Synthesis and Preclinical Evaluation of 177Lu-Labeled Radiohybrid PSMA Ligands for Endoradiotherapy of Prostate Cancer

Visual Abstract

The prostate-specific membrane antigen (PSMA)-targeted radiohybrid (rh) ligand [ 177 Lu]Lu-rhPSMA-7.3 has recently been assessed in a pretherapeutic dosimetry study on prostate cancer patients. In comparison to [ 177 Lu]Lu-PSMA I&T, application of [ 177 Lu]Lu-rhPSMA-7.3 resulted in a significantly improved tumor dose but also higher kidney accumulation. Although rhPSMA-7.3 has been initially selected as the lead compound for diagnostic application based on the characterization of its gallium complex, a systematic comparison of the most promising 177 Lulabeled rhPSMA ligands is still missing. Thus, this study aimed to identify the rhPSMA ligand with the most favorable pharmacokinetics for 177  LNCaP cells, as well as lipophilicity measurements. In addition, we determined the apparent molecular weight (AMW) of each tracer as a parameter for human serum albumin (HSA) binding. Biodistribution studies and small-animal SPECT imaging were performed on LNCaP-tumor bearing mice at 24 h after injection. Results: 177 Lu labeling of the radiohybrids was performed according to the established procedures for the currently established PSMA-targeted ligands. All ligands showed potent binding to PSMA-expressing LNCaP cells, with affinities in the low nanomolar range and high internalization rates. Surprisingly, the most pronounced differences regarded the HSA-related AMW. Although Radi oligand therapy (RLT) of metastatic castration-resistant prostate cancer with ligands targeting prostate-specific membrane antigen (PSMA) holds great promise for patients who have exhausted conventional treatment regimens. Currently, 177 Lu-labeled PSMA-617 (1) and PSMA I&T (2) are the most extensively evaluated agents in this class and have demonstrated favorable safety and good treatment response rates (3,4). Although regulatory approval is still awaited for both agents, their application in compassionate-use programs has recently been reaffirmed by the European Association of Nuclear Medicine procedure guidelines for 177 Lu-PSMA therapy (5). [ 177 Lu]Lu-PSMA-617 has been evaluated by Novartis in a phase 3 clinical trial (NCT 03511664) on patients with metastatic castration-resistant prostate cancer. In that trial, [ 177 Lu]Lu-PSMA-617 was compared with the best standard of care. The investigators recently announced that both primary endpoints-overall and radiographic progression-free survival-were met (6). In addition, an ongoing phase 3 trial investigating [ 177 Lu]Lu-PSMA I&T (NCT 04647526) is evaluating its efficacy versus abiraterone or enzalutamide in delaying radiographic progression in patients with metastatic castration-resistant prostate cancer after second-line hormonal treatment. Retrospective clinical comparison of 177 Lu-labeled PSMA-617 and PSMA I&T point toward nearly identical pharmacokinetics for both tracers, and clinical efficacy is assumed to be similar, with no clear advantage to either compound (7).
Recently, the novel class of radiohybrid (rh) PSMA-targeted ligands was introduced by our group (8). These compounds combine a silicon-fluoride acceptor for 19 F/ 18 F-isotopic exchange radiolabeling and a chelator for complexation of a metal or radiometal (e.g., 177 Lu, 68 Ga, or 225 Ac). Respective ligand pairs of 18 F/nonradioactive metal and 19 F/radiometal, such as [ 18 F]Lu-rhPSMA and [ 177 Lu]Lu-rhPSMA, are chemically identical and thus display identical pharmacokinetics, offering unique possibilities for theranostic applications (Fig. 1).
For initial evaluation of the radiohybrid technology for therapeutic applications, rhPSMA-7.3 was labeled with 177 Lu and compared with [ 177 Lu]Lu-PSMA I&T in biodistribution and dosimetry studies on mice. Both ligands showed similar uptake in healthy organs, resulting in a similar dose to all major organs, including bone marrow and kidney (13). Compared with [ 177 Lu]Lu-PSMA I&T, [ 177 Lu]Lu-rhPSMA-7.3 exhibited a 2.8-and 4.7-fold higher tumor uptake at 1 and 168 h after injection, respectively, resulting in a significantly higher dose at the tumor and a superior treatment response (13).
In a pretherapeutic comparative dosimetry study of 177 Lu-labeled rhPSMA-7.3 and PSMA I&T on a small patient cohort (n 5 6, intraindividual comparison), an approximately 2.4-fold higher mean absorbed dose for tumor lesions of the radiohybrid ligand was found, consistent with the preclinical observations (14). However, contradictory to animal studies, the mean absorbed dose to different healthy organs was also higher-for example, 2.3-fold higher doses to kidneys and 2.2-fold higher doses to bone marrow for [ 177 Lu]Lu-rhPSMA-7.3 versus [ 177 Lu]Lu-PSMA I&T. The authors concluded that the radiohybrid tracer holds promise for therapeutic effects similar to those obtained with [ 177 Lu]Lu-PSMA I&T while offering potential economic advantages by an approximately 2-fold reduction in the injected radioactive doses (14).
The selection of rhPSMA-7.3 as lead compound for diagnostic application was based on the evaluation of gallium chelates [ 18 F]Ga-rhPSMA-7.1, -7.2, -7.3, and -7.4 (12). Since it is known from the literature that the complex structure of the metal chelate (e.g., Given the promising initial data from 177 Lu-labeled rhPSMA-7.3, the aim of the present study was to evaluate whether other isomers of 177 Lu-labeled rhPSMA-7 or the closely related compounds

