Imaging of the Glucose-Dependent Insulinotropic Polypeptide Receptor Using a Novel Radiolabeled Peptide Rationally Designed Based on Endogenous GIP and Synthetic Exendin-4 Sequences

Imaging and radiotherapy targeting the glucose-dependent insulinotropic polypeptide receptor (GIPR) could potentially benefit the management of neuroendocrine neoplasms (NENs), complementing clinically established radiopharmaceuticals. The aim of this study was to evaluate a GIPR-targeting positron emission tomography (PET) radioligand with receptor-specific binding, fast blood clearance, and low liver background uptake. The peptide DOTA-bioconjugate, C803-GIP, was developed based on the sequence of the endogenous GIP(1-30) and synthetic exendin-4 peptides with selective amino acid mutations to combine their specificity for the GIPR and in vivo stability, respectively. The 68Ga-labeled bioconjugate was evaluated in vitro in terms of binding affinity, specificity, and internalization in HEK293 cells transfected with the human GIPR, GLP1, or GCG receptors and in sections of human insulinoma and NENs. In vivo binding specificity, biodistribution, and tissue background were investigated in mice bearing huGIPR-HEK293 xenografts and in a pig. Ex vivo organ distribution, pharmacokinetics, and dosimetry were studied in normal rats. [68Ga]Ga-C803-GIP was stable and demonstrated a high affinity to the huGIPR-HEK293 cells. Binding specificity was demonstrated in vitro in frozen sections of NENs and huGIPR-HEK293 cells. No specific uptake was observed in the negative controls of huGLP1R and huGCGR cells. A novel rationally designed PET radioligand, [68Ga]Ga-C803-GIP, demonstrated promising binding characteristics and specificity towards the GIPR.


Introduction
The glucose-dependent insulinotropic polypeptide receptor (GIPR) is an emerging therapeutic and diagnostic target. Its endogenous ligand is the glucose-dependent insulinotropic polypeptide (GIP), which is a 42-mer peptide released from the intestinal K-cells in response to food intake. The GIP is an incretin hormone and augments insulin secretion in response to hyperglycemia by activating the GIPR on the pancreatic beta cells. The activity of the GIPR has also been shown to affect lipid metabolism as well as bone turnover. The GIPR is expressed physiologically in low density throughout the body [1].
Overexpression of the GIPR has been demonstrated in pathologies, such as a broad spectrum of human gastrointestinal, pancreatic, and bronchial neuroendocrine neoplasms

Radiochemistry
The 68 Ga-labeling procedure was optimized in terms of the radiochemical yield and required optimization of such factors as temperature, heating time, radical scavenger content, buffer concentration, pH, and product purification mean. The labeling was reproducible with a high non-decay corrected radiochemical yield of 61.1 ± 2.9% and a molar activity of 59. 8 ± 4.4 MBq/nmol at the end of the synthesis (Figure 2). The radioactivity at the end of the synthesis was 424 ± 29 MBq with a radiochemical purity of 94.5 ± 0.3% (for the chromatogram, see Supplementary Materials, Figure S1). The respectively high molar activity allowed for the accurate determination of binding affinity and in vivo localization of the target GIPR. Although the radiolysis could be suppressed by the addition of only ethanol (200 µL) to the reaction mixture, the post-production stabilization at ambient temperature required the addition of ascorbic acid with a radiochemical purity of over 91% for at least 2 h. The accurate determination of the peptide concentration is critical for the estimation of binding affinity and was accomplished by UV-HPLC with respective response calibration.  3, 6, 9.6, 19.8, 30, 60, 120, and 240 min post subcutaneous administration in rats (n = 3). The error bars are not seen due to the low SD values. The plasma concentration of Ga-C803-GIP was below the limit of detection at the 240 min time point and was assigned zero value.

Radiochemistry
The 68 Ga-labeling procedure was optimized in terms of the radiochemical yield and required optimization of such factors as temperature, heating time, radical scavenger content, buffer concentration, pH, and product purification mean. The labeling was reproducible with a high non-decay corrected radiochemical yield of 61.1 ± 2.9% and a molar activity of 59.8 ± 4.4 MBq/nmol at the end of the synthesis (Figure 2). The radioactivity at the end of the synthesis was 424 ± 29 MBq with a radiochemical purity of 94.5 ± 0.3% (for the chromatogram, see Supplementary Materials, Figure S1). The respectively high molar activity allowed for the accurate determination of binding affinity and in vivo localization of the target GIPR. Although the radiolysis could be suppressed by the addition of only ethanol (200 µL) to the reaction mixture, the post-production stabilization at ambient temperature required the addition of ascorbic acid with a radiochemical purity of over 91% for at least 2 h. The accurate determination of the peptide concentration is critical for the estimation of binding affinity and was accomplished by UV-HPLC with respective response calibration.

Cellular Internalization and Binding Affinity of [ 68 Ga]Ga-C803-GIP
HEK293 cells transfected with human GIPR were subjected to incubation with [ 68 Ga]Ga-C803-GIP for 0, 30, 60, 90, and 120 min. The radioactivity associated with the cells (%ID/million cells) is presented as a function of time in Figure 3. The tracer bound to the cells reached approximately 75% ID/million cells after 60-120 min. The internalized fraction was approximately 40% of the total binding and could be suppressed almost completely by incubation at low temperature (4 • C). The binding of [ 68 Ga]Ga-C803-GIP to huGIPR could be inhibited by 0.3 µM C803-GIP. The binding saturation assay resulted in K d 18.3 ± 6.5 nM and Bmax 2.4 ± 1.0 pmol/Mcells (n = 3) ( Figure 3B).

