Cyclohexanediamine Triazole (CHDT) Functionalization Enables Labeling of Target Molecules with Al18F/68Ga/111In

The Al18F-labeling approach offers a one-step access to radiofluorinated biomolecules by mimicking the labeling process for radiometals. Although these labeling conditions are considered to be mild compared to classic radiofluorinations, improvements of the chelating units have led to the discovery of (±)-H3RESCA, which allows Al18F-labeling already at ambient temperature. While the suitability of (±)-H3RESCA for functionalization and radiofluorination of proteins is well established, its use for small molecules or peptides is less explored. Herein, we advanced this acyclic pentadentate ligand by introducing an alkyne moiety for the late-stage functionalization of biomolecules via click chemistry. We show that in addition to Al18F-labeling, the cyclohexanediamine triazole (CHDT) moiety allows stable complexation of 68Ga and 111In. Three novel CHDT-functionalized PSMA inhibitors were synthesized and their Al18F-, 68Ga-, and 111In-labeled analogs were subjected to a detailed in vitro radiopharmacological characterization. Stability studies in vitro in human serum revealed among others a high kinetic inertness of all radiometal complexes. Furthermore, the Al18F-labeled PSMA ligands were characterized for their biodistribution in a LNCaP derived tumor xenograft mouse model by PET imaging. One radioligand, Al[18F]F-CHDT-PSMA-1, bearing a small azidoacetyl linker at the glutamate-urea-lysine motif, provided an in vivo performance comparable to that of [18F]PSMA-1007 but with even higher tumor-to-blood and tumor-to-muscle ratios at 120 min p.i. Overall, our results highlight the suitability of the novel CHDT moiety for functionalization and radiolabeling of small molecules or peptides with Al18F, 68Ga, and 111In and the triazole ring seems to entail favorable pharmacokinetic properties for molecular imaging purposes.


■ INTRODUCTION
Noninvasive imaging by positron emission tomography (PET) and single photon emission computed tomography (SPECT) has become an essential technique in nuclear medicine for the diagnosis of pathological conditions such as neurodegenerative diseases, 1,2 cardiac diseases, 3 inflammatory or infectious diseases, 4 and cancer. 5For PET applications, a suitable vector molecule (small molecule, peptidomimetic, peptide, or protein) that targets the respective process of interest, e.g., a ligand that binds to a cell surface receptor, is equipped with a positron (β + ) emitting radionuclide.Although a broad range of such radionuclides with suitable nuclear properties for PET imaging exists and their production is established, 6 fluorine-18 is still one of the most frequently applied radionuclides.This originates from its sufficiently long half-life (109.8 min) that enables even multistep radiosyntheses and facilitates the transport of the final radiopharmaceutical to distant application sites within a satellite concept.Furthermore, the high percentage of β + emission (97%) and the low positron energy (E max = 0.635 MeV) favors the quality of the PET images. 7,8iven by its almost ideal nuclear properties for PET imaging, the (radio)chemical toolbox for the introduction of fluorine-18 has expanded tremendously allowing the radiofluorination of basically all kinds of target molecules from small molecules to proteins. 9Generally, 18 F-labeled radiotracers are accessed via direct labeling of their respective precursors 10 or via indirect labeling using initially prepared 18 F-labeled building blocks (prosthetic groups). 8In case of [ 18 F]fluoride as 18 F-species, direct labeling usually requires the workup of the crude aqueous solution via anion exchange and subsequent drying to provide reactive [ 18 F]fluoride for efficient labeling reactions. 11Subsequent 18 F-labeling is often performed in water-free organic solvents and at high temperatures.However, such conditions would usually not maintain the structural integrity of peptides and proteins.Furthermore, the structural complexity of these molecules bearing several functional groups lowers the labeling efficiency. 12Therefore, to enable the efficient direct labeling of such complex biomolecules, labeling strategies that rely on the formation of heteroatomand metal−[ 18 F]fluoride bond formation have been developed.In contrast to C−F bond formation, the introduction of [ 18 F]fluoride is possible even under mild conditions and in the presence of water. 12,13f these alternative 18 F-labeling strategies, the Al 18 F-labeling approach appears particularly appealing. 14,15Based on the strong binding of fluoride to several metals, McBride et al. 16 discovered the successful complexation of (Al[ 18 F]F) 2+ by common chelators including DTPA, NOTA, and NODA.−19 Similar to labeling reactions with radiometals, the Al 18 Flabeling approach offers the opportunity for radiofluorination of target molecules via lyophilized kits. 20However, efficient Al 18 F-labeling usually entails temperatures of 100 °C, which is comparable to the conditions often applied for 68 Ga-or 64 Cu- labeling.Such high temperatures are tolerated by small peptides but certainly not by proteins. 