(R)-NODAGA-PSMA: A Versatile Precursor for Radiometal Labeling and Nuclear Imaging of PSMA-Positive Tumors

Purpose The present study aims at developing and evaluating an urea-based prostate specific membrane antigen (PSMA) inhibitor suitable for labeling with 111In for SPECT and intraoperative applications as well as 68Ga and 64Cu for PET imaging. Methods The PSMA-based inhibitor-lysine-urea-glutamate-coupled to the spacer Phe-Phe-D-Lys(suberoyl) and functionalized with the enantiomerically pure prochelator (R)-1-(1-carboxy-3-carbotertbutoxypropyl)-4,7-carbotartbutoxymethyl)-1,4,7-triazacyclononane ((R)-NODAGA(tBu)3), to obtain (R)-NODAGA-Phe-Phe-D-Lys(suberoyl)-Lys-urea-Glu (CC34). CC34 was labeled with 111In, 68Ga and 64Cu. The radioconjugates were further evaluated in vitro and in vivo in LNCaP xenografts by biodistribution and PET studies. Biodistribution studies were also performed with 68Ga-HBED-CC-PSMA (HBED-CC: N,N′-bis[2-hydroxy-5-(carboxyethyl)benzyl]ethylenediamine-N,N′-diacetic acid) and 111In-PSMA-617 for comparison. Results 68Ga-CC34, 64Cu-CC34, and 111In-CC34 were prepared in radiochemical purity >95%. 68/natGa-CC34, 64/natCu-CC34, 111/natIn-CC34, 68/natGa-HBED-CC-PSMA, and 111/natIn-PSMA-617 exhibited high affinity for the LNCaP cells, with Kd values of 19.3±2.5 nM, 27.5±2.7 nM, 5.5±0.9 nM, 2.9±0.6 nM and 5.4±0.8 nM, respectively. They revealed comparable internalization profiles with approximately 75% of the total cell associated activity internalized after 3 h of incubation. 68Ga-CC34 showed very high stability after its administration in mice. Tumor uptake of 68Ga-CC34 (14.5±2.9% IA/g) in LNCaP xenografts at 1 h p.i. was comparable to 68Ga-HBED-CC-PSMA (15.8±1.4% IA/g) (P = 0.67). The tumor-to-normal tissue ratios at 1 and 2 h p.i of 68Ga-CC34 were also comparable to 68Ga-HBED-CC-PSMA (P>0.05). Tumor uptake of 111In-CC34 (28.5±2.6% IA/g) at 1 h p.i. was lower than 111In-PSMA-617 (52.1±6.5% IA/g) (P = 0.02). The acquisition of PET-images with 64Cu-CC34 at later time points showed wash-out from the kidneys, while tumor uptake still remained relatively high. This resulted in an increased tumor-to-kidney ratio over time. Conclusions 68Ga-CC34 is comparable to 68Ga-HBED-CC-PSMA in terms of tumor uptake and tumor to normal tissue ratios. 64Cu-CC34 could enable high contrast imaging of PSMA positive tissues characterized by elevated expression of PSMA or when delayed imaging is required. 64Cu-CC34 is currently being prepared for clinical translation.


Introduction
Prostate cancer is the most common malignancy found in men and the second leading cause of cancer death in the US. In 2015, it is estimated that a total of 220,800 new prostate cancer cases will be diagnosed while 27,540 prostate cancer deaths are predicted to occur [1]. Therefore, efforts to discover and evaluate new diagnostic and therapeutic biomarkers for prostate cancer continue. Prostate specific membrane antigen (PSMA) is a well-established target for diagnostic and potential therapeutic applications. PSMA is a 750 amino acid type II integral membrane glycoprotein which is primarily expressed in healthy human prostate epithelium and in nonprostatic solid tumor vasculature without being shed into the circulation [2]. It is overexpressed by almost all prostate cancers with an increased expression by a factor of about 1000 in poorly differentiated, metastatic, and hormone-refractory cases [3][4][5]. The proposed pharmacophore for PSMA active site can be divided into three parts; three carboxylic groups, a carbonyl oxygen as part of the zinc complexation and nearby aromatic residues [6].
PSMA was originally targeted by the 111 In-labeled monoclonal antibody 7E11-C5 (ProstaScint 1 ; Cytogen Corporation, Princeton, NJ) which specifically binds to the PSMA+ human adenocarcinoma cell line LNCaP [7]. 7E11-C5 only binds to the intracellular site of PSMA (amino terminus) only accessible on necrotic tumors [8], therefore this tracer did lack wide acceptance in the field of nuclear medicine for the detection of prostate cancer. Radiolabeled monoclonal antibodies which bind to the extracellular site of PSMA were further developed [9][10][11][12][13]. Their successful preclinical evaluation and promising clinical assessment justified the utility of PSMA in the diagnosis and the potential therapy of prostate cancer [14,15]. As part of the ongoing efforts of several groups to develop new PSMA-specific ligands which outperform the disadvantages of antibodies such as inadequate pharmacokinetics and tissue accessibility, several chemical scaffolds such as PSMA inhibitors of low molecular weight have been synthesized and evaluated [16]. Among them, the urea-based PSMA inhibitors, functionalized to be used for imaging with Single-Photon Emission Computed Tomography (SPECT) or Positron Emission Tomography (PET), were able to successfully image PSMA-expressing xenografted mice [17][18][19][20][21][22][23][24][25][26][27][28][29].

