The effect of purification of Ga-68-labeled exendin on in vivo distribution

Background Ga-labeled radiotracers are increasingly used for PET imaging. During the labeling procedure, formation of 68Ga-colloid may occur. Upon i.v. injection, 68Ga-colloid will accumulate rapidly in the liver, spleen, and bone marrow, resulting in reduced target-to-background ratios. In this study, we applied a thin layer chromatography (TLC) method to measure colloid content and we studied the effect of the purification method on the in vivo characteristics of 68Ga-labeled DOTA-exendin-3. DOTA-exendin-3 was labeled with 68Ga, and the colloid content was measured by TLC on silica gel ITLC with two mobile phases. The labeling mixture was purified by gel filtration on a 5-ml G25M column, by reversed-phase high-performance liquid chromatography (RP-HPLC) using a C8 column or by solid phase extraction (SPE) on an HLB cartridge. The in vivo characteristics of the preparations were determined in BALB/c nude mice, and PET images were acquired 1 h p.i. using a microPET scanner. In these studies, unpurified 68Ga-DOTA-exendin-3 and 111In-DOTA-exendin-3 were used as a reference. Results The colloid content of 111In-DOTA-exendin-3 and unpurified, gel filtration, RP-HPLC- and SPE-purified 68Ga-DOTA exendin-3 was <3, 7, 9, <3, and <3 %, respectively. Unpurified 68Ga-DOTA exendin-3 showed high liver and spleen uptake. Gel filtration partly removed 68Ga-colloid from the preparation, resulting in moderate liver and spleen SPE-purified 68Ga-DOTA exendin-3 showed very low liver and spleen uptake, that was similar to that of RP-HPLC purified 68Ga-DOTA exendin-3. Conclusions We showed that the colloid content can be measured by TLC and that solid phase extraction and HPLC completely remove 68Ga-colloid from 68Ga-labeled tracer preparations, resulting in very low liver and spleen uptake. This study clearly shows the importance of removal of 68Ga-colloid from preparations.

Background 68 Ga-labeled peptides are increasingly used for positron emission tomography (PET), since 68 Ga is a readily available PET radionuclide. Because 68 Ga is a generatorproduced positron emitter, it is widely available and relatively cheap. PET imaging is advantageous over conventional scintigraphy and SPECT because of its excellent sensitivity in combination with its superior spatial resolution. A recent study in patients with neuro-endocrine tumors (NET) showed that PET imaging with the somatostatin analog 68 Ga-DOTA-TOC was more sensitive for detecting NET lesions than conventional somatostatin receptor scintigraphy with 111 In-octreotide [1]. Moreover, 68 Ga-labeling of peptides conjugated with a chelator (e.g., DOTA, NOTA) is a fast and efficient one-step reaction. Because of all these characteristics, there is an increasing interest in the application of 68 Ga-labeled peptides.
It has been shown that for efficient receptor targeting, low peptide doses should be administered, since higher peptide doses could lead to receptor saturation and reduced uptake in the target tissue [2][3][4][5], especially in preclinical imaging studies. In rodents, relatively high activity doses (3)(4)(5)(6)(7)(8)(9)(10) have to be administered to acquire PET images with adequate image quality. Therefore, 68 Ga-labeled peptides with a high specific activity should be produced to administer high activity doses at a low amount of peptide. However, when producing 68 Ga-labeled compounds with a high specific activity, the formation of insoluble 68 Ga-species, such as 68 Ga(OH) 3 may occur. These insoluble 68 Ga-species, generally referred to as " 68 Ga-colloid" will accumulate in the liver, spleen, and bone marrow. Indeed, previous studies showed enhanced uptake in liver and spleen of 68 Ga-labeled tracers as compared to the 111 In-labeled compounds [3,6,7]. This enhanced tracer uptake in liver and spleen might result in decreased target-tobackground ratios. We previously showed that insoluble 68 Ga-species can be removed from the labeling reaction of 68 Ga-labeled DOTA-exendin-3, a tracer targeting the glucagon-like peptide-1 receptor (GLP-1R), using (preparative) reversed-phase high-performance liquid chromatography (RP-HPLC) [3]. However, this purification method is time consuming and the solution of purified 68 Ga-labeled peptide is diluted, making postpurification concentration necessary. Due to the short half-life of 68 Ga (68 min), this method is not convenient in clinical practice. Moreover, purification with RP-HPLC requires expensive equipment. Solid phase extraction is an alternative method for purification of radiolabeled compounds and is a fast, simple, and cheap purification method that is now routinely used for 68 Ga-tracer purification [8,9].
In this study, we examined the effect of the purification method on the in vivo characteristics of 68 Ga-DOTAexendin-3 in BALB/c nude mice. 68 Ga-DOTA-exendin-3 was purified by RP-HPLC, gel filtration, or solid phase extraction. Unpurified 68 Ga-DOTA-exendin-3 and 111 In-DOTA-exendin-3 were used as a reference in this study. In addition, we describe a quality control method, based on instant thin layer chromatography (ITLC), to determine the colloid content of the 68 Ga-labeled tracer [10].
Radiolabeling DOTA-exendin-3 was labeled with 68 Ga and 111 In as previously described [3]. Briefly, 120 MBq 68 Ga in 1000 μl Ultrapure 0.1 N HCl was added to 10 μg DOTA-exendin-3 in 120 μl 2.5 M HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, Sigma Aldrich, St. Louis, MO, USA). After 20-min incubation at 95°C, EDTA was added to a final concentration of 5 mM and the reaction mixture was incubated at room temperature for another 5 min. Subsequently, 10 % Tween-80 (Sigma Aldrich, St. Louis, MO, USA) was added to a final concentration of 0.1 % to prevent sticking of the radiolabeled peptide to the vessel wall and quality control was performed as described below.

