Synthesis of 68Ga-Labeled cNGR-Based Glycopeptides and In Vivo Evaluation by PET Imaging

Tumor hypoxia induces angiogenesis, which is required for tumor cell survival. The aminopeptidase N receptor (APN/CD13) is an excellent marker of angiogenesis since it is overexpressed in angiogenic blood vessels and in tumor cells. Asparagine-glycine-arginine (NGR) peptide analogs bind selectively to the APN/CD13 recepto, therefore, they are important vector molecules in the development of a PET radiotracer which is capable of detecting APN-rich tumors. To investigate the effect of glycosylation and pegylation on in-vivo efficacy of an NGR-based radiotracer, two 68Ga-labeled radioglycopeptides were synthesized. A lactosamine derivative was applied to glycosylation of the NGR derivative and PEG4 moiety was used for pegylation. The receptor targeting potential and biodistribution of the radiopeptides were evaluated with in vivo PET imaging studies and ex vivo tissue distribution studies using B16-F10 melanoma tumor-bearing mice. According to these studies, all synthesized radiopeptides were capable of detecting APN expression in B16-F10 melanoma tumor. In addition, lower hepatic uptake, higher tumor-to background (T/M) ratio and prolonged circulation time were observed for the novel [68Ga]-10 radiotracer due to pegylation and glycosylation, resulting in more contrasting PET imaging. These in vivo PET imaging results correlated well with the ex vivo tissue distribution data.


Introduction
Initially, small tumors do not have their own vascular network; therefore, tumor growth can stop due to a lack of oxygen and nutrients. However, oxygen deficiency induces angiogenesis and new capillaries are formed from existing vessels to allow tumor cells to survive [1]. Since angiogenesis leads to tumor growth and malignant development, inhibiting of tumor neovascularization is a valuable approach to anticancer therapy. This this study is the first report of the synthesis and in vivo evaluation of 68 Ga-labeled NGRbased glycopeptides for determination of aminopeptidase N (APN/CD13) and galectin-3 expression with PET imaging.
After stirring overnight at room temperature, the reaction mixture was diluted with water, frozen and lyophilized. The crude product was purified by semipreparative RP-HPLC. The conditions of semipreparative RP-HPLC were the following: the Supelco Discovery ® Bio Wide Pore C18 column 10 µm (150 mm × 10mm) was eluted at a flow rate 4 mL/min. Linear gradient elution (0 min 10% B; 5 min 10% B; 30 min 90% B) was used with eluent A: 0.1% TFA in water and eluent B: 0.1% TFA in ACN/H 2 O (95:5, v/v). Peaks were detected at 254 nm. The product was collected between 11.1 and 11.5 min, frozen and lyophilized to give compound 11 (  A 68 Ge/ 68 Ga-generator (Eckert-Ziegler, GalliaPharm ® ) was eluted with 0.1 M aq. ultra-purified hydrochloric acid. A volume of 1000 µL of 68 Ga eluate (~80 MBq) was transformed into an Eppendorf vial, then 200 µL of NH 4 OAc (3 M, pH = 4) and 20 µL of aq. stock solution of compound 3, compound 9, and compound 10 (1 mg/mL) were added to, respectively. The reaction was performed at 95 • C for 15 min. Then the reaction mixture was passed through a pre-conditioned (5 mL EtOH, 10 mL water) SPE cartridge (Oasis HLB 1cc). After washing of the cartridge with 1 mL of water, the radiolabeled product was eluted with 1:1 mixture of ethanol and saline (2 × 250 µL). The obtained solution of the labeled compounds was concentrated with stream of nitrogen, then the residue was diluted with 100 µL saline. For quality control, the radiochemical purity of the radiotracers was investigated by radio-HPLC on Alliance module with Luna C18 3 µm column. Solvent A: 0.1% oxalic acid, solvent B: 95% acetonitrile, gradient: 0 min: 100% A, 1 min: 100% A, 10 min: 100% B, 11 min: 100% B, 12 min: 100% A. The radiochemical purity was found to be better than 95% for all labeled compounds and the molar activity was 3. A 68 Ge/ 68 Ga-generator (Eckert-Ziegler, GalliaPharm ® ) was eluted with 0.1 M aq. ultra-purified hydrochloric acid. A volume of 1000 µL of 68 Ga eluate was transformed into an Eppendorf vial, then 160 µL of NaOAc (1 M) and 3 µL of aq. solution of compound 11 (6 mM) were added to. The reaction was performed at 95 • C for 5 min. Then the reaction mixture was passed through a pre-conditioned (5 mL EtOH, 10 mL water) SPE cartridge (Oasis HLB 1cc). After washing of the cartridge with 1 mL of water, the radiolabeled product was eluted with 2:1 mixture of ethanol and saline (300 µL). For quality control, the radiochemical purity of the radiotracer was investigated by radio-HPLC on KNAUER HPLC system with an analytical Halo C18 5 µm column (100 mm × 3.0 mm) and 0.7 mL/min flow rate. Eluent A: 0.1% TFA in water and eluent B: 0.1% TFA in ACN/H 2 O (95:5, v/v), gradient: 0 min 4% B; 2 min 4% B; 5 min 10% B; 10 min 30% B; 13 min 45% B. [ 68 Ga]-11 was produced with high molar activity (8.51 ± 0.09 GBq/µmol) and with good radiochemical purity was better than 95%, in all cases. (Supplementary Material Figure S12).  12-week-old male C57BL/6 (n = 20) mice were housed in IVC system (Techniplast, Akronom Ltd., Budapest, Hungary) at 26 ± 2 • C with 55 ± 10% humidity and artificial lighting with a circadian cycle of 12 h. Semi-synthetic diet (SDS VRF-1, Animalab Ltd., Naszály út, Hungary) and tap water were available ad libitum to all the animals. The animals received human care and the experiments were authorized by the Ethical Committee for Animal Research, University of Debrecen, Hungary. Experimental animals were kept and treated in compliance with all applicable sections of the Hungarian Laws and directions and regulations of the European Union.