MATERIALS AND METHODS
A detailed description of the chemical synthesis of rhPSMA and the analytic instruments is provided in the supplemental materials (available at http://jnm.snmjournals.org).

Radiolabeling
Radiolabeling with no-carrier-added 177 Lu was performed according to the established procedures for PSMA-targeted ligands (1,2). Briefly, the precursor (1.0 nmol, 10 mL, 0.1 mM in dimethylsulfoxide) was added to 10 mL of 1.0 M aqueous NaOAc buffer (pH 5.5). Subsequently, 20-50 MBq of [ 177 Lu]LuCl 3 (specific activity . 3,000 GBq/mg at the time of radiolabeling, 740 MBq/mL, 0.04 M HCl; ITM) was added, and the mixture was filled up to 100 mL with 0.04 M HCl. The reaction mixture was heated for 20-30 min at 90 C, and the radiochemical purity was determined using radio-high-performance liquid chromatography and radio-thin-layer chromatography with 0.1 M sodium citrate buffer on instant thin-layer chromatography-silica gel chromatography paper (Agilent) and 1.0 M NH 4 OAc/dimethylformamide buffer (1/1; v/v) on thin-layer chromatography silica gel 60 F 254 plates (Merck Millipore).

Lipophilicity
Approximately 1 MBq of the 177 Lu-labeled PSMA ligand was dissolved in 1 mL of a 1:1 mixture (v/v) of phosphate-buffered saline (pH 7.4) and n-octanol (n 5 6). After vigorous mixing of the suspension for 3 min, the vial was centrifuged at 15,000g for 3 min, and 100-mL aliquots of both layers were measured in a g-counter. Finally, the ratio of the radioactivity detected in the n-octanol sample and the phosphate-buffered saline buffer was calculated and expressed as distribution ratio log D 7.4 .

Binding to HSA
Binding of 177 Lu-labeled ligands to HSA was assessed by albuminmediated size-exclusion chromatography (AMSEC), a novel method that has recently been developed by our group to determine the apparent molecular weight (AMW) of a compound in the presence of HSA. A dedicated and detailed description of the AMSEC method will be published elsewhere to cover the context and the development process of this method in its entirety. Briefly, a gel filtration size-exclusion column (Superdex 75 Increase 10/300 GL; fractionation range, 70-3 kDa; GE Healthcare) was calibrated as recommended by the manufacturer using a commercially available set of proteins (Gel Filtration LMW Calibration Kit; GE Healthcare). AMSEC experiments were performed by injection of the various radioligands using an HSA buffer at physiologic concentration (Biowest) as the mobile phase at room temperature. Depending on the strength of the HSA/ligand interaction during the chromatographic procedure, an injected radioligand (1.0 MBq, 10-20 GBq/mmol) can show a reduced retention time that correlates to AMWs higher than the actual, physical, molecular weight (the latter being for all investigated ligands , 2 kDa, and thus below the column fractionation range). The stronger this interaction, the longer the mean time the ligand is bound to HSA during the chromatographic process and the faster the ligand is Internalization studies of the radiolabeled ligands (1.0 nM/well) were performed on LNCaP cells (1.25 3 10 5 cells in 0.25 mL/well) at 37 C for 1 h and accompanied by ([ 125 I]IBA)KuE (0.2 nM/well) as a reference. Data were corrected for nonspecific binding and normalized to the specific internalization observed for the reference (n 5 3), as previously published (8).

