Novel agonist and antagonist radioligands for the GLP‐2 receptor. Useful tools for studies of basic GLP‐2 receptor pharmacology

Background Glucagon‐like peptide‐2 (GLP‐2) is a pro‐glucagon‐derived hormone secreted from intestinal enteroendocrine L cells with actions on gut and bones. GLP‐2(1–33) is cleaved by DPP‐4, forming GLP‐2(3–33), having low intrinsic activity and competitive antagonism properties at GLP‐2 receptors. We created radioligands based on these two molecules. Experimental approach The methionine in position 10 of GLP‐2(1–33) and GLP‐2(3–33) was substituted with tyrosine (M10Y) enabling oxidative iodination, creating [125I]‐hGLP‐2(1–33,M10Y) and [125I]‐hGLP‐2(3–33,M10Y). Both were characterized by competition binding, on‐and‐off‐rate determination and receptor activation. Receptor expression was determined by target‐tissue autoradiography and immunohistochemistry. Key results Both M10Y‐substituted peptides induced cAMP production via the GLP‐2 receptor comparable to the wildtype peptides. GLP‐2(3–33,M10Y) maintained the antagonistic properties of GLP‐2(3–33). However, hGLP‐2(1–33,M10Y) had lower arrestin recruitment than hGLP‐2(1–33). High affinities for the hGLP‐2 receptor were observed using [125I]‐hGLP‐2(1–33,M10Y) and [125I]‐hGLP‐2(3–33,M10Y) with K D values of 59.3 and 40.6 nM. The latter (with antagonistic properties) had higher B max and faster on and off rates compared to the former (full agonist). Both bound the hGLP‐1 receptor with low affinity (K i of 130 and 330 nM, respectively). Autoradiography in wildtype mice revealed strong labelling of subepithelial myofibroblasts, confirmed by immunohistochemistry using a GLP‐2 receptor specific antibody that in turn was confirmed in GLP‐2 receptor knock‐out mice. Conclusion and implications Two new radioligands with different binding kinetics, one a full agonist and the other a weak partial agonist with antagonistic properties were developed and subepithelial myofibroblasts identified as a major site for GLP‐2 receptor expression.

The metabolite GLP-2(3-33) has been shown to display low intrinsic activity in cAMP accumulation with an E max of 15% of GLP-2 (1-33) and an EC 50 of $6 nM, thus acting as a partial agonist of the GLP-2 receptor (Thulesen et al., 2002). In the same study, it was shown to inhibit the activity of hGLP-2(1-33), thus also displaying antagonistic properties in vitro and in vivo. Structurally, GLP-2 is closely related to the peptide hormones glucagon-like peptide-1 (GLP-1) and gastric inhibitory peptide/glucose-dependent insulinotropic polypeptide (GIP). GIP (secreted from enteroendocrine K cells) and GLP-1 (co-secreted with GLP-2 from L cells) are important insulinotropic hormones, whereas GLP-2 is inactive in this respect (Schiellerup et al., 2019). GLP-1 analogues are widely used as treatment for type 2 diabetes mellitus and obesity, and more recently, a dual-agonist of GLP-1 and GIP showed promising effects within this field (Coskun et al., 2018).
The GLP-2 receptor is a G protein-coupled receptor (GPCR), belonging to the subclass B1 of the GPCR family, which comprises 15 receptors including the GLP-1 receptor the GIP receptor, the glucagon receptor, the secretin receptor and the vasoactive intestinal peptide 1 and 2 receptors (VPAC 1 and VPAC 2 ) (Fredriksson et al., 2003). High resolution structures of class B1 GPCRs combined with mutation studies have enabled the analysis of the active, intermediate and inactive conformations of the receptors, thereby revealing residues that are essential for ligand binding and/or activation Sun et al., 2020;Zhang et al., 2017).
For many years, the leading paradigm regarding ligand binding to class B1 GPCRs was the 'two-step' binding mechanism, suggesting that the C-terminus of the peptide ligand switches between an overall disordered and a more ordered alpha-helical secondary structure. The receptor recognizes and binds the ordered conformation of the peptide ligand, which initiates receptor changes and activation (Parthier et al., 2007). Today, this view has been expanded to include a complex network of conformational changes that takes place upon receptor activation (Liang, Khoshouei, Deganutti, et al., 2018;Venneti & Hewage, 2011;Wu et al., 2020;Zhang et al., 2017). Signalling through class B1 receptors, including the GLP-2 receptor, mainly occurs through Gα s coupling, thereby evoking multiple signalling cascades,

