Preclinical comparison of four [18F, natGa]rhPSMA-7 isomers: influence of the stereoconfiguration on pharmacokinetics

Introduction Radiohybrid (rh) ligands, a novel class of prostate-specific membrane antigen (PSMA)-targeted radiopharmaceuticals, can be labeled either with [18F]fluorine via isotopic exchange or with radiometals (such as [68Ga]Gallium, [177Lu]Lutetium, [225Ac]Actinium). Among these, [18F, natGa]rhPSMA-7 has recently entered clinical assessment. Aim Since [18F, natGa]rhPSMA-7 is composed of four stereoisomers ([18F, natGa]rhPSMA-7.1, -7.2, -7.3 and -7.4), we initiated a preclinical selection process to identify the isomer with the most favorable pharmacokinetics for further clinical investigation. Methods A synthetic protocol for enantiopure [19F, natGa]rhPSMA-7 isomers has been developed. The comparative evaluation of the four isomers comprised human serum albumin binding, lipophilicity, IC50, internalization and classical biodistribution studies and competition experiments in LNCaP tumor-bearing CB-17 SCID mice. In addition, a radio high-performance liquid chromatography-based method was developed allowing quantitative, intraindividual comparison of [18F, natGa]rhPSMA-7.1 to -7.4 in LNCaP tumor-bearing mice. Results Cell studies revealed high PSMA affinity and internalization for [18/19F, natGa]rhPSMA-7.2, -7.3 and -7.4, whereas [18/19F, natGa]rhPSMA-7.1 showed approximately twofold lower values. Although the biodistribution profile obtained was typical of PSMA inhibitors, it did not allow for selection of a lead candidate for clinical studies. Thus, an intraindividual comparison of all four isomers in LNCaP tumor-bearing mice was carried out by injection of a diastereomeric mixture, followed by analysis of the differential uptake and excretion pattern of each isomer. Based on its high tumor accumulation and low uptake in blood, liver and kidneys, [18F, natGa]rhPSMA-7.3 was identified as the preferred isomer and transferred into clinical studies. Conclusion [18F, natGa]rhPSMA-7.3 has been selected as a lead compound for clinical development of a [18F]rhPSMA-based candidate. The intraindividual differential uptake and excretion analysis in vivo allowed for an accurate comparison and assessment of radiopharmaceuticals.


Introduction
During the last decade, advancements in the field of prostate-specific membrane antigen (PSMA)-targeting radiopharmaceuticals have had significant impact on the clinical management of patients suffering from prostate cancer [1][2][3]. [ 68 Ga]Ga-PSMA-11 [4,5] in particular has been the subject of extensive evaluation and has already proved its eligibility, especially for positron emission tomography (PET)-based detection of biochemical recurrence [6,7], resulting in its broad clinical use [8,9].
Due to the superior nuclear properties of the fluorine-18 radionuclide and accompanying logistic and economic advantages [10,11], a shift of interest from 68 Ga-labeled PSMA tracers toward 18 F-labeled analogues has been observed in recent years [12][13][14]. In this context radiohybrid (rh) PSMA ligands, developed by our group, form a novel class of radiopharmaceuticals, which combine a Silicon-Fluoride-Acceptor (SiFA) and a metal chelate (or a chelator) in a single molecule [15]. Such rhPSMA ligands can either be labeled with fluorine-18 by isotopic exchange at the SiFA-moiety in the presence of a non-radioactive metal chelate (e.g., nat Ga-or nat Luchelate), or with a radiometal (e.g., [ 68 Ga]Gallium, [ 177 Lu] Lutetium, [ 225 Ac]Actinium) by means of the chelator, while the SiFA moiety is non-radioactive [15]. The chemical identity of the 18 7) has already been assessed for PET imaging of primary and metastatic castration-resistant prostate cancer. Biodistribution of [ 18 F, nat Ga]rhPSMA-7 was found to be similar to that of established PSMA ligands, and [ 18 F, nat Ga]rhPSMA-7 PET/CT demonstrated high detection rates in early biochemical recurrence after radical prostatectomy, especially among patients with low prostate-specific antigen values [16]. Furthermore, the novel tracer outperformed morphologic imaging for N-staging of high-risk primary prostate cancer, with efficacy comparable to the literature data for [ 68 Ga]Ga-PSMA-11 [17]. The highest imaging quality of this tracer was obtained at 50-70 min post injection in PET/CT [18].
Here, we report a preclinical comparison of the four rhPSMA-7 isomers, [ 18

Automated 18 F-labeling
The production of [ 18 F, nat Ga]rhPSMA-7 with starting activities of 50-100 GBq was performed by means of a fully-automated procedure at the Klinikum rechts der Isar (see Additional file 1: Fig. 6) [15].

