Background

Gastrin-releasing peptide receptor (GRPR) is a G protein-coupled receptor, expressed in pancreas, gastrointestinal tract, and central nervous system, and involved in physiological functions such as synaptic plasticity, hormone secretion, and smooth muscle contraction (Jensen et al. 2008; Bitar and Zhu 1993; Weber 2009). Overexpression of GRPR has been reported to induce cancer cell proliferation and facilitate malignant neoplasm development (Jensen et al. 2008; Weber 2009; Cornelio et al. 2007; Hajri et al. 1996; Moody et al. 1996; Preston et al. 1995, 1994; Gugger and Reubi 1999; Markwalder and Reubi 1999; Roesler et al. 2006; Shimoda 1992; Qin et al. 1994). The overexpression of GRPR in various tumors makes it a promising target for the design of targeted radiopharmaceuticals for diagnosis and radioligand therapy of GRPR-expressing cancers.

Two natural ligands, gastrin-releasing peptide (GRP) and bombesin (BBN) show high binding affinity towards GRPR and share the same heptapeptide sequence (Trp-Ala-Val-Gly-His-Leu-Met-NH2) at the C-terminus (Erspamer et al. 1972a, 1972b; McDonald et al. 1978). The C-terminal heptapeptide of GRP and BBN has been used as a template for designing GRPR-targeted radiopharmaceuticals for decades (Varvarigou et al. 2004; Baum et al. 2007; Kähkönen et al. 2013; Stoykow et al. 2016; Baratto et al. 2021; Kurth et al. 2020; Nock et al. 2017; Marsouvanidis et al. 2013). The derivatives of GRP and BBN have been radiolabeled for imaging with single photon emission computed tomography (SPECT) and positron emission tomography (PET), and some of them have also been radiolabeled with beta and alpha emitters for radiotherapeutic applications (McDonald et al. 1978; Baum et al. 2007; Kurth et al. 2020; Nock et al. 2017; Minamimoto et al. 2016; Lin et al. 2004). However, the current clinically validated GRPR-targeted radioligands show an extremely high uptake in pancreas (Baum et al. 2007; Kähkönen et al. 2013; Kurth et al. 2020; Nock et al. 2017; Minamimoto et al. 2016), which not only limits the detection of cancer lesions located in or adjacent to pancreas, but also lowers the maximum tolerated dose for targeted radioligand therapy. Our group recently reported 68Ga-labeled TacsBOMB2 based on a known pseudopeptide-bond-containing antagonist sequence [Leu13ψThz14]Bombesin(7–14), which showed significant lower pancreas uptake than the clinically validated [68Ga]Ga-RM2 (Wang et al. 2022). Replacing the reduced peptide bond (Leu13ψThz14) with an amide bond restores the GRPR agonist characterizations and the derived [68Ga]Ga-TacBOMB2 retained high GRPR binding affinity and low uptake in mouse pancreas (Wang et al. 2023).

The development of GRPR-targeted radiopharmaceuticals has been focused on using antagonist sequences as targeting vectors because of their potentially higher tumor uptake due to higher in vivo stability (Ghosh et al. 2019) and more binding sites than those available for agonists (Mansi et al. 2009), and/or less short term adverse effects (Chatalic et al. 2016; Mansi et al. 2013). However, agonists can be internalized upon binding to GRPR and lead to a longer tumor retention (Jensen et al. 2008; Mansi et al. 2013; Yang et al. 2011), which might be preferable especially for the development of radiotherapeutic agents. The in vivo instability of GRPR-targeted ligands is caused by enzymatic degradation especially by neutral endopeptidase 24.11 (NEP) (Chatalic et al. 2016; Nock et al. 2014). The reported cleavage sites including His12-Leu13, Trp8-Ala9 and Gln7-Trp8 for AMBA derivatives and Trp8-Ala9, Ala9-Val10 and Gln7-Trp8 for RM2 derivatives (Kähkönen et al. 2013; Linder et al. 2009).

