Influence of charge on cell permeability and tumor imaging of GPR30-targeted ¹¹¹In-labeled non-steroidal imaging agents

Abstract

Recent clinical studies implicate the role of G protein-coupled estrogen receptor, GPR30 in aggressive forms of breast, ovarian and endometrial cancers. However, the functional role of GPR30 at cellular and molecular level remains less clear and controversial, particularly its subcellular location. The primary objective of this study was to develop radiolabeled neutral and charged GPR30-targeted non-steroidal analogues to understand the influence of ligand charge on cell binding, cellular permeability and in vivo tumor imaging. Therefore, we developed a series of GPR30-targeted ^111/113In(III)-labeled analogues using macrocyclic and acyclic polyamino-polycarboxylate chelate designs that would render either a net negative or neutral charge. In vitro biological evaluations were performed to determine the role of negatively charged analogs on receptor binding and activation using calcium mobilization and phosphoinositide 3-kinase assays. In vivo evaluations were performed on GPR30-expressing human endometrial Hec50 tumor-bearing mice to characterize the biodistribution and potential application of GPR30-targeted imaging agents for translational research. In vitro functional assays revealed an effect of charge, such that only the neutral analogue activated GPR30-mediated rapid signaling pathways. These observations are consistent with expectations for initial rates of membrane permeability and suggest an intracellular rather than the cell surface location of functional receptor. In vivo studies revealed receptor-mediated uptake of the radiotracer in target organs and tumors; however, further structural modifications will be required for the development of future generations of GPR30-targeted imaging agents with enhanced metabolic properties and decreased non-specific localization to the intestines.

INTRODUCTION

Estrogen plays an important role in a variety of normal physiological and pathological processes. Estrogen exerts many of its effects via the two known nuclear estrogen receptors (ERs), ER〈 and ER® (1). In 2000, GPR30, an orphan seven transmembrane G protein-coupled receptor (GPCR) was shown to be involved in estrogen-mediated activation of ERK1/2 in cells lacking ERα but expressing GPR30, implicating a role for GPR30 in rapid estrogen-mediated cellular responses (2). In 2005, two studies reported estrogen binding to GPR30 and further characterized relationships between GPR30 expression and rapid estrogen-mediated signaling events (3, 4). Novel fluorescent estrogen derivatives (E2-Alexas) were used to examine the cellular and sub-cellular localization of GPR30 using confocal microscopy (3). The microscopy studies revealed that E2-Alexas detected ER〈 and ER® in the nucleus of the cells, whereas GPR30 was predominantly located in the endoplasmic reticulum with no detectable signal at the plasma membrane. Most GPCRs are localized in the plasma membrane; therefore, to investigate this unexpected observation further, we developed a novel class of 17〈-substituted small-molecule estrogen derivatives exhibiting differential cell permeability (5). This new class of estrogen derivatives revealed that positively charged cell-impermeable molecules did not activate rapid GPR30 or ER signaling, whereas the neutral cell-permeable molecules rapidly activated both ER and GPR30 in cell-based functional assays (5). These results confirmed the predominantly intracellular location of functional GPR30, as indicated by previous confocal microscopy studies (3). However, other studies have reported the presence of GPR30 on the cell surface, suggesting that a small fraction of total cellular receptor may be present at the cell membrane as a result of limited export and/or rapid constitutive internalization (4, 6). Nevertheless, it appears that estrogen must traverse the plasma membrane for functional activity of GPR30.

We have previously described a GPR30-selective agonist, G-1 (1) and recently a structurally related GPR30-selective antagonist, G15 (2) (Figure 1) (7, 8). Significant biological roles of GPR30 in physiological and pathological processes have been revealed using these GPR30-selective ligands (9-12). GPR30 knockout mice (in some cases in conjunction with GPR30-selective ligands) are also beginning to address the functions of GPR30 in vivo (13, 14). Recent clinical studies indicate that GPR30 is expressed in and associated with aggressive forms of breast, endometrial and ovarian cancer that display low survival rates (15-17).

Figure 1.

