- Email: [email protected]
S] donor atom set
Nuclear Medicine and Biology 31 (2004) 785–793
www.elsevier.com/locate/nucmedbio
Synthesis and biological evaluation of silylated mixed-ligand complexes with the [PNS/S] donor atom set
99m
Tc
C. Fernandesa, T. Kniessa, L. Ganoa, S. Seifertb, H. Spiesb, I. Santosa,* b
a Departamento de Quı´mica, ITN, Estrada Nacional 10, 2686-953, Sacave´m Codex, Portugal Institute of Bioinorganic & Radiopharmaceutical Chemistry, Forschungszentrum Rossendorf, D-01314 Dresden, Germany
Received 4 September 2003; received in revised form 12 March 2004; accepted 26 March 2004
Abstract New oxotechnetium complexes of general formula [99mTc(O)(PNS)(S(CH2)nOSiR3)] (4–6) were synthesized by direct reduction of [ TcO4]⫺ with stannous chloride, in the presence of the tridentate heterofunctionalized phosphine H2PNS and of the monodentate silylated thiols [HS(CH2)nOSiR3] (n⫽2, R⫽Ph (1); n⫽3, R⫽Ph (2); n⫽3, R⫽Et (3)). The mixed-ligand rhenium and technetium complexes of general formula [M(O)(PNS)(S(CH2)nOH)] (n⫽2: M⫽99mTc, (7), M⫽Re, (7a); n⫽3: M⫽99mTc, (8), M⫽Re, (8a)) were also prepared. All the 99mTc complexes were obtained with high radiochemical purity (⬎95%), after purification by HPLC, and were characterized by comparison of their HPLC profiles with the ones obtained for the corresponding Re compounds. The silylated compounds 4–6 are stable in phosphate saline buffer (PBS) pH 7.4, rat plasma, human serum and whole blood, and do not bind to plasmatic proteins, and also do not challenge with glutathione. The biological behavior of [99mTc(O)(PNS)(S(CH2)nOH)] (7, 8) and [99mTc(O)(PNS)(S(CH2)nOSiR3)] (4–6) was studied. The effect of the pH on the cleavage of the O-Si bond in complexes 4–6 was also evaluated. © 2004 Elsevier Inc. All rights reserved. 99m
Keywords: Mixed-ligand complexes;
99m
Tc; Biological evaluation
1. Introduction The transportation and accumulation into the target organ is an important issue for 99mTc radiopharmaceuticals development. Previous studies have shown that silylation of biologically active molecules facilitates their transportation in the organism [1– 4]. Applying the same principle to rhenium and 99mTc complexes it was expected to improve their bioavailability, especially in terms of passing the blood-brain barrier with consequent increase of brain uptake. The accumulation into the target organ would be enhanced if we were able to find silyl ethers stable in physiological conditions but hydrolysable in the target organ. The first attempts for preparing silylated mixed-ligand Re and Tc complexes were performed with the tridentate 3-thiapentane-1,5-dithiol (HSSSH) [5–7]. However, the instability of these compounds [8,9] led us to explore a new family of tridentate heterofunctionalized phosphines, namely 2-(diphenylphosphanyl)-N-(2-thioethyl)benzamide * Corresponding author. Tel.: ⫹00351219946201; fax: ⫹00351219941455. E-mail address: [email protected] (I.Santos). 0969-8051/04/$ – see front matter © 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.nucmedbio.2004.03.011
(H2PNS), 2-(diphenylphosphanyl)-N-(2-aminoethyl)benzamide (H2PNN) and 2-(diphenylphosphanyl)-N-(2-hydroxyethyl)benzamide (H2PNO), recently introduced by our group [10 –13]. These chelators are very versatile stabilizing tricarbonyl compounds as well as mixed-ligand oxocomplexes, some of them bearing monodentate silylated coligands [10 –14]. Herein, we report on the preparation of 99m Tc mixed-ligand oxocomplexes with monodentate silylated co-ligands, [99mTc(O)(PNS)(S(CH2)nOSiR3)] (4–6). The stability and biological behavior of these compounds will be also reported. Taking into account that in vitro or in vivo hydrolysis of 4–6 can promote the breaking of the O-Si bond, we have also synthesized and characterized the complexes [M(O)(PNS)(S(CH2)nOH)] (M⫽Re, 99mTc; n⫽2, 3) (7,8), which are also described herein. 2. Experimental 2.1. Materials and methods Unless otherwise stated, all chemicals were of reagent grade and were used without further purification. 2-mer-
786
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
capto-ethanol, 3-mercapto-1-propanol, chloro triethylsilane, chloro triphenylsilane, imidazole and stannous chloride were obtained from Aldrich. (99mTc)TcO4⫺ in saline solution was eluted from a 99Mo/99mTc generator from MDS Nordion S.A. (Belgium). Elemental analysis was performed on a Perkin Elmer automatic analyzer. The 1H and 31P NMR spectra were recorded on a VARIAN Unity 300MHz spectrometer; 1H chemical shifts were referenced relative to tetramethylsilane and the 31P chemical shifts to external 85% H3PO4 solution. The NMR samples were prepared in CDCl3, chemical shifts are given in parts per million. Chemical reactions were monitored by thin-layer chromatography (TLC) on Merck plates precoated with silica gel 60 F254. Spots were visualized either by UV light or Ellman’s reagent for thiols. Column chromatography was performed in silica gel 60 (Merck). HPLC investigations were carried out using an RP C18 column (Hypersil ODS, 4.0⫻250 mm, 10 m) and using as eluent a gradient of acetonitrile/water with a flow rate of 1.0 ml/min. The identification of the compounds was made using a UV-VIS detector (LC 290, Perkin Elmer) or a ␥-detector (LB 509, Berthold). [nBu4N][Re(O)Cl4] was prepared according to literature methods [15]. The hydroxyl silylated monodentate thiols (1–3) and the 2-(diphenylphosphanyl)-N-(2-thioethyl)benzamide (H2PNS) were synthesized as previously reported [6,11]. 2.2. General procedure for the preparation of complexes 7a and 8a [nBu4N][Re(O)Cl4] (174 mol) was dissolved in 8 ml of dichloromethane and stirred at room temperature (r.t.) under a nitrogen atmosphere. A solution of 174 mol of H2PNS, 250 mol of 3-mercapto-1-propanol and 100 l triethylamine (720 mol) in 4 ml dichloromethane was added drop wise and the mixture stirred for 6 hours. The solvent was evaporated and the complexes purified by column chromatography (silicagel, dichloromethane/ethyl acetate (90:10)). 2.2.1. [Re(O)(3-PNS)(1-S-CH2-CH2-OH)] (7a) Yield 15 %. 1H NMR (␦, ppm, CDCl3): 2.02 (m, 2H, CH 2), 2.22 (m, 1H, CHtrid), 2.70 (m, 1H, CHtrid), 2.95 (m, 1H, CHtrid), 3.80 –3.94 (m, 2H, CH2), 5.08 (m, 1H, CHtrid), 6.83 (m, 1H, CHar), 7.20 –7.66 (m, 12H, CHar), 8.12 (m, 1H, CH ar). Elemental analysis: C23H23O3S2NPRe, calc., C 42.98, H 3.61, N 2.18 S 9.98, found, C 40.88, H 4.17, N 2.02, S 8.78. 2.2.2. [Re(O)(3-PNS)(1-S-CH2-CH2-CH2-OH)] (8a) Yield 20 %. 1H NMR (␦, ppm, CDCl3): 2.04 (m, 2H, CH 2), 2.23 (m, 1H, CHtrid), 2.73 (m, 1H, CHtrid), 2.99 (m, 1H, CHtrid), 3.77–3.98 (m, 4H, 2⫻CH2), 5.08 (m, 1H, CHtrid), 6.80 (m, 1H, CHar), 7.26 –7.60 (m, 12H, CHar), 8.12 (m, 1H, CHar). Elemental analysis: C24H25O3S2NPRe, calc., C
43.89, H 3.84, N 2.13, S 9.76, found, C 43.09, H 4.50, N 2.11, S 9.17. 2.3. General procedure for the preparation of complexes: complexes 4–6
99m
Tc