We hypothesized that (1) replacing amino acids at the potential cleavage sites of our previously reported GRPR agonist [68Ga]Ga-TacBOMB2 ([68Ga]Ga-DOTA-Pip-D-Phe6-Gln7-Trp8-Ala9-Val10-Gly11-His12-Leu13-Thz14-NH2) with unnatural amino acids can improve in vivo stability and retain the agonist characteristics; and (2) the resulting stabilized [68Ga]Ga-TacBOMB2 derivatives can also retain the minimal pancreas uptake characteristics. Thus, in this study we first synthesized the GRPR-targeted sequence of TacBOMB2 (LW01085, D-Phe6-Gln7-Trp8-Ala9-Val10-Gly11-His12-Leu13-Thz14-NH2, Fig. 1) and systematically substituted the amino acids (Gln7, Trp8, Ala9, Val10, Gly11 and His12) at its potential cleavage sites with an unnatural amino acid. The derivatives with high GRPR binding affinity were coupled with the DOTA chelator and 4-amino-(1-carboxymethyl)piperidine (Pip) linker. The binding affinities of their Ga-complexed standards were further confirmed by in vitro competition assays, and their agonists characteristics were confirmed by calcium release assays. Finally, the lead candidates were radiolabeled with 68Ga and evaluated by PET imaging and ex vivo biodistribution studies using the GRPR-expressing PC-3 prostate cancer model.

Fig. 1
figure 1

Chemical structures and GRPR binding affinities (Ki, mean ± SD, n = 3) of A LW01085 and its derivatives with an unnatural amino acid substitution at B His12, C Val10, D Ala9, E Gln7, F Val10-Gly11, and G Trp8. The potential cleavage sites of LW01085 are pointed by black arrows

Full size image

Methods

Synthesis of GRPR-targeted ligands

Detailed procedures for the synthesis, purification, and characterization of GRPR-targeted ligands and their natGa/68Ga-complexed analogs are provided in the Supplementary Information (Additional file 1: Figs. S1-S46 and Tables S1-S4).

Cell culture

The PC-3 cells obtained from ATCC (via Cedarlane, Burlington, Canada) were cultured in RPMI 1640 medium (Life Technologies Corporations, Carlsbad, CA, USA) supplemented with 10% FBS, penicillin (100 U/mL) and streptomycin (100 μg/mL) at 37 °C in a Panasonic Healthcare (Tokyo, Japan) MCO-19AIC humidified incubator containing 5% CO2 and 95% air. The cells were confirmed pathogen-free via IMPACT Rodent Pathogen Test (IDEXX BioAnalytics, Columbia, MO, USA). Cells grown to 80–90% confluence were washed with sterile Dulbecco’s phosphate-buffered saline (DPBS, pH 7.4) and collected after 1-min trypsinization. The cell concentration was counted using Moxi mini automated cell counter (ORFLO Technologies, Ketchum, IDUSA).

In vitro competition binding assays

Inhibition constants (Ki) of GRPR-targeted ligands to GRPR were measured by in vitro competition binding assay using PC-3 cells and [125I-Tyr4]Bombesin as the radioligand following previously published procedures (Wang et al. 2022, 2023; Bratanovic et al. 2022). The assays were conducted in triplicate with varied concentrations (10 μM to 1 pM) of tested ligands. Briefly, PC-3 cells were seeded in 24-well poly-D-lysine plates at 2 × 105 cells/well 24–48 h prior to the assay. The growth medium was replaced with 400 μL of reaction medium (RPMI 1640 containing 2 mg/mL BSA, and 20 mM HEPES), then the plates were incubated at 37 °C for 60 min. The tested ligands in 50 μL reaction medium and 50 μL of 0.01 nM [125I-Tyr4]Bombesin were added into the wells followed by incubation with moderate agitation for 1 h at 37 °C. Cells were gently washed with ice-cold DPBS twice, harvested by trypsinization, and counted for radioactivity on a Perkin Elmer (Waltham, MA, USA) Wizard2 2480 automatic gamma counter. Data were analyzed using nonlinear regression with GraphPad (San Diego, CA, USA) Prism 8 software.