12iming at Al 18 F-labeling at lower temperatures, Cleeren et al. 21developed acyclic pentadentate ligands with a N 2 O 3 donor set.Of these new ligands, H 3 L3 (Figure 1) provided an acceptable stability of the Al 18 F-complex in vitro and bone uptake of Al[ 18 F]F-L3 and a PSMA ligand conjugate was low at 60 min p.i. Later, the same group introduced a cyclohexane moiety in their ligand scaffold, resulting in the ligand called (±)-H 3 RESCA or rac-H 3 RESCA (REStrained Complexing Agent, Figure 1) which enabled Al 18 F-labeling at very mild conditions (pH 4.6, ambient temperature) in a short period of time (12 min). 22Moreover, the stability of the Al[ 18 F]F-RESCA-complex, in particular with regards to defluorination, is sufficient for in vivo applications.The introduction into biomolecules is realized by an TFP or maleimide functionalized (±)-H 3 RESCA, 23 which are commercially available.−26 Inspired by the general suitability of the Al 18 F-RESCAmethod, we envisaged a complementary approach for the introduction of RESCA-like complexing agents into biomolecules.For this purpose, we sought to structurally modify (±)-H 3 RESCA by replacing the 2-(p-tolyl))acetic acid moiety by a propargyl moiety, which would transform into a 1,4disubstituted-1,2,3-triazole ring upon coupling to azidefunctionalized biomolecules via CuAAC (copper(I)-catalyzed azide−alkyne cycloaddition, Figure 1).Apart from this alternative functionalization strategy, we hypothesized that the triazole ring could participiate in coordination to metal ions (resulting in a N 3 O 3 donor set), which could support thus the efficient and stable complexation of other metal ions in addition to Al 3+ .1,2,3-Triazoles are known to participate as ligands in metal complexes including complexes of rhenium, technetium, copper, zinc, and platinum, 27−30 but there are also occasional reports on coordination to aluminum 31 and indium. 32To the best of our knowledge, studies on complexation of other metal ions, in particular radiometals for PET and SPECT imaging, by (±)-H 3 RESCA have not been reported so far.Furthermore, the 1,2,3-triazole ring entails an increased hydrophilicity compared to the benzyl ring in (±)-H 3 RESCA which might positively affect the pharmacokinetics of respective conjugates. 33erein, we present the synthesis of the novel cyclohexanediamine alkyne (±)-CHDA-tBu (CycloHexaneDiamine Alkyne) in which the three acetic acid moieties are tBu-protected to allow an efficient CuAAC with azide-functionalized biomolecules.(±)-CHDA-tBu was then coupled to two simple model azides (5-azido pentanoic acid and benzyl azide) to generate the CHDT (CycloHexaneDiamine Triazole)-functionalized molecules CHDT-Pe and CHDT-Bn.Al 18 F-labeling but also labeling with radiometals such as 68 Ga 3+ and 64 Cu 2+ was tested and the Al 18 F-labeled compounds were subjected to ex vivo biodistribution studies after injection into healthy mice.Subsequently, (±)-CHDA-tBu was used to prepare three novel CHDT-functionalized PSMA ligands (CHDT-PSMA-1/ 2/3, Scheme 1).PSMA (prostate specific membrane antigen) is a type II transmembrane glycoprotein with cocatalytic metallopeptidase activity 34 and emerged as attractive target for the radionuclide diagnosis and treatment of prostate cancer, 35,36 owing to its high abundance in this type of cancer. 37,38Although numerous radiolabeled PSMA ligands were developed in the past and some compounds are already in clinical use, there is still an interest in novel radioligands, in particular 18 F-labeled ligands. 39CHDT-PSMA-1/2/3 were labeled with (Al 18 F) 2+ but also 68 Ga 3+ and 111 In 3+ to demonstrate the potential of the CHDT moiety for complexing other metal ions in addition to Al 3+ .The respective nine radioligands were radiopharmacologically characterized in vitro with a focus on serum stability, saturation binding analyses and internalization behavior using monolayers and spheroids of PSMA-positive LNCaP cells.Moreover, the in vivo radiopharmacological characterization of Al 18 F-labeled CHDT-PSMA-1/2/3 was performed in LNCaP-tumor bearing mice in comparison to the well-known PSMA ligand [ 18 F]PSMA-1007.

■ RESULTS AND DISCUSSION
Synthesis of (±)-CHDA-tBu and the CHDT-Functionalized Compounds.The synthesis of (±)-CHDA-tBu was accomplished in four steps starting with racemic trans-1,2diaminocyclohexane in orientation to the procedure for the preparation of (±)-H 3 RESCA-TFP or (±)-H 3 RESCA-Mal. 40,41The first step comprised the monobenzylation by the use of benzaldehyde and sodium borohydride.This was followed by alkylation with tert-butyl bromoacetate under basic conditions (DIPEA).Subsequently, the benzyl group was removed by hydrogenolysis using Pd/C and ammonium formate.For the last step to (±)-CHDA-tBu, the propargyl group was introduced by reaction with propargyl bromide and triethyl amine.All steps proceeded in good to excellent yields and (±)-CHDA-tBu was obtained in an overall yield of 41%.The tert-butyl groups had to be kept for the introduction into biomolecules to enable an efficient coupling via CuAAC as the unprotected CHDT moiety is also able to complex Cu 2+ (see below).