Material and Methods
The supplier information for all reagents and details of instruments used are provided in the S1 Appendix.

Synthesis of the urea-based compound CC34
The PSMA-based inhibitor, CC34, was synthesized using standard Fmoc chemistry. Description of the synthesis is given in the S2 Appendix.
The 64 Cu-and 111 In-labeled radiotracers were prepared within 30 min at 95°C and were used without any further purification (S3 Appendix).

Saturation binding / Internalization studies
For receptor saturation analysis, PSMA positive LNCaP cells (metastatic lesion of human prostatic adenocarcinoma, ATCC) were seeded at a density of 0.8-1 million cells per well in 6-well poly-L-lysine PLL-coated plates and incubated overnight with medium (RPMI Medium 1640 -GlutaMAX containing 1% FBS, 100 U/mL penicillin and 100 μg/mL streptomycin, sodiumpyruvat 1 mM). The next day, the medium was removed, the cells washed and incubated for 1 h at 37°C with fresh binding buffer (RPMI Medium 1640-GlutaMAX containing 1% FBS, 100 U/mL penicillin and 100 μg/mL streptomycin, 50 mM Hepes, 50 μg/mL bacitracin, 0.5% BSA). Afterwards, the plates were placed on ice for 30 min followed by incubation with increasing concentrations of either 68/nat Ga-CC34, 64/nat Cu-CC34, 111/nat In-CC34 and 111/nat In-PSMA-617 (1-100 nM) in phosphate-buffered saline binding buffer pH 7.4. After the addition of the radioligands, the cells were incubated for 120 min at 4°C. Non-specific binding was determined in the presence of 2-(phosphonomethyl)-pentanedioic acid (PMPA) at a final concentration of 1 μM. The cells were washed twice with ice-cold PBS, followed by solubilization with 1 N NaOH and the cell-associated radioactivity was measured using a gamma-counter. Specific binding was plotted against the total molar concentration of the added radiotracer. The Kd values and the concentration of the radiotracers required to saturate the receptors (Bmax) were determined by nonlinear regression using GraphPad (Prism 5 Graph Pad Software, San Diego, CA). For all the cell studies the values are normalized for 1x10 6 cells per well and all data are from two independent experiments with triplicates in each experiment.
For internalization experiments, LNCaP cells were seeded into 6-well plates and treated as described above. Approximately 0.25 pmol of the radiopeptides ( 64 Cu-CC34, 68 Ga-CC34, 111 In-CC34, 68 Ga-HBED-CC-PSMA and 111 In-PSMA-617), were added to the binding buffer and the cells were incubated (in triplicates) for 0.5, 1, 2, 4 and 6 h at 37°C, 5% CO 2 in case of 64 Cu-CC34, 111 In-CC34 and 111 In-PSMA-617 and for 10, 30, 60, 90, 120 and 180 min at 37°C, 5% CO 2 in case of 68 Ga-CC34 and 68 Ga-HBED-CC-PSMA. To determine nonspecific membrane binding and internalization, excess of PMPA (final concentration 1 μM) was added to selected wells. At each time point, the internalization was stopped by removing the medium and washing the cells twice with ice-cold PBS. To remove the receptor-bound radioligand, an acid wash was carried out twice with a 0.1 M glycine buffer pH 2.8 for 5 min on ice. Finally, cells were solubilized with 1 N NaOH. The radioactivity of the culture medium, the receptorbound, and the internalized fractions were measured in a γ-counter.

