Synthesis, Physicochemical, Labeling and In Vivo Characterization of 44Sc-Labeled DO3AM-NI as a Hypoxia-Sensitive PET Probe

Hypoxia promotes angiogenesis, which is crucial for tumor growth, and induces malignant progression and increases the therapeutic resistance. Positron emission tomography (PET) enables the detection of the hypoxic regions in tumors using 2-nitroimidazole-based radiopharmaceuticals. We describe here a physicochemical study of the Sc(DO3AM-NI) complex, which indicates: (a) relatively slow formation of the Sc(DO3AM-NI) chelate in acidic solution; (b) lower thermodynamic stability than the reference Sc(DOTA); (c) however, it is substantially more inert and consequently can be regarded as an excellent Sc-binder system. In addition, we report a comparison of 44Sc-labeled DO3AM-NI with its known 68Ga-labeled analog as a hypoxia PET probe. The in vivo and ex vivo biodistributions of 44Sc- and 68Ga-labeled DO3AM-NI in healthy and KB tumor-bearing SCID mice were examined 90 and 240 min after intravenous injection. No significant difference was found between the accumulation of 44Sc- and 68Ga-labeled DO3AM-NI in KB tumors. However, a significantly higher accumulation of [68Ga]Ga(DO3AM-NI) was found in liver, spleen, kidney, intestine, lung, heart and brain than for [44Sc]Sc(DO3AM-NI), leading to a lower tumor/background ratio. The tumor-to-muscle (T/M) ratio of [44Sc]Sc(DO3AM-NI) was approximately 10–15-fold higher than that of [68Ga]Ga(DO3AM-NI) at all time points. Thus, [44Sc]Sc(DO3AM-NI) allows the visualization of KB tumors with higher resolution, making it a promising hypoxia-specific PET radiotracer.


Introduction
Accurate tumor locating, staging and monitoring the impact of anticancer treatment are important issues in the fight against cancer. Positron emission tomography (PET) is a sensitive noninvasive imaging technique that can detect pathological processes at the molecular level in living systems. Hypoxia that develops during tumor growth induces cellular changes that result in malignant progression [1] and furthermore causes the following problems for anticancer treatments [2]. Oxygen deficiency triggers angiogenesis in solid tumors, but the formed vessels are abnormal; therefore, the delivery of the anticancer drugs to tumor cells is reduced [3]. Low-oxygenated cells are 2-3 times more resistant to ionizing radiation than healthy cells, which decreases the effectiveness of radiotherapy [4]. Consequently, the determination of the hypoxic region of tumors is key to achieving effective anticancer treatment. Because hypoxia is a significant challenge for oncologists, the development of radiopharmaceuticals capable of detecting low-oxygenated tumor cells is becoming increasingly important.
2-Nitroimidazole (NI) derivatives are excellent targeting vectors for imaging the hypoxic regions of tumors because they are reduced and entrapped by nitroreductase enzymes in hypoxic cells. In normal tissues, however, they are reoxidized and eliminated [5]. To date, some 18 F-labeled radiopharmaceuticals containing a 2-nitroimidazole moiety were developed [6][7][8]. The two best known are [ 18 F]-fluoromisonidazole ([ 18 F]-FMISO) [9], which was the first PET radiopharmaceutical to be most widely evaluated in tumor hypoxia imaging, and [ 18 F]-fluoroazomycin arabinoside ([ 18 F]-FAZA) [10]. The lipophilic features of these PET agents provide diffusion into cells but cause slow clearance from healthy tissues, resulting in a low tumor-to-background ratio.
Faster clearance is achieved with radiolabeled chelator-conjugated 2-nitroimidazole derivatives [11]. In addition, the incorporation of radiometals into a suitable ligand offers a more straightforward avenue of radiolabeling than the nucleophilic radiofluorination, which often requires anhydrous conditions and a complex labeling procedure, resulting in low radiochemical yield. Therefore, radiometalation has recently become a favored process for peptides and other hydrophilic biomolecules [12][13][14][15][16][17]. Scandium-44 (t 1/2 = 3.97 h, I = 94.27%, E mean (β+) = 0.63 MeV) for radiolabeling has sparked our interest because of its excellent nuclear properties for PET imaging [12]. In addition, this positron emitter radionuclide has several advantages over the widely used gallium-68 (t 1/2 = 68 min, I = 89%, E max (β+) = 1.92 MeV) [13]. The shorter-lived 68 Ga radiometal can only be used in-house, while the relatively long half-life and large-scale production of the 44 Sc by a cyclotron via 44 Ca (p, n) 44 Sc reaction [14] allow transportation of 44 Sc-labeled radiopharmaceuticals. Some studies have reported that radiolabeling with the 44 Sc nuclide improves the pharmacokinetic features of the labeled complexes compared to 68 Ga-labeled analogs [15]. Furthermore, the utilization of 44 Sc-labeled radiotracer is also advantageous if a longer observation time is required for PET imaging (e.g., proteins, antibodies) [16]. In addition, beta emitter 47 Sc is a valuable therapeutic match to the 44 Sc nuclide for radiotheragnostic applications [12,17].
Hypoxia-selective uptake of 68 Ga-labeled DO3AM-NI has been proved by Hoigebazar et al. [11]. Therefore, we have decided to achieve the radiolabeling of this DOTAconjugated NI ligand (Figure 1) with the positron-emitting 44 Sc isotope and compare the pharmacokinetic properties of this 44 Sc-labeled complex with the known 68 Ga-labeled version by in vivo PET imaging and ex vivo biodistribution. We have also conducted a detailed physicochemical study of the scandium (III) complex of DO3AM-NI to demonstrate its suitable thermodynamic stability, formation kinetics and kinetic inertness for in vivo use. The radiolabeling of 2-nitroimidizole derivatives with 44 Sc isotope and the evaluation of the utility of the radiotracer for PET hypoxia imaging have not been previously reported.