Quality control
Quality control was performed using RP-HPLC on a C 18 reversed-phase column (Zorbax Rx-C18; 4.6 mm × 25 cm; Agilent Technologies, Palo Alto, CA, USA) and instant thin layer chromatography (ITLC). The column was eluted with mixture of water containing 0.1 % trifluoroacetic acid (TFA) and acetonitrile with a linear gradient from 3 to 100 % acetonitrile in 10 min (flow rate 1 ml/min). ITLC was performed on silica gel ITLC Validation of the quality control by ITLC for the detection of 68 Ga-colloid 68 Ga-colloid was prepared by adding 1250 μl 2.5 M HEPES to 500 μl 68 GaCl 3 in 0.1 M HCl. The final pH of this mixture was approximately 6. The mixture was incubated at 95°C for 15 min, and EDTA was added to a final concentration of 5 mM. The amount of 68 Ga-colloid was determined by TLC with 0.1 M EDTA in 0.25 M NH 4 Ac, pH 5.5 as a mobile phase (R f = 0: 68 Gacolloid, R f = 1 68 Ga-EDTA) and 1.25 M NH 4 Ac, pH 5.5: DMF (1:1) as a mobile phase (R f = 0: 68 Ga-colloid, R f = 1: 68 Ga-EDTA). The reaction mixture was applied on a disposable PD-10 desalting column, containing Sephadex TM G-25 medium (GE Life Sciences, Diegem, Belgium) and was eluted with 6 ml phosphate-buffered saline (PBS) containing 5 mM EDTA. The fraction containing the majority of the radioactivity was collected (from 3-4 ml), representing 68 Ga-colloid, and quality control was performed by ITLC as described above. 68 Ga-DOTA-exendin-3 was purified by RP-HPLC as described below, and various amounts of the 68 Ga-colloid were added to obtain final 68 Ga-colloid concentrations of 1, 2, 3, 4, 5, 10, 20, 40, and 80 % (n = 4). The amount of 68 Ga-colloid was determined by ITLC with 1.25 M NH 4 Ac, pH 5.5: DMF (1:1) as the mobile phase (R f = 0: 68 Ga-colloid, R f = 1: 68 Ga-DOTA-exendin-3 and 68 Ga-EDTA). RP-HPLC-purified 68 Ga-DOTA-exendin-3, 68 Ga-EDTA, and 68 Ga-colloid were used as controls. ITLC strips were exposed to an imaging plate (Fuji Film BAS-SR 2025, Raytest, Straubenhardt, Germany) for 1 min. Images were acquired with a radioluminography laser imager (Fuji Film BAS 1800 II system, Raytest, Straubenhardt, Germany) and analyzed with Aida Image Analyzer software (Raytest). Correlation between the measured 68 Ga-colloid fraction and the added 68 Ga-colloid content was determined by linear regression using GraphPad Prism 5. The detection limit of the ITLC method was defined by the Y-intercept as determined by linear regression analysis.

