Improved Tumor-Targeting with Peptidomimetic Analogs of Minigastrin 177Lu-PP-F11N

Simple Summary Several radiolabeled peptides targeting CCK2R-positive types of cancer (such as medullary thyroid cancer and small cell lung cancer) have been reported in the last 25 years, some of which have entered clinical trials. In an effort to improve its tumor-targeting properties, we applied chemical modifications to the backbone of the peptide 177Lu-PP-F11N, an analog of minigastrin in clinical trials. The generated radiolabeled peptidomimetics showed significantly improved characteristics in mice bearing CCK2R-positive tumor xenografts, such as higher tumor uptake, slower tumor washout, and increased tumor-to-kidney ratios. These properties make the novel compounds promising candidates for the imaging and therapy of CCK2R-positive tumors and metastases. Abstract The cholecystokinin-2 receptor (CCK2R) is an attractive target in nuclear medicine due to its overexpression by different tumors. Several radiolabeled peptidic ligands targeting the CCK2R have been investigated in the past; however, their low stability against proteases can limit their uptake in tumors and metastases. Substitution of single or multiple amide bonds with metabolically stable 1,4-disubstituted 1,2,3-triazoles as amide bond bioisosteres proved a promising strategy for improving the tumor-targeting properties of a truncated analog of minigastrin. In this study, we applied the previously studied structural modifications to improve the pharmacokinetic and pharmacodynamic properties of PP-F11N, a minigastrin analog currently in clinical trials. Novel minigastrins (NMGs) as analogs of PP-F11N with one or two amide bonds substituted by 1,2,3-triazoles were synthesized, radiolabeled with 177Lu3+, and subjected to full evaluation in vitro (cell internalization, receptor affinity, stability in blood plasma) and in vivo (stability, biodistribution, SPECT/CT imaging). NMGs with triazoles inserted between the amino acids DGlu10-Ala11 and/or Tyr12-Gly13 showed a significantly increased cellular uptake and affinity toward the CCK2R in vitro. Resistance against the metabolic degradation of the NMGs was comparable to those of the clinical candidate PP-F11N. Imaging by SPECT/CT and biodistribution studies demonstrated a higher uptake in CCK2R-positive tumors but also in the CCK2R-positive stomach. The peptidomimetic compounds showed a slow tumor washout and high tumor-to-kidney ratios. The structural modifications led to the identification of analogs with promising properties for progression to clinical applications in the diagnosis and therapy of CCK2R-positive neoplasms.


Introduction
The cholecystokinin-2 receptor (CCK2R) is a target of high interest in nuclear medicine. Several types of cancer (e.g., medullary thyroid cancer (MTC) and small cell lung cancer (SCLC)) have been reported to express the CCK2R with high incidence and density [1,2]. As these types of cancers can remain asymptomatic over long periods of time, patients are often diagnosed only at late and metastasized stages of the disease [3,4]. Typical treatment options such as chemotherapy, surgical resection, and external beam radiation show only a moderate anti-tumor effect, leading to an overall low survival [5,6]. In light of these facts, reliable methods for early diagnosis and efficient treatment options are urgently needed. Several radiolabeled peptide conjugates targeting the CCK2R have been investigated during the past two decades, [7] some of which have meanwhile entered clinical trials [8][9][10][11]. Although remarkable progress has been achieved, two main challenges in the development of CCK2R-targeting radiopeptides remain: First, the accumulation of radioactivity in the kidneys due to renal elimination and tubular reabsorption mechanisms is a major concern for peptide receptor radionuclide therapy (PRRT) as it can cause nephrotoxicity [12,13]. Second, the rapid proteolytic degradation of the radiotracer after intravenous administration reduces the amount of intact conjugate reaching the tumor, thus impairing the uptake of radiolabeled minigastrins in CCK2R-positive tissues, such as a tumor or the stomach [14,15]. These two challenges must be addressed for the development of new radioligands targeting the CCK2R.
In this study, we applied these amide-to-triazole substitutions to PP-F11N (DOTA-(DGlu) 6 -Ala 11 -Tyr 12 -Gly 13 -Trp 14 -Nle 15 -Asp 16 -Phe 17 -NH 2 ) [21], a minigastrin derivative with five additional DGlu at the N-terminus in comparison to [Nle 15 ]MG11. The hexa-D-glutamate linker improves the metabolic stability of PP-F11N without causing high renal uptake observed for the hexa-L-glutamate analog [22]. Results from phase 0 clinical studies indicate that 177 Lu-labeled PP-F11N can accumulate in tumors of MTC in potentially therapeutic doses [10]. These promising results led to the progression of the radio-conjugate to phase I clinical trials, which are currently ongoing [8].

