Evaluation of 134Ce/134La as a PET Imaging Theranostic Pair for 225Ac α-Radiotherapeutics

Visual Abstract

Advances in targeted molecular imaging and radionuclide therapy have given rise to the field of targeted theranostics (1). In this paradigm, a molecular agent with a PET or SPECT imaging isotope (e.g., 64 Cu, 89 Zr, or 123 I) is paired with a cognate radionuclide therapy agent (e.g., 177 Lu, 225 Ac, or 131 I) (2). a-emitting radiotherapies with isotopes, including 227 Th, 225 Ac, 213 Bi, 212 Pb/ 212 Bi, 211 At, and 149 Tb, have demonstrated promise in human trials (3,4). a-particles have a shorter range in tissue (40-100 mm) and higher linear energy transfer than b-particles (5).
To date, 225 Ac is one of the most promising radionuclides for targeted a-therapy (6). However, an imaging isotope to match with 225 Ac to measure pharmacokinetics and dosimetry has been elusive (7). Actinium has 2 short-lived daughter isotopes, 221 Fr and 213 Bi, that emit low-energy g-rays, which are challenging to image with SPECT (8). Thus, 225 Ac therapy is commonly paired with 68 Ga, 89 Zr, or 111 In for imaging-based pharmacokinetic or dosimetry information. However, because of substantial differences in halflife (t 1/2 ) ( 68 Ga) or chelation chemistry ( 89 Zr), these are imperfect PET imaging surrogates for 225 Ac. To overcome these limitations, lanthanum-based PET imaging agents such as 132 La (t 1/2 5 4.8 h, 42% b 1 ) and 133 La (t 1/2 5 3.9 h, 7% b 1 ) have emerged as potential imaging surrogates for 225 Ac (9,10). Unfortunately, the t 1/2 values of these isotopes are considerably shorter than for 225 Ac, restricting their translation to longer-t 1/2 macromolecule-based PET imaging.
In this context, the Department of Energy isotope program (11) has recently initiated the production of 134 Ce, an isotope with a 3.2-d t 1/2 that decays by electron capture to 134 La with the emission of low-energy Auger electrons. The 134 La is a positron emitter (63% b 1 ; endpoint energy, 2.69 MeV) with a t 1/2 of 6.45 min. The unique relationship between the t 1/2 values of 134 Ce and 134 La establishes a secular equilibrium (12). In pioneering work, 134 Ce cation in the 13 oxidation state has been shown to complex with diethylenetriamine pentaacetate (DTPA) (11) and DOTA (13) and to be used for in vivo PET imaging of the chelate as well as the antibody trastuzumab. It was suggested that the similar chemical characteristics between 225 Ac 31 and 134 Ce 31 and the longer 134 Ce t 1/2 (3.2-d) might be advantageous for tracking in vivo pharmacokinetics, especially at later time points. However, DOTA and DTPA require higher molar ratios and elevated temperatures for isotope complexation. Alternatively, MACROPA has demonstrated superior chelate properties for 225 Ac and a high stability (K LnL 5 15.1) for nonradioactive cerium (14), suggesting that it may function well for 134 Ce/ 225 Ac theranostic development (15). 225 Ac-based radiopharmaceutical therapy has recently attracted great interest in prostate cancer, particularly 225 Ac-PSMA-617 in small trials, demonstrating great efficacy, especially in the context of resistance to 177 Lu-PSMA-617 (16,17). Our own laboratories have identified the antibody YS5, which targets a tumor-selective epitope, CD46, that is highly expressed in prostate cancer (18). An immuno-PET agent, 89 Zr-DFO-YS5, has successfully imaged both PSMA-positive and PSMA-negative tumor xenografts and patientderived PDX models (19). Development of cognate 225 Ac-YS5 radiopharmaceuticals for therapy is currently under way (20)(21)(22). These therapeutic approaches would significantly benefit from a companion imaging agent.
