AAZTA5/AAZTA5-TOC: synthesis and radiochemical evaluation with 68Ga, 44Sc and 177Lu

Purpose AAZTA (1,4-bis (carboxymethyl)-6-[bis (carboxymethyl)]amino-6-methylperhydro-1,4-diazepine) based chelators were initially developed in the context of magnetic resonance imaging. First radiochemical studies showed the capability of AAZTA to form stable complexes with radiolanthanides and moderately stable complexes with 68Ga. For a systematic comparison of the labelling capabilities with current diagnostic and therapeutic trivalent radiometals, AAZTA5 (1,4-bis (carboxymethyl)-6-[bis (carboxymethyl)]amino-6-[pentanoic-acid]perhydro-1,4-diazepine) was synthesized representing a bifunctional version with a pentanoic acid at the carbon-6 atom. To evaluate the effect of adding a targeting vector (TV) to the bifunctional chelator on the complex formation, AAZTA5-TOC was synthesized, radiolabeled and tested in comparison to the uncoupled AAZTA5. Methods AAZTA5 was synthesized in a 5-step synthesis. It was coupled to the cyclic peptide TOC (Phe1-Tyr3 octreotide) via amide bound formation. AAZTA and AAZTA5-TOC complex formations with 68Ga, 44Sc and 177Lu were investigated at different pH, temperature and precursor amounts. Stability studies against human serum, PBS buffer, EDTA and DTPA were performed. Results AAZTA5 and AAZTA5-TOC achieved quantitative labelling (> 95%) at room temperature in less than 5 min with all three nuclides at pH ranges from 4 to 5.5 with low precursor amounts of 1 to 10 nmol. [44Sc]Sc-AAZTA5 complexes as well as [44Sc]Sc-AAZTA5-TOC were completely stable. The 177Lu complexes of AAZTA5 and AAZTA5-TOC showed high stability comparable to the 44Sc complexes. In contrast, the [68Ga]Ga-AAZTA5 complex stability was rather low, but interestingly, [68Ga]Ga-AAZTA5-TOC was completely stable. Conclusion AAZTA5 appears to be a promising bifunctional chelator for 68Ga, 44Sc and 177Lu with outstanding labelling capabilities at room temperature. Complex stabilities are high in the case of 44Sc and 177Lu. While [68Ga]Ga-AAZTA complexes alone lacking stability, [68Ga]Ga-AAZTA5-TOC demonstrated high stability. The latter indicates an interesting feature of [68Ga]Ga-AAZTA5–labelled radiopharmaceuticals.


Introduction
The ongoing development in molecular imaging is focusing more and more on theranostic approaches as they link imaging directly with therapy as well as monitoring of the treatment (Baum and Kulkarni 2012;Baum and Rösch 2013;Rösch et al. 2017). In this context, patient-individual dosimetry is important for application of long-lived therapeutic nuclides. This emphasizes the advantage of a longer-lived PET-nuclide like 44 Sc over 68 Ga (Khawar et al. 2018;Roesch 2012a;Kerdjoudj et al. 2016).
Most of the commonly used nuclides for PET imaging and therapy are trivalent metals Me(III), thus have a positive charge of 3 as ions in solution. The most frequently used bifunctional chelators (BFC) are macrocyclic derivatives based on DOTA and NOTA. Bifunctionalization of DOTA normally uses one of the acid groups to connect to the targeting vector (TV). DOTA-conjugated TV's are typically 68 Ga-labeled at ca. 95°C for 10-15 min to achieve high labelling yields (Tsionou et al. 2017;Price and Orvig 2014;Roesch 2012b). NOTA-or NODAGA-conjugated analogues allow 68 Ga-labelling at lower temperature, if larger amounts of the labelling precursor are utilized, yet for many compounds at lower concentration heating to 95°C is common to achieve fast and quantitative labelling (Eisenwiener et al. 2002). For theranostic strategies with 177 Lu only DOTA can be used, since NOTA derivatives are not suitable for complexing 177 Lu. Again, 177 Lu-labellig of DOTA-peptides is achieved at temperatures close to 100°C and optimal labelling pH is around 4.
In addition, some relevant molecular targeting vectors (e.g. based on proteins) are temperature and pH sensitive. Accordingly, optimal chelators for radiolabeling sensitive biomolecules with trivalent radiometals ( 68 Ga, 44 Sc, 177 Lu, 90 Y, 213 Bi, 225 Ac) at room temperature and pH 5-6 would be extremely important. There are some proper candidates for labelling at room temperature. HBED (Schuhmacher et al. 1992), THP (Berry et al. 2011), H 2 dedpa (Boros et al. 2010(Boros et al. , 2012, DATA (Waldron et al. 2013;Seemann et al. 2017;Farkas et al. 2017;Seemann et al. 2015) show the desired labelling capabilities mostly for 68 Ga, while THP and DATA have optimum labelling parameters reaching fast quantitative yields at room temperature on wide pH ranges up to pH 6 (Tsionou et al. 2017). AAZTA 5 /AAZTA CN were reported to provide good labelling with not only 68 Ga, but also 44 Sc and 177 Lu (Pfister et al. 2015;Nagy et al. 2017;Manzoni et al. 2012).
In this work, we want to evaluate the AAZTA lead structure in several aspects. First, a bifunctional derivative (AAZTA 5 ), needed to later covalently attach the chelator to a TV, was synthesized. Here, our strategy is to replicate the approach we successfully introduced for the DATA chelator (Seemann et al. 2017). Second, we systematically optimize radiolabeling protocols for Me(III)-AAZTA 5 complexes with 68 Ga, 44 Sc and 177 Lu. Third, we correlate labelling yield with Me(III)-AAZTA 5 complex stability in vitro. Fourth, we synthesized a proof-of-principle radiopharmaceutical, namely AAZTA 5 -TOC, to again determine radiolabeling efficiency and in vitro stability.