Cellular Internalization and Binding Affinity of [ 68 Ga]Ga-C803-GIP
HEK293 cells transfected with human GIPR were subjected to incubatio [ 68 Ga]Ga-C803-GIP for 0, 30, 60, 90, and 120 min. The radioactivity associated w cells (%ID/million cells) is presented as a function of time in Figure 3. The tracer bo the cells reached approximately 75% ID/million cells after 60-120 min. The inter fraction was approximately 40% of the total binding and could be suppressed almo pletely by incubation at low temperature (4 °C). The binding of [ 68 Ga]Ga-C803 huGIPR could be inhibited by 0.3 µM C803-GIP. The binding saturation assay resu Kd 18.3 ± 6.5 nM and Bmax 2.4 ± 1.0 pmol/Mcells (n = 3) ( Figure 3B). The total binding (blue), non-specific (grey), and specific binding (red, calculated by the subtraction of non-specific binding from binding). Data from a representative experiment are shown.

In Vitro Binding of [ 68 Ga]Ga-C803-GIP to Frozen Sections and Stability in Human P
The binding of [ 68 Ga]Ga-C803-GIP to sections of a frozen pellet of cells tran with the GIPR was strong and could be abolished by co-incubation with 10 µM GI The total binding (blue), non-specific binding (grey), and specific binding (red, calculated by the subtraction of non-specific binding from the total binding). Data from a representative experiment are shown.

In Vitro Binding of [ 68 Ga]Ga-C803-GIP to Frozen Sections and Stability in Human Plasma
The binding of [ 68 Ga]Ga-C803-GIP to sections of a frozen pellet of cells transfected with the GIPR was strong and could be abolished by co-incubation with 10 µM GIP(1-42) and 10 µM C803-GIP but not by GLP1(7-36) ( Figure 4). No uptake of [ 68 Ga]Ga-C803-GIP and no blocking effect could be detected in the pellets of cells transfected with the hGLP1R and hGCGR. One set of frozen sections of biopsies of human INS and ileum NETs was subjected to the incubation with [ 68 Ga]Ga-C803-GIP for the elucidation of total binding. Sections of the hGIPR-HEK293 cell pellet were included as positive controls and hGLP1R-HEK293 cell pellet sections as negative controls. The other set of the sections was pre-incubated with GIP  prior to the addition of [ 68 Ga]Ga-C803-GIP in order to elucidate the binding specificity. A slight blocking effect could be achieved in the frozen sections of human INS but not ileum NETs.
[ 68 Ga]Ga-GIP-C803 was stable in human plasma for at least 90 min (See Supplementary Materials, Figure S2).
Sections of the hGIPR-HEK293 cell pellet were included as positive controls and hGLP1 HEK293 cell pellet sections as negative controls. The other set of the sections was pr incubated with GIP  prior to the addition of [ 68 Ga]Ga-C803-GIP in order to elucida the binding specificity. A slight blocking effect could be achieved in the frozen sections human INS but not ileum NETs. [ 68 Ga]Ga-GIP-C803 was stable in human plasma for at least 90 min (See Suppleme tary Materials, Figure S2).

In Vivo Organ Distribution of [ 68 Ga]Ga-C803-GIP in Rat
The organ distribution kinetics of [ 68 Ga]Ga-C803-GIP was studied in male Spragu Dawley rats (n = 16). Two animals were sacrificed at each of the eight time points as d picted in Figure 5A. Most tissues exhibited a rapid clearance of radioactive uptake exce for the kidney, which is an excretory organ. The liver-to-blood ratio increased with tim indicating faster blood clearance compared to the washout from the organ ( Figure 5B).
Four rats received a co-injection of 1 mg/kg of C803-GIP precursor in the competitio experiment. The in vivo competition showed no significant displacement in the liver pancreas or in any other organ in the presence of an excess of non-labeled precursor C80 GIP at 20 min post injection ( Figure 5C).

In Vivo Organ Distribution of [ 68 Ga]Ga-C803-GIP in Rat
The organ distribution kinetics of [ 68 Ga]Ga-C803-GIP was studied in male Sprague Dawley rats (n = 16). Two animals were sacrificed at each of the eight time points as depicted in Figure 5A. Most tissues exhibited a rapid clearance of radioactive uptake except for the kidney, which is an excretory organ. The liver-to-blood ratio increased with time indicating faster blood clearance compared to the washout from the organ ( Figure 5B).
Four rats received a co-injection of 1 mg/kg of C803-GIP precursor in the competition experiment. The in vivo competition showed no significant displacement in the liver or pancreas or in any other organ in the presence of an excess of non-labeled precursor C803-GIP at 20 min post injection ( Figure 5C).

Time, [min]
Blood Liver Liver-to-blood ratio

Organ Distribution of [ 68 Ga]Ga-C803-GIP in GIPR-Expressing Xenograft Mice
The binding to the GIPR-expressing xenografts was low ( Figure 6A,C). The expression of the GIPR in the inoculated cell line was unexpectedly decreased after in vivo transplantation but still present as determined by an ex vivo autoradiography assay (in the range of 3 fmol/mm 3 ). The tumor-to-tissue ratio was in the range of 1.5 for blood and liver ( Figure 6B). Furthermore, there was only a limited blocking effect (on average, approximately a 30% decrease) seen by co-injection of 1 mg/kg C803-GIP ( Figure 6A,D).