For model compounds CHDT-Pe and CHDT-Bn, 5azidopentaoic acid and benzyl azide, respectively, were coupled via CuAAC.Subsequent removal of the tert-butyl groups by TFA treatment afforded both compounds in good yields of 46% and 81%, respectively, over these two steps.A similar procedure was followed for the synthesis of the PSMA ligands CHDT-PSMA-1/2/3, for which the respective tert-butyl protected and azide-functionalized precursor compounds (4− 6) were synthesized (see Chemistry section in Supporting Information).CHDT-PSMA-1/2/3 were obtained in overall yields of 22−65% for the coupling and deprotection steps.Regarding the linker entities between the glutamyl-urea-lysine (KuE) PSMA binding motif and the CHDT moiety, CHDT-PSMA-1 was designed to bear the smallest possible azide linker (apart from the use of ε-azido-norleucine).In contrast, CHDT-PSMA-3 was designed in orientation to PSMA-617 (L-2NaI and trans-4-(aminomethyl)cyclohexanecarboxylic acid (AMCH)) by substituting the DOTA chelator with the CHDT moiety and a hexanoyl linker.For CHDT-PSMA-2, D-1NaI was chosen instead of L-2NaI and AMCH was omitted.Benesova et al. 42 previously demonstrated that L-2NaI is better tolerated than L-1NaI due to a more favorable orientation within the binding pocket of PSMA.Moreover, L-2NaI is superior to D-2NaI.However, D-1NaI was not tested, which prompted us to test this particular amino acid in the context of the present study.S1 in Supporting Information).Subsequently, we also envisaged complexation of [ 68 Ga]Ga 3+ and [ 64 Cu]Cu 2+ by the CHDT moiety.To our delight, complexation of these two radiometals was indeed possible (shown for CHDT-Pe in Figure S1) under similar conditions as applied for Al 18 Flabeling.This could indicate that the triazole ring participates in binding of these radiometal species as hypothesized by us.However, it is not excluded that (±)-H 3 RESCA is also able to complex other radiometal species but such data are not reported so far.Motivated by the initial radiolabeling experiments with CHDT-Pe and CHDT-Bn, radiolabeling of the novel CHDT-functionalized PSMA ligands, CHDT-PSMA-1/2/3, was performed.In addition to (Al[ 18 F]F) 2+ , [ 68 Ga]Ga 3+ , and [ 64 Cu]Cu 2+ , we also tried the complexation of [ 111 In]In 3+ .Complexation at 40 °C and a precursor amount of 10 μg (7.5, 8.3, and 10.5 nmol for CHDT-PSMA-1/2/3, respectively) resulted in high radiochemical conversions (>95%) for all radiometal species (Figure S2 in Supporting Information).Overall, we demonstrated that the transformation of the original (±)-H 3 RESCA ligand to the novel CHDT moiety still allows Al 18 F-labeling, but also labeling with [ 68 Ga]Ga 3+ , [ 64 Cu]Cu 2+ , and [ 111 In]In 3+ .

Complexation of (Al[
While the 68 Ga/ 64 Cu/ 111 In-labeled compounds were not further processed for their radiopharmacological characterization, residual unbound 18 F-species were removed from the reaction mixture by the addition of hydroxyapatite.It is worth noting, that the treatment with hydroxyapatite should be conducted rather short (30 s) as it not only binds unbound 18 F-species but leads also to decomplexation of the Al 18 Fcomplex, which ultimately lowers the radiochemical yield and apparent molar activity.Alternatively, unbound 18 F-species could also be removed by solid-phase extraction using suitable cartridges.This processing would also improve radiochemical yield and molar activity values.However, herein we decided to follow the hydroxyapatite treatment to ensure [ 18 F]fluoride (and Al 18 F species)-free radioligand formulations.For the radiolabeled CHDT-PSMA-1/2/3, radiochemical purities of >95% (Figure S3 in Supporting Information) and apparent molar activities of 10 (±1, Al 18 F), 24 (±3, 68 Ga), and 24 (±5, 111 In) MBq/nmol, respectively (mean ± SD values over all compounds), were achieved.
Distribution Coefficients (logD 7.4 ) and Stability Studies.The distribution coefficients logD 7.4 of the Al 18 F-, 68 Ga-, and 111 In-labeled PSMA conjugates were determined and are summarized in Table 1.Among the three PSMA conjugates, radiolabeled CHDT-PSMA-1 is the most hydrophilic compound, irrespective of the radionuclide, which appears reasonable due to the short azido alkyl linker and the absence of a naphthylalanine residue compared to the other two compounds.
Stability studies of the Al 18 F-, 68 Ga-, and 111 In-labeled PSMA conjugates were conducted in human serum and the percentages of intact radioligand are summarized in Table 1.Analysis of serum samples was performed by radio-TLC and radio-HPLC to correctly assess disintegration of the radiolabeled molecules caused by released radiometal species and metabolization.All compounds showed an excellent stability over 3 h with values for residual intact radioligand >96%.Moreover, the 111 In-labeled compounds appeared also to be stable even after a prolonged incubation period of 24 and 48 h (Table 1).These results illustrate that the novel CHDT moiety provides a high kinetic inertness of the respective Al 18 F-, 68 Ga-, and 111 In-complexes under physiological conditions in vitro.
Although [ 64 Cu]Cu-complexation by the CHDT moiety is also possible, the kinetic inertness of the resulting complex appears rather low.This was exemplarily demonstrated for [ 64 Cu]Cu-CHDT-PSMA-1 in a protein challenge experiments using human serum and analysis by radio-SDS-PAGE (Figure S4 in Supporting Information).There was a significant transchelation of [ 64 Cu]Cu 2+ to albumin (20%), while other copper chelators such as TETA, DOTA, NOTA, cyclam, or diamSar show only minimal 64 Cu-transchelation under the same conditions. 43In contrast, for [ 68 Ga]Ga-CHDT-PSMA-1 under similar conditions no transchelation of [ 68 Ga]Ga 3+ in human serum to transferrin, which has a high affinity for Ga 3+ , 44 was observed (Figure S5 in Supporting Information), which highlights the higher kinetic inertness of the 68 Gacomplex compared to the 64 Cu-complex.In this context, Ga 3+ and In 3+ can be classified as hard acidic cations according to the Pearson ' s hard−soft acid−base theory and favor hard donor atoms (e.g., anionic oxygen), while Cu 2+ is rather a borderline acid and favors soft donor atoms (e.g., nitrogen). 45onsequently, it appears comprehensible that the acyclic CHDT moiety with N 2 O 3 donor set (without triazole ring) is sufficient for stable complexation of Ga 3+ and In 3+ .In contrast, although the 1,2,3-triazole was intended to provide an additional nitrogen donor for metal complexation, this coordination (if coordination occurs at all) is not sufficient for stably complexing Cu 2+ .It is worth noting that previously a DTPA analog bearing a trans-1,2-diaminocyclohexane backbone (CHX-A″) has been developed for complexation of 212 Bi, 46,47 but this chelator is also able to complex a series of other radiometals including 177 Lu, 86 Y, 68 Ga, and 111 In. 45,48,49 Furthermore, an EDTA analog bearing a trans-1,2-diaminocyclohexane backbone (4-ICE) allows stable complexation of radiometals, such as 111 In, 57 Co, and 47 Sc, but produced a less stable complexation of 67 Cu as seen by a higher activity uptake in the liver of a respective antibody conjugate compared to conjugates bearing other chelators. 50In this context, it is known that a strong binding of copper ions is favored by macrocyclic ligands due to the so-called macrocyclic effect, 51 which adds to the above-mentioned HSAB concept for explaining the low kinetic inertness of the [ 64 Cu]Cu-CHDT complex.
