^99mTc-labeling of Peptidomimetic Antagonist to Selectively Target α_vβ₃ Receptor-Positive Tumor: Comparison of PDA and EDDA as co-Ligands

Abstract

Objectives

The aim of this research was to synthesize radiolabeled peptidomimetic integrin α_vβ₃ antagonist with ^99mTc for rapid targeting of integrin α_vβ₃ receptors in tumor to produce a high tumor to background ratio.

Methods

The amino terminus of 4-[2-(3,4,5,6-tetra-hydropyrimidin-2-ylamino)-ethyloxy]benzoyl-2-(S)-[N-(3-amino-neopenta-1-carbamyl)]-aminoethylsulfonyl-amino-β-alanine hydrochloride (IAC) was conjugated with N-hydroxysuccinimide ester of HYNIC and labeled with ^99mTc using tricine with either 1,5-pyridinedicarboxylic acid (PDA) or ethylenediamine-N,N′-diacetic acid (EDDA) as the co-ligand. The products, ^99mTc EDDA₂/HYNIC-IAC (P1) and ^99mTc PDA (tricin)/HYNIC-IAC (P2) were subjected to in vitro serum stability, receptor-binding, biodistribution and imaging studies.

Results

P1 and P2 were synthesized with an overall yield of >80%. P1 was slightly more stable than P2 when incubated in serum at 37 °C for 18 hrs (84 vs 77% intact). The In vitro receptor-binding of P1 was higher than that of P2 (78.02 ± 13.48 vs 51.05 ± 14.05%) when incubated with α_vβ₃ at a molar excess (0.8 μM). This receptor binding was completely blocked by a molar excess of an unlabeled peptidomimetic antagonist. Their differences shown in serum stability and the receptor-binding appeared to be related to their biological behaviors in tumor uptake and retention; the 1 h tumor uptakes of P1 and P2 were 3.17±0.52 and 2.13±0.17 % ID/g, respectively. P1 was retained in the tumor longer than P2. P1 was excreted primarily through the renal system whereas P2 complex was excreted equally via both renal and hepatobiliary systems. Thus, P1 was retained in the whole-body with 27.25 ± 3.67% ID at 4 h whereas 54.04 ± 3.57% ID of P2 remained in the whole-body at 4 h. This higher whole-body retention of P2 appeared to be resulted from a higher amount of radioactivity retained in liver and intestine. These findings were supported by imaging studies showing higher tumor-to-abdominal contrast for P1 than for P2 at 3 h postinjection.

Conclusions

P1 showed good tumor targeting properties with a rapid tumor uptake, prolonged tumor retention and fast whole-body clearance kinetics. These findings warrant further investigation of the HYNIC method of ^99mTc labeling of other peptidomimetic antagonists using EDDA as a coligand.

INTRODUCTION

Integrin α_vβ₃ is a heterodimeric transmembrane glycoprotein and a receptor for extracellular matrix proteins including vitronectin, fibronectin and fibrinogen which contain an arginine-glycine-aspartic acid (RGD) sequence [1, 2]. Integrin α_vβ₃ is overexpressed in tumor induced angiogenic vessels and various malignant human tumors. To assess angiogenesis and receptor status in tumors by selective targeting of this receptor, various RGD peptides and peptidomimetic antagonists have been labeled with gamma and positron emitters for scintigraphic detection [2–9] and beta emitters for radiotherapy of tumors [10]. Steady progress has been reported in optimizing the labeling methodologies. To increase tumor-to-non-tumor tissue ratios, especially for the tumor-to-liver ratio, the organ clearance pharmacokinetics was optimized by increasing the hydrophilicity of the product via glycosylation and PEGylation [11–14]. Dimers and tetramers of cyclic RGD derivatives have been synthesized to increase the receptor binding affinity, thereby improving the tumor targeting and retention [3, 15, 16].