Chemical structures of parent and functionalized GPR30-targeted analogues: GPR30-agonist G-1 (1); GPR30-antagonist G15 (2); functionalized GPR30 derivative G-amine (3); GPR30-targeted analogue based on an acyclic polyamino-polycarboxylate DTPA chelate design: G-Bz-DTPA (4) and GPR30-targeted analogues based on macrocylic polyamino-polycarboxylate DOTA chelate designs: G-Bz-DOTA (5) and G-DOTA (6).

Estrogen has been a widely studied radiopharmaceutical target over the past 35 years. The most successful radiolabeled estrogen derivative, 16α-¹⁸F-17®-estradiol (FES) is well characterized in patients with breast, uterine, ovarian and endometrial cancer (18-21). FES has been used clinically with promising results in imaging ER-expressing tumors and to evaluate responsiveness of tumors to anti-estrogen drugs (18, 22, 23). To study the in vivo distribution and role of GPR30, we have developed GPR30-selective radioiodinated derivatives (24). However, these radioiodinated derivatives were unsuitable for in vivo use due to rapid metabolism, high lipophilicity and poor in vivo targeting characteristics.

ER-targeted imaging agents employing macrocyclic and acyclic polyamino-polycarboxylate chelates such as DOTA and DTPA have been previously reported (25-28). An aspartic acid-linked 4-hydroxytamoxifen-DTPA ligand demonstrated receptor-specificity; however, the binding affinity was 10-fold lower than that of tamoxifen (25). Another DTPA-tamoxifen analogue with similar affinity to tamoxifen was evaluated for imaging ER-positive lesions (27). Biodistribution, autoradiography, and radionuclide imaging demonstrated in vivo receptor-specificity of the ¹¹¹In-labeled DTPA-tamoxifen conjugate and tumors could be clearly visualized even after 48 h (27). In another report, estradiol was labeled with ¹⁷⁷Lu using p-SCN-DOTA as the chelating agent and in vitro cell binding studies demonstrated receptor-specificity (26). In spite of the success of polyamino-polycarboxylate chelate designs for ER-targeting, a major concern is the possibility that a net charge can hinder receptor-binding kinetics resulting from limited cell permeability.

We have previously shown that a net positive charge can hinder cell permeability of a derivatized estrogen molecule in short term experiments (seconds to minutes), although the ligand retains receptor-specificity and can activate cell signaling over prolonged time periods (hours) (5). The goals in the current study were to understand the influence of charge of the chelate ligand on cell binding, permeability and tumor imaging. Therefore, we conjugated the tetrahydro-3H-cyclopenta[c]quinoline amine (3) based on the chemical scaffold of the novel GPR30-selective ligands G-1 and G15, with p-SCN-DTPA, p-SCN-DOTA and DOTA to provide the corresponding aminocarboxylate chelate ligands (4-6). These chelates were labeled with ^111/113In and evaluated using cell-based ligand-binding and functional biological assays to characterize the role of anionic charged chelate complexes [¹¹¹In-G-Bz-DTPA]²⁻, [¹¹¹In-G-Bz-DOTA]⁻ on receptor binding and activation in comparison with a neutral complex [¹¹¹In-G-DOTA]. We then investigated the use of ¹¹¹In-labeled GPR30-targeted analogues to detect GPR30 in vivo using radioisotope-imaging modalities. In this report, we describe radiochemistry experiments (examining the role of pH and incubation time on labeling efficiency), cell permeability and GPR30 functional assays (to assess ligand-chelate activity) as well as in vivo biodistribution and imaging studies on GPR30-expressing human endometrial cancer-bearing female mice.

RESULTS AND DISCUSSION

In this study, we developed neutral and negatively charged GPR30-selective indium-labeled polyamino-polycarboxylate compounds for in vivo targeting of GPR30 and potential use as cancer diagnostic and therapeutic agents. We designed this series of aminocarboxylate chelates based on the tetrahydro-3H-cyclopenta[c]quinoline scaffold that we have previously demonstrated binds to GPR30 with high selectively. The removal of the ethanone functional group at the C8 position of G-1, yielding G15, changes the GPR30-mediated activity from agonism to antagonism, while maintaining strong relative binding affinity for GPR30 (8). Therefore, we anticipated that introduction of an alkylamine linkage at this C8 position would provide a derivative suitable for conjugation with amine-reactive chelates derivatives such as DOTA-NHS, p-SCN-Bz-DOTA, and p-SCN-Bz-DTPA to provide the desired GPR30-targeted aminocarboxylate chelates.

