High yield cyclotron production of a novel 133/135La theranostic pair for nuclear medicine

This study reports the high-yield production of a novel 133/135La theranostic pair at a 22 MeV proton beam energy as an attractive alternative to the recently introduced 132/135La pair, demonstrating over an order of magnitude production increase of 133/135La (231 ± 8 MBq 133La and 166 ± 5 MBq 135La at End of Bombardment (EOB)) compared to 11.9 MeV production of 132/135La (0.82 ± 0.06 MBq 132La and 19.0 ± 1.2 MBq 135La) for 500 µA·min irradiations. A new sealed solid cyclotron target is introduced, which is fast to assemble, easy to handle, storable, and contains reusable components. Radiolabeling with macrocyclic chelators DOTA and macropa achieved full incorporation, with respective apparent 133La molar activites of 33 ± 5 GBq/µmol and 30 ± 4 GBq/µmol. PET centers with access to a 22 MeV capable cyclotron could produce clinically-relevant doses of 133/135La, via natBa irradiation, as a standalone theranostic agent for PET imaging and Auger electron therapy. With lower positron energies and less energetic and abundant gamma rays than 68Ga, 44Sc and 132La, 133La appears to be an attractive radiometal candidate for PET applications requiring a higher scanning resolution, a relatively long isotopic half-life, ease of handling, and a low patient dose.

Theranostics in nuclear medicine is a technique whereby a site specific pharmaceutical is radiolabelled first with a radioisotope for diagnostic imaging. After analysis, the same pharmaceutical is labelled with a particle emitting radioisotope for therapeutic application 1 . The complementary isotopes used are called theranostic pairs. It is essential that the two isotopes have very similar chemical properties with the ideal case being that they are different isotopes of the same element. Auger electron-emitting isotopes have potential as a high linear energy transfer (LET) therapeutic agent to destroy cancer cells by depositing their ionizing emission energy over a very short path length, damaging DNA by inducing various types of DNA damage, including double-strand breaks. This holds advantages over lower LET therapy such as βtherapy where emissions can travel over 1 cm, and may unnecessarily irradiate healthy tissue 2,3 . High LET Auger electron emissions have achieved encouraging clinical results, with 111 In-DTPA-octreotide and 125 I-IUdR causing tumor remissions in patients with lower normal tissue toxicity, and improvements in the survival of glioblastoma patients using 125 I-mAb 425 with minimal normal tissue toxicity 4 . A recently developed theranostic pair is 132/135 La, where positron emissions from 132 La are used for PET imaging while the Auger electrons from 135 La have the potential for use in Auger electron therapy (AET) [5][6][7] . Theranostic La pairs are not only inherently useful but also can serve as surrogates for potential future study relating to 225 Ac alpha-particle therapy. 225 Ac-labeled compounds have seen significant recent clinical successes in treating aggressive tumor metastases 2,8 .
However, 132/135 La has limitations for PET imaging due to its fundamental positron and gamma emission properties, and current cyclotron production methods. The average and maximum 132 La positron energies of 1.29 MeV and 3.67 MeV are significantly higher than those of other commonly used PET isotopes such as 18 F (E mean = 0.250 MeV, E max = 0.634 MeV), 64 Cu (E mean = 0.278 MeV, E max = 0.653 MeV), 68 Ga (E mean = 0.829 MeV, E max = 1.90 MeV), or 44 Sc (E mean = 0.632 MeV, E max = 1.47 MeV) 9 . The higher positron energy of 132 La implies reduced PET image spatial resolution for tumor imaging, especially when imaging smaller tumors and metastases. Furthermore, 132 La emits high abundance gamma rays within typical 511 keV PET scanner energy windows that can contribute to spurious coincidences, as well as high energy gamma rays that may complicate handling.
Using nat Ba target material, current 132 La cyclotron production methods via 132 Ba(p,n) 132 La require long irradiation times and generate reduced activity due to the very low natural abundance of 132 Ba (0.1%).