In Vivo Experiments
All animal experiments were conducted in accordance with general animal welfare regulations in Germany (German animal protection act, as amended on May 18, 2018, article 141 G, version March 29, 2017, I 626, approval 55.2-1-54-2532-71-13) and the institutional guidelines for the care and use of animals. LNCaP tumor xenografts were established in 6-to 8-wk-old male CB-17 SCID mice as described previously (8).
All rhPSMA ligands were evaluated during the same period (first quarter of 2020), whereas 177 Lu-labeled PSMA-617 and PSMA I&T (17) were assessed in 2016, using the identical cell line, mouse model, and experimental procedure.
Small-Animal SPECT/CT Imaging. Static images of 177 Lu-labeled ligands in euthanized mice were recorded 24 h after injection directly after blood collection, with an acquisition time of 45 min using a highenergy, general-purpose rat-and-mouse collimator and a stepwise multiplanar bed movement. For imaging studies, a VECTor4 small-animal SPECT/PET/optical imaging/CT device from MILabs was applied. Data were reconstructed using MILabs.Rec software (version 10.02) and PMOD software (version 4.0; PMOD Technologies LLC). After imaging, the mice underwent biodistribution studies.

Synthesis and Radiolabeling
Uncomplexed PSMA ligands were obtained via a solid-phase/ solution-phase synthetic strategy with chemical purities of more than 97% as determined by high-performance liquid chromatography (absorbance at 220 nm). Identity was confirmed by mass spectrometry. Complexation with a 2.5-fold molar excess of LuCl 3 led to quantitative formation of the respective lutetium-PSMA ligands, which were used for in vitro studies. 177 Lu labeling of PSMA ligands according to standard manual procedures resulted in a radiochemical purity of more than 95%, determined by radio-high-performance liquid chromatography and radio-thin-layer chromatography (Supplemental Table 1).

In Vitro Characterization
Results of the in vitro evaluation of all rhPSMAs and the reference ligands PSMA-617 (1) and PSMA I&T (2) are summarized in Figure 3 and Supplemental Table 2. PSMA binding affinity (halfmaximal inhibitory concentration; Fig. 3A) was high and in the low nanomolar range for all lutetium-rhPSMA ligands (range, 2.

In Vivo Characterization
Biodistribution Studies. Overall, the comparative biodistribution study of the 177 Lu-labeled PSMA ligands in LNCaP tumor-bearing mice at 24 h after injection revealed a quite similar distribution pattern with high tumor uptake, fast excretion from background organs, but a varying degree of activity retention in the kidneys ( Fig. 4