Organ Distribution of [ 68 Ga]Ga-C803-GIP in GIPR-Expressing Xenograft Mice
The binding to the GIPR-expressing xenografts was low ( Figure 6A,C). The expression of the GIPR in the inoculated cell line was unexpectedly decreased after in vivo transplantation but still present as determined by an ex vivo autoradiography assay (in the range of 3 fmol/mm 3 ). The tumor-to-tissue ratio was in the range of 1.5 for blood and liver ( Figure 6B). Furthermore, there was only a limited blocking effect (on average, approximately a 30% decrease) seen by co-injection of 1 mg/kg C803-GIP ( Figure 6A,D).

In Vivo Organ Distribution of [ 68 Ga]Ga-C803-GIP in Piglets
[ 68 Ga]Ga-C803-GIP demonstrated a rapid clearance from the circulation and most tissues after administration as assessed by dynamic PET/CT ( Figure 7). Uptake in the liver was SUV <1 after 30 min and was approximately SUV ≈ 0.5 after 90 min, indicating low non-specific accumulation due to metabolism. Excretion occurred almost exclusively via the urine as demonstrated by the strong and irreversible retention in the kidney cortex. Binding in the pancreas was low and similar in magnitude to that in blood and the spleen.

In Vivo Organ Distribution of [ 68 Ga]Ga-C803-GIP in Piglets
[ 68 Ga]Ga-C803-GIP demonstrated a rapid clearance from the circulation and most tissues after administration as assessed by dynamic PET/CT ( Figure 7). Uptake in the liver was SUV <1 after 30 min and was approximately SUV ≈ 0.5 after 90 min, indicating low non-specific accumulation due to metabolism. Excretion occurred almost exclusively via the urine as demonstrated by the strong and irreversible retention in the kidney cortex. Binding in the pancreas was low and similar in magnitude to that in blood and the spleen.

In Vivo Organ Distribution of [ 68 Ga]Ga-C803-GIP in Piglets
[ 68 Ga]Ga-C803-GIP demonstrated a rapid clearance from the circulation and most tissues after administration as assessed by dynamic PET/CT (Figure 7). Uptake in the liver was SUV <1 after 30 min and was approximately SUV ≈ 0.5 after 90 min, indicating low non-specific accumulation due to metabolism. Excretion occurred almost exclusively via the urine as demonstrated by the strong and irreversible retention in the kidney cortex. Binding in the pancreas was low and similar in magnitude to that in blood and the spleen.

Human Predicted Dosimetry
The rat organ distribution of [ 68 Ga]Ga-C803-GIP data were extrapolated to human tissues and used for the calculation of the predicted absorbed radiation dose in humans ( Table 3). The doses were highest in the kidneys for both males (0.713 mGy/MBq) and females (0.773 mGy/MBq). The whole-body effective dose was 0.0217 mSv/MBq for males and 0.0237 mSv/MBq for females.

