), which were then integrated from zero to infinity. The time-integrated activity coefficient (ã) was obtained by multiplying the mean tissue mass (i.e., m = 0.42 g for tumor and m = 0.26 g for kidney) by the respective time-integrated activity.
The mean absorbed dose for tumor and kidney was obtained by multiplying ã and the S values (0.0962 and 0.328 mGy/MBq.s for tumor and kidney, respectively). The S values were calculated for 153 Sm radiation decay data using the Monte Carlo code MCNP6.2 (Los Alamos National Laboratory, Los Alamos, NM, USA). Spectrum data from 153 Sm were modelled according to the decay data reported in ICRP Publication 107 [8]. A water sphere of 0.42 g (for the tumor) and a realistic model of a mouse kidney [12] were used to model the tissue geometry in the Monte Carlo simulations. Additionally, the S value for kidneys was scaled to the mean measured kidney tissue mass.

SPECT Imaging
To evaluate the imaging possibilities of [ 153 Sm]Sm-DOTA-TATE, one CA20948 tumor xenograft mouse was subjected to a SPECT scan. The mouse was anaesthetized using isoflurane (2.5% in O 2 , flow rate 1 L/min) and injected with about 20 MBq of [ 153 Sm]Sm-DOTA-TATE via the tail vein. At 4 h and 24 h after injection, the mouse was scanned for approximately 2 h, followed by a 3 min CT scan. The in vivo imaging was performed in a dedicated small animal SPECT-CT (U-SPECT6-CT, Milabs, Utrecht, The Netherlands). The scans were acquired in list-mode using a general-purpose mouse collimator with 0.6 mm diameter pinholes.
The images were reconstructed and co-registered using the Milabs provided manufacturer software. SPECT images were reconstructed using a 20% energy window centered on the 103 keV photopeak. Scatter correction was applied using the triple energy window method with two adjacent background-scatter windows of 4.2% width. SPECT image reconstruction was performed using the similarity-regulated ordered-subsets expectation maximization (SROSEM) algorithm [13], with four iterations, 128 subsets and a voxel size of 0.4 mm 3 .
SPECT images were registered to the CT images, for a final voxel size of 0.16 mm 3 , and corrected for photon attenuation based on the CT data. Data post-processing and analysis was performed using PMOD 4.101 software (PMOD, Zürich, Switzerland).
SPECT data in reconstructed counts per seconds (cps) units were calibrated to activity units (MBq/cc) using a calibration factor (CF). The CF was calculated from a SPECT scan of a syringe filled with a known activity concentration of [ 153 Sm]SmCl 3 . The activity concentration was traceable to high-resolution gamma spectrometry measurements performed on-site.

Radiolabeling and Stability
Radiolabeling of [ 153 Sm]Sm-DOTA-TATE at 95 • C was highly efficient, resulting in high radiochemical conversion rates (>99%) with an A m of up to 10 MBq/nmol. The shelf life of [ 153 Sm]Sm-DOTA-TATE was tested by storing the radiopharmaceutical at room temperature in its radiolabeling buffer. Up to 7 days after incubation in radiolabeling buffer,

Radiolabeling and Stability
Radiolabeling of [ 153 Sm]Sm-DOTA-TATE at 95 °C was highly efficient, resulting in high radiochemical conversion rates (>99%) with an Am of up to 10 MBq/nmol. The shelf life of [ 153 Sm]Sm-DOTA-TATE was tested by storing the radiopharmaceutical at room temperature in its radiolabeling buffer. Up to 7 days after incubation in radiolabeling buffer, no degradation of the [ 153 Sm]Sm-DOTA-TATE complex was observed (radiochemical purity > 98%) ( Figure 1). Additionally, incubating [ 153 Sm]Sm-DOTA-TATE in biologically relevant media like PBS and human serum at 37 °C did not result in any dissociation of 153 Sm from the DOTA-complex (radiochemical purity > 99%).