We have investigated the use of peptidomimetic antagonists, 4-[2-(3,4,5,6-tetrahydropyrimidine-2-ylamino)ethyloxy]benzoyl-2-(S)-aminoethylsulfonyl-amino-β-alanine (IA) and 4-[2-(3,4,5,6-tetrahydropyrimidin-2-ylamino)-ethyloxy]benzoyl-2-(S)-[N-(3-aminoneopenta-1-carbamyl)]-aminoethylsulfonylamino-β-alanine hydrochloride (IAC), which is a higher affinity carbamate derivative of a first generation peptidomimetic integrin α_vβ₃ antagonist, IA, for molecular imaging of tumors via binding to α_vβ₃ receptor. We previously labeled IA and IAC with ¹¹¹In using 2-(p-isothiocyanatobenzyl)-DOTA as a bifunctional chelator and investigated comparative in vivo studies in nude mice bearing the receptor-positive M21 tumors [17]. The comparative biodistribution and imaging studies indicated that both ¹¹¹In labeled-IA and -IAC using DOTA as a bifunctional chelator were primarily excreted via the renal system and a minor portion of the labeled injectants was excreted via the hepatobiliary route. The biodistribution study also showed that the ¹¹¹In labeled-IAC with a higher affinity produced higher tumor uptake and higher tumor-to-non-tumor radioactivity ratios than the ¹¹¹In labeled-IA. The tumor was clearly visualized by static imaging 3 hrs after the injection of the ¹¹¹In labeled-IAC. In this current research, we investigated the ^99mTc labeling of IAC using the HYNIC method with either EDDA or PDA as a co-ligand and studied the biodistribution of the resulting products in nude mice bearing the receptor-positive M21 tumors. We selected the HYNIC method because this method was shown to produce a hydrophilic, stable ^99mTc labeled RGD with a high affinity to α_vβ₃ receptor in vitro and good tumor targeting properties in vivo [18–24].

MATERIALS AND METHODS

Synthesis of IAC-HYNIC

The synthesis of 4-[2-(3, 4, 5, 6-tetrahydropyrimidin-2-ylamino)-ethyloxy]benzoyl-2-(S)-[N-(3-amino-neopenta-1-carbamyl)]-aminoethylsulfonylamino-β-alanine hydrochloride (IAC) was reported previously [25]. HYNIC-NHS (1.67 mM; Solulink, San Diego, CA) was conjugated to the amino terminus of IAC (16.7 mM), by reacting the two reagents with gentle shaking in a solvent mixture containing 67% DMF, 17% DMSO and 16% aqueous medium with the final concentration of 83 mM sodium bicarbonate overnight at room temperature (Fig. 1). The reaction resulted in the IAC-HYNIC conjugate with a 97% yield calculated based on HYNIC-NHS. The yield of IAC-HYNIC was determined based on the distribution of ^99mTc between IAC-HYNIC and HYNIC as described below.

FIGURE 1. Synthesis of ^99mTc labeled IAC.

HYNIC-NHS (1.67 mM) was conjugated to the amino terminus of IAC (16.7 mM) in 83 mM sodium bicarbonate (pH 8.4) at room temperature overnight. IAC-HYNIC was incubated with ^99mTcO₄⁻¹ in 0.3 ml at pH 6 containing tricine (112 mM, 20 mg/ml), sodium succinate (25 mM) and stannous chloride dehydrate (100 μg/ml) at 37 °C for 30 min. To this solution, 0.3 ml aqueous medium containing 6 mg of EDDA or PDA was added. Then the mixture was heated at 80 °C for 25 min.

^99mTc labeling of IAC-HYNIC using EDDA or PDA as a co-ligand

In a rubber-sealed vial, IAC-HYNIC (33.4 μM) was incubated with ^99mTcO₄⁻¹ (31 mCi; 1147 MBq) in 0.3 ml aqueous medium at pH 6 containing tricine (112 mM, 20 mg/ml), sodium succinate (25 mM) and stannous chloride dehydrate (100 μg/ml) at 37 °C for 30 min. To this solution, 0.3 ml aqueous medium containing 6 mg of EDDA or PDA was added. Then the mixture was heated at 80°C for 25 min (Fig. 1). As reported for the ^99mTc complex of HYNIC-RGD peptides [20, 24, 26, 27], we hypothesize that the ^99mTc labeling of HYNIC-IAC in the presence of EDDA produced a ^99mTc complex involving one or two EDDA molecules bonded to ^99mTc-HYNIC-IAC core (P1) and the ^99mTc labeling in the presence of PDA produced a ternary ^99mTc complex involving each of PDA and Tricine bonded to ^99mTc-HYNIC-IAC core (P2) (Fig. 1).

