|Home | About | Journals | Submit | Contact Us | Français|
A simple column chromatographic method was developed to isolate 77As (94 ± 6% (EtOH/HCl); 74 ± 11 (MeOH)) from germanium for potential use in radioimmunotherapy. The separation of arsenic from germanium was based on their relative affinities for different chromatographic materials in aqueous and organic environments. Using an organic or mixed mobile phase, germanium was selectively retained on a silica gel column as germanate, while arsenic was eluted from the column as arsenate. Subsequently, enriched 76Ge (98 ± 2) was recovered for reuse by elution with aqueous solution (neutral to basic). Greater than 98% radiolabeling yield of a 77As-trithiol was observed from methanol separated [77As]arsenate .
Radionuclides are widely used in nuclear medicine and represent powerful tools in both diagnosis and therapy of disease [1-3]. Reactor-produced radionuclides are often of low specific activity because they originate from (n,γ) reactions on the same element; for example, 153Sm is produced by the reaction 152Sm(n,γ)153Sm with the product not separable from the target, and only 1-2% of the atoms are converted to 153Sm. High specific activity radionuclides can be made indirectly in a reactor by producing the parent radionuclide via the (n,γ) reaction, which then beta (β−) decays to the radionuclide of interest (a different element) and that can be chemically separated from the target material. One such radionuclide is 77As (t1/2 = 38.8 h, β−, 100%, Eβ max = 0.683 MeV), which is available from the decay of 77Ge (t1/2 = 11.30 h, β−, 100%, Eβ max = 2.7 MeV); 77Ge is produced via the 76Ge(n, γ)77Ge reaction in a reactor, and then β− decays to 77As. Arsenic-77 is a beta-emitting radionuclide of interest for therapeutic applications as a matched pair to the diagnostic, positron-emitting radionuclide 72As (t1/2 = 26.0 h, β+, 88%, EC, 12%). Together these two radioisotopes provide an excellent diagnostic/therapeutic pair. The relatively long half-lives of both 72As and 77As make them amenable for radiolabeling antibodies and proteins, which require more time (up to several days) to localize to their targeted tissues, thus enabling a wider range of applications . The chemistry of As is versatile allowing for the development of new potential radiopharmaceuticals. However, As radionuclides have limited availability, which is hindered by issues with the separation methods available to produce a high purity product.
Many methods to isolate As from Ge have been reported including solvent extraction [5, 6], distillation [7, 8], ion exchange [8-10], solid phase extraction [11-13], and thin layer chromatography . The solvent extraction methods reported often involve several labor intensive steps, including back extractions, and are not readily adapted for high activity separations requiring remote handling. Distillation methods require heating, some to high temperature (1105 °C ), which increases sample processing time and may require further radioarsenic purification . Chromatographic methods involving ion exchange or solid phase adsorption provide an efficient, robust purification method more readily adapted to automated, remote handling required with high levels of radioactivity. Many of the current literature methods involve the use of strong acids (e.g., >8 M HNO3, HF, HCl) in the separation process, which are undesirable for radionuclides that will be used for synthesizing radiopharmaceuticals [8-12]. Chakravarty  reported a two-column separation for high purity radioarsenic from a GeO2 target using nanocrystalline titania. This method requires the synthesis and strict quality control of the nanocrystalline titania, which is not commercially available. We report a simple, single column Ge/As separation method under mild conditions in an effort to make high purity, high specific activity 77As more readily available. The method minimizes handling of the target material, reduction in operator dose, and use of mild reagents throughout the separation process.
All reagents and solvents were purchased from Alfa Aesar (Ward Hill, MA) and Fisher Scientific (Pittsburg, PA). Enriched 76GeO2 (96.2% or 98.5% enrichment) was purchased from Trace Sciences International (Richmond Hill, ON, Canada). Analytical grade silica gel, SilicAR, was purchased from Mallinckrodt (St. Louis, MO). Commercial 60Å silica gel was purchased from Acros Organics (Geel, Belgium). Acidic alumina was purchased from Fisher Scientific. Zirconium oxide resin (HZO-01) and poly-prep columns (0.8 cm ID; 10 mL reservoir) were purchased from Bio-Rad Corporation (Hercules, CA). All materials were used as received without further purification. All water used was purified on-site (deionized water from a Millipore system to greater than 18 MΩcm).