The present work describes high yield 133/135 La production through 22 MeV proton irradiation of nat Ba metal encapsulated within a convenient sealed cyclotron target. Irradiating nat Ba at 22 MeV generates much higher Instrumentation. Sample activity was measured using an Atomlab 500 Dose Calibrator (Biodex, Shirley, NY, USA). Radionuclidic purity was assessed using a GEM35P4-70-SMP high-purity germanium detector (ORTEC, Oak Ridge, TN, USA) with ORTEC GammaVision software. Elemental purity was assessed using a 720 Series ICP-OES (Agilent Technologies, Santa Clara, CA, USA). A NEPTIS Mosaic-LC synthesis unit (Optimized Radiochemical Applications, Belgium) was used to separate and purify the 133/135 La from the dissolved Ba target solution. An AR-2000 Radio-TLC Imaging Scanner (Eckert & Ziegler, Hopkinton, MA, USA) was employed to quantify the fraction of chelator-bound 133/135 La after the reaction. The solid targets were manufactured using a Model 6318 hydraulic press (Carver, Wabash, IN, USA), and the nat Ba metal was pressed inside a 10 mm (I.D.) EQ-Die-10D-B hardened steel die (MTI Corporation, Richmond, CA, USA). A S90013A optical light microscope (Fisher Scientific, Waltham, MA, USA) was employed to inspect the seal integrity of each sealed solid target after manufacturing.
Cyclotron targetry and irradiation. A completed sealed cyclotron target is depicted in Fig. 1. Cyclotron targets were prepared from 200 mg of nat Ba metal, an Ag disc (24 mm diameter, 1.5 mm thick) cut from an Ag rod, In wire (1 mm diameter), and Al foil (25 µm thick). A 10 mm diameter depression was machined into the center of each disc to a 100 µm depth, and a 1 mm wide annulus with an inner diameter of 15 mm was machined www.nature.com/scientificreports/ to a depth of 100 µm. Using a method similar to the target production described in 10 , nat Ba metal was quickly loaded into a hardened stainless steel die to minimize exposure to the atmosphere, and a force of 15 kN was applied using a hydraulic press, producing a 10 mm diameter pellet with a thickness of 0.8 mm. Pellets were produced in large quantities (> 10/batch) and removed quickly from the die and sealed in a vial with an argon atmosphere to prevent oxidation during storage. A 23 mm diameter Al foil cover was cut out with a flap extension to facilitate post-irradiation removal by peeling. Individual pellets were then placed in the central Ag disc depression and pressed at a force of 20 kN on the hydraulic press to secure the pellets in the depression. 5.5 cm of In wire was then laid into the annulus depression with 1 mm of overlap at the ends, the target assembly was quickly covered by the Al cover, and a force of 25 kN was applied using the hydraulic press to compress the In wire to form an air-tight bond between the Ag disc and Al cover. Following pressing, the target was observed under an optical light microscope to confirm target seal integrity, verifying there were no pinholes present in the Al cover. The target was stored under regular atmospheric conditions ready for on-demand irradiation.
Targets were irradiated at 22 MeV using a 24 MeV TR-24 cyclotron (Advanced Cyclotron Systems Inc., Richmond B.C., Canada) for 25-200 min with a maximum proton beam current of 20 µA at current densities of 25.5 µA/cm 2 . A pneumatically actuated TA-1186 solid target assembly (Advanced Cyclotron Systems Inc., Richmond B.C., Canada) was used with the target disc perpendicular to the proton beam. O-rings within the assembly provided a helium gas seal on the front and water seal on the back for both cooling streams. The Ag target was designed to be at least 0.6 mm thick behind the 0.8 mm thick nat Ba pellet so that the exit beam energy leaving the Ag disc was degraded below 6 MeV, as simulated by SRIM 2013 11 . This design consideration was to avoid the production of 13 N (t 1/2 = 9.97 min) in the cyclotron cooling water circuit via the 16 O(p,α) 13 N reaction. A 250 µm thick Ag degrader was added to the cyclotron beamline after the Al vacuum foil so that extracting the cyclotron beam at 17 MeV resulted in the target incident energy being degraded to 11.9 MeV. These irradiations at 11.9 MeV served to provide a comparison to the 132/135 La isotope production introduced by Aluicio-Sarduy et al. 5 .
After allowing 1-2 h post-irradiation for decay of short-lived La isotopes, the target assembly was opened pneumatically, and the sealed target slid down a plastic guide tube into a lead shield. The lead shield was brought to a dose calibrator where its activity was measured, followed by placement into a lead castle containing a NEPTIS automated separation unit. At a 22 MeV target incident beam energy, the simulation suggests significant 135 La and 133 La cross sections for the 137 Ba(p,3n) 135 La, 136 Ba(p,2n) 135 La, 135 Ba(p,3n) 133 La, and 134 Ba(p,2n) 133 La reactions. The 132 Ba(p,n) 132 La cross-section is over two orders of magnitude lower at 22 MeV compared to at 11.9 MeV, and the 134 Ba(p,3n) 132 La reaction cross-section does not begin until just above 22 MeV. Irradiating nat Ba at 22 MeV should therefore maximize the production of 133 La and 135 La, bypass the majority of 132 La production from the 132 Ba(p,n) 132 La reaction, and just avoid the onset of the significant 134 Ba(p,3n) 132 La reaction. Due to the higher natural abundances of 134 Ba (2.42%) and 135 Ba (7.59%) compared to 132 Ba (0.10%), 133 La production potential is much greater compared to 132 La, illustrated in the difference between the absolute and isotopically weighted cross-sections shown in Figs. 2 and 3, respectively.