Synthesis of the NMGs
NMGs 1-3 were synthesized and radiolabeled following published procedures (Supporting Information (SI), Scheme S1) [16,17]. Details on the syntheses and analytical data for all compounds can be found in the Supplementary Materials Table S1.

(Radio) Metal Labeling, and Evaluation In Vitro
Procedures for (radio)metal labeling of the DOTA-minigastrin conjugates with 177 Lu 3+ and nat Lu 3+ , cell culturing of A431-CCK2R cells [23], cell internalization (including blocking experiments with excess minigastrin (H-LE5AYGWMDF-NH2)), competition binding assays (IC50), stability measurements in human blood plasma, and determinations of logDpH 7.4 have been reported previously for the in vitro evaluation of TZMGs and were applied to NMGs without alteration [16,17].

In Vivo Stability
All animal experiments were performed in compliance with Swiss laws on animal protection and approved by the veterinary office of the canton of Aargau (license-Nr. AG 75700).

Synthesis of the NMGs
NMGs 1-3 were synthesized and radiolabeled following published procedures (Supporting Information (SI), Scheme S1) [16,17]. Details on the syntheses and analytical data for all compounds can be found in the Supplementary Materials Table S1.

(Radio) Metal Labeling, and Evaluation In Vitro
Procedures for (radio)metal labeling of the DOTA-minigastrin conjugates with 177 Lu 3+ and nat Lu 3+ , cell culturing of A431-CCK2R cells [23], cell internalization (including blocking experiments with excess minigastrin (H-LE 5 AYGWMDF-NH 2 )), competition binding assays (IC 50 ), stability measurements in human blood plasma, and determinations of logD pH 7.4 have been reported previously for the in vitro evaluation of TZMGs and were applied to NMGs without alteration [16,17].

Biodistribution
Procedures for biodistribution studies with 177 Lu-labeled conjugates (10 pmol, 100 µL, 0.1 µM in PBS, 0.5 MBq, 1 µg/kg) in female CD1-nu mice with xenografts of A431-CCK2R cells at 1, 4, and 24 h p.i., as well as blocking experiments at 4 h p.i. with a 6000-fold excess of minigastrin have been reported previously and were followed without modifications [16]. Statistical analysis was performed via GraphPad Prism using two-sided analysis of variance. n = 4 animals per group. 2.5. SPECT/CT Imaging 10-14 days prior to the experiment, 6-week-old female CD1-nu mice were inoculated with 5 × 10 6 A431-CCK2R cells (0.1 mL, 5 × 10 7 cells/mL in PBS) subcutaneously in both shoulders. The mice were randomly assigned to groups of two animals. On the day of the experiment, 177 Lu-labeled conjugates (200 pmol, 100 µL, 2 µM in PBS, 20 MBq, 20 µg/kg) were injected via the tail vein. The mice were kept with food and water ad libitum. SPECT scans of 40 min (1 and 4 h p.i., frame duration 60 sec) or 60 min (24 h p.i., frame duration 80 sec) were acquired after an initial CT scan of 7.5 min (CT tube voltage 55 kV, CT tube current 145 µA) with a multi-pinhole NanoSPECT/CTplus (Bioscan Inc., France, and Mediso, Hungary). For the duration of the SPECT/CT scans, mice were kept under anesthesia with a mixture of isoflurane (1.5-2.5%) and oxygen and warmed by a constant air flow of 37 • C. The images were reconstructed with HiSPECT and analyzed in VivoQuant. A post-reconstruction filter was applied (Gauss, full width at half max, 1 mm) to the SPECT images, and the scale was adjusted to visualize important organs and tissues (0.001-0.004 kBq/voxel). n = 2 animals per group.