Characterization of PSMA Binding.Motivated by the stability results for Al 18 F, 68 Ga, and 111 In-labeled CHDT-PSMA-1/2/3, we envisaged their further radiopharmacological characterization.First, the binding affinities of the novel compounds and three reference ligands (PSMA-1007, PSMA-617, and KuE (lysine-urea-glutamate, see Chemistry section in Supporting Information)) to PSMA were determined in a competition binding assay using LNCaP cell homogenates and [ 177 Lu]Lu-PSMA-617 as radioligand (Figure 2 and Table 2).
In this context, LNCaP cells are known for their high amount of PSMA, 52 which was also herein confirmed by Western blotting (Figure S6 in Supporting Information).While PSMA-1007, PSMA-617, and CHDT-PSMA-3 show binding affinities by means of K i values in the single-digit nanomolar range, CHDT-PSMA-1/2 exhibit significantly lower affinities (49 and 115 nM).The comparable binding affinity of CHDT-PSMA-3 and PSMA-617 appears reasonable due to the high structure similarity, in particular the L-2-NaI and AMCH moieties.Similarly, the lower binding affinity of CHDT-PSMA-2 can be rationalized due to the unfavorable orientation of the naphthyl moiety in 1-NaI within the PSMA binding site, 42 which is obviously independent of the configuration (L or D) of the 1naphthylalanine residue.CHDT-PSMA-1 exhibits an even lower binding affinity than CHDT-PSMA-2 indicating that the attachment of the CHDT moiety close to the KuE binding motif is detrimental to the binding to PSMA.In this context, although other PSMA ligands with a chelator moiety close to the KuE binding motif and a high binding affinity are known, e.g., PSMA-11 (6-aminohexanoyl linker between KuE and HBED-CC), the binding affinity seems to depend on the identity of the chelator moiety (K i values of 12 and 37.6 nM for PSMA-11 and its DOTA analog, respectively). 53The lowest binding affinity was observed for the unmodified lysineurea-motif (KuE), which agrees with the original report for this PSMA binding motif by Maresca et al. 54 In addition to assessing the binding affinities for nonlabeled CHDT-PSMA-1/2/3, we sought to characterize the saturation binding of the Al 18 F-, 68 Ga-, and 111 In-labeled compounds using LNCaP cells cultured in monolayers and as spheroids.The saturation binding curves for Al[ 18 F]F-CHDT-PSMA-1/ 2/3 are depicted in Figure 3, while the graphs for [ 18 F]PSMA-1007 and 68 Ga-and 111 In-labeled CHDT-PSMA-1/2/3 can be found in Figures S7−S9 in the Supporting Information.The binding parameters (K d and B max ) for Al[ 18 F]F-CHDT-PSMA-1/2/3 and [ 18 F]PSMA-1007 are summarized in Table 3 (for the 68 Ga-and 111 In-labeled compounds see Tables S4 and S5 in the Supporting Information).For the saturation binding assays, we recorded surface-bound radioligand (named "binding"), which refers to bound radioligand that can be removed by acid wash, and internalized or acid-resistant radioligand (named "internalization"), which refers to radioligand that remains cell-bound after acid wash.Consequently, two saturation binding plots were obtained and two sets of binding parameters (K d /B max and K d,int /B max,int ) were derived.
Although the binding affinities of nonlabeled CHDT-PSMA-1/2 were almost 1 order of magnitude lower than that of CHDT-PSMA-3, the K d values (cell surface binding) of the radiolabeled analogs, independent of the particular radiolabel (Al 18 F, 68 Ga, 111 In), were comparable (factor < 4).For example, the K d values for the Al 18 F-labeled compounds were 26.4,12.9, and 19.7 nM, respectively, using LNCaP cells as monolayer.This might indicate in case of CHDT-PSMA-1/2 that the CHDT moiety as complex with one of the abovementioned radiometal species is better tolerated than the nonlabeled moiety.The most obvious difference after complexation is the overall charge of this moiety, with three  The results for the cell surface binding of the radioligands are largely conserved when analyzing the saturation binding data for the internalized fraction.This demonstrates that internalization is also concentration-dependent and that a  constant fraction of around 50% of total-bound radioligand is internalized after 2 h (as B max and B max,int values are comparable).In contrast to the pronounced binding of all radioligands to LNCaP cells, binding to PSMA-negative PC3 cells was negligible (exemplarily shown for Al[ 18 F]F-CHDT-PSMA-1 in Figure S10 in Supporting Information), which further confirms their PSMA-specificity.
The results obtained for saturation binding using LNCaP cells cultured in monolayers are also conserved when using LNCaP spheroids.However, the K d values were consistently higher and B max values lower when comparing the data sets obtained for monolayer and spheroids.When the cells form a spheroid, the environment of the individual cells changes.−57 For LNCaP cells, abundance of PSMA was previously shown to be conserved in spheroids compared to monolayers. 58herefore, regarding the somewhat reduced binding capacity toward spheroids observed herein originates most likely from a limited permeation potency of the radioligands into the spheroids and thus, the apparent amount of binding sites is reduced.