DISCUSSION
Although it has recently been demonstrated in patients that the uptake of [ 177 Lu]Lu-rhPSMA-7.3 in tumor lesions was on average 2-to 3-fold higher than that of [ 177 Lu]Lu-PSMA I&T, the slower clearance resulted in a comparatively higher dose to the kidney as the organ at risk (14). In retrospect, this result is hardly surprising, since the selection process of the best diagnostic rhPSMA-7 isomer was based on criteria that are considered suboptimal for therapy: fast blood clearance to reach high tumor-to-background ratios at early time points, predominantly renal clearance, and high kidney retention to ensure low activity in the bladder at early time points. In contrast, the selection criteria for the best therapeutic isomer are different. Compared with today's therapeutic ligands, a slightly delayed blood clearance is preferred. The ligand should be excreted renally while showing almost no retention in the kidneys. It must also be considered that the change in the isotope (lutetium for RLT instead of gallium for the diagnostic compound) results in a different structure and charge at the chelate-metal complex (gallium-DOTA: hexadentate, zwitterionic; lutetium-DOTA: octadentate, uncharged) (15). This matter obviously influences the pharmacokinetic properties of the entire ligand, as demonstrated by the prolonged clearance kinetics of the 177 Lu-labeled "best diagnostic isomer" rhPSMA-7.3 in patients (14).
With the aim of addressing these therapeutic criteria and of identifying a 177 Lu-labeled rhPSMA tracer with more favorable characteristics for RLT, we performed a coevaluation of 6 different rhPSMA ligands (4 rhPSMA-7 isomers and 2 rhPSMA-10 isomers) and compared the results with preclinical data on the 2 reference compounds, PSMA-617 and PSMA I&T.
All lutetium-complexed radiohybrid tracers and the external references, PSMA I&T and PSMA-617, demonstrated potent binding to LNCaP cells with an excellent affinity in the low-nanomolar range and high internalization rates, which did not allow prioritization of certain candidates for further evaluation.
In the context of PSMA-targeted RLT, the kidney and then the bone marrow are still considered the main organs at risk, and uptake in these should be carefully considered (18).
In our comparative biodistribution studies, pronounced differences in kidney uptake values were observed. Whereas our internal reference, D-Dap-S-DOTAGA-configured [ 177 Lu]Lu-rhPSMA-7.3, showed a kidney uptake of 9.    (24). Until further investigations improve our understanding of species-dependent renal handling of PSMA tracers, and in the absence of alternative and more valid selection criteria, the evaluation of the biodistribution in mice, including the assessment of the different excretion behavior, will remain our only viable option-although we should treat such data with appropriate caution.
Regarding important nontarget organs such as liver, muscle, and heart, all ligands demonstrated almost identical and complete clearance 24 h after injection. Even though only low activity levels were found in the blood pool for all ligands, [ 177 Lu]Lu-rhPSMA-10.1 showed the best clearance of all investigated PSMA ligands, and this superior clearance is also expressed by the highest tumor-toblood ratio (11,498): 3 times higher than for [ 177 Lu]Lu-rhPSMA-7.3 and 8 times higher than for [ 177 Lu]Lu-PSMA-617.
The tendency of both DOTA-conjugated [ 177 Lu]Lu-rhPSMA-10 isomers to clear more quickly can at least in part be explained by the HSA-binding experiments and our newly introduced in vitro parameter, AMW. The molecular weight of a molecule is known to have a direct implication in the glomerular sieving coefficient (GSC) (as a rule of thumb, the lower the molecular weight, the higher the GSC and the faster the excretion kinetics) (25). Thus, the stronger the interaction of a molecule with HSA or the higher the ratio of HSAbound ligand to free ligand, the less ligand is subjected to glomerular filtration and thus the less ligand is excreted. In our assay, this ratio is indirectly determined by calculating the AMW of each compound (details on these methods will be described elsewhere).
Looking at the AMW of 177 Lu-labeled PSMA-617 (13.7 kDa) and PSMA-I&T (5.3 kDa) as key reference points, the 2.3-fold lower AMW of PSMA I&T appears prima facie unproportional: Kulkarni (7). However, taking into account the nonlinear correlation of the molecular weight and GSC and the tiny changes in the GSC at low molecular weights, the differences in the AMW of PSMA I&T and PSMA-617 result in only slightly different GSCs, which explains the similar kidney excretion kinetics of these 2 ligands in patients (25). In contrast, the higher AMW of [ 177 Lu]Lu-rhPSMA-7.3 (molecular weight, 30.4) would translate into a markedly lower GSC and thus a delayed clearance, which has been confirmed by clinical results from Feuerecker et al. (14). On the basis of these results, we expect that the blood clearance kinetics of . Thus, we are optimistic that we will be able to obtain a similarly improved tumor uptake in patients during the clinical assessment of [ 177 Lu]Lu-rhPSMA-10.1 and thus tumor doses higher than those currently obtained with the state-of-the art nonhybrid ligands (14).
Certainly there are also other effects determining different clearance kinetics of radiopharmaceuticals in mice and patients that must be considered-for example, species differences in drug binding to serum albumin (26) and differences in magnitude and binding affinities of the tracers to plasma proteins other than HSA, such as a-1-acid glycoprotein (27), transthyretin (28), or lipoproteins (29). Moreover, individual differences in uptake of PSMA ligands into the kidneys (30), varying relative proportions of hepatobiliary to renal clearance, and effects of species differences between mice and humans must be considered. In summary, however, we are optimistic that the promising biodistribution profile of [ 177 Lu]Lu-rhPSMA-10.1 observed in mice, together with its low AMW, will translate into improved tumor doses and tumor-to-kidney dose ratios of this isomer in patients.