Introduction

Astatine-211 is one of the most attractive radionuclides for targeted alpha therapy (TAT)1,2,3,4. It has a 7.21 h half-life, very low abundance of high energy gamma-ray emissions and 100% alpha emission. Astatine-211 can be produced by bombardment of high purity bismuth metal (naturally monoisotopic) targets with 28–29 MeV alpha particles. Dry distillation is widely used for isolation of 211At from irradiated Bi targets. Various dry distillation set-ups have been described in the literature5,6,7,8,9,10,11, but handling of radioactive gases has raised concerns by institutional radiation safety officials. Moreover, if large quantities of 211At are to be produced, the size of bismuth target must be increased significantly. This increase in size of target is due to the low melting point of Bi (271.5 °C), requiring that the alpha beam be spread over a large area to provide efficient cooling during irradiation. Implementing the dry distillation method when using larger bismuth targets can present significant challenges due to the fact that increasing the size of the quartz tube for distillation of Bi can make it difficult to find commercial tube ovens for this purpose and the larger tubes can make the distillation process less efficient12. An alternative to placing large irradiated bismuth targets (and their backing material) in an oven is to mechanically remove the irradiated bismuth from the target backing, followed by placing the bismuth in a high temperature oven (650–750 °C)5. The safety of this process has also raised concerns from radiation safety officials.

An alternative method of isolating 211At from irradiated bismuth targets involves “wet chemistry” liquid-liquid extraction processes13,14,15,16,17. We have been able to consistently obtain high isolation yields using this method, but it is a 2.5–3 hour process where distillation of concentration HNO3 and multiple liquid-liquid extraction steps are required12. In the isolation process irradiated bismuth targets are dissolved in concentrated HNO3, but the organic phase used, diisopropyl ether (DIPE), cannot effectively extract 211At from HNO3 solutions. Therefore, the HNO3 is removed by distillation and 8 M HCl is used to re-dissolve the Bi(NO3)3 salt residue. Liquid-liquid extractions are then performed to isolate 211At from the 211At/Bi3+ mixture and to remove Bi3+ salts using DIPE and 8 M HCl. We have investigated automation of this wet chemistry, liquid-liquid exaction method for 211At isolation, and although it has been technically challenging, we have had some success in automation18. Unfortunately, thus far we have not been able to decrease the time required using the automated “wet chemistry” 211At isolation process from that achieved in the manual separation procedure (unpublished data). In an effort to simplify the isolation process and decrease the time to obtain the isolated 211At, we looked for alternative isolation methods.

During the separation process in the wet chemistry method, astatine undergoes several changes in its oxidation state, possibly from At(+5) to At(+3), then to At(0), and finally to At(−1)14. Along with the change in oxidation state there are likely different chemical species produced. In addition to astatide, four other unknown astatine species have been observed at different times by anion exchange radio-HPLC12. The inconsistency in the radiochemical purity of the 211At isolated using the DIPE extraction method can result in poor radiolabeling yields, which can be a major problem in fulfilling prescribed doses in the clinical setting.

Here we report a new 211At-isolation approach based on a tellurium-packed column (Te column) previously described in the literature. Bochvarova et al. reported a method of using two tellurium metal packed columns to effectively isolate 211At from 660 MeV proton beam irradiated thorium targets19. Astatine-211 can be rapidly absorbed on metallic tellurium in HCl in the presence of SnCl2 and eluted by a solution of 1–2 M NaOH. In order to adapt the Te column method to irradiated bismuth targets, we used NH2OHHCl to convert the HNO3 solution containing the dissolved target to a solution of HCl. We also demonstrated this new method can be readily automated and can provide [211At]NaAt of consistent and high radiochemical purity.

Results

At-211 isolation using Te columns

Figure 1 summarizes the steps involved in the process of isolating 211At from irradiated Bi targets using Te columns. Step 1 involves dissolution of the bismuth target using 10 M HNO3. This is the same initial step used in the automation of the wet chemistry isolation process18. Step 2 involves addition of NH2OH·HCl to the HNO3 solution containing dissolved Bi dropwise until complete cessation of bubbling is noted. Steps 3–6 are conducted as shown in Fig. 1. However, in our initial studies steps 3 and 5 were slightly different, as a reductant, SnCl2 was used to assure that the 211At was in the astatide form. The steps initially used were as follows: In step 3, 0.1 M SnCl2 in 6 M HCl was used instead of 1.5 M HCl and in step 5, the column was washed by 0.1 M SnCl2 in 6 M HCl, 6 M HCl and deionized (D.I.) H2O sequentially. As NH2OH·HCl is a strong reducing agent, the use of 0.1 M SnCl2 seemed to be redundant but we thought it would be best to evaluate isolation yields with and without having SnCl2 present.