Discussion
Most NENs express the GIPR, GLP-1R, or/and somatostatin receptor (SSTR) [13]. Targeting the SSTR for SPECT and PET imaging as well as radiotherapy is an established method for the management of NENs [14][15][16]. SPECT and PET imaging in combination with exendin-based radiopharmaceuticals have also been proven clinically relevant [17][18][19]. The expression of the GIPR, and thus the potential for targeting imaging, has been demonstrated for pancreatic and gastrointestinal NENs [20,21], and the development of radioligands for specific receptor binding has been initiated [6,22]. The endogenous GIP(1-42) was truncated to GIP(1-30) and functionalized with a DOTA chelator for the subsequent labeling with In-111 and Ga-68 (EG2, EG4) [6], and the mouse xenografts (INR1G9-huGIPr) were visualized, though the affinity of the radioligand was reduced compared to GIP . In an attempt to maintain the binding affinity, the endogenous GIP(1-42) was functionalized at Lys37 with a DTPA chelator ([Lys37(111In-DTPA)]N-acetyl-GIP1-42), and the resulting radioligand visualized the mouse xenografts (BHK-GIPR) [22]. Another GIP analog, consisting of the truncated part of GIP(1-31) with a change of 13 amino acid residues at various positions essential for binding and stability, was labeled with Gallium-68 ([ 68 Ga]Ga-S02-GIP-T4) and demonstrated sub-nanomolar affinity for the GIPR [7]. Importantly, cross-reactivity for the GLP-1R was negligible, indicating high selectivity of the radioligand for the GIPR. The sensitivity of the imaging was high allowing for the first-time visualization of the GIPR density in the physiological pancreas, most probably due to a magnitude higher affinity compared to previously published analogs. However, while [ 68 Ga]Ga-S02-GIP-T4 enables the sensitive detection of the pancreatic GIPR, it demonstrated a liver background signal of approximately SUV 8 in non-human primates after 90 min, potentially making detection of the GIPR in this tissue challenging, e.g., in the form of NEN metastasis. Given the knowledge of the liver being one of the most common sites for metastases [8], there is still room for novel GIPR-targeting radioligands with a specific focus on the detection of primary and metastatic NENs and low liver background uptake.
The C803-GIP reported in this work is based on the GIP(1-30) and exendin-4 sequences. The aim of the design was a peptide of improved water solubility, in vivo stability, and binding-site recognition. Eight of the amino acid residues of GIP(1-30) at positions 2, 14, 21, 23, 24, 27, 29, and 30 were substituted with other AAs unique to C803-GIP. The exchange of L-Ala(2) to D-Ala(2) was aimed to improve the stability against enzymatic processing. In order to exclude the oxidative radiolysis, Met (14) was substituted with Leu (14). Substitution of aspartic acid with glutamic acid is known to increase the unfolding transition temperature of a protein [23]; thus, the introduction of Glu(21) instead of Asp (21) was considered. Substitution of Val (23) to Ile(23) would introduce more bulkiness adjacent to the protein backbone and restrict the conformation to β-sheets. The exchanges of Asn (24) to Glu (24) and Leu (27) to Lys (27) were to improve solubility. Both Gln (29) and Lys (30) were substituted with Gly residues to support the tryptophan cage conformation. The C-terminal tail (31-39), originating from exendin-4, provided stability and solubility [24], while no interaction with the receptor was proven by the absence of cross-activation in the affinity studies (Table 2). Pro(31) was exchanged to His(31) in order to reduce the constraint of the structure and add a positive charge for better binding functionality.
In comparison to the previously described S02-GIP-T4 analog, the potency of C803-GIP was seemingly improved with regard to the human and NHP GIPR ( Table 2). This effect may be due to the exchange of Aib (2) [7]. The affinity of [ 68 Ga]Ga-C803-GIP was comparable in magnitude to that of EG2 and EG4 (Kd values of 10.6 ± 1.6 nM for 111/natIn-EG2 and 8.5 ± 1.0 nM for 111/natIn-EG4) [6] and higher than that of [Lys37(111In-DTPA)]N-acetyl-GIP1-42 (IC 50 = 4.8 µM) [22]. The notion according to the binding potential (B max /K d ) on the linear improvement of imaging of a certain receptor density (B max ) with the improvement of radioligand affinity (K d ) is reflected in these results.
The binding specificity of [ 68 Ga]Ga-C803-GIP was investigated in vitro in cell and frozen tissue sections as well as in vivo in mice bearing xenografts. Several sets of frozen sections of HEK293 cell pellets were used for the investigation of the binding specificity: one transfected with huGIPR and the other with huGCGR or huGLP1R. The experiment outline also allowed the investigation of the cross-activity. In both cases, the blocking of the total uptake of [ 68 Ga]Ga-C803-GIP was conducted using an excess of GIP(1-42), C803-GIP, or GLP(7-36). The blocking effect of native GIP(1-42) and non-labeled precursor C803-GIP was strong, respectively by 99 and 92%. There was no baseline accumulation or blocking effect observed in the huGLP1R-HEK293 or huGCGR-HEK293 cell pellet sections, demonstrating the specificity of [ 68 Ga]Ga-C803-GIP binding towards the GIPR and absence of cross-reactivity towards the GLP1R and GCGR. A similar tendency was observed in the frozen sections of the human biopsies of the INS and ileum of pancreatic NET, wherein the inhibition of the [ 68 Ga]Ga-C803-GIP binding by GIP(1-42) was detected. Thus, in summary, the in vitro data demonstrated strong GIPR-mediated binding to cell lines and relevant tumor biopsies supporting further in vivo evaluation.
[ 68 Ga]Ga-C803-GIP uptake, background binding, and washout were investigated in healthy rats to ensure low physiological uptake in the background tissues necessary for providing sufficiently high-contrast images with lesion-elevated uptake. The distribution indicated a fast clearance from all tissues except for the kidneys as the organ of excretion. The organ most commonly susceptible to metastases is the liver [8], and thus a low physiological uptake is essential for achieving an image contrast. The liver uptake was below SUV <1 already after 5 min post-administration, and the washout with respect to the 5 min time point was 50% within 1 h and 75% within 3 h, promising favorable imaging conditions to detect GIPR-expressing hepatic lesions. [ 68 Ga]Ga-C803-GIP also displayed a rapid clearance from blood. Essentially, the only organ displaying a high SUV increasing with time up to 85.5 ± 1.6 within 3 h was the kidney indicating renal excretion. Thus, no uptake significantly higher than the background signal could be detected for the rest of the vital organs. Moreover, no statistically significant difference could be found in the organ distribution between the baseline, wherein only the radioligand was administered, and the competition study, wherein an excess of C803-GIP was co-administered with the radioligand. The low physiological uptake and absence of the blocking effect indicate a non-existent expression of the GIPR and thus strong potential for high contrast of the images with NENs. In particular, the ratio of SUVs for liver-to-blood increased with the time, which is important given that the liver is a usual organ for metastases. However, the liver-to-blood ratio is driven primarily by the rapid clearance of [ 68 Ga]Ga-C803-GIP from blood, and there may be an increased uncertainty for this value as the liver uptake is divided by a very low blood SUV value. Although the affinity of [ 68 Ga]Ga-C803-GIP was comparable or somewhat lower to the GIPR targeting analogs previously reported [6,7,22], the liver uptake was lower for [ 68 Ga]Ga-C803-GIP presenting a crucial advantage for detection of hepatic lesions.
The in vivo binding specificity of [ 68 Ga]Ga-C803-GIP was lastly studied in vivo in mice bearing xenografts of GIPRhu HEK293 cells, wherein the uptake in the xenografts was partly precluded by the excess of C803-GIP as compared to the baseline uptake. The uptake in the muscle, liver, and lung was low and non-specific potentially allowing images of high contrast in these tissues, which are potential locations for NEN metastasis. However, the tumor uptake in absolute terms was very low (SUV < 1), which is hardly suitable for sensitive in vivo detection of GIPR-expressing neoplasms. Furthermore, the tumor-to-tissue ratio was just above 1.
A similar biodistribution pattern was seen by PET/CT scanning in pig. A very low background signal was seen in most tissues. The liver uptake was SUV ≈ 0.5 after 90 min, which was approximately 16-times lower than for a previously reported GIPR high-affinity PET tracer in a large animal model [7]. However, the binding of [ 68 Ga]Ga-C803-GIP to tissues with a known distribution of the GIPR such as the pancreas was low or negligible. Furthermore, the blocking of the GIPR by an infusion of GIP(1-42) did not affect the uptake pattern in pig.
The radiotoxicity was investigated in order to determine the possibility of repeated examinations for follow-up and longitudinal studies. The organ distribution of [ 68 Ga]Ga-C803-GIP in healthy rats was thus analyzed for the extrapolation to human radiation dosimetry. The calculation of the organ-equivalent doses and effective dose was based on the assumptions that the biodistribution pattern is similar in a human and rat, and that radioactivity is distributed homogeneously throughout a given organ. The longest incubation time of 180 min corresponded to over three effective half-lives covering both the physical decay and biological clearance, and thus providing high accuracy for dosimetry calculations. The kidney was found to be the critical organ tissue in the human predicted dosimetry as it was for [ 68 Ga]Ga-S02-GIP-T4 calculated based on rat biodistribution data; however, the absorbed dose was 1.5-times higher for the former. It was somewhat higher also for adrenals and the spleen (app. 1.2 times), while the absorbed dose of [ 68 Ga]Ga-S02-GIP-T4 was considerably higher in the small intestine (2.2 times), heart wall (2.4 times), liver (1.5 times), lung (1.3 times), red marrow (2 times), osteogenic cells (1.9 times), ovaries (1.5 times), thymus (2 times), thyroid (1.7 times), and urinary bladder wall (2.5 times) as compared to the [ 68 Ga]Ga-C-803-GIP absorbed dose to those organs. Although the absorbed dose was higher for [ 68 Ga]Ga-S02-GIP-T4 in most of the organs, the total effective dose was still somewhat lower. Importantly, both [ 68 Ga]Ga-C-803-GIP and [ 68 Ga]Ga-S02-GIP-T4 would potentially allow for several PET scans per year with regard to the total effective dose. Renal dosing should ideally be decreased to enable additional PET scanning. Additionally, renal dosimetry is crucial for future potential development of radiotherapeutic applications of the current compound. Kidney uptake may potentially be decreased by the pre-administration of amino acids or plasma volume expanders, similar to how these strategies are employed for established Peptide Receptor Radiotherapies. We also hypothesize that renal uptake can be modulated by altering the rapid clearance of [ 68 Ga]Ga-C803-GIP by adding a fatty acid moiety or albumin-binding domain. A long circulatory half-life may simultaneously increase the binding capacity and tissue exposure of [ 68 Ga]Ga-C803-GIP.
Thus, [ 68 Ga]Ga-C803-GIP demonstrated strong and GIPR-specific binding in vitro in cells and tissues. Of importance, we demonstrated GIPR-mediated binding in sections from human insulinoma and ileum NENs. The in vivo tissue background and excretion patterns were favorable for potential tumor and liver metastases localization, but binding to the GIPR-dense tissues was low. The reason for poor translation of the promising in vitro data to the in vivo setting may have several reasons, but the primary culprits are likely the limited affinity (15-20 nM) towards the GIPR, perhaps in combination with low tissue exposure due to rapid clearance. However, the advantage of [ 68 Ga]Ga-C-803-GIP was the in vivo low liver uptake, potentially allowing high-contrast imaging of liver metastatic lesions. Thus, further development of [ 68 Ga]Ga-C-803-GIP as a lead compound is warranted with the major aims of developing a radiotracer with improved binding to GIPRrich tissues in combination with a reduced rate of clearance and improved biodistribution while preserving low liver uptake.

