Synthesis of DOTA-pyridine chelates for 64Cu coordination and radiolabeling of αMSH peptide

Background 64Cu is one of the few radioisotopes that can be used for both imaging and therapy, enabling theranostics with identical chemical composition. Development of stable chelators is essential to harness the potential of this isotope, challenged by the presence of endogenous copper chelators. Pyridyl type chelators show good coordination ability with copper, prompting the present study of a series of chelates DOTA-xPy (x = 1–4) that sequentially substitute carboxyl moieties with pyridyl moieties on a DOTA backbone. Results We found that the presence of pyridyl groups significantly increases 64Cu labeling conversion yield, with DOTA-2Py, −3Py and -4Py quantitatively complexing 64Cu at room temperature within 5 min (1 × 10− 4 M). [64Cu]Cu-DOTA-xPy (x = 2–4) exhibited good stability in human serum up to 24 h. When challenged with 1000 eq. of NOTA, no transmetallation was observed for all three 64Cu complexes. DOTA-xPy (x = 1–3) were conjugated to a cyclized α-melanocyte-stimulating hormone (αMSH) peptide by using one of the pendant carboxyl groups as a bifunctional handle. [64Cu]Cu-DOTA-xPy-αMSH retained good serum stability (> 96% in 24 h) and showed high binding affinity (Ki = 2.1–3.7 nM) towards the melanocortin 1 receptor. Conclusion DOTA-xPy (x = 1–3) are promising chelators for 64Cu. Further in vivo evaluation is necessary to assess the full potential of these chelators as a tool to enable further theranostic radiopharmaceutical development. Supplementary Information The online version contains supplementary material available at 10.1186/s41181-020-00119-4.


Yang et al. EJNMMI Radiopharmacy and Chemistry
(2021) 6:3 Page 3 of 16 Zhang et al. 2019;Zhang et al. 2018;Zhang et al. 2017). In all cases, tracers achieved good tumor visualization in PET images with excellent tumor-to-normal tissue ratios. Moreover, we recently evaluated 225 Ac labeled αMSH derivatives, which also exhibited excellent tumor-to-normal tissue ratios (Ramogida et al. 2019;Yang et al. 2020). Knowing the potential of tumor accumulation and internalization, and low uptake in normal tissues, we were interested in labeling an αMSH peptide with 64 Cu to assess the unique theranostic power of 64 Cu. [ 64 Cu]Cu-αMSH peptide will allow for real-time monitoring of physiological processes during therapy studies, and when coupled with the longer half-life of 64 Cu over other PET isotopes (i.e. 68 Ga and 18 F), it will enable later time points for imaging and biodistribution studies, and a more accurate assessment of the longer-term biological fate of both peptide and radionuclide.

