Synthesis and Evaluation of Bombesin Derivatives on the Basis of Pan-Bombesin Peptides Labeled with Indium-111, Lutetium-177, and Yttrium-90 for Targeting Bombesin Receptor-Expressing Tumors Free

The development of ligand-targeted therapeutics in anticancer therapy including drug-ligand conjugates has gained momentum in recent years (1). Systemic cytotoxic chemotherapy shows little selectivity and is limited by potentially serious side effects. One strategy to improve the lack of selectivity is to couple therapeutics to vectors like monoclonal antibodies, their fragments, or even smaller molecules (2). The cytotoxic drug part of conjugates used in ligand-targeted therapeutics is often composed of a therapeutic radiometal encapsulated by its bifunctional chelator.

A very promising group of small targeting ligands is composed of regulatory peptides (3). A high number of peptide receptors were shown to be overexpressed in various human tumors (4). They are promising targets for molecular imaging and targeted therapy of cancer, because they are located on the plasma membrane and, upon binding of a ligand, the receptor-ligand complex is internalized. These findings were the basis for the development of diagnostic and therapeutic radiopeptides useful in peptide receptor scintigraphy and targeted radiotherapy (5, 6, 7, 8, 9, 10). Among the most relevant peptide receptors, the bombesin receptors are of major interest, because they were found to be overexpressed in various important cancers like prostate (11, 12), breast (13, 14), and small cell lung cancer (15). The human counterparts of bombesin, namely gastrin-releasing peptide (16) and neuromedin B (17), have been found in mammalian tissue. They bind to different bombesin receptor subtypes, such as the neuromedin B preferring receptor (BB1 receptor; ref. 18), the gastrin-releasing peptide preferring receptor (BB2; ref. 19), as well as the orphan bombesin receptor subtype-3 (BB3 receptor; ref. 20) and the BB4 receptor (21). The BB1, BB2, and BB3 receptors have been shown recently to be overexpressed on different human tumors (22). Gastrin-releasing peptide receptors were predominantly expressed in human prostate cancer (100%), gastrinoma (100%), and breast cancer (70%), whereas concomitant expression of gastrin-releasing peptide receptor (33%) and BB3 receptor (40%) were found in small cell lung cancer. Also gastrin-releasing peptide receptor (40%) and BB3 (25%) were found concomitantly in renal cell carcinoma. Preferential expression of BB1 was found in intestinal carcinoids (11 of 24), and bronchial carcinoids had preferential BB3 receptor expression (9 of 26).

These findings provide a possibility to apply bombesin-like peptides as a vehicle for delivering cytotoxic drugs (23, 24, 25) into tumor cells. In addition, radiolabeling may allow us to diagnose and treat these tumors (10, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37). The sequence bombesin(7, 8, 9, 10, 11, 12, 13, 14) was regarded to be sufficient for the specific binding interaction with the gastrin-releasing peptide receptor (38, 39). Therefore, most radiolabeled bombesin-like peptides are based on the sequence bombesin (7, 8, 9, 10, 11, 12, 13, 14)(10, 28, 29, 30, 31, 33, 34, 35) . For example, different conjugates were developed using bifunctional chelators for labeling with ^99mTc, like N₂S₂ (29), N₃S (31), N_α-histidinyl acetate (35), and diaminopropionic acid (36), using the carbonyl approach. Also, diethylenetriaminepentaacetic acid (DTPA) and 1,4,7,10-tetraazacyclododecane-N,N′,N″,N‴-tetraacetic acid (DOTA) were coupled to this sequence for labeling with hard Lewis acid radiometals like ¹¹¹In, ^{67, 68}Ga, ⁹⁰Y, and the lanthanides. Some ^99mTc-labeled peptides have been or are currently being investigated in gastrin-releasing peptide receptor-positive tumors in patients (30, 31, 33).