The labeling yield was determined by reverse-phase HPLC (Ace 5 C18 column, 4.6 × 100 mm, 5 μm, MAC-MOD Analytical Inc., Chadds Ford, PA). HPLC solvents consisted of 0.05 M triethylammonium phosphate (TEAP) buffer, pH 2.25 (solvent A) and acetonitrile (solvent B). The applied gradient was: 0 to 2 min, 100% A; 2 to 5 min, from 100 to 80% A; 5 to 12 min, from 80 to 65% A; 12 to 20 min, from 35 to 100% B; 20 to 25 min, 100% B. The labeled products were purified on a Sep-Pak C-18 cartridge. Each of the labeled products was diluted to 20 ml with distilled water and was loaded on a Sep-Pak C-18 cartridge (Long Body Sep-Pak plus, Waters, Milford MA); the Sep-Pak C-18 cartridge was preconditioned prior to the sample purification by washing with 20 ml of acetonitrile followed by 20 ml of distilled water. The cartridge was eluted stepwise with distilled water (10 ml), 5% acetonitrile in water (5 ml) and 10% acetonitrile in water (5 ml) and 20% acetonitrile in water. The product eluted out with 20% acetonitrile was collected. The organic solvent was removed by lyophilization and the product was re-dissolved in 0.02 M sodium phosphate in 0.15 M sodium chloride, pH 7.2 (PBS).

Stability study

The stability was tested in serum; 400 μl of mouse serum with 0.04% sodium azide was mixed with 50 μl of 0.5 M sodium phosphate, pH 7.2 and 50 μl (116 μCi, 4.29 MBq) of P1 or P2. The solution was incubated at 37°C. Aliquots were taken out at designated time points and analyzed by size exclusion HPLC equipped with a TSK gel G3000SW_XL column (7.8 × 300 mm, 5μm, TOSOH Bioscience, Japan; 0.067 M sodium phosphate/0.10 M potasium chloride, pH 6.8; 1.0 ml/min), a UV monitor and an online flow radioactivity detector (Bioscan Inc., Washington DC).

Receptor-binding assay

To assess the receptor-binding ability, the ^99mTc labeled products (1.1×10⁶ cpm, <2.5 pmol) were incubated with various amounts (2.6 to 51.2 pmol) of human integrin α_vβ₃ (MW 237,000; Chemicon, Temecula, CA) in a total volume of 64 μl PBS, pH 7.2 at 37 °C for 3 hrs [9]. To assess the percentage of non-specific binding, the binding assay was performed in the presence of a 250 times molar excess of IA to integrin α_vβ_3. The binding percentage to the integrin α_vβ₃ was analyzed by size exclusion HPLC equipped with a TSK gel G3000SW_XL column. The retention time of the receptor-bound and the free ^99mTc labeled antagonists was 7 min and 15 min, respectively.

Tumor Model

The human M21 melanoma cell line expressing α_vβ₃ was used to make an in vivo tumor model. Human M21 melanoma cells initially received from Dr. David Cheresh (Scripps Research Institutes, La Jolla, CA) were cultured in RPMI 1640 supplemented with 10% FBS and 1% penicillin-streptomycin under a humidified atmosphere with 5% CO₂ [9]. Tumor xenografts were established by s.c. inoculation of 3 × 10⁶ cells into the right flank of athymic mice (NCI-DCT, Frederick, MD).

Biodistribution Study

Animal experiments were performed under an NIH Animal Care and Use Committee approved protocol. Biodistribution studies were performed in nude mice (n=4~15 per time point) bearing receptor-positive M21 human melanoma xenografts. The mice received intravenous (i.v.) P1 or P2 (6.25 μCi, 0.231 MBq/6 pmol in 0.2 ml PBS containing 1% BSA when the tumor size was approximately 200 mm³. The mice were euthanized at designated times after injection for the biodistribution study. In a separate experiment, a group of mice was coinjected with 0 ~ 200 μg of unlabeled IA and euthanized at 1 h to test if the tumor uptake was blocked by the excess amount of IA.