Figure 1
figure 1

Flow chart for the isolation process of 211At from irradiated Bi targets using Te columns.

Table 1 shows that the astatine adsorption kinetics of the Te column is very fast and efficient, irrespective of whether SnCl2 was used or not. Essentially all of the 211At was absorbed by the Te column from 1.5 M HCl even at a high flow rate of 5–10 mL/min. Moreover, very little activity was found in the washes (<0.1%) and isolation yields ~75% were obtained when the 211At/Bi mixture was adjusted to 0.1 M SnCl2 in 6 M HCl. Comparable or even higher 211At isolation yields were obtained without the use of SnCl2 (Table 1). Therefore, the later 211At isolation experiments were conducted without SnCl2, as outlined in Fig. 1.

Table 1 Isolation yields of 211At using Te columns in the presence or absence of SnCl2 in the loading mixturea.

Semi-automated 211At isolation runs

The results of the manual 211At isolation runs were very encouraging, so the procedure was adapted to a semi-automated system to demonstrate potential application of the Te column method for routine 211At isolation. Three proof-of-concept semi-automated 211At isolation runs were conducted using irradiated bismuth targets following the procedure outlined in Fig. 1. The schematic of the semi-automated process is shown in Fig. 2. A picture of our prototype is shown in Fig. S1. Results of the semi-automated runs are summarized in Table 2. The irradiated bismuth targets each contained about 962 MBq (26 mCi) and the overall 211At isolation process run time was 90–100 minutes. On average, the system was able to recover 93±4% of the 211At in 2 mL of 1 M NaOH. For the 1st and 2nd semi-automated runs, 80 mL of 35% aqueous NH2OHHCl was used and the columns were washed with 20 mL of 1.5 M HCl, followed by 20 mL of deionized (D.I.) H2O. For the 3rd semi-automated run, the volume of 35% NH2OHHCl was reduced by 15 mL which appears to be enough to destroy all of the nitrate. Decreasing the volume of aqueous NH2OHHCl reduced the overall run time, even though an additional 20 mL of 1.5 M HCl and D.I. H2O was used for washing the column in the 3rd semi-automated run.

Figure 2
figure 2

Schematic of the semi-automated process of 211At isolation using a Te column. In the process two syringe pumps move the solvents and mixtures through valves (A–E). Numbered solutions are as follows: 1: 10 M HNO3; 2: air; 3: D.I. H2O; 4: 35% NH2OHHCl; 5: 8 M HCl; 6: 1.5 M HCl; 7: 1 M NaOH. Waste 1: non-radioactive waste; Waste 2: radioactive waste. (Detailed description of equipment and reagents can be found in Fig. S1).

Table 2 Quantities of reagents, isolated yields and run times from three semi-automated 211At isolation experiments.

Optimization of the amount of NH2OH·HCl

Although the 211At isolation yields were encouraging when an excess of NH2OH·HCl was used, we obtained inconsistent radiolabeling yields, ranging from 10.4% to 94.7%, when using the isolated 211At solutions to label isothiocyanato-phenethyl-closo-decaborate (2-) (B10-NCS)-conjugated monoclonal antibodies (MAbs) at room temperature in the absence of an oxidant. We also observed that addition of the oxidant chloramine-T improved the MAb-B10 radiolabeling yield, suggesting that there might be reductive impurities in the isolated 211At solution.

We hypothesized optimizing the amount of NH2OH·HCl might mitigate this problem. Experiments were conducted to determine the minimal volume of 35% NH2OH·HCl required to destroy all the NO3 (Vmin) in the HNO3 solution containing the dissolved Bi target, as it seemed to be impossible to accurately assign defined stoichiometry to the reaction between NH2OH·HCl and HNO3. Minimally, about 52 mL of 35% NH2OH·HCl is needed for a Bi target that is dissolved in 15 mL of 10 M HNO3. A series of manual 211At isolations were conducted using different percentages of Vmin. Table 3 shows that although the 211At isolation yield decreased as the volume of 35% NH2OH·HCl was reduced, the radiolabeling yields for B10-conjugated MAb increased significantly. Reducing the amount of NH2OH·HCl to 52% of Vmin appears to be optimal, as the radiolabeling yield increased to 82.5%, equivalent to the normal radiolabeling yield achieved with 211At isolated using the DIPE method. And at 52% of Vmin, an isolation yield of about 80% could still be obtained.