In addition to the saturation binding analyses of cell surfacebound and internalized radioligand, the time-dependent internalization of the radiolabeled CHDT-PSMA-1/2/3 was investigated (10 min, 1 h, and 2 h).Specific internalization of the Al 18 F-labeled CHDT-PSMA-1/2/3 and [ 18 F]PSMA-1007 are shown in Figure 4 (the data for the 68 Ga-and 111 In-labeled compounds are shown in Figure S11 in the Supporting  S8 in the Supporting Information for the corresponding data), each performed in triplicate or quadruplicate.Same color coding as in Figure 3. Information).Internalization increased for all radiolabeled compounds over time with the highest values among the novel PSMA ligands obtained for CHDT-PSMA-3, irrespective of the particular radiolabel.The internalization of Al[ 18 F]F-CHDT-PSMA-3 was even comparable to that of [ 18 F]PSMA-1007.In contrast, radiolabeled CHDT-PSMA-1/2 exhibited a lower internalization.In agreement with the aforementioned high B max values for the Al 18 F-labeled compounds, internalization was consistently higher for the Al 18 F-labeled CHDT-PSMA-1/2/3 compared to their 68 Ga-and 111 In-labeled counterparts.
−61 Previously, Matthias et al. 62 demonstrated for fluorophore-labeled PSMA inhibitors and by using stimulated emission depletion (STED) nanoscopy that PSMA internalization upon inhibitor binding proceeds also via clathrin-dependent endocytosis.Herein, we also characterized the time-dependent internalization of the radiolabeled compounds in the presence of the methylated cyclic oligosaccharide methyl-β-cyclodextrin (MβCD), 63,64 a known endocytosis inhibitor.Treatment with MβCD was associated with a significantly reduction in internalization for all PSMA radioligands examined herein, which indicates that the novel radioligands induce upon binding PSMA internalization via endocytosis.
Ex Vivo Biodistribution and PET Imaging.Due to the original use of (±)-H 3 RESCA for complexation of (Al[ 18 F]-F) 2+ , we sought to subject the Al 18 F-labeled compounds developed herein bearing the novel CHDT moiety to a detailed in vivo radiopharmacological characterization.First, the biodistribution of the two radiolabeled model compounds, Al[ 18 F]F-CHDT-Pe and Al[ 18 F]F-CHDT-Bn, was studied at 60 and 240 min p.i. in healthy SKH1 mice to primarily assess the activity uptake in bone tissue and thus to get information about the in vivo stability of the Al[ 18 F]F-CHDT complex (Tables S9 and S10 in the Supporting Information).While Al[ 18 F]F-CHDT-Pe was mainly excreted into the urine (79% ID in urine at 60 min p.i.), the benzyl analog was predominantly excreted via the hepatobiliary route (51% ID in intestine at 60 min p.i.).This different in vivo behavior is reasonable considering the structures of both compounds with the aliphatic carboxylic acid moiety likely mediating the renal excretion pathway of Al[ 18 F]F-CHDT-Pe.Activity uptake in bone tissue was assessed on the basis of the femur.Values of around 0.4%ID/g were measured, which did not increase up to 240 min p.i.This suggests a sufficient stability of Al[ 18 F]F-CHDT complex in vivo regarding the release of [ 18 F]F − or related species that would accumulate permanently in the bone tissue.
We furthermore investigated the biodistribution and suitability for targeting tumor-associated PSMA of Al 18 Flabeled CHDT-PSMA-1/2/3 by PET imaging up to 120 min p.i.For this purpose, the PSMA-positive LNCaP tumor xenograft was used.Representative PET images at selected time points p.i. are shown in Figure 5, time-activity curves and total tissue uptakes based on PET images are illustrated in Figure 6.For comparison, [ 18 F]PSMA-1007 was also examined in this tumor model.For all radiofluorinated PSMA ligands, the LNCaP tumor was clearly visible and PSMA-specific uptake was confirmed by coadministration of 2-(phosphonomethyl)pentanedioic acid) (2-PMPA), which significantly reduced the tumor uptake (Figure 6C).Total tumor uptake was comparable for Al[ 18 F]F-CHDT-PSMA-1/ 3 and [ 18 F]PSMA-1007, while uptake of Al[ 18 F]F-CHDT-PSMA-2 was significantly lower (Figure 6A and C).Considering rather comparable binding affinities of all novel Al 18 F-labeled PSMA ligands (Table 3), the inferior tumor targeting capability of Al[ 18 F]F-CHDT-PSMA-2 might primarily be a consequence of its altered pharmacokinetic behavior.[ 18 F]PSMA-1007 and Al[ 18 F]F-CHDT-PSMA-1/3 were mainly renally excreted, while Al[ 18 F]F-CHDT-PSMA-2 exhibited a higher uptake in the organs associated with hepatobiliary excretion compared to kidneys and bladder (Figure 6B and D).Apparently, the presence of the naphthylalanine residue shifts the excretion route to hepatobiliary excretion, which is partly compensated for Al[ 18 F]F-CHDT-PSMA-3 by the AMCH moiety.Although total uptake in bone tissue was low for all compounds (Figure 6C), the average SUVmean values at 120 min p.i. appeared to be slightly higher for the CHDT-functionalized PSMA ligands (0.30−0.51) compared to [ 18 F]PSMA-1007 (0.21, Figure 6A).This is in agreement with the results from ex vivo biodistribution studies of the two radiofluorinated model compounds and further supports our view that the Al[ 18 F]F-CHDT complex is sufficiently stable in vivo.