In Vivo PET Imaging and Image Analysis
For in vivo biodistribution studies B16-F10 tumor-bearing mice were anaesthetized with 3% isoflurane and were injected with 6.6 ± 0.

Statistical Analysis
Data were presented as mean ± SD. The significance was calculated using Student's t-test (two-tailed), two-way ANOVA, and Mann-Whitney U-test. The significance level was set at p ≤ 0.05 unless otherwise indicated.

Chemistry
First, we accomplished the synthesis of chelator-conjugated lactosamine derivative Since the conjugation of the peptide to chelator-bearing lactosamine 3 was designed with copper-free click reaction, the cKNGRE (4) was functionalized with cyclooctyne moiety using DBCO-NHS (5) in DMSO and in the presence of diisopropylethylamine (DIPEA) to give compound 6. For the synthesis of PEG4-linker-containing radiotracer the peptide 4 was reacted with DBCO-PEG4-NHS (7) in the same way as described above to yield compound 8. (Scheme 2). Subsequently, compound 6 and compound 8 were conjugated to chelator-bearing lactosamine 3 with a strain promoted click reaction in DMSO, respectively (Scheme 3). Since the conjugation of the peptide to chelator-bearing lactosamine 3 was designed with copper-free click reaction, the cKNGRE (4) was functionalized with cyclooctyne moiety using DBCO-NHS (5) in DMSO and in the presence of diisopropylethylamine (DIPEA) to give compound 6. For the synthesis of PEG 4 -linker-containing radiotracer the peptide 4 was reacted with DBCO-PEG 4 -NHS (7) in the same way as described above to yield compound 8. (Scheme 2). Since the conjugation of the peptide to chelator-bearing lactosamine 3 was designed with copper-free click reaction, the cKNGRE (4) was functionalized with cyclooctyne moiety using DBCO-NHS (5) in DMSO and in the presence of diisopropylethylamine (DIPEA) to give compound 6. For the synthesis of PEG4-linker-containing radiotracer the peptide 4 was reacted with DBCO-PEG4-NHS (7) in the same way as described above to yield compound 8. (Scheme 2). Subsequently, compound 6 and compound 8 were conjugated to chelator-bearing lactosamine 3 with a strain promoted click reaction in DMSO, respectively (Scheme 3). Subsequently, compound 6 and compound 8 were conjugated to chelator-bearing lactosamine 3 with a strain promoted click reaction in DMSO, respectively (Scheme 3).
To investigate the effect of the structural changing on the in vivo performance of the radioglycopeptides we carried out the synthesis of the DOTAGA-cKNGRE (11) as a reference. Thus, p-NCS-Bn-DOTA-GA (2) was coupled with cKNGRE (4) in the presence of triethylamine in dimethylformamide and water (Scheme 4).
The glycopeptides (9 and 10) and compound 11 are suitable precursors for radiolabeling with 68 Ga radioisotope. The synthetized compounds were characterized by analytical RP-HPLC and HRMS ESI. To investigate the effect of the structural changing on the in vivo performance of the radioglycopeptides we carried out the synthesis of the DOTAGA-cKNGRE (11) as a reference. Thus, p-NCS-Bn-DOTA-GA (2) was coupled with cKNGRE (4) in the presence of triethylamine in dimethylformamide and water (Scheme 4). The glycopeptides (9 and 10) and compound 11 are suitable precursors for radiolabeling with 68 Ga radioisotope. The synthetized compounds were characterized by analytical RP-HPLC and HRMS ESI.