Physicochemical Studies
The first step in the physicochemical studies of the Sc(III) chelate was to assess the rate of complex formation at an acidic pH, since based on equilibrium data published in the literature for Sc(III) complexes of DOTA, DTPA or AAZTA ligands one can expect that the formation of the (likely quite stable) Sc(DO3AM-NI) complex is expected at a low pH [19,20]. 1 H and 45 Sc NMR measurements performed at pH = 1.52 show that the complex formation reaches equilibrium in about twelve hours ( Figures S1-S3). This is promising for the labeling studies, yet it also underlines that the so-called "batch-method" must be applied in very acidic samples for the determination of the stability constant of the Sc(DO3AM-NI) chelate.
Equilibrium measurements were commenced by determining the protonation constants of the ligand, measured at 37 °C (I = 0.15 M NaCl) owing to the relevance of the chelate under in vivo conditions. The protonation constants obtained by fitting the pHpotentiometric data for the ligand DO3AM-NI are shown in Table 1 along with the literature data published for DOTA and certain DO3A-monoamides (DO3A-mono-N-buthylamide as well as that of a hypoxia-sensitive MRI probe reported by some of us a couple of years ago) [21]. Five protonation steps were observed by performing the titration in the pH range of 1.72-11.85 for the DO3AM-NI monoamide type ligand, two of which (characterized by high constants) describe the protonation of nitrogen atoms in the macrocycle (positioned trans-to each other), while the remaining log Ki H values (much lower) likely reflect the protonation of the carboxylates. The replacement of an acetate group of DOTA by an amide moiety results in a decrease in the protonation sites of the ligand, and as a consequence the overall basicity of the ligand (expressed as the sum of the protonation constants, log β015) was as expected, predicted from the data published in literature. What is more, the formation of a weak Na + complex (originating from the ionic strength) competing with the first protonation process may be responsible of the lower value of the first protonation constant. Altogether, the basicity value is significantly lower than that for DOTA and comparable to/slightly lower than that of DO3AM-monoamides. The decrease in ligand basicity resulted in a decrease in the stability constants of the Ln(III) complexes, which is also expected for Sc(III) chelate. We carried out the synthesis of the DO3AM-NI ligand according to the literature's methods. First, the preparation of 2-(2-nitro-imidazol-1-yl) ethanamine was performed by the slight modification of the procedure reported by Zha et al. [18]. Then, this compound was conjugated to 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) by using the reaction conditions that were described by Hoigebazar et al. [11].