Synthesis of AAZTA 5
All standard chemicals were acquired from Sigma-Aldrich, Merck and VWR. AAZTA 5 (2,2′-(6-(bis (carboxymethyl)amino)-6-(4-carboxybutyl)-1,4-diazepane-1,4-diyl) diacetic acid) was synthesized in a 5-step synthesis. The reactions to form the diazepine ring 1 via ring opening of nitrocyclohexanon and the full reduction to 2 has been published before as it is also part of the DATA synthesis ( Fig. 1) (Seemann et al. 2017 To deprotect the bifunctional acid group, the methyl ester is cleaved with LiOH in dioxane/water to afford 4, while for the evaluation of the free chelator 5 the protective tert-butyl groups were cleaved with TFA leaving the methyl ester intact.

Radiolabeling with 68 Ga, 44 Sc and 177 Lu
For radiolabeling with 68 Ga a 68 Ge/ 68 Ga generator (TiO 2 -based matrix, Cyclotron Co., Obninsk, Russia) was used with acetone post-processing separating iron and zinc impurities as well as 68 Ge breakthrough (Zhernosekov et al. 2007). Radiolabeling with 44 Sc was performed with a 44 Sc/ 44 Ti generator Pruszyński et al. 2010) located at the Institute of Nuclear Chemistry in Mainz. 177 Lu was provided by itg Munich following the carrier-free production pathway 176 Yb(n,γ) 177 Yb → 177 Lu (Lebedev et al. 2000).
Radiolabeling for all radiometals was performed in 1 ml of 0.2 M ammonium acetate buffer pH 4.5 and 5.5. Due to the post processing for 44 Sc the nuclide is provided in 0.25 M ammonium acetate buffer pH 4 (Pruszyński et al. 2010) and for first studies the pH was not adjusted above pH 4. With 1 ml labeling volume for 68 Ga and 44 Sc precursor concentrations of 5 and 10 pmol/l and for 177 Lu lower concentrations of 0.5 and 1 pmol/l were used. To show the mild labelling capability of the AAZTA 5 chelator, the reaction temperature was adjusted to 25°C with a BT 03 heater from HLC BioTech (Germany). Labelling studies were performed with 50 MBq for 44 Sc and 100 MBq for 68 Ga and 177 Lu. At different time points of 1, 3, 5 and 10 min aliquots for TLC and HPLC analytics were taken. The pH was controlled at start of labelling and after labelling, was finished.
For reaction control TLC (TLC Silica gel 60 F 254 Merck®) with citrate buffer (pH 7) and ammonium acetate buffer (pH 7)/MeOH 50/50 was used and compared to radio HPLC (Chromolith flush column, water: ACN with 0.1% TFA, 5 to 95% ACN in 10 min). TLCs were measured in RITA TLC imager (Elysia Raytest). The citrate TLC showed free radio metal with an Rf of 0.9 and all labelled compound sticked to an Rf of 0.1 to 0.3. As TLC would show colloidal radiometals sticking to an Rf of 0.1 to 0.2 as labelled compound, radio-HPLC was used to exclude the presence of colloidal radiometals, as colloidal radiometals cannot be eluted from the used HPLC columns.

Stability studies
Stability studies were performed in HS, PBS and EDTA/DTPA solution (pH adjusted to 7 by PBS buffer) in triplicate, starting from Me(III)-AAZTA 5 / AAZTA 5 -TOC batches of > 95% radiochemical purity. Time points were adjusted to the nuclides half-life: for 68 Ga 0.5, 1, 2 h; for 44 Sc 0.5, 1, 4, 8 h, 24 h; for 177 Lu 0.5, 1 and 2 h, 1 and 7 days. Stabilities against HS were also tested for adsorption to HS as fractions of the stability study at given time points were added to acetonitrile. After precipitation and centrifugation, the supernatant solution was removed and residue was measured for activity. HS (human male AB plasma, USA origin) was bought from Sigma Aldrich, PBS was prepared with a BupH™ Phosphate Buffered Saline Pack (PIERCE), EDTA and DTPA solution were prepared using the prepared PBS buffer by adding EDTTA and DTPA to a 0.01 M concentration.