In Vitro Specificity and Cellular Internalization of [ 153 Sm]Sm-DOTA-TATE
The specific binding of [ 153 Sm]Sm-DOTA-TATE to SSTR 2 receptors was evaluated in CA20948 cells. After 1 h of incubation, 0.3% of the added activity was associated with the CA20948 cells, of which 79 ± 2% was internalized ( Figure 2). Co-incubation of, respectively, a 4-and a 4000-fold excess of unlabeled DOTA-TATE drastically decreased the binding, with about 40% and 92%, proving the binding specificity.

In Vitro Specificity and Cellular Internalization of [ 153 Sm]Sm-DOTA-TATE
The specific binding of [ 153 Sm]Sm-DOTA-TATE to SSTR2 receptors was evaluated in CA20948 cells. After 1 h of incubation, 0.3% of the added activity was associated with the CA20948 cells, of which 79 ± 2% was internalized ( Figure 2). Co-incubation of, respectively, a 4-and a 4000-fold excess of unlabeled DOTA-TATE drastically decreased the binding, with about 40% and 92%, proving the binding specificity.

In Vitro Biological Evaluation of [ 153 Sm]Sm-DOTA-TATE
Cell viability was assessed after exposure to increasing concentrations of [ 153 Sm]Sm-DOTA-TATE. MTS assays revealed that [ 153 Sm]Sm-DOTA-TATE treatment significantly reduced cell viability down to 48 ± 6% in an activity concentration-dependent manner when compared to untreated cells ( Figure 3). Notably, treatment with [ 153 Sm]Sm-DTPA also significantly lowered cell viability down to 45 ± 5% in an activity concentration-dependent manner when compared to untreated cells ( Figure 3). However, no effect on cell viability (85 ± 10%) was observed with the lowest tested activity concentration (

Biodistribution Studies
The biodistribution of [ 153 Sm]Sm-DOTA-TATE in healthy BALB/c mice showed rapid excretion of radioactivity from the body (Supplementary Figure S1 and Table S1). After 1 h, already 82.0 ± 4.8% of the activity was cleared from the body. After 72 h, the excreted fraction increased to 97.8 ± 0.2 %IA. The initial uptake of [ 153 Sm]Sm-DOTA-TATE in kidneys (SUV 1h = 2.1 ± 0.2; SUV 4h = 1.8 ± 0.1) and a limited liver and intestinal uptake suggest renal clearance ( Figure 5 and Supplementary Table S2). Apart from some initial kidney uptake, only slight retention of radioactivity was observed in spleen.

Biodistribution Studies
The biodistribution of [ 153 Sm]Sm-DOTA-TATE in healthy BALB/c mice showed rap excretion of radioactivity from the body (Supplementary Figure S1 and Supplementa Table S1). After 1 h, already 82.0 ± 4.8% of the activity was cleared from the body. Aft 72 h, the excreted fraction increased to 97.8 ± 0.2 %IA. The initial uptake of [ 153 Sm]Sm DOTA-TATE in kidneys (SUV1h = 2.1 ± 0.2; SUV4h = 1.8 ± 0.1) and a limited liver an intestinal uptake suggest renal clearance ( Figure 5 and Supplementary Table S2). Apa from some initial kidney uptake, only slight retention of radioactivity was observed spleen.  Unlabeled [ 153 Sm]SmCl 3 , on the other hand, showed high accumulation in liver and bone tissue (SUV 72h = 7.7 ± 1.8 and 3.1 ± 1.3 for liver and bone, respectively) (Supplementary Figures S2 and S3 and Supplementary Tables S3 and S4). This contrasts with the limited bone uptake observed in [ 153 Sm]Sm-DOTA-TATE-injected animals (SUV 72h = 0.1 ± 0.0, Figure 5 and Supplementary Table S2).
In order to evaluate the pharmacokinetic profile in tumor bearing mice, [ 153 Sm] Sm-DOTA-TATE was injected in CA20948 xenograft mice. Again, rapid excretion was observed with about 80 %IA already excreted after 1 h p.i. (Supplementary Figure S4 and Table S5). At the initial time points, mild uptake was observed in kidneys (SUV 1h = 1.8 ± 0.1) and pancreas (SUV 1h = 1.2 ± 0.2), which rapidly washed out from these organs ( Figure 6 and Table 2). Substantial tumor uptake was observed 1 h p.i. (SUV 1h = 4.5), which only slowly decreased over the different time points (SUV 72h = 1.7). The limited kidney SUV uptake and good retention in the tumor resulted in high tumor-to-kidney SUV ratios ( Table 2). The lack of uptake in bone tissue (SUV 72h = 0.0 ± 0.0) again confirms the in vivo stability of the [ 153 Sm]Sm-DOTA-TATE complex.