Metabolic studies
Normal female athymic Balb/c nude mice (2) were injected with approximately 80 pmol / 9 MBq of 68 Ga-CC34 in a total volume of 0.1 mL of PBS and sacrificed 15 min p.i.. Blood was collected in heparinized tubes and centrifuged (5 min, 1,700g) for plasma isolation. Sample of plasma (300 mL) was transferred to an ultrafiltration device (Vivacon 500; 30,000 molecular weight cutoff [Sartorius Stedium Biotech GmbH]), followed by centrifugation (10 min, 9,660g) for the separation of proteins. Samples from the ultrafiltrate and 68 Ga-CC34 solution were analyzed by Reversed-Phase High-Performance Liquid Chromatography (RP-HPLC).

Cell line / Animal model
The PSMA + LNCaP cell line was cultured at 37°C and 5% CO 2 in (RPMI Medium 1640-Gluta-MAX containing 1% FBS, 100 U/mL penicillin and 100 μg/mL streptomycin, sodium-pyruvate 1 mM. Female athymic Balb/C nude mice (age: 4-6 weeks, weight: 17-20 g) were purchased from Janvier, France. For implantation, the tumor cells were harvested by trypsinization and 5x10 6  The organs of interest were dissected and weighted, and the radioactivity in tissue samples was counted in a γ-counter. Biodistribution data are given as percent of injected activity per gram of tissue (% IA/g) and are means ± SD (n = 4).
To demonstrate the specificity of binding, LNCaP mice were co-injected with 10 pmol of the radiotracers along with 20 nmol of PMPA. Animals were sacrificed at 1 h after injection by isoflurane anesthesia.

PET studies in LNCaP xenografts
For the 68 Ga-labeled compounds static images were acquired for a time period of 15 min at 1 and 2 h p.i., and for 64 Cu-CC34 at 1, 4, 24 and 48 h p.i. PET blocking studies of the 68 Ga-and 64 Cu-labeled radioconjougates, were performed as described above and static scans were obtained as previously described. PET-images were corrected for 68 Ga or 64 Cu decay and reconstructed with an ordered-subset expectation maximization algorithm provided by the manufacturer. PET images generated by the AMIDE software.
Tracer uptake is expressed as percentage of decay-corrected %IA/g, with a color scale set from 0% to 250% for qualitative comparison among the images.

Statistical Analysis
All data are expressed as the mean of values±standard deviation (mean±SD). Prism 5 Software (GraphPad Software) was used to determine statistical significance at the 95% confidence level, with a P value of less than 0.05 being considered significantly different.

Radiochemistry / Lipophilicity
The PSMA urea-based conjugates were labeled with 68 Ga, 64 Cu and 111 In with a labeling yield >98%. The specific activities were ranging between 75 and 80 MBq/nmol for the 68 Ga-labeled tracers, approximately 40 MBq/nmol for 64 Cu-CC34 and about 10 MBq/nmol for the 111 Inlabeled radioligands.

Saturation binding / Internalization studies-Metabolic stability
All radiotracers exhibited high affinity for the PSMA + LNCaP cells (Fig 2). 68  Internalization studies showed specific internalization at 37°C and the internalized activity always exceeded the surface bound activity (Fig 3). Approximately 75% of the total cell associated activity was internalized after 3 h of incubation.

Biodistribution in LNCaP xenografts
The biodistribution data are summarized in Tables 1 to 4

Small-animal PET studies
Representative PET images obtained upon injection of 68 Ga-HBED-CC-PSMA and 68 Ga-CC34 in LNCaP xenografts at 1 and 2 h p.i. (Fig 5), with specific tumor and kidney uptake, as shown by the blocking studies at 1 h p.i..
The pharmacokinetics of 64 Cu-CC34 was also determined by small-animal PET (Fig 6) at 1, 4, 24 and 48 h in the same animals. The radiotracer is specifically taken up by the PSMA positive organs at early time points. The PET studies of 64 Cu-CC34 showed good pharmacokinetics, with gradually reduced uptake in kidneys at 24 and 48 h, respectively. The 64 Cu-labeled conjugate showed low liver uptake.

