Instant kit preparation of 68Ga-radiopharmaceuticals via the hybrid chelator DATA: clinical translation of [68Ga]Ga-DATA-TOC

Purpose The widespread use of 68Ga for positron emission tomography (PET) relies on the development of radiopharmaceutical precursors that can be radiolabelled and dispensed in a simple, quick, and convenient manner. The DATA (6-amino-1,4-diazapine-triacetate) scaffold represents a novel hybrid chelator architecture possessing both cyclic and acyclic character that may allow for facile access to 68Ga-labelled tracers in the clinic. We report the first bifunctional DATA chelator conjugated to [Tyr3]octreotide (TOC), a somatostatin subtype 2 receptor (SST2)-targeting vector for imaging and functional characterisation of SSTR2 expressing tumours. Methods The radiopharmaceutical precursor, DATA-TOC, was synthesised as previously described and used to complex natGa(III) and 68Ga(III). Competition binding assays of [natGa]Ga-DATA-TOC or [natGa]Ga-DOTA-TOC against [125I-Tyr25]LTT-SS28 were conducted in membranes of HEK293 cells transfected to stably express one of the hSST2,3,5 receptor subtypes (HEK293-hSST2/3/5 cells). First in vivo studies were performed in female NMRI-nude mice bearing SST2-positive mouse phaeochromocytoma mCherry (MPC-mCherry) tumours to compare the in vivo SST2-specific tumour-targeting of [68Ga]Ga-DATA-TOC and its overall pharmacokinetics versus the [68Ga]Ga-DOTA-TOC reference. A direct comparison of [68Ga]Ga-DATA-TOC with the well-established PET radiotracer [68Ga]Ga-DOTA-TOC was additionally performed in a 46-year-old male patient with a well-differentiated NET (neuroendocrine tumour), representing the first in human administration of [68Ga]Ga-DATA-TOC. Results DATA-TOC was labelled with 68Ga with a radiolabelling efficiency of > 95% in less than 10 min at ambient temperature. A molar activity up to 35 MBq/nmol was achieved. The hSST2-affinities of [natGa]Ga-DATA-TOC and [natGa]Ga-DOTA-TOC were found similar with only sub-nanomolar differences in the respective IC50 values. In mice, [68Ga]Ga-DATA-TOC was able to visualise the tumour lesions, showing standardised uptake values (SUVs) similar to [68Ga]Ga-DOTA-TOC. Direct comparison of the two PET tracers in a NET patient revealed very similar tumour uptake for the two 68Ga-radiotracers, but with a higher tumour-to-liver contrast for [68Ga]Ga-DATA-TOC. Conclusion [68Ga]Ga-DATA-TOC was prepared, to a quality appropriate for in vivo use, following a highly efficient kit type process. Furthermore, the novel radiopharmaceutical was comparable or better than [68Ga]Ga-DOTA-TOC in all preclinical tests, achieving a higher tumour-to-liver contrast in a NET-patient. The results illustrate the potential of the DATA-chelator to facilitate the access to and preparation of 68Ga-radiotracers in a routine clinical radiopharmacy setting. Electronic supplementary material The online version of this article (10.1186/s13550-019-0516-7) contains supplementary material, which is available to authorized users.