Dosimetry
The

Dosimetry
The absorbed dose estimated for a spherical tumor (m = 0.42 g) was 30.7 mGy/MBq·s. This results in 6 mGy/s dose to the tumor for a mouse injected with 2 MBq of [ 153 Sm]Sm-DOTA-TATE. For the kidneys, the absorbed dose was 3.35 mGy/MBq·s, which results in 6.7 mGy/s dose to the kidneys for a mouse injected with 2 MBq of [ 153 Sm]Sm-DOTA-TATE.

SPECT-CT Imaging
Small-animal SPECT-CT scans were acquired at 4 and 24 h p.i. of~20 MBq of [ 153 Sm]Sm-DOTA-TATE in CA20948 tumor-bearing mice resulting in the images shown in Figure 7. The images are displayed as maximum intensity projection (MIP) and are expressed in units of MBq/cc. In the figure, the tumor (Tu) can be observed in the right shoulder of the mouse. Other organs visible in the image are kidneys (Ki), liver (Li) and bladder (Bl).  [ 153 Sm]Sm-DOTA-TATE showed a time-dependent tumor uptake of 1.01 MBq/cc and 0.56 MBq/cc at 4 and 24 h p.i., respectively. Rapid renal clearance is observed, with the bladder showing a high accumulation of activity at 4 h p.i. which has mostly cleared at 24 h p.i.. Other organs observed in the SPECT-CT images were kidneys and liver, the latter with limited retention of activity. Additionally, no bone uptake was observed. The findings on the in vivo imaging are in line with the biodistribution data.

