The tetraamine chelator outperforms HYNIC in a new technetium-99m-labelled somatostatin receptor 2 antagonist

Background Somatostatin receptor targeting radiopeptides are successfully being used to image, stage, and monitor patients with neuroendocrine tumours. They are exclusively agonists that internalise upon binding to the relevant receptor. According to recent reports, antagonists may be preferable to agonists. To date, 99mTc-labelled somatostatin receptor antagonists have attracted little attention. Here, we report on a new somatostatin receptor subtype 2 (sst2) antagonist, SS-01 (p-Cl-Phe-cyclo(D-Cys-Tyr-D-Trp-Lys-Thr-Cys)D-Tyr-NH2), with the aim of developing 99mTc-labelled ligands for SPECT/CT imaging. SS-01 was prepared using Fmoc solid-phase synthesis and subsequently coupled to the chelators 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA), 6-carboxy-1,4,8,11-tetraazaundecane (N4), and 6-hydrazinonicotinic acid (HYNIC) to form the corresponding peptide-chelator conjugates SS-03, SS-04, and SS-05, respectively. SS-04 and SS-05 were radiolabelled with 99mTc and SS-03 with 177Lu. Binding affinity and antagonistic properties were determined using autoradiography and immunofluorescence microscopy. Biodistribution and small animal SPECT/CT studies were performed on mice bearing HEK293-rsst2 xenografts. Results The conjugates showed low nanomolar sst2 affinity and antagonistic properties. 177Lu-DOTA-SS-01 (177Lu-SS-03) and 99mTc-N4-SS-01 (99mTc-SS-04) demonstrated high cell binding and low internalisation, whereas 99mTc-HYNIC/edda-SS-01 (99mTc-SS-05) showed practically no cellular uptake in vitro. The 99mTc-SS-04 demonstrated impressive tumour uptake at early time points, with 47% injected activity per gram tumour (%IA/g) at 1 h post-injection. The tumour uptake persisted after 4 h and was 32.5 %IA/g at 24 h. The uptake in all other organs decreased much more rapidly leading to high tumour-to-normal organ ratios, which was reflected in high-contrast SPECT/CT images. Conclusions These data indicate a very promising 99mTc-labelled sst2-targeting antagonist. The results demonstrate high sensitivity of the 99mTc-labelling strategy, which was shown to strongly influence the receptor affinity, contrary to corresponding agonists. 99mTc-SS-04 exhibits excellent pharmacokinetics and imaging properties and appears to be a suitable candidate for SPECT/CT clinical translation. Electronic supplementary material The online version of this article (10.1186/s13550-018-0428-y) contains supplementary material, which is available to authorized users.