Radiochemistry
Radiolabeling of the synthetized precursors (3, 9, 10, and 11) was performed using 68 Ge/ 68 Ga-generator produced 68 Ga isotope with high radiochemical purity (>95%). The labeled complexes were purified with solid phase extraction (SPE) using Oasis HLB 1cc cartridge. The radiochemical purity of the obtained radiopharmaceuticals was analyzed by radio-HPLC (>95%). The octanol/water partition coefficients (logP) of 68 Ga-labeled radioligands were determined. The logP value was found to be −3. Furthermore, the stability of the labeled ligands was investigated; therefore, the solution of the radiotracers was mixed with a solution of human serum, oxalic acid (0.01 M) and Na2EDTA (0.01 M) at room temperature, respectively. Samples were then taken at various time points (0, 60, 120 min) and analyzed by radio-HPLC. Radiochromatograms showed that none of the radiotracers was degraded at the time points studied, as the ra- To investigate the effect of the structural changing on the in vivo performance of the radioglycopeptides we carried out the synthesis of the DOTAGA-cKNGRE (11) as a reference. Thus, p-NCS-Bn-DOTA-GA (2) was coupled with cKNGRE (4) in the presence of triethylamine in dimethylformamide and water (Scheme 4). The glycopeptides (9 and 10) and compound 11 are suitable precursors for radiolabeling with 68 Ga radioisotope. The synthetized compounds were characterized by analytical RP-HPLC and HRMS ESI.

Radiochemistry
Radiolabeling of the synthetized precursors (3, 9, 10, and 11) was performed using 68 Ge/ 68 Ga-generator produced 68 Ga isotope with high radiochemical purity (>95%). The labeled complexes were purified with solid phase extraction (SPE) using Oasis HLB 1cc cartridge. The radiochemical purity of the obtained radiopharmaceuticals was analyzed by radio-HPLC (>95%). The octanol/water partition coefficients (logP) of 68 Ga-labeled radioligands were determined. The logP value was found to be −3. Furthermore, the stability of the labeled ligands was investigated; therefore, the solution of the radiotracers was mixed with a solution of human serum, oxalic acid (0.01 M) and Na2EDTA (0.01 M) at room temperature, respectively. Samples were then taken at various time points (0, 60, 120 min) and analyzed by radio-HPLC. Radiochromatograms showed that none of the radiotracers was degraded at the time points studied, as the radiochemical purity of all samples was >95%.

