Synthesis, Complexation Properties, and Evaluation of New Aminodiphosphonic Acids as Vector Molecules for 68Ga Radiopharmaceuticals

Two new aminodiphosphonic acids derived from salicylic acid and its phosphonic analogue were prepared through a simple and efficient synthesis. 2-[(2-Amino-2,2-diphosphono)ethyloxy]-benzoic acid 8 and 2-[(2-amino-2,2-diphosphono)ethyloxy]-5-ethyl-phenylphosphonic acid 9 were evaluated for their applicability as 68Ga binding bone-seeking agents. Protonation constants of 8 and 9 and stability constants of the Ga3+ complexes with 8 and 9 in water were determined. The stability constant of Ga3+ complex with fully phosphorylated acid 9 (logKGaL = 31.92 ± 0.32) significantly exceeds stability constant of Ga3+ complex with 8 (logKGaL = 26.63 ± 0.24). Ligands 8 and 9 are as effective for Ga3+ cation binding as ethylenediamine-N,N’-diacetic-N,N’-bis(methy1enephosphonic) acid and ethylenediamine-N,N,N’,N’-tetrakis(methylenephosphonic) acid, respectively. The labelling process and stability of [68Ga]Ga-8 and [68Ga]Ga-9 were studied. Both 8 and 9 readily form 68Ga-complexes stable to ten-fold dilution with saline. However, in fetal bovine serum, only [68Ga]Ga-9 was stable enough to be subject to biological evaluation. It was injected into rats with bone pathology and aseptic inflammation of soft tissues. For [68Ga]Ga-9 in animals with a bone pathology model in 60 and 120 min after injection, a slight accumulation in the pathology site, stable blood percentage level, and moderate accumulation in the liver were observed. For animals with an aseptic inflammation, the accumulation of [68Ga]Ga-9 in the pathology site was higher than that in animals with bone pathology. Moreover, the accumulation of [68Ga]Ga-9 in inflammation sites was more stable than that for [68Ga]Ga-citrate. The percentage of [68Ga]Ga-9 in the blood decreased from 3.1% ID/g (60 min) to 1.5% ID/g (120 min). Accumulation in the liver was comparable to that obtained for [68Ga]Ga-citrate.


Introduction
Generator-produced positron-emitting 68 Ga (T 1/2 = 67.71 min) is one of the most promising radionuclides [1,2]. 68 Ga radiopharmaceuticals are used for diagnostics of neuroendocrine tumors and prostate cancer, or for visualization of infection and inflammation [3][4][5]. Many studies are devoted to the search of osteotropic 68 Ga radiopharmaceuticals 5]. Many studies are devoted to the search of osteotropic 68 Ga radiopharmaceuticals for imaging bone metastases during the early stages [6][7][8][9]. Phosphonate ligands are commonly used for this purpose as they work with other radionuclides ( 99m Tc, 111 In, 153 Sm, 188 Re) [10,11]. 1-Amino-1,1-diphosphonic acids having fine complexation properties are of interest as ligands for radiopharmaceuticals. These compounds are structural diphosphonic analogues of physiologically important aminocarboxylic acids and have low toxicity [12,13]. Besides, they can be easily prepared by the addition of two phosphorous acid molecules to a nitrile group.
Salicylic acid and its phosphonic analogue were functionalized by alkylation of phenolic oxygen using monochloroacetonitrile. The obtained nitrile group was easily converted to the aminodiphosphonic group by the addition of two phosphorous acid molecules, resulting in acids 8 and 9 ( Figure 2). Previously, we used this reaction of phosphorous acid addition to a nitrile group in order to obtain a number of aminodiphosphonic acids, promising as physiologically active compounds [17,20]. Molecules 8 and 9 contain a fragment of salicylic acid 5a or its phosphonic analogue 5b and a structural diphosphonic isostere of serine, linked by an ether bond.