Purification of 68 Ga-DOTA-exendin-3 by gel filtration
Purification by gel filtration was performed on disposable PD-10 desalting columns, containing Sephadex TM G-25 medium (GE Life Sciences, Diegem, Belgium). The column was preconditioned by eluting with 10 ml PBS containing 0.5 % (v/w) BSA, and the reaction mixture was loaded onto the column. The column was eluted with PBS-BSA (0.5 %), and 0.5 ml fractions were collected. The majority of the radioactivity, representing 68 Ga-DOTAexendin-3, was collected in fractions 5 and 6 (2-3 ml). The radiochemical purity of the tracer in the fractions was analyzed by RP-HPLC and ITLC as described above.
Purification of 68 Ga-DOTA-exendin-3 by solid phase extraction Solid phase extraction was performed using a hydrophiliclipophilic balance (HLB) reversed-phase sorbent cartridge (Waters Oasis © , Milford, MA, USA). The cartridge was activated by elution with 1 ml ethanol, the residual ethanol was removed with 1 ml water (Versol, Lyon, France), and the column was conditioned with 1 ml 0.1 N HCl:2.5 M HEPES (8:1, similar to the reaction mixture). The reaction mixture was loaded, and 68 Ga-EDTA was washed from the column with 2 ml 0.1 N HCl:2.5 M HEPES (8:1). After removal of the residual HCl-HEPES mixture with 1 ml water, 68 Ga-DOTA-exendin-3 was eluted with 200 μl ethanol. The radiochemical purity of the tracer in the fractions was analyzed by RP-HPLC and ITLC as described above. Before injection into mice, the eluate containing 68 Ga-DOTA-exendin-3 was diluted with PBS containing 0.5 % BSA to a final ethanol concentration of less than 10 %.

Biodistribution studies
Animal experiments were performed after approval of the local ethical committee (RUDEC) for animal experiments. The biodistribution of unpurified, gel filtration-, RP-HPLC-, and SPE-purified 68 Ga-DOTA-exendin-3 was determined in BALB/c nude mice. Mice (n = 5) were injected intravenously with 3 MBq 68 Ga-labeled exendin-3 at a peptide dose of 0.3 μg (60 pmol). As a control, another group of mice received 370 kBq 111 In-DTPA-exendin-3 (60 pmol). The mice were euthanized 1 h post-injection by CO 2 /O 2 suffocation, a blood sample was taken, and samples of relevant tissues were dissected, weighed, and counted.

Statistical analysis
All mean values are expressed as mean ± standard deviation (SD). Statistical analysis was performed using unpaired two-tailed t test using GraphPad Prism (version 5). In order to determine whether there was a overall difference in blood and kidney accumulation of the various preparations, a one-way ANOVA was performed. The level of significance was set at p < 0.05.

Validation of TLC method
The results of the validation of the TLC method to determine the colloid content in preparations of 68 Ga-DOTA-exendin-3 are shown in Fig. 1. The colloid content measured by TLC correlated linearly with the amount of colloid present in the reaction mixtures (R 2 = 0.98). The detection limit of the TLC method was a 68 Ga-colloid content of 3.2 ± 0.9 %, as determined by the Y-intercept of the trend line.