Here, we aim to evaluate the potential of positron-emitting 134 Ce/ 134 La as a PET imaging surrogate for 225 Ac. We describe methods for efficient chelation of 134 Ce using the MACROPA and DOTA chelators and demonstrate the stability of the conjugates. The imaging and distribution characteristics of the 134 Ce-labeled tumor-targeting agents PSMA-617 and MACROPA-PEG 4 -YS5 are evaluated in prostate cancer models. These studies demonstrate the feasibility and applicability of 134 Ce-based radiopharmaceuticals for cancer imaging.

MATERIALS AND METHODS
Radiolabeling of DOTA, MACROPA.NH 2 , and PSMA-617 with 134 CeCl 3 134 Ce(NO 3 ) 3 in 0.1 M HCl was produced at the Isotope Production Facility of Los Alamos National Laboratory as previously described (11). Test batches were supplied by the Department of Energy isotope program for our studies. Radiolabeling reactions of DOTA, MACRO-PA.NH 2 , and PSMA-617 at various ligand-to-metal molar ratios were performed using 2 M NH 4 OAc buffer, pH 8.0, except when the product was used for animal injections (0.1 M NH 4 OAc, pH 8.0). For radiolabeling, aliquots of 134 CeCl 3 in 0.1 M HCl (5.17 mL) were mixed with MACROPA.NH 2 (23 mL, 630 mg/mL in 2 M NH 4 OAc buffer) or DOTA (20 mL, 375 mg/mL in 2 M NH 4 OAc buffer) in 2 M NH 4 OAc buffer, pH 8.0 (100 mL) at 25 C for 30 min and PSMA-617 (1.5 mL, 0.8 mg, 500 mg/mL) at 60 C for 1 h. The reaction solution was analyzed by radio-thinlayer chromatography (TLC) using C 18  MACROPA-PEG 4 -YS5 (221.4 mg; 1:1 total metal-to-YS5 molar ratio) was incubated with an aliquot of 134 CeCl 3 (105 mL, 48.1 MBq) in 2 M NH 4 OAc (pH 8.0) at 25 C for 1 h. The radiolabeling progress was monitored by instant thin-layer chromatography (iTLC) on Varian iTLC silica gel strips using 50 mM ethylenediaminetetraacetic acid, pH 5.5, as an eluent. The reaction mixture was purified over PD10 column gel filtration eluting with 0.9% saline solution.
Small-Animal PET Imaging 134 Ce-MACROPA.NH 2 and 134 Ce-DOTA reactions in 0.1 M NH 4 OAc buffer were diluted in saline (1:1 ratio), and 4.81-5.92 MBq in 100 mL were administered via the tail vein to 5-to 6-wk-old wild-type C57BL/6 male mice under isoflurane anesthesia. The specific and molar activities were 19.24 GBq/mg and 20.4 GBq/mmol, respectively, for 134 Ce-MACROPA.NH 2 and 3.7 GBq/mg and 1.9 GBq/mmol, respectively, for 134 Ce-DOTA. Dynamic small-animal PET/CT (Inveon; Siemens Medical Solutions) was performed for 1 h simultaneously on 3 mice for both 134 Ce-MACROPA.NH 2 and 134 Ce-DOTA. Free 134 CeCl 3 (4.81-5.92 MBq) in saline (100 mL) was injected similarly to the method described above, to a group of 2 mice for dynamic smallanimal PET/CT and a group of 3 mice for static small-animal PET/CT (20-min PET acquisition) at 2 h and 24 h.
For tumor imaging studies, 134 Ce-PSMA-617 (4.3 MBq) in saline (100 mL) was injected via the tail vein into 22Rv1 tumor-bearing mice, and the mice were imaged at 1 h after injection using smallanimal PET/CT. For 134 Ce-MACROPA-PEG 4 -YS5 (4.44 MBq), the conjugate was injected intravenously into mice implanted with 22Rv1 xenografts and imaged at 4 h and then at 1, 2, 4, and 7 d after injection. Small-animal PET/CT was performed with 20 min of PET at earlier time points (4 h, 1d, and 2 d) and with 60 min of PET at later time points (4 and 7 d). The specific and molar activities were 2.58 GBq/mg and 2.67 GBq/mmol, respectively, for 134 Ce-PSMA-617 and 0.18 GBq/mg and 26.94 GBq/mmol, respectively, for 134 Ce-MACROPA-PEG 4 -YS5.