Background
There has been a surge in the development of 68 Ga-radiopharmaceuticals over the last decade initiated by the clinical and commercial success of [ 68 68 Ga generators now fulfilling pharmaceutical standards [1][2][3][4][5][6].
As a result, [ 68 Ga]Ga-DOTA-TOC and [ 68 Ga]Ga-DO-TA-TATE are currently being used in clinical settings for the diagnosis of neuroendocrine tumours (NETs). Furthermore, [ 68 Ga]Ga-DOTA-TATE acquired FDA approval as a diagnostic PET radiopharmaceutical for the visualisation of NET lesions (FDA News Release, June 1, 2016), following the 'orphan drug' designation to [ 68 Ga]Ga-DOTA-TOC by FDA [7], and [ 68 Ga]Ga-DO-TA-TOC was approved by European Medicines Agency. Due to the availability of 68 Ga via commercial 68 Ge/ 68 Ga generators and favourable emission characteristics for PET imaging (β + = 89%, E β,max = 1.9 MeV), the facile and efficient access to 68 Ga-radiopharmaceuticals is expected to drive the use of 68 Ga in PET centres [1,[8][9][10][11]. A key step in this direction is the development of simple, effective, robust, and reliable labelling protocols, which depend primarily on the chelating moiety of the bifunctional chelator (BFC) attached to the vector of interest. Established BFCs based on a DOTA or DO3A scaffold for 68 Ga require relatively harsh conditions (a balance of high temperatures, low pH, and high concentrations of the precursor) for efficient radiolabelling [2,12]. This restriction inherently limits the portfolio of 68 Ga-radiopharmaceuticals, because several promising peptide-, protein-, and antibody-based vectors for application in nuclear oncology are temperature and/or pH sensitive [13]. Thus, the radiolabelling of such molecules imposes stringent requirements on the BFC, i.e. > 95% labelling efficiency at ambient temperatures, less acidic conditions, and at high molar activities. Moreover, in the case of short-lived radionuclides, like 68 Ga (t 1/2 = 67.7 min), shorter labelling times and simple preparations are highly desirable, leading to a ready-for-injection radiolabelled product that does not require further purification prior to use. The development of such labelling protocols should be seen as mandatory to fully exploit the aforementioned advantages of 68 Ga, but presents significant challenges in the design of suitable BFCs [14].
In general, chelators (Additional file 1: Figure S1) can be distinguished as cyclic (DOTA, NOTA, TRAP), associated with high thermodynamic stability, or acyclic (DFO, DTPA, HBED, THP), linked to a high kinetic stability that allows for higher labelling efficiency [12,[15][16][17][18][19]. For example, Blower et al. demonstrated that THP derivatives are superior in terms of labelling kinetics [20] compared to BFCs with cyclic chelating functionalities and the novel THP-conjugated radiopharmaceuticals are under evaluation to prove their full viability in vivo for different targeting vectors.
Special chelators like TRAP offer good properties in general, but are predominantly seen in the context of multivalent applications. The DATA scaffold represents a unique approach to chelator design in that the chelating moiety is a hybrid, possessing both cyclic and acyclic character. It is believed that flexibility of the acyclic portion (6′ nitrogen and associated acetate function) facilitates rapid complexation, whilst the preorganised cyclic portion minimises the energy barrier to complexation and inhibits decomplexation processes [21,22]. The favourable radiolabelling kinetics of the DATA chelators, ambient temperature, and pH 4-6.5, along with the excellent stability of the forming 68 Ga-chelates, justified the development of a bifunctional derivative [23]. We recently reported on the synthesis and 68 Ga-radiolabelling of the first DATA peptide conjugate, DATA-TOC ( Fig. 1) [23].
Following the promising results of the initial work with uncoupled DATA-BFCs, the aim was to evaluate the suitability of a DATA-BFC with an established vector for comparison with the current clinical standard. Therefore, [ 68 Ga]Ga-DATA-TOC was selected as the first target for comparison with [ 68 Ga]Ga-DOTA-TOC as the clinically established reference in a series of biological in vitro and in vivo models expressing the somatostatin subtype 2 receptor (hSST 2 ), specifically (i) competition binding assays in human SST 2/3/5 -positive cell membranes, (ii) biodistribution and small animal PET imaging in a preclinical mCherry-expressing mouse phaeochromocytoma (MPC-mCherry) model with high SST 2 density [24,25], and (iii) clinical studies in a patient previously diagnosed with NETs. This direct comparison will reveal the influence of a DOTA-to-DATA chelator-switch on the biological behaviour of 68 Ga-labelled DOTA-TOC.

Cell culture and in vitro assays
The HEK293 cell line was transfected to stably express each of the hSST 2/3/5 and the resultant HEK293-hSST 2/3/5 cells used for receptor affinity assessments were donated by Prof. S. Schultz (Institute of Pharmacology and Toxicology, University Hospital, Friedrich Schiller University Jena, Germany). Cells were cultured at 37°C and 5% CO 2 in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 100 U/mL penicillin, 100 mg/mL streptomycin and 500 mg/mL G418, as previously described [28,30]. Culture reagents A B were from Gibco BRL, Life Technologies, and Biochrom KG Seromed. The genetically modified MPC-mCherry cells [25] derived from MPC cells (clone 4/30PRR [31]) characterised by a high density expression of mouse SST 2 [24] were cultured and prepared for in vivo application as previously described [24,25].