To compare 133/135 La to 132/135 La in this study, irradiations were performed with a target incident beam energy of 11.9 MeV. Figure 2 suggests irradiations at 11.9 MeV would result in the production of 135 La and 132 La via the Automated separation of 133/135 La. 133/135 La was separated using modified aspects of a method described by Aluicio-Sarudy et al. 5 . The reactor vessel within its shield was transferred into the lead castle, the sealed target was opened by peeling back the Al cover, and a suction line was attached. The reactor vessel was filled with 10 mL of 18 MΩ·cm water, dissolving the nat Ba target material in 5 min. The Ag target disc was removed, and 10 mL of 6 N HNO 3 was added to the reactor to bring the overall concentration to 3 N HNO 3 . 3 N HNO 3 was selected to reduce possible degradative effects of concentrated 6 N HNO 3 on the branched DGA resin. The target solution was withdrawn from the reactor and passed through two Acrodisc 32 mm diameter filters with 5 µm Supor membranes in parallel to capture any solid material such as nat Ba salts and oxides resulting from the dissolution stage. Following filtration, the target solution was passed through a SPE cartridge containing 0.25 g of branched DGA resin, and washed with 50 mL of 3 N HNO 3 to remove residual Ba and other metal impurities, followed by 5 mL of 0.5 N HNO 3 . [ 133/135 La]LaCl 3 was eluted using 1 mL of 0.1 N HCl. Following a decay period of 5 days (to permit the decay of the short-lived 107 Cd and longer-lived 106m Ag) the Ag disc was removed and cleaned in reagent grade 10 N HCl for reuse. For the comparative aspects, 132/135 La was separated using the same process.
Activity measurement and radionuclidic purity analysis. After separating the [ 133/135 La]LaCl 3 product, its radionuclidic purity was determined by gamma-ray spectroscopy using a high purity germanium (HPGe) detector. Calibrations for efficiency and energy were performed using NIST traceable Eckert & Ziegler Isotope Products Inc. γ-ray sources. Activities of La isotopes of interest were quantified using the efficiency-corrected HPGe measurements.

Results
Cyclotron targetry. Prior to longer irradiations, initial tests were performed with nat Ba targets at beam currents ranging from 1-20 µA to investigate target properties and durability. After irradiation and automated separation, HPGe analysis was performed on the Ag targets. For 11.9 MeV runs, analysis indicated small activities of 107 Cd (t 1/2 = 6.5 h) and 109 Cd (t 1/2 = 461.4 d) were produced via the 107 Ag(p,n) 107 Cd and 109 Ag(p,n) 109 Cd reactions. For 22 MeV runs, following the 3-h decay period, analysis indicated small activities of 107 Cd, 109 Cd, and 106m Ag (t 1/2 = 8.28 d). For both beam energies in this study, the targets did not activate significantly, and the majority of the activity present was 107 Cd and 106m Ag, which decayed significantly after several days. Following a 5-day decay period the targets were deemed acceptable for handling and reuse after placing the target in 10 N HCl to clean its surface. For all irradiations, none of the sealed Ag targets showed signs of physical degradation, with multiple target discs being reused upwards of 10 times.  Table 1, and several ratios of La isotopes of interest are given as a function of time after EOB in Table 2. At 22 MeV, 500 µA·min runs (n = 3) yielded 231 ± 8 MBq 133 La, and 166 ± 5 MBq 135 La. Saturated yields were 161 ± 5.5 MBq/µA for 133 La, and 561 ± 17 MBq/µA for 135 La. Significant amounts of 134 La and 136 La were present at EOB (1191 ± 96 MBq and 3914 ± 384 MBq, respectively), however owing to their short half-lives (6.45 min and 9.87 min, respectively), they decayed to negligible levels after 3-h post-EOB. Short-lived 130 La (8.7 min halflife) was observed and undetectable after the 3-h decay period. 132 La was produced (0.38 ± 0.03 MBq at EOB), indicating its production reactions were largely bypassed. Co-production of 131 La was observed (19.0 ± 1.2 MBq at EOB), however owing to its relatively short half-life (59.2 min), it decayed significantly during the 3-h decay period. TENDL 2019 cross sections indicated production of long-lived 137 La and 138 La, however, this was not quantified due to their extremely long half-lives.