Recently, a universal ligand, (d-Tyr⁶, β-Ala¹¹, Phe¹³, Nle¹⁴) bombesin (6, 7, 8, 9, 10, 11, 12, 13, 14), has been developed by Mantey et al. (40) and Pradhan et al. (41), which has high affinity to all of the bombesin receptor subtypes. The finding that not only the gastrin-releasing peptide receptor is overexpressed on human tumors but in some cases also neuromedin B and BB3 receptor subtypes prompted us to develop conjugates based on the slightly modified (Thi¹³ versus Phe¹³) universal bombesin ligand [d-Tyr⁶, β-Ala¹¹, Phe¹³, Nle¹⁴] bombesin (6, 7, 8, 9, 10, 11, 12, 13, 14) that can be labeled with hard Lewis acid-type metallic radionuclides like ¹¹¹In, ⁹⁰Y, and ¹⁷⁷Lu. The Thi¹³ versus Phe¹³ modification was done because preliminary data from our laboratory have shown an increased metabolic stability of this peptide over the Phe¹³ analogue. We studied these (radio)metallopeptides with regard to the bombesin receptor subtype profile. The internalization rate of the two chelated peptides, labeled with the indicated radiometals, into AR4–2J (rat pancreatic tumor cells bearing the gastrin-releasing peptide receptor) and PC-3 (human prostate cancer cell line) cells was studied as well. We also report on the metabolic stability in human blood serum and the identification of metabolites. In addition, the biodistribution of the ¹¹¹In-labeled peptides was studied. This work is the first one of a pan-bombesin ligand aimed at radiotargeted diagnosis and therapy.

All chemicals were obtained from commercial sources and used without additional purification. DOTA-tris(tBu ester) was commercially available (Macrocyclics, Dallas, TX) or was synthesized according to Heppeler et al. (42). DTPA(tris-tBu) was received from Mallinckrodt Medical (Dr. Ananth Srinivasan, St. Louis, MO). Rink amide MBHA resin and all of the Fmoc-protected amino acids were commercially available from NovaBiochem (Läufelfingen, Switzerland). [¹¹¹In]Cl₃ was purchased from Mallinckrodt Medical (Petten, the Netherlands), [⁹⁰Y]Cl₃ from Perkin-Elmer Life Sciences Inc. (Boston, MA), and [¹⁷⁷Lu]Cl₃ from I.D.B. (Petten, the Netherlands). Electrospray ionization mass spectroscopy was carried out with a Finnigan SSQ 7000 spectrometer or fast atom bombardment mass spectroscopy with a VG 70SE spectrometer and matrix-assisted laser desorption ionization-mass spectrometry measurements on a Voyager sSTR equipment with a Nd:YAG laser (Applied Biosystems, Framingham, NY). Analytical high-performance liquid chromatography (HPLC) was performed on a Hewlett Packard 1050 HPLC system with a multiwavelength detector and a flow-through Berthold LB 506 Cl γ-detector using a Macherey-Nagel Nucleosil 120 C₁₈ column. Preparative HPLC was performed on a Metrohm HPLC system LC-CaDI 22–14 with a Macherey-Nagel VP 250/21 Nucleosil 100–5 C₁₈ column. Quantitative gamma counting was performed on a COBRA 5003 γ-system well counter from Packard Instruments. Solid-phase peptide synthesis was performed on a semiautomatic peptide synthesizer commercially available from Rink Combichem (Bubendorf, Switzerland). The cell culture medium was DMEM with 10% FCS from Bioconcept.

The peptide synthesis was performed on a semiautomatic peptide synthesizer according to a general procedure described previously (ref. 43; Fig. 1). Standard Fmoc chemistry was used throughout (44); the peptide was assembled on a Rink amide MBHA resin. Trt and tBu were used as protecting groups of His and d-Tyr, respectively, and Boc for Trp. The chelators were coupled as follows. Three equivalents DTPA(tBu)₃ were preincubated with N,N′-diisopropylcarbodiimide in 1-methyl-2-pyrrolidone for 30 min and incubated with the resin-assembled peptide until the 2,4,6-trinitrobenzenesulfonic acid test was negative (∼3 hours). DOTA(tBu)₃ was coupled as described (43). The peptide chelator conjugates were cleaved from the resin and deprotected by incubation with trifluoroacetic acid—thioanisole–water 92:6:2 for 4 to 6 hours at room temperature and precipitated in isopropyl ether–petroleum ether (1:1). The crude peptide-chelator conjugate was purified by preparative HPLC (Macherey-Nagel Nucleosil 100–5 C₁₈, flow: 15 mL/min; eluents: A = 0.1% trifluoroacetic acid in water and B = acetonitrile; nonlinear gradient: 0 min, 70% A; 10 min, 50% A). Mass spectrometry [(−)electrospray ionization, matrix-assisted laser desorption ionization] was used to determine the composition of the conjugates.

Diethylenetriaminepentaacetic acid-γaminobutyric acid (DTPA-GABA), DTPA-GABA-d-Tyr, DTPA-GABA-d-Tyr-Gln, DTPA-GABA-d-Tyr-Gln-Trp, DTPA-GABA-d-Tyr-Gln-Trp-Ala-Val, DTPA-GABA-d-Tyr-Gln-Trp-Ala-Val-β-Ala, and DTPA-GABA-d-Tyr-Gln-Trp-Ala-Val-β-Ala-His were synthesized in parallel using the same protocol as described above.