A more detailed view on the biodistribution and tumor uptake of Al[ 18 F]F-CHDT-PSMA-1 reveals that its in vivo performance is at least comparable to that of [ 18 F]PSMA-1007.In fact, the tumor and kidney uptakes were similar but blood clearance was faster, which resulted in significantly higher tumor-to-blood and tumor-to-muscle ratios at 120 min p.i. (Table 4).The good in vivo performance of Al[ 18 F]F-CHDT-PSMA-1 is also striking considering the results of previously developed PSMA ligands by Cleeren et al. bearing acyclic pentadentate chelating units, i.e., Glu-urea-Lys(Ahx)-L3 and PSMA-RESCA1 (Figure 7). 21,65A PSMA-specific tumor uptake was also seen for both Al 18 F-labeled ligands, but Al[ 18 F]F-PSMA-RESCA1 underlay a significant hepatobiliary excretion, which is less favorable for imaging purposes.Consequently, our objective of making the chelating unit more hydrophilic by the introduced triazole ring seems to be successful, although we should mention that the Ahx linker present in PSMA-RESCA1 might also add a certain hydrophobicity.In contrast, Al[ 18 F]F-Glu-urea-Lys(Ahx)L3 showed a more pronounced uptake in bone tissue.Overall, our data for Al[ 18 F]F-CHDT-PSMA-1 indicate that this radioligand is an interesting candidate for clinical translation.

■ CONCLUSION
Previously reported acyclic pentadentate chelating units including (±)-H 3 RESCA turned out to facilitate efficient Al 18 F-labeling at low temperatures with sufficient complex stabilities for in vivo applications.Based on (±)-H 3 RESCA, our newly developed alkyne-functionalized building block allows the straightforward functionalization of biomolecules via click chemistry and simultaneously forms the actual complexing unit with a triazole ring (CHDT) upon conjugation.Apart from its conserved suitability to complex (Al[ 18 F]F) 2+ , we showed that the CHDT moiety also enables an efficient complexation of [ 68 Ga]Ga 3+ and [ 111 In]In 3+ and that the respective complexes exhibit a high kinetic inertness at least under physiological conditions in vitro.Furthermore, the preclinical characterization of the Al 18 F-labeled CHDTfunctionalized PSMA ligands revealed that the triazole ring ■ EXPERIMENTAL SECTION General.For the syntheses of the PSMA conjugates and model compounds all chemicals were obtained by commercial suppliers and used without further purification.Solvents were obtained by Fisher Scientific and anhydrous solvents were supplied by Sigma-Aldrich.PSMA-617 was purchased from ABX (Germany).KuE was prepared as previously described. 66,67Nuclear magnetic resonance spectra were recorded on an Agilent Technologies 400 MR spectrometer consisting of 400/54 premium compact magnet, 400 MR console and 400 MHz OneNMRProbe PT probe head (400 MHz for 1 H, 101 MHz for 13 C and 376 MHz for 19 F).Spectra were processed by using the program MestreNova (version 14.2.1-27684).NMR chemical shifts were referenced to the residual solvent resonances relative to tetramethylsilane (TMS; 1 H and 13 C) and trichlorofluoromethane (CFCl 3 ; 19 F).Mass spectra (ESI) were obtained on a Waters Xevo TQ-S mass spectrometer (driven by the Mass Lynx software) or an Advion ExpressIon CMS spectrometer.The following molar masses of the final compounds were used for calculations (all molar masses were calculated including two TFA molecules): CHDT-Pe (697.53 g/mol), CHDT-Bn (687.54 g/mol), CHDT-PSMA-1 (956.75 g/mol), CHDT-PSMA-2 (1210.10g/mol, CHDT-PSMA-3 (1335.27g/mol).
Radiolabeling. [ 18 F]fluoride (no-carrier-added) was produced on a cyclotron (30 MeV TR-Flex-Cyclotron, Advanced Cyclotron Systems Inc., Canada) by irradiation of [ 18 O]H 2 O via the 18 O(p,n) 18 F nuclear reaction.An aliquot of the aqueous [ 18 F]fluoride solution (600 MBq) was withdrawn and sodium acetate buffer (0.1 M, pH 4.0) was added to reach a volume of 300 μL.For formation of (Al[ 18 F]F) 2+ , 10 μL of AlCl 3 (2 mM, in 0.1 M sodium acetate, pH 4) was added and the mixture was incubated for 10 min at 22 °C.Subsequently, 10 μL of the respective CHDT-functionalized compound (stock of 1 μg/ μL) were added and the mixture was incubated for 20 min at 40 °C (300 rpm, Thermomixer Comfort, Eppendorf, Germany).For processing after successful labeling, a few crumbs of hydroxyapatite were added to the mixture.After 30 s, the suspension was shortly centrifuged (10 s at 10000g, Eppendorf 5415R), and the supernatant was transferred to a clean vial.[ 18 F]PSMA-1007 was prepared according to a protocol from Cardinale et al. 69 68 Ga was eluted as [ 68 Ga]GaCl 3 from a 68 Ge/ 68 Ga generator of iThemba Laboratories (South Africa) in 0.6 M HCl.To 300 MBq of [ 68 Ga]GaCl 3 , 1 M MES buffer (pH 6.0) was added until a pH of 5 to 6 was reached.Subsequently, 10 μL of the respective CHDT-functionalized compound (stock of 1 μg/ μL) were added and the mixture was incubated for 20 min at 40 °C. 111In was purchased as [ 111 In]InCl 3 from Curium (United Kingdom).To 50 MBq of [ 111 In]InCl 3 , 390 μL MES buffer (0.1 M, pH 5.5) and 10 μL of the respective CHDTfunctionalized compound (stock of 1 μg/μL) were added and the mixture was incubated for 20 min at 40 °C.
n-Octanol/PBS Distribution Coefficient (logD 7.4 Value).The log D 7.4 values were determined as follows: 500 μL of n-octanol and 450 μL of PBS (pH 7.4) were premixed prior to the addition of 50 μL of the respective radiolabeled compound.The mixture was vigorously stirred for 30 min at 22 °C, and thereafter centrifuged for 5 min at 7,500g at 22 °C.Aliquots of the n-octanol and the aqueous phases were transferred to measuring tubes, and the activity counts in both phases were measured in a gamma counter (Hidex Deutschland Vertrieb GmbH, Germany).The logarithm of the quotient of the activity counts from the n-octanol and the aqueous phase was calculated, which is equal to the log D 7.4 value.