The 99mTc complexes were prepared by direct reduction of pertechnetate with stannous chloride. The following general procedure was used: 50 l of propylene glycol (PG), 1300 l of ethanol, 0.5 mg of the monodentate ligand (1–3), 0.03 mg (8.2⫻10⫺8 mol) of H2PNS ligand as well as 10 l of a SnCl2 solution (2.0 –3.0 mg SnCl2 dissolved in 5 ml ethanol) were added to 0.5 ml (100 –200 MBq) of [99mTc]pertechnetate eluate. The reaction mixture was heated at 40 –50°C for 5–15 min, depending on the monodentate ligand. The 99mTc complexes were analyzed by HPLC using a reverse-phase Hypersil 120 ODS column (250⫻4 mm, 10 m). The column was eluted using acetonitrile/water as mobile phase at 1.0 ml/min flow rate and with a linear gradient of 70 –100% acetonitrile over 5 min. followed by 10 min at 100% of acetonitrile. These complexes were also analyzed by TLC using two different chromatographic systems: (a) silicagel/acetone and (b) silicagel/saline. The compounds 4–6 were purified on a semi-preparative Hypersil ODS column (250⫻8 mm, 10 m) using acetonitrile/water as mobile phase (flow rate: 2 ml/min) and with a linear gradient of 70 –100% acetonitrile over 5 min followed by 20 min at 100% of acetonitrile. The fractions were collected and analyzed in the analytical column. The purified complexes, after removal of the organic solvent, were reconstituted with 0.1 M PBS 7.4 and 100 l of PG was added. The characterization of the 99mTc complexes was accomplished by chromatographic correlation (HPLC) with the corresponding rhenium complexes [14]. 2.4. Complexes 7 and 8 The complexes 7 and 8 were synthesized using two different procedures: (a) Incubation of complexes 4 and 6 with a diluted HCl solution, at 50°C. (b) Heating of complexes 4 and 6 at 80°C during 20 min, immediately after labeling. The hydrolyzed complexes (7, 8) were purified by semi-preparative HPLC, using the experimental conditions above described. 2.5. General procedures for stability studies 2.5.1. Challenge with GSH 100 l (approximately 10 MBq) of the purified 99mTc complex was mixed with 100 l of a 2 or 20 mM solution of GSH in 0.1 M phosphate buffer (pH 7.4). The mixture was incubated at 37°C for 5–120 min and analyzed by HPLC. Samples without GSH were also analyzed and used for comparison. HPLC analyzes were carried out using the experimental conditions previously described. Compounds:
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
4, 87%, 2 hours, 10 mM; 5, 92%, 2 hours, 10mM; 6, 88%, 2 hours, 10mM. 2.5.2. Stability in rat plasma Blood collected from mice in heparinized polypropylene tubes were immediately centrifuged for 15 min at 2,000 rpm at 4°C and the plasma collected. To 1.0 ml of fresh rat plasma 100 l (approximately 10 MBq) of the 99mTc complex, were added and the mixture was incubated at 37°C. At appropriate periods of time (5, 30 min, 1, 2, and 4 hours), 100 l aliquots (in duplicate) were removed and treated with 200 l of ethanol to precipitate the proteins. Samples were then cooled at 4°C and centrifuged at 3,000 rpm for 15 min at 4°C. The supernatant was separated from the precipitate and the sediment washed twice with 1 ml of ethanol and counted in a gamma counter. The activity in the sediment was compared with the total activity used and the percent of complex bound to proteins calculated. The supernatant was analyzed by HPLC, using the experimental conditions above described, to test the stability of the complex in rat plasma. Compounds: 4, 95%, 4 hours; 5, 98%, 4 hours; 6, 90%, 2 hours. 2.5.3. Stability in whole human blood To 1 ml of whole human blood, collected in heparinized polypropylene tubes was added a solution of the 99mTc complex (⬇10 MBq) and the mixture incubated at 37°C. Samples were taken, 5 min, 1 hour, and 4 hours after incubation, and centrifuged 15 min at 2,000 rpm at 4°C. The plasma was separated and ethanol was added in a 2:1 (v/v) ratio. The samples were centrifuged at 3,000 rpm and the supernatant analyzed by HPLC. Compounds: 4, 93%, 4 hours; 5, 90%, 4 hours; 6, 93%, 4 hours. 2.6. Biodistribution studies The in vivo behavior of the 99mTc complexes was evaluated in groups of 3–5 female CD-1 mice (randomly bred, Charles River) weighting approximately 20 –25 g each. Animals were intravenously injected with 100 l (1.5– 4.5 MBq) of each preparation via the tail vein and were maintained on normal diet ad libitum. At 5 min and 1 hour mice were killed by cervical dislocation. The radioactive dosage administered and the radioactivity in the sacrificed animal was measured in a dose calibrator (Aloka, Curiemeter IGC-3, Tokyo, Japan). The difference between the radioactivity in the injected and sacrificed animal was assumed to be due to excretion, mainly urinary excretion. Blood samples were taken by cardiac puncture at sacrifice. Tissue samples of the main organs were then removed and counted in a gamma counter (Berthold). Biodistribution results were expressed as percent of injected dose per organ (% I.D./total organ). For blood, bone and muscle, total activity was calculated assuming that these organs constitute 6, 10, and 40% of the total weight, respectively. The remaining activity in the carcass was also measured in a dose calibrator.
787
Afterwards blood was centrifuged and the serum separated for RP-HPLC analysis.
3. Results and discussion The 99mTc complexes [99mTc(O)(PNS)(S(CH2)nOSiR3)] (4–6) were obtained by direct reduction of [99mTcO4]⫺ with stannous chloride in the presence of the H2PNS and of the monodentate [HS(CH2)nOSiR3] ligands (n⫽2, R⫽Ph (1); n⫽3, R⫽Ph (2); n⫽3, R⫽Et (3)) (Scheme 1). The characterization of these complexes has been made by comparing their HPLC profiles with the profiles of the analogous Re compounds, recently reported by our group [14]. As indicated in Scheme 1, the preparation of complexes 4–6 was achieved using stannous chloride dissolved in ethanol, freshly prepared, and by heating the mixture for 5–15 min, at 40 –50°C. This methodology is quite different from the ones described for preparing hydroxyl silylated mixedligand complexes with the [SSS/S] donor atom set or for preparing 3⫹1 99mTc complexes with the [PNS/S] donor atom set [6,13]. In fact, the mixed-ligand complexes with the [SSS/S] donor atom set were obtained by gluconate transchelatation and the compounds anchored in the [PNS/ S] donor atom set were obtained by reducing the [99mTcO ⫺ 4] with stannous chloride in aqueous acidic media. When we tried to prepare complexes 4–6 using the first methodology no labeling was observed, and using the second procedure only hydrolyzed species could be isolated. Another important point for preparing complexes 4–6, with relatively high yield, is the use of a large excess of the monodentate ligand, relatively to the tridentate H2PNS. These experimental conditions are important as they avoid the formation of complexes 4–6 contaminated with the complex [99mTc(O)(3-PNS)(1-HPNS)]. This radiochemical impurity has been characterized by comparing its HPLC profile with the profile of the analogous Re complex, previously described [11]. Complexes 4–6 were obtained in 60 –90 % yield, as indicated by HPLC and by TLC. Using TLC (silicagel/ saline) only minor amounts of reduced hydrolyzed technetium could be detected (2–5%). Considering that one of the radiochemical impurities in the preparation of complexes 4–6 could be the hydrolyzed complexes [M(O)(PNS)(S(CH2)nOH] (M⫽Re, 99mTc) (7, n⫽2; 8a, n⫽3) we also prepared and characterized these compounds with Re and 99mTc. The rhenium complexes with mercaptoethanol and mercaptopropanol [Re(O)(PNS)(S(CH2)nOH] (7a, n⫽2; 8n⫽3) were obtained by reacting [nBu4N][Re(O)Cl4] with H2PNS and mercaptoethanol or mercaptopropanol, in the presence of a slight excess of triethylamine (Scheme 1). The synthesis was carried out under nitrogen atmosphere, to avoid the oxidation of the thiol or of the phosphine ligands. Compounds 7a and 8a, after purification by column chromatography and dried under vacuum, were obtained as grey green
788
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
Scheme 1. Synthesis of the complexes.