Synthetic chemistry

The pendant aminoethyl derivative (G-amine, 3) was prepared in 95% yield from the Sc(OTf)₃ catalyzed Povarov cyclization of 6-bromopiperonal, cyclopentadiene and the ^tBoc-protected aminophenethylamine in acetonitrile, followed by deprotection with TFA/CH₂Cl₂ as previously described (8) and shown schematically in Suppl. Figure 1. The G-amine derivative was then coupled with the activated phenylisothiocyanate groups attached to the alkyl backbone of the aminocarboxylate chelates p-SCN-Bz-DTPA and p-SCN-Bz-DOTA to provide the thiourea-linked derivatives displaying free penta- and tetra-acetic acid groups respectively (4 and 5) (Suppl. Figure 2). These derivatives were isolated by precipitation with ethanol and filtration, followed by water washes to provide the pure solid derivatives. The mono-N-hydroxysuccinimide ester of DOTA was coupled with the G-amine derivative in dry DMF to provide the tri-acetic acid G-DOTA derivative (6) (Suppl. Figure 2). The connecting linkage of this G-DOTA chelate lacks the hydrophobic phenyl group that is present in the other derivatives, and therefore exhibited increased water solubility. The G-DOTA derivative was purified by reverse phase chromatography using a C18 column eluted with methanol/water (60:40).

Radiochemistry

After incubation at room temperature for 60 min., the labeling efficiency and radiochemical purity for ¹¹¹In-labeled G-Bz-DTPA were over 99% and 95%, respectively, as determined by ITLC and HPLC (Suppl. Figure 3a). Over 99% incorporation yield was obtained for ¹¹¹In-labeled G-Bz-DOTA and G-DOTA, when the reaction mixture was heated at 80°C for 20 min at pH 4.5. The radiochemical purities as assessed by HPLC of ¹¹¹In-labeled G-DOTA (Suppl. Figure 3b) and ¹¹¹In-labeled G-Bz-DOTA (Suppl. Figure 3c) were over 95%. The HPLC analysis revealed the presence of two predominant isomeric complexes of the ¹¹¹In-labeled G-DOTA and ¹¹¹In labeled G-Bz-DOTA compounds as seen in the representative chromatograms, while a single DTPA construct was observed under the reverse-phase chromatographic conditions used. The specific activity ranged from 35-60 MBq nmol⁻¹ for all the derivatives. However, when the labeling reaction was performed at 100°C for the DOTA derivatives, a degradation product was obtained decreasing the radiochemical purity to as low as 70%. Very low incorporation yields for the DOTA derivatives were obtained when the reaction was carried out at 60°C with an incubation time of 20 min.

Stability and solubility

The ¹¹¹In-labeled G-DOTA, G-Bz-DOTA and G-Bz-DTPA derivatives exhibited good stability (over 90%) after incubation at 37°C for up to 4 days in PBS (pH 7.4) or mouse plasma. The ¹¹¹In-labeled G-Bz-DTPA would be expected to form a monomeric dianionic (2-) charged complex with In(III) at physiological conditions, and exhibited a correspondingly low log P_(o/w) value of 1.5 ± 0.1. The macrocyclic DOTA derivatives were significantly more lipophilic than the acyclic DTPA derivative, such that the neutral complex G-DOTA exhibited a log P_(o/w) value of 4.8 ± 0.3, while the monoanionic benzyl-thiourea linked complex G-Bz-DOTA exhibited an increased log P_(o/w) value of 5.2 ± 0.3.