What is already known
• Need of high-affinity radioligands for the GLP-2 receptor.
• GLP-2 receptor mRNA transcript expressed in both intestinal and extraintestinal tissues.

That this study adds
• Description of two GLP-2 based radioligands with different binding kinetics and dual selectivity.
• GLP-2 receptor expression at the protein level in intestinal tissue.

Clinical significance
• Confirms GLP-2 receptor GI-tract protein expression supporting the therapeutic use of GLP-2 in short-bowelsyndrome.
• Different binding kinetics of peptides with different pharmacological properties.
including increased levels of the downstream second messenger cyclic adenosine monophosphate (cAMP). Furthermore, by immunofluorescence microscopy, Estall et al. (2004Estall et al. ( , 2005 showed that the C-terminus of the GLP-2 receptor recruits β-arrestin-2 following agonist stimulation but that this recruitment is not required for desensitization or receptor endocytosis of the GLP-2 receptor. Functional consequences of β-arrestin recruitment by the GLP-2 receptor have not yet been described, although important effects hereof have been demonstrated for other class B1 GPCRs, such as the GIP receptor    (Bjerknes & Cheng, 2001;El-Jamal et al., 2014;Ørskov et al., 2005;Pedersen et al., 2015;Yusta et al., 2000;Yusta et al., 2019) and in the intestinal subepithelial myofibroblasts (SEMF) cell line, CCD-19Co (El-Jamal et al., 2014). Further, mRNA transcripts of the GLP-2 receptor have been reported in various extraintestinal tissues (fat, lymph nodes, bladder, spleen, liver and hepatocytes cells) (El-Jamal et al., 2014;Yusta et al., 2000) including human and rat pancreas (de Heer et al., 2007), a tissue known for high expression levels of the GLP-1 receptor (Richards et al., 2014).
In the present study, we investigated two novel radioligands with tyrosine (Tyr)-substitution at position 10 with methionine (Met) (referred to as M10Y) in the two naturally occurring human GLP-2 (hGLP-2) peptides, the agonist GLP-2(1-33) and its metabolite GLP-2 (3-33), the partial agonist/antagonist. With these, we determined differential binding kinetics in vitro. We performed autoradiography studies in mice and hereby showed GLP-2 receptor protein in the GI  Before binding assays were performed, the eluted fractions were tested in homologous competition binding (see next section for method).

| Kinetic binding experiments
The association assays were performed by preparing a mixture of  gamma-counter to determine the amount of radioactivity injected into the animals. Ten minutes after peptide injection, the thorax was opened, after which the vascular system was perfused at a constant flow with 0.9% saline with an outlet through the right ventricle. The body temperature of the animals during the precedure were maintained using a heat lamp while no breating assistence were provided. Next, the mice were fixated by flushing the system with icecold 4% paraformaldehyde. After fixation, the pancreas, small intestine and kidneys (as positive control) were removed and stored in 45% paraformaldehyde until further processing.

| Autoradiography
Small intestinal, pancreatic and kidney tissue samples were embedded in paraffin, and histological 4 μm sections were cut with a microtome and placed on glass slides. The sections were dewaxed and coated in a dark room with 43-45 C Kodak NTB emulsion (VWR, Herlev, Denmark) diluted 1:1 with 43-45 C water and subsequently dried and stored in light-proof boxes at 5 C for 6 weeks. After 6 weeks, the tissue sections were developed in a dark room in Kodak D-19 developer (VWR, Herlev, Denmark) for 5 min, dipped 10 times in 0.5% acetic acid and fixated in 30% sodium thiosulphate for 10 min. The sections were then washed, first in water for 10 min and then in 70% ethanol. Finally, the sections were lightly counterstained with haematoxylin and examined with a light microscope (Orthoplan, leitz).
Images were taken with an AxioCam ICc5 camera (Zeiss) connected to the light microscope.

