Background

Compared to β- or γ-radiation, tissue interaction of α-particles is characterized by a higher linear energy transfer (LET) and a much higher cell toxicity due to an enhanced probability of causing DNA double strand breaks [1]. In addition, the low tissue penetration depth of α-radiation (3–4 cell diameters) entails a more localized therapeutic effect, ideal for killing remaining single cancer cells or micrometastases which, in most conventional treatment regimes, can survive and later function as nuclei of tumor recurrence. Nevertheless, radionuclide therapy of cancer using radiopharmaceuticals labeled with α-emitting radionuclides (referred to as “targeted alpha therapy,” TAT) has hitherto played only a limited clinical role, although the therapeutic potential of the α-emitter 225Ac (T½ = 9.92 d) [2] has been emphasized already in 2001 [3]. The recent approval and market entry of 223Ra chloride as an α-emitting therapeutic radiopharmaceutical [4] and successful application of 225Ac-labeled inhibitors of prostate-specific membrane antigen (PSMA) for treatment of prostate cancer [5] highlighted the clinical potential of α-therapy and led to a tremendous boost of attention for TAT.

However, despite of proven suitability of 225Ac for treatment of terminal cases, such as β-refractory prostate cancer patients [5], there are still concerns regarding safe applicability of this nuclide for other than palliative use. There are four α-decays in its multistep decay scheme, while the recoil of the first α-emission releases the nuclide from the binding site (typically a chelate) [6]. In adverse cases, slow or incomplete internalization into cells may lead to uncontrolled distribution of the various α-emitting daughter nuclides in the body, causing the risk of severe side effects (e.g., kidney toxicity or carcinogenesis) owing to undesirable irradiation of healthy tissue [7]. In view of these issues, 213Bi (T½ = 46 min) [8], a late daughter nuclide of 225Ac, appears to be a valuable alternative. Diffusion after recoil is not a problem because a stable isotope, 209Bi, is obtained after decay via nearly simultaneous α- and β-emissions, while an additional 440 keV γ-line enables scintigraphic imaging. 213Bi is conveniently obtained from 225Ac/213Bi generators, small shielded chromatographic benchtop devices containing 225AcIII adsorbed to an organic matrix, from which 213BiIII is eluted with iodide solution in form of the [213BiI4]− and [213BiI5]2− complexes [9]. Hence, 213Bi has been exploited for various therapeutic applications [10,11,12], none of which, however, reached clinical routine so far, above all, due to very limited availability of 213Bi. Nevertheless, the awakened interest in 225Ac and the foreseeable expansion of global 225Ac production capacity will also entail a wide availability of 213Bi generators in the near future [13]. Overall, a higher inherent safety of 213Bi, resulting from its short half-life and a decay scheme involving only a single α-decay, renders this nuclide attractive for future development of α-therapeuticals which, in turn, calls for improvement of the corresponding labelling chemistry that hitherto received only little attention.

Materials and methods

General

p-NH2-CHX-A"-DTPA was purchased from Macrocyclics (Plano, TX, USA). 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) was purchased from CheMatech (Dijon, France). Previously published, optimized protocols were applied for synthesis of the chelators 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetrakis[methylene(2-carboxyethylphosphinic acid)] (DOTPI) [14] as well as for 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetrakis(methylenephosphonic acid) (DOTP), 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetrakis(methylenephosphinic acid) (DOTPH), and 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetrakis[methylene(phosphonic acid monoethyl ester)] (DOTPOEt) [15].

Radio-TLC was performed on glass microfiber chromatography papers impregnated with silica (ITLC®, Agilent). Using 0.1 M aq. sodium citrate as eluent, non-incorporated 213BiIII is transformed into the citrate complex which moves with the solvent front, while all chelates are sufficiently retained to enable ground-line separation (Rf < 0.5). Readout of chromatograms was done using a Bioscan TLC scanner, consisting of B-MS-1000 scanner, and B-EC-1000 detector with a B-FC-3600 GM tube.

213Bi labelling

213BiIII was eluted with a mixture of 0.2 M aq. HCl (0.3 mL) and 0.2 M aq. sodium iodide (0.3 mL) as anionic species [213BiI4]− and [213BiI5]2−) from a225Ac/213Bi generator system with an initial activity of 150 MBq as provided by the Institute for Transuranium Elements (Karlsruhe, Germany) [16]. The eluate was adjusted to pH 5.5 with 1 M aq. NaOAc buffer (1.6 mL). Labelling was performed by addition of the buffered eluate (90 μL) into an Eppendorf cup containing the precursor solution (10 nM–1 mM, 10 μL), resulting in final chelator concentrations of 0.001–100 μM. After 5 min of incubation at ambient temperature (approx. 25 °C) or at 95 °C, the fraction of complexed 213BiIII was evaluated by radio-TLC.

Stability studies

Stability of 213BiIII-complexes was tested in human plasma or 0.1 M aq. Na-DTPA (pH 7.5) by addition of 10 μL of the labelling solution (with 1 mM ligand concentration), containing the radiometal complex, to 90 μL of the competing medium at 37 °C. The fraction of intact chelate was evaluated by radio-TLC. Values were normalized to the fraction of radiometal complex at t = 0.