Table 3 Using various amounts of NH2OHHCl for 211At isolation and its influence on the B10-conjugated MAb labeling yield.

Chemical and radiochemical purity

The Te column method can effectively isolate 211At from large-sized Bi targets containing 4–5 g of Bi metal, but trace levels of Te and Bi impurities in the isolated 211At solutions were anticipated. The concentrations of Te and Bi impurities in the 211At solutions were evaluated by ICP-MS after the 211At decayed. Calibration curves were generated using Bi and Te standard solutions at concentrations of 1, 10, 50 100 ppb, with R2 ≥0.9995. Recoveries of the internal standards (ISTDs) were very close to 100% for all samples including calibration standards. On average, the concentration of residual Bi in the isolated 211At solution was 3.0 ppm, which is slightly higher than that of the 211At isolated using the DIPE method (Table 4). The concentration of the Te contaminant in the isolated 211At solution is rather high, about 32.8 ppm on average. However, it should be noted that 211At-labeled MAb prepared from Te column isolated 211At and purified by a size-exclusion (PD-10) column had significantly reduced Bi and Te concentrations of about 0.05 and 0.04 ppm, respectively (Table 4).

Table 4 Levels of Te and Bi impurities in isolated 211At and 211At-labeled MAb.

Anion exchange radio-HPLC analyses were performed to evaluate the radiochemical purity of the isolated 211At. A representative radio-chromatogram is shown in Fig. 3. Only one radiopeak at 9.5 min has been observed for several 211At solutions purified using the Te column method (n > 10), which suggests radiochemical purity >99% can be consistently achieved using the Te column method.

Figure 3
figure 3

Representative anion exchange radio-HPLC chromatogram of the [211At]NaAt isolated using the Te column method.

Discussion

Astatine-211 can be rapidly absorbed on metallic Te in HCl in the presence of SnCl2 and eluted by a solution of 1–2 M NaOH. The high affinity of At to elemental Te in HCl might be the result of the formation of a coordination bond between the surface Te and the highly polarizable At− 19. In the literature, Te columns were used for separation of 211At from irradiated thorium targets19 or polonium impurities, when 211At is produced via high energy proton induced spallation of thorium19 or the 209Bi(7Li, 5n)211Rn → 211At route20, respectively. In those scenarios, the solutions containing 211At/impurities do not contain large amounts of HNO3 and can be readily diluted and constituted to approximately 0.1 M SnCl2 in 6 M HCl prior to loading onto the Te column. However, dissolution of our Bi targets which contain 4–5 grams of Bi metal requires the use of 15–17 mL of 10 M or concentrated HNO3. Hydroxylamine can reduce HNO3 to HNO2 which further reacts with NH3OH+ and produce gaseous N2O (g), and N2 (g), so it is used to convert the nitrate matrix to HCl21,22. The addition of NH2OHHCl also eliminates the need for using SnCl2 in the solution transferred onto the Te column. It is likely that NH2OHHCl reduces astatine in other oxidation states to astatide, the astatine species that might be required for the Te column method to work properly.

Compared to the DIPE extraction method, using NH2OHHCl for converting the HNO3 solution containing the dissolved Bi target to a HCl matrix is not only easier to automate, but can be faster than distilling the HNO3 to dryness. It takes ~25 min to completely destroy the nitrate, adding the NH2OHHCl solution (Vmin = 52 mL) using the semi-automated system, which is comparable to the time it takes to remove the HNO3 by distillation (~30 min)12. However, we found that not all of the nitrate needs to be destroyed to obtain good isolation yields. In fact, adding 52% of Vmin provides high 211At isolation yields as well as high B10-conjugated MAb labeling yields (Table 3).

The overall run times of the semi-automated 211At isolation experiments are 20–30 min shorter than those of the DIPE extraction process (Table 2). The high affinity of At to elemental Te allows adsorption of 211At onto Te columns and washing the Te columns with HCl and H2O at high flow rates of 6 mL/min. Also, the elution of 211At from the Te column using NaOH is very efficient, averaging 93% of the 211At in 2 mL volume at a flow rate of 60 mL/min. It should be noted the NaOH back extraction step in the DIPE extraction method can takes 10–20 min to finish12. The fast flow rates used in the Te column isolation process are critical for achieving good 211At isolation yields in a reasonable amount of time, especially, considering the volume of 211At solution passing over the Te columns is rather large.