Physicochemical Studies
The first step in the physicochemical studies of the Sc(III) chelate was to assess the rate of complex formation at an acidic pH, since based on equilibrium data published in the literature for Sc(III) complexes of DOTA, DTPA or AAZTA ligands one can expect that the formation of the (likely quite stable) Sc(DO3AM-NI) complex is expected at a low pH [19,20]. 1 H and 45 Sc NMR measurements performed at pH = 1.52 show that the complex formation reaches equilibrium in about twelve hours ( Figures S1-S3). This is promising for the labeling studies, yet it also underlines that the so-called "batch-method" must be applied in very acidic samples for the determination of the stability constant of the Sc(DO3AM-NI) chelate.
Equilibrium measurements were commenced by determining the protonation constants of the ligand, measured at 37 • C (I = 0.15 M NaCl) owing to the relevance of the chelate under in vivo conditions. The protonation constants obtained by fitting the pHpotentiometric data for the ligand DO3AM-NI are shown in Table 1 along with the literature data published for DOTA and certain DO3A-monoamides (DO3A-mono-N-buthylamide as well as that of a hypoxia-sensitive MRI probe reported by some of us a couple of years ago) [21]. Five protonation steps were observed by performing the titration in the pH range of 1.72-11.85 for the DO3AM-NI monoamide type ligand, two of which (characterized by high constants) describe the protonation of nitrogen atoms in the macrocycle (positioned transto each other), while the remaining log K i H values (much lower) likely reflect the protonation of the carboxylates. The replacement of an acetate group of DOTA by an amide moiety results in a decrease in the protonation sites of the ligand, and as a consequence the overall basicity of the ligand (expressed as the sum of the protonation constants, log β 015 ) was as expected, predicted from the data published in literature. What is more, the formation of a weak Na + complex (originating from the ionic strength) competing with the first protonation process may be responsible of the lower value of the first protonation constant. Altogether, the basicity value is significantly lower than that for DOTA and comparable to/slightly lower than that of DO3AM-monoamides. The decrease in ligand basicity resulted in a decrease in the stability constants of the Ln(III) complexes, which is also expected for Sc(III) chelate.  [20]; [g] Ref. [21].
The stability constant of the Sc(DO3AM-NI) complex was assessed by 1 H and 45 Sc-NMR studies using samples prepared in the 0.055-0.195 M acid concentration range. These data were supplemented by pH-potentiometric titration data obtained for the preformed complex to ensure the absence/presence of protonated and ternary hydroxydo species in solution at low and high pH, respectively. Absence of protonation and the presence of a ternary hydroxydo species formation with K ScL•OH = 10.88 (5) were detected at our experimental conditions. The NMR samples contained 3.32 mM ligand, ScCl 3 and a constant ionic strength of 0.15 M NaCl. The samples were thermostated at 37 • C, and their 1 H-and 45 Sc-NMR spectra were recorded after 60 days. In order to ensure that the equilibrium in the samples was attained, the 1 H and 45 Sc-NMR measurements were repeated again 90 days later. The spectra were practically identical with the spectra recorded at 60 days ( Figures S4 and S5). As expected, the thermodynamic stability constant of the Sc(DO3AM-NI) complex (log K ScL = 22.36(4)) was considerably lower than those of Sc(III) complexes formed with DOTA, DTPA or AAZTA ligands (Table 1). A moderate decrease in terms of thermodynamic stability constants was already observed for the complexes of DO3A-monamides formed with Gd(III) ions in comparison to the stability of parent Gd(DOTA) − chelate [21,22]. Despite the lower stability constant of the complex, the Sc(III) ion was complexed completely around pH = 2.0 (in a millimolar concentration range, see Figure 2). However, it might be misleading to directly compare the stability constants of the complexes formed by ligands of different basicity (as it depends on the basicity of the ligand); therefore, we calculated and compared the pSc values (which could be treated as conditional constants) for the complexes shown in Table 1  chelates, yet the drop was much less pronounced compared to the difference of five to eight log units observed in the thermodynamic stability constant (i.e., significant part of the drop in the stability constant of Sc(DO3AM-NI) can be attributed to the differences in experimental conditions rather than the use of an amide moiety instead of a carboxylate metal binding unit causing a moderate drop in the conditional stability of the complex). Sc(DO3AM-NI) can be attributed to the differences in experimental conditions rather than the use of an amide moiety instead of a carboxylate metal binding unit causing a moderate drop in the conditional stability of the complex). Safe in vivo applications of metal chelates require the utilization of very robust complexes characterized by high thermodynamic stability and inertness to avoid the transmetalation/transchelation reactions occurring with competing endogenous metal ions/ligands in the biological milieu. Since the stability of the Sc(DO3AM-NI) complex is lower than those of the complexes used as comparative benchmarks, we have probed its inertness. In most of the cases, the acid-assisted dissociation is recognized as a major dissociation path responsible for metal ion loss from the complexes of DOTA derivative ligands. Therefore, we have followed the dissociation rate of Sc(DO3AM-NI) complex in the presence of strong acid (1 M HCl) in order to gain some information about its inertness. The given study performed in duplicate gives rate constants of (1.55 ± 0.04) × 10 −6 s −1 and (1.67 ± 0.05) × 10 −6 s −1 ( Figure S6), which is one-fourth of that observed for the Sc(DOTA)complex under identical conditions and several orders of magnitude smaller than the value previously found for the Sc(AAZTA) -complex (0.1 M −1 s −1 ) [19,20]. These data clearly show that the Sc(DO3AM-NI) chelate possesses superior inertness (slightly better than Sc(DOTA) − ) and can be recommended for in vivo studies.