Lipophilicity and binding to human serum albumin
Approximately 1 MBq of the 18 F-ligand was dissolved in 1 mL of a 1:1 mixture (v/v) of phosphate-buffered saline (PBS, pH 7.4) and n-octanol (n = 18-23). After vigorous mixing of the suspension for 3 min, the vial was centrifuged at 15,000 × g for 3 min and 100 μL aliquots of both layers were measured in a γ-counter. HSA binding of 19 F-nat Ga-rhPSMA ligands was determined according to a previously published procedure [15,21].

In vivo experiments
All animal experiments were conducted in accordance with general animal welfare regulations in Germany (German animal protection act, as amended on 18.05.2018, Art. 141 G v. 29.3.2017 I 626, approval no. 55.2-1-54-2532-71-13 by the General Administration of the Free State of Bavaria) and the institutional guidelines for the care and use of animals.

Differential uptake and excretion pattern
With the intention to identify even marginal differences in the in vivo behavior, all four [ 18 F, nat Ga]rhPSMA-7 isomers were co-evaluated in one single tumor-bearing mouse in order to quantify the differential uptake and excretion pattern of each isomer. For this purpose, the diastereomeric mixture [ 18 F, nat Ga]rhPSMA-7 was produced in high molar activity (A M ) of 247-349 GBq/ µmol (at end of synthesis) and the abundance of each isomer was determined by radio-HPLC. Mice (n = 4) were injected with [ 18 F, nat Ga]rhPSMA-7 (180-280 MBq, 0.9-1.0 nmol) and left under anesthesia. After 30 min p.i., they were sacrificed and urine, blood, liver, kidneys and tumor were collected and processed. Radioactivity was extracted from solid tissues by means of a Potter-Elvehjem tissue grinder (n = 1) or a ball mill (n = 3) and separated from the protein fraction by subsequent cartridge-based solid-phase extraction (SPE). Blood samples were diluted, centrifuged and the supernatant was purified by SPE (see supporting information). Urine and extracts from tissue samples were then analyzed by radio-HPLC to quantify the fraction of each isomer in each sample. Finally, the relative percentage difference between the abundance of each isomer in the injected diastereomeric mixture (values taken from the quality control) and the abundance of the respective isomer in each analyzed tissue or body fluid was calculated. To accurately quantify the abundance of the isomers, especially in cases when no baseline separation was achieved in radio-HPLC, all radio-HPLC profiles were processed by the Systat (San Jose, US) software package PeakFit.
PeakFit allows for automated nonlinear separation, analysis and quantification of HPLC elution profiles by deconvolution procedures that uses a Gaussian response function with a Fourier deconvolution/filtering algorithm (see Additional file 1: Fig. 7).

Chemical synthesis
When employing standard coupling conditions (HOBt, TBTU and DIPEA) during the initial production of [ 19 F, nat Ga]rhPSMA-7, conversion of D-to L-Dap was observed. Optimization was carried out by either substitution of the coupling reagents HOBt and TBTU by HATU or replacement of DIPEA by the weaker base 2,4,6-trimethylpyridine. Whereas the use of HATU already reduced the conversion to about 12%, it was almost eliminated (< 2%) using 2,4,6-trimethylpyridine.
To eliminate the second source of isomers, racemic DOTA-GA employed in [ 19 F, nat Ga]-rhPSMA-7 was substituted by the enantiomerically pure S-and R-DOTA-GA. After non-radioactive metal complexation with Ga(NO 3 ) 3 and final purification, HPLC analysis (UV absorbance at 220 nm) revealed > 98% purity for all 19 F-nat Ga-rhPSMA-7 isomers and also confirmed the absence of undesired diastereomers.