Fluorometric calcium release assays

Following previously published procedures (Bratanovic et al. 2022; Lau et al. 2019), 5 × 104 PC-3 cells were seeded in 96-well clear bottom black plates 24 h prior to the assay. The growth medium was removed and replaced with a loading buffer containing a calcium-sensitive dye (FLIPR Calcium 6 assay kit from Molecular Devices, San Jose, CA, USA). After incubated at 37 °C for 30 min, the plates were placed in a FlexStation 3 microplate reader (Molecular Devices). Tested ligands (50 nM) or DPBS (negative control) were added to the cells and the fluorescent signals were acquired for 2 min. Agonistic/antagonistic properties of the tested ligands were determined based on the relative fluorescent unit (RFU = max – min) of their generated fluorescent signals.

LogD7.4 measurements

The LogD7.4 values of 68Ga-labeled tracers were measured using the shake flask method as previously published (Lin et al. 2015). Briefly, aliquots (2 μL) of the 68Ga-labeled tracers were added into a 15 mL falcon tube containing 3 mL of n-octanol and 3 mL of DPBS (pH 7.4). The mixture was vortexed for 1 min and then centrifuged at 3,000 rpm for 15 min. Samples of the n-octanol (1 mL) and DPBS (1 mL) layers were collected and measured in a Perkin Elmer Wizard2 2480 automatic gamma counter. LogD7.4 was calculated with the following equation: LogD7.4 = log10[(counts in n-octanol phase)/(counts in DPBS phase)].

Biodistribution, PET imaging, and in vivo stability studies

PET/CT imaging, biodistribution, and in vivo stability studies were conducted on male NOD.Cg-Rag1tm1Mom Il2rgtm1Wjl/SzJ (NRG) mice following previously published procedures (Bratanovic et al. 2022; Lau et al. 2019; Lin et al. 2015; Kuo et al. 2018). The experiments were conducted according to the guidelines established by the Canadian Council on Animal Care and approved by Animal Ethics Committee of the University of British Columbia. The mice were anaesthetized by inhalation of 2.5% isoflurane in 2 mL/min oxygen, and implanted subcutaneously with 5 × 106 PC-3 cells (100 µL; 1:1 PBS/Matrigel) behind the left shoulder. Mice were used for PET/CT imaging and biodistribution studies when the tumor grew to 5–8 mm in diameter over around 4 weeks.

PET imaging experiments were conducted using a Siemens Inveon (Knoxville, TN, USA) micro PET/CT scanner. Each tumor bearing mouse was injected with 3–5 MBq (90.6–166.8 ng) of 68Ga-labeled tracer via the lateral caudal tail vein under anaesthesia (2% isoflurane in oxygen). For blocking, the mice were co-injected with 100 μg of [D-Phe6,Leu-NHEt13,des-Met14]Bombesin(6–14). The mice were allowed to recover and roam freely in their cages. After 50 min, the mice were sedated again with 2% isoflurane in oxygen inhalation and positioned on the scanner. A 10-min CT scan was conducted first for localization and attenuation correction after segmentation for reconstructing the PET images, followed by a 10-min static PET imaging acquisition.

For biodistribution studies, the mice were injected with 2–4 MBq (50.2–106.8 ng) of radiotracer as described above. For blocking, the mice were co-injected with [D-Phe6,Leu-NHEt13,des-Met14]Bombesin(6–14) (100 μg). At 1 h post-injection, the mice were anesthetized with 2% isoflurane inhalation, and euthanized by CO2 inhalation. Blood was withdrawn by cardiac puncture, and organs/tissues of interest were collected. The collected organs/tissues were rinsed with PBS, blotted dry, weighed, and counted using the automatic gamma counter.

For in vivo stability studies, the 68Ga-labeled ligand (6–15 MBq) was injected via the lateral caudal vein into healthy male NRG mice (n = 3). At 15 min post-injection, mice were euthanized, and the urine and blood samples were collected. The plasma was extracted from whole blood samples by the addition of CH3CN (500 μL), 1-min vortex, 20-min centrifugation, and the separation of supernatant. The plasma and urine samples were analyzed via radio-HPLC using the conditions for quality control (Additional file 1: Table S4).

