A Third Generation Potentially Bifunctional Trithiol Chelate, Its ^nat,1XXSb(III) Complex, and Selective Chelation of Radioantimony (¹¹⁹Sb) from Its Sn Target

Abstract

The therapeutic potential of the Meitner-Auger- and conversion-electron emitting radionuclide ¹¹⁹Sb remains unexplored because of the difficulty of incorporating it into biologically targeted compounds. To address this challenge, we report the development of ¹¹⁹Sb production from electroplated tin cyclotron targets and its complexation by a novel trithiol chelate. The chelation reaction occurs in harsh solvent conditions even in the presence of large quantities of tin, which are necessary for production on small, low energy (16 MeV) cyclotrons. The ¹¹⁹Sb-trithiol complex has high stability and can be purified by HPLC. The third generation trithiol chelate and the analogous stable ^natSb-trithiol compound were synthesized and characterized, including by single-crystal X-ray diffraction analyses.

Graphical Abstract

INTRODUCTION

Targeted radionuclide therapy (TRT) has gained momentum as an attractive cancer treatment method, as seen in recent clinical trials.^1,2 Among many TRT radionuclide choices, Meitner-Auger Electron (MAE) emitting radionuclides have unique properties to be advantageously leveraged in cancer TRT.^3–7 Antimony-119 (¹¹⁹Sb, t_1/2 = 38.19 h, EC = 100% (MAE))⁸ is considered one of the most promising MAE emitting radionuclides for TRT applications^9,10 and has potential application in theranostic nuclear medicine with its imaging radioisotope congener antimony-117 (¹¹⁷Sb, t_1/2 = 2.8 h, γ = 85.9%, 158.56 keV, EC = 97.3%, β⁺ = 1.8%).¹¹

Antimony-119 is a MAE emitting radionuclide that emits 23–24 conversion electrons and MAEs with energies up to 28 keV and no dosimetrically problematic gammas.¹² Electrons with these energies have high linear energy transfer (LET) and consequently high relative biological effectiveness (RBE) in a short range (nm-μm), capable of delivering lethal radiation doses in subcellular ranges while having little off-target toxicity.¹³ The densely ionizing pathlengths up to 15 μm length in tissue¹² reach a cell nucleus from the cell surface.^9,14–17 Its lighter radioisotope ¹¹⁷Sb emits a 158.6 keV gamma ray that is ideal for single photon emission computed tomography (SPECT).¹⁴ Together, ¹¹⁹Sb and ¹¹⁷Sb form a theranostic (therapy and diagnostic) radionuclide pair, facilitating precise measurement of patient specific biodistribution and calculations of in vivo dosimetry. ¹¹⁹Sb can be produced indirectly via a tellurium generator.^18–22 However, direct, no-carrier-added (n.c.a.) production routes of irradiating tin with protons or deuterons avoid medically incompatible eluent systems and dosimetrically problematic tellurium radioisotopes, making the generators difficult to shield.^23–25

Antimony complexes are used for treatment of parasitic infections Schistosomiasis and Leishmaniasis. These complexes deliver pentavalent antimony to the parasites, and it is believed that reduction to Sb(III) interferes with the parasites’ thiol redox metabolism.²⁶ Clinical treatments for antimony and fellow pnictogen arsenic poisoning employ two chelators: meso-2,3-dimercaptosuccinic acid (DMSA) and 2,3-dimercaptopropane-1-sulfonic acid (DMPS).²⁷ Antimony is considered a large,^28,29 soft atom,^30,31 and similar to arsenic, antimony is thiophilic and forms complexes with thiol donor atoms.^32–35 The greatest barrier to medical application of ¹¹⁹Sb is the lack of stable complexation with a bifunctional chelator that has the ability to chelate Sb and also covalently bind to a targeting moiety. The only reported chelation of radioantimony was a ¹¹⁷Sb Potassium Antimonyl Tartrate (PAT) complex published in 1969.³⁶ No complex stability measurements are reported, and the PAT complex is incapable of functionalization to molecular targeting agents.

This work explores stable, selective chelation of radioantimony by a potentially bifunctional trithiol chelator originally developed for the complexation of arsenic.^37–40 The trithiol chelator design includes three thiol functional groups to complex radioantimony and two carboxylic acid arms for potential linker incorporation. The selectivity of the novel trithiol for antimony circumvented usual and cumbersome radiochemical isolation of n.c.a. radioantimony from the tin accelerator target material. We describe a novel, economical production route to small quantities of ¹¹⁹Sb, the synthesis and characterization of a novel trithiol chelate, its chemistry with ^natSb(III), and its radiochemistry with ^1XXSb. Tin target recycling, facilitating isotopically enriched ¹¹⁹Sn targetry and production of radioisotopically pure ¹¹⁹Sb, will be reported in a future article.

EXPERIMENTAL SECTION

Materials and General Methods.

All solutions were prepared with 18 MΩ·cm deionized water and optima grade H₂SO₄, HCl, ethanol, and acetonitrile (MeCN) from Fisher Chemical (Hampton, NH). Dimethyl 5-hydroxyisophthalate, pentaerythritol tetrabromide, tris(2-carboxyethyl)phosphine hydrochloride (TCEP), silica gel 60 Å, potassium thiocyanate, antimony trichloride, and potassium carbonate were purchased from Fisher Scientific (Waltham, MA) or Sigma-Aldrich (St. Louis, MO). Silica gel w/UV 254 TLC plates were purchased from Sorbtech Technologies (Norcross, GA). All solvents and reagent grade acids and bases were purchased from Fisher Scientific or Sigma-Aldrich and used without further purification. Unless otherwise noted, only 18 MΩ·cm water was used.