solids in 15–20 % yield. The low yield of these reactions is due to the formation of some competing complexes. Only one of these species was unambiguously identified as being [Re(O)(3-PNS)-(1-HPNS)] [11]. The other complexes in the mixture must be homoleptic and anchored by the alcohols, as no PNS ligand could be identified by 1H or 31P NMR spectroscopy. The 31P NMR spectra of compounds 7a and 8a show only one singlet at 19.7 and at 19.8 ppm, respectively. These values compare well with the values previously reported for other 3⫹1 oxocomplexes with the [PNS/S] donor atom set [11,14]. In the 1H NMR spectra of complexes 7a and 8a appear four sets of multiplets, integrating for one proton each, due to the methylenic protons of the PNS ligand, which are diastereotopic. The resonances due to the aromatic rings of the PNS and due to the methylenic protons of the co-ligand appear in the expected range [11,14]. The synthesis of the analogous 99mTc complexes 7 and 8 was not possible using the methodology described for 4–6 (Scheme 1). In fact, using this procedure only very hydrophilic species were formed, which were not characterized at the macroscopic level. The quantitative synthesis of 7 and 8 was achieved by hydrolysis of the purified silylated complexes 4 and 6. This hydrolysis was performed by adding diluted HCl to 4 and 6 and by heating at 50°C during 20 min. Another procedure consisted on heating 4 or 6 at 80°C during 20 min., in the labeling conditions, followed by HPLC purification. All the 99mTc complexes described in this work (4–8) were obtained with high radiochemical purity (⬎95%), after HPLC purification. Their characterization was made by
HPLC, comparing their retention time (␥ detection) with the retention time of the well-defined analogous rhenium complexes (UV detection at 254 nm) (Fig. 1 and Table 1). 3.1. Stability The 99mTc complexes, after evaporation of the solvents, were reconstituted with 0.1 M PBS pH 7.4 and their stability studied. We found that complexes 4–8 are stable in 0.1 M PBS 7.4, for at least 4 hours at 37°C, which is an important parameter to pursue with biological studies. As previously reported, a drawback of the 3⫹1 Tc complexes of the type [TcO(SES)(SR)] (E⫽S, O, NR⬘) is their instability when in the presence of thiolated nucleophiles, like glutathione [8,9,16]. We also found that the silylated 99m Tc complexes stabilized by the SSS tridentate ligand, although stable in PBS 7.4 at 37°C, display a high rate of exchange with glutathione (1 mM solutions, 5–120 min at 37°C) (Fig. 2) [17]. Taking this into account, we studied the stability of the silylated 99mTc complexes 4–6 in challenge reactions with glutathione, in rat plasma, in human serum and in whole blood, before proceeding with in vivo biological studies. We found that 4–6 do not challenge with glutathione, even in the presence of concentrated solutions (1 and 10 mM GSH dissolved in 0.1M phosphate buffer pH 7.4, 5–120 min at 37°C) and they also remain intact in fresh rat plasma (37°C, 5 min– 4 hours), and do not bind significantly to plasmatic proteins (Table 1). Even in the presence of whole human blood, with remarkable GSH concentration, all the complexes remained intact in the supernatant and the radioac-
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
789
Fig. 1. HPLC chromatograms for complexes 5 (␥ detection, RT⫽11.6 min) and 5a (UV detection at 254 nm, RT⫽11.2 min).
tivity bound to the proteins is very low (2–13%). These results indicate a high stability for compounds 4–6 showing that no transchelatation or metabolic process did take place (Table 1). This is an interesting result which confirms the importance of the nature of the tridentate ligand on the stability of the mixed-ligand “3⫹1” 99mTc complexes [8,9]. We also investigated the pH influence on the cleavage of the O-Si bond in complexes [99mTc(O)(PNS)(S(CH2)nOSiR3)] (4–6). Fig. 3 shows the hydrolysis curves for complex 6, at pH 6 and 7.4, as well as the HPLC profile of one sample which has been incubated over 1 hour, at pH 6 and at 37°C. As can be seen, in physiological conditions (pH⫽7.4, 37°C) 6 is relatively stable but rapid cleavage of the O-Si bond was observed at pH 6.0, 37°C. However, complexes 4 and 5, with a triphenyl substituent, are stable in the same conditions, requiring a much low pH to be hydrolyzed (Fig. 4). 3.2. Lipophilicity The lipophilicity of all the compounds was evaluated by determination of the partition coefficient (P) in physiological conditions (n-octanol/0.1 M phosphate buffer pH 7.4). The results obtained, expressed as log Po/w, are presented in Table 1. As expected, there is an increase of the lipophilicity when the -OH group is blocked by SiR3 groups.
3.3. Biodistribution The in vivo behavior of the 99mTc complexes (4–8) was evaluated in mice at 5 and 60 min post intravenous injection. Fig. 5 shows the results expressed as % I.D./total organ in the most relevant organs. All the complexes have high initial blood, muscle and liver uptake as expected for lipophilic compounds. Nevertheless, the blood and the muscle clearance is quite fast for all the complexes. Excretion occurs mainly through the hepatobiliary tract with quite low urinary elimination. The liver uptake is higher and faster for the complexes blocked with the triphenyl substituent at the silicium atom (4 and 5) than for the complex with the triethyl substituent (6). In the case of 4 and 5 the activity increases over the period studied, whereas it decreases for complex 6. The introduction of the triethyl substituent in the ligand backbone induces also a significantly faster hepatobiliary excretion and a slightly higher kidney uptake. It was expected that the introduction of the silyl groups in 99mTc complexes would increase their lipophilicity and consequently their bioavailability, brain uptake and brain retention promoted by hydrolyze in the target organ. In fact, the lipophilicity of the 99mTc complexes was significantly increased by silylation (Table 1) but no effect was found in the brain uptake (⬍0.09 % I.D.) or brain retention. We found that, despite their high lipophilicity, compounds 4–6 present a lower brain uptake than the model
Table 1 Some properties of compounds 4 – 8 (nd-not determined) Compound
4/4a 5/5a 6/6a 7/7a 8/8a
Labelling yield (%)
70–90 60–90 60–70 85–90 85–90
RT (min)
11.4/11.5 11.1/11.6 11.9/12.1 3.6/3.8 3.8/4.0
log Po/w
2.26 2.40 2.32 1.70 1.81
Protein binding (%) 5 min.