Discussion
Prostate cancer is a complex and biologically heterogeneous disease and therefore cannot be fully assessed with conventional imaging alone. Radionuclide molecular imaging with positron emission tomography (PET) is poised to fill this unmet need through noninvasive detection of the multiple molecular and cellular processes that are active in prostate cancer patients [30].
PSMA is primarily expressed in the human prostate epithelium, salivary and lacrimal glands as well as kidneys with enhanced expression by almost all prostate cancers and further up-regulation in poorly differentiated, metastatic and hormone-refractory carcinomas [31]. These characteristics render PSMA as promising target for prostate cancer imaging and potential therapy. Many efforts have been made especially during the last decade in regard to the development of PSMA-based imaging agents with particular focus on the low molecular weight PSMA inhibitors. A variety of chelators and spacers were introduced to the urea-based PSMA inhibitors and the derived conjugates were labeled with a series of radionuclides for SPECT    and PET imaging [17][18][19][20][21][22][23][24][25][26][27][28][29]. A representative example of this class of radiotracers is the peptidomimetic structure Lys-NH-CO-NH-Glu when coupled to the spacer 6-amino-hexanoic acid (Ahx) and functionalized with the chelator N,N 0 -dis [2-hydroxy-5-(carboxyethyl)benzyl] ethylenediamine-N,N 0 -diacetic acid (HBED-CC) to obtain HBED-CC-Ahx-Lys-NH-CO-NH-Glu (HBED-CC-PSMA). The ability to image PSMA using 68 Ga-HBED-CC-PSMA shows great promise preclinically and clinically [25,32].
Our goal was to develop and evaluate a versatile probe suitable for imaging of PSMA-positive tumors. Thus, (R)-NODAGA(tBu) 3 was conjugated to the well-established Lys-urea-Glu PSMA-inhibitor, through a Phe-Phe-D-Lys(suberoyl) linker [17,18,20,28]. Because of the small volume of the triazacyclononane cage and its ability to coordinate metal ions, NODAGA is particularly attractive for the rapid and stable chelation of 111 In for SPECT, as well as 68 Ga and 64 Cu applied in PET imaging. The choice of the Phe-Phe-Lys(suberoyl) linker has already proven to be successful by Banerjee et al. [28]. CC34 was labeled with 68 Ga, 64 Cu and 111 In, to be used for PET and SPECT imaging and intraoperative applications. Due to its availability from generator systems, the positron emitter 68 Ga has gained rapidly increasing interest in the field of radiopharmaceutical chemistry [33]. 64 Cu is a positron emitter with a longer half-life (12.7 h) compared to 68 Ga (67.8 min) and can give PET-images at later time points with improved tumor-to-background ratios [34]. 64 Curadiopharmaceuticals can be produced centrally and shipped to distant hospitals. 111 In is used as an important SPECT label. Additionally, intraoperative gamma probes are now an important, well-established technology in the management of cancer, particularly in the detection of sentinel lymph nodes.
CC34, being labeled with 68 Ga, 64 Cu and 111 In, showed high affinity towards PSMA on LNCaP cells, with 111/nat In-CC34 exhibiting a significantly increased affinity compared to the 68/nat Ga-and 64/nat Cu-conjugates. In our saturation binding assay, 68/nat Ga-HBED-CC-PSMA revealed the highest affinity towards PSMA while 111/nat In-PSMA-617 was as affine as 111/nat In-CC34.
To figure out if CC34 is subjected to in vivo metabolic degradation and to which extent, RP-HPLC metabolite analysis of plasma samples was performed. Fifteen min p.i. of 68 Ga-CC34 more than 95% of the remaining circulating activity corresponds to intact radiotracer. When the Lys-NH-CO-NH-Glu peptidomimetic structure was coupled to the spacer Phe-Phe-Lysine-suberoyl (L-amino acid spacer) and functionalized with the chelator (1-(1,3-carboxypropyl)-4,7,10(carboxymethyl)-1,4,7,10 tetraazacyclo-dodecane (DOTAGA) [23], rapid in vivo metabolization of the 68 Ga-labeled radiovector was demonstrated. In the same report, the D-amino acids spacer led to an in vivo metabolic stable radiotracer. In our study we proved that only the substitution of L-with D-Lysine in the spacer, resulted in high in vivo stability of 68 Ga-CC34.