Generation of Sequence and DOTA Conjugation
The peptide conjugate C803-GIP was developed based on the hGIP(1-30) and exendin-4 sequences (Table 1). Selective amino acid mutations were applied to attain selectivity for the GIPR and improved stability. As seen in Table 1, C803-GIP has 10 residues shared with both GIP(1-30) and exendin-4. Twelve residues, and especially residues 12-20, are conserved from GIP  to retain the strong affinity and potency towards the GIPR. Finally, 10 residues towards the C-terminus are conserved from exendin-4 to take advantage of the established high in vivo stability of this peptide. The alanine was changed to d-alanine in order to minimize enzymatic degradation at the N-terminal. The peptide conjugate comprises the chelating moiety DO3A attached via a cysteine amide on the C-terminus and a diethylsulfone linker. The peptide was synthesized in-house via standard solid-phase peptide synthesis (Sanofi). LC-ESI-TOF-MS analysis of C803-GIP was performed using an LCT Premier mass analyzer. The analysis was performed with positive-mode scanning and selected ion recording, detecting  Figure S4).

In Vitro Potency
The potency of C803-GIP loaded with nonradioactive gallium was assessed by a functional cAMP assay in the HEK293 cells transfected with the huGIPR (EvoTech). The details of the cell line and potency assay procedure have been described previously [25,26].