1h
8.7 2.4 2.8 nd nd
12.1 12.5 8.1 nd nd
790
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
Fig. 2. Stability of complex [99mTc(O)(SSS/S(CH2)3OSiPh3] in 0.1 M PBS pH 7.4 and in presence of 1 mM GSH solution at 37°C.
complex [99mTcO(PNS/SPh)] (log P⫽1.7, I.D./organ ⫽0.14% at 5 min p.a.) [13]. This difference can probably be explained by the high molecular weight of the silylated
complexes 4–6. The results obtained clearly indicated that not only the lipophilicity of the complex determines the brain uptake. Several other parameters, such as charge, complex size and molecular weight, are important for blood-brain barrier penetration. The biodistribution of the hydrolyzed complexes (7 and 8) was also studied. As shown in Fig. 5, these complexes also present a hepatobiliar behavior. However, the liver uptake is significantly slower, especially for complex 7. They also present a slightly higher kidney uptake and a faster excretion of the radioactivity than 4–6. The higher liver uptake and the slower excretion of complex 8, relative to complex 7, may be explained by the long alkyl chain in the ligand backbone. For the silylated complexes the in vivo stability in mice
Fig. 3. Hydrolysis of complex 6: (a) pH 6.0 and 7.4, 37°C and (b) HPLC profile of 6 (RT⫽12.2 min) after 1 hour at pH 6.0 and at 37°C.
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
791
was also evaluated. The blood collected at 5 min and 1 hour post-injection was centrifuged, after precipitation of the proteins with ethanol, and the supernatant analyzed by HPLC. For all the compounds the radioactive HPLC chromatograms show mainly the intact complex, being present only a small amount of a more hydrophilic metabolite (Fig. 6). Complex 6 was the only one that has shown a small amount of the hydrolyzed form (8).
4. Concluding remarks
Fig. 4. Hydrolysis of complexes 4 and 5 at 37°C and different pH values.
We were able to prepare neutral and lipophilic silylated “3⫹1” mixed ligand 99mTc complexes stabilized by the [PNS/S] donor atom set. Using the optimized labeling conditions the complexes were obtained with relatively high yields and with high radiochemical purity (⬎95%). All the silylated complexes are stable in physiological conditions and in the presence of a large excess of glutathione. The in vitro stability studies performed in rat plasma and human blood show the intact silylated compounds in the supernatant and a low protein binding, indicating that no transchelatation or metabolic process occurs. In vitro the hydrolysis of 4–6, although being strongly dependent on the nature of the substituent at the silicium atom, requires pH values significantly lower than the physiological pH. The silylated complexes are stable in vivo and no significant hydrolysis was observed, as indicated by blood analysis. The 99mTc complexes are eliminated essentially
Fig. 5. Biodistribution results in mice (% injected dose/organ) for complexes 4–8.
792
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
through the hepatobiliary tract with low urinary elimination. The silylation increased significantly the lipophilicity of the compounds, but no effect was found in their brain uptake or brain retention. Although not having a satisfactory brain uptake, we have shown that compounds 4–6, anchored by the tridentate PNS ligand, are quite stable in vitro and in vivo, contrasting with all the previously described mixed-ligand “3⫹1” complexes, anchored by other tridentate ligands. Our results clearly show the importance of the tridentate ligand on the stability in vitro and in vivo of the 3⫹1 compounds. The stability promoted by our heterofunctionalized phosphines encourage further studies to improve blood-brain barrier penetration.
Acknowledgments The authors would like to thank the DAAD and ICCTI for a bilateral project. One of the authors (T.K.) thanks the FCT for a PRAXIS postdoctoral grant.
References
Fig. 6. In vivo stability of complexes 4–6, 1 hour spot-injection (hydrophilic metabolite is indicated by an asterisk).
[1] Beckett AH, Taylor DC, Garrod JW. Trialkylsily moieties as potencial pharmacokinetic modifying groups for aminoalchohols. J Pharm Pharmacol 1975;27:588 –93. [2] Millership JS, Shanks ML. Prodrugs utilizing organosilyl derivation: an investigation of the long-term androgenic and myotrophic activities of silyl derivatives of testosterone. J Pharm Sci 1988; 77:116 –9. [3] Przuntek H, Westarp ME, Vohl ML, Gerlach M, Jutzi P, Wekerle H. The effect of silyl substituted methotrexate in the treatment of experimental allergic encephalomyelitis in rats mediated by T-line cells. Neuropharmacology 1987;26:255– 60. [4] Tsareva TA, Kramarova EP, Zharkovskii AM. Pharmacological characteristics of the silyl derivatives of alpha-pyrrolidone. Farmakol Toksikol 1982;45:20 –3. [5] Fietz Th, Spies H, Zablotskaya A, Scheller D. Reactions of hydroxy-group containing ‘3 ⫹ 1’ mixed-ligand rhenium(V) compounds. Part 1: Silylation of ‘3 ⫹ 1’ mixed-ligand complexes. Inst Bioanorg Radiopharm Chemie, FZ-Rossendorf, Annual Report, 1997;113–16. [6] Kniess T, Spies H, Santos I, Zablotskaya A. Synthesis of hydroxyl silylated rhenium and (99mTc)technetium ‘3 ⫹ 1’ mixed ligand complexes. J Label Comp Radiopharm 2002;45:629 –36. [7] Spies H, Fietz Th, Zablotska A, Belyakov S, Lukevics E. Silyl modifications of biologically active compounds. Organosilicon complexes of rhenium(V) with mixed ligands. Chem Het Comp 1999;35: 112–20. [8] Gupta A, Seifert S, Syhre R, Scheunemann P, Brust P, Johannsen B. Reactivity of 99mTc(V) “3 ⫹ 1” mixed-ligand complexes towards glutathione. Radiochim Acta 2001;89:43–9. [9] Syhre R, Seifert S, Spies H, Gupta A, Johannsen B. Stability versus reactivity of “3 ⫹ 1” mixed-ligand technetium-99m complexes in vitro and in vivo. Eur J Nucl Med 1998;25:793– 6. [10] Correia JDG, Domingos A, Santos I, Alberto R, Ortner K. Re tricarbonyl complexes with ligands containing P,N,N and P,N,O donor atom sets: synthesis and structural characterization. Inorg Chem 2001;40:5147–51.
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793 [11] Correia JDG, Domingos A, Santos I, Spies H. Synthesis and characterization of mixed-ligand oxorhenium(V) complexes with new [(PNO/S)/ (S)] donor atom sets. J Chem Soc Dalton Trans 2001;2001:2245–50. [12] Correia JDG, Santos I, Alberto R, Ortner K, Spies H, Drews A. Heterofunctionalized phosphines as anchor groups for coupling biomolecules to the fac-[M(CO)3]⫹ (M⫽Re,Tc) moiety. J Label Comp Radiopharm 2001;44:S507–9. [13] Fernandes C, Correia JDG, Gano L, Santos I, Seifert S, Spies H, Syhre R. Tc oxocomplexes with the PNO/S and PNS/S donor atom sets: labeling of a 5HT1A receptor-binding ligand. J Labeled Comp Radiopharm 2001;44:S518 –20. [14] Kniess T, Fernandes C, Santos I, Spies H, Kraus W. Silylated mixed-
793
ligand rhenium complexes with the [PNS/S] donor atom set. Inorg Chim Acta 2003;348:237– 41. [15] Alberto R, Schibli A, Egli A, August Schubiger P, Herrmann WA, Artus G, Abram U, Kaden TA. Metal carbonyl synthesis XXII. Low pressure carbonylation of [MOCl4]⫺ and [MO4]⫺: the technetium(I) and rhenium(I) complexes [NEt4][MCl3(CO)3]. J Organomet Chem 1995;493:119 –27. [16] Seifert S, Gupta A, Syhre R, Spies H, Johannsen B. Ligand-exchange reaction of labile “3 ⫹1” 99mTc(V) complexes with SH group-containing proteins. Appl Radiat Isot 2001;54:637– 44. [17] Kniess T, Fernandes C, Seifert S, Spies H, Santos I. unpublished results.