Results and discussions
The chelator synthesis started with the cyclen-xPy (x = 1-4, compounds 1-4, Scheme 1). All derivatives were synthesized using a single reaction followed by purification by HPLC. Each compound was then reacted with tert-butyl bromoacetate and deprotected to yield the final product DOTA-xPy (x = 1-3, compounds 5-7) (Subat and König 2001;Veiga et al. 2013). Compounds cyclen-1Py (Aime et al. 1999;Aoki et al. 2004;Subat et al. 2007), cyclen-2Py (El Hajj et al. 2009), cyclen-4Py (Natrajan et al. 2010) and DOTA-1Py (Aime et al. 1999) were previously reported, while compounds cyclen-3Py, DOTA-2Py and DOTA-3Py are new compounds to the best of our knowledge. DOTA-1Py and DOTA-3Py were obtained as a single compound. DOTA-2Py could potentially maintain both cisand trans-stereoisomer configurations. Because 1 H NMR of cyclen-2Py indicated a cisderivative, DOTA-2Py was assumed to be cis-. The complexation of nat Cu to DOTA-xPy (x = 1-3) was confirmed by IR and UV-vis spectra ( Fig. S1 and S2 in Supporting Information). In the UV-vis spectra, we observed increased absorbance between 200 and 400 nm, with a new shoulder peak emerging at around 290 nm, suggesting metal complexation formation. A new peak at 720-740 nm and the blue color of the product indicated the formation of Cu(II). Also, for all three chelates, UV-vis spectra showed saturation at 1.2 eq. Cu(II), suggesting the complex has 1:1 Cu(II):chelate ratio. Infrared (IR) analysis showed the decrease of the stretching absorption at 1700 cm − 1 for C=O and 1200 cm − 1 for C-N (Spreckelmeyer et al. 2017), indicating that both COO and pyridyl moieties bind to the metal. In attempting to interpret the coordination geometry, we considered both the relative frequency and intensities of both asymmetric (ν as ,~1600 cm − 1 ) and symmetric (ν s ,1 400 cm − 1 ) COO stretching vibrations for the free and bound chelates (Berestova et al. 2017). For the DOTA-1Py and -2Py chelates, we observed the evolution of dual maxima upon metal binding, suggesting there exist multiple binding configurations for the COO moieties around the metal. That said, the spectroscopic data did not allow for the elucidation of a definitive orientation of binding motifs around the metal. Future studies will focus on a full structural analysis to aid in determining the coordination behavior for these chelates. The labeled peptide [ 64 Cu]Cu-DOTA-2Py-αMSH showed a double peak on HPLC gamma trace under isocratic conditions, which could indicate the formation of stereoisomers upon metal coordination. 64 Cu was produced by proton irradiation of enriched 64 Ni solid metal targets using TRIUMF's TR13 cyclotron. After acid dissolution and purification by cation exchange Yang et al. EJNMMI Radiopharmacy and Chemistry (2021)  resin (AG1-X8), 64 Cu was obtained in dilute HCl. 64 Cu labeling of DOTA, DOTA-1Py, DOTA-2Py, DOTA-3Py, and cylen-4Py were examined at various ligand concentrations. The results ( Fig. 2 and Fig. 3) indicated that the subsequent addition of pyridine moiety served to enhance 64 Cu labeling conversion in general. While DOTA showed little complexation with 64 Cu at ambient temperature even at ligand concentrations of 1 × 10 − 3 M (100 nmol ligand, 3 MBq 64 Cu), having one, two or three pyridyls significantly improved radiochemical conversion (RCC), with DOTA-3Py demonstrating efficient complexation of 64 Cu (RCC close to 95%) at ligand concentrations of 1 × 10 − 5 M (1 nmol ligand, 3 MBq 64 Cu) and at ambient temperature. The differences in 64 Cu labeling performance between DOTA-2Py and -4Py were not significant. We then studied the reaction kinetics for labeling DOTA-2Py, −3Py and -4Py at 1 × 10 − 4 M (10 nmol ligand, 3 MBq 64 Cu, DOTA-1Py not studied because the labeling was less efficient). Under the same conditions as described above, the reactions were monitored at 1, 5, 10, 20 and 30 min by radioTLC. All three chelators demonstrated near quantitative metal incorporation within 5 min at ambient temperature (Fig. 4).
The stability of [ 64 Cu]Cu-DOTA-2Py, −3Py and -4Py were examined in phosphate buffer and human serum at 1 h and 24 h at 37°C. All three ligands were largely stable (> 94%) in both phosphate buffer and human serum in 24 h (Table 1).
NOTA has been the widely used chelate for labeling with 64 Cu as it forms kinetically stable complexes (Cooper et al. 2012). For comparison, we examined the kinetic inertness of [ 64 Cu]Cu-DOTA-2Py, −3Py and -4Py using NOTA as the transchelation challenge reagent. After the formation of [ 64 Cu]Cu-DOTA-2Py, −3Py and -4Py at ligand concentration of 1 × 10 − 4 M (10 nmol ligand, 3 MBq 64 Cu), 1000 eq. of p-SCN-Bn-NOTA were added and the mixture was monitored by radio-HPLC over 4 h. During   With radiolabeling efficiency, stability and kinetic inertness of the Cu-complexes established, we attached DOTA-1Py, −2Py, −3Py to an αMSH peptide through a carboxylic moiety of each chelator (Fig. 5). Although DOTA-1Py-αMSH was expected to complex 64 Cu less efficiently, it was included to study the trend with increasing pyridyls. The peptide was synthesized via the step-wise addition of amino acids followed by on-resin cyclization. The final Fmoc protecting group was removed by mixing with HOBt, HBTU and isopropyl ethylamine and agitated at ambient temperature overnight. The deprotection and purification by HPLC was done by standard procedures (Zhang et al. 2017) and characterized by mass spectrometry.
Modifications of peptides may adversely impact target binding affinity (Di 2015). We studied the binding affinity of nat Cu-DOTA-xPy-αMSH (x = 1-3) and found that binding of all three peptides towards MC1R remained in the in low nanomolar level  (Table 2), although lower than Ga-DOTA-αMSH complex (K i = 0.31 ± 0.06 for CCZ01048) (Zhang et al. 2017).

Conclusion
We have synthesized a new set of chelators DOTA-1Py, −2Py, −3Py and -4Py and demonstrated that substitution of the carboxylic groups with one or two pyridine groups on a DOTA backbone can significantly improve the binding affinity towards 64 Cu. We covalently attached the chelators to an αMSH peptide and demonstrated that DOTA-xPy-αMSH (x = 1-3) can be quantitatively labeled by 64 Cu at ambient temperature. Moreover, all three nat Cu-DOTA-xPy-αMSH (x = 1-3) retained good binding affinity towards MC1R. Metabolic stability and biodistribution evaluation of [ 64 Cu]Cu-DOTA-xPy-αMSH (x = 1-3) is underway.