Background
Somatostatin receptors are important biomarkers for imaging and targeted radionuclide therapy of human cancers. They belong to the large family of G-protein coupled receptors, which currently account for 30-40% of marketed drugs [1] and are not only overexpressed in neuroendocrine tumours in particular but also in non-neuroendocrine tumours [2]. The receptors make ideal targets for imaging, as they are easily accessible on the plasma membrane of the tumour cell. Their action is mediated through two mechanisms: G-protein activation and β-arrestin function. Among its signalling roles, the latter promotes receptor internalisation-an important mechanism for radiolabelled agonist ligand uptake, accumulation, and retention [3,4].
In contrast, binding of neutral antagonists does not lead to internalisation but potentially involves many more binding sites than agonist-based approaches. This may lead to the binding of a higher number of radiovectors and thus to a stronger signal originating from the tumour. Ginj et al. have shown that in both in vitro and in vivo animal models, somatostatin-based radioantagonists may indeed be superior to radioagonists [5]. These findings have recently been duplicated with antagonists conjugated with DOTA and NODAGA and labelled with the positron-emitting radiometals 68 Ga and 64 Cu and other 3+ (radio)metals [6][7][8] and further supported by first-in-human imaging and therapy studies [9][10][11][12]. Similarly, preclinical [13][14][15][16] and clinical [17] studies of bombesin-based radioantagonists showed that using antagonists may be advantageous over agonists for targeted imaging and therapy of GRP receptor-express ing tumours.
Despite the fact that the first somatostatin-based radioantagonists have been studied in humans and proven to be superior to registered radioagonists, little is known about the influence of bioconjugation, labelling strategies, and other modifications on the in vitro and in vivo pharmacology of radiolabelled antagonists. Wadas et al. compared a potent agonist, TATE ([Tyr 3 ,-Thr 8 ]octreotide) labelled with 64 Cu using CB-TE2A, a cross-bridged cyclam-14 derivative, with the antagonist (sst2-ANT, p-NO 2 -Phe-cyclo(D-Cys-Tyr-D-Trp-Lys-Thr-Cys)D-Tyr-NH 2 ), labelled using the same strategy [18]. They showed that the radioantagonist had lower tumour uptake despite the much higher receptor numbers found in the AR42J tumour model. AR42J cells endogenously express sst2 receptors. The authors also reported a much lower receptor affinity for the antagonist, which might be the reason for the difference rather than the origin of the receptor (natural or transfected), as proposed by the authors.
Even more intriguing is a study by Dude et al. [19], who compared the antagonist 68 Ga-NODAGA-JR11 with the two agonists 68 Ga-DOTATOC and 68 Ga-DOTATA TE. Surprisingly, the authors found that 68 Ga-NODA-GA-JR11 has the lowest tumour uptake in their human ZR-75-1 breast tumour model. In addition, the number of receptors using 68 Ga-DOTATOC was more than twofold higher than 68 Ga-DOTATATE and significantly higher than the antagonist.
More clinically relevant, however, is the fact that the phenomenon of higher tumour uptake of antagonists was also observed in human tumours. Reubi et al. performed quantitative autoradiography in neuroendocrine and non-neuroendocrine tumour specimens using the 125 I-DOTA-JR11 antagonist and 125 I-Tyr 3 octreotide agonist and found up to tenfold higher uptake of the antagonist radioligand compared to the agonist. The authors concluded that all renal cell cancers, most breast cancers, non-Hodgkin lymphomas, and medullary thyroid cancers appear to be novel targets for in vivo targeting with sst2 radioantagonists [20].
These different observations emphasise the need to study the influence of overall structure on pharmacology in vitro and in vivo in different cell and tumour models. In particular, modifications allowing 99m Tc-labelling have only been studied in one case [21], although 99m Tc-labelled radiopharmaceuticals are still the mainstay of nuclear medicine. We report here the synthesis of a new somatostatin-based antagonist, 4-Cl-Phe-cyclo(D-Cys-Tyr-D-Trp-Lys-Thr-Cys)D-Tyr-NH 2 (SS-01), which was designed for labelling with 99m Tc using two different strategies, namely N4: 6-carboxy-1,4,8,11-tetraazaundecane (SS-04) and HYNIC: 6-hydrazinopyridine-3-carboxylic acid (SS-05), as metal-binding domains. The choice of the chemical structure of the antagonist was based on our previous experiences. We and others observed that the chirality change of amino acids 1 and 2 (aa 1 , aa 2 ) and C-terminal amidation compared to octreotide type octapeptides transforms an agonist into an antagonist [22,23]. In addition, we chose the most easily accessible amino acids leading to antagonistic peptides. To serve as a control, we also modified the peptide with DOTA for labelling with 177 Lu ( 177 Lu-SS-03).

Methods
The supplier information for reagents, radiolabelling protocols, and log D determination, as well as details about instrumentation, are provided in the Additional file 1.

Synthesis of chelator-peptide conjugates, radiochemistry
The peptides were assembled on the Rink-Amide methylbenzylhydryl (MBHA) resin employing standard Fmoc strategy. The coupling reactions were performed on a semiautomatic peptide synthesiser (RinkCombichem, Bubendorf, Switzerland) with a threefold excess of Fmocamino acids, using DIC/HOBt as activating agents in DMF/NMP for 2 h (see Additional file 1 for details, including labelling protocols and corresponding highperformance liquid chromatography (HPLC) data).