www.elsevier.com/locate/nucmedbio
Synthesis and biological evaluation of silylated mixed-ligand complexes with the [PNS/S] donor atom set
99m
Tc
C. Fernandesa, T. Kniessa, L. Ganoa, S. Seifertb, H. Spiesb, I. Santosa,* b
a Departamento de Quı´mica, ITN, Estrada Nacional 10, 2686-953, Sacave´m Codex, Portugal Institute of Bioinorganic & Radiopharmaceutical Chemistry, Forschungszentrum Rossendorf, D-01314 Dresden, Germany
Received 4 September 2003; received in revised form 12 March 2004; accepted 26 March 2004
Abstract New oxotechnetium complexes of general formula [99mTc(O)(PNS)(S(CH2)nOSiR3)] (4–6) were synthesized by direct reduction of [ TcO4]⫺ with stannous chloride, in the presence of the tridentate heterofunctionalized phosphine H2PNS and of the monodentate silylated thiols [HS(CH2)nOSiR3] (n⫽2, R⫽Ph (1); n⫽3, R⫽Ph (2); n⫽3, R⫽Et (3)). The mixed-ligand rhenium and technetium complexes of general formula [M(O)(PNS)(S(CH2)nOH)] (n⫽2: M⫽99mTc, (7), M⫽Re, (7a); n⫽3: M⫽99mTc, (8), M⫽Re, (8a)) were also prepared. All the 99mTc complexes were obtained with high radiochemical purity (⬎95%), after purification by HPLC, and were characterized by comparison of their HPLC profiles with the ones obtained for the corresponding Re compounds. The silylated compounds 4–6 are stable in phosphate saline buffer (PBS) pH 7.4, rat plasma, human serum and whole blood, and do not bind to plasmatic proteins, and also do not challenge with glutathione. The biological behavior of [99mTc(O)(PNS)(S(CH2)nOH)] (7, 8) and [99mTc(O)(PNS)(S(CH2)nOSiR3)] (4–6) was studied. The effect of the pH on the cleavage of the O-Si bond in complexes 4–6 was also evaluated. © 2004 Elsevier Inc. All rights reserved. 99m
Keywords: Mixed-ligand complexes;
99m
Tc; Biological evaluation
1. Introduction The transportation and accumulation into the target organ is an important issue for 99mTc radiopharmaceuticals development. Previous studies have shown that silylation of biologically active molecules facilitates their transportation in the organism [1– 4]. Applying the same principle to rhenium and 99mTc complexes it was expected to improve their bioavailability, especially in terms of passing the blood-brain barrier with consequent increase of brain uptake. The accumulation into the target organ would be enhanced if we were able to find silyl ethers stable in physiological conditions but hydrolysable in the target organ. The first attempts for preparing silylated mixed-ligand Re and Tc complexes were performed with the tridentate 3-thiapentane-1,5-dithiol (HSSSH) [5–7]. However, the instability of these compounds [8,9] led us to explore a new family of tridentate heterofunctionalized phosphines, namely 2-(diphenylphosphanyl)-N-(2-thioethyl)benzamide * Corresponding author. Tel.: ⫹00351219946201; fax: ⫹00351219941455. E-mail address: [email protected] (I.Santos). 0969-8051/04/$ – see front matter © 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.nucmedbio.2004.03.011
(H2PNS), 2-(diphenylphosphanyl)-N-(2-aminoethyl)benzamide (H2PNN) and 2-(diphenylphosphanyl)-N-(2-hydroxyethyl)benzamide (H2PNO), recently introduced by our group [10 –13]. These chelators are very versatile stabilizing tricarbonyl compounds as well as mixed-ligand oxocomplexes, some of them bearing monodentate silylated coligands [10 –14]. Herein, we report on the preparation of 99m Tc mixed-ligand oxocomplexes with monodentate silylated co-ligands, [99mTc(O)(PNS)(S(CH2)nOSiR3)] (4–6). The stability and biological behavior of these compounds will be also reported. Taking into account that in vitro or in vivo hydrolysis of 4–6 can promote the breaking of the O-Si bond, we have also synthesized and characterized the complexes [M(O)(PNS)(S(CH2)nOH)] (M⫽Re, 99mTc; n⫽2, 3) (7,8), which are also described herein. 2. Experimental 2.1. Materials and methods Unless otherwise stated, all chemicals were of reagent grade and were used without further purification. 2-mer-
786
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
capto-ethanol, 3-mercapto-1-propanol, chloro triethylsilane, chloro triphenylsilane, imidazole and stannous chloride were obtained from Aldrich. (99mTc)TcO4⫺ in saline solution was eluted from a 99Mo/99mTc generator from MDS Nordion S.A. (Belgium). Elemental analysis was performed on a Perkin Elmer automatic analyzer. The 1H and 31P NMR spectra were recorded on a VARIAN Unity 300MHz spectrometer; 1H chemical shifts were referenced relative to tetramethylsilane and the 31P chemical shifts to external 85% H3PO4 solution. The NMR samples were prepared in CDCl3, chemical shifts are given in parts per million. Chemical reactions were monitored by thin-layer chromatography (TLC) on Merck plates precoated with silica gel 60 F254. Spots were visualized either by UV light or Ellman’s reagent for thiols. Column chromatography was performed in silica gel 60 (Merck). HPLC investigations were carried out using an RP C18 column (Hypersil ODS, 4.0⫻250 mm, 10 m) and using as eluent a gradient of acetonitrile/water with a flow rate of 1.0 ml/min. The identification of the compounds was made using a UV-VIS detector (LC 290, Perkin Elmer) or a ␥-detector (LB 509, Berthold). [nBu4N][Re(O)Cl4] was prepared according to literature methods [15]. The hydroxyl silylated monodentate thiols (1–3) and the 2-(diphenylphosphanyl)-N-(2-thioethyl)benzamide (H2PNS) were synthesized as previously reported [6,11]. 2.2. General procedure for the preparation of complexes 7a and 8a [nBu4N][Re(O)Cl4] (174 mol) was dissolved in 8 ml of dichloromethane and stirred at room temperature (r.t.) under a nitrogen atmosphere. A solution of 174 mol of H2PNS, 250 mol of 3-mercapto-1-propanol and 100 l triethylamine (720 mol) in 4 ml dichloromethane was added drop wise and the mixture stirred for 6 hours. The solvent was evaporated and the complexes purified by column chromatography (silicagel, dichloromethane/ethyl acetate (90:10)). 2.2.1. [Re(O)(3-PNS)(1-S-CH2-CH2-OH)] (7a) Yield 15 %. 1H NMR (␦, ppm, CDCl3): 2.02 (m, 2H, CH 2), 2.22 (m, 1H, CHtrid), 2.70 (m, 1H, CHtrid), 2.95 (m, 1H, CHtrid), 3.80 –3.94 (m, 2H, CH2), 5.08 (m, 1H, CHtrid), 6.83 (m, 1H, CHar), 7.20 –7.66 (m, 12H, CHar), 8.12 (m, 1H, CH ar). Elemental analysis: C23H23O3S2NPRe, calc., C 42.98, H 3.61, N 2.18 S 9.98, found, C 40.88, H 4.17, N 2.02, S 8.78. 2.2.2. [Re(O)(3-PNS)(1-S-CH2-CH2-CH2-OH)] (8a) Yield 20 %. 1H NMR (␦, ppm, CDCl3): 2.04 (m, 2H, CH 2), 2.23 (m, 1H, CHtrid), 2.73 (m, 1H, CHtrid), 2.99 (m, 1H, CHtrid), 3.77–3.98 (m, 4H, 2⫻CH2), 5.08 (m, 1H, CHtrid), 6.80 (m, 1H, CHar), 7.26 –7.60 (m, 12H, CHar), 8.12 (m, 1H, CHar). Elemental analysis: C24H25O3S2NPRe, calc., C
43.89, H 3.84, N 2.13, S 9.76, found, C 43.09, H 4.50, N 2.11, S 9.17. 2.3. General procedure for the preparation of complexes: complexes 4–6
99m
Tc