Imaging

Nuclear imaging study of tumor-bearing mice was performed using a micro-SPECT single photon emission computed tomography system (A-SPECT, Gamma Medica Instruments, Northridge, CA) with a 2 mm pin-hole collimator for high resolution animal imaging. The camera was equipped with an array of 2 mm X 2 mm X 6 mm NaI (Tl) pixels coupled to a position-sensitive photo-multiplier tube readout. Mice were i.v. injected with 33.3 MBq (900 μCi) of either P1 or P2. Then the animals were anaesthetized with ketamine (60 mg/kg)/xylazine (10 mg/kg) immediately before being placed supine in a gantry positioned in front of the pin-hole collimator, and they were centered on the field of view of the micro-SPECT. The static imaging was performed for 10 ~ 40 min at 3 h post-injection.

Statistical analysis

Statistical analysis was performed using ANOVA [9] for comparing multiple groups, and Student’s t test was performed for unpaired data between two groups. All tests were two-sided, and a probability value (p) of less than 0.05 was considered significant.

RESULTS

In vitro studies

The IAC-HYNIC was synthesized with a 97% yield based on HYNIC-NHS and was labeled with ^99mTc using EDDA or PDA as a co-ligand without further purification. The conjugation yield of HYNIC to the amino terminus of IAC was proportional to the concentration of IAC as determined by the distribution of ^99mTc between HYNIC-IAC and HYNIC. A quantitative conjugation yield was obtained when the IAC concentration was > 20 mM and the reaction was performed at an IAC to chelator molar ratio of 10 at pH 8.4 overnight. This high conjugation yield rendered the purification step of HYNIC-IAC from the free chelator unnecessary. The ^99mTc labeled product P1 or P2 was purified on a Sep-Pak C-18 cartridge and were eluted out with 20% acetonitrile. The organic solvent was then removed by lyophilization and the product was redissolved in 0.02 M sodium phosphate-0.15 M NaCl (PBS). The radiochemical purity of the purified P1 or P2 was > 97% (Fig. 2). The radiolabeled products were stable in PBS with ^99mTc radioactivity associated with P1 and P2 for an 18-h period of incubation. The stability study in serum indicated that at 18 h, 13% and 14% of total ^99mTc radioactivity were associated with a protein with its HPLC retention time identical to albumin and 3% and 9% of ^99mTc radioactivity were associated with ^99mTcO₄ ⁻¹ for P1 and P2, respectively (Fig. 3). The receptor-bindability analyzed by size exclusion HPLC showed that 78.02 ± 13.48% (n = 4) of P1 and 51.05 ± 14.05% (n = 4) of P2 bound to 0.8 μM α_vβ₃ in vitro. The receptor-binding was completely blocked by a molar excess of unlabeled α_vβ₃ antagonist.

FIGURE 2.

Reverse-phase HPLC profile of P1 (A) and P2 (B) after Sep-Pak C-18 purification. See the Methods section for the HPLC elution condition.

FIGURE 3.

Stability of P1and P2 in mouse serum at 37 °C. The formation of ^99mTc-bound serum protein and ^99mTcO₄⁻¹ increased over time whereas the percentage of P1 (diamond) and P2 (square) decreased. Mean values were plotted over time. Bar, S.D.; N ≥ 3.