The starting compounds for the synthesis of new aminodiphosphonic acids 8 and 9 were isopropyl salicylate 5a (R = COO-iPr, X = H) and diethyl (2-hydroxy-5ethylphenyl)phosphonate 5b (R = P(O)(OEt)2, X = Et). The alkylation of phenols 5a and 5b with monochloroacetonitrile gave nitriles 6a and 6b, respectively ( Figure 2). The following addition of two molecules of phosphorous acid in dioxane in the presence of phosphorus tribromide led to aminodiphosphonic acids 7a and 7b (as monoethyl ester). The desired water-soluble form of 8 was obtained by hydrolysis of isopropyl ester 7a with 6N hydrochloric acid. An attempt to obtain phosphonic analogue 9 in a similar way led to the break of the phosphorus-carbon bond and the loss of the ortho-phosphonic fragment. Acid 9 was successfully obtained after treatment of ethylphosphonate 7b with trimethylsilyl bromide in an acetonitrile solution ( Figure 2).
According to 13 C and 31 P NMR, molecule 7a is likely a mixture of two conformers having a different orientation (syn:anti~1:4) of the aminodiphosphonic fragment and the isopropyl radical relative to the aromatic ring. The spectral data demonstrate the magnetic equivalence of phosphorus atoms of the aminodiphosphonic fragment. According to 31 P NMR, molecule 7b is a mixture of two conformers with a ratio of ~3:1 (see Materials and Methods: signals of a minor conformer are marked by an asterisk (*)). The 31 P NMR spectrum of each conformer includes two signals with a ratio of 1:2 from arylphosphonic and
Salicylic acid and its phosphonic analogue were functionalized by alkylation of phenolic oxygen using monochloroacetonitrile. The obtained nitrile group was easily converted to the aminodiphosphonic group by the addition of two phosphorous acid molecules, resulting in acids 8 and 9 ( Figure 2). Previously, we used this reaction of phosphorous acid addition to a nitrile group in order to obtain a number of aminodiphosphonic acids, promising as physiologically active compounds [17,20]. Molecules 8 and 9 contain a fragment of salicylic acid 5a or its phosphonic analogue 5b and a structural diphosphonic isostere of serine, linked by an ether bond.
The starting compounds for the synthesis of new aminodiphosphonic acids 8 and 9 were isopropyl salicylate 5a (R = COO-iPr, X = H) and diethyl (2-hydroxy-5-ethylphenyl) phosphonate 5b (R = P(O)(OEt) 2 , X = Et). The alkylation of phenols 5a and 5b with monochloroacetonitrile gave nitriles 6a and 6b, respectively ( Figure 2). The following addition of two molecules of phosphorous acid in dioxane in the presence of phosphorus tribromide led to aminodiphosphonic acids 7a and 7b (as monoethyl ester). The desired water-soluble form of 8 was obtained by hydrolysis of isopropyl ester 7a with 6N hydrochloric acid. An attempt to obtain phosphonic analogue 9 in a similar way led to the break of the phosphorus-carbon bond and the loss of the ortho-phosphonic fragment. Acid 9 was successfully obtained after treatment of ethylphosphonate 7b with trimethylsilyl bromide in an acetonitrile solution ( Figure 2).
According to 13 C and 31 P NMR, molecule 7a is likely a mixture of two conformers having a different orientation (syn:anti~1:4) of the aminodiphosphonic fragment and the isopropyl radical relative to the aromatic ring. The spectral data demonstrate the magnetic equivalence of phosphorus atoms of the aminodiphosphonic fragment. According to 31 P NMR, molecule 7b is a mixture of two conformers with a ratio of~3:1 (see Materials and Methods: signals of a minor conformer are marked by an asterisk (*)). The 31 P NMR spectrum of each conformer includes two signals with a ratio of 1:2 from arylphosphonic and aminodiphosphonic fragments. Thus, phosphonate 7b exists in the form of a mixture of two conformers with the syn-orientation and anti-orientation of the aminodiphosphonic fragment and the ester fragments of the phosphonic function relative to the aromatic ring, similarly to aminodiphosphonic acid 7a. The presence of the second form in the case of acid 9 is likely determined by the syn-orientation or anti-orientation of the methyl group in the ethyl radical and the aminodiphosphonic fragment. Perhaps, this explains the fact that the ratio of conformers, in this case, is~1:10.