Biodistribution studies
The results of the biodistribution studies are summarized in Fig. 2. 111 In-DOTA-exendin-3 showed very low accumulation in the liver and spleen: 0.7 ± 0.1 and 0.3 ± 0.1 %ID/g, respectively, representing the liver and spleen uptake of a preparation without colloid. High uptake in the liver and spleen (6.1 ± 1.0 and 4.5 ± 0.7 %ID, respectively) was observed when unpurified 68 Ga-DOTA-exendin-3 was injected. Gel filtration partially removed 68 Gahydroxide, resulting in intermediate uptake in the liver and spleen: 3.0 ± 0.3 and 1.4 ± 0.3 %ID/g, respectively). Uptake in the liver and spleen of SPE-purified (0.8 ± 0.0 %ID/g and 0.5 ± 0.1 %ID/g, respectively) and RP-HPLCpurified tracer (0.6 ± 0.1 and 0.4 ± 0.1 %ID/g, respectively) was very low and similar to liver and spleen uptake of 111 In-DOTA-exendin-3, indicating very efficient removal of 68 Ga-colloid with these purification methods.
The uptake in the pancreas was similar for all compounds, except of RP-HPLC purified 68 Ga-DOTA-exendin-3, that had a significantly higher pancreatic uptake (10.2 ± 2.5 %ID/g, p < 0.05). A significant difference in radioactivity concentration in the blood and kidneys was found between all groups (p < 0.0001 and p = 0.0035 for blood and kidneys, respectively), probably caused by differences in the colloid content of the various preparations.

Discussion
During the radiolabeling procedure of DOTA-conjugated compounds with 68 Ga, in most cases, insoluble colloidal 68 Ga species are formed, especially when compounds are labeled at a high specific activity. This insoluble 68 Ga-colloid results in enhanced accumulation in the liver and spleen when injected in laboratory animals or patients, resulting in reduced image quality. Therefore, prevention of formation or removal of 68 Ga-colloid is required. We evaluated three purification methods for the removal of 68 Ga-colloid: solid phase extraction (SPE), preparative HPLC, and gel filtration. Solid phase extraction was a fast and simple method that effectively removed 68 Gacolloid from a labeling mixture of 68 Ga-DOTA-exendin-3, resulting in negligible liver and spleen accumulation similar to that of 111 In-DOTA-exendin-3. The major advantage of this technique is that purification of the labeling mixture can be performed within 5 min. Purification by preparative RP-HPLC resulted in similar spleen and liver accumulation. However, this technique is more time-consuming and less suited for purification of 68 Ga-labeled tracers, due to the short half-life of 68 Ga. Remarkably, the uptake of RP-HLPCpurified 68 Ga-DOTA-exendin-3 in the pancreas was significantly higher. This is probably due to a slightly lower peptide dose, which results in higher pancreatic uptake as previously described [3]. Gel filtration removed 68 Ga-colloid less efficiently, resulting in higher Table 1 68 Ga-colloid content determined by ITLC, liver, and spleen uptake of unpurified, gel filtration, RP-HPLC-and SPE-purified liver and spleen uptake than SPE and RP-HPLC purification. It should be noted that gel filtration could be more appropriate for the purification of larger compounds such as proteins. The higher molecular weight of proteins combined with the low molecular weight of 68 Ga-colloid could result in better separation. However, the feasibility to separate 68 Ga-colloid from high molecular weight compounds by gel filtration should be evaluated. High specific activities of 68 Ga-labeled tracers are required to administer high activity doses in combination with low peptide doses. It was shown that high peptide doses lead to reduced uptake in the target tissues due to (partly) saturation of the receptors [2][3][4][5], which is of particular importance in preclinical imaging. Moreover, high peptide doses can lead to (toxic) side effects, especially when biologically active compounds are used as the ligand. The need for a high specific activity generally enhances 68 Ga-colloid formation, since incorporation of 68 Ga present in the labeling mixture in the presence of low amounts of metal binding ligand might be incomplete. Since 68 Ga(OH) 3 could be formed at pH 3 in the absence of chelating agents (such as DOTA, DTPA, and  [11], 68 Ga(OH) 3 is formed in a labeling mixture at a pH between 3.5 and 4 when 68 Ga-incorporation is incomplete. Interestingly, in the pH range generally used for 68 Ga-labeling procedures (pH 3.5-4), Ga(III) is predominantly present as polymeric Ga(III) hydroxides, while Ga(OH) 3 is the predominant form at pH 7 [12]. Furthermore, at higher pH, the formation of insoluble GaO(OH) can occur [13]. GaO(OH) is particularly formed at high temperature [14], which is of importance since most 68 Galabeling procedures are carried out at high temperatures (80-100°C). In contrast to Ga(III) hydroxides, GaO(OH) is only slowly redissolved even at low pH [14].