Caution! Germanium-77 and 76,77As are radioactive and all work involving these radionuclides was carried out in approved laboratories following the appropriate radiation safety procedures. Germanium-77, used as a Ge tracer in column and batch studies, was produced at the University of Missouri Research Reactor (MURR) via the 76Ge(n, γ)77Ge reaction of a quartz encapsulated 76GeO2 target (3-5 mg). High specific activity 77As (370-555 MBq (10-15 mCi) was available from the decay of 77Ge. Carrier-added 76As was produced at MURR via the 75As(n, γ)76As reaction from natAs2O3 (natAs is 100% 75As; 5-10 mg). Table 1 lists the various radionuclides and their nuclear properties.
The irradiated 76GeO2 target was dissolved in 450 μL of 1 M NaOH (0.1 M NaOH is sufficient) with gentle heat (35-45° C) and stirring over 20 minutes and then treated further, as described below, for either batch/column studies with methanolic eluent or for column studies with ethanolic eluents.
Following dissolution, the target was acidified to pH 4-5 with 1 M HCl (0.1 M HCl for 0.1 M NaOH dissolution), and then 100 μL of 30% H2O2 was added to ensure the sample was fully oxidized; the solution was then heated at 60° C to destroy the excess hydrogen peroxide. This solution was used as a stock solution for all batch studies and column chromatography studies using methanol and water as eluents.
Following target dissolution, the pH of the target was acidified to approximately pH 1-2 instead of pH 4-5 and then 100 μL of 30% H2O2 was added to ensure the sample was fully oxidized; the solution was then heated at 60° C to destroy the excess hydrogen peroxide. A 100 μL aliquot of this solution was diluted to a total volume of 500 μL using 0.01 M HCl to maintain a pH of 2. This diluted solution was used for all of the column chromatographic studies using an eluent containing ethanol.
The irradiated natAs2O3 sample was directly dissolved in 500 μL of 1 M NaHCO3.
Radiochemical assays for 76,77As and 77Ge related to the batch studies and methanol-based column chromatographic studies were determined by γ-ray spectroscopy using a Canberra Model GC2018S HPGe detector system with a relative efficiency of 20% at 1.33 MeV. Spectral analyses were performed using a Canberra Model 9660 analyzer. All samples were counted for 900 seconds and decay corrected to the time correlating to either the end of contact time for the batch studies or the first column wash for the column studies. The detector dead time was less than 10% for all samples.
Radiochemical assays for 77As and 77Ge related to the ethanol and mixed-solvent column chromatographic studies were determined by γ-ray spectroscopy on an Ortec Model GEM20-70 HPGe detector system with a relative efficiency of 23% at 1.33 MeV. Spectral analyses were performed using a Canberra DSA LX analyzer. All samples were counted for 600 seconds and decay corrected to the time correlating to count time of the original Ge/As solution. The detector dead time was less than 10% for all samples.
Several media were evaluated in room temperature batch studies for their ability to separate Ge and As following a previously published method . The removal of Ge and As from varying pH (1-13) solutions was measured by mixing 50 mg of each substrate (silica, acidic alumina, and zirconium oxide) with 1.5 mL of a 77Ge- or 76As-spiked solution (154 kBq (4.17 μCi) and 81 kBq (2.19 μCi) per sample, respectively); each solution was adjusted to the desired pH using either 0.1 M HCl or NaOH, 1-2 μL of the spiked solution was added, and the pH retested. The solution pH was determined using pH strips due to the radioactivity present in samples. The liquid-solid system was mixed by vortex for 2 minutes and immediately centrifuged for 2 minutes at 7500 rpm. Two 500 μL aliquots of the contacted solution (As) were transferred into clean 1 mL HDPE vials. Additionally, a 500 μL aliquot of the original uncontacted solution (A0) was transferred to a clean 1 mL HDPE vial as the standard. The activity in each vial was determined by γ-ray spectroscopy and the distribution ratio, Kd, calculated by the following formulation:
A0 is the original Ge or As activity in the aqueous solution. As is the Ge or As activity remaining in solution following contact with the resin. (A0 – As) is the amount of activity adsorbed by the resin. The volume (V) of the aqueous solution is measured in milliliters and the mass (m) is measured in grams resulting in a distribution ratio unit of mL/g. The experimental errors are reported as a function of the standard deviation.