Stability Constants of Gallium(III) Complexes with 8 and 9
One of the most important criteria for the suitability of ligands as components of radiopharmaceuticals is the stability of their complexes with radionuclides [3,21,22]. For this, the protonation constants of the ligands and the stability constants of their complexes with Ga 3+ were determined.
The protonation constants of 8 (H 5 L) and 9 (H 6 L) were determined at 298 Кand ionic strength of I = 0.1 M KCl. Stepwise equilibrium constants of acids are given in Table 1. Full constants are given in Table S1 (see Supplementary Materials). The values of the stepwise constants of 8 and 9 are in good agreement with data for other aminodiphosphonic acids [23]. The first three protonation constants (no. 1-3, Table 1) of 8 are similar to the corresponding constants of (aminoethylene)diphosphonic acid (AEDP), which is essentially a fragment of 8. The complexation reactions of 8 and 9 with Ga 3+ in water at 298 Кand ionic strength of I = 0.1 M KCl is well described by the model, which includes the complexation of Ga 3+ with a deprotonated ligand and addition to Ga 3+ of one to three (8) or one to four (9) protons besides the ligand (Table 2; full constants are given in Table S2, see Supplementary Materials). Stability constant of the Ga 3+ complex with deprotonated ligand 9 (logK GaL = 31.92 ± 0.32) significantly exceeds the corresponding stability constant of the Ga 3+ complex with 8 (logK GaL = 26.63 ± 0.24) ( Table 2). This is very consistent with the fact that acid 9 has the more active P(O)(OH) 2 group for the complexation instead of the COOH group of 8. This result suggests that ligand 9 is preferable over ligand 8 as a radiopharmaceutical component. The stability constants of the gallium(III)-deprotonated ligand complexes for 8 and 9 are similar to the stability of the complexes for ethylenediamine-N,N'-diacetic-N,N'bis(methylenephosphonic) acid (EDDADPO) (logK GaL = 26.82) and for the ethylenediamine-N,N,N',N'-tetrakis(methylenephosphonic) acid (EDTPO) (logK GaL = 31.83) [25] (Table 2 and  Table S2 in Supplementary Materials). Ligands 8 and 9 are effective ligands for gallium(III) as EDDADPO and EDTPO, respectively. Since [ 68 Ga]Ga-EDTPO showed promising results in µ-PET studies, we supposed that a study of osteotropic properties of 8 and 9 is of interest.

Radiolabelling and Stability
Both 8 and 9 ligands readily form complexes with 68 Ga at room temperature when using ligand concentrations >5 mM and pH 5 ± 2 as confirmed by radio-TLC. Both [ 68 Ga]Ga-8 and [ 68 Ga]Ga-9 obtained in this way were stable for ten-fold dilution with saline. However, when ten-fold diluted with fetal bovine serum, up to 90% and 50% of [ 68 Ga]Ga-8 and [ 68 Ga]Ga-9 correspondingly undergo decomposition. Elevation of 9 concentration to 20 mM and raising reaction temperature to 95 • C (30-min reaction time) allowed us to achieve 60-90% stability in a 2 hr interval. This approach did not work out for [ 68 Ga]Ga-8, and regardless of the concentration of 8 in the mixture, temperature, and reaction time, the stability of [ 68 Ga]Ga-8 was very low. These results are in good agreement with the stability constants of gallium(III) complexes with 8 and 9 in water. Thus, it is once more affirmed that phosphonic groups are significantly more favorable for M 3+ radiometals binding over carboxylic groups [26]. In the frame of this study, we decided to carry out further detailed studies using [ 68 Ga]Ga-9. Figure 3 shows the ligand concentration-labelling reaction yield dependence for the ligand 9 and previously studied 1,7-diamino-4-oxaheptane-1,1,7,7-tetraphosphonic acid (1) [10]. The conditions of 68 Ga-labelling for both compounds were the same: рН6, acetate concentration of 0.2 M, 25 • C, and 15 min reaction time. Ligand 1 contains two amidiphosphonic groups separated by a five-atoms ether chain with a weak donor oxygen atom. Ligand 9 contains one amidiphosphonic group and one phosphonic group in orto-position to it. According to potentiometric studies, ligand 1 forms one more protonated complex than ligand 9. The labelling reaction yield at low ligand concentration for [ 68 Ga]Ga-1 is higher than that for [ 68 Ga]Ga-9. Clearly, the structures of the Ga 3+ complexes of these ligands differ significantly. This can be attributed to the fact that amidiphosphonic groups may be more effective for 68 Ga binding than phosphonic groups. Another possible explanation may be the orto-position of the phosphonic group in 9 preventing the amidiphosphonic group from realizing its full binding potential toward 68 Ga.