| Immunohistochemistry
The Immuno-related procedures used comply with the recommendations made by the British Journal of Pharmacology . Specimens of pancreas and small intestine (n = 5) from wild-type mice and GLP-2 receptor KO mice were fixed in formalin buffer 10% and embedded in paraffin. The tissue blocks were cut in sections of 4 μM and dewaxed through xylene to tap water.
For antigen retrieval, the sections were boiled in a microwave oven for 15 min in a 10 mmol Tris-EDTA-buffer pH 9 followed by pre-incubation in 2% BSA for 10 min and an overnight incubation at 4 C with the polyclonal primary rabbit GLP-2 receptor antibody 99,077 diluted 1:16,000 in PBS containing in addition 2% BSA (Ørskov et al., 2005). On day 2, the sections were washed and incubated with biotinylated secondary goat-anti rabbit antibody (Vector

| Design of the experiments and statistical analysis
Studies were designed to generate groups of equal sizes. In some cases, the group size for the in vitro experiments was below 5 as sufficient data were obtained with less experiments. For in vivo autoradiography and immunohistochemistry, five animals were included in each group. Due to the three Rs and the loss of one animal in each group, only four animals per group are shown. For these data, no statistical analysis was conducted. Special randomization and blinded analysis were not necessary for the in vitro studies because the cells used in the experiments could be maintained under the same experimental conditions. To avoid experimental bias, the same experiment was not repeated more than twice a week. The data and statistical analysis comply with the British Journal of Pharmacology on experimental design and analysis in pharmacology (Curtis et al., 2018). Statistical analysis was only carried out for data with a group size of n ≥ 5 for each group. Statistical significances between dose-response curves and B max were analysed using paired Student's t-test with a level of significance of 0.05. The group size, n, value represents the number of independent experiments, and statistical analysis was carried out using independent values. To compensate for inter-assay variations, the data were normalized to the binding of reference radioligand or activation by the endogenous full agonist (see figure legends for further explanation in each assay). All experimental data points were included in the data analysis. The analysis was interpreted using Gra-phPad Prism 8.0 software (Graphpad software, RRID:SCR_002798) to obtain the following parameters: IC 50 , EC 50 , E max , k on , k off , k obs and B max . All sigmoidal curves were fitted with a Hill slope of either 1 for activation curves or À1 for inhibition curves. All data were expressed as mean ± standard error of the mean (SEM), if not otherwise stated.
2.10 | Data analysis B max (the total density of receptors in the sample) were calculated from homologous competitive binding curves according to Equation 1 (Richards et al., 2014): where B 0 is the total specific binding in CPM and [L] is the concentration of radioligand in nM.
The equilibrium dissociation constant (K D ) was also calculated from the homologous competitive binding curves according to Equation 2: The inhibition constant (K i ) was obtained from the heterologous competitive binding curves by using the Cheng and Prusoff equation according to Equation 3 (Cheng & Prusoff, 1973): The association rate constant (k on ) was calculated according to Equation 4 (van der Velden et al., 2020): where k obs is the observed association rate constant (min À1 ) and k off is the dissociation rate constant (min À1 ).
The K D calculated from the kinetic parameters was calculated according to Equation 5: where k off and k on are determined from the membrane binding.
Thus, the EC 50 of hGLP-2(1-33) increased with 3.4-and 11.6-fold in the presence of 100 nM and 1 μM of hGLP-2(3-33), respectively, and 4.8-and 16.3-fold in the presence of similar doses of hGLP-2(3-33,M10Y). Schild plot analysis revealed that both acted as competitive antagonists of G protein-mediated signalling with a Hill slope of 1.15 ± 0.11 and 1.13 ± 0.11 for hGLP-2(3-33) and hGLP-2(3-33,M10Y), respectively, and with pA2 values of 51.9 and 37.6 nM, respectively ( Figure 1f,g). These data demonstrate that the two N-terminally truncated GLP-2 variants act as competitive antagonists in the presence of the endogenous agonist, and alone display weak intrinsic activity and thereby partial agonistic properties in the cAMP accumulation assay.
Since ligand-receptor binding kinetics is considered to be a key determinant of ligand efficacy and onset of action (van der Velden et al., 2020), we determined the association (k on ) and dissociation (k off ) rates, using membranes prepared from cells stably expressing the hGLP-2 receptor. For both radioligands, the kinetic profiles were best fitted with a one-phase association and a one-phase dissociation.