Role of incubation time and pH on labeling efficiency

To study the role of incubation time on incorporation yield and labeling efficiency, G-DOTA and G-Bz-DOTA were incubated with ¹¹¹InCl₃ at 80°C and pH 5. The labeling kinetics were more rapid for G-Bz-DOTA than for G-DOTA (Figure 2a). After 10 min incubation, over 94% of the ¹¹¹In was incorporated into G-Bz-DOTA as compared to 80% for G-DOTA. The reduced rate of coordination to form the ¹¹¹In-G-DOTA complex compared with the ¹¹¹In-G-Bz-DOTA species may result from increased steric contraints and limited conformational flexibility associated with the direct carboxamide linkage between the DOTA-nitrogen and the GPR30-targeting quinoline derivative. Monoclonal antibodies have been successfully radiolabeled with ¹¹¹In at pH 7 and higher with incorporation yields of over 90% (29, 30). However, for hydrophobic small molecules, G-DOTA and G-Bz-DOTA, the best incorporation yields were achieved between pH 4-6, which may be related to the isoelectric point of the agent, offering better solution stabilization and thermodynamics at pH 4-6 (29).

Figure 2.

Labeling efficiency of G-DOTA and G-Bz-DOTA. Influence of (a) incubation time and (b) pH on ¹¹¹In labeling efficiency of G-DOTA and G-Bz-DOTA.

To examine the role of pH, G-DOTA and G-Bz-DOTA were incubated with ¹¹¹In at 80°C for 20 min from pH 4 to 8. Optimum radiolabeling (over 95% incorporation) was achieved between pH 4-6 (Figure 2b). At pH 8, incorporation yields were as low as 25%. At pH values >5, G-DOTA labeling was more dependent on pH than G-Bz-DOTA, presumably due to differences in pKa, deprotonation rates and electrostatic repulsion between these two derivatives (31).

In vitro cellular activity

GPR30 activation by either estrogen or G-1 has been shown to initiate a number of signal transduction cascades in cells. These include intracellular calcium mobilization and activation of a number of kinases, such as epidermal growth factor receptor (EGFR), mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase (PI3K). In order to investigate the functional properties of the chelate series on endogenous GPR30 expressing cells, the corresponding stable isotope ¹¹³In-labeled derivatives were assessed for their ability to mobilize intracellular calcium in GPR30-positive, ERα/β-negative SKBr3 cells. The ¹¹³In-labeled derivatives carrying a net negative charge, ¹¹³In-G-Bz-DOTA and ¹¹³In-G-Bz-DTPA, were unable to significantly mobilize calcium in these cells within a timeframe (3 min) consistent with rapid signaling whereas the neutral ¹¹³In-G-DOTA compound was effective in a timeframe comparable to G-1 (maximal response in 15-20 sec) (Figure 3). This result suggests that, even applied at a higher concentration than their neutral counterparts, the net negative charge associated with these compounds prevents their rapid entry into the cell and thus their rapid signaling via GPR30.

Figure 3.

Mobilization of intracellular calcium by ¹¹³In-labeled derivatives. The effects of ¹¹³In-labeled derivatives and known GPR30 agonists were evaluated in indo1-AM-loaded SKBr3 cells. Calcium mobilized by 50 nM 17®-estradiol at 3 min post stimulation was defined as 100% and the effects of other ligands compared to this response. All data represent the mean ± s.e.m. from three independent experiments.

Additionally, the ¹¹³In-labeled derivatives were used to assess PI3K activation utilizing an assay employing the PIP3-binding pleckstrin homology (PH) domain of AKT fused to mRFP (PH-RFP) to localize PIP3 production and accumulation in SKBr3 cells. Consistent with our findings in the calcium mobilization assay, stimulation of cells with ¹¹³In-G-DOTA for 15 min resulted in the accumulation of the PH-RFP reporter in the nucleus, as has been observed for estrogen and G-1, indicative of GPR30-medated PI3K activation (Figure 4a). Interestingly, the ¹¹³In derivatives with net negative charges, ¹¹³In-G-Bz-DOTA and ¹¹³In-G-Bz-DTPA, were able to elicit a weak PI3K activation when administered to cells at higher doses (1 μM, Figure 4b) but no PI3K activation was detected when cells were stimulated with low doses (10 nM) of these compounds. This result is consistent with our previous finding that charged compounds at high concentrations can activate PI3K via GPR30 when the cells are stimulated for longer times than in the calcium mobilization assay (3 min vs 15 min) (5). This may be due to either the slow entry of a low concentration of compound into the cytoplasm, where it can stimulate GPR30, or due to a small fraction of GPR30 present at the cell surface at some time during cell stimulation, although the latter is less likely due to the lack of calcium mobilization in these same cells over a 3 min time interval.