The influence of the buffering agent type on the labelling reaction yield was analysed before for [ 68 Ga]Ga-1 [14]. The influence of acetate concentration on the labelling reaction yield was demonstrated for DOTA-conjugated molecules [27]. Here, the effect of acetate concentration was studied using four ligand 9 concentrations at a constant pH 5.9 ± 0.4. To observe the effect of ligand 9, the following concentrations were chosen: 0.8, 1.0, 2.0, and 4.0 mM. Results are presented in Figure 4a. In the case of [ 68 Ga]Ga-9, there is a distinct correlation: the lower the acetate concentration is, the higher is the labelling reaction yield with maximal yield achieved in the absence of acetate. This correlation becomes less significant with increasing ligand concentration and becomes statistically insignificant (p > 0.05) at 5 mM of 9. In addition to the results [14], we carried out similar experiments with [ 68 Ga]Ga-1 and found no statistically significant correlation even at a concentration of The influence of the buffering agent type on the labelling reaction yield was analysed before for [ 68 Ga]Ga-1 [14]. The influence of acetate concentration on the labelling reaction yield was demonstrated for DOTA-conjugated molecules [27]. Here, the effect of acetate concentration was studied using four ligand 9 concentrations at a constant pH 5.9 ± 0.4. To observe the effect of ligand 9, the following concentrations were chosen: 0.8, 1.0, 2.0, and 4.0 mM. Results are presented in Figure 4a. In the case of [ 68 Ga]Ga-9, there is a distinct correlation: the lower the acetate concentration is, the higher is the labelling reaction yield with maximal yield achieved in the absence of acetate. This correlation becomes less significant with increasing ligand concentration and becomes statistically insignificant (p > 0.05) at 5 mM of 9. In addition to the results [14], we carried out similar experiments with [ 68 Ga]Ga-1 and found no statistically significant correlation even at a concentration of 1 in the reaction mixture being as low as 0.2 mM. Thus, there are three patterns of acetate concentration influencing the labelling reaction yield. Gallium is known to form weak acetate complexes [28,29]. With logK = 3.7 [30], acetate is not able to compete with 9 for gallium binding. There is a possibility of a ternary Ga-9-OAc complex (or complexes) formation similar to that described for copper [18]. This possibility should be a subject of a separate study. In this study, acetate ion was added to reaction mixtures exclusively in the form of sodium acetate. The additional concentration of sodium in the reaction mixtures due to using NaOH for pH adjusting was ≤ 0.003 M. This suggests that sodium concentration in studied samples is virtually equal to that of acetate. To evaluate the influence of Na + on the labelling process, additional experiments with constant Na + concentrations were carried out. For this purpose, calculated amounts of NaCl were added to reaction mixtures at pH 5.7 ± 0.7 and 1 mM of 9. The comparison of the effects of dynamic and constant sodium concentration on the labelling reaction yield is presented in Figure 4b. At constant Na + concentration of 0.4 M, obtained RCP (radiochemical purity) values were consistently lower than those obtained at dynamic Na + concentration. However, the differences were mostly statistically insignificant (p > 0.05). Still, it is reasonable to assume that the acetate ion is the component responsible for the changes demonstrated in Figure 4a. The changes of water structure induced by an Na + presence were observed in 1 M sodium chloride and sodium acetate solutions [31]. The examination of this sodium concentration in our experiments resulted in much lower RCP values as compared with corresponding samples with dynamic Na + concentration (p < 0.05, Figure 4b). Thus, it is clear that Na + itself has an impact on the Ga-9 complex formation process.
