Introduction

The use of 68Ga labeled peptide in positron emission tomography (PET and PET/CT) to visualize various type of tumors allows better planning of the therapeutic strategy and effective prediction of the treatment outcomes [1]. For example DOTATOC labeled with 68Ga shows high binding affinity for the human somatostatin receptor subtype 2, improving this way tumor imaging capabilities and offering the possibility of low dose imaging, followed by higher dose treatment by DOTATOC labeled with 177Lu or 90Y [2].

However, because of their short half-life (T1/2 = 67.71 min), 68Ga-based radiopharmaceuticals are synthesized and used in-house. Furthermore, the relatively high cost of the generators render this isotope of limited utility in a clinical setting. These circumstances lead to high overall cost and render 68Ga radiopharmaceuticals of limited interest for centralized production and commercial distribution.

Recently there has been steadily growing interest in the medical applications of Sc radioisotopes. The longer half-life of the scandium β+ emitters, 44gSc and 43Sc (T1/2 = 3.97 h and T1/2 = 3.89 h, respectively) compared to 68Ga potentially permits of a centralized production of 44Sc-labeled peptides and their shipment over several hundred kilometers to hospitals with PET centers that do not have a radiopharmacy on site. [3, 4]. In addition, images could be acquired over longer periods. Another advantage in the use of 43,44Sc as diagnostic radioisotopes lies in the other scandium radioisotope, i.e. 47Sc (T1/2 = 3.35 d) which is a promising low-energy β− emitter for targeted radionuclide radiotherapy, and therefore represents an theranostic pair with the β+ emitting 44gSc or 43Sc radioisotopes. The potential of 44Sc/47Sc as a theragnostic pair has been demonstrated in a preclinical pilot study with tumor-bearing mice [5, 6].

Additionally, it has been shown that Sc3+, like Y3+ and Lu3+, forms in solution DOTA complexes with coordination number 8, whereas Ga3+ forms octahedral complexes. Because the difference in the coordination of complexes influences on the lipophilicity of conjugates, Sc-DOTATATE has nearly identical lipophilicity as that of Lu- and Y-DOTATATE, whereas Ga-DOTATATE is more hydrophilic [7]. The difference in the chelates structure determinates the receptor affinity of labeled conjugates. For example, the IC50 value of Ga-DOTATOC for the hsst2 receptor is 2.5 nmol, while that for Y-DOTATOC is 11 nmol [8]. Due to the chemical similarity of Sc3+ to the Lu3+ and Y3+ cations 44Sc-DOTA-bioconjugates will likely demonstrate similar properties in vivo (i.e. receptor affinity, kidney clearance) to the 177Lu- and 90Y-conjugates currently applied in therapy, therefore 44Sc and 43Sc can form theranostic pairs with 177Lu and 90Y.

For receptor radionuclide diagnosis and therapy radiolabeling and molar activity i.e., radionuclide–ligand molar ratio should be as high as possible to minimize receptor occupancy of unlabeled conjugates. Therefore in radiolabeling process the presence of trace amounts of carrier or other cations that form strong complexes with the ligand can significantly reduce the degree of labeling yield and the molar activity of the bioconjugate. In radiopharmaceutical practice, the influence of transition metal contaminants on the labeling of bioconjugates with 177Lu, 90Y and 68Ga has been repeatedly observed since many of the used chelates form strong complexes with a range of metal ions [9]. Particularly, this effect was observed for the cations such as Fe3+, Zn2+and Cu2+. This effect was also observed in our studies on the labeling of bioconjugates with radionuclides 43Sc, 44Sc and 47Sc. While in the case of 68Ga, this problem has been extensively studied [10,11,12], in the case of scandium radionuclides studies have not been conducted. Only in a number of studies, the effect of transition metal cations on 44Sc-DOTA radiobioconjugates stability were studied [13,14,15]. However, due to the kinetic inertness of the Sc-DOTA complexes, both processes are completely different.

This article details investigations into the influence of a variety of metal ions on 44Sc radiolabeling. We have chosen to studies the clinically relevant somatostatin receptor ligand, DOTATATE (DOTA0-Tyr3-octreotate), and tested labeling yield in the presence of varying amounts metal ion impurities. The cation impurities were selected taking into account the possibility of their occurrence in the labeling solution and forming stable complexes with the DOTA ligand, namely Fe2+/3+, Cu2+, Zn2+, Al3+ and Ca2+. If DOTA-functionalized bioconjugates labeled with scandium radionuclides should be used for clinical practice, it will be important to determine the maximum concentration of metal contaminants which would still allow high specific and reproducible labeling. The obtained results can be used to optimize the labelling conditions of clinical radiopharmaceutical production.

Experimental

Materials and methods

All chemicals and solvents were of analytical or pharmaceutical grade unless otherwise specified. Metal salts (ultrapure grade; trace metal content) were obtained from Sigma-Aldrich. DOTATATE (DOTA0-Tyr3-octreotate), was purchased from Polatom (Poland). Since the hydrochloric acid may contain traces of iron ions we decided for its further purification on DistillacidTM BSB-939-IR instrument specially designed for preparation of high purity acids. The distilled HCl contains metallic impurities like Fe2+/3+, Zn2+ in concentrations below 5 ppb. The target was prepared from 100 mg of natural CaCO3 (99.999%, Sigma Aldrich) pressed in graphite.