The metal complex was synthesized according to the methods described previously (42). A mixture of DTPA-GABA-[D, Tyr⁶, β-Ala¹¹, Thi¹³, Nle¹⁴] BN(6, 7, 8, 9, 10, 11, 12, 13, 14) (BZH1; 0.5 μmol) in 500 μL of 0.4 mol/L sodium acetate buffer (pH 5.0) was incubated with 1.5 μmol InCl₃·5H₂O in 0.04 mol/L HCl for 1 h at room temperature and purified over a SepPak C₁₈ cartridge preconditioned with 10 mL of EtOH and 10 mL of water. The cartridge was eluted with 10 mL of water followed by 3 mL of methanol resulting in [¹¹¹In]-BZH1 after evaporation of the methanol. The final product was analyzed by analytical HPLC; the purity was ≥97%. Mass spectrometry [matrix-assisted laser desorption ionization, m/z (%)]: 1711.0 (100, [M+H]⁺), 1732.9 (25, [M+Na]⁺).

[Y^III]-DOTA-GABA-[D-Tyr⁶, β-Ala¹¹, Thi¹³, Nle¹⁴] BN (6, 7, 8, 9, 10, 11, 12, 13, 14) (BZH2) was synthesized using Y(NO₃)₃·5H₂O as described above except that the incubation was at 95°C for 20 to 25 minutes. The final product was analyzed by analytical HPLC; the purity was >95%. Mass spectrometry (matrix-assisted laser desorption ionization; m/z (%)):1696.8 (100, [M+H]⁺), 1718.8 (20, [M+Na]⁺), 1734.8 (10, [M+K]⁺).

[¹¹¹In]-BZH1 was prepared by dissolving 10 μg of BZH1 (6.25 nmol) in sodium acetate buffer [300 μL and 0.4 mol/L (pH 5.0)] and by incubation with ¹¹¹InCl₃ (3 to 6 mCi) for 1 hour at room temperature. A 1.5 molar excess of InCl₃·5H₂O was added and the final solution incubated again at room temperature for 1 hour. Subsequently, radiometalated peptides were purified using a SepPak C₁₈ cartridge as described above affording a very pure ¹¹¹In/^natIn-labeled ligand for internalization studies. For biodistribution and serum stability studies, the labeling was performed accordingly without the addition of cold InCl₃·5H₂O. For injection the solution was prepared by dilution with 0.9% NaCl (0.1% bovine serum albumin) to afford the radioligand solution.

[⁹⁰Y/¹¹¹In/¹⁷⁷Lu]-BZH2 were prepared and purified accordingly by heating at 95°C for 20 to 25 minutes; Y(NO₃)₃·5H₂O, Lu(NO₃)₃·6H₂O, and InCl₃·5H₂O were used.

To 1 mL of freshly prepared human serum, previously equilibrated in a 5% CO₂ (95% air) environment at 37°C, we added 0.6 nmol ¹¹¹In-labeled BZH1 or BZH2 standard solution. The mixture was incubated in a 5% CO₂, 37°C environment. At different time points, 50-μL aliquots (in triplicate) were removed and treated with 60 μL of EtOH. Samples were then cooled (4°C) and centrifuged for 15 min at 500 × g and 4°C to precipitate serum proteins. Fifty μL of supernatant were removed for activity counting in a γ-well counter, the sediment was washed twice with 1 mL of EtOH and counted, and the activity in the supernatant was compared with the activity in the pellet to give the percentage of peptides not bound to proteins or radiometal transferred to serum proteins. The supernatant was analyzed with HPLC (eluents: A = 0.1% trifluoroacetic acid in water and B = acetonitrile; gradient: 0 to 45 minutes, 95% to 60% A; 45 to 46 minutes, 100% B; 46 to 49 minutes, 100% B; 50 minutes, 95% A) to determine relative amounts of metabolites.

The data points were fitted using origin 6 (Microcal Software, Inc., Northampton, MA), assuming a consecutive reaction (Eq. A) and Eq. B and C to calculate the disappearance of intact peptide A as well as the formation and disappearance of metabolite B, respectively.

To identify the metabolites, [¹¹¹In]-BZH1 was used as the leading peptide. The extracted supernatant, obtained as described above, was co-injected with the potential “metabolites,” synthesized as described above. The metabolism in serum was studied by the addition of EDTA (2.4 mmol/L) or carnosine (22.4 mmol/L) to [¹¹¹In]-BZH1 serum solutions.