Statistical analysis

Statistical analyses were performed by Student’s t-test using the Microsoft (Redmond, WA, USA) Excel software. The comparison of biodistribution data between two tracers was conducted via unpaired two-tailed test. Unpaired one-tailed test was used to compare biodistribution data between blocked/unblocked mice injected with the same tracer. The difference was considered statistically significant when the p value was < 0.05.

Results

GRPR binding affinities of LW01085 and its derivatives

As shown in Fig. 1, the GRPR binding affinity (Ki) of LW01085 is 8.77 ± 1.33 nM. Tle10 substitution improves binding affinity (LW01080: 3.56 ± 0.72 nM) while NMe-His12 (LW01088: 9.00 ± 1.99 nM), 2-Me-Trp8 (LW02009: 11.5 ± 1.71 nM), 7-F-Trp8 (LW01177: 9.38 ± 1.68 nM) and 5-Me-Trp8 (LW01182: 10.5 ± 1.86 nM) substitutions lead to analogs with comparable binding affinities. The other substitutions generate analogs with either slightly reduced (Ki = 13.5–53.1 nM for LW02019, LW01078, LW01136, LW01183, LW02007, LW01175, LW01166, LW01180 and LW02015) or greatly reduced binding affinities (Ki > 150 nM for LW02016, LW02011, LW01083, LW01075, LW02030, LW01128, LW01137, LW01191, LW01173, LW01171 and LW02013).

GRPR binding affinities of Ga-TacBOMB2 derivatives

As shown in Fig. 2 and Additional file 1: Figs. S47–S48, compared with the previously reported Ga-TacBOMB2 (Ki = 7.62 ± 0.19 nM) (Wang et al. 2023), the NMe-His12 and Tle10 substitutions, either alone, combined or combined with an additional His7 or 7-F-Trp8 substitution generate analogs with an enhanced binding affinity (Ga-LW01107: 2.98 ± 0.69 nM; Ga-LW01108: 1.34 ± 0.12 nM; Ga-LW01110: 1.39 ± 0.03 nM; Ga-LW01142: 3.19 ± 0.78 nM; Ga-LW02040: 2.87 ± 0.09 nM). The substitutions of 7-F-Trp8, 5-Me-Trp8 and 2-Me-Trp8 lead to Ga-LW02021, Ga-LW02023 and Ga-LW02025, respectively, with a slightly reduced binding affinity (Ki = 13.6–14.9 nM). The derivatives containing an αMe-Trp8 (Ga-LW01149) or NMe-Gly11 substitution (Ga-LW01143) have poor binding affinities (Ki > 300 nM).

Fig. 2
figure 2

Chemical structures and GRPR binding affinities (Ki, mean ± SD, n = 3) of A Ga-TacBOMB2 and its derivatives with an unnatural amino acid substitution at B His12, C Val10, D Trp8, E Val10 and His12, F Gln7, Val10 and His12, G Gln7, Val10, Gly11 and His12, and H Trp8, Val10 and His12

Full size image

Confirmation of agonist characteristics of Ga-TacBOMB2 derivatives

To confirm the agonist characteristics of Ga-TacBOMB2 derivatives, calcium release assays were conducted using PC-3 cells. As shown in Fig. 3, Ga-LW01107, Ga-LW01108, Ga-LW01110, Ga-LW01142, Ga-LW02021, Ga-LW02023, Ga-LW02025, and Ga-LW02040 induced Ca2+ efflux corresponding to 1,004 ± 32.0, 549 ± 46.7, 521 ± 43.7, 559 ± 96.2, 178 ± 53.7, 312 ± 45.9, 197 ± 50.3, and 242 ± 44.1 relative fluorescence units (RFUs), respectively. The RFUs for the blank control (DPBS), antagonist control ([D-Phe6,Leu-NHEt13,des-Met14]Bombesin(6–14), and positive controls (ATP and bombesin) are 6.57 ± 1.66, 38.2 ± 7.20, 253 ± 46.5 and 450 ± 136, respectively. Therefore, all tested Ga-TacBOMB2 derivatives are confirmed to be GRPR agonists as they induced comparable or higher calcium release than the positive control ATP.