Animal studies

Both P1 and P2 accumulated rapidly in the receptor-positive tumor with the uptake values of 3.17±0.52 and 2.13±0.17 % ID/g at 1 h, respectively. The tumor uptake of P1 at 1 h was blocked by unlabeled IA in a concentration dependent manner (Fig. 4). The co-injection of 200 μg of the antagonist decreased the tumor uptake values of P1 and P2 to 0.68 ± 0.06 % ID/g and 0.96 ± 0.30 % ID/g at 1 h, respectively (p < 0.004). This indicated that the tumor uptake was receptor mediated (Fig. 5). There were differences shown in their washout rates from the tumor; the tumor uptake value of P1 remained relatively unchanged for a 4-h period whereas the radioactivity of P2 in the tumor was washed out with 62% of P2 at 1 h remained in the tumor at 4 h (p <0.00005) (Fig. 5). Differences were also found in the whole-body retention; P1 was retained in the whole-body with 39.58 ± 2.91% ID at 1 h and 27.25 ± 3.67% at 4 h, indicating that it was excreted primarily via the renal system. On the other hand, P2 remained in the whole-body with 57.84 ± 5.15% ID at 1 h and 54.04 ± 3.57% ID at 4 h, indicating that it was excreted to a larger extent via the hepatobiliary system (Fig. 6). Comparative biodistribution studies also showed significantly higher accumulation of P2 in liver and intestine than P1 for a 4-h period. The tumor-to-nontumor ratios of both P1 and P2 increased over time, indicating preferential retention of radioactivity in tumor over most other organs except the intestine. Comparing the tumor-to-nontumor ratios of the two labels, the increases in the ratios of P1 were larger than those of P2 (Table 1). These biodistribution results were further substantiated by imaging studies (Fig. 7). The static anterior imaging performed at 3 h post-injection showed that the tumor areas were much brighter than the surrounding background. The non-tumor associated radioactivity was primarily seen in abdomen and bladder areas for P1. Compared to P1, a much higher background was observed in the P2 images.

FIGURE 4.

Effect of IA co-injection on tumor uptake (%ID/g) of P1. A group of mice was coinjected with 0, 10 and 200 μg of unlabeled IA and the tumor uptake was determined at 1 h postinjection. Columns, mean; bar, S.D. N = 5.

FIGURE 5.

Biodistribution (%ID/g, n ≥ 5) of P2 (A) and P1 (B) in melanoma M21-bearing nude mice. Columns, mean; bar, SD.

FIGURE 6.

Whole body retention (%ID/g, n = 5) of P1 (square) and P2 (circle) in melanoma M21-bearing nude mice at 1, 2 and 4 hour post injection. Mean values were plotted over time. Bar, S.D.

Table 1.

Tumor-to-organ ratios^* of P1 and P2 in M21 melanoma bearing nude mice

Organ	Time
	0.3 h	1 h		2 h		4 h
	P1a	P1	P2b	P1	P2	P1	P2
Blood	0.90±0.20	1.38±0.48	1.83±0.16	2.19±0.18	2.58±0.25	17.11±5.18	3.15±0.25
Spleen	0.71±0.07	1.36±0.38	1.09±0.15	1.30±0.12	1.95±0.27	2.83±0.81	1.54±0.40
Liver	0.72±0.12	0.89±0.27	0.46±0.27	1.02±0.12	0.45±0.30	1.59±0.72	0.40±0.17
Kidney	0.51±0.12	0.75±0.28	0.62±0.05	0.85±0.08	0.81±0.07	1.22±0.32	0.78±0.04
Lung	0.72±0.10	1.19±0.43	0.83±0.22	1.50±0.09	1.34±0.51	2.47±0.49	2.71±0.95
Heart	1.91±0.43	2.48±0.75	2.64±0.36	3.23±0.48	3.85±0.57	7.62±1.99	3.88±0.90
Stomach	0.54±0.13	0.40±0.20	0.40±0.12	0.43±0.11	0.29±0.09	1.53±0.80	0.27±0.05
Intestine	0.66±0.07	0.60±0.19	0.15±0.01	0.57±0.07	0.11±0.02	0.46±0.15	0.07±0.00
Bone	1.29±0.23	2.12±1.17	1.48±0.37	2.46±0.19	2.73±0.54	4.78±1.07	3.76±1.36
Muscle	3.04±0.68	6.11±1.80	3.85±1.64	4.68±0.31	5.75±1.56	7.84±1.78	6.33±1.57

^a99mTc EDDA₂/HYNIC-IAC (P1)

^b99mTc PDA (tricin)/HYNIC-IAC (P2)

^*Ratios were calculated by dividing tumor uptake (%ID/g) by organ uptake (%ID/g)

FIGURE 7.

Images of melanoma M21-bearing nude mice at 3 h postinjection. The mice received 33.3 MBq of P1 or P2 by i.v. injection. (A) and (B), P1; (C) and (D), P2. The animals were anaesthetized with ketamine (60 mg/kg)/xylazine (10 mg/kg) immediately before imaging. The static image was acquired for 33 min/40 K counts (A), 43 min/40 K counts (B), 17 min/40 K counts (C), and 13 min/40 K counts (D) to visually compare the tumor-to-background ratios.