The binding affinity profiles of [In^III]-BZH1 and [Y^III]-BZH2 for the three bombesin receptor subtypes were determined in vitro using receptor autoradiography. Human tumors were selected that had been shown previously to express predominantly one single bombesin receptor subtype, namely either neuromedin B receptor, gastrin-releasing peptide receptor or BB₃ receptor. IC₅₀ were measured in competitive binding experiments performed with increasing concentrations of [In^III]-BZH1, [Y^III]-BZH2 and [d-Tyr⁶, β-Ala¹¹, Phe¹³, Nle¹⁴]bombesin(6–14; as reference) in successive tissue sections containing tumors expressing either gastrin-releasing peptide-receptors, neuromedin B receptors or BB3 receptors, using [¹²⁵I-d-Tyr⁶, β-Ala¹¹, Phe¹³, Nle¹⁴]bombesin(6, 7, 8, 9, 10, 11, 12, 13, 14) as universal radioligand, as described in detail previously (22, 45).

AR4–2J rat pancreatic tumor cells and PC-3 cells were cultured in Dulbecco’s minimal essential medium (DMEM). DMEM was supplemented with vitamins, essential and nonessential amino acids, l-glutamine, antibiotics (penicillin/streptomycine), fungicide (amprotencine) and 10% fetal calf serum (FCS) as described elsewhere (46).

Internalization and externalization experiments were performed in 6-well plates as indicated in a previous publication (46). The procedure was the same for both cell lines. Briefly, the cells were washed twice with the internalization medium and allowed to adjust to the medium for 1 h at 37°C. Approximately 1.8 kBq (0.25 pmol) of radioligand were added to the medium and the cells (10⁶ cells per well) and incubated (in triplicates) for 0.5, 1, 2, 4 and 6 h at 37°C, 5% CO₂, with or without excess of cold BZH2 (150 μL of a 5.8 μmol/L solution, final concentration of cold BZH2 was 0.58 μmol/L) to determine nonspecific internalization. The final volume was 1.5 ml. At appropriate time points the internalization was stopped by removal of the medium followed by washing the cells with ice-cold solution composed of 0.9% NaCl/0.01 mol/L Na₂HPO₄/0.01 M KH₂PO₄ (pH 7.2). Cells were then treated 5 min (twice) with glycine buffer (0.05 mol/L glycine solution, pH adjusted to 2.8 with 1 mol/L HCl) to distinguish between cell surface-bound (acid releasable) and internalized (acid resistant) radioligand. Finally, cells were detached from the plates by incubation with 1 mol/L NaOH for 10 min at 37°C, and the radioactivity was measured in a γ-counter.

For externalization studies, the AR4–2J cells were allowed to internalize the radioligands for periods of 2 h and were then exposed to an acid wash, as described in the previous section, to dissociate cell-surface-bound radioligand. One ml of culture medium was added to each well, cells were incubated at 37°C in a 5% CO₂ environment and externalization of the cell-incorporated radioactivity was studied at different times. The culture medium was collected and measured for radioactivity.

Lewis male rats were implanted subcutaneously with 10 millions AR4–2J tumor cells, which were freshly expanded in a sterilized solution of 0.9% NaCl/0.01 mol/L Na₂HPO₄/0.01 mol/L KH₂PO₄ (pH 7.2).

Fourteen days after inoculation the tumors weighed 0.3–1.2 g and the rats were injected into the back leg vein with 0.1 μg radiolabeled peptides (about 0.5 MBq ¹¹¹In or 0.9 MBq ¹⁷⁷Lu), diluted in NaCl (0.1% bovine serum albumin, pH 7.4, total injected volume = 200 μL). For the determination of nonspecific uptake in tumor or receptor positive organs, a group of 4 animals was injected with a mixture of 0.1 μg radiolabeled peptide/50 μg BZH2 in 0.9% NaCl solution (injected volume, 225 μL). At 4 h, 24 h, 48 h, and 72 h rats (in groups of 4–11 rats) were sacrificed, and organs of interest were collected, rinsed of excess blood, blotted, weighed and counted in a γ-counter. The percentage of injected dose per gram (% ID/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.

All animal experiments were performed in compliance with the Swiss regulation for animal treatment (Bundesamt für Veterinärwesen, approval no. 789).

BZH1 and BZH2 (Fig. 1) were synthesized using Fmoc strategy affording an overall yield of approximately 30% based on the removal of the first Fmoc group; the purity analyzed by HPLC was ≥ 97%. BZH1 was labeled with ¹¹¹In by incubation at room temperature (1 h incubation, pH 5, 0.4 mol/L sodium acetate buffer). BZH2 was labeled with ¹¹¹In, ⁹⁰Y, and ¹⁷⁷Lu at elevated temperature (95°C, 20–25 min). In all cases, labeling yields of ≥ 98% at specific activities of > 37 GBq μmol^-1 were achieved.