Fig. 3
figure 3

Intracellular calcium efflux in PC-3 cells induced by GRPR-targeted ligands. Cells were incubated with DPBS or 50 nM of Ga-complexed GRPR-targeted ligand, [D-Phe6,Leu-NHEt13,des-Met14]Bombesin(6–14), bombesin, or ATP

Full size image

PET imaging and biodistribution

The capability of 68Ga-labeled TacBOMB2 derivatives to target GRPR in vivo was evaluated by PET imaging and biodistribution studies in mice bearing GRPR-expressing PC-3 tumor xenografts. As shown in Fig. 4, all 68 Ga-labeled tracers enabled visualization of PC-3 tumors with good tumor-to-background contrasts. These tracers were excreted mainly via the renal pathway and had only low to moderate uptake in pancreas. Higher tumor uptake was observed by using [68Ga]Ga-LW01110, [68Ga]Ga-LW02040 and [68Ga]Ga-LW01142, followed by [68Ga]Ga-LW01107 and [68Ga]Ga-LW01108, and [68Ga]Ga-LW02021 had the lowest tumor uptake. [68Ga]Ga-LW01142 which showed high blood retention at 1 h post-injection was further evaluated at 3 h post-injection (Fig. 4F). The tumor uptake of [68Ga]-LW01142 increased further at 3 h post-injection, leading to an enhanced tumor-to-background contrast. Co-injection of [D-Phe6,Leu-NHEt13,des-Met14]Bombesin(6–14) reduced tumor uptake of both [68Ga]Ga-LW01110 and [68Ga]Ga-LW01142 at 1 h post- injection (Figs. 4C and F).

Fig. 4
figure 4

Representative PET images of A [68Ga]Ga-LW01107, B [68Ga]Ga-LW01108, C [68Ga]Ga-LW01110, D [68Ga]Ga-LW02040, E [68Ga]Ga-LW02021 and F [68Ga]Ga-LW01142 in mice bearing PC-3 tumor xenografts. Blocking study was performed by co-injection with 100 μg of [D-Phe6,Leu-NHEt13,des-Met14]Bombesin(6–14). t: tumor; k: kidney; p: pancreas; bl: urinary bladder

Full size image

The biodistribution data of 68Ga-labeled GRPR-targeted tracers in PC-3 tumor-bearing mice obtained at 1 h post-injection are provided in Additional file 1: Table S5, and the previously reported data obtained from [68Ga]Ga-TacBOMB2, [68Ga]Ga-RM2 and [68Ga]Ga-AMBA are included for comparison (Wang et al. 2022, 2023). In consistent with the observations from PET images, all 68Ga-labeled LW01107, LW01108, LW01110, LW01142, LW02021, and LW02040 had significantly lower uptake in pancreas (0.39 ± 0.03, 9.32 ± 1.97, 8.99 ± 1.54, 4.40 ± 0.27, 1.22 ± 0.18 and 11.7 ± 0.47 %ID/g, respectively) than [68Ga]Ga-RM2 (41.9 ± 10.1 %ID/g) and [68Ga]Ga-AMBA (62.4 ± 4.26 %ID/g). [68Ga]Ga-LW01110 showed the highest tumor uptake (16.6 ± 1.60 %ID/g), followed by [68Ga]Ga-LW02040 (12.3 ± 2.14 %ID/g) and [68Ga]Ga-LW01142 (11.4 ± 1.22 %ID/g). [68Ga]Ga-LW01110 also had the best tumor-to-organ uptake ratios (134 ± 16.7, 119 ± 22.6, 24.7 ± 4.17 and 5.10 ± 0.39 for tumor-to-bone, tumor-to-muscle, tumor-to-blood and tumor-to-kidney, respectively). With the lowest uptake in pancreas (0.39 ± 0.03 %ID/g), [68Ga]Ga-LW01107 showed the highest tumor-to-pancreas uptake ratio (17.9 ± 1.10).