Radiochemistry
Radiolabeling of the synthetized precursors (3, 9, 10, and 11) was performed using 68 Ge/ 68 Ga-generator produced 68 Ga isotope with high radiochemical purity (>95%). The labeled complexes were purified with solid phase extraction (SPE) using Oasis HLB 1cc cartridge. The radiochemical purity of the obtained radiopharmaceuticals was analyzed by radio-HPLC (>95%). The octanol/water partition coefficients (logP) of 68 Ga-labeled radioligands were determined. The logP value was found to be −3. Furthermore, the stability of the labeled ligands was investigated; therefore, the solution of the radiotracers was mixed with a solution of human serum, oxalic acid (0.01 M) and Na 2 EDTA (0.01 M) at room temperature, respectively. Samples were then taken at various time points (0, 60, 120 min) and analyzed by radio-HPLC. Radiochromatograms showed that none of the radiotracers was degraded at the time points studied, as the radiochemical purity of all samples was >95%.
In vivo PET imaging results correlated well with the ex vivo data shown in Figure 2.
The design of the radiotracer, including the chemical structure of the targeting molecule, the chelating agent, and the linker moiety, has a remarkable effect on the biological efficacy of radiolabeled compounds [18,19]. One effective method to improve the in vivo performance of a peptide-targeted radiotracer is glycosylation, because glycosylated peptides are more hydrophilic than the original peptides, which increases their solubility and bioavailability, as well as their enzymatic stability with the endogenous proteases [18]. We chose a lactosamine derivative as a carbohydrate moiety for glycosylation of a NGR

Diccussion
Numerous studies demonstrated with blocking experiments that NGR peptide analogs bind to APN/CD13 receptor with high selectivity, making them excellent vector molecules for the synthesis of peptide-targeted radiopharmaceuticals [13][14][15][16]. 68 Ga-labeling of an NGR derivative conjugated with a bifunctional chelator can provide a PET imaging agent for the detection of APN/CD13-positive tumor cells. For example, to detect APN/CD13 expression, 68 Ga-DOTA-NGR was prepared by Zang et al. Based on MicroPET imaging, high tumor uptake of this radiotracer was observed in the tumor site in A549 tumor xenografts [13]. Furthermore, Chen et al. described the synthesis and preclinical evaluation of two radioligands containing monomeric and dimeric NGR peptides which were radiolabeled with 64 Cu PET isotope (β + 17.8%; t 1/2 = 12.7 h; E(β + ) = 655 keV) [14].  [15]. In another study [16] a c(NGR) analog modified by N-methylation (c[CH 2 -CO-Lys(NODAGA)-Asn-N(Me)Gly-Arg-Cys]-NH 2 ) to prevent asparagine deamidation, was synthesized and radiolabeled with 68 Ga isotope. The pharmacokinetic properties of this radiopeptide were compared with the 68 Ga-labeled non-N-methylated version (c[CH 2 -CO-Lys(NODAGA)-Asn-Gly-Arg-Cys]-NH 2 ) and two other cNGR-based radiotracers, the NOTA-c(NGR) [15] and NODAGA-c(NGR). All labeled compounds showed high tumor accumulation in APN/CD13-positive He/De and Ne/De tumors. These studies confirmed that chelator-conjugated NGR analogs are promising PET agents for APN-overexpressing tumors.
The design of the radiotracer, including the chemical structure of the targeting molecule, the chelating agent, and the linker moiety, has a remarkable effect on the biological efficacy of radiolabeled compounds [18,19]. One effective method to improve the in vivo performance of a peptide-targeted radiotracer is glycosylation, because glycosylated peptides are more hydrophilic than the original peptides, which increases their solubility and bioavailability, as well as their enzymatic stability with the endogenous proteases [18]. We chose a lactosamine derivative as a carbohydrate moiety for glycosylation of a NGR derivatives hence it is natural ligand of galectin-3 receptor which is overexpressed in numerous tumors [22]. This receptor is a carbohydrate-binding protein and can recognize the β-galactoside motif of glycoconjugates [28]. It is involved in different pathological processes during cancer progression, such as cell adhesion and migration, proliferation, differentiation, angiogenesis and metastasis [29]. Therefore, the incorporation of lactosamine moiety into a DOTAGA-conjugated NGR derivetive can result in dual-targeting radiopharmaceutical. Dual-targeting approach is a relatively new method in the field of PET imaging, but previously numerous bispecific antibodies were developed for therapeutic applications [30]. Pegylation can also improve the pharmacokinetic properties of the radiopeptides [19]. To investigate the effect of glycosylation and pegylation on the in vivo properties of a radiotracer, three NGR-based radiopeptides were synthesized. We have previously reported the synthesis and radiochemical investigation of a glycopeptide, 68 Ga-NODAGA-LacN-E[c(RGDfK)] 2 , which may be suitable for the detection of α v β 3 integrin and galectin-3 expression in tumor endothelial and cancer cells by PET imaging [31]. In this report, a similar synthetic approach was used to prepare NGR-based radioligands as previously applied for 68 Ga-NODAGA-LacN-E[c(RGDfK)] 2 . In this case, however, a DOTAGA derivative was chosen to complex the gallium-68 radioisotope hence DOTA-based compounds are excellently chelate of various diagnostic radiometals, such as gallium-68, scandium-44 and copper-64, as well as various therapeutic metals, e.g., lutetium-177 and bismuth-213. Thus, for targeted radionuclide therapy the newly synthetized DOTAGA-conjugated peptides may also be labeled with therapeutic radiometals.
After successful preparation of the NGR-and/or lactosamine-containing precursors (3, 9, 10 and 11), their 68 Ga labeling was performed with high radiochemical purity using conventional procedure. The octanol/water partition coefficients were also determined and similarly low logP values indicated that the radiolabeled compounds were hydrophilic. Interestingly, the logP of the non-glycosylated radiopeptide was lower than that of the glycosylated radiopeptides. The DBCO moiety formed during bioconjugation is likely to compensate for the hydrophilicity of the lactosamine residue. In addition, in vitro stability tests of radiolabeled compounds in human serum have been performed and remained intact for two hours.
By analyzing the liver and kidney uptake, differences were observed between the radiotracers (Figure 1). This finding shows the route of excretion from the body according to the logP values (approximately −3.5); moreover, the accumulation of kidney and liver is greatly influenced by the expression of galectin-3 and APN/CD13 in these tissues. It is known from the literature that the APN/CD13 expression is high in the kidney and liver [33], in addition, low galectin-3 expression was found in the healthy kidney and liver tissues [34][35][36]. Consistent with these observations, we also found higher liver and kidney accumulation using the APN/CD13-specific cKNGRE-containing radiotracers than that of the galectin-3-specific [ 68 Ga]-3 probe. Moreover, the accumulation in the liver was significantly (p ≤ 0.05) lower for the PEG-containing [ 68 Ga]-10 than for the [ 68 Ga]-3, [ 68 Ga]-11 and [ 68 Ga]-9 due to pegylated proteins increase the circulation time in the body and decrease immunogenicity [37]. The prolonged circulation time was also confirmed by our ex vivo study in which the blood showed higher radioactivity concentration for [ 68 Ga]-10 90 min after the injection, compared to the other investigated molecules (Figure 2). Our in vivo PET imaging results correlated well with the ex vivo data shown in Figure 2.