(a) (b) Finally, taking into account previous results, the effect of pH on the labelling reaction yield was studied using reaction mixtures with 50 mM of acetate and 0.8, 1.0, and 2.0 mM of ligand 9. A maximal reaction yield was observed for the samples at pH 3-4 ( Figure 5), which is consistent with the data observed for 68 Ga [27,32]. According to the calculations based on the obtained stability constants, the protonated complex GaH4L + of ligand 9 dominates in this pH range.  Gallium is known to form weak acetate complexes [28,29]. With logK = 3.7 [30], acetate is not able to compete with 9 for gallium binding. There is a possibility of a ternary Ga-9-OAc complex (or complexes) formation similar to that described for copper [18]. This possibility should be a subject of a separate study. In this study, acetate ion was added to reaction mixtures exclusively in the form of sodium acetate. The additional concentration of sodium in the reaction mixtures due to using NaOH for pH adjusting was ≤ 0.003 M. This suggests that sodium concentration in studied samples is virtually equal to that of acetate. To evaluate the influence of Na + on the labelling process, additional experiments with constant Na + concentrations were carried out. For this purpose, calculated amounts of NaCl were added to reaction mixtures at pH 5.7 ± 0.7 and 1 mM of 9. The comparison of the effects of dynamic and constant sodium concentration on the labelling reaction yield is presented in Figure 4b. At constant Na + concentration of 0.4 M, obtained RCP (radiochemical purity) values were consistently lower than those obtained at dynamic Na + concentration. However, the differences were mostly statistically insignificant (p > 0.05). Still, it is reasonable to assume that the acetate ion is the component responsible for the changes demonstrated in Figure 4a. The changes of water structure induced by an Na + presence were observed in 1 M sodium chloride and sodium acetate solutions [31]. The examination of this sodium concentration in our experiments resulted in much lower RCP values as compared with corresponding samples with dynamic Na + concentration (p < 0.05, Figure 4b). Thus, it is clear that Na + itself has an impact on the Ga-9 complex formation process.
Finally, taking into account previous results, the effect of pH on the labelling reaction yield was studied using reaction mixtures with 50 mM of acetate and 0.8, 1.0, and 2.0 mM of ligand 9. A maximal reaction yield was observed for the samples at pH 3-4 ( Figure 5), which is consistent with the data observed for 68 Ga [27,32]. According to the calculations based on the obtained stability constants, the protonated complex GaH 4 L + of ligand 9 dominates in this pH range.  Finally, taking into account previous results, the effect of pH on the labelling reaction yield was studied using reaction mixtures with 50 mM of acetate and 0.8, 1.0, and 2.0 mM of ligand 9. A maximal reaction yield was observed for the samples at pH 3-4 ( Figure 5), which is consistent with the data observed for 68 Ga [27,32]. According to the calculations based on the obtained stability constants, the protonated complex GaH4L + of ligand 9 dominates in this pH range.