Results

Up to now, 213BiIII labelling virtually exclusively relied on well-established acyclic or cyclic polyamino-polycarboxylate ligands, above all, CHX-A"-DTPA [17] or the highly popular and versatile chelator DOTA [18] (Fig. 1). However, we previously noticed that 1,4,7-triazacyclononanes bearing phosphinic acids as N-substituents (TRAP chelators) [19] show superior labelling efficiency for the short-lived trivalent positron emitter 68GaIII [20, 21] in comparison to their parent tricarboxylate NOTA (1,4,7-triazacyclononane-1,4,7-triacetic acid) [22], pointing at potentially superior radiolabelling properties of phosphorus-pendant azamacrocycles in general. Thus, we elucidated the potential of cyclen-based chelators with phosphorus-based N-pendant arm donors for 213BiIII complexation.

Fig. 1
figure 1

Chelator motifs investigated in terms of 213BiIII complexation properties

Full size image

For this purpose, 213BiIII labelling of phosphonic acid derivatives DOTP [23] and DOTPOEt [24] as well as of phosphinic acids DOTPH [25] and DOTPI [14] was compared to the aforementioned standard scaffolds DOTA [26] and CHX-A"-DTPA [27], also under mild conditions (ambient temperature, pH 5.5) compatible with any type of biological targeting vector, including antibodies (Fig. 2; Tables 1 and 2). DOTA shows the poorest performance among all chelators investigated and, as with virtually all other radiometals, apparently cannot be labeled quantitatively with 213Bi within reasonable ranges of concentration and time at ambient temperature. This is why open-chain chelators, particularly CHX-DTPA derivatives, are usually applied for this purpose, despite of inherently lower in vivo stability of their MIII complexes [28]. However, to our surprise, the performance of phosphorus-based cyclens was found at least comparable to CHX-A"-DTPA, while DOTP showed particularly efficient radiolabelling, most likely due to higher affinity of trivalent bismuth to the relatively hard phosphonate oxygen atoms. This is a remarkable finding because in contrast to open-chain ligands, metal ion complexation by cyclic chelators usually occurs slower, via a two-step mechanism [29]. An initially formed out-of-cage complex, wherein the metal ion is coordinating only to side arm oxygen donors and solvent (water) molecules [25] is transformed into the in-cage complex, characterized by a N4O4 coordination mode of the ligand, via a substantial energy barrier. In terms of radiolabelling, this barrier causes a slower activity incorporation, but, on the other hand, is also related to higher kinetic inertness of the radiometal complexes which translates to lower dissociation rates.

Fig. 2
figure 2

Incorporation of 213BiIII by chelate ligands shown in Fig. 1 as functions of ligand concentration. Mean values ± SD, n = 3. Labelling conditions pH 5.5, V = 0.1 mL, reaction time 5 min, at 25 °C (a) and 95 °C (b). For data in numerical form, see Tables 1 and 2

Full size image
Table 1 Percentage of incorporation of 213BiIII by chelate ligands (pH 5.5, V = 0.1 mL, reaction time 5 min at 25 °C). Data represent mean values ± SD, n = 3
Full size table
Table 2 Percentage of incorporation of 213BiIII by chelate ligands (pH 5.5, V = 0.1 mL, reaction time 5 min at 95 °C). Data represent mean values ± SD, n = 3
Full size table

To assess this important parameter, we characterized the stability of the 213Bi chelates in a transchelation challenge against DTPA, and in human plasma at 37 °C. Figure 3 and Table 3 show that in accordance with expectations, the 213BiIII-complex of the open-chain ligand CHX-A"-DTPA exhibits the lowest kinetic inertness, resulting in a larger extent of dissociation than observed for the cyclic systems. Among the latter, all phosphorus ligands show quite similar resistance against demetallation. Notably, their 213BiIII complexes are also more inert than that of DOTA, most likely because they are protonated at lower a pH [15, 30].

Fig. 3
figure 3

Stability of 213Bi chelates. Percentage of intact 213BiIII complexes after 120 min challenge with 0.1 M aq. disodium DTPA at pH 7.5 (black bars) and in human plasma (red bars), both at 37 °C. For more data see Table 3

Full size image
Table 3 Percentage of intact 213BiIII chelates after incubation at 37 °C with 0.1 M aq. sodium DTPA (pH 7.5) or human plasma. Data represent mean values ± SD, n = 3
Full size table

Discussion

With a > 90% stability in plasma over the entire dosimetrically relevant time period of 213Bi (approx. three half-lives), the phosphinate and phosphonate chelators appear better suited for a safe application in 213Bi therapeutics than CHX-A"-DTPA derivatives. In addition, the higher labelling efficiencies, i.e., lower molar amounts of chelator required for the same extent of radiometal incorporation, will provide radiopharmaceuticals with higher specific activity, that is, an improved ratio of labelled vs. non-labelled compound in the final preparation. Hence, by administration of the same amount of, e.g., a 213Bi labelled antibody, a multiple amount of activity could be deposited in the target (tumor) tissue, resulting in a substantially increased radiation dose per tissue volume and, consequently, in a more successful therapy.

Conclusion

In conclusion, we found that 213BiIII complexation properties of cyclen-based phosphinate and particularly of phosphonate ligands are superior to the gold standard acyclic or cyclic chelators for 213BiIII, CHX-A"-DTPA and DOTA, respectively, reaching comparable labelling yields at 2–4 orders of magnitude lower concentrations both at ambient and elevated temperatures. In view of such highly efficient 213Bi incorporation, the phosphorus chelators appear ideal for application in freeze-dried labelling kits as known from 99mTc tracers and in antibody conjugates for immunotherapy where they would offer the benefits of improved in-vivo stability and higher target doses due to higher specific activity. Because at last, targeted α-therapy is widely entering clinical healthcare schemes after remaining in an experimental state for decades [13], our results are expected to support the currently increasing efforts towards advanced 213Bi radiotherapeutics for improved treatment of cancer.