Radiolabeling of Bifunctional Chelators DOTA and MACROPA.NH 2
We assessed the radiolabeling efficiencies of MACROPA.NH 2 and compared with DOTA at varying ligand-to-metal (L/M) ratios ( Fig. 1 left). The L/M ratios were calculated using the stable cerium plus lanthanum present in the 134 CeCl 3 solution as per the certificate of analysis (Supplemental Fig. 1; supplemental materials are available at http://jnm.snmjournals.org). As posited, MACRO-PA.NH 2 complexed all the 134 Ce in greater than 95% yield from 0.5:1 to 10:1 L/M ratios. In contrast, DOTA complexed 94.2% 6 1.8% of the 134 Ce only at the 10:1 L/M ratio ( Fig. 1; Supplemental  Fig. 2). A slight increase in radiolabeling complexation was observed for DOTA using L/M ratios of 2:1 (32.6% vs. 23.3%) and 5:1 (88.2% vs. 72.53%) at an elevated temperature of 60 C (Supplemental Fig. 3). These studies demonstrate that MACRO-PA.NH 2 exhibited a radiolabeling yield superior to that of DOTA, notably allowing rapid, near-quantitative radiolabeling at a 1:1 L/M ratio at room temperature. The 134 Ce-MACROPA.NH 2 (1:1 ratio) radiocomplex was analyzed by reverse-phase radio-high-performance liquid chromatography, and the retention time was compared with the Nat Ce-MACROPA.NH 2 complex (Supplemental Figs. 4-8; Supplemental Scheme 1). However, the radio-high-performance liquid chromatogram showed a tailing behavior, likely due to the ejection of 134 La from the chelate after the decay by its parent, 134 Ce. The stability of the 134 Ce-MACROPA.NH 2 complex was evaluated in physiologic buffers and in human and rat serum. Over 7 d, more than 95% of the complex was intact in all buffers and serum (Supplemental Fig. 9).
In Vivo Stability of 134 Ce-MACROPA.NH 2 and DOTA Demonstrated by PET Imaging and Biodistribution Studies After successful 134 Ce radiolabeling of MACROPA.NH 2 and DOTA, complex pharmacokinetics and stability were studied in healthy wild-type C57BL/6 mice via PET imaging and biodistribution compared with free 134 CeCl 3 . 134 CeCl 3 showed a gradual increase in liver uptake, as well as in bladder and kidney uptake ( Fig Table 1. High liver (71.5 6 4.3 percentage injected dose [%ID]/g) and bone (15.54 6 2.69 %ID/g) uptake was observed for free 134 CeCl 3, with similar results found at 2.5 and 24 h after injection (Supplemental Fig. 15). In contrast, 134 Ce-MACROPA.NH 2 (4.36 6 2.54 %ID/g) and 134 Ce-DOTA (5.17 6 2.33 %ID/g) were equally taken up in the kidney, with low accumulation in the liver and other organs, indicating low nonspecific accumulation and renal clearance. Overall, the PET imaging and biodistribution studies of 134 Ce-MACRO-PA.NH 2 and 134 Ce-DOTA versus free 134 Ce demonstrated high complex in vivo stability.