Irradiation of calcium target

Irradiations of the calcium targets were performed using the GE PETtrace cyclotron at the Radiopharmaceuticals Production and Research Centre put into operation by the Heavy Ion Laboratory, University of Warsaw a few years ago. This cyclotron was recently equipped with an external beam line for solid sample irradiations, also allowing a good cooling conditions for these samples [16]. A 2-h proton irradiation at the energy 16 MeV and 15 μA current were performed. During irradiation process the front side and the back side of the target were cooled.

Separation of 44Sc from calcium target

For separation of 44Sc from calcium target the method recently elaborated by Minegishi et al. [17] in combination with purification on Dowex 50 resin was applied. The separation process is presented in Fig. 1.

Fig. 1
figure 1

Separation scheme of the isolation of 44Sc from the CaCO3 targets

Full size image

The irradiated CaCO3 target was dissolved in 3 mL 1 M HCl solution during 10 min. Then the dissolved target solution was alkalized to pH 10 by aq ammonia solution (25%) In this condition Sc3+ forms Sc(OH)3 which is quantitatively separated from the solution by passed through the Teflon 0.2 µm filter. Subsequently, 5 mL of pure water were passed through the filter to wash out residual Ca2+ and NH4+ cations. 44Sc trapped on the filter was eluted by 0.5 M aqueous HCl (2 mL).

For additional purification of the 44Sc and change pH to appropriate for labeling the eluate from filter was adsorbed on the column filled with DOWEX 50Wx4 cation exchange resin. The elution of 44Sc was performed using 1.5 mL 1 M ammonium acetate aq. solution (pH 4.5). The activity of the eluted fractions was monitored and 0.75 mL fraction with maximum activity of eluate was selected for labeling experiments. The concentration of metallic impurities in labeling solutions was measured by the ICP-MS instrument Elan DRC II from Perkin Elmer (USA).

DOTATATE labeling

DOTATATE was labeled with obtained 44Sc using various amounts of the peptide. The stock solution was prepared by dissolution of 100 µg DOTATE (69 nmol) in 100 ul 1 M ammonium acetate buffer. Next, the most active fraction of 44Sc solution (80 µL, 20 MBq) was combined with 100 µL 1 M ammonium acetate containing 3, 5, 10 and 15 nmol of DOTATATE. Peptide was labeled for 30 min at 95 °C.

Radiochemical yield was estimated by instant thin-layer chromatography (ITLC) using silica gel 60 TLC plates (Merck) in citrate buffer (1.5 M, pH 5). The 2 μL of the solution was spotted on the ITLC plate. Free 44Sc moved with the front boundary of the solution, whereas the labeled bioconjugate stayed at the starting point. The activity on the plate was measure by cyclone Plus Phosphor Imager (Perkin-Elmer, USA). The radiochemical yield of reaction was calculated as the ratio of activity of the plate application part to the whole plate activity.

DOTATATE labeling in presence of metal cations carried out as follows. Different amounts of metal salts was dissolved in 1 M HCl and alkalized to pH ~ 4,5 with ammonium acetate buffer. Then 20 MBq (80 µL) of 44Sc and 10 nmol (14.5 µL) DOTATATE was added. After mixing all components the reaction solution was then heated to 95 °C for 30 min. After that radiochemical yield was checked by ITLC method in citrate buffer as described above. Labeling was repeated 3–4 times in separate experiments.

Results and discussion

Production and separation of 44Sc

After 2 h of proton irradiation of natural CaCO3 target at 16 MeV current gives 419 MBq of 44Sc, 11.4 MBq of 43Sc, 7.3 MBq of 44mSc, 5,6 MBq of 47Sc and 6.5 of 48Sc.

The separation process was completed within 50 min from the end of bombardment (EOB) and involved the following steps: dissolution of target, alkalization with NH3, trapping of 44Sc on the filter, washing of filter, elution of 44Sc, adsorption of eluent on the Dowex 50 column and elution 44Sc with ammonium acetate buffer. Finally we obtained 270 MBq of 44Sc in 0.7 mL of 1 M ammonium acetate buffer (pH ~ 4,5). Scandium recovery after filtration step was: 94.1 ± 1.0% and after additional purification on Dowex 50 column 89.8 ± 1.2%. Separation of 44Sc is relatively simple and can be quickly implemented and automated.

The solution after dissolution of the target and final product after separation process was analyzed with ICP-MS for trace metal contaminants. The results of analysis is presented in Table 1.

Table 1 Concentration of cationic impurities in solution after dissolution of the target and after separation steps
Full size table

The obtained results indicate that the concentration of calcium, iron, zinc, copper and aluminum was relatively low. In the case of Ca2+ and Fe3+ impurities their concentration after separation process decrease by a factor of 103 and 40 respectively, but in the case of Zn2+ by only 1.5.