Microwave reactions were performed using a CEM Discover SP microwave reactor (CEM Corp., Matthews, NC). The ¹H and ¹³C NMR spectra were obtained in CDCl₃ or d₆-DMSO on a Bruker ARX-500 or 600 MHz spectrometer and calibrated with the respective residual solvent. Infrared (IR) spectra were obtained on a Thermo Nicolet Nexus 670 Fourier transform infrared (FT-IR) spectrophotometer. Elemental analyses were performed by Atlantic Microlabs, Inc. (Norcross, GA). High-resolution mass spectral (HRMS) analyses were performed at the University of Missouri Charles W Gehrke Proteomics Center; briefly, the sample was loaded by an EASY-nLC system with methanol solvent and analyzed by nanoelectrospray ionization in positive-ion/negative-ion mode on a ThermoScientific LTQ Orbitrap XL mass spectrometer. The methanol solvent flow rate was set to 600 nL/min. HRMS data were acquired for 5 min per sample (30,000 resolving power, 120–1000 m/z, 1 microscan, maximum inject time of 500 ms, automatic gain control = 5 × 10⁵).

Caution! Antimony and Sn radioisotopes are produced by deuteron irradiation of Sn targets as listed in Table 1. All are gamma emitters. Proper radiation safety procedures and shielding were used when handling these radionuclides in laboratories approved for radioisotope use.

Table 1.

Physical Decay Characteristics of Radionuclides Produced from Deuteron Bombardment of Natural Tin

radionuclide	half-life (h)	emission (keV) (Iγ)
117Sb	2.80	158.56 (0.859)
118mSb	5.00	253.66 (0.996)
119Sb	38.19	25.271 (0.389)
120mSb	138.24	1171.7 (1.0)
122Sb	65.37	564.24 (0.707)
124Sb	1444.8	602.72 (0.978)
125Sb	24181.5	427.87 (0.296)
117mSn	336	158.56 (86.5)

Target Production.

Metallic tin was electroplated as previously reported^41,42 from a stannous sulfate based electrolytic solution⁴³ onto 2 mm thick silver disks (∅24 mm). Briefly, stannous sulfate (90 mg; Strem Chemicals, Newburyport, MA) was dissolved in 100 μL of concentrated H₂SO₄. After addition of phenol sulfonic acid (45 μL; Sigma-Aldrich), gelatin (2 mg; Dot scientific, Burton, MI), and 2-naftol (1 mg; Chem Cruz, Dallas, TX), the solution was gently heated (50 °C) and vortexed. Additional DI water was added to a total volume of 1 mL before adding it into the electrodeposition chamber. Voltage control with a constant 3.0 V applied between the platinum wire anode and silver disk cathode resulted in approximately 50 mg of tin deposited onto an 8 mm diameter exposed surface of the Ag disk over a time period of 48 h.

Target Irradiation.

Using a PETtrace (General Electric, Sweden) cyclotron, antimony radioisotopes (Table 1) were produced by irradiating electroplated tin targets with deuterons at an energy of 8 MeV and beam intensities between 20 and 40 μA, inducing ^natSn-(d,n)^1XXSb and few ^natSn(d,2n)^1XXSb nuclear reactions. An ARTMS Quantm Irradiation System (Richmond, BC) automated retrieval of the irradiated target from the cyclotron, reducing operators’ radiation exposure. For yield measurements and radiolabeling experiments, targets received 10 and 30 μA·h of integrated beam fluence over 0.5 and 1 h, respectively. The radionuclides produced and their decay characteristics are listed in Table 1.⁴⁴

Target Yield Measurements.

Two hours after the end of bombardment (EOB), radionuclide activities were quantified via High Purity Germanium (HPGe) gamma spectrometry using an aluminium-windowed detector (Ortec, Knoxville, Tennessee) coupled to a Canberra (Concord, Ontario) Model 2025 research amplifier and multichannel analyzer calibrated for energy and efficiency using ²⁴¹Am, ¹³³Ba, ¹⁵²Eu, ¹³⁷Cs, and ⁶⁰Co sources (Amersham PLC, Little Chalfont, U.K.). The system’s full width at half-maximum resolution at 1333 keV is 1.8 keV. A second HPGe spectrum collection after all ¹¹⁷Sb decayed (>30 h) allowed quantification of ^117mSn activity and subsequent correction for ^117mSn contribution to 158.56 keV signal. Target yields were measured and decay corrected to EOB. Using a saturation factor dependent upon radionuclide decay properties and length of irradiation, EOB yield can be converted to yield at End of Saturated Bombardment (EOSB).

Synthesis.

Synthesis of Dimethyl 5-(3-Bromo-2,2-bis-(bromomethyl)propoxy)isophthalate [C₁₅H₁₇Br₃O₅], 1.

Synthesis of compound 1 was accomplished following a modified literature procedure.⁴⁵ Pentaerythritol tetrabromide (13.8 g; 35.7 mmol), potassium carbonate (9.83 g; 71.1 mmol), and dimethyl 5hydroxyisophthalate (4.98 g; 23.7 mmol) were dissolved in 100 mL of anhydrous dimethylformamide (DMF) under a nitrogen atmosphere in a 500 mL round-bottom flask. The reaction mixture was stirred and heated in a 70 °C oil bath for 24 h. After cooling to room temperature, DMF was removed under vacuum. Deionized water (500 mL) was added to the reaction, and the mixture was extracted with dichloromethane (DCM; 3 × 300 mL). The organic layers were collected and combined, dried over anhydrous sodium sulfate, filtered, and taken to dryness to afford the crude product. The crude product was purified via silica gel column chromatography using hexanes:DCM (3:1) as the mobile phase (dimethyl 5-hydroxyisophthalate, R_f ≈ 0; pentaerythritol tetrabromide; R_f ≈ 0.5; 1, R_f ≈ 0.15). The product was eluted with DCM, and the fractions were collected, combined, and taken to dryness to afford the pure product as a white solid. Yield: 67%, 8.2 g. ¹H NMR (CDCl₃; 600 MHz) δ ppm: 3.680 (s, 6H, CH₂Br), 3.956 (s, 6H, OCH₃), 4.147 (s, 2H, OCH₂), 7.779 (d, 2H, CH), 8.333 (t, 1H, CH). ¹³C NMR (CDCl₃; 125 MHz) δ ppm: 34.22 (CH₂Br), 43.83 (C(CH₂)4), 52.89 (CH₃), 67.89 (CH₂O), 120.09 (OC=C), 124.05 (O=CC=C), 132.14 (O=CC), 158.24 (OC=C), 166.02 (COO). HRMS (m/z): 514.86899 (514.8704 calc’d for [M + H]⁺ of [C₁₅H₁₇Br₃O₅]). Elemental Anal. Calc’d (found) for C₁₅H₁₇Br₃O₅: C, 34.85 (35.03); H, 3.31 (3.27). (See Supplemental Figures S1–S3.)