Analytical characterization
Previously produced clinical batches of diastereomeric

Radiolabeling
Drying of aqueous [ 18 F]fluoride followed by 18 F-labeling by 18 F-for- 19 F isotopic exchange was carried out as previously described [15,22]

Differential uptake and excretion pattern
With the aim to generate a more robust set of comparative data that might allow for a well-founded selection of the most promising isomer, we decided to co-assess [ 18 F, nat Ga]rhPSMA-7.1 to -7.4 intraindividually (n = 4) by analyzing the differential uptake and excretion pattern in urine, blood, liver, kidneys and tumor after injection of [ 18 F, nat Ga]rhPSMA-7 by means of radio-HPLC analyses.

Tissue extraction
The percentage of the radioactivity which was extracted from tissue samples and obtained after SPE-based removal of the protein fraction is summarized in Table 1. Most of the activity could be recovered from blood (90 ± 4%) and liver (86 ± 2%). When compared with the ball mill, sample extraction via the manual tissue grinder was slightly more efficient for kidney (91% vs. 63 ± 5%) and tumor (90% vs. 64 ± 18%). However, after both procedures a significant amount of the extracted activity could not be trapped on the SPE cartridge, resulting in  a decreased overall extraction efficiency of 43-60% and 42-53% for kidney and tumor, respectively.

Metabolic stability
Radio-HPLC analyses of urine samples and of the radioactivity extracted from the homogenized (kidney, liver, tumor) or diluted (blood) samples did not show any radiolabeled fragments of [ 18 F, nat Ga]rhPSMA-7.

Differential uptake and excretion pattern
The differential uptake and excretion pattern of [ 18 F, nat Ga]rhPSMA-7 isomers in urine, blood, liver, kidneys and tumor is shown in Fig. 5, and expressed as the relative percentage difference between the abundance of the isomer in the quality control (and thus, at the time point of injection into tumor-bearing mice) and the abundance of the isomer in the analyzed tissue or body fluid (also see Additional file 1: Fig. 9). Compared to the time point of injection, the abundance of the D-Dap-containing isomers [ 18 F, nat Ga]rhPSMA-7.1 and -7.3 was found to be higher in blood, liver, kidney and tumor samples in all experiments, thus demonstrating a stronger ´enrich-ment´ of D-Dap-based isomers when compared with the L-Dap-isomers [ 18 F, nat Ga]rhPSMA-7.2 and -7.4. One outlier was found in the second experiment of the blood sample, in which [ 18 F, nat Ga]rhPSMA-7.2 showed higher Table 1 Extraction efficiency of [ 18 F, nat Ga]rhPSMA-7 from blood, liver, kidney and tumor.
Samples were extracted using either a Potter-Elvehjem tissue grinder (n = 1) or a MM-400 Ball Mill (n = 3). The percentage of activity after sample extraction and after SPE purification was quantified, decay-corrected, and the overall extracted activity was calculated (values are expressed as mean ± SD).

Sample extraction SPE purification Overall
Potter-Elvehjem tissue grinder (n = 1)  -7.4 showed the lowest enrichment in blood, liver and kidney, but also the lowest uptake by tumors.