Metal ion competition on labeling DOTATAE with 44Sc

The obtained 44Sc solution was tested for labeling with DOTATATE bioconjugate. As show in Fig. 2 the DOTATATE concentration of 40 nmol is sufficient to achieve the full labeling.

Fig. 2
figure 2

Influence of DOTATATE concentration on its labeling with 44Sc. Volume of solution = 0.25 ml, 44Sc radioactivity 30 MBq

Full size image

The obtained molar activity of 44Sc-DOTATATE does not exceed 3 MBq/nmol. This is related to the using a natural calcium target, from which we obtained relatively low 44Sc activity, which caused a relatively high ratio of metallic impurity concentrations to 44Sc. Therefore, the molar activity was small. Similar molar activity was obtained in others works where natural Ca targets were used [18].

To evaluate the influence of metal cations contaminant on the radiolabeling of DOTATE with 44Sc solutions, the formation of 44Sc-DOTATATE was investigated in the presence of metal cations added to the solution. The cations were selected taking into account formation of stable complexes with the DOTA ligand (Fe2+/3+, Cu2+, Zn2+) and possibility of their occurrence in the labeling solution (Ca2+ from target material, Al3+ from aluminum foil and holder used for irradiation). The results are shown in Fig. 3.

Fig. 3
figure 3

Influence of Ca2+, Fe2+/3+, Zn2+ and Cu2+ concentration on DOTATATE labeling with 44Sc. Concentration of DOTATATE—40 nmol/ml, volume of solution = 0.25 ml

Full size image

We compared obtained results to relevant stability constants for formation of M(DOTA) complexes presented in Table 2. In M3+-DOTA molecule complexation occurs through four N donor atoms in the ring and four carboxylic groups. In the case of the DOTATATE conjugate, the coordination sphere of M3+ cations is filled by four donor nitrogen atoms, three carboxyl groups and one carbonyl group. To the best of our knowledge no stability or formation constants for [M(DOTATATE)] type complexes have been determined. More data is for DOTA-amides derivatives, but they are usually for DOTA-tetraamides derivatives. Due to the similarity of complex structures we decided to use stability constants determined for the DOTA ligand. We believe, that the stability constants values for monoamide complexes may be slightly smaller, but the same trends in stability constants should be maintained.

Table 2 Log K values for the M + DOTA = M(DOTA) reactions for selected cations
Full size table

For Ca2+ and Al3+ no effect was detectable up to 104 nmol of Ca2+and Al3+. Similar results were obtained also by Pruszyński et al. [13], where transmetallation reactions were not observable for 44Sc labeled DOTA-peptides, even at high Ca2+ concentrations of 400 ppm. This is understandable considering the low stability constant of Ca2+-DOTA and Al3+-DOTA complexes (17.2 and 17.0 respectively). In addition, in the case of Al3+ the formation of solid Al(OH)3 is observed. However for Cu2+, Zn2+ and Fe2+/3+ cations, we can see a much higher impact on labeling efficiency.

Analyzing the results obtained it is clear that there is no correlation between the stability constants of [M(DOTA)] complexes and influence M cations on radiolabeling. In particular, on the basis of the stability constants values it is not possible to predict the stronger effect of the Zn2+ on the DOTATATE labeling with the 44Sc radionuclide. On the contrary the stability constants of DOTA complexes suggest less impact of Zn2+ than Cu2+ on DOTATE labeling. In fact, since the stability constant of the Fe3+-DOTA is more than hundred times higher than those of Sc3+ we expected also a much greater influence of iron than Zn2+ cations. Lower impact of iron for labeling can be explained by the reduction of Fe3+ to Fe2+ in the labeling conditions. As see in Table 2 the DOTA complex with Fe2+ are 9 orders of magnitude less stable than those of Fe3+ and Fe2+ much less compete for coordination sites of the DOTA ligand. Such effect was also observed by Asti et al. [9] studying the influence of metals cations on the labeling of DOTATATE conjugate with 90Y and 177Lu, and Šimeček et al. [11] examining DOTA complex formation in presence of competing cations. It should be noted that while the Fe2+,3+, Zn2+ and Cu2+ cations strongly influence on the labeling process of DOTA radiobioconjugates with 44Sc, their impact on the stability of the formed complexes is negligible [13,14,15]. This is associated with kinetic inertness of DOTA complexes.

Conclusions

While studying the influence of Fe2+/3+, Cu2+, Zn2+, Al3+ and Ca2+ on formation 44Sc-DOTATATE conjugate we found that labeling efficiency strongly decreases in presence of trace concentration (> 2 nmol) of iron and zinc cations. Cu2+ cations reduce the labeling to a lesser extent (at concentrations greater than 10 nmol) and in the case of Ca2+ and Al3+ we do not observe any effect on labeling up to 2 μmol. Since 44Sc labeled radiopharmaceuticals started to be used on the patients the obtained results can be transferred to radiopharmaceutical preparation and used to optimize labeling conditions.