Animal studies
A number of 2 × 10 6 MPC-mCherry cells were transplanted subcutaneously into the right shoulder of female NMRI-nude mice (8 to 10 weeks old, RjOrl:NMRI-Foxn1 nu /Foxn1 nu , Janvier Labs). Tumour growth was monitored by fluorescence imaging using the in vivo Xtreme optical imaging system (Bruker) [24] under anaesthesia that was induced and maintained by inhalation of 12% and 9% (v/v) desflurane in 30/10% (v/v) oxygen/ air, respectively. Animals were studied when the tumour diameter was 6 to 9 mm.
For biodistribution studies with [ 68 Ga]Ga-DATA-TOC 17 (control n = 9, blocked n = 8) and [ 68 Ga]Ga-DOTA-TOC 12 (control n = 5, blocked n = 7) female mice (body weight 36.3 ± 2.1 g) were injected intravenously into a tail vein with approximately 2.3 MBq (62 μCi)/0.35 nmol peptide (DATA-TOC 11.2 nmol/kg body weight and DOTA-TOC 10.5 nmol/kg body weight) in 0.1 mL electrolyte solution E-153 (Serumwerk Bernburg AG) without (control) or with simultaneous injection of 100 μg/ mouse [Nal 3 ]octreotide acetate (blocked). Animals were sacrificed at 60 min post-injection (p.i.). Blood, tumour, and the major organs were collected, weighed, and counted in a cross-calibrated γ-counter (Isomed 1000, Isomed GmbH) and Wallac WIZARD Automatic Gamma Counter (PerkinElmer). The activity of the tissue samples was decay-corrected and calibrated by comparing the counts in tissue with the counts in aliquots of the injected radiotracer that had been measured in the γ-counter at the same time. The activity in the selected organs was expressed as percent-injected activity per organ (%IA) and the activity concentration in tissues and organs as standardised uptake value (SUV in [MBq activity/g tissue]/[MBq injected activity/g body weight]). Values are quoted as mean ± standard deviation for each group of animals.
PET scans were performed using a dedicated rodent PET/CT scanner (NanoPET/CT, Mediso). Anaesthetised mice (two animals per group) bearing subcutaneous MPC-mCherry-tumours on the right shoulder were positioned on a warmed bed along the scanner axis. The 68 Ga-labelled product, 10 MBq/0.26 nmol/300 μL (8.6 nmol DATA-TOC/kg body weight) and 10 MBq/ 0.26 nmol/300 μL (14.1 nmol DOTA-TOC/kg body weight), was infused over 1 min into a tail vein. PET images were acquired beginning with the injection on a Mediso NanoPET/CT camera and were reconstructed in dynamic mode with 38 frames and 0.5 mm 3 voxel size. Total scan time was 2 h. Region-of-interest (ROI) quantification was performed with ROVER (ABX GmbH). The ROI values were not corrected for recovery and partial volume effects. For each nanoPET/CT scan, 3D ROIs were drawn over the tumour, heart, muscle, liver, and kidneys in decay-corrected whole-body orthogonal images.
Statistical analyses were carried out with GraphPad Prism version 6 (GraphPad Software). The data expressed as mean ± SEM was submitted to a one-way analysis of variance (ANOVA) with post hoc Tukey's multiple comparisons test, with a single pooled variance. Values of P < 0.05 were considered statistically significant and indicated by an asterisk (*).

Small animal PET and biodistribution
Micro-PET imaging: specific tumour binding In dynamic PET studies in NMRI-nude mice, the implanted allogenic subcutaneous MPC-mCherry tumour was clearly visible with both radiotracers. Figure 3 shows coronal sections of dynamic PET images summarised from 1 to 2 h p.i. (midframe time 90 min) for one animal each. In vivo data for [ 68 Ga]Ga-DATA-TOC and [ 68 Ga]Ga-DOTA-TOC are illustrated for one animal each under A and C, respectively. For both radiotracers, the micro-PET data show a high accumulation of the radiotracers in the tumours. On a quantitative scale, the SUV (given in Fig. 3) appears to be higher for [ 68 Ga]Ga-DOTA-TOC. However, the micro-PET data are affected by photon energies of 68 Ga, by partial volume and spill-over effects. Accordingly, for a quantitative and statistically relevant comparison, we performed ex vivo organ distributions with n = 9 animals, see below.
In addition to the absolute tumour uptake of the two tracers it is important to address the specificity of the binding. Figure 3  Concerning the kidneys visualised in Fig. 3, kidney uptake is dependent in part on the individual hydration, urine flow of the mouse, and level of the anaesthesia. The figure shows individual mice at a specific time point during the PET study. The accumulation of the radiotracers in the kidney may differ across mice and from timepoint to timepoint. Consequently, kidney uptake is also addressed in the ex vivo biodistribution studies.

Micro-PET imaging: kinetics of tumour binding
The in vivo PET studies allow kinetic data for the SUV in several organs at different timepoints p.i. to be collected. Figure 4 shows the ratios between tumour and blood as SUV mean (tumour)/SUV mean (blood). The kinetic tumour-to-blood ratios of [ 68 Tables S2  and S3) for quantitative comparison of tumour accumulation, distribution, and elimination in control and blocked state. Figure 5a shows values of uptake in terms of %ID, whilst Fig. 5b shows values in terms of SUV. Both graphs also show ratios derived from the results of the blocking studies.
The tumour uptake of [ 68 Ga]Ga-DATA-TOC and [ 68 Ga]Ga-DOTA-TOC at 1 h after injection was in the same range with SUVs of 3.41 ± 1.43 and 4.52 ± 1.96 (P = 0.2838), respectively. These quantitative and statistically relevant ex vivo data are consistent with the in vivo PET data shown in Fig. 3.