Table 1 shows the bombesin receptor subtype binding profile of the 2 new metallopeptides. As reference peptide [d-Tyr⁶, β-Ala¹¹, Phe¹³, Nle¹⁴] bombesin(6, 7, 8, 9, 10, 11, 12, 13, 14) was used, which binds with high affinity to all 3 bombesin receptor subtypes. Although less potent than the reference peptide, the 2 new bombesin analogues still have retained high affinity to all three receptor subtypes. The IC₅₀ values of [In^III]-BZH1 are 3.47 ± 0.32 nm to the gastrin-releasing peptide receptor, 10.5 ± 3.03 nm to the neuromedin B receptor, and 41.7 ± 22.2 nm to the BB3 receptor. The respective values for [Y^III]-BZH2 are 1.40 ± 0.10 nm, 4.93 ± 1.03 nm, and 10.7 ± 4.2 nm.

After incubation of the radiolabeled peptides with fresh serum, three metabolites were determined using HPLC and radiometric analysis. The proteic fraction obtained as pellet contained <5% of radioactivity. [¹¹¹In]-BZH1 was metabolized to [¹¹¹In]-DTPA-GABA-d-Tyr-Gln-Trp-Ala-Val-β-Ala (corresponding to B) due to the cleavage between β-Ala¹¹ and His¹², and then this metabolite decomposed to [¹¹¹In]-DTPA-GABA-d-Tyr-Gln (C) and [¹¹¹In]-DTPA (D; Fig. 2). These metabolites were identified using reverse-phase HPLC by co-injection of presynthesized chelator-conjugated peptides. The similarity of HPLC elution times indicates that [¹¹¹In]-BZH2 is metabolized to the same breakdown products. There was no indication of radiometal transfer to serum proteins during incubation studies.

Fig. 3 shows the kinetics of metabolic degradation of [¹¹¹In]-BZH1 and [¹¹¹In]-BZH2. Curve fitting data gave rate parameters for the disappearance of intact peptides (A, A′) and the built-up of metabolites (B, C, D; B′, C′, D′).

The respective half-lives for [¹¹¹In]-BZH1 and [¹¹¹In]-BZH2 were determined to 2.0 hours and 2.3 hours; however, the first metabolite B′ of [¹¹¹In]-BZH2 was more stable than B of [¹¹¹In]-BZH1, and their half-lives of disappearance are 26 hours and 46 hours, respectively (Table 2). When carnosine or EDTA were incubated together with [¹¹¹In]-BZH1 in serum, [¹¹¹In]-BZH1 is significantly more stable than without these inhibitors (Fig. 4).

Both ¹¹¹In-labeled radiopeptides showed a fast cell uptake, which did not reach a plateau within 6 hours of incubation at 37°C in AR4–2J cells (Figs. 5 and 6). Internalization was strongly reduced in the presence of 0.57 μmol/L unlabeled BZH2 (data not shown). Nonspecific internalization was <1% of the added activity. The surface bound peptide (acid removable) was <7% of the added activity (data not shown). BZH2 labeled with different radioisotopes (¹¹¹In, ¹⁷⁷Lu, and ⁹⁰Y) exhibited the same internalization rates (Fig. 5). [¹¹¹In]-BZH1 was internalized by AR4–2J and PC-3 cells with a similar kinetics (Fig. 4). Preliminary studies using in vitro receptor autoradiography showed that AR4–2J and PC-3 tumors express predominantly the gastrin-releasing peptide receptor.³

The kinetics of externalization of both peptides was studied with cells exposed for 2 hours to the radioligand as described for internalization. Within 2 hours, 45% of [¹¹¹In]-BZH1 and 30% of [¹¹¹In]-BZH2 were released from the AR4–2J cells. The efflux of [¹¹¹In]-BZH1 was faster than that of [¹¹¹In]-BZH2 (Fig. 7). To identify the composition of externalized peptide, [¹¹¹In]-BZH1 was used as a leading peptide. The externalized radioactivities corresponded to the intact peptide and one metabolite [¹¹¹In(DTPA)]²⁻, identified by HPLC using co-injection.

Upon 2 hours of internalization and acid wash, the externalized radioactivity already after 30 min consisted of ∼50% [¹¹¹In(DTPA)]²⁻ as the only metabolite and 50% intact radiolabeled peptide. At 4 hours, the externalized radioactivity consisted of the metabolite [¹¹¹In(DTPA)]²⁻ (87.5%) and intact peptide (12.5%), whereas at 24 hours, only the metabolite was found.