Astatine-211 solutions obtained using the DIPE liquid-liquid extraction method can have multiple astatine species which can lead to low radiochemical purity12. In contrast, 211At solutions isolated using Te columns consistently provide only astatide in a radiochemical purity >99%. This might be due to 211At being reduced to astatide by NH2OHHCl before being transferred onto the column. However, it must be noted that a small amount of tellurium metal is dissolved in 1 M NaOH as 211At is eluted off the column and Na2TeO3 is a weak reducing agent. While untested at this time, the presence of Na2TeO3 might cause problems for astatine labeling, especially electrophilic astatination reactions. Thus, methods for purifying 211At from the Te (and possibly Bi) impurities need to be evaluated for applications that require higher purity.

In conclusion, Te columns provide an alternative method for efficiently isolating 211At from irradiated Bi targets. The isolated 211At solution is of high radiochemical purity and is suitable for B10-conjugated MAb labeling. A semi-automated process based on the Te column method has been demonstrated. Studies to evaluate the influence of the mesh size of the Te powder used on the 211At isolation yield are on-going. As future work, the geometry of the Te column, the volumes of reagents including 1.5 M HCl, D.I. H2O and 1 M NaOH need to be optimized to minimize the overall run time and to reduce the volume of the final product to 0.5 mL or less.

Methods

Reagents and general procedures

The chemicals and reagents used were purchased from VWR International (Radnor, PA), Sigma Aldrich (St. Louis, MO) or Fisher Scientific (Pittsburgh, PA), and were used without further purification unless otherwise specified. Empty Mini Spe-ed column cartridges were obtained from Applied Separations (Allentown, PA). ICP-MS tuning solution and 10 μg/mL standard solutions of Bi and Te were obtained from Inorganic Ventures (Christiansburg, VA). Monoclonal antibodies were obtained from the Fred Hutchinson Cancer Research Center Biologics Production Facility and isothiocyanato-phenethyl-closo-decaborate (2-) (B10-NCS)-conjugated MAbs were prepared in house as previously described23. Astatine-211 was produced by irradiation of Bi metal on an aluminum target support with 29 MeV α-particles using the Scanditronix MC50 cyclotron as previously described24. Various solutions containing 211At generated before and after Te column separation were measured in a Capintec CRC-55tR dose calibrator using the calibration setting number 44. Astatination of the B10-conjugated MAbs with 211At isolated using Te columns were conducted as previously described25.

Determine the minimal volume of NH2OH·HCl

Determining the production rate of each of the products generated by the HNO3 and NH2OH·HCl redox reaction was not attempted. It seemed that accurately assigning a defined stoichiometry to this reaction was not possible, thus the minimal volume of 35% NH2OH·HCl required to destroy all the nitrate ions in the dissolved Bi target solution (Vmin) was determined experimentally.

To mimic the semi-automated isolation process, 4.25 g of Bi metal was dissolved in 15 mL of 10 M HNO3. The resultant solution was split into three 5-mL fractions and 35% aqueous NH2OH·HCl was added dropwise at a flow rate of approximately 2 mL/min using a 25-mL burette. The completion of the NH2OH·HCl and nitric acid reaction was observed as a cessation of bubbling. Nitrate/nitrite test paper (EMD MilliporeTM) was used to verify the nitrate ion was below the detection limit (10 ppm).

Manual 211At isolation using Te columns

Each Mini Spe-ed column cartridge (Applied Separations, Allentown, PA) was filled with 1 g (dry weight) of tellurium metal powder (−200 mesh). The columns were washed sequentially by 20 mL of 2 M NaOH, 3 M HCl and D.I. H2O. Prior to use, the Te columns were pre-equilibrated with 20 mL of 0.1 M SnCl2 in 6 M HCl or 1.5 M HCl.

An irradiated Bi target containing ~0.96 GBq (~26 mCi) of 211At was placed Bi face down in a plastic container. A total of 17 mL of concentrated HNO3 was manually added to dissolve the Bi and 211At. Because 10 M HNO3 rather than concentrated HNO3 would be used for target dissolution in the semi-automated 211At isolation process, 1 g of high purity (99.999% trace metal basis) Bi beads were dissolved in 3 mL of 10 M nitric acid to mimic the Bi3+ and NO3 concentration in the dissolved Bi target solution obtained using the semi-automated process. The resultant solution was spiked with 0.5–1 mL concentrated HNO3 of dissolved irradiated Bi targets. Depending on the chromatographic conditions to be evaluated, the obtained solution was adjusted accordingly.