Radiochemistry
The radiolabeling of the DO3AM-NI ligand with [ 68 Ga]Ga 3+ was previously described by Hoigebazar et al. [11]. For radiolabeling, we treated the [ 68 Ga]GaCl3 solution, which Safe in vivo applications of metal chelates require the utilization of very robust complexes characterized by high thermodynamic stability and inertness to avoid the transmetalation/transchelation reactions occurring with competing endogenous metal ions/ligands in the biological milieu. Since the stability of the Sc(DO3AM-NI) complex is lower than those of the complexes used as comparative benchmarks, we have probed its inertness. In most of the cases, the acid-assisted dissociation is recognized as a major dissociation path responsible for metal ion loss from the complexes of DOTA derivative ligands. Therefore, we have followed the dissociation rate of Sc(DO3AM-NI) complex in the presence of strong acid (1 M HCl) in order to gain some information about its inertness. The given study performed in duplicate gives rate constants of (1.55 ± 0.04) × 10 −6 s −1 and (1.67 ± 0.05) × 10 −6 s −1 (Figure S6), which is one-fourth of that observed for the Sc(DOTA)complex under identical conditions and several orders of magnitude smaller than the value previously found for the Sc(AAZTA)complex (0.1 M −1 s −1 ) [19,20]. These data clearly show that the Sc(DO3AM-NI) chelate possesses superior inertness (slightly better than Sc(DOTA) − ) and can be recommended for in vivo studies.

Radiochemistry
The radiolabeling of the DO3AM-NI ligand with [ 68 Ga]Ga 3+ was previously described by Hoigebazar et al. [11]. For radiolabeling, we treated the [ 68 Ga]GaCl 3 solution, which Pharmaceuticals 2022, 15, 666 6 of 16 was eluted from a 68 Ge/ 68 Ga-generator, with a post-processing method suggested by Eppard et al. [24] to reduce the eluate volume and remove metal tracer impurities. Then, the DO3AM-NI precursor was labeled with the conventional labeling procedure with good labeling yield (>90%). After radiolabeling, the radiotracer was purified by solid phase extraction (SPE) using a reversed phase C-18 SPE cartridge (SepPak C18 plus, Waters, Milford, USA). The radiochemical purity (RCP) of the purified labeled complex was analyzed by radio-HPLC and found to be >96%.
Cyclotron production of positron-emitting 44 Sc was performed by proton irradiation of natural calcium targets via the 44 Ca(p, n) 44 Sc reaction [14]. The 44 Sc was separated from target materials and other metal-based impurities by the methods described by Happel et al. [25]. The radiolabeling with 44 Sc was carried out by the modified version of the previously published procedure [15] with a high labeling yield (>93%). The radiotracer was purified by solid-phase extraction using a LiChrolute EN cartridge (Merck, Darmstadt, Germany). The RCP of the labeled product was determined by radio-HPLC, and it was over 95% ( Figure S7).
Subsequently, the octanol/water partition coefficient (logP) of the labeled complexes was determined to be −3.89 for [ 68 Ga]Ga(DO3AM-NI) and −2.59 for [ 44 Sc]Sc(DO3AM-NI). The low logP values indicated that both radiotracers were hydrophilic. To assess the labeled compounds' stability, the chelates were mixed with the solution of mouse plasma, Na 2 H 2 EDTA and oxalic acid, respectively. Aliquots were then taken at different time points (0, 60, 120 and 240 min) and injected into the radio-HPLC column, and the chromatograms were analyzed. After 240 min, the RCP of the samples was above 93% in all three cases. Figure S8 shows the radio-HPLC chromatogram of the serum stability test at 240 min. These results show that both radioactive tracers possess high stability under the conditions studied.