Results from biodistribution studies using the ¹¹¹In-labeled peptides performed with Lewis rats bearing the AR4–2J pancreatic tumor are presented in Table 3 as the percentage of injected dose per gram of tissue (% ID/g). Tumor uptake and retention of [¹¹¹In]-BZH2 is also shown in Fig. 8.

Both [¹¹¹In]-BZH1 and [¹¹¹In]-BZH2 displayed rapid blood clearance with <0.02% ID/g and 0.01% ID/g at 4 hours, respectively. Fast clearance from the gastrin-releasing peptide receptor-negative tissues except the kidneys was found as well. [¹¹¹In]-BZH1 and [¹¹¹In]-BZH2 show high uptake values in the AR4–2J tumor and in the gastrin-releasing peptide receptor-positive organs, e.g., at 4 hours: tumor, 1.71 ± 0.51% ID/g versus 0.79 ± 0.07% ID/g; and pancreas, 3.92 ± 0.86% ID/g versus 2.63 ± 0.59% ID/g. The tumor uptake of [¹¹¹In]-BZH2 dropped from 0.79 ± 0.07% ID/g at 4 hours to 0.44 ± 0.07% ID/g at 24 hours, 0.42 ± 0.05 at 48 hours, and 0.33 ± 0.08 at 72 hours. The respective pancreas values are 2.63 ± 0.59 at 4 hours, 2.13 ± 0.68 at 24 hours, 2.07 ± 0.22 at 48 hours, and 1.65 ± 0.32 at 72 hours.

In vivo competition experiments using 50 μg of BZH2 co-injected with [¹¹¹In]-BZH1 (data for [¹¹¹In]-BZH2 in brackets) resulted in a >93% (89%) reduction of tumor uptake and also in a reduction of the uptake in normal gastrin-releasing peptide receptor-positive organs (>97% for both radiopeptides in the pancreas). Also, the uptake in other organs could be blocked to a high degree: stomach 84% (65%), bowel 72% (82%), and adrenals 73% (66%). The injection of the blocking dose had no significant influence on the uptake in nontarget organs except the kidneys where the uptake decreased by a factor of 1.5 on blocking. Due to the fast clearance of both peptides, high tumor–background ratios were found (Table 3).

¹¹¹In was used as a surrogate of ⁹⁰Y for studying the pharmacokinetics of BZH2, because the latter is a pure β-emitter. This is the strategy most often used in such studies, although there may be some differences between the two due to structural differences (47, 48). Therefore, we also studied [¹⁷⁷Lu]-BZH2 as a radiotherapeutic peptide radiopharmaceutical. ¹⁷⁷Lu is a low energy β-emitter that has two γ-lines at 133 keV (7%) and 208 keV (11%) allowing convenient localization. The organ uptake values of [¹⁷⁷Lu]-BZH2 were very similar to those of [¹¹¹In]-BZH2. The values for the most important organs are: tumor 0.67 ± 0.04% ID/g (4 hours) and 0.42 ± 0.03% ID/g (24 hours); pancreas 2.39 ± 0.24% ID/g (4 hours) and 1.92 ± 0.25% ID/g (24 hours); and kidneys 1.17 ± 0.37% ID/g (4 hours) and 0.54 ± 0.07% ID/g (24 hours).

This study describes two very promising pan-bombesin peptidic ligands for radiolabeling with diagnostic and therapeutic radiometals. They may be used for targeted diagnosis and radiotherapy of bombesin receptor-positive tumors like prostate and breast cancer.

A variety of radiopeptides are currently being developed for the targeting of tumors (5, 6, 7, 8, 9, 10). Those based on bombesin are of interest, because bombesin receptors were shown to be overexpressed on a variety of frequently occurring tumors like breast cancer and prostate cancer. In addition, Markwalder and Reubi (11) not only found a massive gastrin-releasing peptide receptor overexpression in invasive prostate cancer tissue but also in the early stage of the disease, i.e., the prostatic intraepithelial neoplasia, which may open the possibility to localize an early event in prostate carcinogenesis.

Many reports on bombesin-based radiopeptides have been published recently (26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36), but there has been no study thus far on radiopeptides based on pan-bombesin ligands. These may be of interest, because not only the gastrin-releasing peptide (BB2)-receptor but also BB1-receptor and BB3-receptor were found to be overexpressed on human tumors as well (22).