Synthesis of 5-(3-Bromo-2,2-bis(bromomethyl)propoxy)-isophthalic Acid [C₁₃H₁₃Br₃O₅], 2.

Synthesis of compound 2 was achieved by hydrolysis of 1. Compound 1 (2.6 g, 5 mmol) and NaOH (2.016 g, 50 mmol) were dissolved in 30 mL of MeCN in a 100 mL round-bottom flask. The reaction mixture was stirred and refluxed at 94 °C for 18 h. Silica gel TLC was used to monitor the reaction progress (ethyl acetate (EtOAc):DCM:glacial acetic acid (20:80:0.5 v/v/v); 1, R_f ≈ 0.95; 2, R_f ≈ 0.4). The reaction was cooled to room temperature, and the MeCN was removed under vacuum. DI water (50 mL) was added to the reaction flask, the mixture was extracted with EtOAc (3 × 50 mL), and the combined organic fractions were back-extracted with DI water (2 × 50 mL). The combined aqueous phase was adjusted to pH 2 with 3 M HCl. A white precipitate formed, and it was collected by vacuum filtration and dried under vacuum. Yield: 1.222 g; 50%. The purity was over 95% by ¹H NMR, and no further purification was performed. ¹H NMR (DMSO; 600 MHz) δ ppm: 3.701 (s, 6H, CH₂Br), 4.113 (s, 2H, OCH₂), 7.753 (d, 2H, CH), 8.114 (t, 1H, CH). ¹³C NMR (CDCl₃; 150 MHz) δ ppm: 34.66 (CH₂Br), 43.36 (C(CH₂)4), 67.38 (CH₂O), 119.36 (OC=C), 122.98 (O=CC=C), 132.80 (O=CC), 158.21 (OC=C), 166.30 (COO). HRMS (m/z): 484.82146 (484.822938 calc’d for [M –H]⁻ of [C₁₃H₁₃Br₃O₅]). Elemental Anal. Calc’d (found) for C₁₃H₁₃Br₃O₅: C, 31.93 (32.80); H, 2.68 (2.76). (See SupplementalFigures S4–S6.)

Synthesis of 5-(3-Thiocyanate-2,2-bis(thiocyanatomethyl)-propoxy)isophthalic Acid, [C₁₆H₁₃N₃O₅S₃], 3.

Synthesis of compound 3 was performed by thiocyanation of 2. Compound 2 (350 mg, 0.83 mmol) and KSCN (725.2 mg, 8.3 mmol) were suspended in 4 mL of DMF in a 10 mL microwave reaction vessel outfitted with a stir bar. The reaction vessel was placed into the microwave at a fixed power of 1 kW, temperature of 120 °C, and reaction time of 20 min. The reaction mixture was cooled to room temperature, poured into ice water, and then placed in the freezer overnight. The solid was collected by vacuum filtration, dissolved in EtOAc, and then dried over anhydrous sodium sulfate. The solvent was removed under vacuum to yield the product as a white solid. Yield: 140 mg, 40%. X-ray quality crystals were grown by dissolving 3 in a 70/30 (v/v) MeCN/H₂O mix at 70 °C and cooling to room temperature. ¹H NMR (DMSO; 600 MHz) δ ppm: 3.626 (s, 6H, CH₂SCN), 4.288 (s, 2H, OCH₂), 7.733 (d, 2H, CH), 8.117 (t, 1H, CH), 13.326 (s, COOH). ¹³C NMR (DMSO; 150 MHz) δ ppm: 37.34 (CH₂SCN), 44.49 (C(CH₂)4), 68.76 (CH₂O), 113.20 (CH₂SCN), 119.36 (OC=C), 122.97 (O=CC=C), 132.55 (O=CC), 157.40 (OC=C), 166.30 (COO). HRMS (m/z): 445.98983 (445.99095 calc’d for [M + Na]⁺ of [C₁₆H₁₃N₃O₅S₃]). Elemental Anal. Calc’d (found) for C₁₆H₁₃N₃O₅S₃: C, 45.38 (45.38); H, 3.09 (3.17); N, 9.92 (9.49); S, 22.71 (21.82). (See Supplemental Figures S7–S9.)

Synthesis of 5-(3-Mercapto-2,2-bis(mercaptomethyl)propoxy)-isophthalic Acid [C₁₃H₁₆O₅S₃], 4.