Synthesis of AAZTA 5 and AAZTA 5 -TOC
The established synthesis for DATA 5m and DATA 5m -TOC (Seemann et al. 2017) was successfully transferred to AAZTA 5 and AAZTA 5 -TOC. The key step, attaching four tert-butyl-acetate-arms to the diazepine backbone, was optimized to work in good yields by forcing the reaction with slightly increased temperatures of 40°C and the addition of KI. Without the exchange from bromine to iodine on the tert-butyl bromoacetate nearly no alkylation was observed. Even with KI the reaction is still relatively slow and needs to be stirred over two days. Deprotection of the methyl group on the AAZTA 5 was done similar to that of DATA 5m . Purification by extraction (dichloromethane against 0.1 M NaCO 3 in water) gave the pure product ready for coupling in the dichloromethane phase. Ready for coupling AAZTA 5 could be isolated with a yield of 35% calculated over all reaction steps from the N,N′-dibenzylethylenediamine used in the first reaction step.
Formation of the AAZTA 5 active ester with HBTU (in DMF with 7 eq of DIPEA) was fast and could be monitored by LC-MS showing the active ester. After 15 min and positive LC-MS control, the active ester was added to the TOC solved in DMF. After one night, the DMF was removed and residues were resolved in 95% TFA/2.5% water/2.5% TiS (triisopropylesilane) to start the deprotection step. Lower concentrations of TFA with dichloromethane as solvent were not acidic enough to fully deprotect the chelator, leaving 1-2 tert-butyl groups even after 1 night. After 3-4 h of reaction the TFA were removed and the residue solved in 80% water/20% acetonitrile for HPLC. Yields after the HPLC were in a solid range of 30-40%.

Radiochemical evaluations with 68 Ga
For both the free chelator AAZTA 5 as well as the AAZTA 5 -TOC, radiolabelling with 10 nmol precursor showed an almost quantitative yield of > 98% in less than 5 min; most of the time reaching > 95% even after 1 min at room temperature ( Fig. 2). At 5 nmol of precursor amount, a slower kinetic was observed, but still yielding > 90% after 15 min most of the time, yet reproducibility became a problem. Variation of the pH from 4.5 to 5.5 showed no difference in the kinetics or the overall yield. Thus, 10 nmol was found to be the optimal precursor amount to prepare [ 68 Ga]Ga-AAZTA 5 and [ 68 Ga]Ga-AAZTA 5 -TOC, ready to be investigated in subsequent stability studies.
Stability of [ 68 Ga]Ga-AAZTA 5 against HS and PBS were high for 1 h, but decreased slowly at 2 h to a value of 85% and 79%, respectively. Adding EDTA or DTPA decreased the stability further up to 40-50% after 2 h. In contrast to the free chelator, the [ 68 Ga]Ga-AAZTA 5 -TOC was completely stable against HS, PBS, EDTA and DTPA over two hours (Table 1).

Radiochemical evaluation with 44 Sc
Labelling both AAZTA 5 and AAZTA 5 -TOC with 44 Sc showed a quantitative yield of > 97% with 10 nmol in less than 5 min, most of the time reaching > 95% even after 1 min, Sinnes et al. EJNMMI Radiopharmacy and Chemistry (2019)  all at room temperature. Lowering the precursor amount to five nmol induced slower kinetics for the free chelator, still yielding > 95% after 10 min for AAZTA 5 . In contrast, radiolabelling with five nmol of AAZTA 5 -TOC was always > 95% in less than 5 min at room temperature. Labelling with five nmol was stable enough to use it for stability studies. Stability studies showed for both the free chelator as well as the AAZTA 5 -TOC stability against HS, PBS, EDTA and DTPA over 8 h. [ 44 Sc]Sc-AAZTA 5 -TOC remained stable with > 90% in HS and PBS even over 24 h (Table 2).