The 99mTc complexes were prepared by direct reduction of pertechnetate with stannous chloride. The following general procedure was used: 50 l of propylene glycol (PG), 1300 l of ethanol, 0.5 mg of the monodentate ligand (1–3), 0.03 mg (8.2⫻10⫺8 mol) of H2PNS ligand as well as 10 l of a SnCl2 solution (2.0 –3.0 mg SnCl2 dissolved in 5 ml ethanol) were added to 0.5 ml (100 –200 MBq) of [99mTc]pertechnetate eluate. The reaction mixture was heated at 40 –50°C for 5–15 min, depending on the monodentate ligand. The 99mTc complexes were analyzed by HPLC using a reverse-phase Hypersil 120 ODS column (250⫻4 mm, 10 m). The column was eluted using acetonitrile/water as mobile phase at 1.0 ml/min flow rate and with a linear gradient of 70 –100% acetonitrile over 5 min. followed by 10 min at 100% of acetonitrile. These complexes were also analyzed by TLC using two different chromatographic systems: (a) silicagel/acetone and (b) silicagel/saline. The compounds 4–6 were purified on a semi-preparative Hypersil ODS column (250⫻8 mm, 10 m) using acetonitrile/water as mobile phase (flow rate: 2 ml/min) and with a linear gradient of 70 –100% acetonitrile over 5 min followed by 20 min at 100% of acetonitrile. The fractions were collected and analyzed in the analytical column. The purified complexes, after removal of the organic solvent, were reconstituted with 0.1 M PBS 7.4 and 100 l of PG was added. The characterization of the 99mTc complexes was accomplished by chromatographic correlation (HPLC) with the corresponding rhenium complexes [14]. 2.4. Complexes 7 and 8 The complexes 7 and 8 were synthesized using two different procedures: (a) Incubation of complexes 4 and 6 with a diluted HCl solution, at 50°C. (b) Heating of complexes 4 and 6 at 80°C during 20 min, immediately after labeling. The hydrolyzed complexes (7, 8) were purified by semi-preparative HPLC, using the experimental conditions above described. 2.5. General procedures for stability studies 2.5.1. Challenge with GSH 100 l (approximately 10 MBq) of the purified 99mTc complex was mixed with 100 l of a 2 or 20 mM solution of GSH in 0.1 M phosphate buffer (pH 7.4). The mixture was incubated at 37°C for 5–120 min and analyzed by HPLC. Samples without GSH were also analyzed and used for comparison. HPLC analyzes were carried out using the experimental conditions previously described. Compounds:
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
4, 87%, 2 hours, 10 mM; 5, 92%, 2 hours, 10mM; 6, 88%, 2 hours, 10mM. 2.5.2. Stability in rat plasma Blood collected from mice in heparinized polypropylene tubes were immediately centrifuged for 15 min at 2,000 rpm at 4°C and the plasma collected. To 1.0 ml of fresh rat plasma 100 l (approximately 10 MBq) of the 99mTc complex, were added and the mixture was incubated at 37°C. At appropriate periods of time (5, 30 min, 1, 2, and 4 hours), 100 l aliquots (in duplicate) were removed and treated with 200 l of ethanol to precipitate the proteins. Samples were then cooled at 4°C and centrifuged at 3,000 rpm for 15 min at 4°C. The supernatant was separated from the precipitate and the sediment washed twice with 1 ml of ethanol and counted in a gamma counter. The activity in the sediment was compared with the total activity used and the percent of complex bound to proteins calculated. The supernatant was analyzed by HPLC, using the experimental conditions above described, to test the stability of the complex in rat plasma. Compounds: 4, 95%, 4 hours; 5, 98%, 4 hours; 6, 90%, 2 hours. 2.5.3. Stability in whole human blood To 1 ml of whole human blood, collected in heparinized polypropylene tubes was added a solution of the 99mTc complex (⬇10 MBq) and the mixture incubated at 37°C. Samples were taken, 5 min, 1 hour, and 4 hours after incubation, and centrifuged 15 min at 2,000 rpm at 4°C. The plasma was separated and ethanol was added in a 2:1 (v/v) ratio. The samples were centrifuged at 3,000 rpm and the supernatant analyzed by HPLC. Compounds: 4, 93%, 4 hours; 5, 90%, 4 hours; 6, 93%, 4 hours. 2.6. Biodistribution studies The in vivo behavior of the 99mTc complexes was evaluated in groups of 3–5 female CD-1 mice (randomly bred, Charles River) weighting approximately 20 –25 g each. Animals were intravenously injected with 100 l (1.5– 4.5 MBq) of each preparation via the tail vein and were maintained on normal diet ad libitum. At 5 min and 1 hour mice were killed by cervical dislocation. The radioactive dosage administered and the radioactivity in the sacrificed animal was measured in a dose calibrator (Aloka, Curiemeter IGC-3, Tokyo, Japan). The difference between the radioactivity in the injected and sacrificed animal was assumed to be due to excretion, mainly urinary excretion. Blood samples were taken by cardiac puncture at sacrifice. Tissue samples of the main organs were then removed and counted in a gamma counter (Berthold). Biodistribution results were expressed as percent of injected dose per organ (% I.D./total organ). For blood, bone and muscle, total activity was calculated assuming that these organs constitute 6, 10, and 40% of the total weight, respectively. The remaining activity in the carcass was also measured in a dose calibrator.
787
Afterwards blood was centrifuged and the serum separated for RP-HPLC analysis.
3. Results and discussion The 99mTc complexes [99mTc(O)(PNS)(S(CH2)nOSiR3)] (4–6) were obtained by direct reduction of [99mTcO4]⫺ with stannous chloride in the presence of the H2PNS and of the monodentate [HS(CH2)nOSiR3] ligands (n⫽2, R⫽Ph (1); n⫽3, R⫽Ph (2); n⫽3, R⫽Et (3)) (Scheme 1). The characterization of these complexes has been made by comparing their HPLC profiles with the profiles of the analogous Re compounds, recently reported by our group [14]. As indicated in Scheme 1, the preparation of complexes 4–6 was achieved using stannous chloride dissolved in ethanol, freshly prepared, and by heating the mixture for 5–15 min, at 40 –50°C. This methodology is quite different from the ones described for preparing hydroxyl silylated mixedligand complexes with the [SSS/S] donor atom set or for preparing 3⫹1 99mTc complexes with the [PNS/S] donor atom set [6,13]. In fact, the mixed-ligand complexes with the [SSS/S] donor atom set were obtained by gluconate transchelatation and the compounds anchored in the [PNS/ S] donor atom set were obtained by reducing the [99mTcO ⫺ 4] with stannous chloride in aqueous acidic media. When we tried to prepare complexes 4–6 using the first methodology no labeling was observed, and using the second procedure only hydrolyzed species could be isolated. Another important point for preparing complexes 4–6, with relatively high yield, is the use of a large excess of the monodentate ligand, relatively to the tridentate H2PNS. These experimental conditions are important as they avoid the formation of complexes 4–6 contaminated with the complex [99mTc(O)(3-PNS)(1-HPNS)]. This radiochemical impurity has been characterized by comparing its HPLC profile with the profile of the analogous Re complex, previously described [11]. Complexes 4–6 were obtained in 60 –90 % yield, as indicated by HPLC and by TLC. Using TLC (silicagel/ saline) only minor amounts of reduced hydrolyzed technetium could be detected (2–5%). Considering that one of the radiochemical impurities in the preparation of complexes 4–6 could be the hydrolyzed complexes [M(O)(PNS)(S(CH2)nOH] (M⫽Re, 99mTc) (7, n⫽2; 8a, n⫽3) we also prepared and characterized these compounds with Re and 99mTc. The rhenium complexes with mercaptoethanol and mercaptopropanol [Re(O)(PNS)(S(CH2)nOH] (7a, n⫽2; 8n⫽3) were obtained by reacting [nBu4N][Re(O)Cl4] with H2PNS and mercaptoethanol or mercaptopropanol, in the presence of a slight excess of triethylamine (Scheme 1). The synthesis was carried out under nitrogen atmosphere, to avoid the oxidation of the thiol or of the phosphine ligands. Compounds 7a and 8a, after purification by column chromatography and dried under vacuum, were obtained as grey green
788
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
Scheme 1. Synthesis of the complexes.