Biodistribution of [ 68 Ga]Ga-9
In Table 3, the biodistribution data of [ 68 Ga]Ga-9 and [ 68 Ga]Ga-acetate in animals with fractures are presented. Non-target biodistribution pathways and bone pathology uptake for [ 68 Ga]Ga-9 are comparable to those of [ 68 Ga]Ga-1 and [ 68 Ga]Ga-2 [14]. The pathology site/intact bone ratio for [ 68 Ga]Ga-9 is inferior to that of [ 68 Ga]Ga-oxa-bis-ethylenenitrile tetra(methylene phosphonic acid) (3) [33] and even to that of [ 68 Ga]Ga-acetate studied in this experiment (Table 3). Moreover, [ 68 Ga]Ga-3 uptake in blood, liver, intestine, and kidneys is lower than that of [ 68 Ga]Ga-9. Likely, it depends on the different stability of 68 Ga-labelled complexes in vivo and requires additional research. Since fracture healing may be accompanied by an inflammatory process, animals with an aseptic inflammation model were also studied. In Table 4, the data on the biodistribution of [ 68 Ga]Ga-9 in animals with aseptic inflammation are presented along with data for [ 68 Ga]Ga-citrate, which is known to have an inflammation imaging potential [34]. During the comparison of biodistribution dynamics, it was found that [ 68 Ga]Ga-9 can be a potential agent for aseptic inflammation imaging more promising than [ 68 Ga]Ga-citrate. Activity in blood 120 min after injection in comparison to a 60-min time point decreases three and two times for [ 68 Ga]Ga-9 and [ 68 Ga]Ga-citrate, respectively. [ 68 Ga]Ga-9 pathology site/muscular tissue ratio is almost constant during the time of observation. It will allow imaging pathology foci 1 h after i.v. injection (for [ 68 Ga]Ga-citrate-2 h). a Specific activity accumulation was measured as a fraction (%) of the injected dose per gram of the considered organ or tissue (%ID/g). b Activity accumulation was measured as a fraction (%) of the injected dose per the considered organ (%ID/organ).

Synthesis
The progress of the reactions was monitored by 31 P NMR spectroscopy. All chemicals and solvents were purchased from Acros Organics (Acrus, Moscow, Russia) and Alfa Aesar (Reakor, Moscow, Russia). The 1 H, 31   NMR spectra of synthesized compounds are given in Supplementary materials.

Stability Constant Measurements and Calculations
The potentiometric titration technique using the OP-300 Radelkis potentiometer was described earlier [35]. Solutions of 8 and 9 were titrated with a standard 0.1 МNaOH solution at 298 ± 0.1 K and ionic strength of I = 0.1 M KCl. The initial volume of solutions was 160 mL. Titrations were performed in the range of pH 3.0-11.6 (8) and 3.5-11.5 (9). Experiments included from 33 to 52 (8) and from 31 to 66 (9) data points. The initial analytical concentrations were 0.35-0.92 mM (8) and 0.27-0.48 mM (9). The protonation constants were estimated from four titrations using the CHEMEQUI program [36] freely available on the server [37]. CHEMEQUI evaluates equilibrium constants using four different algorithms: EQ, SIMPLEX, MONTE-CARLO, and the genetic algorithm SDE [38]. Estimation of the constants was performed based on each titration. In the case of significant correlations between the protonation constants, resulting in a shift of the constants, the calculations were performed simultaneously based on several titrations. All the computational results were used to calculate the average values of the estimated full constants logβ and their standard deviations. The average values were determined from 17 (8) and 12 (9) calculations based on several titrations and algorithms. The errors in the stepwise logK values are evaluated using standard deviations in estimated full equilibrium constants logβ and an error propagation rule for several titrations and applied algorithm calculations.
The Hamilton's R-factor (HRF) and the squared determination coefficient (R 2 det ) were used as the agreement criteria of the proposed set of equilibrium reactions and calculated constants with the experimental data [35]. Depending on the experiment and the program algorithm, HRF was 0.30-0.96% (8), 0.46-0.92% (9) at the calculations of protonation constants, 0.99-1.50% (8 + Ga 3+ ), and 1.41-3.18% (9 + Ga 3+ ) at the calculations of stability constants of the Ga 3+ complexes.

Radiolabelling, Stability, and Quality Control
All chemicals used for labelling were of a "reagent pure" or "extra pure" grade (Sigma-Aldrich, Panreac). The 68 Ge/ 68 Ga generators (Cyclotron Ltd., Obninsk, Russia) with activity 20 and 50 mCi were used (0-12 months after the production). The generator was eluted with 0.1 M HCl as per the manufacturer's instruction.