Radiochemical evaluation with 177 Lu
General labelling procedures for DOTA conjugated radiopharmaceuticals with n.c.a. 177 Lu showed that DOTA is capable of complexing 177 Lu in molar ratios from 1: 5 to 1: 10 for save reproducible quantitative labelling. 100 MBq of 177 Lu is around 0.1 nmol. 177 Lu-labelling with 0.5 nmol AAZTA 5 (ratio 1: 5) showed quantitative yields of > 98% already after 1 min at room temperature. Formation of [ 177 Lu]Lu-AAZTA 5 -TOC had lower kinetics, yet also leading to yields of > 95% after 5 min. Both showed rapid labelling with 177 Lu in less than 1 min at one nmol precursor. Labelling behaviour was the same for pH 4.5 and 5.5. Labelling of 500 MBq 177 Lu with 5 nmol precursor (for stability studies) were also performed yielding > 98% after 1 min reaction time at room temperature (Table 3). Stability studies showed for both the free chelator as well as the AAZTA 5 -TOC good stabilities of > 90% against HS and PBS over 24 h. Stabilities against EDTA and DTPA were slightly lower with > 85% after 24 h. Longer studies of 7 d proved good stability against PBS with > 90% and a small degradation in HS with > 85% (mostly adsorption on serum proteins, visible by precipitation). 7 d stability values for EDTA and DTPA gave increasing instability down to 75 and 70% respectively for AAZTA 5 -TOC and to 60 and 50% respectively for the free chelator.

Discussion
Our experience concerning the synthesis of DATA 5m -TOC (Seemann et al. 2017) could be transferred, with adjustments on the alkylation step, to prepare the free chelator AAZTA 5 and its TOC derivative AAZTA 5 -TOC in good yields. Radiolabelling with 68 Ga showed good and rapid complexation at room temperature with quantitative  yields in less than 5 min for both the free chelator as well as the AAZTA 5 -TOC. The complexation also worked on pH 5.5 without showing issues forming colloidal 68 Ga, giving the same high yields after 5 min as for pH 4.5. This proves that elevation of the pH does not affect the labelling in a negative way. This rapid labelling at room temperature represents a significant advantage of AAZTA 5 over DOTA. Second, fast labelling kinetics of DOTA also demand higher temperatures and precursor amounts and this demonstrates the most relevant feature of AAZTA 5 compared to DOTAderivatives.
Stability of the free [ 68 Ga]Ga-AAZTA 5 complex was not 100% over the 2 h, releasing 10-15% of the 68 Ga against HS and PBS consistent with literature (Waldron et al. 2013;Baranyai et al. 2013). However, and extremely interesting, [ 68 Ga]Ga-AAZTA 5 -TOC was completely stable over two hours against HS and PBS, showing a positive effect of the TOC on the stability of the 68 Ga-complex. Due to the short spacer between the chelator and the TV, this could be a steric influence of the peptide moiety, but also the contribution of the amide formed in the coupling process may influence 68 Ga coordination. Adding EDTA or DTPA to the PBS increased the instability for AAZTA 5 , whereas AAZTA 5 -TOC remained fully stable. It appears to be necessary to study the impact of moieties other then TOC to further understand this phenomenon.
Rapid and quantitative 44 Sc-AAZTA 5 complex formation (both AATZA 5 and AAZTA 5 -TOC) at room temperature with low precursor amounts of five nmol demonstrate that the AAZTA 5 structure is almost ideal for labelling 44 Sc. With 100 MBq, 68 Ga in contrast to 50 MBq 44 Sc the effective amount of radiometal used per labelling was even higher for 44 Sc due to its longer half-life. In fact, 50 MBq 44 Sc contain near double the nmol of radio metal compared to 100 MBq 68 Ga. Using half of the precursor amount on double the nmol radio metal leads to a chelator to radio metal ratio in the labelling solution that is 4 times lower in the 44 Sc labelling than for 68 Ga one. In addition, less precursor is also lowering the chelator concentration in the labelling solution. Consequently, AAZTA 5 and AAZTA 5 -TOC have outstanding labelling capabilities for 44 Sc. [ 44 Sc]Sc-AAZTA 5 and [ 44 Sc]Sc-AAZTA 5 -TOC were completely stable against HS and PBS over two times the half-lives of 44 Sc and even stable after 24 h. Only the addition of EDTA and DTPA showed a small release of the 44 Sc after 24 h while the complexes are stable over 8 h. Combining the rapid quantitative radiolabelling with the good stability, AAZTA 5 and AAZTA 5 -TOC offer optimal labelling capabilities for 44 Sc.
Radiolabelling with 177 Lu was performed in equimolar ratios showing that even a ratio of 1: 5 between radiometal and chelator gave quantitative yields. Using ratios of 1: 10, rapid labelling is observed after one minute already at room temperature. Adjusting the pH to 5.5 had no influence on labelling speed or yields showcasing the same stable labelling as seen before with 68 Ga and 44 Sc. Stability over 24 h showed both the AAZTA 5 as well as the AAZTA 5 -TOC to be stable against HS and PBS with a slight decomplexation against ETDA and DTPA. After seven days, both complexes stay intact against PBS whereas the degradation by addition of EDTA and DTPA increases. For HS, some adsorption to serum proteins was observed by precipitation, while most of the complex stays intact. Overall, the complex stability is good for PBS and needs further evaluation in vivo to see the influence of the data from the HS.