General
All reactions were carried out with commercial solvents and reagents that were used as received. Concentration and removal of trace solvents was done via a Büchi rotary evaporator using dry ice/acetone condenser, and vacuum applied from an aspirator or Büchi V-500 pump. Nuclear magnetic resonance (NMR) spectra were recorded using D 2 O or DMSO-d 6 or MeOD-d 4 . Signal positions are given in parts per million from

IR and UV-vis titration of DOTA-xPy (x = 1-3)
A PerkinElmer Fourier transfer-infrared spectrometer (FT-IR) was used to obtain the spectra of three chelators and their respective metal complexes with Cu 2+ metal ions, at 298 K, covering wavenumbers 650-4000 cm − 1 . The six samples had been dried to be solid samples as TFA salts. Samples were re-dissolved in ACN and dried three times to remove extra TFA. Approximately 2-5 mg of each sample was used to cover the 2 mm diameter crystal on the attenuated total reflectance (ATR) top plate. UV-vis absorbance was recorded on an Agilent Cary100 UVvisible spectrophotometer. 5 mg of each chelator was dissolved in water to make a solution of 0.26-29 mM. 0.2 eq. of Cu(II) aqueous solution (trace metal basis) was added for each titration. UV-vis spectra at 200-800 nm was recorded using a 1 cm pathlength quartz cell.

Production of 64 Cu
Isotopically enriched 64 Ni metal powder (50 mg) was dissolved in HCl. Aqueous ammonia was added to adjust the pH to~9. 64 Ni is then electroplated onto a rhodium backing disk (35 mm diameter, 1 mm thickness) using a small custom plating apparatus overnight. The plated nickel layer was approximately 8 mm diameter × 110 μm thickness. The target was irradiated for 1 h at 10 μA proton current 13 MeV and left in the cyclotron for 2 h to allow for the decay of the short-lived isotopes. The major co-produced isotopes were 61 Co (2.9%) and 103 Pb (2.0%) after cool-down. The 64 Ni layer of the target was dissolved in hot concentrated HCl. After evaporation to dryness, the residue was re-dissolved in 6 M HCl and loaded onto a 1 cm AG1-X8 solid extraction column pre-conditioned with 6 M HCl. The column was washed with 6 M HCl (10 mL) and 64 Cu was eluted with milli-Q water. Radionuclide purity (> 99%) was confirmed by gamma spectroscopy.

RadioTLC
RadioTLC was carried out using BioScan system 200 Image Scanner. iTLC-SA plates used as stationary phase, developed using 0.2 M EDTA (pH 5). Under these conditions uncomplexed 64 Cu II migrates with the solvent front (R f = 1), while the 64 Cu-complex remains at the baseline (R f = 0).

Kinetics of radiolabeling studies
To an aqueous solution of ligand (10 μL, 1 × 10 − 3 M) in MES buffer (83 μL, 0.4 M, pH = 6.2) was added [ 64 Cu]CuCl 2 (3 MBq, 7 μL in 0.01 M HCl). The reaction mixture was kept at room temperature. Aliquots of the mixtures were analyzed by radio-HPLC and radio-iTLC using the conditions described above to determine the radiolabeling yields.

In vitro stability of radio-ligands
Pre-formed 64 Cu-labeled complexes (100 μL) were added to phosphate buffer (300 μL, pH 7.4) or human serum (300 μL). After 1 h or 24 h of incubation at 37°C, the buffer samples were analyzed by semi-preparative HPLC. The serum samples were firstly treated with ethanol to precipitate and separate proteins (twice the volume of the mixture) and centrifuged for 1 min (13,200 rpm). The filtrate was decanted and further filtered (VWR sterile syringe filter, 0.2 μm PES), and subsequently analyzed by radio-iTLC using the conditions outlined above.

In vitro stability of radiolabeled peptides
Pre-formed 64 Cu-labeled bioconjugate (100 μL) was mixed in phosphate buffer (300 μL, pH 7.4) or human serum (300 μL). After 1 h or 4 h of incubation at 37°C, the buffer samples were analyzed by semi-preparative HPLC. The serum samples were firstly treated with ethanol to precipitate and separate proteins (twice the volume of the mixture) and centrifuged for 1 min (13,200 rpm). The filtrate was decanted and further filtered (VWR sterile syringe filter, 0.2 μm PES), and subsequently analyzed by radio-iTLC using the conditions outlined above.
Binding affinity of nat Cu-DOTA-xPy towards MC1R K i was measured using the same method previously reported (Zhang et al. 2017). Briefly, 5 × 10 5 B16F10 cells/well were seeded in complete growth medium in a 24 well Poly-D-Lysine plate (Corning BioCoat™,Ref No. 354414) and grown to 85-90% confluency at standard conditions. The growth medium was replaced Yang et al. EJNMMI Radiopharmacy and Chemistry (2021)  with reaction medium (RPMI, 2 mg/ml BSA, 20 mM HEPES) and the nonradioactive ligand added at a final concentration of 0.5 pM to 5 μM, along with 0.1 nM [ 125 I]I-αMSH. The cells were incubated for 1 h at 25°C with mild shaking. The supernatant was removed and cells were washed twice with ice-cold PBS. Cells were harvested with trypsin and counted on WIZARD 2480 gamma counter. The measurements were repeated in three separate experiments and reported as mean ± standard deviation.