The ex vivo biodistribution of 225 AcCl 3 , 225 Ac-MACROPA.NH 2 , and 225 Ac-DOTA (Supplemental Fig. 16; Supplemental Table 2) was assessed and compared with the respective 134 Ce complexes. Free 225 Ac accumulates primarily in the liver (38.33 6 6.75 %ID/g) and bone (29.56 6 2.40 %ID/g), similarly to 134 Ce (Fig. 2B). 225 Ac-MACROPA.NH 2 (3.54% 6 1.07%) and DOTA (3.07 6 0.99 %ID/g) complexes displayed a higher uptake in the kidney, with minimal uptake in the liver (0.74 6 0.19 and 0.28 6 0.008 %ID/g), similarly to 134 Ce-MACROPA.NH 2  Given the encouraging in vivo results in normal mice, we investigated the 134 Ce radiochemistry of cancer-targeting radiopharmaceuticals, including the small-molecule prostate-specific membrane antigen (PSMA)-targeting agent PSMA-617 (23) and the CD46targeting antibody derivative MACROPA-PEG 4 -YS5. For PSMA-617, higher L/M ratios were required for quantitative 134 Ce-labeling, as 24.3%, 81.0%, and 100% radiolabeling yields were noted by radio-TLC for 2:1, 5:1, and 10:1 L/M ratios, respectively ( Fig. 3A; Supplemental Fig. 17). The radiolabeling yields were comparable to the similar ratios (10:1) of 225 Ac-PSMA-617 based on the prior literature (24). After 1 h of incubation of 134 Ce with PSMA-617 (Fig. 3), iTLC showed 94.1% radiolabeling yield. Surprisingly, the radiolabeling yields were apparently reduced to about 53.2% when the reaction was diluted in saline. However, when the same TLC plate was allowed to decay and rescanned, quantitative labeling was again observed. Similarly, when the apparently 94.1% pure 134 Ce-PSMA-617 was analyzed on reverse-phase radio-highperformance liquid chromatography (Supplemental Fig. 18), a significant tailing behavior was observed between 4 and 9 min. These data are consistent with the release of 134 La due to the dechelation or recoil effect after the decay of the parent, 134 Ce. On the basis of the favorable model labeling studies, we hypothesized that MACROPA would be a superior chelator to enable 134 Ce immuno-PET imaging. To facilitate the bioconjugation of MACROPA to the YS5 antibody, we prepared a bifunctional chelator containing MACROPA with a short PEG 4 linker with an activated TFP ester. MACROPA-PEG 4 -TFP (7 g) was synthesized over 7 steps in 56.  Fig. 38). Optimized conditions for MACROPA 134 Ce-labeling were applied, and the radiochemical yield was 96.4% as confirmed by radio-iTLC, with 69.3% isolated yield after purification and a specific activity of 0.18 GBq/mg (Figs. 3C and 3D). In contrast, DOTA-YS5 was unable to complex 134 Ce even at higher molar ratios (L/M ratio, 2 or 4) and 40 C (Supplemental Fig. 39). Calculation of the ligand-to-metal ratios was based on the number of chelators per antibody YS5. Unexpectedly, the purified eluted fraction of 134 Ce-MACROPA-PEG 4 -YS5 showed an apparent decrease in radiochemical purity to about 53.8% (Fig. 3D). As seen in the case of labeled PSMA-617, when the same TLC plate was scanned after decaying for 1 h, 100% radiochemical yield was observed (Fig. 3D). Sizeexclusion chromatography demonstrated no evidence of aggregation, whereas an elevated baseline was noticed between the product peak at 9.65 to 25 min, indicating the possible dechelation of daughter isotope 134 La (Supplemental Fig. 40). The release of daughter 134 La was also evident when these reaction mixtures were diluted in saline either for purification or for mouse injections, irrespective of MACROPA or DOTA ligands (Fig. 3E).