A simplified method using only the silica gel column was evaluated. A small, preparative column packed with a 1 mL bed volume of silica gel was conditioned with methanol (~10 mL). A 50 μL aliquot of the 77Ge/ 77As stock solution of known activities was added to the top of the column followed by twenty 1-mL methanol elutions. The column was then stripped with three 1-mL DI H2O and four 1-mL 0.01 M NaOH elutions (Table 2). A 500 μL aliquot of each fraction eluted was counted by γ-ray spectroscopy.
Various ratios (Table 3) of HCl:ethanol were evaluated for arsenic elution from the column while maintaining germanium retention. A small, preparative column was wet-packed with a 1 mL bed volume of silica gel (~510 mg) slurried with ethanol and was conditioned with the mobile phase of interest (~10 bed volumes). A 50 μL aliquot of the 77Ge/77As stock solution, which had been oxidized and pH adjusted to 1.5-2.5, was added to the top of the column followed by elution with the mobile phase of interest (4 × 1 mL). All fractions were assayed by γ-ray spectroscopy along with an aliquot of the load sample.
The synthesis of this [77As]-labeled trithiol from methanol separated [77As]arsenate has been reported . Briefly, aqueous ammonium mercaptoacetate (25 mM), and 2-ethyl-2-(mercaptomethyl)propane-1,3-dithiol (trithiol ligand; 15 μM in absolute ethanol) were combined in a reaction vial. To this solution was added an aliquot of the [77As]arsenate in DI water (the methanol had been removed at 50 °C and then DI water added). The reaction mixture was heated at 60 °C for 30 min. RP-HPLC analysis as previously reported gave >98% radiochemical yield.
Radiochemical separations can be challenging because of the very large mass difference (often 106 or greater) between the desired radionuclide produced and the irradiated target. Additionally, the target material is often recovered because it is highly enriched in the particular isotope that generated the desired radionuclide, and therefore quite precious (costly). A separation was developed to isolate no carrier added (nca) 77As and recover the highly enriched (96-99%) 76Ge target for reuse. Of the methods reported for the separation of nca radioarsenic from Ge targets, most include a step to ensure that the radioarsenic is in oxidation state +3. Since nca As(III) oxidizes to nca As(V) within minutes, a separation was developed to isolate nca As(V). Maki and Murakami  reported a silica gel thin-layer chromatographic (TLC) method to separate 77As(III) and 77As(V) from its GeO2 target with methanol and HCl as the eluent. Caletka and Kotas  reported that at acid concentrations > 8 M, 68Ge was retained on silica gel columns while 68Ga and other radionuclides (including 77As) were eluted. Bokhari et al.  reported a column chromatographic method using hydrous zirconium oxide (HZO) to selectively adsorb Ge under basic conditions (0.1 M NaOH) while ~90% of the radioarsenic passed through the HZO, most likely as As(V). The prior studies suggested that silica gel might allow a simple, straightforward separation for 77As from its enriched 76Ge target. With the aim of incorporating the 77As into bifunctional chelates following purification, > 8M acid was not considered an optimal separation medium. Thus, several chromatographic media were investigated with eluents over the entire pH range and with methanol, and finally with ethanol to make the system more biocompatible. An additional consideration in radiochemical separations is minimization of the eluent volume required to isolate the desired radionuclide so that subsequent handling prior to use is minimized.