Wild et al. showed a similar biodistribution of 68 Ga-labeled DOTA-exendin-4 [15] as 68 Ga-DOTA-exendin-3  in our study. In the former study, no purification was performed of the 68 Ga-labeled compound, and this lead to slightly higher uptake in the spleen as compared to the SPE and RP-HPLC purified tracer used in our study. However, the splenic uptake of 68 Ga-DOTA-exendin-4 was lower than that of the unpurified 68 Ga-DOTA-exendin-3 in our study. This lower uptake in the spleen of 68 Ga-DOTA-exendin-4 reported by Wild et al. is probably due to a lower 68 Ga-colloid content in the labeling mixture. The lower 68 Ga-colloid content might be due to a lower specific activity of 68 Ga-DOTA-exendin-4 and a different labeling protocol. In the study performed by Wild et al., the 68 Ga-eluate was purified and the labeling was performed in a microwave for 5 min. The shorter labeling time in combination with the lower specific activity may lead to faster and more complete incorporation of 68 Ga, reducing the risk of 68 Ga-colloid formation [11]. Several other labeling methods [16,17] and chelators [18][19][20][21] for labeling compounds with 68 Ga are described with more efficient labeling yields and faster labeling kinetics. These new strategies for 68 Ga-labeling of compounds might also reduce the risk of 68 Ga-colloid formation.
We validated a TLC method for the detection of 68 Gacolloid in labeling mixtures of 68 Ga-DOTA-exendin-3 and showed a clear linear correlation between added 68 Ga-colloid to the labeling mixtures (free of 68 Ga-colloid) and the measured 68 Ga-colloid content. These results suggest accurate quantification of 68 Ga-colloid in the labeling mixture. Determination of the colloid content is of great importance to calculate the specific activity. Neglecting the presence of 68 Ga-colloid might result in overestimation of the specific activity and thus administration of a higher peptide dose as initially planned. This is especially true for exendin, where very low peptide doses need to be administered to prevent receptor saturation resulting in lower uptake in GLP-1R-positive tissues [3]. The detection limit of this method is approximately 3 % making detecting of very low concentrations of colloid difficult. The rather low sensitivity is due to tailing of the radiolabeling peptide hampering the clear delineation between the peak representing the 68 Ga-colloid and the peak representing 68 Ga-DOTAexendin-3. The sensitivity of this method for the determination of 68 Ga-colloid for other peptides might be different due to different migratory characteristics of the 68 Ga-peptide on this TLC system compared to exendin.
Enhanced liver and spleen uptake will hamper the detection of lesions in the upper abdomen. Radiolabeled exendin was successfully used for detection of insulinomas [22] and could potentially be used for detection of the pancreatic beta cell mass. Enhanced liver and spleen uptake might reduce the sensitivity of these methods. Previous studies explored the feasibility of tumor imaging and atherosclerotic plaques with 68 GaCl 3 in mouse models for pancreatic adenocarcinoma and artherosclerosis, respectively [23,24]. Although the tumor and atherosclerotic plaques could be detected, the PET images suffered from high background signal in the liver and lungs as well as high concentration in the blood. The binding of 68 Ga to transferrin (prolonging the circulation time) and the formation of 68 Ga-colloid could explain the enhanced background and is in line with our study.

Conclusions
Solid phase extraction using a HLB cartridge is a fast and simple method to remove 68 Ga-colloid. The uptake in the liver and spleen of the SPE-purified product was similar to that of 111 In-DOTA-exendin-3 or HPLCpurified 68 Ga-DOTA-exendin-3, indicating sufficient removal of insoluble 68 Ga-colloid. Gel filtration only partly removed 68 Ga-colloid species and is not suitable for purification of 68 Ga-labeled exendin-3 and most likely other peptides. Moreover, SPE cartridges are routinely integrated in GMP-grade synthesis modules for 68 Ga-labeling of peptide, and therefore, this method can be used for purification of 68 Ga-labeled compounds for human use. Possibly, gel filtration might be suitable for larger compounds (e.g., antibodies, proteins), since the performance of size exclusion chromatography is expected to be superior for larger compounds. However, the feasibility for the removal of 68 Ga-colloid by gel filtration of proteins and compounds where solid phase extraction is not possible should be verified. The studies presented here show the importance of complete removal of 68 Gacolloid before in vivo use.