Discussion
The novel radiohybrid PSMA-targeting ligand [ 18 F, nat Ga] rhPSMA-7 already demonstrated promising results in staging and restaging of prostate cancer [16,17]. Since the compound is composed of four stereoisomers ([ 18 F, nat Ga]rhPSMA-7.1, -7.2, -7.3 and -7.4), this study aimed to identify the one isomer with the most promising characteristics for further clinical investigation.
Prior to the actual selection process, the chemical synthesis needed to be improved to produce the individual [ 19 F, nat Ga]rhPSMA-7 isomers. Suppression of unexpected conversion of D-to L-Dap was achieved by using the weaker base 2,4,6-trimethylpyridine. The benefits of 2,4,6-trimethylpyridine, particularly when used in conjunction with structurally-related cysteine analogues, have already been described in the literature [27]. The authors speculate that formation of an active-ester of Fmoc-Dap(Dde)-OH during pre-activation increases acidity of the proton at the α-carbon, which then is abstracted by DIPEA. The abstraction is also promoted by the strong electron-withdrawing effect of the Ddeprotecting group, resulting in a partial positive charge at the β-amine of Dap. In contrary 2,4,6-trimethylpyridine seems to be too weak and bulky for proton abstraction.
This key step and the application of the respective enantiopure DOTA-GA chelator allowed for production of the [ 19 F, nat Ga]rhPSMA-7 isomers in excellent final purities of > 98%.
The following classic comparative in vivo evaluation of the four [ 18 F, nat Ga]rhPSMA-7 isomers resulted in less reproducible data sets, and the differences between the biodistribution of each isomer did not allow to select one compound. In addition, there were certain discrepancies between the in vivo and in vitro results. As an example, [ 18 F, nat Ga]rhPSMA-7.1 showed the lowest PSMA affinity and the lowest internalization, while its tumor uptake at 1 h p.i. was almost identical to that of [ 18 F, nat Ga] rhPSMA-7.3 and -7.4. On the other hand, the more than twofold lower tumor uptake of [ 18 F, nat Ga]rhPSMA-7.2 did not correspond with its high PSMA-affinity and internalization rate.
One potential explanation for these results might be differences in the metabolism in individual mice, which become most apparent in experiments involving small animal cohorts (e.g., n = 3 for [ 18 F, nat Ga]rhPSMA-7). Furthermore, in vivo degradation of [ 18 F, nat Ga]rhPSMA-7 isomers, especially of L-Dap-comprising compounds (7.2 and 7.4), by peptidases must be considered [28,29].
In order to overcome these limitations, we established a radio-HPLC-based analysis to intraindividually determine the differential uptake and excretion pattern of the four isomers with the aim to precisely assess their different in vivo behavior. This method allows to minimize methodological errors, to discover potential metabolites and to reduce the number of experimental animals.
Interestingly, no hydrophilic or lipophilic radiolabeled fragments were detected by radio-HPLC analyses of urine and extracted samples from blood, liver, kidneys and tumor. Due to fewer carboxylic acids and high lipophilicity of the SiFA-moiety (log P = 3.6 [30,31]), all radioactive species formed by metabolic cleavage should exhibit increased lipophilicity (Additional file 1: Fig. 10). In this context one could argue that potential lipophilic metabolites might not be quantitatively extracted from samples and thus remain unaccounted for. However, such lipophilic species formed in vivo would either bind to plasma proteins and/or show hepatobiliary excretion, resulting in noticeable accumulation in the liver and gastrointestinal system [32,33]. Due to the high extraction efficiencies in blood (90 ± 4%) and liver (86 ± 2%), such speculations seem unsubstantiated. Moreover, no elevated uptake in these organs was observed in biodistribution studies, indicating high metabolic stability of all isomers. Similarly, defluorination by hydrolysis of the [ 18 F]SiFA moiety would have resulted in elevated activity accumulation in bone [34]; again not detected in either biodistribution studies in mice, or in clinical PET scans using [ 18 F, nat Ga]rhPSMA-7 [18].
Based on the analysis of the differential uptake and excretion pattern, we favored the D-Dap isomers [ 18 F, nat Ga]rhPSMA-7.1 and -7.3. These isomers showed higher tumor uptake than L-Dap isomers [ 18 F, nat Ga] rhPSMA-7.2 and -7.4, which showed lowest accumulation in all analyzed tissues and blood.
Overall, out of the more favorable D-Dap-isomers, [ 18 F, nat Ga]rhPSMA-7.3 seems to be the isomer of choice, as it shows lower enrichment in blood, liver and kidney, yet still displaying high tumor uptake. In addition, this isomer represented almost ca. 40% of the former diastereomeric mixture [ 18 F, nat Ga]rhPSMA-7. Thus, significant problems, such as metabolism, unfavorable organ distribution or tumor uptake in men would have been already detected during the clinical PET investigations with diastereomeric mixture [ 18 F, nat Ga]rhPSMA-7.
The intraindividual co-assessment of tracers or isomers, when experimentally feasible, seems to be the evaluation method of choice, as that method takes into account all known or unknown parameters that finally influence the uptake and excretion of a tracer in vivo.