Inorganic media such as acidic alumina, silica gel and hydrous zirconium oxide (HZO-01) are often considered for use in radiochemical separations because they are more resistant to radiation damage than organic-based resins. These media were evaluated for their ability to retain arsenate and germanate with equilibrium distribution ratios acquired as a function of pH (1 – 13). Distribution ratios were determined for As and Ge at each pH and in methanol using the batch method, and are reported as mean values and their calculated experimental errors. Separation factors were calculated for each condition from the distribution ratios. For acidic alumina, a maximum 77Ge/76As separation factor of 9 ± 1 was observed at pH 11; this value is somewhat low and would require a large column and elution volumes or risk tailing of one species into the other (Figure 1). For silica gel, a substantial 77Ge/76As separation factor of 82 ± 16 was observed using methanol as the eluent, which could be used to recover the bulk of the Ge target material while eluting arsenate with methanol (Figure 2). Germanate adsorbed strongly to the silica gel in methanol and had no retention under any aqueous conditions, while arsenate showed little affinity under the conditions evaluated. A usable 76As/77Ge separation factor of 19 ± 8 was observed using HZO-01 and methanol as the eluent, which could allow for the isolation of arsenate from Ge (Figure 3). Arsenate was retained by HZO-01 in methanol and had little affinity in either very high or very low pH (pH ≤ 1 or ≥ 11) aqueous conditions, while germanate had little affinity in methanol and greater affinity under high pH aqueous conditions.
To apply the distribution coefficients in a practical application, a tandem column assembly using silica gel and HZO-01 resin was evaluated. The distribution coefficients indicated the silica gel should trap the bulk Ge, while the HZO-01 should trap the As when methanol was the eluent; any Ge breakthrough from the silica gel should pass through the HZO-01 leaving a high purity 77As product behind. Additionally, any residual Ge on the HZO-01 column would be trapped using a pH 11 solution, while eluting the As based on their relative distribution coefficients (430 ± 20 and 39± 1, respectively). Arsenic adsorbed more strongly to the HZO-01 resin than the distribution coefficients suggested. No As elution was observed using a pH 11 solution and required a pH 13 solution to elute the trapped arsenate. Even under these highly basic conditions, a large portion of the As remained on the column. Although the As elution sample was free of detectable Ge, significant loss of As was observed using the tandem column assembly.
The distribution coefficients and high separation factor suggested a facile separation of As and Ge using methanol. A reasonable recovery of the 77As (74 ± 11%) eluted in 4 mL of methanol with very little, if any, 77Ge breakthrough (0.09% ± 0.08) observed. A newly opened bottle of methanol resulted in lower recovery (~60% vs ~80%) likely due to the water content (methanol is quite hygroscopic). Efficient recovery of Ge is critical due to the high cost associated with the enriched 76GeO2 starting material, which is recycled for reuse. Recovery of greater than 86% of the 77Ge in 5 mL of water mitigates this concern; 98% recovery is observed with 0.01 M NaOH following the water elution (Table 2). The elution profile is shown in Figure 4. The large loss of 77As likely resulted from the precipitation of Ge during pH adjustment with HCl. This has been visibly observed at the macroscopic scale and quite possibly occurs at the microscale, though not visibly. As the Ge precipitates, it traps the tracer As in the process; this was further indicated by the co-elution of a larger portion of As with the Ge when the eluent was changed to water.
If As was simply lost to silica adsorption, the recovery on a second column should have been approximately 1/3, as observed with the first column. To further investigate this loss of As, the arsenic-containing fractions were passed through a second silica gel column. Interestingly, greater than 80% of the As loaded onto the second silica gel column was recovered, while germanium was reduced to non-detectable levels. This result further supports loss related to Ge precipitation.
The 76GeO2 targets were dissolved in base (NaOH) and acidified prior to loading onto the silica gel columns. To evaluate the effect of acid used for acidification to pH 4-5, HCl, HNO3, and H3PO4 were compared. The results are shown in Figures 5--77. Each data set was normalized to the percentage of recovered material to eliminate variations resulting from overall recovery with each mobile phase. The HCl-adjusted sample exhibited very little Ge breakthrough in the first 10 mL (0.6 %). However a sizable percentage (12.7%) of the 77As co-eluted with the 77Ge, suggesting it was trapped within a Ge precipitate. Use of HNO3 resulted in less Ge breakthrough (0.04% in the first 10 mL), but a larger portion of 77As was trapped within the Ge precipitate (25.7%), suggesting more of a Ge precipitate with HNO3. Finally, H3PO4 was unique compared to use of the other two acids as a much higher percentage of 77As was recovered quickly (67.6% in 1 mL) but Ge breakthrough was more substantial (0.3% with the first mL and 6.1% in the second mL). Additionally, very little 77As co-eluted with Ge (0.4%) suggesting less Ge precipitation occurred with this system. Overall HCl outperformed HNO3 and H3PO4 since a large portion of the Ge was successfully retained on the column without sacrificing As recovery.