Figure 4.

GPR30-mediated activation of PI3K by ¹¹³In-labeled derivatives. The activity of ¹¹³In-labeled derivatives was evaluated using SKBr3 cells transfected with the Akt-PH-mRFP1 reporter. 17®-estradiol, ¹¹³In-G-DOTA, ¹¹³In-G-Bz-DOTA and ¹¹³In-G-Bz-DTPA were used at 10 nM (a) or 1 μM (b). Data are representative of three independent experiments.

Radioligand binding assays using intact GPR30-expressing, ER〈/®-negative Hec50 cells show that the ¹¹³In-G-DOTA compound binds to GPR30 with an IC₅₀ of 33.9 ± 2.4 nM. To evaluate the possible non-selective binding of ¹¹³In-labeled derivatives to the classical estrogen receptor ER〈, binding assays using ERα-transfected COS7 cells were performed. No detectable binding of any of the three ¹¹³In-labeled derivatives to ERα was observed (data not shown) indicating that the In-labeled derivatives are highly selective for GPR30.

Biodistribution in Hec50 tumor-bearing animals

Based on the in vitro analysis, biodistribution was performed in Hec50 endometrial tumor-bearing animals by injecting 1.8 MBq of ¹¹¹In-G-DOTA in the tail vein of conscious mice. ¹¹¹In-G-DOTA demonstrated moderate tumor/blood and tumor/muscle ratios and rapid clearance. The tumor/blood ratio of 0.71 at 1 h PI increased to 2.08 at 4 h PI; similarly, the tumor/muscle ratio of 3.55 at 1 h PI increased to 6.00 at 4 h PI. The liver and the intestines exhibited the highest uptakes values (Table 1), although the uptake values in liver rapidly decreased over time (Figure 5a). Due to the hepatobiliary excretion, an opposite trend was observed in the intestines where the values increased over time (Figure 5a). The observed pharmacokinetic trend in the liver, gall bladder, kidney, urinary bladder and the intestines suggests hepatobiliary and urinary excretion and possible metabolism. Hydrophilic radiometabolites were observed in chromatograms of the collected urine samples (Suppl. Figure 4). Over 60% of the injected dose was excreted via the urinary and hepatobillary systems within 4 h PI and the calculated half-life (t_1/2) was 2.9 h. At 0.5 and 1 h PI, high uptake was observed in non-specific target organs such as stomach, bone and lungs; however, these values decreased over time (Table 1 and Figure 5a). A similar trend was observed in the tumors, where uptake decreased from 3.81 % ID/g to 0.48 % ID/g between 0.5 h and 4 h PI. Receptor-mediated uptake, determined by co-injection with excess unlabelled G-1, was observed in adrenal glands in addition to uterus, mammary tissue and the Hec50 tumor (Figure 5b). The uptake values of the radiotracer in the uterus, mammary glands, tumor and adrenal glands were significant when compared to the uptake values in the presence of excess unlabeled G-1.

Table 1. Time-dependent biodistribution of ¹¹¹In-G-DOTA.

Biodistribution of the ¹¹¹In-G-DOTA derivative in selected organs of ovariectomized female athymic (NCr) nu/nu mice bearing GPR30-expressing human endometrial Hec50 tumors.