Synthesis and Radio-labeling
The NMGs were synthesized by solid-phase peptide synthesis in satisfactory yields (11-21%) and high purities (>95%; Supplementary Materials Table S1). Radio-labeling with 177 Lu 3+ was achieved in high radiochemical yields and purities of >95% as determined by γ-HPLC and at molar activities of up to 100 MBq/nmol (peptide: nuclide ratio = 7:1; SI, Figures S1-S5). Table 1 and Figure 2 summarize the physio-chemical properties of the NMGs in vitro in comparison with PP-F11N. NMGs 2 and 3 showed a better affinity toward the CCK2R than the PP-F11N, which correlated with an increased cell internalization at all time points (SI, Figure S6). In contrast, NMG 1 had a higher IC 50 value compared to PP-F11N, and the cell internalization was reduced. More than 95% of intact radiolabeled conjugates were found for all compounds after 24 h of incubation in human blood plasma in vitro. Internalization was mediated by the CCK2R, as demonstrated by blocking experiments (SI, Tables S2-S5, Figure S7). Minor deviations of logD pH 7.4 values were not statistically significant (Supplementary Materials Table S6).  [24] and are included for comparison. f Expressed as % of applied radioactivity. g In vitro = 24 h incubation time, in vivo = 10 min p.i. of the radiotracer. Ψ[Tz] stands for a 1,4-disubstituted 1,2,3-triazole between two amino acids. a Lu = 175 Lu. b Lu = 177 Lu. c Data are presented as mean values (95% confidence interval (CI) of nonlinear regression). d Data are presented as mean values ± standard deviations (SD). e Data were reported previously [24] and are included for comparison. f Expressed as % of applied radioactivity. g In vitro = 24 h incubation time, in vivo = 10 min p.i. of the radiotracer. Ψ[Tz] stands for a 1,4-disubstituted 1,2,3-triazole between two amino acids.

Evaluation In Vivo
Metabolites of the 177 Lu-labeled conjugates were observed in samples of blood and urine of mice at 10 min p.i. of the radiotracers. The amount of intact conjugate (in %) detected in blood decreased in the following order: NMG 1 (52%) > PP-F11N (43%) > NMG 3 (39%) > NMG 2 (21%). Only negligible amounts of intact conjugates were observed in the urine. Representative γ-chromatograms are shown in Figure 3.

Evaluation In Vivo
Metabolites of the 177 Lu-labeled conjugates were observed in samples of blood and urine of mice at 10 min p.i. of the radiotracers. The amount of intact conjugate (in %) detected in blood decreased in the following order: NMG 1 (52%) > PP-F11N (43%) > NMG 3 (39%) > NMG 2 (21%). Only negligible amounts of intact conjugates were observed in the urine. Representative γ-chromatograms are shown in Figure 3.  Biodistribution experiments were conducted at 1, 4, and 24 h p.i. in mice with CCK2R-positive xenografts. All 177 Lu-labeled peptide conjugates showed rapid clearance from blood and receptor-negative organs and tissues (for tables of biodistribution and blocking experiments see SI, Figure S8, Tables S7-S12). The uptakes in a CCK2R-positive Biodistribution experiments were conducted at 1, 4, and 24 h p.i. in mice with CCK2Rpositive xenografts. All 177 Lu-labeled peptide conjugates showed rapid clearance from blood and receptor-negative organs and tissues (for tables of biodistribution and blocking experiments see SI, Figure S8, Tables S7-S12). The uptakes in a CCK2R-positive stomach and tumor were significantly reduced by a co-injection of excess amounts of minigastrin confirming receptor-specific uptake (Supplementary Materials Table S13). Differences in tumor uptakes (% i.A./g) between 177 Lu-labeled NMGs and PP-F11N were significant at 4 h for NMG 2 (p < 0.001) and at all time points for NMG 3 (p ≤ 0.002, Figure 4). Uptake in the stomach was higher for NMGs than for PP-F11N, whereas accumulation of radioactivity in the kidneys was slightly-but not with statistical significance-reduced for NMGs in comparison to PP-F11N. The tumor-to-kidney ratios were increased for the triazole-derivatives NMG 2 and 3, while the tumor-to-stomach ratio was higher for PP-F11N, particularly at 1 h p.i.   Figure S8).
SPECT/CT imaging was performed with 177 Lu-labeled PP-F11N, NMGs 2 and 3 at 1, 4, and 24 h p.i., and it confirmed the results from the biodistribution studies. Representative SPECT/CT images (normalized maximal intensity projection, MIP) are displayed in Figure 5. All radiotracers clearly visualized the tumor xenografts already at 1 h p.i. 177 Lulabeled NMGs 2 and 3 showed a higher accumulation of radioactivity in the xenografts at all time points compared to PP-F11N. As a result, the xenografts were still clearly visualized at 24 h p.i. for NMGs 2 and 3. Due to the renal elimination of the radiotracers, an accumulation of radioactivity was observed in the kidneys and the urinary bladder at 1 and 4 h p.i. and was almost completely eliminated at 24 h p.i. NMG 2 showed low amounts of radioactivity in the gastrointestinal region (CCK2R-positive stomach, intestines) at 1 and 4 h p.i., which was cleared within 24 h.  Figure S8). SPECT/CT imaging was performed with 177 Lu-labeled PP-F11N, NMGs 2 and 3 at 1, 4, and 24 h p.i., and it confirmed the results from the biodistribution studies. Representative SPECT/CT images (normalized maximal intensity projection, MIP) are displayed in Figure 5. All radiotracers clearly visualized the tumor xenografts already at 1 h p.i. 177 Lulabeled NMGs 2 and 3 showed a higher accumulation of radioactivity in the xenografts at all time points compared to PP-F11N. As a result, the xenografts were still clearly visualized at 24 h p.i. for NMGs 2 and 3. Due to the renal elimination of the radiotracers, an accumulation of radioactivity was observed in the kidneys and the urinary bladder at 1 and 4 h p.i. and was almost completely eliminated at 24 h p.i. NMG 2 showed low amounts of radioactivity in the gastrointestinal region (CCK2R-positive stomach, intestines) at 1 and 4 h p.i., which was cleared within 24 h.