As [68Ga]Ga-LW01142 had high blood pool uptake at 1 h post-injection (6.88 ± 0.29 %ID/g), its biodistribution was further evaluated at 3 h post-injection. The tumor uptake increased (11.4 ± 1.22 to 15.3 ± 2.45 %ID/g) and the uptake in other organs/tissues decreased at 3 h post-injection (Figs. 4F and 5; Additional file 1: Table S6), leading to enhanced tumor-to-background contrast ratios. The tumor-to-bone, tumor-to-muscle, tumor-to-blood, tumor-to-kidney, and tumor-to-pancreas ratios of [68Ga]Ga-LW01142 at 3 h post-injection were 91.6 ± 12.2, 86.6 ± 25.1, 6.40 ± 1.70, 3.17 ± 0.46 and 7.36 ± 1.17, respectively.

Fig. 5
figure 5

Uptake (mean ± SD, n = 4) of [68Ga]Ga-LW01142 at 1 and 3 h post-injection in PC-3 tumor-bearing mice

Full size image

Blocking studies of [68Ga]Ga-LW01110 and [68Ga]Ga-LW01142 were conducted at 1 h post-injection (Additional file 1: Table S6). The results showed that co-injection of [D-Phe6,Leu-NHEt13,des-Met14]Bombesin(6–14) reduced their average tumor uptake values by 44% and 31%, respectively. The average pancreas uptake values of [68Ga]Ga-LW01110 and [68Ga]Ga-LW01142 were also reduced by 42% and 30%, respectively.

LogD7.4 measurement and in vivo stability

LogD7.4 measurements were conducted for the 68Ga-labeled tracers, and as shown in Table 1, these 68Ga-labeled tracers are highly hydrophilic with average LogD7.4 values in the range of -3.10 to -1.81. In vivo stability studies showed that, compared with the previously reported [68Ga]Ga-TacBOMB2 (12.7 ± 2.93% intact at 15 min post-injection), Tle10 ([68Ga]Ga-LW01108) and NMe-His12 ([68Ga]Ga-LW01107) substitutions increased the intact tracer fraction in mouse plasma to 35.3 ± 0.93 and 66.2 ± 12.4%, respectively (Additional file 1: Figs. S49-S50 and Table 1). Further improvement was obtained by combining both Tle10 and NMe-His12 as ≥ 89% intact was observed for [68Ga]Ga-LW01110, [68Ga]Ga-LW01142 and [68Ga]Ga-LW02040 (Additional file 1: Figs. S51-S53 and Table 1). No intact tracer was detected in mouse urine samples for all the tested GRPR-targeted ligands (Additional file 1: Figs. S49-S53).

Table 1 LogD7.4 values and in vivo stability of GRPR-targeted tracers
Full size table

Discussion

In this study, we first compared LW01085 (D-Phe-[Thz14]Bombesin(7–14), the pharmacophore of our previously reported [68Ga]Ga-TacBOMB2) with 25 derivatives with an unnatural amino acid substitution at the potential cleavage sites (Fig. 1). In vitro competition binding assays showed that NMe-His12 substitution (LW01088) is tolerable, which is consistent with a previous report by Horwell, et al. (1996) that Ac-Bombesin(7–14) and Ac-[NMe-His12]Bombesin have similar GRPR binding affinities (Ki = 0.7 vs 0.4 nM). In addition, we discovered that 2-Me-Trp8 (LW02009), 7-F-Trp8 (LW01177), 5-Me-Trp8 (LW01182) and Tle10 (LW01080) substitutions also led to derivatives with enhanced or comparable GRPR binding affinities. Therefore, these unnatural amino acid substitutions were selected and subsequently compared/combined with two reported unnatural amino acid substitutions (αMe-Trp8 and NMe-Gly11) for the design of Ga-DOTA-complexed Pip-linker-containing GRPR-targeted ligands (Fig. 2).