Compound 4 was prepared by reductive deprotection of 3. Compound 3 (80 mg, 0.189 mmol) was dissolved in 7 mL of 90% MeCN in water in a round-bottom flask and placed in a 55 °C water bath. TCEP (538 mg, 1.89 mmol) dissolved in 70% MeCN in water (7 mL) was added. Ethanol can be used instead of acetonitrile in the deprotection of 3. After 2 h, the reaction was cooled to room temperature, and the solvent was removed under vacuum. The crude product was extracted with EtOAc (3 × 20 mL), which was collected and back extracted with water (3 × 20 mL). The organic layer was collected and dried over anhydrous sodium sulfate. The solvent was removed under vacuum to yield the product as a white powder. Yield: 26.3 mg; 40%. The trithiol product is easily oxidized, and thus the crude reaction was used in situ in the complexation with antimony. ¹H NMR (DMSO; 600 MHz) δ ppm: 2.439 (t, 3H, CH₂SH), 2.680 (d, 6H, CH₂SH), 3.992 (s, 2H, OCH₂), 7.705 (d, 2H, CH), 8.084 (t, 1H, CH), 13.294 (s, COOH). (See Supplemental Figure S10.)

Synthesis of 5-((2,6,7-Trithia-1-stibabicyclo[2.2.2]octan-4-yl)-methoxy)isothalic Acid, 5.

Compound 5 was synthesized by two methods. Method a: Compound 3 (15 mg, 0.0354 mmol) and TCEP (101.4 mg, 0.3538 mmol) were dissolved in 5 mL of 70% ethanol in water and stirred on a hot plate at 55 °C for 2 h. After the reductive deprotection reaction was complete, SbCl₃ (11.81 mg, 0.0425 mmol) dissolved in 1 mL of ethanol was added, and the reaction mixture was kept at 55 °C for another 45 min. A white precipitate formed during this time. The reaction mixture was cooled, the white precipitate was collected by filtration, washed with water, followed by diethyl ether, and dried in vacuo. Yield: 14.08 mg; 85%. X-ray quality crystals were grown by dissolving 5 (5 mg) in DMSO (2 mL) containing H₂O (200 mL) at 70 °C and allowing the mixture to sit at room temperature for 2 weeks. ¹H NMR (DMSO; 600 MHz) δ ppm: 3.193 (s, 6H, CH₂S), 3.828 (s, 2H, OCH₂), 7.649 (d, 2H, CH), 8.093 (t, 1H, CH). ¹³C NMR (DMSO; 150 MHz) δ ppm: 28.66 (CH₂S), 56.02 (C(CH₂)₃), 76.46 (OCH₂), 119.28 (CH), 122.56 (CH), 132.60 (CCO), 158.58 (COCH₂), 166.32 (COOH). HRMS (m/z): 464.88606 (464.88795 calc’d for [M –H]⁻ of [C₁₃H₁₃O₅S₃Sb]). Elemental Anal. Calc’d (found) for C₁₃H₁₃O₅S₃Sb: C, 33.42 (30.49); H, 2.80 (2.74); S, 20.59 (17.32). FT-IR (cm⁻¹): 1711 (C=O), 1197 (C–O). Note, ¹H NMR shows excess residual water at 3.30 ppm. Elemental Anal. Calc’d (found) for C₁₃H₁₃O₅S₃Sb·2.5H₂O: C, 30.48(30.49); H, 3.54 (2.74); S, 18.78 (17.32). (See Supplemental Figures S11–S14.)

Method b:

Compound 3 (15 mg, 0.0354 mmol) and TCEP (101.4 mg, 0.3538 mmol) were dissolved in 2 mL of 70% ethanol in water in a 10 mL microwave vessel. The vial was capped and placed into the microwave unit and set at a fixed power of 15 kW and a temperature of 70 °C for 5 min to generate 4. Following the reduction reaction, SbCl₃ (11.81 mg, 0.0425 mmol) dissolved in 1 mL of ethanol was transferred into the reaction vessel via syringe. The reaction vessel was then microwaved at a fixed power of 15 kW and a control temperature of 70 °C for 5 min. After cooling, the reaction mixture was centrifuged, and the white precipitate was filtered and washed three times with water and three times with ether. Yield: 14.88 mg; 90%.

Direct Radioantimony Labeling and Purification.

The trithiol chelator (3, 10 mM in MeCN) was deprotected in 1:1 MeCN:H₂O using TCEP (100 mM in H₂O) to yield 4. After irradiation, the tin/radioantimony target (∼50 mg) was dissolved in 3 mL of concentrated HCl (3 h, 90 °C), and without purification, the ^1XXSb reacted in situ (30 min, 25 °C) with 0.01–1 mM 4. A C18 Sep-Pak (55–105 μm particle size, 125 Å pore size; Waters Corporation, Milford, MA) was preconditioned with 5 mL of ethanol (Fisher Chemical) and 5 mL of H₂O. The labeled target solution was diluted (1/20) with H₂O and passed through the preconditioned C18 cartridge. Five milliliters of H₂O was passed through the cartridge to remove unchelated tin target material, and [^1XXSb]5 was eluted in 3 mL of MeCN and dried under nitrogen. Next, 20 sequential 30 min HPGe activity assays quantified ^118mSb, ^120mSb, ¹²²Sb, and ¹²⁴Sb, and, by fitting the decay of ¹¹⁷Sb and ^117mSn’s shared 158.56 keV gamma to double exponential decay equations, determined ¹¹⁷Sb and ^117mSn activities in the final purified fraction.

Stability and Cysteine Challenges.

The stability of the [^1XXSb] 5 complex was determined by challenging with various chelating agents endogenous to biological systems. After drying, the purified [^1XXSb]5 was resuspended in either phosphate buffered saline (PBS; Thermo Scientific), 25 mM cysteine (Thermo Scientific) in PBS, or fetal bovine serum (FBS; ATCC, Manassas, VA) and allowed to sit at room temperature. At various time points (0, 24, 72 h), aliquots were analyzed by reversed phase high performance liquid chromatography (RP-HPLC) using a C18 Acclaim column (4.6 mm I.D. × 250 mm, 5 μm particle size, 120 Å pore size; DIONEX, Sunnyvale, CA) and 1 mL/min flow rate with the following H₂O:MeCN gradient: 25% MeCN 0–3.5 min; 25–50% MeCN 3.5–23.5 min; 50–90% MeCN 23.5–24 min; 90% MeCN 24–29 min; 90–25% MeCN 29–30 min; 25% MeCN 30–35 min. For the FBS challenge solutions, an equal volume of MeCN was added to the aliquot to precipitate large serum proteins, which were removed by centrifugation (12,000 rpm, 5 min; Beckman Coulter Microfuge 22R Centrifuge, Brea, CA) prior to HPLC analysis.