Serum Stability Assay.To investigate the serum stability of the radiolabeled CHDTA-PSMA-1/2/3, human serum (frozen, from male AB clotted whole blood; Sigma-Aldrich/ Merck) was centrifuged for 5 min at 20,000g at 4 °C after thawing.The supernatant was sterile filtrated (filter pore 0.2 μm).Four parts serum and one part of radioligand were incubated at 37 °C with shaking.Aliquots were taken after 3 h and for the 111 In-labeled compounds also after 24 and 48 h, and residual intact radiolabeled compound was determined by radio-TLC and/or radio-HPLC.For analysis by radio-HPLC, the aliquots were mixed with Supersol (mixture for protein precipitation that consists of 20% ethanol, 0.5% Triton X-100, 5 mM EDTA and 0.1% saponin) 72 and kept on ice for 5 min.After centrifugation for 5 min at 20,000g at 4 °C, the supernatant was used for radio-HPLC analysis.
Biological In Vitro Assays.Binding assays were conducted using the high PSMA synthesizing human prostate carcinoma cell line LNCaP (ATCC CRL-1740) and the PSMA negative cell line PC3 (ATCC CRL-1435).The cells were grown as monolayer at 37 °C in a humidified atmosphere comprising 5% CO 2 and 95% air in RPMI medium including 10% FCS (Biochrom AG).After washing the confluent cells twice with PBS and detaching them with trypsin/EDTA (0.05%/0.02%), the cells were suspended in RPMI and counted (Casy TT, Omni Life Science, Germany).
For the formation of spheroids, the LNCaP cells were incubated overnight in a normal culture bottle with biocompatible nanoparticles (1 × 10 −6 cells with 80 μL of a suspension consisting of Au, Fe 2 O 3 and poly-L-lysine called Nanoshuttle, Greiner Bio-One GmbH, Germany).After harvesting and cell counting, 10,000 cells/well of a 24-well plate or 7,000 cells/well of a 96-well plate (cell-repellent plates) were sown.The well plates were placed on matching plates with button magnets (Greiner Bio-One GmbH, Germany).The magnetized LNCaP cells formed suitable spheroids that were used after 5 d.
In order to assess the PSMA synthesis of LNCaP and PC3 cells, Western blotting was carried out.Cell lysates were prepared with RIPA buffer and 10 μg (protein level determined by Bio-Rad protein assay) was separated by SDS-PAGE (12% separating gel).This was followed by blotting onto a polyvinylidene fluoride membrane and blocking with 5% bovine serum albumin (BSA) in phosphate-buffered saline +0.1% Tween 20 (PBST) for 1 h, followed by an 1 hincubation with appropriate primary antibodies (rabbit anti-PSMA 1:5,000; rabbit anti-β-actin 1:5,000, Cell Signaling Technology, USA) diluted with 5% BSA in PBST.After washing three times in PBST (à5 min), the blots were incubated with a horseradish peroxidase (HRP)-conjugated secondary antibody (antirabbit-HRP 1:20,000, Cell signaling Technology Inc., Danvers, USA) for 1 h in 5% BSA-TBST.The visualization of the HRPenzyme activity was accomplished with Western blotting luminol reagent according to manufacturer's instructions (Santa Cruz Biotechnology, Dallas, TX, USA).Enhanced chemiluminescence was detected with a blot scanner (C-DiGit Blot Scanner, LI-COR Bioscience GmbH, Germany).
Competition Binding Assay.The binding affinity of the nonlabeled CHDT-PSMA-1/2/3, PSMA-1007, PSMA-617, and KuE was assessed in competition binding assays using LNCaP cell homogenates (∼1 × 10 6 cells/mL using Potter-Elvehjem) and the radioligand [ 177  The incubation was stopped by washing the homogenate with cold PBS four times using a filter (Whatman GF/C, 90 min presoaked in 0.3% polyethylenimine) in a cell harvester (Brandel, USA).The radioactivity bound to the filter was measured in a gamma counter.Filter binding was determined for adjacent samples without cell homogenate.The experiments were performed two to four times in triplicates or quadruplicates.Inhibitory constants (K i ) were derived from nonlinear regression according to the model of "One site − Fit Ki" as implemented in GraphPad Prism 10.The K d value of [ 177 Lu]Lu-PSMA-617 was determined in saturation binding assays using LNCaP cell homogenates to be 15.2 (±2.1) nM.