For the comparison 132/135 La production runs at 11.9 MeV, 500 µA·min runs (n = 3) yielded 0.82 ± 0.06 MBq 132 La and 17.9 ± 0.8 MBq 135 La at EOB. Saturated yields were 0.70 ± 0.03 MBq/µA for 132 La, and 60.6 ± 2.8 MBq/  , which decayed to undetectable levels after the 3-h decay period. Cross-sections generated by TENDL 2019 indicated the production of long-lived 137 La and 138 La. However, production was also not quantified owing to their long half-lives. As shown in Table 2, the activity ratio of 135 La to 133 La at 22 MeV is much lower than the ratio of 135 La to 132 La at 11.9 MeV, resulting in a much greater PET imaging potential for a given total activity. At 22 MeV, the activity ratio of 133 La to 132 La remains large throughout the time intervals, suggesting that the production of the 132 La impurity was minimized.
Automated separation of 133/135 La. To determine dissolution time, several Ba targets were dissolved in the reactor with 10 mL of water, with the time required to completely dissolve the target ranging from 4 to 5 min. A dissolution time of 5 min was selected for production run separations to provide a sufficient time margin. The DGA resin was preconditioned with 3 N HNO 3 so the NEPTIS unit was prepared to receive the activity. The final product elution in 1 mL of 0.1 N HCl was calibrated to capture the maximum 133/135 La activity while avoiding excess dilution of the solution.
From the start of NEPTIS separation to the completion of product elution took ~ 35 min. Over 88% of decaycorrected 133/135 La activity was consistently recovered from the automated synthesis. Residual decay activities were 3% of the total in the branched DGA resin, 3% in the dissolution reactor, 2% in the two reactor filters, with the remainder (≤ 4%) in the waste.

Radionuclidic and elemental purity analysis. For irradiations at 22
MeV beam energies, small amounts of 131 La and 132 La were detected by HPGe gamma-ray spectroscopy performed on the 133/135 La eluate product after NEPTIS separation and a 3-h decay period. For 500 µA·min runs (n = 3) at 22 MeV, the 131 La and 132 La activities back-calculated to EOB were 19 ± 1.2 MBq and 0.38 ± 0.03 MBq, respectively.
The decay of 133 La resulted in small activities of its daughter nucleus 133 Ba (t 1/2 = 10.6 y). However, the resulting activity of 133 Ba after the complete decay of 133 La was approximately three orders of magnitude lower than the IAEA 1 MBq consignment level exemption limits 13 . No other radionuclidic impurities were observed in the 133/135 La product.
After allowing the 133/135 La eluate to decay for 10 days, an ICP-OES analysis was performed to investigate trace metal contaminants against a known mixture standard containing Zn, Fe, Al, Ba, Ag, In, Sn, and Cu. Metal impurities (n = 3 runs) are presented in Table 3. Table 4  Radiolabeling with the eighteen-membered macrocyclic chelator macropa was performed with 133/135 La at room temperature (22 ºC) for 10 min, and analyzed with radio-TLC.

Discussion
This study presents a high-yield cyclotron production avenue for a novel 133/135 La theranostic pair using a new sealed target design. Automated separation and purification produced a chemically pure product, with radiochemistry validating the feasibility of the 133/135 La theranostic pair using several common radiometal chelators. Table 5 outlines the positron decay characteristics and notable gamma rays for 133 La, 132 La, and several other common isotopes used for PET. 132 La has a higher positron branching ratio (41.2%) compared to 133 La (7.2%), producing more 511 keV emissions for a given sample activity. Initially, this higher branching ratio would seem advantageous for PET imaging. However, positrons emitted by 132 La have a much higher 1.29 MeV average and 3.67 MeV maximum energy compared to 133 La positron emissions, which have a low, more desirable 0.463 MeV average and 1.02 MeV maximum positron energy. Since higher positron energies are correlated with lower PET imaging spatial resolution 14,15 , this implies that 133 La would have superior PET imaging quality compared to 132 La.