Discussion
Based on its favorable decay characteristics, 153 Sm can be a valuable radionuclide for targeted radionuclide therapy (TRNT). However, due to the current standard production route, only low-A m 153 Sm can be obtained, rendering it unsuitable for TRNT. This changed when the Belgian Nuclear Research Centre and the MEDICIS research branch of CERN joined forces to produce high-A m 153 Sm. In a previous study, we successfully used mass separation in combination with laser resonance ionization to extract 153 Sm from its irradiation matrix resulting in non-carrier-added, high-A m 153 Sm [6]. Initial radiolabeling studies showed that 153 Sm could be incorporated in p-SCN-Bn-DOTA with high yields. The applicability of the produced high-A m 153 Sm in TRNT was yet to be proven. Therefore, this proof-of-concept study aimed to investigate the potential of 153 Sm for TRNT. DOTA-TATE was chosen as the preferred vector for the evaluation of the potential of 153 Sm in TRNT. The FDA-approved [ 177 Lu]Lu-DOTA-TATE (Luthatera ® ) has a proven effect on the progression-free survival of patients suffering from GEP-NETs [7]. With the development of the high-A m production method, 153 Sm could be added to the group of radionuclides used in TRNT. In a clinical study comparing 153 Sm-EDTMP and 177 Lu-EDTMP for bone pain palliation, it was observed that both radionuclides were equally effective and, more importantly, could be used interchangeably [14]. Compared to 177 Lu, 153 Sm could lead to improved therapeutic efficacy for TRNT of small and medium-to-large tumors. This is attributed to the emission of higher-energy β − particles and a significant amount of Auger-Meitner and conversion electrons (Table 1) [1]. This high Auger-Meitner decay, combined with a shorter half-life and emission of gamma-rays, makes 153 Sm a versatile radioisotope. Efforts are currently being made to scale up the production of high-A m 153 Sm. We successfully radiolabeled [ 153 Sm]Sm-DOTA-TATE at fair A m that showed no leaching of 153 Sm from the DOTA complex for at least 7 days. This was in line with earlier data showing that DOTA forms a very stable complex with several lanthanides, including 153 Sm [15]. The stability observed in vitro was also confirmed in our in vivo results. Our data showed that 153 SmCl 3 highly accumulated in bone and liver. When injecting [ 153 Sm]Sm-DOTA-TATE, these organs did not show any uptake in the SPECT images and biodistribution studies. Therefore, we can safely assume that the complex between 153 Sm and DOTA-TATE is highly stable both in vitro and in vivo.
Our cell binding experiments showed specific binding to the SSTR 2 -expressing CA20948 cells. Both CA20948 and SSTR 2 -transfected U2OS cells are routinely used in preclinical studies concerning cancers overexpressing SSTR 2 [16][17][18]. CA20948 cells are derived from rat pancreas and express SSTR 2 endogenously [18]. These cells can be easily used to generate xenografts, whereas this is not the case for SSTR 2 -transfected U2OS cells. Clear and specific binding of [ 153 Sm]Sm-DOTA-TATE to the CA20948 cells was observed, of which the largest part was internalized. These data are in line with Dalm et al. [19]. The absolute uptake in our study was lower compared to the reported literature in the same cell line, which could be attributed to the lower A m used in our experiments. In parallel, the comparison of uptake between [ 177 Lu]Lu-DOTA-TATE and lower A m [ 213 Bi]Bi-DOTA-TATE also yielded roughly a 100-fold-lower uptake of the low-A m compound [20].
An activity-dependent effect of [ 153 Sm]Sm-DOTA-TATE was observed on CA20948 cell viability and clonogenic potential. In parallel, we performed the same experiment with [ 153 Sm]Sm-DTPA at the same radioactive concentrations, as a surrogate for radioactive effects originating from the medium. DTPA is commonly used since it easily complexes lanthanides and is not expected to show any affinity towards CA20948 cells. Although the uptake of [ 153 Sm]Sm-DTPA in the CA20948 cells was not studied, the literature results testing the binding of [ 177 Lu]Lu-DTPA showed a complete lack of affinity towards SSTR 2 in CA20948 cells [20,21]. As such, we assumed that, upon changing the radionuclide, [ 153 Sm]Sm-DTPA would also not be internalized by CA20948 cells. In the study by Chan et al. [20], the clonogenic potential of CA20948 It is well known that the A m plays an important role in the efficacy of TRNT. Performing experiments with lower-A m radiopharmaceuticals will have consequences for biological evaluation, which is illustrated by the MTS and CFU assays. The low A m in our experiments is, however, caused by the current production method of 153 Sm. Although the A m of the produced 153 Sm is rather high, it is currently only produced in relatively low amounts, making it hard to generate sufficient amounts of the radiopharmaceutical. Therefore, the [ 153 Sm]Sm-DOTA-TATE in our studies is only obtained in fairly low A m (2-10 MBq/nmol) compared to, respectively, clinical (± 40 MBq/nmol) [17] and preclinical in vitro A m (±50 MBq/nmol) [19] of [ 177 Lu]Lu-DOTA-TATE. Efforts are currently underway to increase production efficiency and to ramp up production capacity. This will allow larger scale productions of 153 Sm, enabling higher-A m radiolabelings. Since we already see distinct cell binding and therapeutic effect with [ 153 Sm]Sm-DOTA-TATE at our current low A m , this effect is only expected to be greater when higher A m can be reached.