Single-Crystal X-ray Diffraction Analysis.

Single-crystal X-ray diffraction data were collected on a Bruker X8 Prospector diffractometer (Bruker-AXS, Inc., Madison, WI, USA) using Cu K_α radiation (l = 1.54178 Å) from a microfocus source. The crystals were cooled to 100 K during collection using a Cryostream 700 cryostat (Oxford Cryosystems, Oxford, UK). Hemispheres of data were collected out to resolutions of at least 0.81 Å using strategies of scans about the phi and omega axes. Unit cell determinations, data reduction, absorption corrections, and scaling were performed using the Bruker Apex3 software suite.⁴⁶ The crystal structure of 3 was solved by an iterative dual space approach as implemented in SHELXT,⁴⁷ and 5 was solved by direct methods.⁴⁸ Both structures were refined by full-matrix least-squares refinement using SHELXL⁴⁹ implemented via Olex2.⁵⁰ Non-hydrogen atoms were located from the difference maps and refined anisotropically. Hydrogen atoms were placed in calculated positions, and their coordinates and thermal parameters were constrained to ride on the carrier atoms. The crystal structure of 3 was found to contain regions of disordered solvent that could not be accurately modeled; these were treated by applying a solvent mask as implemented in PLATON SQUEEZE.⁵¹ Six hundred thirty-seven electrons were removed from a total void volume of 2014 ų per unit cell, equivalent to 1.5 acetonitrile molecules per formula unit. Crystal data, structure refinements, and bond distances and angles are reported in the Supporting Information (Tables S1–S4).

RESULTS AND DISCUSSION

In targeted radionuclide therapy contexts, ¹¹⁹Sb has potential to effectively eliminate disease when targeting cellular locations outside of the cell nucleus¹⁴ while mitigating off target radiation toxicities.^9,15 However, exploration of this potential is limited to in silico studies due to a lack of stable complexing agents capable of functionalization to a targeting moiety. No literature reports of stable radioantimony complexation by a bifunctional chelator exist.⁵² Complexation of radioantimony builds upon previous work complexing fellow pnictogen arsenic. Antimony is the heavier congener of arsenic. Both arsenic and antimony are thiophilic, generally found in nature as their sulfides (M₂S₃) and dithiolates (e.g., DMSA, DMPS, British anti-Lewisite (BAL)) are used for treating arsenic and antimony poisoning.⁵³ Thiolate compounds of Sb(III) have been reported, including a recent trithiolate complex.^54–57 Previously, various trithiol ligands and their ^natAs and ⁷⁷As complexes were reported.^37,39,40 This is the third iteration in our development of a hydrophilic bifunctional trithiol chelate.

Trithiol Ligand Synthesis.

A trithiocyanate protected trithiol ligand 3 was synthesized as shown in Scheme 1 and deprotected to 4 just prior to use as free thiols (−SH) are readily oxidized. Since alkyl bromides are good leaving groups and are readily substituted with the thiocyanate thiol-protecting group, pentaerythritol tetrabromide was used as the starting material. The two ester groups were then hydrolyzed to their carboxylic acids with one carboxylic acid group available for bioconjugation with a targeting moiety (e.g., peptide, antibody) and the other carboxylic acid improving hydrophilicity. Thiocyanate is a very stable thiol-protecting group and is easily reduced to the free thiol with the mild reducing reagent TCEP.

Scheme 1.

Synthetic Route for Trithiol Ligand 4

Compound 3 was reduced with TCEP in MeCN or ethanol with 20–50% water to form the reduced form of the free thiol −SH (compound 4) just prior to complexation with Sb(III) to form compound 5. Free thiol groups are readily oxidized and must be used in situ. The advantage of this synthetic route is that, following reduction to the free-thiol, no separation was needed for the metal complexation reaction with Sb(III), either at the macroscopic or radiotracer (nM or less) level, similar to observations with As(III).^37,39

Compounds 1–3 were characterized by their elemental analyses, ¹H NMR and ¹³C NMR spectra and HRMS mass spectra. Trithiol 4 was only characterized by its ¹H NMR spectra due to its ease of oxidation. Compound 3 was also characterized by single-crystal X-ray diffraction (Figure 1, Supplemental Figure S15). This trithiol chelator is expected to further improve the hydrophilicity of the resultant radiometal bioconjugate compared to its previous trithiol analogues.^37,39,40 The previous two trithiol bioconjugates were quite lipophilic based on the high hepatobiliary clearance of their ⁷⁷As labeled trithiol-bioconjugates in mice.^37,39 The presence of two carboxylic acid groups in 3 (Figure 1) allows one to be conjugated to a targeting moiety while leaving the second one to increase hydrophilicity relative to the previous analogues.

Figure 1.

X-Seed representation of 3 (CCDC #2071806) showing 50% ellipsoids.

Sb-Trithiol Complex.