The potential for improved PET scanning resolution of 133 La over 132 La could permit more accurate imaging to track the treatment of small tumors and metastases, complementing high LET targeted radionuclide therapy such as alpha particle or AET, which are both well suited for eradicating small metastatic tumors.
As shown in Table 5, 132 La has high energy gammas with a significant abundance, whereas 133 La has lower energy gammas with a much lower abundance. 132 La has a maximum gamma energy of 1909.91 keV at 9% abundance, whereas 133 La has a maximum gamma energy of 1099 keV with a 0.2% abundance. The lower energy and much lower abundance of the 133 La gamma rays should simplify handling and reduce the dose to patients upon injection for equivalent imaging activities, even though a greater activity of 133 La might be required due to the lower positron branching ratio of 133 La. In addition to potentially reducing the patient dose, the gamma ray energy distribution of 133 La could improve PET scanner imaging spatial resolution.
The 132 La 465 keV (76%) and 567 keV (14.7%) high abundance gamma rays are within a typical 350-650 keV PET scanner energy window used to detect the 511 keV annihilation gamma rays 15 , which could lead to excess spurious coincidences within the scanner timing window, and interfere with image quality. 133 La has no gamma rays with energies within a typical PET scanner energy window, which should result in no spurious coincidences. Additionally, as previously depicted in Table 2, the much lower activity ratio of 135 La to 133 La produced at 22 MeV, compared to the ratio of 135 La to 132 La produced at 11.9 MeV, should significantly reduce the relative amount of spurious coincidences in the PET scanner energy window from the 135 La 480.5 keV gamma ray.
Comparing 133 La to other PET isotopes in Table 5 shows that its respective mean and maximum positron energies of 0.461 MeV and 1.02 MeV are higher than those of 64 Cu and 18 F, comparable to those of 11 C, and 89 Zr, and lower than those of 132 La, 68 Ga, 44 Sc, and 82 Rb.
The ubiquitous 18 F has a very low positron energy that provides a sharp image, and 11 C has a similar positron energy to 133 La. However, the shorter half-lives of 18 F and 11 C limit investigating longer biological processes. 64 Cu has low energy positron emissions, a longer half-life, and βemissions that enable theranostics, however cyclotron production requires expensive isotopically enriched target material due to the low 0.009% natural abundance of 64 Ni. 89 Zr has the longest half-life of the listed isotopes, permitting users to examine longer biological processes, however, it has several high energy gamma rays (909 keV (99%), 1713 keV (0.75%), and 1744 keV (0.12%)), which greatly increase the patient dose and shielding requirements. 68 Ga has become a widely used radiometal for PET owing to its high positron branching ratio, sufficient halflife, and demonstrated chemistry. 68 Ga is easily accessible via 68 Ge/ 68 Ga generators, and alternative cyclotron Existing low current 11.9 MeV cyclotron 132/135 La production requires several-hours of long irradiations to produce small activities for limited pre-clinical applications. In contrast, much higher cross-sections for 133/135 La at 22 MeV allow a significantly shorter irradiation time producing over an order of magnitude more 133/135 La compared to 132/135 La, and significantly, large amounts of 133 La relative to 135 La as previously depicted in Table 2. This large-scale 133 La production compensates for the lower positron branching ratio of 133 La compared to 132 La. Additionally, compared to the small 132 La/ 135 La ratio shortly after EOB, the far larger 133 La/ 135 La ratio allows more flexibility with imaging and therapy.
There is a significant potential increase in PET imaging when using the 133/135 La product soon after the 3-h decay period post-EOB, as well as allowing large amounts of pure Auger therapy with a longer decay period after EOB.
A typical 18 F activity of 300-400 MBq is used for clinical PET imaging 20 , and a typical 68 Ga activity of 1.59 MBq/kg is suggested 21 . It would be a challenge to produce a 132 La activity equivalent to a typical 18 F or 68 Ga dose with current 132/135 La production methods unless isotopically enriched Ba target material was used. In contrast, it should be far easier to reach a clinically relevant 133/135 La activity with a 22 MeV irradiation of a nat Ba target. The much greater yield of 133/135 La with our 22 MeV higher energy production method should enable clinically relevant amounts of activity to be produced with relatively short irradiations.
It should be noted that not all PET centers have access to a cyclotron that can reach 22 MeV, so 133/135 La production will be limited to those centers with sufficiently high beam energy. However, the relatively long half-lives of 133 La (3.9 h) and 135 La (19.5 h) would permit regional distribution of the 133/135 La theranostic pair.