Conclusions
We have successfully completed the synthesis and radiolabeling of two NGR-based glycopeptides (9 and 10) and DOTAGA-cKNGRE (11), which are suitable for the detection of APN-positive tumors by PET imaging. The in vivo properties of the novel radioligands were compared by PET imaging and ex vivo biodistribution. These studies showed that all three NGR-based radiopeptides were able to visualize APN/CD13 receptor expression in B16-F10 melanoma tumors by PET imaging. The glycosylation and pegylation of NGR derivatives resulted in slightly enhanced tumor uptake. However, for the [ 68 Ga]-10 tracer, liver uptake was significantly lower and tumor-to-background (muscle) ratio (T/M) was higher due to pegylation and glycosylation, resulting in a more contrasting PET image. These pharmacokinetic properties are advantageous for potential radiotherapeutic use if the precursor molecule 10 is labeled with a therapeutic radiometal. In addition, the suitability of [ 68 Ga]-3 for the determination of galectin-3 levels was investigated, but the use of this radiotracer resulted in moderate tumor uptake, which may be explained by the fact that this lactosamine derivative is not a sufficiently specific ligand for galactin-3 receptor. Presumably, therefore, the synergetic effect of the dual-targeting approach was also not significant. However, the application of pegylation and glycosylation of the NGR derivative improved the pharmacokinetic characteristics of the radiopeptide.