In Vivo and Ex Vivo Studies
The biodistribution of 44 Sc-and 68 Ga-labeled DO3AM-NI was investigated using healthy control and KB tumor-bearing SCID mice. PET/MRI imaging and ex vivo studies were performed 90 and 240 min after intravenous injection of 44 Sc-or 68 Ga-labeled DO3AM-NI. Representative coronal PET/MRI images of a healthy control mouse are shown in Figure 3. By the qualitative analysis of the PET/MRI images, the kidneys and the bladder with urine were clearly visualized using both radiotracers at each investigated time point; however, in other abdominal and thoracic organs, low uptake was observed. After 90 min of incubation time, a relatively high accumulation was observed in the liver using [ 68 Ga]Ga(DO3AM-NI) ( Figure 3B). After the quantitative SUV analysis of the in vivo PET/MRI images, we found significantly (p ≤ 0.01) higher [ 68 Ga]Ga(DO3AM-NI) uptake in the liver, spleen, kidney, intestines, lung, heart and brain than that of the [ 44 Sc]Sc(DO3AM-NI) accumulation in the same organs 90 and 240 min after tracer injection ( Figure 3E,F).
The results of the quantitative SUV analysis are correlated well with the ex vivo data shown in Table 2. Ex vivo biodistribution studies were carried out 90 min and 240 min after intravenous injection of 44 Sc-or 68 Ga-labeled DO3AM-NI, and the accumulated activities of the organs and tissues were determined using a gamma counter. By the quantitative analysis of the ex vivo data, significant differences were found between the %ID/g values when 44 Sc-and 68 Ga-labeled DO3AM-NI accumulations were compared. Significantly (p ≤ 0.01) lower %ID/g uptake values were measured in all the selected organs using [ 44 Sc]Sc(DO3AM-NI) than that of the [ 68 Ga]Ga(DO3AM-NI) at each investigated time point. Furthermore, notable accumulation was observed in the kidneys and urine using both radiotracers (Table 2), as it was expected from the logP values confirming that these radiolabeled complexes were excreted by the kidneys. The results of biodistribution studies correlated well with the ones obtained formerly for 68 Ga-labeled nitroimidazole derivatives [26,27], which also showed low uptake and accumulation in abdominal and thoracic organs and were mainly excreted through the kidney. In contrast, e.g., [  probes [31], due to their chemical properties, were mainly metabolized via the liver, and the large abdominal accumulation impaired the evaluation of abdominal tumors in the PET images. The 44 Sc-and 68 Ga-labeled nitroimidazole derivatives had more favorable properties in terms of the reporting and evaluation of PET images and due to their rapid elimination through the urinary system. The results of the quantitative SUV analysis are correlated well with the ex vivo shown in Table 2. Ex vivo biodistribution studies were carried out 90 min and 240 after intravenous injection of 44 Sc-or 68 Ga-labeled DO3AM-NI, and the accumulated    Figure 4F).
Overall, we found that the subcutaneously growing KB tumors were clearly visualized with excellent image contrast and higher SUVs, and T/M ratios were found by using the 44 Sc-labeled molecule. Similar differences were found between the 68 Ga-and 44 Sc-labeled DOTA-and NODAGA-RGD molecules [32] or between the 68 Ga-and 44 Sc-labeled DOTA-NAPamide probes [15], when the tumor uptake and tumor-to-background ratios were investigated using in vivo tumor models, and higher accumulation was found using the 44 Sc-labeled molecules. The difference between the 44 Sc-and 68 Ga-labeled DO3AM-NI was also due to the chemical properties and its longer half-life of 44 Sc-DO3AM-labeled molecule.
For the assessment of the accumulation in the tumors of [ 44 Sc]Sc(DO3AM-NI) and [ 68 Ga]Ga(DO3AM-NI), ex vivo biodistribution studies were performed 90 and 240 min post injection using KB tumor xenografts. Table 3 demonstrates that 90 and 240 min after tracer injection no significant differences were found between the 44 Sc-and 68 Ga-labeled DO3AM-NI accumulations in KB tumors; however, the [ 44 Sc]Sc(DO3AM-NI) uptake was relatively higher at each investigated time point. In contrast, when the tumor-to-muscle ratios (T/M) were calculated we found that the T/M ratio of [ 44 Sc]Sc(DO3AM-NI) was approximately 10-15-fold higher at each time point than that of the T/M ratio of [ 68 Ga]Ga(DO3AM-NI), and this difference was significant (p ≤ 0.01).
was also due to the chemical properties and its longer half-life of 44 Sc-DO3AM-labeled molecule.  Table 3 demonstrates that 90 and 240 min after tracer injection no significant differences were found between the 44 Sc-and 68 Ga-labeled DO3AM-NI accumulations in KB tumors; however, the [ 44 Sc]Sc(DO3AM-NI) uptake was relatively higher at each investigated time point. In contrast, when the tumor-to-muscle ratios (T/M) were calculated we found that the T/M ratio of [ 44 Sc]Sc(DO3AM-NI) was approximately 10-15-fold higher at each time point than that of the T/M ratio of [ 68 Ga]Ga(DO3AM-NI), and this difference was significant (p ≤ 0.01).