Despite the relatively low A m , our in vivo evaluation showed some promising results. In the biodistribution study, a clear and prolonged tumor uptake was observed with a rapid renal excretion, the main excretion pathway of [ 153 Sm]Sm-DOTA-TATE. At later time points, a gradual wash out of activity was observed from the tumor tissue. This was comparable to data obtained in a similar animal model [18]. In vivo SPECT-CT images were successfully acquired with the A m achieved and they confirmed the data seen in the biodistribution study at 4 and 24 h p.i. The tumor-to-kidney SUV ratios obtained in our biodistribution studies were remarkably higher compared to [ 177 Lu]Lu-DOTA-TATE in a similar animal model [18]. Other organs which express the SSTR 2 , such as the spleen, show a similar uptake of [ 153 Sm]Sm-DOTA-TATE compared to [ 177 Lu]Lu-DOTA-TATE, including initial high pancreas uptake that was rapidly washed out [18]. Tumor-to-kidney ratios were compared in a AR42J (endogenous expression of SSTR 2 ) xenografted mouse model injected with 55 Co, 64 Cu and 68 Ga-labeled DOTA-TATE. [ 55 Co]Co-DOTA-TATE had, respectively, a 4-and 8-fold-higher tumor-to-kidney ratio compared to 64 Cu and 68 Ga at 24 h p.i., although the initial absolute uptake did not significantly differ. Further preclinical studies on mice will be of interest to investigate the microscopic distribution over time of [ 153 Sm]Sm-DOTA-TATE, the resulting absorbed doses within tissues and the associated radiobiological effects (end points of tumor control and toxicity). This will allow assessment of the potential of TRNT with [ 153 Sm]Sm-DOTA-TATE in relation to the clinically used [ 177 Lu]Lu-DOTA-TATE.
Interestingly, in healthy mice we observed a noticeable retention of [ 153 Sm]Sm-DOTA-TATE by the liver and spleen. The expression of SSTRs is highly homologous between mice and humans, and moderate expression of SSTR 2 is seen in the red pulp of the spleen and has been reported as a site of off-target specific binding [22,23]. SSTR 2 is expressed in liver to a lesser extent [24] and uptake can be attributed to non-specific binding (metabolic enzymes) or blood circulation and clearance. Uptake in these organs is reduced in tumorbearing animals. This can be attributed to a 'sink' effect influenced by overexpression of SSTR2 in the tumor [22]. This increase in local, high-affinity receptor density will withdraw [ 153 Sm]Sm-DOTA-TATE from circulation. Hence, the tumor-to-organ ratio will increase as normally the expression of SSTR 2 in non-tumor organs remains unchanged. The lower A m of [ 153 Sm]Sm-DOTA-TATE can be beneficial in tumor targeting as the fraction cleared by the first pass effect of the liver can be reduced. Part of the high-capacity low-affinity receptors or metabolic enzymes in the liver can be occupied by unlabeled DOTA-TATE (in excess vs. [ 153 Sm]Sm-DOTA-TATE), which leads to a higher probability of [ 153 Sm]Sm-DOTA-TATE binding to tumor sites [25]. Similar effects can be expected in the kidneys, partially explaining the much higher kidney-to-tissue SUV ratios observed in our experiments. In the earlier time points, pronounced uptake in pancreas was observed in both healthy and tumor-bearing animals, in line with endogenous expression of the SSTR 2 in this organ [26].
In conclusion, we successfully evaluated, for the first time, high-A m 153 Sm for TRNT using DOTA-TATE as a vector molecule in vitro and in vivo. Although we radiolabeled at limited A m , a clear uptake in tumor cells and a mouse xenograft model clearly indicates the potential of [ 153 Sm]Sm-DOTA-TATE in TRNT. This proof-of-concept study using massseparated isotopes could open doors towards wider application of mass separation in medical isotope production.  Table S4: Biodistribution data expressed as standard uptake value (SUV) of [ 153 Sm]SmCl 3 in male BALB/c mice. Values represent mean ± standard deviation (n = 3-4); Figure S4: Ex vivo biodistribution over time expressed as percentage of injected activity (%IA) of [ 153 Sm]Sm-DOTA-TATE in CA20948 xenograft mice. Values represent mean ± standard deviation (n = 3-4); Table S5: Biodistribution data expressed as percentage of injected activity (%IA) of [153Sm]Sm-DOTA-TATE in CA20498 tumor bearing male NMRI-nude mice. Values represent mean ± standard deviation (n = 3-4); Table S6: Biodistribution data expressed as standard uptake value (SUV) of [ 153 Sm]Sm-DOTA-TATE in CA20498 tumor bearing male NMRI-nude mice. Values represent mean ± standard deviation (n = 3-4).