Organs	0.5 hr PI	1 hr PI	2 hr PI	2 hr PI (Block- 5 μg G-1)	4 hr PI
Heart	5.87 ± 0.55a	1.46 ± 0.19	0.25 ± 0.04	0.19 ± 0.03	0.17 ± 0.02
Blood	3.94 ± 0.54	2.22 ± 0.34	1.21 ± 0.63	1.11 ± 0.17	0.23 ± 0.02
Lungs	4.98 ± 0.84	2.50 ± 0.44	0.80 ± 0.11	1.00 ± 0.23	0.48 ± 0.10
Liver	14.28 ± 1.50	12.04 ± 1.43	3.87 ± 0.41	4.98 ± 0.71	3.04 ± 0.73
Spleen	0.76 ± 0.06	1.04 ± 0.25	0.26 ± 0.07	0.91 ± 0.71	0.36 ± 0.10
Large intestine	1.29 ± 0.28	8.32 ± 1.03	15.27 ± 4.78	20.12 ± 3.19	8.48 ± 0.67
Small intestine	11.13 ± 0.84	19.43 ± 5.22	6.12 ± 3.96	2.32 ± 0.76	1.25 ± 0.11
Stomach	6.80 ± 0.99	5.05 ± 1.75	0.61 ± 0.18	0.61 ± 0.35	0.30 ± 0.08
Kidneys	2.55 ± 0.30	3.95 ± 0.43	1.62 ± 0.53	1.00 ± 0.16	1.74 ± 0.46
Adrenals	7.11 ± 0.51	5.02 ± 0.86	1.33 ± 0.08	0.55 ± 0.14	0.57 ± 0.08
Bone	1.32 ± 0.21	0.85 ± 0.39	0.17 ± 0.05	0.21 ± 0.08	0.12 ± 0.02
Muscle	0.69 ± 0.10	0.43 ± 0.11	0.13 ± 0.06	0.15 ± 0.03	0.08 ± 0.02
Uterus	1.46 ± 0.23	0.73 ± 0.09	0.69 ± 0.08	0.31 ± 0.06	0.66 ± 0.03
Mammary	1.16 ± 0.32	0.72 ± 0.24	0.44 ± 0.03	0.16 ± 0.02	0.50 ± 0.07
Pituitary	0.88 ± 0.06	0.60 ± 0.05	0.04 ± 0.03	0.15 ± 0.10	0.01 ± 0.00
Brain	0.33 ± 0.01	0.16 ± 0.08	0.06 ± 0.01	0.11 ± 0.06	0.05 ± 0.02
Urinary bladder	4.13 ± 0.39	1.50 ± 0.21	0.15 ± 0.05	0.50 ± 0.16	0.36 ± 0.16
Hec50Tumor	3.81 ± 0.18	1.58 ± 0.41	0.60 ± 0.03	0.44 ± 0.04	0.48 ± 0.02

^aUptake values are expressed as % ID/g. Data represent the mean value ± SEM from at least 3 determinations.

Figure 5.

Biodistribution of ¹¹¹In-labeled G-DOTA. (a) Time-activity curve and (b) in vivo GPR30-specificity of ¹¹¹In-labeled G-DOTA in selected organs of ovariectomized female athymic (NCr) nu/nu mice bearing GPR30-expressing human endometrial Hec50 tumors. All uptake values are expressed as % ID/g. Data represent the mean value ± SEM from at least 3 determinations.

Imaging studies

Imaging studies were carried out after injecting 12.9 MBq ¹¹¹In-G-DOTA via the tail vein of Hec50 tumor bearing mice. Whole body 60 s/projection imaging studies were carried out under 1.5-1.7% isoflurane using a temperature controlled bed (36-38°C). Imaging studies revealed very high consistent activity in the intestines (Suppl. Figure 5a). Upon image quantification, the liver uptake decreased from 14.48% ID (1 h PI) to 3.13% ID (4 h PI); whereas the intestine uptake increased from 64.91% ID (1 h PI) to 83.88% ID (4 h PI). Tumors were visualized at 1 h PI; however, the non-specific activity was very high, particularly in the bladder and the intestine (Suppl. Figure 5a). Nevertheless, activity in shoulder tumors (Suppl. Figure 5b) and flank tumors (Suppl. Figure 5c) was clearly visualized in transverse slices. At the same image threshold, the tumors were better visualized when the bladder was cropped from the image (Suppl. Figure 5d). Without decreasing the image threshold, shoulder tumors were clearly visualized on 200 s/projection focused studies (Figure 6), demonstrating the uptake of the radiotracer in the tumor. It is often desirable for a tracer to rapidly wash out from the non-target organs and retain activity in the target organs for better target/background ratios. Although ¹¹¹In G-DOTA was rapidly washed out from non-target organs, the activity was not retained in the target organs. Metabolism of the imaging agent that results in cleavage of the linkage, producing a metabolite that is excreted by an active transport pathway, or otherwise significantly decreases GPR30 binding affinity, may be implicated. Based on these results, ¹¹¹In G-DOTA does not represent an ideal imaging agent. Nevertheless, valuable information was obtained from the animal studies of these first generation G-polyamino-polycarboxylate acyclic and macrocyclic derivatives. Subsequent development should address structural modifications that increase the uptake in target tissues, maximize residence time and decrease metabolism and non-target organ uptake.