solids in 15–20 % yield. The low yield of these reactions is due to the formation of some competing complexes. Only one of these species was unambiguously identified as being [Re(O)(3-PNS)-(1-HPNS)] [11]. The other complexes in the mixture must be homoleptic and anchored by the alcohols, as no PNS ligand could be identified by 1H or 31P NMR spectroscopy. The 31P NMR spectra of compounds 7a and 8a show only one singlet at 19.7 and at 19.8 ppm, respectively. These values compare well with the values previously reported for other 3⫹1 oxocomplexes with the [PNS/S] donor atom set [11,14]. In the 1H NMR spectra of complexes 7a and 8a appear four sets of multiplets, integrating for one proton each, due to the methylenic protons of the PNS ligand, which are diastereotopic. The resonances due to the aromatic rings of the PNS and due to the methylenic protons of the co-ligand appear in the expected range [11,14]. The synthesis of the analogous 99mTc complexes 7 and 8 was not possible using the methodology described for 4–6 (Scheme 1). In fact, using this procedure only very hydrophilic species were formed, which were not characterized at the macroscopic level. The quantitative synthesis of 7 and 8 was achieved by hydrolysis of the purified silylated complexes 4 and 6. This hydrolysis was performed by adding diluted HCl to 4 and 6 and by heating at 50°C during 20 min. Another procedure consisted on heating 4 or 6 at 80°C during 20 min., in the labeling conditions, followed by HPLC purification. All the 99mTc complexes described in this work (4–8) were obtained with high radiochemical purity (⬎95%), after HPLC purification. Their characterization was made by
HPLC, comparing their retention time (␥ detection) with the retention time of the well-defined analogous rhenium complexes (UV detection at 254 nm) (Fig. 1 and Table 1). 3.1. Stability The 99mTc complexes, after evaporation of the solvents, were reconstituted with 0.1 M PBS pH 7.4 and their stability studied. We found that complexes 4–8 are stable in 0.1 M PBS 7.4, for at least 4 hours at 37°C, which is an important parameter to pursue with biological studies. As previously reported, a drawback of the 3⫹1 Tc complexes of the type [TcO(SES)(SR)] (E⫽S, O, NR⬘) is their instability when in the presence of thiolated nucleophiles, like glutathione [8,9,16]. We also found that the silylated 99m Tc complexes stabilized by the SSS tridentate ligand, although stable in PBS 7.4 at 37°C, display a high rate of exchange with glutathione (1 mM solutions, 5–120 min at 37°C) (Fig. 2) [17]. Taking this into account, we studied the stability of the silylated 99mTc complexes 4–6 in challenge reactions with glutathione, in rat plasma, in human serum and in whole blood, before proceeding with in vivo biological studies. We found that 4–6 do not challenge with glutathione, even in the presence of concentrated solutions (1 and 10 mM GSH dissolved in 0.1M phosphate buffer pH 7.4, 5–120 min at 37°C) and they also remain intact in fresh rat plasma (37°C, 5 min– 4 hours), and do not bind significantly to plasmatic proteins (Table 1). Even in the presence of whole human blood, with remarkable GSH concentration, all the complexes remained intact in the supernatant and the radioac-
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
789
Fig. 1. HPLC chromatograms for complexes 5 (␥ detection, RT⫽11.6 min) and 5a (UV detection at 254 nm, RT⫽11.2 min).
tivity bound to the proteins is very low (2–13%). These results indicate a high stability for compounds 4–6 showing that no transchelatation or metabolic process did take place (Table 1). This is an interesting result which confirms the importance of the nature of the tridentate ligand on the stability of the mixed-ligand “3⫹1” 99mTc complexes [8,9]. We also investigated the pH influence on the cleavage of the O-Si bond in complexes [99mTc(O)(PNS)(S(CH2)nOSiR3)] (4–6). Fig. 3 shows the hydrolysis curves for complex 6, at pH 6 and 7.4, as well as the HPLC profile of one sample which has been incubated over 1 hour, at pH 6 and at 37°C. As can be seen, in physiological conditions (pH⫽7.4, 37°C) 6 is relatively stable but rapid cleavage of the O-Si bond was observed at pH 6.0, 37°C. However, complexes 4 and 5, with a triphenyl substituent, are stable in the same conditions, requiring a much low pH to be hydrolyzed (Fig. 4). 3.2. Lipophilicity The lipophilicity of all the compounds was evaluated by determination of the partition coefficient (P) in physiological conditions (n-octanol/0.1 M phosphate buffer pH 7.4). The results obtained, expressed as log Po/w, are presented in Table 1. As expected, there is an increase of the lipophilicity when the -OH group is blocked by SiR3 groups.
3.3. Biodistribution The in vivo behavior of the 99mTc complexes (4–8) was evaluated in mice at 5 and 60 min post intravenous injection. Fig. 5 shows the results expressed as % I.D./total organ in the most relevant organs. All the complexes have high initial blood, muscle and liver uptake as expected for lipophilic compounds. Nevertheless, the blood and the muscle clearance is quite fast for all the complexes. Excretion occurs mainly through the hepatobiliary tract with quite low urinary elimination. The liver uptake is higher and faster for the complexes blocked with the triphenyl substituent at the silicium atom (4 and 5) than for the complex with the triethyl substituent (6). In the case of 4 and 5 the activity increases over the period studied, whereas it decreases for complex 6. The introduction of the triethyl substituent in the ligand backbone induces also a significantly faster hepatobiliary excretion and a slightly higher kidney uptake. It was expected that the introduction of the silyl groups in 99mTc complexes would increase their lipophilicity and consequently their bioavailability, brain uptake and brain retention promoted by hydrolyze in the target organ. In fact, the lipophilicity of the 99mTc complexes was significantly increased by silylation (Table 1) but no effect was found in the brain uptake (⬍0.09 % I.D.) or brain retention. We found that, despite their high lipophilicity, compounds 4–6 present a lower brain uptake than the model
Table 1 Some properties of compounds 4 – 8 (nd-not determined) Compound
4/4a 5/5a 6/6a 7/7a 8/8a
Labelling yield (%)
70–90 60–90 60–70 85–90 85–90
RT (min)
11.4/11.5 11.1/11.6 11.9/12.1 3.6/3.8 3.8/4.0
log Po/w
2.26 2.40 2.32 1.70 1.81
Protein binding (%) 5 min.