Radiolabeling of C803-GIP
The 68 Ge/ 68 Ga generator (pharmaceutical grade, GalliaPharm ® , Eckert & Ziegler, Berlin, Germany) eluate was fractionated, and the top fraction of 3.5 mL containing approximately 90% of the radioactivity was collected for the radiolabeling. The solution was buffered with sodium acetate buffer (1 M, 300 µL, pH 6-7) to achieve a pH of 4.6-5.0. To suppress the radiolysis, ethanol (200 µL) was added to the reaction mixture, and ascorbic acid (8-10 mg) was added to the final product. The peptide conjugate C803-GIP (20-40 nanomole) dissolved in acetate buffer (pH 6-7) was added to the buffered 68 Gaeluate solution, and the reaction mixture was heated at 75 • C for 15 min. The reaction mixture was purified using a solid-phase extraction cartridge (HLB, Oasis) to assure the elimination of possible hydrophilic radioactive impurities, germanium-68, and colloids. The product was eluted with 1 mL of 60% ethanol in phosphate buffer solution. The radiochemical purity and the peptide concentration in the product were determined by high-pressure liquid chromatography (HPLC). The HPLC system (LaChrom, Hitachi, VWR, Atlanta, GA, USA) consisted of an L-2130 pump, UV detector (L-2400), and a radiation flow detector (Bioscan, Wheaton, IL, USA) coupled in series for product quality control. Separation of the analytes was accomplished using an end-capped analytical column with stationary reversed phase (C-4; Vaydac-C4; 50 × 4.6 mm; particle size: 3 µm). The following system was used: A = 10 mM TFA; B = 30% water: 70% acetonitrile: 10 mM TFA with UV-detection at 220 nm; linear gradient elution: 2-9.9 min from 20% to 100% B followed by re-equilibration 9.9-10 min from 100% to 20% B; flow rate was 1.0 mL/min. Data acquisition and handling were performed using the EZChrom Elite Software Package. The stability of the product at room temperature was monitored for 2 h and assessed by UV-Radio-HPLC. The final product was formulated dependent on the biological assay. The total radioactivity of the product was then measured in an ionization chamber. The radioactivity recovery from the analytical column was controlled by performing an analysis of the product, generator eluate, and the product spiked with the generator eluate with and without column. The fractions were collected from the outlet of the HPLC system for the subsequent measurement of the radioactivity in a well-type NaI(Tl) scintillation counter corrected for dead-time and for radioactive decay. The recovery was calculated as a fraction of the radioactivity collected from the run with the column in relation to the radioactivity collected from the run without the column. The recovery of the product and generator eluate was, respectively, over 98% and nearly 100%, indicating high reliability of the analysis.

Cellular Internalization and Binding Affinity of [ 68 Ga]Ga-C803-GIP
The cell internalization assay was performed as described in detail previously [25]. Briefly, HEK293 cells transfected with the human GIPR (approximately 300,000 cells) were incubated with 5 nM [ 68 Ga]Ga-C803-GIP in complete media for 0, 30, 60, 90, and 120 min, either at 37 • C or at 4 • C. To measure the membrane-bound and internalized radioligand, the cells were first treated with 0.2 M glycine buffer containing 4 M urea (pH 2.5) (acid wash buffer) for 5 min, and the supernatant (containing the membrane-bound fraction) was measured in a gamma counter. Thereafter, the cells were treated with 1 M NaOH (basic wash buffer) for 30 min, and the lysed cells were measured by a gamma counter (internalized fraction). The assay was repeated three times. The total and internalized fractions of [ 68 Ga]Ga-C803-GIP at 37 • C or at 4 • C were calculated.
For the affinity assessment, the GIPR-transfected HEK293 cells (0.5-0.6 million cells per dish) were incubated with seven concentrations (0.3, 1.0, 3.0, 10, 30, 100, and 300 nM) of [ 68 Ga]Ga-C803-GIP G in 1 mL complete media for 60 min. The cells were incubated with radiotracer in the presence and absence of 10 µM of human GIP(1-42) (to block the GIPR and assess non-specific binding). The assay was performed at 4 • C to suppress internalization. After incubation, the cells were washed two times with complete media. For each dish, the cells were then trypsinized, resuspended, and counted, and their radioactivity was measured in a gamma counter. The well counter data were decay corrected to the start of each experiment and converted to pmol bound tracer per million cells (pmol/M cells). Specific binding was determined by subtracting the non-specific binding from the total binding. All samples were in triplicate, and each experiment was repeated at least three times with different batches of radioligand. The values for K d and B max were calculated using non-linear curve fitting in GraphPad Prism 6.05 (GraphPad, La Jolla, CA, USA).

Assessment of Binding Specificity In Vitro by Autoradiography
The binding specificity was assessed by an in vitro autoradiography binding assay using frozen sections of cell pellets of huGIPR-HEK293, huGCGR-HEK293, or huGLP1R-HEK293 and frozen sections of human biopsy samples of gastroenteropancreatic neuroendocrine tumor and insulinoma tumor. Biopsies from surgically removed tumors were For the competition studies, human GIP(1-42), C803-GIP, or human GLP1(7-36) ligands were used, wherein the latter served as a negative control. [ 68 Ga]Ga-C803-GIP at a concentration of 5 nM was incubated with frozen sections of pellets of huGIPR-HEK293, huGCGR-HEK293, or huGLP1R-HEK293 cells as described previously [27]. Another set of the sections was co-incubated with an excess of blocking ligands: 5 µM GIP(1-42), 1 µM GLP1(7-36), or 1 µM unlabeled C803-GIP. The sections were incubated at room temperature (RT) for 60 min in phosphate-buffered saline at pH 7.4 (PBS) containing 1% Bovine Serum Albumin (BSA). Next, the sections were washed 1 × 1 min in assay buffer and 2 × 1 min in PBS followed by air drying at 37 • C. The sections were then exposed to a phosphor-imager screen overnight and digitalized using a Cyclone Phosphor Imager system (PerkinElmer). Autoradiogram images were visualized and analyzed with ImageJ (NIH, US).
Frozen tissue sections (10 µm) from human insulinoma and pancreatic tumors were incubated with 5-10 nM [ 68 Ga]Ga-C803-GIP in PBS containing 1% BSA for 60 min at RT. The sections were incubated with radiotracer alone or together with 5 µM GIP(1-42) or 1 µM unlabeled C803-GIP to assess the binding specificity. A set of frozen sections were incubated with the tracer in the presence of 1 µM GLP1(7-36) to evaluate potential crossbinding of the radiotracer to the GLP1R. Duplicate sections were used, and each experiment was repeated at least three times with different batches of radioligand. After incubation, the sections were washed three times in PBS, carefully dried at 37 • C, and thereafter exposed to a digital phosphorimager plate overnight together with a 10 µL droplet of radioactive reference (cross-calibrated against a gamma counter) on absorbent paper attached to an object glass. The phosphor imager plates were scanned using a Cyclone Plus Phosphor imager (Perkin Elmer) at 600 dpi, and the resulting autoradiograms were analyzed by ImageJ. The pixel values in counts/mm 2 were converted to Bq/mm 2 by the included reference. Bq/mm 2 was further converted to fmol/mm 2 by the known molar activity (Bq/fmol) of [ 68 Ga]Ga-C803-GIP.