pH-Potentiometric Studies
The equilibrium measurements were performed at constant ionic strength maintained by 0.15 M NaCl at 37 • C. For determining the protonation constants of DO3AM-NI, two parallel pH-potentiometric titrations were performed with 0.1611 M NaOH in a 5.00 cm 3 sample containing the ligand at the concentration of 2.76 mM. The stability constant of Sc(DO3AM-NI) was assessed by 1 H and 45 Sc NMR in 0.7 mL "batch" samples containing the metal and the ligand at 3.32 mM and different concentrations of strong acid (0.05-0.20 M) aged 90 days (see below). The formation of protonated and ternary hydroxydo complex species was probed by pH-potentiometric titration by titrating the preformed complex (at a concentration of 1.95 mM) with standardized NaOH solution. For the calculation of the logK MH-1L value, the V NaOH -pH data used were obtained in the pH range 1.68-11.85. The pH-potentiometric titrations were carried out with a Metrohm 785 DMP Titrino titration workstation with the use of a Metrohm-6.0233.100-combined electrode calibrated via a two-point calibration routine using KH-phthalate (pH = 4.005) and borax (pH = 9.177) buffers. The titrated samples were stirred with a magnetic stirrer, and N 2 gas was bubbled through the solutions to avoid the effect of CO 2 . For the calculation of [H + ] from the measured pH values, the method proposed by Irving et al. was applied [33]. For this, a 0.01M HCl solution was titrated with the standardized NaOH solution in the presence of 0.15 M NaCl ionic strength at 37 • C. The differences between the measured (pH read ) and calculated pH (-log[H + ]) values were used to obtain the equilibrium H + concentration from the pH meas values during the titrations. The ionic product of water (pK w ) at 37 • C in 0.15 M NaCI was also calculated form these data and found to be pKw = 13.424. For the calculation of the equilibrium constants, the PSEQUAD program was used [34]. The dissociation kinetics of the Sc(DO3AM-NI) complex was probed in 1 M HCl solution by following the appearance and increase of 45 Sc-NMR signal of Sc(III) aq in the sample containing 7.30 mM Sc(DO3AM-NI) complex purified by HPLC.