Discussion
Targeting of the CCK2R with radiolabeled peptide conjugates is a promising strategy for the diagnosis and endoradiotherapy of tumors overexpressing this receptor. Recent advances in CCK2R-targeting radiopeptides have provided promising results for future use as diagnostic and therapeutic alternatives in the management of patients with MTC and SCLC [8,9,14,25,26]. A general requirement for applications of CCK2R-targeting peptides in nuclear medicine is a high tumor uptake with minimal distribution to healthy organs and tissues. We have recently reported that substitutions of single or multiple amide bonds with metabolically stable 1,4 or 1,5-disubstituted 1,2,3-triazoles improved the tumor-targeting characteristics of the truncated minigastrin analog [Nle 15 ]MG11 [16,17,27]. The triazole-containing analogs of [Nle 15 ]MG11 exhibited higher affinities toward the CCK2R and increased stability against enzymatic degradation in vitro. In the present work, these structural modifications were applied to PP-F11N, a 177 Lu-labeled CCK2R-targeting peptide currently under evaluation in phase I clinical trials [8].
Evaluation of the novel NMGs in vitro revealed interesting properties of the radiolabeled peptidomimetics for targeting of the CCK2R (Table 1, Figure 2). NMGs 2 and 3 showed accelerated rates of receptor-specific cell binding and enhanced cell internalization, reaching almost 80% of cell-associated radioactivity after 4 h of incubation (SI, Tables S2-S5). NMG 1 exhibited a reduced cellular uptake after 4 h of incubation as compared to the reference compound PP-F11N. The results were in agreement with the receptor affinities (IC 50 ) of the nat Lu-labeled peptidomimetics as determined by competition binding assays. In comparison to PP-F11N (10.1 nM), the IC 50 values were considerably lower for NMGs 2 and 3 (4.2 nM and 2.0 nM, respectively) and higher for NMG 1 (22.8 nM). In previous studies, we have shown that the substitution of the amide bond at position 6 (Tyr 12 -Ψ[Tz]-Gly 13 ) of [Nle 15 ]MG11 can result in an additional cation-π interaction between the peptidomimetics and the CCK2R, leading to the increased affinity [16]. The observations with NMGs 2 and 3 confirm these results.
Stability studies in vitro using human blood plasma did not show any metabolites for radiolabeled PP-F11N and NMGs after 24 h of incubation (Figure 3), likely due to the stabilizing effect of the N-terminal hexaglutamate moiety in vitro [28]. This is in contrast to previously evaluated triazole-bearing analogs of minigastrin lacking the N-terminal extension [16,17,27]. To evaluate whether the amide-to-triazole substitutions stabilize the analogs of PP-F11N in vivo, we investigated samples of blood and urine of mice at 10 min p.i. of the radiotracers. In contrast to the studies in vitro, metabolites of the 177 Lu-labeled NMGs were observed in vivo ( Figure 3). Membrane-bound proteases (e.g., neutral endopeptidase, angiotensin-converting enzyme) are absent in blood plasma used for in vitro stability studies but are involved in the degradation of various radiolabeled peptides in vivo [15,29,30]. This should be taken into consideration for assessing the metabolic stability of peptidic radiotracers. In samples of blood, 50-80% of the compounds were found metabolized within 10 min; NMG 1 showed the highest stability, PP-F11N and NMG 3 were found to exhibit a similar stability, and NMG 2 had the lowest stability. Negligible quantities of intact radio-conjugates were found in urine samples. These results confirm our previous findings that not all amide-to-triazole substitutions necessarily lead to an increased resistance of the peptidomimetics against metabolic degradation, [16,17] however, can result in the improvement of other biologically relevant characteristics. 177 Lu-labeled NMGs 2 and 3, as well as reference compound PP-F11N, were evaluated in vivo by biodistribution experiments in mice bearing CCK2R-positive tumor xenografts. NMG 1 was not included in the studies due to its inferior receptor affinity and cell internalization in vitro. Previous studies have shown that enhanced stability with reduced receptor affinity leads to minor or no improvement in the biodistribution of the peptidomimetics [16,17]. Therefore, unless the conjugates are functionalized with pharmacokinetic modifiers (preventing them from rapid renal elimination [31]), a favorable receptor interaction outweighs an increase in stability. Already, at 1 h p.i., the 177 Lu-labeled PP-F11N, and NMG 2 and 3 were cleared