Figure 6.

In vivo imaging of ¹¹¹In-labeled G-DOTA. Focused SPECT/CT 1h PI images of ¹¹¹In-labeled G-DOTA in female in selected organs of ovariectomized female athymic (NCr) nu/nu mice bearing GPR30-expressing human endometrial Hec50 tumors. (a) Maximum intensity projection, (b) coronal slice of the tumor xenograft implanted on the shoulder and (c) transverse slice of the tumor xenograft.

CONCLUSION

In this work, we have successfully synthesized and evaluated the first generation of non-steroidal ¹¹¹In-labeled GPR30-targeted analogues for cancer imaging. The ¹¹³In-labeled analogues demonstrated the intracellular functionality of GPR30 in whole cell-based assays. The neutral ¹¹³In-G-DOTA complex stimulated calcium mobilization consistent with rapid signaling via GPR30, while both anionic charged species failed to elicit a response. Analogous results were obtained using PI3-kinase activation assays, although the charged species [¹¹³In-G-Bz-DOTA]⁻ and [¹¹³In-G-Bz-DTPA]²⁻ could elicit weak responses at very high doses (1 μM). These results are consistent with previous findings and expectations for reduced membrane permeability of the charged species. The GPR30 binding affinity IC₅₀ of 33.9 ± 2.4 nM of the neutral ¹¹³In-G-DOTA complex was comparable to the similar small molecules G-1 (~11 nM) and G15 (~20 nM) from which the targeting agent was derived, and indicates that the presence of the neutral DOTA chelate structure at the C8 position does not significantly compromise receptor binding interactions. The radiochemistry experiments demonstrated efficient metal incorporation in the acyclic and macrocyclic polyaminocarboxylate chelate derivatives conjugated to a small, hydrophobic targeting agent. The biodistribution and imaging studies with ¹¹¹In-G-DOTA revealed moderate tumor/blood and tumor/muscle ratios, although unfavorable in vivo targeting characteristics and rapid clearance from the tumor were observed with this agent. Our results suggest hepatobiliary and urinary excretion and formation of polar metabolites were problematic. These studies provide a foundation for the further development of GPR30 targeted imaging agents.

METHODS

Details of chemical synthesis, purification and characterization are provided in Supporting Information.

Cell culture

ER 〈/®-negative and GPR30-expressing human endometrial carcinoma Hec50 cells, ER 〈/®-negative and GPR30-expressing human breast carcinoma SKBr3 and ER 〈/® and GPR30-negative monkey kidney Cos-7 cells were cultured in DMEM tissue media (Hec50, Cos7) or RPMI-1640 tissue media (SKBr3), with fetal bovine serum (10%) and 100 units/mL penicillin and 100 µg/mL streptomycin. Cells were grown as a monolayer at 37°C, in a humidified atmosphere of 5% CO₂ and 95% air. Lipofectamine2000 was used according to manufacturer’s directions for all transfections.

Intracellular calcium mobilization

SKBr3 cells (1 × 10⁷/mL) were incubated in HBSS containing 3 μM Indo1-AM (Invitrogen) and 0.05% pluronic acid F-127 for 1 h at RT. Cells were then washed twice with HBSS, incubated at RT for 20 min, washed again with HBSS, resuspended in HBSS at a density of 10⁸cells/mL and kept on ice until assay, performed at a density of 2 × 10⁶ cells/mL. Ca²⁺ mobilization was determined ratiometrically using λ_ex 340 nm and λ_em 400/490 nm at 37°C in a spectrofluorometer (QM-2000-2, Photon Technology International) equipped with a magnetic stirrer.