In Vitro Analysis and In Vivo Distribution of Prostate-Targeting Agent PSMA-617
The cell-binding assay of 134 Ce-PSMA-617 was performed with different concentrations using the 22Rv1 cell line. The percentage of cell-bound activity was significantly higher for all the concentrations than for blocking controls. A decrease in cell-bound activity percentage for a higher concentration (0.8 nM) was observed because of the cold mass effect (Supplemental Fig. 41) (26). Small-animal PET/CT was performed on a 22Rv1 tumor-bearing mouse at 1 h after injection. As shown in Figure 4, most of the activity was in the bladder and kidney at 1 h after injection, with low uptake in the tumor, whereas almost all the activity was eliminated from the other organs. This pattern of tumor uptake is similar to that found using other PSMA-targeting agents in 22Rv1 tumors, which express moderate levels of PSMA (27,28).

DISCUSSION
In the design of theranostic agents, it is essential to match the structure and biodistribution of the imaging molecule to that of the radiotherapeutic. Recently, lanthanides have been proposed as nonradioactive surrogates for actinium because of similar chemical properties. 132 La (t 1/2 5 4.8 h) and 133 La (t 1/2 5 3.9 h) have been studied as complementary PET imaging isotopes for targeted a-therapy with 225 Ac (t 1/2 5 9.9 d) (9,10). Aluicio-Sarduy et al. reported cyclotron-produced 132 La-labeled alkyl phosphocholine (NM600) in a 4T1 tumor and showed in vivo uptake characteristics similar to those of 225 Ac (9). Similarly, Nelson et al. described a high-yield cyclotron method to produce 133 La using natural barium and isotopically enriched 135 BaCO 3 targets (10). Potential limitations of 132 La and 133 La include shorter t 1/2 values than for 225 Ac (t 1/2 5 9.92 d) and elevated temperatures (80 C-90 C) required for higher radiochemical conversions (.95%). Although these may be more suitable for fast-clearing small molecules, antibody fragments, or small peptides, their t 1/2 values limit the ability to monitor the pharmacokinetics of macromolecules such as antibodies.  Ce has emerged as an isotope that may be complexed by the same chelates as actinium and thorium. Its decay to 134 La provides an in situ generator of a positron-emitting isotope with the apparent t 1/2 of its parent. The pioneering study by Bailey et al. highlighted the cyclotron production of 134 Ce/ 134 La from a natural lanthanum target and established the radiochemistry with ligands DTPA (as a potential surrogate for 225 Ac) and hydroxypyridinone (as a potential surrogate for 227 Th) (11). Later, the same group demonstrated the in vivo distribution of 134 Ce-DOTA-trastuzumab, an internalizing antibody (13). In the present study, imaging and biodistribution of a small-molecule conjugate, PSMA-617, and the antibody YS5 conjugated with MACROPA (MACROPA-PEG 4 -YS5) were conducted on prostate cancer xenografts. Similar tumor uptake was observed between the 134 Ce-and 225 Ac-labeled MACROPA-PEG 4 -YS5. The 134 Ce/ 134 La pair allows lengthy in vivo monitoring of molecules because of its extended t 1/2 of 3.2 d, which is not possible with 132/133 La radioisotopes.
Broadly speaking, the radiolabeling findings and stability using MACROPA and DOTA chelators with 134 Ce recapitulate the prior reports using the same chelators with 225 Ac (15). Radiolabeling efficiency of greater than 95% was achieved with 1:1 ligand-tometal ratios for MACROPA.NH 2 and 10:1 for DOTA at room temperature. Dynamic PET imaging and ex vivo biodistribution studies of both 134 Ce-MACROPA.NH 2 and 134 Ce-DOTA confirm in vivo stability and a biodistribution similar to that of 225 Ac-MACROPA.NH 2 and DOTA complexes. Overall, the radiolabeling methodologies show that MACROPA.NH 2 was more efficient than DOTA and that both complexes showed excellent overall stability.