Despite being popularly used for the design of GRPR-targeted antagonist ligands (Richter et al. 2016; Sah et al. 2015), NMe-Gly11 substitution was reported to cause > 30-fold reduction in GRPR binding affinity for an agonist sequence (Ki = 0.7 vs 25 nM for Ac-Bombesin(7–14) and Ac-[NMe-Gly11]Bombesin(7–14), respectively) (Horwell et al. 1996). Consistent with the previous report, we also observed a dramatic reduction in GRPR binding affinity with the NMe-Gly11 substitution (Figs. 2F and G, Ki = 3.19 vs 12,790 nM for Ga-LW01142 and Ga-LW01143, respectively). αMe-Trp8 substitution has been successfully used by the Wester group for the design of potent and stable radiolabeled GRPR-targeted antagonists derived from RM2 (Guenther et al. 2022). However, for agonist Ga-TacBOMB2, αMe-Trp8 substitution in Ga-LW01149 caused significant loss of binding affinity (Ki = 7.62 vs 342 nM, Figs. 2A and D). Our data suggest that GRPR agonists and antagonists might bind to the receptors in different configurations as αMe-Trp8 and NMe-Gly11 substitutions which are commonly used for antagonist modifications hinder the binding of agonists to the receptors.

His7 (the amino acid at the corresponding position in GRP), 2-Me-Trp8, 7-F-Trp8, 5-Me-Trp8, Tle10 and NMe-His12 substitutions, either alone or in combination, still led to GRPR-targeted ligands with potent binding affinities (Ki = 1.34–14.9 nM, Fig. 2). This suggests that compared with the targeted peptide sequences presented in Fig. 1, the addition of Ga-DOTA complex and the Pip linker does not affect their binding affinity. Similarly, based on the results of calcium release assays (Fig. 3), His7, 2-Me-Trp8, 7-F-Trp8, 5-Me-Trp8, Tle10 and NMe-His12 substitutions, either alone or in combination, do not change their agonist characteristics.

Subsequently, we radiolabeled potent candidates and evaluated their potential for prostate cancer imaging. As shown in Fig. 4, all 68Ga-labeled tracers were successfully used to visualize PC-3 tumor xenografts in their PET images, confirming good GRPR targeting capabilities of these tracers. A lower tumor uptake was observed for [68Ga]Ga-LW02021, which could be due to its relatively weaker GRPR binding affinity compared with those of others (Ki = 13.6 vs 1.34 – 3.19 nM, Fig. 2). The clearance of these tracers was mainly via the renal pathway, consistent with the highly hydrophilic nature of these tracers (LogD7.4 values ≤ -1.81). A higher blood retention was observed for [68Ga]Ga-LW01142 at 1 h post-injection, which could be due to its relatively higher lipophilicity than other tracers (LogD7.4 = -1.81 vs -2.46 – -3.10, Table 1).

Ex vivo biodistribution studies were also conducted to better quantify uptake in tumors and normal organs/tissues. As shown in Additional file 1: Table S5, except [68Ga]Ga-LW02021 (3.08 ± 0.48 %ID/g at 1 h post-injection), all other evaluated tracers had comparable or improved uptake in PC-3 tumors when compared to that of the previously reported [68Ga]Ga-TacBOMB2 (5.95 ± 0.05 %ID/g). Notably, while Tle10 substitution led to [68Ga]Ga-LW01108 (5.90 ± 0.68 %ID/g) with a comparable tumor uptake, NMe-His12 led to [68Ga]Ga-LW01107 (7.05 ± 0.71 %ID/g) with an improved tumor uptake. Most importantly, the combination of both Tle10 and NMe-His12 with and without an addition substitution (His7 or 7-F-Trp8) led to [68Ga]Ga-LW01110 (16.6 ± 1.66 %ID/g), [68Ga]Ga-LW01142 (11.4 ± 1.22 %ID/g) and [68Ga]Ga-LW02040 (12.3 ± 2.14 %ID/g) with a further improved tumor uptake. Since 68Ga-labeled LW01107, LW01108, LW01110, LW01142 and LW02040 have comparable GRPR binding affinities (Ki = 1.34–3.19 nM), we suspected that the greatly improved tumor uptake for tracers with at least both Tle10 and NMe-His12 substitutions could be mainly due to their improved in vivo stability.