In our present study we evaluated the slightly modified pan-bombesin ligand [d-Tyr⁶, β-Ala¹¹, Phe¹³, Nle¹⁴] bombesin(6, 7, 8, 9, 10, 11, 12, 13, 14) developed by the Jensen group (40, 41), by attaching DTPA(BZH1) and DOTA(BZH2) via a GABA spacer to the octapeptide. The chelators allow a high specific activity labeling with ¹¹¹In (DTPA). DOTA chelates a large number of radiometals with extremely high kinetic stability like ⁶⁸Ga for PET-studies; ¹¹¹In for SPECT; and ⁹⁰Y, ¹⁷⁷Lu, and other lanthanides for therapeutic applications.

The receptor binding profiles of [In^III]-BZH1 and [Y^III]-BZH2 to the three bombesin receptor subtypes BB1, BB2, and BB3 were tested on human tumor specimen preferentially expressing each of the three receptor subtypes. Both metallopeptides show high binding affinity to the gastrin-releasing peptide receptor and slightly lower binding affinities to BB1 receptor and BB3 receptor. The [Y^III]-DOTA-derivative shows a distinctly and significantly higher binding affinity to all three of the receptor subtypes compared with the [In^III]-DTPA-analogue. We assume that this is due to the extra negative charge at the NH₂ terminus, which was shown to lower the binding affinity of bombesin-based radiopeptides.⁴

Both radiopeptides internalize rapidly, and no significant difference could be found between the two when labeled with ¹¹¹In. In addition, no significant difference was observed when different radionuclides ([¹¹¹In/⁹⁰Y/¹⁷⁷Lu]-BZH2) were used with the DOTA-based peptide, indicating that there are no structural differences among these three radiopeptides.

For an optimized use of radiopharmaceuticals in targeted radiotherapy, not only an efficient internalization is of importance but also the trapping (residualization) of the radioactivity adds to the potential success of the treatment via targeted radiotherapy. Therefore, we studied the externalization of [¹¹¹In]-BZH1 as well as the identity of the externalized radioactivity. The rate of efflux of both compounds was shown to be rather fast, even faster for [¹¹¹In]-BZH1 than for [¹¹¹In]-BZH2. The externalized radioactivity consisted preferentially of [¹¹¹In(DTPA)]²⁻ as the only metabolite. This is in contrast to the metabolites found when incubated in human serum, and it is also in contrast to the chelator-conjugated somatostatin analogues that show only intact externalized peptide at least within the first 4 hours of externalization (43). At present we have no information on the particular enzymes responsible for the metabolism inside the cell.

An additional important aspect for the suitability of a radiopeptide used in targeted radiotherapy is its metabolic stability in human serum. High metabolic stability allows the radiopeptide to reach the target intact and in optimal concentrations. Both peptides were studied in the form of their ¹¹¹In-labeled versions in fresh human serum. Fig. 2 shows the disappearance kinetics of the two radiopeptides and the build-up and decay of the first metabolite (B). The metabolic stability of both radiopeptides is relatively low with half-lives of 2.0 hours for [¹¹¹In]-BZH1 and 2.3 hours for [¹¹¹In]-BZH2, respectively. The curve fitting procedure gave similar results for k₁ (k₁′) independent of whether it was derived from the disappearance of A or the build-up and decay of the first intermediate B. Whereas the k₁-values show little difference between the two radiopeptides, the rate of metabolic decay of the intermediate is influenced by the chelate. Reverse-phase HPLC analysis of the metabolites along with co-injection of the predefined metabolites allowed an identification of the decay products.

The definition of the cleavage sites and the characterization of the metabolites are of importance, because this knowledge allows us to define and synthesize peptides of enhanced metabolic stability. Because of lack of reliable access to liquid chromatography/mass spectrometry, we mainly relied on the synthesis of potential metabolites and their co-injection using reverse-phase HPLC and radiometric detection. This procedure allowed us to define the first degradation site of both peptides between β-Ala¹¹ and His¹² and the second degradation site between Gln⁷ and Trp⁸. The question arises as to which enzyme may be responsible for the processing of this amino acid sequence. A first experiment adding EDTA to the human serum/radiopeptide mixture slowed down the metabolic degradation distinctly indicating that the responsible enzyme is a metalloenzyme. The addition of a large excess of carnosine (β-Ala-His), a dipeptide, which is cleaved by carnosinase (49), competitively inhibits the metabolic process (Fig. 3). Therefore, we assume that carnosinase, which is present in human serum, is the enzyme cleaving the peptide at β-Ala¹¹-His¹². This result also supports the finding that the first enzymatic cleavage occurs between these 2 amino acids.