from the blood (<0.5% i.A./g), and only low levels of radioactivity were found in CCK2R-negative tissues or organs, resulting in an excellent tumor-to-background ratio. Differences between the NMGs and PP-F11N were observed for the receptor-positive stomach and tumors, reflecting the difference in receptor affinities (IC 50 , Figure 4, Table 1). NMGs 2 and 3 showed an enhanced uptake in the stomach as opposed to PP-F11N. To our delight, the uptake in tumor xenografts was increased accordingly. For example, NMGs 2 and 3 had tumor uptakes that were 1.7 and 1.9 times higher than those of PP-F11N at 24 h p.i. (4.7 ± 1.5, 5.2 ± 2.0 and 2.7 ± 0.8 % i.A./g, respectively). This illustrates the improvement of the CCK2R-targeting properties of NMGs, especially for NMG 3. Furthermore, the tumor washout for NMGs 2 and 3 was decelerated. For NMG 3, 56% of the radioactivity accumulated in the tumor at 1 h p.i. was still found in the xenografts after 24 h, whereas only 39% of the radioactivity was retained for PP-F11N (Supplementary Materials Table S11). In addition, the uptake of radioactivity in the kidneys due to the renal excretion of the radiotracers was lower for the NMGs than for PP-F11N. As a result, tumor-to-kidney ratios were higher for NMGs than for PP-F11N at all time points (SI, Tables S6-S11), which is important for applications of the radio-conjugates in PRRT. Another important aspect for therapeutic applications is the uptake of radiolabeled minigastrins in the stomach, which was reported as the dose-limiting organ in clinical trials with PP-F11N [10]. In this regard, the clinical candidate PP-F11N was superior over NMGs 2 and 3. PP-F11N showed the highest tumor-to-stomach ratios at all measured time points, particularly at 1 h p.i. (Supplementary Materials Table S9-S11). Co-injection of excess minigastrin reduced the uptake in receptor-positive stomach and tumors by more than 50% (SI, Tables S6-S11), hence demonstrating CCK2R-specific uptake.
The potential of 177 Lu-labeled NMGs 2 and 3 for targeting CCK2R-positive tumors in vivo was further demonstrated by SPECT/CT imaging studies ( Figure 5). These studies confirmed the results of the biodistribution experiments. The tumors were clearly visualized by all radiotracers already at 1 h p.i., as were the urinary bladder and the kidneys. The high kidney uptake visible in Figure 5 for NMG 3 at 1 h p.i. was not supported by the results of previous biodistribution studies and may be attributed to a lower water consumption by the animal. The radioactivity in the kidneys and the urinary bladder dropped substantially at 4 h p.i. for all conjugates and was completely cleared at 24 h p.i. NMG 2 also showed low accumulation in the CCK2R-positive stomach and in the intestines at 1 h p.i. only, which was eliminated within 24 h. Both NMGs showed higher uptake of radioactivity in the tumor xenografts at all time points in comparison to the reference compound PP-F11N. The slower tumor washout of NMGs led to higher residual radioactivity in the tumor xenografts at 24 h p.i., providing the possibility for SPECT imaging also at later time points. The good image quality of the SPECT/CT scans with low accumulation in non-targeted tissues could also allow the application of the 177 Lu-labeled NMGs 2 and 3 for imaging of CCK2R-positive neoplasms via SPECT, although other radionuclides are better suited for imaging applications because of their physical properties (e.g., 68 Ga, 111 In). The enhanced uptake and prolonged retention of radioactivity in the tumor, the higher tumor-to-kidney ratios, and the efficient clearance from the kidneys indicate that NMGs 2 and 3 could be suited for both imaging and PRRT. However, the increased uptake of the novel triazole analogs of PP-F11N in the stomach as dose-limiting organ must be considered and requires further investigation.
Several promising CCK2R-targeting radiopeptide conjugates have been reported in the recent years, some of which are currently being investigated in the clinic (PP-F11N, CP04, MGS5) [8,9,25]. In the present work, we evaluated promising NMGs under standardized experimental conditions. To accelerate the clinical translation of minigastrin-based radiopharmaceuticals, a direct, preclinical comparison of the most promising candidates would be desirable. Such comparative studies would streamline and accelerate the efforts of researchers involved in the development of radiolabeled CCK2R-specific probes for routine applications in the diagnosis and endoradiotherapy of CCK2R-positive neoplasms.