Radiolabeling DO3AM-NI with 68 Ga 3+
A 68 Ge/ 68 Ga-generator was eluted with 0.1 M aq. HCl, then the eluate containing 68 Ga (100-120 MBq), was treated with the following post-processing method proposed by Eppard et al. [24]. 68 Ga was trapped on the cation exchange resin (Strata SCX) and washed with 0.15 M HCl in an 8:2 mixture of ethanol and water (1 mL). The purified 68 Ga was eluted with 0.9 M HCl in a 9:1 mixture of ethanol and water (3x100 µL). A fraction of 100 µL was transferred into an Eppendorf vial, then 60 µL of NaOAc/HOAc puffer (3M, pH = 4), 40 µL 5% NaOH, as well as 40 µL aq. stock solution of DO3A-NI (1 mg/mL) were added. The reaction mixture was kept at 95 • C for 15 min, then diluted with 1 mL water and passed through an SPE cartridge (SepPak ® C18 plus, Waters) preconditioned with 5 mL ethanol and 10 mL water. After purging of the cartridge with 1 mL of water, the labeled compound was eluted with a 1:1 mixture of ethanol and water (0.5 mL). The eluate was concentrated, then labeled product was dissolved in 100 µL of saline. The radiochemical purity of the product was determined with radio-HPLC on a Waters LC Module 1 HPLC with a Luna A total of 120 mg of natural calcium (99.99%) was pressed into a pellet and pushed into the cavity of an aluminum target holder. Then, 60 min of irradiation with 30 µA beam current yielded approx. 300 MBq 44 Sc. The irradiated Ca disc was dissolved in 3 M u.p. HCl (4 mL), and the solution was transferred onto a DGA cartridge preconditioned with 3 M u.p (3 mL) HCl, containing 70 mg resin. The cartridge was washed with 3 mL 3 M u.p. HCl and 3 mL 1 M HNO 3 and eluted with 2 mL 0.1 M u.p. HCl in 200 µL fractions. The highest activity fractions were merged, and a fraction of 500 µL (100-150 MBq) was transformed into an Eppendorf vial, then 100 µL of NaOAc/HOAc puffer (3M, pH = 4), 20 µL 5% NaOH, as well as 5 µL aq. Stock solution of DO3AM-NI (1 mg/mL) were added. The reaction was performed at 95 • C for 15 min. Then, the reaction mixture was passed through a pre-conditioned SPE cartridge (LiChrolut EN, Merck). After purging of the cartridge with 1 mL of water, the labeled compound was eluted with a 1:1 mixture of ethanol and water (0.5 mL). The eluate was concentrated, and then labeled product was dissolved in 100 µL of saline. The radio-HPLC analysis was performed as described above using a Kinetex C18 2.6 µm (100 × 4.6 mm) column.

Determination of logP Value of [ 68 Ga]Ga(DO3AM-NI) and [ 44 Sc]Sc(DO3AM-NI)
A 10 µL volume of [ 68 Ga]Ga(DO3AM-NI) and [ 44 Sc]Sc(DO3AM-NI) solution (~5 MBq) was mixed with 500 µL of 1-octanol and 490 µL of water in an Eppendorf vial, respectively. The mixture was shaken with a vortex shaker (600 rpm) for 10 min and centrifuged (6000 rpm) for 5 min. Then, 100 µL from the octanol phase and 1 µL from the aqueous phase were pipetted into vials, and the aqueous aliquot was diluted to 100 µL with water in order to minimize the effects of sample geometry and the high difference of activity concentrations in the two solvents. The radioactivity of the fractions was determined with a gamma counter. The measurements were performed in triplicates for both labeled compounds.  . Cells were cultured in Eagle's Minimum Essential Medium (Sigma-Aldrich) supplemented with 10% fetal bovine serum (FBS) and 1% Antibiotic Antimycotic Solution (Sigma-Aldrich, Saint Louis, MO, USA). For tumor induction, the cells were used after 6-8 passages and 85% confluence. The viability of the cells was always higher than 90%, as assessed by the trypan blue exclusion test.

Experimental Tumor Model
Adult male CB17 SCID mice (n = 25; from Charles River Laboratories by Innovo Kft., Hungary) were used at the age of 18 weeks. Mice were housed under sterile conditions in IVC cage system (Techniplast, Italy) at a temperature of 26 ± 3 • C, with 52 ± 10% humidity and artificial lighting with a circadian cycle of 12 h. Sterile semi-synthetic diet (Akronom Ltd., Budapest, Hungary) and sterile drinking water were available ad libitum to all the For tumor induction, SCID mice were injected with 5x10 6 KB tumor cells in 0.9% NaCl (100 µL) subcutaneously into the left shoulder area. The tumor growth was assessed by caliper measurements, and tumor size was calculated using the following formula: (largest diameter × smallest diameter 2 )/2. In vivo experiments were carried out approximately 13 ± 1 days after intravenous injection of tumor cells at the tumor volume of approximately 110 mm 3 .