The antimony trithiol 5 was successfully synthesized according to reaction Scheme 2. Precursor 3 (1 equiv) was reduced to trithiol 4 with TCEP (10 equiv) after which SbCl₃ (1.2 equiv) in ethanol solution was added. The product immediately precipitated from solution. The synthesis was evaluated by both conventional and microwave heating methods. Both reaction methods provided high yields over 85% of the product; however, the microwave reaction was more time efficient (10 min vs 45 min). The antimony trithiol (5) was characterized by ¹H NMR and ¹³C NMR spectroscopy, FT-IR spectroscopy, HRMS, and elemental analysis. X-ray quality crystals of 5 were obtained from a DMSO/H₂O mix (Figure 2; Supplemental Figures S16–S17). The ¹H NMR spectrum shows the expected disappearance of the −SH protons and a downfield shift of the −CH₂S protons of 4 on coordination to Sb. The FT-IR shows the expected stretches from the −COOH groups at 1711 cm⁻¹ (C=O) and 1197 cm⁻¹ (C–O).

Scheme 2.

Reaction Scheme for Synthesis of 5, Sb-Trithiol

Figure 2.

X-Seed representation of 5 (CCDC #2071807).

X-ray Diffraction Studies.

Compound 3 crystallized from MeCN and water in the rhombohedral $R\overline{3}$ space group. It packed forming discrete hydrogen bonded rings containing six molecules, all interacting through the carboxylic acid groups. The six-molecule rings stacked above and below each other and interfaced through the thiocyanate groups, which interact with each other through electric dipole interactions and with the phenyl π system. The disordered solvent is modeled as 1.5 acetonitrile molecules per formula unit. Bond lengths and angles are very similar to the previously characterized trithiocyanate protected trithiol ligands.⁴⁰ Selected bond distances and angles are listed in Table 2.

Table 2.

Selected Bond Distances (Å) and Angles (deg) for 3 and 5

compound 3 (CCDC #2071806)		compound 5 (CCDC#2071807)
bond distances (Å)
S1-C9	1.836(5)	Sb-S1	2.4416(5)
S1-C14	1.681(6)	Sb-S2	2.4314(5)
S2-C10	1.858(5)	Sb-S3	2.4465(5)
S2-C15	1.658(5)	S1-C9	1.832(2)
S3-C11	1.806(5)	S2-C10	1.832(2)
S3-C16	1.672(5)	S3-C11	1.832(2)
N1-C14	1.172(7)
N2-C15	1.198(12)	O2(acid) ....O1S(DMSO)	2.602(2)
N3-C16	1.171(6)	O4(acid) ....O1S(DMSO)	2.562(2)
bond angles (deg)
N1-C14-	S1	177.6(6)	S1-Sb-S2	92.25(2)
N2-C15-	S2	173.9(8)	S1-Sb-S3	91.965(16)
N3-C16-	S3	176.1(5)	S2-Sb-S3	90.62(2)

Compound 5 crystallized from DMSO and water over 2 weeks. The crystal structure of compound 5 (Figure 2) confirms the expected molecular structure, which is consistent with its solution structure based on NMR and HRMS. The molecule crystallized with 2 equiv of DMSO in the monoclinic space group P2₁/c. Both acid groups are protonated, and the three thiols are coordinated trigonally to the Sb(III), making it a discrete, neutral molecule. The geometry about the Sb atom is trigonal pyramidal, but the S–Sb–S bond angles are all closer to 90° (an octahedron missing 3 vertices) than they are to 109.5° (a tetrahedron missing one vertex). The S–Sb–S bond angles in the structure are in agreement with those previously reported.^54,58,59 Previously characterized As-trithiol complexes have bond angles about arsenic close to 97°.⁴⁰ The average Sb–S bond length is 2.440(8) Å, which is in agreement with the previously reported average Sb–S (2.447(7) Å) bond distance.^54,58,59 Selected bond distances and angles are listed in Table 2.

Several different interactions influence the crystal packing in this structure. An important interaction appears to be between the Sb atom and the S atoms of adjacent molecules, which organize the molecules into chains along a (Supplemental Figure S17). Each Sb atom interacts with 2 S atoms from one neighboring molecule (Sb·····S2 = 3.6987(6) Å, Sb·····S3 = 3.5370(6) Å) and a third S atom from a second neighboring molecule (Sb·····S1 = 3.2529(5) Å), although they are not octahedrally arranged about the Sb. This seems to be an interaction of the Lewis acidic Sb and the Lewis basic nonbonding S lone pair and has been observed in previously reported structures.^54,58,59 There are also two short S·····S distances between neighboring molecules (S1·····S2 = 3.3212(7) Å, S2·····S3 = 3.3829(7) Å). These interactions polymerize the structure into an infinite chain parallel to the a axis (Supplemental Figure S17). Additionally, the acids groups each donate a hydrogen bond to the oxygen of a DMSO molecule (Table 2) and accept a hydrogen bond from the methyl group of a different DMSO molecule. These hydrogen bonds cross-link the chains into a network.

Target Irradiation and Chelator-Based Separation.

A choice cyclotron target provides strong thermal adherence and electrical conductance, dissipating kW/cm² beam thermal power, conducting tens of microamperes of electrical current, and withstanding transitions from microtorr pressures to standard atmosphere. Electrodeposition uses electrical reduction to transition metal ions in solution to a solid phase, forming a metallic material that is electrically and thermally well-joined to a durable, rigid backing. Electroplated ^natSn targets with lineal mass density 88–128 mg/cm² withstood 8 MeV energy deuterons at a beam current of 40 μA. End of saturated bombardment (EOSB) theoretical, estimated, and measured yields and EOB measured yields are reported in Table 3. The yield at the end of bombardment, A_EOB, is described mathematically according to eq 1, where N is the target atom density, x is the thickness of material slice under consideration, I is the accelerated particle current, σ is the nuclear reaction cross section at specified particle energy, λ is the decay factor, and t is the time duration of irradiation.

Table 3.