Organ Distribution of [ 68 Ga]Ga-C803-GIP in Rat
All animal experiments in this study were authorized by the Animal Ethics Committee of the Swedish Animal Welfare Agency and carried out according to the ARRIVE and institutional guidelines ("Uppsala university guidelines on animal experimentation", UFV 2007/724).
The animals were kept at a constant temperature (20 • C) and humidity (50%) in a 12 h light-dark cycle. Food and water were provided ad libitum.
The in vivo organ distribution of [ 68 Ga]Ga-C803-GIP was assessed in Sprague Dawley rats (324 ± 36 g, male, n = 16). The radiotracer was administered into the tail vein of unsedated animals as a bolus in 0.5-0.6 mL of PBS as the vehicle. The animals received 4.1 ± 2.2 MBq/kg 68 Ga]Ga-C803-GIP corresponding to 2.4 ± 1.3 µg/kg peptide mass.
The in vivo organ distribution was conducted for 5 to 180 min, euthanizing two animals at each of eight time points (5,10,20,40,60,90, 120, or 180 min). Samples of blood, heart, lung, liver, pancreas, spleen, adrenal, kidney, small intestine (without/with its content), large intestine (without its content), feces, urinary bladder (rinsed), testis, muscle, bone, bone marrow, thyroid, and brain were immediately harvested, weighed, and measured for radioactivity (gamma counter). The organ uptake of [ 68 Ga]Ga-C803-GIP was corrected for radioactivity decay to the time point of administration and expressed as standardized uptake values (SUVs, Equation (1) The competition experiments were designed and guided by the results of the organ distribution study. Based on the observed rapid clearance of [ 68 Ga]Ga-C803-GIP from the pancreas (known to contain the GIPR), 20 min post-administration was the time point selected for the blocking studies. The rats (351 ± 23 g, male, n = 8) were administered [ 68 Ga]Ga-C803-GIP (18.4 ± 3.7 MBq/kg; 10.7 ± 2.2 µg/kg). A high radioactive dose was selected to potentially increase the tissue signal. Half of them (n = 4) received [ 68 Ga]Ga-C803-GIP alone, while the other half was co-injected with 1 mg/kg of non-labeled C803-GIP. The animals were sacrificed 20 min after the administration of [ 68 Ga]Ga-C803-GIP, and the tissue uptake of the tracer was analyzed as described above.

Organ Distribution [ 68 Ga]Ga-C803-GIP in GIPR-Expressing Xenograft Mice
BALB/c nu/nu mice (n = 12, 18 ± 1 g) were inoculated with huGIPR-HEK293 cells (3 million) in the right hind leg and examined with [ 68 Ga]Ga-C803-GIP after two weeks. Eight mice were used for the ex vivo organ distribution experiment, wherein one group of mice (n = 4) was injected in the tail vein with 0.5 MBq [ 68 Ga]Ga-C803-GIP (corresponding to 4 µg/kg), and the second group of mice was injected in the tail vein with 0.5 MBq [ 68 Ga]Ga-C803-GIP (corresponding to 4 µg/kg) spiked with 1 mg/kg C803-GIP. After 60 min, the mice were euthanized, and the xenograft and selected organs were excised, weighed, and measured for radioactivity. The organ uptake was expressed as SUVs.
Four additional mice were used for in vivo PET imaging, wherein one set of mice (n = 2) was injected in the tail vein with a target dose of 1 MBq [ 68 Ga]Ga-C803-GIP alone, and the second group of mice was injected in the tail vein with 1 MBq [ 68 Ga]Ga-C803-GIP spiked with 1 mg/kg C803-GIP. The injections were given to anesthetized animals who were positioned on the PET scanner bed. The dynamic PET was initiated at the same time as the tracer was administered and lasted for 60 min. Thereafter, the mice were euthanized, the organs were excised, weighed, and measured for radioactivity, and the organ uptake was expressed as SUVs. PET images were reconstructed and summed over 60 min. PET image analysis and generation of Maximum Intensity Projections (MIPS) were performed in PMOD (PMOD Technologies, Zürich, Switzerland). Coronal projection images were taken over the plane of the tumors.