To make the process biocompatible, ethanol was evaluated as a substitute for methanol. Arsenic was more strongly retained by the silica gel with ethanol as the eluent. Studies varying the loading conditions were performed and it was observed that acidifying the loading solution resulted in lower breakthrough of germanium. The pH of the loading solution was lowered from 4-5 to 1-2, which resulted in a recovery of 7 ± 2% of the 77As eluting in ethanol with no detectable breakthrough of Ge. Additional studies were performed with HCl incorporated directly into the ethanol or with a mixed aqueous/ethanol mobile phase. In the mixed aqueous HCl/ethanol mobile phases the addition of water had a substantial impact on the recovery of As. Adding too much water resulted in breakthrough of Ge. This effect was counteracted by using a higher acid concentration in lower volumes. Similar recovery of 77As was observed when concentrated HCl was directly added to ethanol, 91 ± 4% (Tables 3 and and44; note that the two different lots of enriched 76GeO2 obtained from Trace Sciences International unexpectedly behaved quite differently). The HCl:ethanol mobile phase showed excellent separation of 77As and recovery of Ge especially when water content was minimized.
The utility of the [77As]arsenate separated by this method is demonstrated by radiolabeling studies. A trithiol ligand, namely 2-ethyl-2-(mercaptomethyl)propane-1,3-dithiol, was carried out as previously reported  in >98% yield of the 77As-trithiol product (Figure 8).
The availability of high purity radioarsenic, such as 72As and 77As, would provide a longer-lived, non-metallic radionuclide pair for radioimmunoimaging and radioimmunotherapy. The versatile chemistry of arsenic would enable the development of potential new radiopharmaceuticals. A quick, robust method for the separation of nca 77As from its GeO2 target material was developed to increase the availability of high purity 77As, with the enriched 76Ge starting material easily recovered for reuse.
Distribution coefficient studies showed germanium has a strong affinity for silica gel using methanol as the mobile phase while arsenic had limited affinity. Isolation of 74 MBq (2 mCi) of 77As was consistently eluted free of Ge when 210 MBq (5.7 mCi) of 77As/76,77Ge was loaded onto 520 mg of silica gel. Fine tuning the column chromatographic method based on the observed distribution coefficients resulted in the optimal separation method using a single silica gel column and a mixed mobile phase with a 1:10 ratio of 0.1 M HCl and ethanol; this system recovered 94 ± 6% of the arsenic in a small volume with no detectible germanium. Likewise a high recovery rate (91 ± 4%) was observed using a single silica gel column and a mixed mobile phase with a 0.01 M HCl solution prepared directly in ethanol using concentrated HCl. Germanium was easily recovered from the column by elution with water. With the 1:10 ratio of 0.1 M HCl:EtOH eluent using 520 mg of silica gel, 74 MBq (2.0 mCi) of 77As is consistently eluted free of Ge when 89MBq (2.4 mCi) of 77As/76,77Ge was loaded. This method gives a higher percentage recovery of 77As than does the methanol eluent (83% versus 74%).
Separation of large mass targets containing higher activities is required to demonstrate the utility of this method for routine production. If scale-up is successful, the separated 77As would be available for use as a tracer for medical, toxicological, and environmental studies, which is the ultimate goal of this project.
The authors acknowledge support from the Department of Energy, Office of Basic Sciences, Isotope Research Program under grants DE-SC0003851 and DE-SC0010283. Additional trainee support was provided from the National Science Foundation under IGERT award DGE-0965983 (M.D. Gott) and the National Institutes of Health under NIBIB Training Grant 5 T32-EB004822 (A.J. DeGraffenreid).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.