Radiolabelling and evaluation of [ 68 Ga]Ga-8 and [ 68 Ga]Ga-9 stability were carried out in triplicate using procedures similar to those described in Reference [14]. In short, for labelling with 68 Ga sodium acetate (buffering agent) solutions of various concentrations, (0.1, 0.5, or 1 M) and 68 Ge/ 68 Ga generator eluate were added to the Eppendorf tubes containing 0.1-20 µmol of 8 or 9. pH of the mixtures was then adjusted using NaOH and HCl of pre-estimated concentrations. The reaction mixtures were stirred at 25 or 95 • C for 15-30 min. To evaluate the stability of the complexes obtained, 100 µL of a sample was added to 1000 µL of saline or fetal bovine serum. The mixtures were stirred at 37 • C for 2 h.

Biodistribution Studies
All experiments involving animals were performed following the ethical standards, Russian animal protection laws, and guidelines for scientific animal trials [43].
Animal studies were performed using female outbred albino rats with model pathologies. Animals with fractures (active bone callus formation) [44] had been grouped (N = 3) and [ 68 Ga]Ga-9 (100 µL per rat) was i.v. injected into the tail vein. At preselected time points (60, 120 min), animals were obtained from the experiment using partial decapitation. The organs of interest were collected, blotted dry, and weighed. Radioactivity in samples of organs/tissues was counted using a WIZARD 2 automatic γ-counter (PerkinElmer). The results are expressed as the percentage of injected activity dose per gram (mean % ID or mean % ID/g ± SD) for each organ/tissue. For comparison, the mixture of 68 Ge/ 68 Ga generator eluate with sodium acetate solution (pH 6.5 ± 0.5, 0.18 M total acetate concentration) was also injected into animals with bone pathology.
In addition, animals with a model of aseptic inflammation were used. The site of aseptic soft tissue inflammation was modelled by intramuscular injection of 0.2 mL of turpentine into the rat pelvic limb. An acute inflammatory reaction was observed 7 days after administration. The inflammation foci were marked by swelling of the tissue, which is sharply painful on palpation. An autopsy revealed a burn of soft tissues with elements of necrosis, pronounced as a vascular pattern. [ 68 Ga]Ga-9 was studied using these animals in the same way it was done for animals with fractures. For comparison, the mixture of the 68 Ge/ 68 Ga generator eluate with sodium citrate solution (pH 5.0 ± 0.5, 0.084 M total citrate concentration) was also injected into animals.

Conclusions
The combination of aminodiphosphonic fragment with salicylic acid or its phosphonic analogue into one molecule is a promising way to develop radiopharmaceuticals. According to this technique, two new ligands with high complexation ability to gallium(III) were synthesized. Introducing phosphoryl fragment instead of carbonyl increases stability constants of the gallium(III) complexes in water. Stability constant of the Ga 3+ complex with fully phosphorylated acid 9 (logK GaL = 31.92 ± 0.32) significantly exceeds stability constant of Ga 3+ complex with 8 (logK GaL = 26.63 ± 0.24). Ligands 8 and 9 are as effective for Ga 3+ cation binding as ethylenediamine-N,N'-diacetic-N,N'-bis(methy1enephosphonic) acid and ethylenediamine-N,N,N',N'-tetrakis(methylenephosphonic) acid, respectively.
Both new molecules readily form 68 Ga-complexes stable by ten-fold dilution with saline. However, in fetal bovine serum only, [ 68 Ga]Ga-9 was stable enough to be subject to biological evaluation. It was injected into rats with bone pathology and aseptic inflammation of soft tissues. In vivo studies revealed that [ 68 Ga]Ga-9 is not suitable as a bone-seeking agent, but it can be used for inflammation imaging. To an extent, as inflammation imaging, [ 68 Ga]Ga-9 is preferable over [ 68 Ga]Ga-Citrate due to delayed free 68 Ga release from the complex.
In addition, the 68 Ga-labelling reaction with 9 was studied in detail. A correlation of acetate concentration in the reaction mixture and labelling reaction was found (up to 5 mM of 9): the lower the acetate concentration is, the higher the labelling reaction yield is.