Ex vivo biodistribution: specific tumour binding
The simultaneous injection of excess [Nal 3 ]octreotide clearly blocked the tumour accumulation for both radiotracers. The resulting activity concentrations were not statistically significantly different with 0. 36   0.1027). Ratios for tumour-to-blood and tumour-to-muscle at 1 h p.i. are graphically represented in Fig. 5c. This graph also shows ratios derived from the results of the blocking studies.
Pancreas: The pancreas expresses SST 2 and was therefore investigated as well. Similar to the tumour, there was a higher uptake of [ 68    post-tracer injection, demonstrated a similar, very intense hSST 2 -uptake in the primary pancreatic tumour (Fig. 6). There was a notable lower uptake of 68 Ga]Ga-DATA-TOC in normal liver (Table 1)

Discussion
The novel TOC-conjugate, DATA-TOC, showed the potential to establish an instant kit-type labelling routine of clinically relevant vectors with 68 Ga [23,33]. To establish that the DATA chelator does not negatively affect the receptor affinity and the in vivo performance of the targeting vector, [ 68/nat Ga]Ga-DATA-TOC was directly compared to [ 68/nat Ga]Ga-DOTA-TOC in a series of in vitro and in vivo studies.
Radiolabelling with 68 Ga for animal studies was completed quantitatively at 20°C for DATA-TOC, whereas for DOTA-TOC a higher temperature was required to achieve comparable labelling efficiency. This finding corroborates previously reported radiochemical data for convenient and simple kit-type labelling of DATA-TOC with 68 Ga [23].
[ nat Ga]Ga-DATA-TOC and [ nat Ga]Ga-DOTA-TOC showed high affinity for the hSST 2 . Although [ nat Ga]-Ga-DOTA-TOC displayed fivefold higher affinity than [ nat Ga]Ga-DATA-TOC in this assay, absolute differences in the pertinent IC 50 values were < nM (Fig. 3). Based on previous studies, such differences can be considered minimal for clinical translation given that many other critical factors (such as agonism or antagonism, stability, pharmacokinetics, or tumour residence times) greatly affect final clinical outcomes [28,30,34,35]. Therefore, it is reasonable to conclude that the exchange of DOTA for the DATA chelator was well tolerated by the hSST 2 -target. Furthermore, first in vivo small animal PET studies comparing [ 68 Ga]Ga-DATA-TOC to [ 68 Ga]Ga-DOTA-TOC showed similar biodistribution and kinetic profiles. The uptake in the tumours was specific, reaching comparable values and following similar kinetics. The tumour accumulation of both radiotracers was blocked by [Nal 3 ]octreotide acetate with similiar activity concentration, suggesting an SST 2 -specific process. Ex vivo organ distribution data was collected to mitigate misleading micro-PET data that can be affected by photon energies of 68 [39]. NOPO was attached to the octreotide derivative NaI 3 -octreotide (NOC), labelled with 68 Ga and evaluated in athymic CD-1 nude mice with AR42J tumours using micro-PET imaging and ex vivo biodistribution [40]. Uptake of [ 68 Ga]Ga-NOPO-NOC in the tumours was high and specific, whilst uptake in other organs and tissue was low with the exception of the kidneys. It is not surprising that for the same molecular targeting vector (e.g. octreotide) and the radionuclide (e.g. 68 Ga), the chelate will make a difference to the characteristics of the resulting radiotracer. Therefore in the development of new radiotracers, it is critical to quantify and understand the impact any new chelate may have on binding affinity, internalisation, organ distribution, uptake kinetics, and excretion pathways of a certain type of radiopharmaceutical in head-to-head assays. This will demonstrate to what extent the ease of radiolabelling demonstrated for a new group of chelators can be translated into clinical application, challenging the state-of-the-art chelators such as DOTA and ultimately to the benefit of patients.

Conclusion
It has been shown that [ 68 Ga]Ga-DATA-TOC can be prepared in a simple kit-type manner, and under milder conditions than the DOTA-based counterpart. The described small animal studies and first-in-human study showed [ 68 Ga]Ga-DATA-TOC equally able, and in some cases slightly better, for the visualisation of NET lesions compared to [ 68 Ga]Ga-DOTA-TOC. Combining these results with the in vitro data, the chelator-switch from DOTA to DATA did not negatively affect the biological efficiency of the 68 Ga-labelled TOC. Thus, this proof-of-principle study demonstrated the practical advantages of DATA for instant kit-type labelling without negatively affecting the efficacy.
These advantages highlight the potential of the DATA chelator as a promising tool for 68 Ga-radiolabelling in general, but especially for radiolabelling of heat-and/or pH-sensitive vectors. As a future perspective, the instant-kit type labelling of DATA-based molecular vectors will be broadened to include other medically interesting molecules, such as antibody fragments.