Immunofluorescence microscopy
An immunofluorescence microscopy-based internalisation assay was performed on HEK293-rsst2 cells, as previously described [22]. Briefly, the cells were treated with different antagonist chelator-peptide conjugates and/or [Tyr 3 ]octreotide (TOC) (the sst2 agonist) at 37°C for 30 min. After fixation and permeabilisation, cells were stained with the sst2-specific primary antibody R2-88 (provided by Dr. Agnes Schonbrunn, McGovern Medical School, University of Texas Health Science Center, Houston, Texas, USA) as described previously [25,26]. The cells were imaged using a Leica DM RB immunofluorescence microscope, and the images were acquired using an Olympus DP10 camera.

Receptor-binding, internalisation, and dissociation kinetics
The receptor binding, internalisation, and dissociation rates of 177 Lu-SS-03, 99m Tc-SS-04, and 99m Tc-SS-05 were studied in HEK293-rsst2 cells seeded in six-well plates, as described previously [6]. Briefly, the radiopeptide (0.25 pmol/well) was added and the cells were incubated at 37°C. At different time points (0.5, 1, 2, and 4 h), the cellular uptake was stopped by washing twice with ice-cold PBS. The membrane-bound and internalised fractions were collected with ice-cold glycine buffer, pH 2.8 and 1 M NaOH, respectively.
For the dissociation experiments, the plates were placed on ice for 30 min. The radiopeptide (0.25 pmol/ well) was added to the cells and allowed to bind for 2 h at 4°C. The cells were then quickly washed with ice-cold PBS, and fresh pre-warmed (37°C) medium was added. The cells were incubated at 37°C for 10, 20, 30, 60, 120, and 240 min. The medium was collected for quantification, and the cells were treated as described previously [6].
The activity in each fraction was measured in a γ-counter (Cobra II). Nonspecific uptake was determined in the presence of 250 pmol/well [Tyr 3 ]octreotide. The results were expressed as a percentage of the applied radioactivity.
Biodistribution studies in HEK293-rsst2-bearing mice All animal experiments were conducted in compliance with Swiss animal protection laws and associated regulations. The protocol was approved by the Cantonal Veterinary Ethics Committee of the University of Basel (approval #789). Female athymic nude mice (4-6 weeks old) were injected subcutaneously (s.c.) in the right shoulder with 10 7 HEK293-rsst2 cells in 100 μL sterile PBS. The tumours were allowed to grow for 14-18 days (tumour weight 100-200 mg).
The mice were injected into the tail vein with 100 μL/ 10 pmol/0.37 MBq of 177 Lu-SS-03 or 99m Tc-SS-04 and were euthanised at 1, 4, and 24 h post-injection (p.i.). Organs of interest and blood were collected, rinsed of excess blood, blotted dry, weighed, and counted in a γ-counter. To determine nonspecific uptake, three animals were pre-injected with 20 nmol of the relevant unlabelled peptide in 0.9% NaCl solution (0.1 mL); after 5 min, the radiopeptide was injected and the percentage of injected activity per gram (%IA/g) was calculated for each tissue. The total counts injected per animal were determined by extrapolation from counts of an aliquot taken from the injected solution as a standard.

SPECT/CT imaging study
Mice bearing HEK-rsst 2 tumours were euthanised 4 h after intravenous injection of 15 MBq (150 pmoles) of 99m Tc-SS-04 and imaged supine, head first, using a SPECT/CT system dedicated to imaging small animals (NanoSPECT/CT™ Bioscan Inc.). Topogram and helical CT scan of the whole mouse was first acquired using the following parameters: X-ray tube current 177 μA, X-ray tube voltage 45 kVp, 90 s and 180 frames per rotation, pitch 1. The helical SPECT scan was then acquired from head to toe using multi-purpose pinhole collimators (APT1). The energy window width was 20% centred symmetrically over the energy peak of 99m Tc at 140 keV. Twenty-four projections (200 s per projection) were used, allowing the acquisition of at least 50 kilocounts/ projection.

Data analysis
Statistical analysis was performed using an unpaired two-tailed t test with Prism software (Prism 5.01, September 2007, GraphPad Software Inc.). Differences at the 95% confidence level (P < 0.05) were considered significant.