Conclusions
We have successfully applied the amide-to-triazole substitution strategy to the clinically tested minigastrin analog PP-F11N in an effort to improve its performance in molecular imaging and PRRT. The novel peptidomimetic analogs of PP-F11N (NMGs) described in this work exhibit improved affinities toward the CCK2R and increased cell internalization in vitro as well as superior tumor-targeting properties in vivo. The structural modifications of PP-F11N led to an increased uptake of the novel radiotracers in CCK2Rpositive xenografts in mice, a slower washout of radioactivity from the tumors, and an improved tumor-to-kidney ratio; however, it also resulted in a reduced tumor-to-stomach ratio that needs to be taken into consideration. In particular, the newly identified lead compound NMG 3 holds promising potential for clinical translation and application in nuclear medicine for the diagnosis and therapy of CCK2R-positive tumors.  Table S1: Sequences of the peptide conjugates involved in the study, synthesis yields and data from high-resolution mass spectrometry (HRMS), Table S2: Complete Internalization Data for 177 Lu-labeled PP-F11N as mean ± standard deviation from n = 3-4 in triplicates, Table S3: Complete Internalization Data for 177 Lu-labeled NMG 1 as mean ± standard deviation from n = 3-4 in triplicates, Table S4: Complete Internalization Data for 177 Lu-labeled NMG 2 as mean ± standard deviation from n = 3-4 in triplicates, Table S5: Complete Internalization Data for 177 Lu-labeled NMG 3 as mean ± standard deviation from n = 3-4 in triplicates,  Table S10: Selected tumor-to-nontumor ratios for 177 Lu-labeled PPF11N at 1, 4, and 24 h. (n = 4 animals per group), Table S11: Tumor washout of 177 Lu-labeled PP-F11N, NMGs 2, and 3, Table S13: Two-way analysis of variance (ANOVA) of reduction of uptake of radioactivity in receptor-positive tumor and stomach by co-injection of excess amount of minigastrin. Significance and p-values according to Bonferroni multiple comparisons test performed by GraphPad Prism 8.1.2.