Relevant Antimony Radioisotopes Produced from Deuteron Bombardment of Natural Enrichment Tin and Theoretical, Estimated 119Sb, and Measured EOSB Yields^a

antimony radioisotope	theoretical EOSB yield (MBq·μA−1)		measured EOSB yield (MBq·μA−1)	measured EOB yield (kBq·μAh1-)
117Sb	d,n	32.2
	d,2n	7.9
	total	40.1	48.0 ± 6.2	680 ± 100
118mSb	d,n	4.0
	d,2n	0.3
	total	4.3	1.2 ± 0.4	60 ± 2
119Sb	d,n	50.9
	d,2n	14.8
	total	65.7	70–80b	240–250b
120mSb	d,n	4.1
	d,2n	4.0
	total	8.1	1.60 ± 0.07	4.9 ± 0.8
122Sb	d,2n	6.5	7.3 ± 1.5	39 ± 8
124Sb	d,2n	7.1	12 ± 4	2.9 ± 0.9
125Sb	d,n	2.2	<2.6c	<0.7c

^aUncertainty is presented as the standard deviation of replicate measurements (N = 3).

^bEstimated ¹¹⁹Sb EOSB and EOB yields.

^cNo detected ¹²⁵Sb activity (limit of detection = 7 kBq)⁶¹.

$$A_{\text{EOB}}=Nx\sigma I\left(1-e^{-\lambda t}\right)$$ (1)

A column was added to Table 3 listing EOB values for comparison.

The yield at EOSB, or “end of saturated bombardment”, A_EOSB, describes the yield when the rate of production is equal to the rate of radioactive decay. It is convenient for comparing yields of short-lived radionuclides. EOSB yield is described by eq 2.

$$A_{\text{EOSB}}=Nx\sigma I$$ (2)

Conversion between EOB and EOSB is described in eq 3 and is dependent only upon the half-life of the radionuclide and the length of time duration of the irradiation.

$$A_{\text{EOSB}}=Nx\sigma I=A_{\text{EOB}}/\left(1-e^{-\lambda t}\right)$$ (3)

We report a measured EOSB yield as opposed to a theoretical number predicted from a modeled nuclear excitation function.

Proton irradiation of ^natSn produces larger quantities of long-lived radioantimony impurities: ¹²⁴Sb (t_1/2 = 60.2 d), ¹²²Sb (t_1/2 = 2.72 d), and ^120mSb (t_1/2 = 5.76 d). Deuteron irradiation of ^natSn produces only small quantities of ^120mSb, ¹²²Sb, ¹²⁴Sb, the relatively short-lived ¹¹⁷Sb (t_1/2 = 2.8 h), and ^118mSb (t_1/2 = 5.0 h) in addition to desired ¹¹⁹Sb, dramatically reducing troublesome longer-lived contaminants. For example, deuteron irradiation of the most abundant tin isotope, ¹²⁰Sn (32.58%), predominantly forms stable ¹²¹Sb via the ¹²⁰Sn(d,n)¹²¹Sb reaction. With a reaction threshold of 4.7 MeV, the ¹²²Sn-(d,2n)¹²²Sb nuclear reaction is initiated when bombarding ^natSn, producing less long-lived ¹²²Sb than proton bombardment of ^natSn. Using TALYS theoretical excitation functions,⁶⁰ a typical 1 h, 40 μA 8 MeV deuteron irradiation of a 100 mg ^natSn target theoretically produces approximately 47 MBq of ¹¹⁹Sb, 350 MBq of ¹¹⁷Sb, 22 MBq of ^118mSb, 1.6 MBq of ^120mSb, 2.8 MBq of ¹²²Sb, 0.1 MBq of ¹²⁴Sb, and 2.6 kBq of ¹²⁵Sb at EOB. A similar length and intensity 12 MeV proton irradiation would produce a factor of 6 more long-lived ^120mSb (10.4 MBq), 14 times more ¹²²Sb (40 mBq), and over 6 times the activity of very-long-lived ¹²⁴Sb (0.9 MBq). Although less suitable than proton irradiation for producing large quantities of ¹¹⁹Sb, deuteron irradiation produces relatively low ¹²²Sb and ¹²⁴Sb activities, the longest lived and most problematic antimony radioisotopes in a working laboratory environment, while also producing relatively larger amounts of ¹¹⁹Sb. The deuteron targets also require less ^natSn material because deuterons have a shorter range than protons, reducing possible contaminant mass in processing. Thus, deuteron irradiation of ^natSn is an economical route to developmental, preclinical quantities of ¹¹⁹Sb and ^120mSb.

Due to the inability to resolve the low characteristic photon emission of ¹¹⁹Sb (23.87 keV), the EOSB yield of ¹¹⁹Sb was not measured directly. Instead, ¹¹⁹Sb EOSB yield was estimated using the ratio of ¹¹⁷Sb to ¹¹⁹Sb TALYS theoretical excitation functions⁶⁰ and the measured ¹¹⁷Sb EOSB yield. Because ¹¹⁷Sb and ^117mSn share a 158.56 keV characteristic emission, ^117mSn activities within irradiated targets were quantified after ¹¹⁷Sb decayed away (>30 h), and measured ¹¹⁷Sb EOSB yields were corrected for ^117mSn 158.56 keV signal contribution.

Radiolabeling.