Results
Synthesis, radiolabelling, and distribution coefficients (log D) All conjugates ( Fig. 1) were synthesised with the maximum yield of 30-40% and purity ≥ 97%. The conjugates were characterised by analytical reversed phase HPLC and ESI-MS (Table 1, Additional file 1).
SS-03 was labelled with 177 Lu with labelling yields of > 97% at a maximum specific activity of 50 GBq/μmol. The conjugates SS-04 and SS-05 were labelled with 99m Tc at room temperature (30 min) and elevated temperature (95°C, 10 min), respectively. Tin(II)chloride was used as the reducing agent and citrate as an intermediate supporting Tc(V) ligand for the labelling of SS-04. For SS-05, 99m Tc labelling was performed using edda (ethylenediamine, N,N′-diacetic acid) as coligand. The radiolabelling yields of both 99m Tc-SS-04 and 99m Tc-SS-05 were > 97% at a specific activity of approximately 100 GBq/μmol.
Binding affinity and immunofluorescence microscopy Table 2  Immunofluorescence-based internalisation was performed using HEK-sst2 cells to demonstrate the antagonistic property of the conjugates. Figure 2 illustrates that 10 nM of the agonist [Tyr 3 ]octreotide (TOC) triggers massive receptor internalisation, whereas SS-03 or SS-04 at the much higher concentration of 1000 nM does not stimulate receptor internalisation. However, at a concentration of 1 μM together with 10 nM of TOC, the conjugates were able to prevent the agonist-induced receptor internalisation. Receptor-binding, internalisation, and dissociation kinetics Figure 3 shows the cellular uptake profiles of the radioligands as measured with HEK-rsst2 cells. 177 Lu-SS-03 and 99m Tc-SS-04 showed high uptake and blocking studies demonstrated that the uptake was receptor-mediated. Within 2 h of incubation, the specifically bound fraction levelled off at 48%, demonstrating rapid binding of 177 Lu-SS-03 and 99m Tc-SS-04 to the receptors. The internalised fraction was about 8-10% and increased very slowly. Surprisingly, 99m Tc-SS-05 showed very low cellular uptake of < 1% even after 4 h of incubation.

Biodistribution studies and SPECT imaging
The SPECT/CT images (Fig. 5) were acquired at 4 h after intravenous injection of 99m Tc-SS-04. The highest uptake was visible in the tumour and the kidneys. To a lesser extent, uptake was also visible in the abdomen due to tracer accumulation in the sst2-expressing organs.