In Vivo PET/MRI Imaging
For in vivo imaging studies, healthy control and KB tumor-bearing CB17 SCID mice were injected with 8.42 ± 0.38 MBq of 68 Ga-or 44 Sc-labeled DO3AM-NI via the lateral tail vein 13 ± 1 days after the inoculation of KB tumor cells. Then, 65 and 225 min after radiotracer injection, mice were anaesthetized by 3% isoflurane (Forane) with a dedicated small animal anesthesia device and whole-body T1-weighted MRI scans were performed (3D GRE EXT multi-FOV; TR/TE 15/2 ms; phase: 100; FOV 60 mm; NEX 2) using the preclinical nanoScan PET/MRI system with 1 Tesla magnetic field (Mediso Ltd., Hungary). After MRI imaging, 20 min static whole-body PET scans were acquired (90 min and 4 h after radiotracer injection).

Quantitative PET Data Analysis
Quantitative radiotracer uptake was expressed in terms of standardized uptake value (SUV), SUV = [VOI activity (Bq/mL)]/[injected activity (Bq)/animal weight (g)], assuming a density of 1 g/mL. Volumes of interest (VOI) were manually drawn around the edge of the organ or tumor activity using the InterView™ FUSION (Mediso Ltd., Hungary) image analysis software. Tumor-to-muscle (T/M) ratios were computed as the ratio between the activity in the tumor VOI and in the background (muscle) VOI. Skeletal muscles of the right shoulder area were used as background.

Ex Vivo Biodistribution Studies
For the determination of ex vivo biodistribution of the radiotracers, control and KB tumor-bearing mice were injected intravenously with 8.42 ± 0.38 MBq of 68 Ga-or 44 Sclabeled DO3AM-NI via the tail vein. Mice were euthanized with 5% isoflurane 90 min or 4 h after intravenous radiotracer injection. Three tissue samples were taken from the selected organs, and their weight and radioactivity were measured with a calibrated gamma counter (Perkin-Elmer Packard Cobra, Waltham, MA, USA). The radiotracer uptake was expressed as %ID/g tissue.

Data Analysis
Significance was calculated by two-way ANOVA, Mann-Whitney U-test and Student's t-test (two-tailed). IBM SPSS Statistics and Microsoft Excel software were used for the statistical analysis. The significance level was set at p ≤ 0.05, unless otherwise indicated. Data are presented as mean ± SD of at least three independent experiments.

Conclusions
We have performed the physicochemical studies of the Sc(DO3AM-NI) complex, and the present work shows that substitution of an acetamide functionality for the acetate pendant arm on DOTA significantly changes the properties of the ligand. We found that the moderate decrease in stability and of the Sc(DO3AM-NI) complex together with the great inertness compared to the Sc(DOTA) − complex clearly classify the DO3AM-NI ligand as an excellent Sc binder.
Furthermore, the radiolabeling of the DO3AM-NI ligand with a cyclotron produced 44 Sc radiometal, which was successfully carried out with a high labeling yield and radiochemical purity. In addition, we compared the pharmacokinetic properties of [ 44 Sc]Sc(DO3AM-NI) with the known hypoxia-specific [ 68 Ga]Ga(DO3AM-NI) radiotracer by in vivo PET/MRI studies and ex vivo biodistribution studies. No significant difference was found between the tumor-specific accumulation of 44 Sc-and 68 Ga-labeled DO3AM-NI in KB tumors. However, higher accumulation of [ 68 Ga]Ga(DO3AM-NI) was found in non-target tissues and organs compared to the accumulation of the 44 Sc-labeled analog, resulting in a higher tumor-tobackground ratio. Based on these results, we can conclude that the new [ 44 Sc]Sc(DO3AM-NI) radiotracer is a promising molecular probe for PET imaging of tumor hypoxia due to its favorable features and high specificity.

Informed Consent Statement: Not applicable.
Data Availability Statement: All data are contained in this article and its related supplementary information.

Conflicts of Interest:
The authors declare no conflict of interest.