1h
8.7 2.4 2.8 nd nd
12.1 12.5 8.1 nd nd
790
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
Fig. 2. Stability of complex [99mTc(O)(SSS/S(CH2)3OSiPh3] in 0.1 M PBS pH 7.4 and in presence of 1 mM GSH solution at 37°C.
complex [99mTcO(PNS/SPh)] (log P⫽1.7, I.D./organ ⫽0.14% at 5 min p.a.) [13]. This difference can probably be explained by the high molecular weight of the silylated
complexes 4–6. The results obtained clearly indicated that not only the lipophilicity of the complex determines the brain uptake. Several other parameters, such as charge, complex size and molecular weight, are important for blood-brain barrier penetration. The biodistribution of the hydrolyzed complexes (7 and 8) was also studied. As shown in Fig. 5, these complexes also present a hepatobiliar behavior. However, the liver uptake is significantly slower, especially for complex 7. They also present a slightly higher kidney uptake and a faster excretion of the radioactivity than 4–6. The higher liver uptake and the slower excretion of complex 8, relative to complex 7, may be explained by the long alkyl chain in the ligand backbone. For the silylated complexes the in vivo stability in mice
Fig. 3. Hydrolysis of complex 6: (a) pH 6.0 and 7.4, 37°C and (b) HPLC profile of 6 (RT⫽12.2 min) after 1 hour at pH 6.0 and at 37°C.
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
791
was also evaluated. The blood collected at 5 min and 1 hour post-injection was centrifuged, after precipitation of the proteins with ethanol, and the supernatant analyzed by HPLC. For all the compounds the radioactive HPLC chromatograms show mainly the intact complex, being present only a small amount of a more hydrophilic metabolite (Fig. 6). Complex 6 was the only one that has shown a small amount of the hydrolyzed form (8).
4. Concluding remarks
Fig. 4. Hydrolysis of complexes 4 and 5 at 37°C and different pH values.
We were able to prepare neutral and lipophilic silylated “3⫹1” mixed ligand 99mTc complexes stabilized by the [PNS/S] donor atom set. Using the optimized labeling conditions the complexes were obtained with relatively high yields and with high radiochemical purity (⬎95%). All the silylated complexes are stable in physiological conditions and in the presence of a large excess of glutathione. The in vitro stability studies performed in rat plasma and human blood show the intact silylated compounds in the supernatant and a low protein binding, indicating that no transchelatation or metabolic process occurs. In vitro the hydrolysis of 4–6, although being strongly dependent on the nature of the substituent at the silicium atom, requires pH values significantly lower than the physiological pH. The silylated complexes are stable in vivo and no significant hydrolysis was observed, as indicated by blood analysis. The 99mTc complexes are eliminated essentially
Fig. 5. Biodistribution results in mice (% injected dose/organ) for complexes 4–8.
792
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793
through the hepatobiliary tract with low urinary elimination. The silylation increased significantly the lipophilicity of the compounds, but no effect was found in their brain uptake or brain retention. Although not having a satisfactory brain uptake, we have shown that compounds 4–6, anchored by the tridentate PNS ligand, are quite stable in vitro and in vivo, contrasting with all the previously described mixed-ligand “3⫹1” complexes, anchored by other tridentate ligands. Our results clearly show the importance of the tridentate ligand on the stability in vitro and in vivo of the 3⫹1 compounds. The stability promoted by our heterofunctionalized phosphines encourage further studies to improve blood-brain barrier penetration.
Acknowledgments The authors would like to thank the DAAD and ICCTI for a bilateral project. One of the authors (T.K.) thanks the FCT for a PRAXIS postdoctoral grant.
References
Fig. 6. In vivo stability of complexes 4–6, 1 hour spot-injection (hydrophilic metabolite is indicated by an asterisk).
[1] Beckett AH, Taylor DC, Garrod JW. Trialkylsily moieties as potencial pharmacokinetic modifying groups for aminoalchohols. J Pharm Pharmacol 1975;27:588 –93. [2] Millership JS, Shanks ML. Prodrugs utilizing organosilyl derivation: an investigation of the long-term androgenic and myotrophic activities of silyl derivatives of testosterone. J Pharm Sci 1988; 77:116 –9. [3] Przuntek H, Westarp ME, Vohl ML, Gerlach M, Jutzi P, Wekerle H. The effect of silyl substituted methotrexate in the treatment of experimental allergic encephalomyelitis in rats mediated by T-line cells. Neuropharmacology 1987;26:255– 60. [4] Tsareva TA, Kramarova EP, Zharkovskii AM. Pharmacological characteristics of the silyl derivatives of alpha-pyrrolidone. Farmakol Toksikol 1982;45:20 –3. [5] Fietz Th, Spies H, Zablotskaya A, Scheller D. Reactions of hydroxy-group containing ‘3 ⫹ 1’ mixed-ligand rhenium(V) compounds. Part 1: Silylation of ‘3 ⫹ 1’ mixed-ligand complexes. Inst Bioanorg Radiopharm Chemie, FZ-Rossendorf, Annual Report, 1997;113–16. [6] Kniess T, Spies H, Santos I, Zablotskaya A. Synthesis of hydroxyl silylated rhenium and (99mTc)technetium ‘3 ⫹ 1’ mixed ligand complexes. J Label Comp Radiopharm 2002;45:629 –36. [7] Spies H, Fietz Th, Zablotska A, Belyakov S, Lukevics E. Silyl modifications of biologically active compounds. Organosilicon complexes of rhenium(V) with mixed ligands. Chem Het Comp 1999;35: 112–20. [8] Gupta A, Seifert S, Syhre R, Scheunemann P, Brust P, Johannsen B. Reactivity of 99mTc(V) “3 ⫹ 1” mixed-ligand complexes towards glutathione. Radiochim Acta 2001;89:43–9. [9] Syhre R, Seifert S, Spies H, Gupta A, Johannsen B. Stability versus reactivity of “3 ⫹ 1” mixed-ligand technetium-99m complexes in vitro and in vivo. Eur J Nucl Med 1998;25:793– 6. [10] Correia JDG, Domingos A, Santos I, Alberto R, Ortner K. Re tricarbonyl complexes with ligands containing P,N,N and P,N,O donor atom sets: synthesis and structural characterization. Inorg Chem 2001;40:5147–51.
C. Fernandes et al. / Nuclear Medicine and Biology 31 (2004) 785–793 [11] Correia JDG, Domingos A, Santos I, Spies H. Synthesis and characterization of mixed-ligand oxorhenium(V) complexes with new [(PNO/S)/ (S)] donor atom sets. J Chem Soc Dalton Trans 2001;2001:2245–50. [12] Correia JDG, Santos I, Alberto R, Ortner K, Spies H, Drews A. Heterofunctionalized phosphines as anchor groups for coupling biomolecules to the fac-[M(CO)3]⫹ (M⫽Re,Tc) moiety. J Label Comp Radiopharm 2001;44:S507–9. [13] Fernandes C, Correia JDG, Gano L, Santos I, Seifert S, Spies H, Syhre R. Tc oxocomplexes with the PNO/S and PNS/S donor atom sets: labeling of a 5HT1A receptor-binding ligand. J Labeled Comp Radiopharm 2001;44:S518 –20. [14] Kniess T, Fernandes C, Santos I, Spies H, Kraus W. Silylated mixed-
793
ligand rhenium complexes with the [PNS/S] donor atom set. Inorg Chim Acta 2003;348:237– 41. [15] Alberto R, Schibli A, Egli A, August Schubiger P, Herrmann WA, Artus G, Abram U, Kaden TA. Metal carbonyl synthesis XXII. Low pressure carbonylation of [MOCl4]⫺ and [MO4]⫺: the technetium(I) and rhenium(I) complexes [NEt4][MCl3(CO)3]. J Organomet Chem 1995;493:119 –27. [16] Seifert S, Gupta A, Syhre R, Spies H, Johannsen B. Ligand-exchange reaction of labile “3 ⫹1” 99mTc(V) complexes with SH group-containing proteins. Appl Radiat Isot 2001;54:637– 44. [17] Kniess T, Fernandes C, Seifert S, Spies H, Santos I. unpublished results.