After radiolabeling and purification into saline of the tumortargeting agents PSMA-617 and MACROPA-PEG 4 -YS5 for mouse administration, we chromatographically observed the release of the daughter radionuclide 134 La from the chelate. In the reaction mixture, before dilution or purification, the 134 La may be rechelated after recoil effect if excess ligand is present (Fig. 3E). However, the rechelation may not occur in vivo even if the excess ligand is present, leading to possible 134 La redistribution. Though the stability constants were high for Nat La-MACROPA (14.91) and Nat Ce-MACROPA (15.11) (14), the 134 Ce bond dissociation occurs because of the nuclear recoil effect through electron capture decay and subsequent Auger electron emission (29). A similar phenomenon was seen by Severin et al. for another in vivo PET generator, 140 Nd (t 1/2 5 3.4 d, Electron capture (EC)/ 140 Pr (t 1/2 5 3.4 m, b 1 ), with DOTA-LM3 (small peptide) and DTPA-ATN 291 (antibody). In their work, small differences in tissue distribution were noted via pre-and postmortem imaging-differences that were attributed to redistribution of the daughter. The differences were greater for noninternalizing agents (30,31). Our imaging findings are also consistent with these prior reports.
The imaging properties of 134 Ce/ 134 La have been evaluated in prostate cancer models using PSMA-617 and MACROPA-PEG 4 -YS5. Low to moderate tumor uptake of 134 Ce-PSMA-617 was observed at 1 h after administration. High kidney uptake of PSMAbased targeting vectors is known, as they tend to excrete through renal elimination and the mouse kidneys express PSMA (27,28). In contrast, 134 Ce-MACROPA-PEG 4 -YS5 showed elevated tumor uptake. Our findings are consistent with our prior report demonstrating elevated uptake of 89 Zr-DFO-YS5, compared against 68 Ga-PSMA-11 in the 22Rv1 xenograft model (19).
Remarkably, biodistribution studies of 134 Ce-MACROPA-PEG 4 -YS5 showed tissue distribution almost identical to that of 225 Ac-MACROPA-PEG 4 -YS5 except for the liver and spleen. The high liver uptake observed in early images at 24 h (Fig. 5B) may be due to redistribution of daughter 134 La after ejection from the chelate. This possibility will be further investigated in future studies by conducting pre-and postmortem imaging and comparing it with 225 Ac more systematically. One notable advantage to using 134 Ce is that it allows facile imaging of conjugates bearing the MACROPA chelate, which was previously limited to therapeutic radionuclides. The similar chemical properties of these radionuclides ( 134 Ce/ 225 Ac) may allow a single molecular platform by complexing with the ligands DOTA or MACROPA. This complexation could facilitate predicting the tumor distribution of 225 Ac-labeled targeting vectors ( 225 Ac-PSMA-617 or MACROPA-PEG 4 -YS5) based on the ( 134 Ce-PSMA-617 or MACROPA-PEG 4 -YS5) PET imaging results. Hence, this methodology addresses an important challenge in radiopharmaceutical sciences, namely the study of the biodistribution of 225 Ac radiopharmaceuticals. Overall, these studies support our premise that 134 Ce/ 134 La may serve as an imaging radionuclide to pair with 225 Ac. CONCLUSION MACROPA.NH 2 showed exceptional radiolabeling efficiency with 134 Ce at room temperature. PET imaging of 134 Ce-MACRO-PA.NH 2 and 134 Ce-DOTA revealed that both tracers are highly stable in vivo. The ex vivo biodistributions of both 134 Ce-DOTA and MACROPA.NH 2 were almost identical to the respective 225 Ac complexes. 134 Ce-PSMA-617 shows high binding affinity and uptake in prostate cancer 22Rv1 xenografts. A bifunctional analog for MACROPA was synthesized, conjugated with antibody YS5, and radiolabeled with 134 Ce and 225 Ac. Both the PET imaging and the biodistribution of 134 Ce-MACROPA-PEG 4 -YS5 demonstrate elevated tumor retention in 22Rv1 prostate cancer xenografts. The ex vivo biodistribution is consistent with the 225 Ac-MACROPA-PEG 4 -YS5 distribution in most tissues, including the tumor. These studies support the future development of 134 Ce-radiopharmaceuticals for cancer imaging as a companion to paired a-particle radiotherapeutics.