DISCUSSION

The goal of this research was to synthesize ^99mTc labeled peptidomimetic antagonists that achieve a high target to non-target radioactivity ratio within a short time after injection. In this study, a second generation peptidomimetic α_vβ₃ antagonist, IAC (IC₅₀ 2.94 nM), was conjugated with HYNIC for ^99mTc labeling. We optimized the conjugation reaction at pH 8.4 and an IAC to succinimidyl HYNIC molar ratio of 10 to conjugate HYNIC solely to the amino terminus of IAC and not to the guanidino group of IAC. We believe that under this optimal condition, the guanidino group (pKa ~12) is completely protonated and thus, non-reactive with succinimidyl HYNIC. The resulting HYNIC-IAC was labeled with ^99mTc by stepwise reactions to produce either P1 or P2. P1 or P2 was produced when Tricine was first reacted and then EDDA or PDA was added as a coligand. Both P1 and P2 had an overall yield of >80% yield. Both P1 and P2 were quite stable. P1 was also produced in one step reaction without Tricine. However, the labeling yield from the one step reaction was much lower than the 2-step reaction with Tricine (unpublished observation), similar to that reported by Wang, et al. [26]. When P1 and P2 were incubated in serum for 18 h at 37º C, about 13% of the total ^99mTc radioactivity was bound to a serum protein with a HPLC retention time similar to that of albumin, and 4 to 9% of the radioactivity was associated with a peak with a HPLC retention time identical to that of ^99mTcO₄⁻¹. These results are similar to those reported for the corresponding ^99mTc labeled peptides [28]. The structure of the ^99mTc complex of EDDA/HYNIC has not been unequivocally defined. von Guggenberg, et al., postulated a ^99mTc complex involving two EDDA molecules bonded to ^99mTc-HYNIC core based on the analysis by Mass Spectroscopy [27] whereas Wang, et al., proposed a ^99mTc complex involving one EDDA molecules bonded to ^99mTc-HYNIC core [26]. A ternary complex of PDA (Tricine)/HYNIC was first postulated by Liu, et al. [20]. However, we have not further investigated on the precise structure of these ^99mTc complexes in this current study.

The in vivo studies, P1 showed better tumor targeting properties than P2. Although two ^99mTc radioligands rapidly targeted tumor via a receptor-mediated uptake, P1 remained longer in the tumor. It was excreted primarily via the renal system, thereby resulting in lower whole-body background radioactivity for a 4-h period. On the other hand, P2 was excreted equally via the hepatobiliary and the renal systems, thereby showing higher liver and intestinal radioactivity during a 4-h study. The imaging studies with these two ^99mTc radioligands also clearly visualized tumors at 3 h postinjection. However, a better tumor-to-abdominal contrast was shown with P1. The whole-body clearance of P1 was similar to that of ¹¹¹In labeled IAC using 2-p-isothiocyanatobenzyl-DOTA as a bifunctional chelator [28]. Although the tumor uptake values were lower for P1 than for the ¹¹¹In labeled IAC, the tumor-to-organ ratios of P1 at 4 h became similar to the corresponding ratios of the ¹¹¹In labeled IAC at 2 h because of the preferential retention of P1 in the tumor than in the organs. The tumor imaging property of P1 was comparable to that of ¹¹¹In labeled IAC using 2-p-isothiocyanatobenzyl-DOTA as a bifunctional chelator at 3 h postinjection.

CONCLUSION

P1 showed good tumor targeting properties with a rapid tumor uptake kinetics, prolonged tumor retention and rapid whole-body clearance kinetics. These favorable biological behavior in combining with excellent physical properties (t1/2 of 6 h and 140 KeV photon energy) of ^99mTc warrant further investigation of ^99mTc labeling for the receptor-mediated tumor targeting utilizing the HYNIC method for other peptidomimetic antagonists with even higher affinity produced by the dimerization or tetramerization of this peptidomimetic antagonist, similar to that shown for cyclic RGDfK derivatives [3, 26].