Trithiol 4 rapidly complexes n.c.a. pmol quantities of antimony in 30 min at room temperature in the presence of mmol quantities of the tin target (10⁸-fold excess) at ligand concentrations down to 0.01 mM. At low radiolabeling concentrations of 10 μM, 30 nmol of ligand was used to complex 5.8 pmol of various antimony isotopes—calculated from HPGe gamma spectroscopy measured radioantimony activities, theoretical accelerator produced stable antimony isotopes, and trace metal contaminant limits of detection from reagents within the dissolved target solution. After an irradiation of 30 μAh 8 MeV deuterons, the calculated molar activity for 49 MBq of ¹¹⁷Sb is 8.4 MBq/pmol, and the calculated molar activity for an estimated 240 kBq ¹¹⁹Sb is 42 kBq/pmol. HPLC traces of radioantimony unbound and bound by 4 are shown (Figure 3). The nonradioactive [^natSb]5 standard coelutes with radioactive [^1XXSb]5 with a retention time of 23.7 min (Figure 4). The trithiol chelator 4 has a strong selectivity for antimony over tin as seen by the full chelation of radioantimony (<nM) in the presence of macroscopic quantities of tin and harsh (∼6 M HCl) solvent conditions. The ability to directly radiolabel radioantimony quantitatively from unseparated target material greatly simplifies radiopharmaceutical production. Greater than 99% of [^1XXSb]5 activity was trapped onto the C18 cartridge and eluted with a radiochemical yield of 65% ± 20% (N = 3).

Figure 3.

HPLC radiation detector traces showing chelation of radioantimony using trithiol chelator 4: a) unchelated radioantimony R_t = 1.5 min, b) direct radiolabeling of radioantimony from an unseparated target solution using 4 R_t = 24.0 min post C18 cartridge purification, c) 24 h 25 mM cysteine challenge R_t = 24.0 min, and d) 72 h FBS challenge R_t = 23.6 min.

Figure 4.

HPLC traces of [^1XXSb]5 and [^natSb]5: a) 280 nm UV–vis spectrum of compound 5 R_t = 24.5 min post C18 cartridge purification, b) 280 nm UV–vis spectrum of [^natSb]5 standard R_t = 23.8 min, c) 280 nm UV–vis spectrum of coinjection containing both C18 cartridge purified [^1XXSb]5 and [^natSb]5 standard R_t = 23.7 min, and d) radiotrace of matching coinjected [^nat/1XXSb]5 R_t = 23.7 min.

[^1XXSb]5 was stable over 72 h when challenged with biologically relevant complexing agents. HPLC analyses showed [^1XXSb]5 to be 91% ± 9% (N = 3) intact in 25 mM cysteine at 72 h and 97.5% ± 1.6% (N = 3) intact in FBS at 72 h. HPLC traces of [^1XXSb]5 challenged with 25 mM cysteine at 72 h and FBS at 24 h are reported in the Supporting Information (Supplemental Figure S18). [^1XXSb]5 is thus expected to be stable in vivo. Unchelated compound 4 has a retention time of 24.5 min (Figure 4), which is approximately 0.5 min longer than [^1XXSb]5, and the two peaks would be distinguishable and separable. The molar activity of the final solution can be increased by removing nonradiolabeled chelator from [^1XXSb]5. The reaction of trithiol chelator 4 with antimony is the first reported complexation of radioantimony with a bifunctional chelator capable of both stably retaining the radiometal and providing a linker group that can be conjugated to disease targeting moieties.

Fractions isolated from the HPLC [^1XXSb]5 peak, R_t = 24.0 min, were assayed by HPGe, and the spectra are shown in Figure 5 left. To determine the degree to which 4 complexes solely antimony and not tin target material, ^117mSn activity within a HPLC purified fraction was measured. Antimony-117 decays to ground state ¹¹⁷Sn.¹¹ The fitted, logged decay of the sample enabled a half-life measurement that distinguishes ¹¹⁷Sb (2.80 h) from ^117mSn (14 d). Fitting the 158.56 keV photopeak over a 30-h span constructed the decay curve (Figure 5 right) and measured a half-life of 2.86 ± 0.02 h, 2.3% larger than the true ¹¹⁷Sb half-life, 2.80 h.¹¹ A measured half-life larger than the true half-life indicates the presence of a longer-lived radionuclide, in this case, ^117mSn.

Figure 5.

HPGe gamma spectrum and produced activity curves. (left) HPGe gamma spectrum of C18 purified sample labeled with characteristic emissions, displaying radionuclidic purity. ¹²²Sb (692.65 keV) and ¹¹⁷Sb (861.3 keV, 1004.5 keV) peaks are observed but unlabeled for graphic clarity. (right) Double exponential equation fitting of time activity curve.

Double exponential decay equations describe and quantify relative activities of mixed radionuclide samples in time; using this fit, the ^117mSn activity coeluting with [^1XXSb]5 at HPLC separation was calculated to be 11.29 ± 0.12 Bq, and comparing initial reaction to final purified ^117mSn activity provides a tin decontamination factor of 1.41 × 10³. This decontamination factor, describing the high level at which tin target material was removed from the radiopharmaceutical, is impressive for a nontraditional radiochemical production and resulted in a final formulation of tin mass that is orders of magnitude below the estimated daily intake of 4.003 mg inorganic tin for an adult in the United States.⁶² For a target with lineal mass density of 120 mg/cm², the final sample holds an estimated 45 μg of tin. The Agency for Toxic Substances and Disease Registry (ATSDR) reports no evidence that inorganic tin is a neurotoxin, mutagen, carcinogen, or immunotoxin or affects reproduction or development in humans.⁶² No radionuclidic impurities were observed besides the various radioantimony isotopes (useful for radioantimony activity quantification) and ^117mSn (Figure 5 left).

CONCLUSIONS

Deuteron irradiation of inexpensive ^natSn targets produces a profile of radioantimony isotopes with far shorter half-lives compared to proton bombardment while still producing preclinical quantities of ¹¹⁹Sb, making it a better production route for research and development purposes. The various radioantimony isotopes allow useful tracking and quantification of radioantimony activity. No unexpected radionuclide impurities were observed in the final product. We report a method for radioantimony chelation using a functionalizable trithiol ligand, circumventing usual radionuclide isolation from dissolved accelerator targets. This is the first report of radioantimony complexation with a chelator capable of bifunctionalization—an essential step toward exploration of ¹¹⁷Sb and ¹¹⁹Sb in theranostic applications of targeted radionuclide therapeutic contexts.