The enhanced PSMA affinity, 111/nat In-CC34 and 111/nat In-PSMA-617 resulted in two-to five-fold increased internalization rate and higher but also longer tumor retention compared to 64/nat Cu-and 68/nat Ga-CC34 and 68/nat Ga-HBED-CC-PSMA. Although the binding affinity can be considered as the most crucial parameter which greatly influences the tumor uptake, the overall pharmacokinetic performance of a radiotracer is determined by many other factors which certainly need to be taken into consideration. In particular, parameters such as, lipophilicity, charge, plasma protein binding and molecular weight also influence the pharmacokinetic performance of a radiotracer. CC34 can be considered as a strong evidence of the above assumption. 68 Ga-CC34 and 64 Cu-CC34 demonstrated fast clearance from blood and normal tissues, even kidneys as proved by biodistribution and PET-studies in the case of the copper-64 labeled compounds. The superiority of NODAGA compared to HBED-CC to be efficiently labeled with 64 Cu allows for the conduction of biodistribution/imaging studies at later time points and therefore the complete accomplishment of the pharmacokinetic performance of the radiotracer. 64 Cu-CC34 showed faster wash out from the PSMA positive organs compared to the tumor leading to improved tumor to background ratios. The very low liver uptake of 64 Cu-CC34 at all time points in combination with the short blood circulation need to be pointed out since this is a strong indication of the excellent in vivo stability of the 64 Cu-NODAGA complex. Unfortunately, this was not the case for 111 In-CC34. The affinity of 111/nat In-CC34 towards PSMA was higher compared to 68/nat Ga-CC34 and 64/nat Cu-CC34 by a factor of about 4 and as it was anticipated tumor uptake was higher and longer retention was also observed. Surprisingly, 111 In-CC34 exhibited long blood circulation and highly retained activity in the kidneys even at 48 h p.i.. Unfortunately, this undesired in vivo profile eliminates its applicability as a SPECT imaging agent. The pharmacokinetics of the alanine-containing DOTA-conjugated PSMA inhibitor, PSMA-617, labeled with 111 In, on the other hand, was excellent with impressive tumor to background ratios over time. Protein binding studies (data not shown) 15 min p. i. of 111 In-CC34 and 111 In-PSMA-617 in mice revealed that 83 and 53%, respectively, of the remaining circulating activity was bound to the proteins. This biodistribution profile of 111 In-CC34 cannot be explained by its hydrophilicity (LogD octanol/PBS = -3.32). NODAGA, although is not usual, might form a seven-coordinate complex with 111 In, and if this is the case here, 111 In-CC34 with an additional charge appears to bind strongly to proteins [35]. The seventh coordination site could be occupied by a Lys side chain of a protein. It is worth noting that 177 Lu-PSMA-617, exhibited similar pharmacokinetics compared to 111 In-PSMA-617, which is also characterized by high and retained tumor uptake and almost complete elimination of the radioactivity through kidneys within 24 h [26].
The chelators 1,4,8,11-tetraazabicyclo[6.6.2]-hexadecane-4,11-diacetic acid (CB2-TE2A) and 1,4,7,10-tetraazacyclododecane-N,N 0 ,N@,N‴-tetraacetic acid (DOTA) were used for the functionalization of the Phe-Phe-L-Lys(suberoyl-Lys-urea-Glu). Significantly high in vivo stability, as evidenced by the low liver uptake and fast blood clearance was demonstrated for the 64 Cu-CB2-TE2A-conjugated radiotracer with comparable tumor to blood and tumor to muscles ratios with 64 Cu-CC34 at 1 h p.i.. High liver uptake and slow blood clearance for the 64 Cu-DOTA-conjugated radiotracer was indicative of free Cu(II), which is accumulated in liver [29]. Kidney uptake of the 64 Cu-CB2-TE2A-PSMA-based tracer was two-fold lower compared to 64 Cu-CC34 and very fast wash out was demonstrated. Although the PSMA-mediated renal uptake has been shown to be specific by us and others, it is interesting that variations are observed not only in regard to the absolute kidney uptake but also as far as the renal elimination concerns.

Conclusions
Our data document the versatility of the (R)-NODAGA-functionalized PSMA-based radiotracer, CC34, labeled with 68 Ga and 64 Cu, as a highly specific targeted PET imaging agent. In particular, the favorable pharmacokinetic performance of the 64 Cu-labeled radiotracer renders it a promising tracer that would enable sufficient imaging of PSMA positive tissues which are characterized by elevated expression of PSMA or in the case that delayed imaging is required. This tracer is currently prepared for clinical translation.
Supporting Information S1 Appendix. Reagents and Instrumentation.