Loading in 0.1 M SnCl2 in 6 M HCl. Thirty five percent NH2OH·HCl was added dropwise until the cessation of bubbling. The solution was then adjusted to 0.1 M SnCl2 in 6 M HCl using 1 M SnCl2 and concentrated HCl. After loading the resultant solution onto the Te column, the column was eluted with 20 mL of 0.1 M SnCl2 in 6 M HCl, followed by 20 mL of 6 M HCl, D.I. H2O and 2 mL of 1–2 M NaOH.

Loading in 1.5 M HCl. Various amounts of 35% aqueous NH2OH·HCl were added dropwise, followed by 3–4.35 mL of 8 M HCl to reconstitute the solution to approximately 1.5 M HCl. After loading the resultant solution onto the Te column, the column was eluted by 20 mL of 1.5 M HCl, D.I. H2O and then 2 mL of 1–2 M NaOH.

The semi-automated 211At isolation process

The schematic of the semi-automated system is shown in Fig. 2 and component details of the system are provided in Fig. S1. An irradiated Bi target containing ~0.96 GBq (~26 mCi) of 211At was manually put into the polyetherimide (ULTEMTM) dissolution chamber18, then the semi-automated 211At isolation process was started from the computer. The magnetic stirrer was turned on manually. Using the 10-mL syringe installed on the Hamilton syringe pump A, 15 mL of 10 M HNO3 (1) was pumped through the dissolution chamber to dissolve the target at a flow rate of 2.4 mL/min. The HNO3 solution obtained from the dissolved target was collected in the 250-mL round bottom flask. Air (2) was pushed through the dissolution chamber to ensure all the HNO3 was transferred into the round bottom flask. D. I. H2O (3) was used to rinse the 10-mL syringe and was moved into Waste 1. Using the same 10-mL syringe, at a flow rate of 2.4 mL/min, 65 mL or 80 mL of 35% NH2OH·HCl (4) was added to the dissolved target solution to destroy the nitrate ions, followed by 17.1 mL or 22 mL of 8 M HCl (5) to make the solution 1.5 M in HCl. The total volume after these steps was about 97 mL or 117 mL, depending on volume of 35% NH2OH·HCl used. Using a 25-mL syringe on Hamilton syringe pump B, 25 mL of 1.5 M HCl (6) was passed over the Te column into Waste 1 to pre-equilibrate the column. After the column was equilibrated, 21.5 mL of the dissolved target solution was transferred into a 25-mL loop, then flowed through the Te column at a flow rate of 6 mL/min. This step was repeated multiple times to load all the radioactive solution onto the column. During this process, the effluent from the column containing Bi and other impurities was delivered into Waste 2. The Te column was then rinsed with 20 or 40 mL of 1.5 M HCl, followed by 20 or 40 mL of D.I. H2O. Finally, the column was eluted using 5–7 mL of 2 M NaOH in 1 mL fractions at a flow rate of 60 mL/min.

Radio-HPLC and ICP-MS analysis

Radio-HPLC. Radio-HPLC analyses were performed on the 211At solutions using a Hewlett-Packard model 1050 HPLC (Hewlett-Packard Company, Palo Alto, CA) with a Beckman Model 170 Radioisotope Detector (Beckman-Coulter, Brea, CA). Isocratic analyses were conducted on a Dionex IonPac AS-20 anion exchange column with a Dionex AG-20 guard column (Dionex, Sunnyvale, CA), eluting with a 50 mM NaOH solution at 1.3 mL/min.

ICP-MS. Determination of the concentrations of Bi and Te impurities in the isolated 211At product was carried out using an Agilent 7900 ICP-MS and the associated Mass Hunter Workstation software package. High purity HNO3 and ultrapure H2O (≥17.8 MΩ·cm) were used to prepare ICP-MS samples. Ten μL aliquots of 211At solutions were diluted 1,000 times in 5% HNO3. Samples containing antibodies were heated for 1 h in concentrated high purity HNO3 at 99 °C before diluting to 5% HNO3. A 100-ppb solution of Co, Y, Ce, and Tl was used as internal standard (ISTD). The ISTD was aspirated by a separate line and mixed with samples or standard solutions before the nebulizer using an online ISTD addition connector.