In Vivo Organ Distribution of [ 68 Ga]Ga-C803-GIP in Piglets
An anesthetized pig (25 kg) was positioned to include the abdomen with pancreas in the center of the 25 cm axial field of view of a Discovery MI PET/CT scanner (GE Healthcare, MI, USA) by assistance of a low-dose CT scout view (140 kV, 10 mAs). The details of the animal handling, preparation, and anesthesia have been previously described [28].
Attenuation correction was acquired by a 140 kV, 10-80 mA CT examination. The administration of [ 68 Ga]Ga-C803-GIP (17.3 MBq, corresponding to approximately 2.5 µg peptide or 0.10 µg/kg) was made upon the start of the baseline dynamic scan lasting for 90 min (30 frames; 12 × 10 s, 6 × 30 s, 5 × 120 s, 5 × 300 s, 2 × 600 s). The dynamic scan was followed by a static scan over the brain from 90-120 min post administration. A second PET investigation was performed after an additional 2 h to allow the radioactivity from the first scan to fade from the circulation by decay and excretion.
Before the second PET examination, an infusion with human GIP(1-42) (20 pmol/kg/min) was conducted for a total of 150 min. Thirty minutes after the start of the GIP(1-42) infusion, the pig was administered with [ 68 Ga]Ga-C803-GIP (20.6 MBq, corresponding to 0.13 µg peptide/kg) and scanned according to the same protocol as for the baseline scanning. The GIP(1-42) infusion was continued until the end of the PET scan. Finally, after the last PET scan, a contrast-enhanced CT scan was performed to assist delineation of the tissues, e.g., the pancreas.
Segmentation of the tissues of interest (descending aorta, pancreas, spleen, kidney, liver, and muscle) was performed in PMOD (PMOD Technologies), and time-activity curves were expressed as SUV as detailed above.

Stability of [ 68 Ga]Ga-GIP-C803 in Human Plasma
Stability of [ 68 Ga]Ga-GIP-C803 was studied in human plasma, wherein the agent was incubated at 37 • C for under 30 s, 5 min, 45 min, and 90 min. Thereafter, the samples were analyzed by polyacrylamide gel electrophoresis. To enable the interpretation of the results, a number of references were applied along with the molecular weight marker ( Figure S2). The gel was imaged by GelDoc Go Imaging System (Biorad, Hercules, CA, USA) for protein/peptide signals and, thereafter, exposed to a phosphor-imager screen overnight and digitalized using the Cyclone Phosphor Imager system (PerkinElmer) for radioactivity signals. The amount of the sample radioactivity loaded on the gel was controlled by a reference drop of the same sample volume on a filter paper exposed to the same screen. The images were analyzed by ImageJ (NIH, Bethesda, MD, USA) and Image Lab (Biorad) software.

Human Predicted Dosimetry
The predicted dosimetry in humans was calculated using [ 68 Ga]Ga-C803-GIP organ distribution data obtained in healthy male Sprague Dawley rats. The rat decay-uncorrected SUVs (SUV A ) in various organs were normalized to the whole-body adult reference phantom weights (Equation (2)) [29] in order to assess the residence times (MBq-h/MBq) of the tracer in those organs [30]. The data from the measured time points (5, 10, 20, 40, 60, 90, 120, and 180 min) were used for trapezoidal approximation followed by the extrapolation from the last point to infinity by a single exponential fit. The bone marrow blood volume model was used for the determination of the residence time in bone marrow. The absorbed doses were calculated using OLINDA/EXM version 1.1 with reference to adult male and female phantoms (International Commission on Radiological Protection 60). % organ human = SUV A × g organ kg weight human (2)

Statistics
The data on group level are reported as means ± SD. The differences between the groups were assessed by one-way ANOVA.

Conclusions
A novel peptide construct, C803-GIP, was strategically designed based on the sequences of GIP(1-30) and exendin-4. The radiolabeled analog [ 68 Ga]Ga-C803-GIP demonstrated potential for the visualization of NENs having in vitro binding and selectivity towards the GIPR. However, in vivo binding of [ 68 Ga]Ga-C803-GIP in GIPR-expressing tissues, including xenografts and the pancreas, was relatively low, indicating limited potential in diagnostics. However, in vivo liver uptake was low, potentially providing high-contrast localization of liver metastases. These results may further improve the understanding of the structure-activity relationship for future GIPR-targeting PET tracers and further development of analogs with improved circulatory half-life and tissue exposure, as well as with higher binding affinity while preserving low liver background uptake.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ph16010061/s1, Figure S1: Typical HPLC-chromatograms with UV absorption profile (lower panel) and radioactivity signal profile (upper panel) of [ 68 Ga]Ga-C803-GIP; Figure  Author Contributions: I.V. Researched data, designed the study, interpreted the results, and wrote the manuscript; M.B. conceptualized the study, secured funding, designed the study, interpreted the results, and reviewed the manuscript; T.H. designed and synthesized precursor and reviewed the manuscript; I.L. conceptualized and reviewed the manuscript; S.E. designed, researched, analyzed, wrote, reviewed, and edited the manuscript; L.J. conceptualized the study, secured funding, designed the study, interpreted the results, and revised the manuscript; S.P. conceptualized, designed, and reviewed; M.W. conceptualized the study, secured funding, designed the study, interpreted the results, and revised the manuscript; O.E. conceptualized the study, researched data, designed the study, and wrote the manuscript. O.E. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors have read and agreed to the published version of the manuscript.

Funding:
The study was sponsored by Sanofi and performed in collaboration with Antaros Medical AB. Parts of the study were sponsored by Uppsala Diabetes Center (UDC), ExoDiab, Diabetesfonden, and Barndiabetesfonden.
Institutional Review Board Statement: Biopsies from surgically removed tumors were obtained from Uppsala Biobank (sample collection 827) and their use was approved by Swedish Ethical Review Authority (Dnr 2020-00049).

Informed Consent Statement: Not applicable.
Data Availability Statement: Data is contained within the article and supplementary material.