| Selectively binding profile of the two radioligands for related class B1 GPCRs
Given the high sequence similarity between class B1 receptors and their peptide ligands, we next determined whether the two radioligands cross-reacted with the closely related class B1 receptors hGIP receptor, hGLP-1 receptor, hglucagon, hsecretin receptor, hVPAC 1 receptor and hVPAC 2 receptor (Figure 4a). While we observed no specific binding for five of the six receptors, a low but significant binding was observed for both radioligands to the hGLP-1 receptor (Figures 4b and S1). This cross-reaction intrigued us to test the opposite pairing with binding of [ 125 I]-hGLP-1(7-36) to the hGLP-2 receptor, which turned out to be undetectable (Figure 4c). In addition to GLP-2(1-33) and GLP-2(3-33) identified in our study, a broad range of other peptides are known to bind to the GLP-1 receptor (glucagon, oxyntomodulin, besides GLP-1 receptor) Jorgensen et al., 2007;Skov-Jeppesen et al., 2019). In contrast, this broad specificity in binding does not seems to be the case for the GLP-2 receptor, which seemingly exhibits a narrower binding of only GLP-2-based ligands.
We found that the M10Y-modification in hGLP-2(1-33) had a minor impairment on the binding affinity and potency in cAMP accumulation, and a larger effect on β-arrestin recruitment as both potencies and efficacies were affected on β-arrestin 1 as well as β-arrestin 2 recruitment. This could reflect the general weaker arrestin recruitment compared to Gα s coupling of class B1 receptors as described for the GIP receptor    Jorgensen et al., 2007) and the activation of the GLP-1 receptor by glucagon (Svendsen et al., 2018). Thus, cross-activation is a common phenomenon within class B1 GPCRs, which is reflected in the high sequence similarities observed among the receptors and across species. For rodent GLP-2 receptors, 81% and 79% sequence identities are found for the mGLP-2 receptor and rGLP-2 receptor to the hGLP-2 receptor, respectively, explaining the high-affinity binding observed for both radioligands to the rodent GLP-2 receptors.
In conclusion, we developed two new radioligands for the GLP-2 receptor; both with high affinity to the human, rat and mouse GLP-2 receptor, and with low affinity for the mouse and human GLP-1 receptor. With these, we show differential binding kinetics of full agonist and partial agonist with antagonistic properties to the GLP-2 receptor and confirm GLP-2 receptor expression at the protein level in the GI tract's subepithelial myofibroblasts.
Our observations are of importance for tissue localization and structural characterization for not only the GLP-2 receptor, but also for other class B1 GPCRs.
Larsen for the assistance with oxidation iodination and Siv

CONFLICT OF INTEREST
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. M.M.R, J.J.H and B. H are founders of Bainan Biotech but declare that the research was conducted in the absence of any commercial or financial relationships.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request. Some data may not be made available because of privacy or ethical restrictions.

REGOUR
This declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research as stated in the BJP guidelines for Design and Analysis, Immunoblotting and Immunochemistry and Animal Experimentation, and as recommended by funding agencies, publishers and other organizations engaged with supporting research.