Discussion
Radiolabelled peptides targeting G-protein-coupled receptors have been an important focus in the radiopharmaceutical and nuclear oncology fields over the last 20 years [27][28][29]. The clinically employed somatostatin receptortargeting peptides are agonists, which exhibit nanomolar binding affinities and fast receptor-mediated internalisation in vitro and in vivo via endocytosis. More recently, sst2 antagonists were studied in animal models [6][7][8]18] and in humans [9][10][11][12]. Surprisingly, the antagonists showed high and long-lasting tumour uptake in most cases. However, antagonist superiority cannot be generalised, as shown by Wadas et al. [18] and Dude et al. [19].
To date, many peptides labelled with gamma-, positron-, and beta-emitters ( 111 In, 64 Cu, 68 Ga, and 177 Lu)   have been described, but only one 99m Tc-labelled antagonist has been reported to date [21]. We selected two of the more popular and successful chelators for 99m Tc: HYNIC with the co-ligand edda (ethylenediamine-N,N′-diacetic acid) and the bifunctional tetraamine N4. The HYNIC core has been widely used for the labelling of octreotide-based somatostatin receptor-target ing peptides, and the [HYNIC, Tyr 3 ]octreotide/edda kit (HYNIC-TOC) is registered in some European countries. The [O=Tc=O] + core is formed with linear and macrocyclic tetraamines such as 1,4,8,11-tetraazaundecane and cyclam-14. This core exhibits high kinetic stability, can be labelled at room temperature, is highly hydrophilic, and with different functions in 6-position, such as carboxylic acid can be easily coupled to the N-termini of biomolecules [30]. We have conjugated these two chelators to a new antagonist, p-Cl-Phe-cyclo(D-Cys-Tyr-D-Trp-Lys-Thr-Cys)D-TyrNH 2 . As mentioned earlier, we chose the most easily accessible amino acids leading to antagonistic peptides. Our new antagonist is based on a modification of the first radiolabelled sst2 antagonists (DOTA-BASS) [5,9,31], where the p-NO 2 -Phe has been replaced by p-Cl-Phe. We also conjugated DOTA to this antagonist and labelled the conjugate with 177 Lu for comparative in vitro and in vivo studies.
The antagonistic properties of SS-03 and SS-04 were investigated using immunofluorescence microscopy; the two compounds inhibit receptor internalisation triggered by the potent agonist [Tyr 3 ]octreotide but do not trigger internalisation on their own. Somewhat surprisingly, but as previously reported [6,14,17], we observed low but significant cell internalisation of 177 Lu-SS-03 and 99m Tc-SS-04 (around 10% of the added radiopeptide activity per one million cells at 4 h). This uptake can be blocked by excess unlabelled peptide. The internalisation rate is significantly lower than that of somatostatinbased radiolabelled agonists [14,22,32]. The reason for the difference between the two assays is unclear.
In contrast, the N4-conjugated and 99m Tc-labelled radiopeptide showed superior properties, with very high cell uptake in vitro and about the highest tumour uptake at 1 and 4 h of any somatostatin-based radiopeptide studied to date in this xenograft model [6,7]. The higher tumour uptake of 99m Tc-SS-04, compared with 177 Lu-SS-03, may be attributed to the somewhat slower blood clearance of 99m Tc-SS-04, even though at 4 h p.i, the blood values are at the same level for both radiopeptides. In addition, it shows very good tumour retention, no change between 1 and 4 h, and only about 30% release between 1 and 24 h p.i. The target-to-relevant organ ratios are > 5 at 24 h, reflected in excellent SPECT/CT images. For both radiotracers, the kidney uptake is somewhat high but the tumour-to-kidney ratios are still within a reasonable range when compared to potent radioagonists. We are currently studying somewhat different peptide motifs coupled to tetraamine chelators for 99m Tc-labelling to improve the overall pharmacokinetics. The hypothesised improvement is based on the properties of 177 Lu-SS-03, which exhibits pharmacokinetics somewhat inferior to the previously published 111 In-DOTA and 177 Lu-DOTA conjugated antagonists [31]. As previously mentioned, Radford et al. studied a somewhat different sst2 antagonist peptide labelled using the 99m Tc tricarbonyl strategy and an NSN-type chelator [21]. The Re-complexed congener showed reasonably good receptor affinity indicating that this strategy, other than the HYNIC strategy, retains binding affinity [21]. The problem with the radioligand is its high abdominal uptake and low tumour-to-normal organ ratios, which is below 1 for some relevant tissues. As the authors pointed out, this is due to  Tc-based sst2 antagonist 99m Tc-SS-04 is hydrophilic, as indicated by its log D value. Consequently, it has limited accumulation in the abdomen and rather high renal excretion. The biodistribution profile of 99m Tc-SS-04, together   with its very high tumour uptake, favours this tracer for clinical translation.

Conclusions
Sst2 antagonists were labelled with 99m Tc using two common chelators. Surprisingly, the widely used 99m Tc ligand HYNIC along with the co-ligand edda resulted in almost complete loss of sst2 binding affinity, whereas 99m Tc-SS-04 demonstrated impressive tumour uptake and high tumour-to-normal organ ratios. Therefore, 99m Tc-SS-04 appears to be an excellent candidate for SPECT imaging of sst2-positive tumours and is a promising option for further modification of the peptide motif. The pharmacokinetics of 99m Tc-SS-04 demonstrate once again that the 99m TcO 2 (N4) core offers favourable pharmacokinetic features for small peptides and may potentially be the 99m Tc-labelling strategy of choice.