In vivo stability studies were subsequently conducted to verify our hypothesis. As shown in Table 1, compared with the previously reported [68Ga]Ga-TacBOMB2 (12.7 ± 2.93% intact tracer at 15 min post-injection), Tle10 and NMe-His12 substitutions led to [68Ga]Ga-LW01108 (35.3 ± 0.93% intact) and [68Ga]Ga-LW01107 (66.2 ± 12.4% intact) with an improved in vivo stability. Combination of at least both Tle10 and NMe-His12 substitutions further led to [68Ga]Ga-LW01110, [68Ga]Ga-LW01142 and [68Ga]Ga-LW02040 with an average ≥ 89% intact tracer at 15 min post-injection. These data are consistent with the trend of their tumor uptake observed from the ex vivo biodistribution studies: [68Ga]Ga-TacBOMB2 ≈ [68Ga]Ga-LW01108 < [68Ga]Ga-LW01107 < [68Ga]Ga-LW01110, [68Ga]Ga-LW01142 and [68Ga]Ga-LW02040. In addition, our in vivo stability data also suggest that His12-Leu13 is the major cleavage site of GRPR-targeted ligands, followed by Ala9-Val10, and then Gln7-Trp8/Trp8-Ala9. This is also consistent with the fact that most of reported GRPR-targeted radioligands had modifications to avoid the cleavage at His12-Leu13 such as using Sta13 substitution for the RM2 derivatives (Kähkönen et al. 2013; Mansi et al. 2011) and Leu13ψThz14 in our previously reported TacsBOMB derivatives (Wang et al. 2022). Contrary to the improved stability observed in plasma, no intact tracer was detected in urine samples even for ligands with both Tle10 and NMe-His12 substitutions (Additional file 1: Figs. S49–S53). This is due to the facts that GRPR-targeted ligands are cleaved mainly by NEP and kidneys have the highest NEP expression level (Jiang et al. 2004). Therefore, GRPR-targeted tracers which remain intact in plasma are completely metabolized by NEP in kidneys before being excreted into the urinary bladder.

Compared with the clinically validated [68Ga]Ga-RM2 and [68Ga]Ga-AMBA (Wang et al. 2022, 2023), our stabilized tracers ([68Ga]Ga-LW01110, [68Ga]Ga-LW01142 and [68Ga]Ga-LW02040) have not only higher tumor uptake, but also comparable or even higher tumor-to-background uptake ratios (Additional file 1: Tables S2–S3). Most importantly, they also have a much lower pancreas uptake than [68Ga]Ga-RM2 and [68Ga]Ga-AMBA (4.40–11.7 vs 41.9–62.4%ID/g at 1 h post-injection). Therefore, these tracers are expected to have a higher sensitivity for detecting cancer lesions in or adjacent to the pancreas, and can achieve better treatment efficacy and cause less damage to the pancreas when radiolabeled with an α- or β-emitter for radiotherapeutic applications.

Conclusions

We systematically replaced the amino acids (Gln7, Trp8, Ala9, Val10, Gly11 and His12) at potential cleavage sites of the previously reported sequence of [68Ga]Ga-TacBOMB2, and identified that Tle10 and NMe-His12 substitutions, either alone or in combination, led to derivatives with comparable/enhanced GRPR binding affinities. In vivo stability and ex vivo biodistribution studies confirmed the improved stability resulted from unnatural amino acid substitutions, which further led to enhanced tumor uptake. With both Tle10 and NMe-His12 substitutions, the top candidate [68Ga]Ga-LW01110 has higher in vivo stability, tumor uptake and tumor-to-background uptake ratios than clinically validated [68Ga]Ga-RM2 and [68Ga]Ga-AMBA, and is promising for use for detecting GRPR-expressing tumors with PET. Due to the observed lower pancreas uptake and foreseeable longer tumor retention as being agonists, our optimized sequence, [Tle10,NMe-His12,Thz14]Bombesin(7–14), is a promising template for use for the design of GRPR-targeted radiotherapeutic agents.