For saturation assays in well plates, monolayers of LNCaP (sowing 5 × 10 4 cells/well in 48-well plates 2 days before the assay), as well as LNCaP spheroids were incubated with radiolabeled CHDT-PSMA-1/2/3 and PSMA-1007 (0.2 nM to 100 nM) with adjacent samples receiving KuE (0.8 mM) at 37 °C for 2 h.The number of experiments can be derived from the data in the Tables S4, S5, and S7 in the Supporting Information.After removing the incubation medium and washing of cells with ice-cold PBS (containing Mg 2+ and Ca 2+ ) the surface-bound activity was stripped with 4 °C cold acid-wash buffer (0.2 M glycine, 0.15 M NaCl, pH 3.0) for 10 min.The acid wash buffer was transferred from the wells to measuring tubes, as was the PBS buffer after washing once.Cytosolic activity (internalized radioligand) was determined after treatment with cell lysis buffer (1% SDS in 0.1 M NaOH) at 37 °C for 30 min.Cell surface and cytosolic activity were measured separately in a gamma counter.Time-depending internalization assays were performed only with cells in monolayer.The incubation (radioligand concentration 13 nM) for these assays was terminated after 10 min, 1 and 3 h, with 111 In and 177 Lu additionally after 24 and 48 h at 37 °C, respectively.During the internalization assays adjacent samples received KuE (0.8 mM) for PSMA blocking, methyl-βcyclodextrin (MβCD; 3 mM) or KuE together with MβCD, 1 h before radioligand application.The number of experiments can be derived from the data of Figure S11 and associated table and from Table S8 in the Supporting Information.
Dissociations constants (K d and K d,int ), as well as maximal binding capacities (B max and B max,int ), were derived from specific binding data by nonlinear regression according to the model of "One site-specific binding" as implemented in GraphPad Prism 10.Protein content was determined using the bicinchoninic acid protein assay kit (Pierce, Thermo Scientific) to calculate the binding capacities in pmol/mg and for the time-depending internalization the internalized percent of administrated dose per μg protein (% AD/μg protein).
Small Animal PET/CT Imaging.Tumor Xenograft Model.All animal experiments were performed in accordance with the German Animal Welfare regulations and were approved by the local ethics committee for animal experiments (DD24.1-51-31/449/49,Oct 26, 2018).A PSMA-positive prostate cancer xenograft model was generated by subcutaneous injection of human LNCaP cells into the right shoulder of 8−12 week-old male nude mice (Rj:NMRI-Foxn1 nu/nu , Janvier Laboratories, Le Genest-Saint-Isle, France).Imaging studies were performed when subcutaneous tumors reached a diameter of at least 6 mm.General anesthesia of the animals was induced and maintained by inhalation of 10% (v/v) desflurane in 30/70% (v/v) oxygen/air.Animals were continuously warmed at 37 °C during anesthesia.

Figure 2 .
Figure 2. Competition binding of CHDT-PSMA-1/2/3 compared to known PSMA ligands using LNCaP cell homogenates [ 177 Lu]Lu-PSMA-617 was used as radioligand (12 nM).Data shown are mean values (±SEM) of 2−8 separate experiments (see Experimental Section for details and TableS2in the Supporting Information for the corresponding data), each performed in triplicate or quadruplicate.

Figure 4 .
Figure 4. Internalization behavior of the Al 18 F-labeled CHDT-PSMA-1/2/3 (A−C) and [ 18 F]PSMA-1007 (D) Time-depending internalization using LNCaP cells as monolayer with specific internalization and specific internalization in the presence of MβCD (3 mM).Data shown are mean values (±SEM) of 3 or 7 separate experiments (see TableS8in the Supporting Information for the corresponding data), each performed in triplicate or quadruplicate.Same color coding as in Figure3.

Table 1 .
Summary of logD7.4ValuesandSerumStability of the Radiolabeled PSMA Ligands a Data shown are mean values (±SEM) of three experiments ( # two experiments).Each experiment was performed in triplicate.Percentage of intact111In-labeled conjugates after 24 h (*) and after 48 h (**).Percentage of intact radioligand was assessed by radio-TLC using iTLC-SG strips as stationary phase and 2 M ammonium acetate/methanol (1:1, v/v) as eluent.The data of the individual experiments are listed in TableS1in Supporting Information. a

Table 2 .
K i Values of CHDT-PSMA-1/2/3 Compared to Known PSMA Ligands Determined from Competition Binding Assays a moiety in CHDT-PSMA-3 compared to the other two analogs might separate the CHDT moiety too far away from the PSMA binding site to exert any interactions with PSMA.Consequently, the complexation state of CHDT is not as crucial for the binding affinity as for the other two compounds.While there is no clear trend for the K d values in dependence on the particular radiolabel, the B max values were consistently higher d value (5.4 nM) but a comparable B max value (1.11 pmol/mg) compared to the Al 18 F-labeled CHDT-PSMA-1/2/3.

Table 3 .
Parameters for Extracellular and Intracellular Saturation Binding of Al[ 18 F]F-CHDT-PSMA-1/2/3 and [ 18 F]PSMA-1007 a Data shown are mean values (±SEM) of 2−4 separate experiments (see TableS7in the Supporting Information for the corresponding data).n.d.denotes not determined, M denotes monolayer, and S denotes spheroids. a

Table 4 .
65,65-to-Organ Contrast in PET Imaging with Al18F-Labeled CHDT-PSMA-1/2/3 Compared to [ 18 F]PSMA-1007 a Structures of PSMA ligands Glu-urea-Lys(Ahx)L321,65and PSMA-RESCA165might also positively affect the in vivo behavior of small molecule or peptide-derived conjugates.In particular, the in vivo performance of Al[ 18 F]F-CHDT-PSMA-1 is even comparable to that of [ 18 F]PSMA-1007 rendering this novel PSMA ligand a promising candidate for clinical translation.Moreover, the activity uptake in bone tissue was really low and highlights the high stability of the Al[ 18 F]F-CHDT-complex in vivo.For the [ 68 Ga]Ga-and [ 111 In]In-CHDT-complexes, the kinetic inertness in vivo needs to be further investigated.