PI3K activation

The PIP3 binding domain of Akt fused to mRFP1 (PH-RFP) was used to localize cellular PIP3. SKBr3 (transfected with PH-RFP) were plated on coverslips and serum starved for 24 h followed by stimulation with ligands as indicated. The cells were fixed with 2% PFA in PBS, washed, mounted in Vectashield containing DAPI (Vector Labs) and analyzed by confocal microscopy using a Zeiss LSM510 confocal fluorescence microscope.

Receptor-binding studies

To evaluate ligand binding to GPR30 expressed in Hec50 cells, competition binding was performed on selected ¹¹³In-labeled GPR30-targeted derivative with a radioiodinated GPR30-targeted analogue (8). Hec50 cells (50-60% confluent) were washed twice with PBS solution and the medium was replaced with phenol-red free medium containing 10% charcoal stripped fetal bovine serum and incubated overnight. Cells were washed three times with PBS and trypsinized. Approximately 75,000 cells/plate were plated in 24 well plates and incubated overnight in phenol red-free, charcoal-stripped serum-containing medium. On the day of the experiment, the plated cells were washed twice with PBS and increasing doses of ¹¹³In-labeled derivatives were added and incubated for 30 min. After the 30 min incubation period, ¹²⁵I-labeled GPR30-targeted analogue was added to each well and incubated for 1 h at 37°C. Cells were washed three times with PBS to remove non-specifically bound ¹²⁵I derivative. To extract the cells, ethanol was added to each well and incubated at 37°C for 5 min. The radioactivity associated with the collected extract was then counted using a Wallac Wizard 1480 automatic gamma counter. To determine nonspecific binding, the cells were incubated with 10 µM G-1. The IC₅₀ values were determined using in-built non-linear regression analysis in GraphPad Prism version 5 software (San Diego, CA, USA).

Binding assays for ERα were performed as previously described (Revankar, 2005). Briefly, Cos7 cells were transiently transfected with ERα-GFP. Following serum starvation for 24 h, cells (~5×10⁴) were incubated with ¹¹³In-labeled derivatives for 10 min. in a final volume of 10 µL prior to addition of 10 μL 20 nM E2-Alexa633 in saponin-based permeabilization buffer. Following 5 min at RT, cells were washed once with 1 mL PBS/2%BSA, resuspended in 200 μL and analyzed on a FACS Calibur flow cytometer (BD Biosciences).

Animal and tumor models

The human endometrial carcinoma Hec50 tumor model was developed by injecting 3-4 million Hec50 cells subcutaneously in 8 week-old female ovariectomized athymic (NCr) nu/nu mice (Harlan Inc., Indianapolis, IN, USA). After 4-6 weeks, tumors ranging from 0.5-0.7 cm in diameter were observed.

Biodistribution and SPECT/CT imaging studies

Conscious Hec50 tumor-bearing mice were injected intravenously (tail vein) with selected ¹¹¹In-labeled derivatives. To determine GPR30-receptor specificity, 5 μg G-1 was premixed with the radiotracer and the two compounds were co-injected intravenously. At the desired time points, the animals were sacrificed by CO₂ inhalation. Urine and blood samples were collected for metabolism studies. After sacrificing the animals, organs were carefully removed and isolated to determine the biodistribution characteristics of the tracer. The organ samples were weighed and the corresponding radioactivity was measured using an automated gamma counter after verifying the counting efficiency with standards. The percent-injected dose per gram of tissue (%ID/g) was calculated by comparison with standards representing the injected dose per animal.

NanoSPECT/CT Imaging studies were performed using a multi-pinhole NanoSPECT/CT small animal imager (Bioscan Inc, Washington DC, USA). Whole body imaging studies were carried out using 1.5-1.7% isofluorane on a temperature controlled bed. All animal experiments were conducted in compliance with the guidelines and approved protocols established by the UNM Institutional Animal Care and Use Committee.

Statistical Analysis

All numerical data were expressed as the mean of the values ± the standard error of mean (s.e.m). GraphPad Prism version 5 (San Diego, CA) was used for statistical analysis and a P value less than 0.05 was considered statistically significant.