In biodistribution studies, a strong accumulation of all of the radiotracers in bombesin receptor-positive tissues and the xenografted tumor was observed. As a tumor model the AR4–2J rat pancreatic carcinoma cell line was used, which is known to express high levels of bombesin receptors (50). Lewis rats bearing the AR4–2J solid tumor showed a high and specific uptake of both ¹¹¹In-labeled peptides as well as [¹⁷⁷Lu]-BZH2 in the tumor (and other bombesin receptor-bearing organs and tissues like the pancreas, the stomach, and the intestines). The blood clearance of both radiopeptides is very fast with <0.015% ID/g remaining in the blood at 4 hours. Somewhat unexpectedly, the uptake in receptor-positive organs was higher for [¹¹¹In]-BZH1 compared with [¹¹¹In]-BZH2 by almost a factor of 2, although the latter has a higher binding affinity to all three of the receptor subtypes, and the rate of internalization is comparable for the two. BZH2 was developed for radiolabeling with therapeutic radionuclides like ⁹⁰Y and the lanthanides. ¹¹¹In was used as a surrogate of ⁹⁰Y, because it can be followed and measured with more certainty than the pure β-emitter ⁹⁰Y. The two show similar chemistry; nevertheless, some differences in the pharmacokinetics of radiopharmaceuticals labeled with the two and studied in comparison have been reported. ¹⁷⁷Lu, another therapeutic radiometal [β_energy = 0.49 MeV; γ-emissions 133 keV (7%); 208 keV (11%)], shows great promise in targeted radiotherapy. Therefore, we studied the biodistribution of [¹⁷⁷Lu]-BZH2. There was no significant difference between [¹¹¹In]-BZH2 and [¹⁷⁷Lu]-BZH2 indicating that we may expect similar results with [⁹⁰Y]-BZH2.

The competitive binding studies with cold peptides clearly demonstrated that the uptake of both radiopeptides in relevant target tissues is specific and receptor mediated. The residence times of both peptides in the tumor are not very long; this may originate from a somewhat low metabolic stability in the respective cells and may parallel the relatively low serum stability. Relatively fast tumor washout was also found for other radiometal-labeled, bombesin-based radiopeptides, for instance, ^99mTc-labeled bombesin(7, 8, 9, 10, 11, 12, 13, 14) derivatives (34, 35). The identification of the “weakest” peptide bond at β-Ala¹¹-His¹² will help to improve the stability in future developments.

Tumor–normal tissue radioactivity ratios were very high for both radiopeptides. Ratios of tumor–blood of and tumor–muscle of >100 indicate that an early scintigraphic detection with a low background should be feasible. These ratios are higher than for other bombesin-based radiopeptides published recently using the PC-3 (33, 34, 36) and AR4–2J (51) tumor mouse model and argue for an early human use of these compounds. For instance, the tumor–muscle ratio for ^99mTc-RP527 [N,N-dimethyl-Gly-Ser-Cys-Gly-5-aminovaleric acid-bombesin(7, 8, 9, 10, 11, 12, 13, 14)] in a AR4–2J mouse tumor model was 23.5 at 4 hours and 73 at 24 hours. Our values of 171 (for ¹¹¹In-BZH1) and 103 (for ¹¹¹In-BZH2) at 4 hours and 90 (98) at 24 hours compare well with these data. In addition, both radioligands appear to interact with the human gastrin-releasing peptide receptor (PC-3 cell line; 2.5 ± 0.6 × 10⁵ gastrin-releasing peptide receptor-binding sites per cell; ref. 52) in a way very similar to that with the rat gastrin-releasing peptide receptor (∼1.5 × 10⁵ gastrin-releasing peptide receptor binding sites per cell; ref. 50). Indeed, first clinical applications in breast and prostate cancer patients, if labeled with ⁶⁸Ga or ¹¹¹In, showed very promising tumor localizations. Because we do not expect that the performance with the therapeutic radionuclides is much different, [⁹⁰Y, ¹⁷⁷Lu]-BZH2 may already be good candidates for targeted radiotherapy in patients. However, because the localization of neuromedin B receptor and BB3 receptor in normal human tissues is virtually unexplored, one cannot yet exclude the appearance of unwanted side effects related to yet unknown physiologic bombesin targets. In addition, one may argue that ¹⁷⁷Lu with its long physical half-life of 6.65 days may not be the ideal therapeutic radionuclide considering the relatively low metabolic stability of the two radiopeptides. ⁹⁰Y (half-life = 64 hours) and other radiolanthanides like ¹⁶⁶Ho (half-life = 27 hours) or ¹⁴⁹Pm (half-life = 53 hours), which also can be labeled to the DOTA-modified peptides, may be more suitable.

We thank Evelyn Peter for reviewing the English grammar.