Design of potent and proteolytically stable biaryl-stapled GLP-1R/GIPR peptide dual agonists

Abstract

Recent clinical trials have revealed that the chimeric peptide hormones simultaneously activating glucagon-like peptide-1 receptor (GLP-1R) and glucose-dependent insulinotropic polypeptide receptor (GIPR) demonstrate superior efficacy in glycemic control and bodyweight reduction, better than those activating the GLP-1R alone. However, the linear peptide-based GLP-1R/GIPR dual agonists are susceptible to proteolytic cleavage by common digestive enzymes present in the gastrointestinal tract and thus not suitable for oral administration. Here, we report the design and synthesis of biaryl-stapled peptides, with and without fatty diacid attachment, that showed potent GLP-1R/GIPR dual agonist activities. Compared to a linear peptide dual agonist and semaglutide, the biaryl-stapled peptides displayed drastically improved proteolytic stability against the common digestive enzymes. Furthermore, two stapled peptides showed excellent efficacy in an oral glucose tolerance test in mice, owing to their potent receptor activity in vitro and excellent pharmacokinetics exposure upon subcutaneous injection. By exploring a more comprehensive set of biaryl staplers, we expect this stapling method could facilitate the design of the stapled peptide-based dual agonists suitable for oral administration.

Graphical Abstract

INTRODUCTION

The incidence of diabetes has risen to an epidemic proportion in today’s world. It was estimated that 463 million people had diabetes in 2019, with type 2 accounting for more than 90%.^1–2 Since the approval of exenatide by the FDA in 2005, the glucagon-like peptide-1 receptor agonists (GLP-1 RAs) have continued to expand to include once-a-day injectable liraglutide and once-weekly injectables dulaglutide and semaglutide. Collectively, GLP-1 RAs have offered millions of diabetes patients excellent treatment options to control their blood glucose and bodyweight. GLP-1 RAs lower blood glucose by increasing insulin secretion in a glucose-dependent manner and suppressing glucagon secretion, and reduce bodyweight by suppressing appetite and delaying gastric emptying.^3–6 Despite this success, many patients receiving GLP-1 RA treatment still could not reach their glycemic target, suggesting more efficacious agents are needed.⁷ To this end, recent studies showed that peptides activating both GLP-1R and glucose-dependent insulinotropic polypeptide receptor (GIPR) showed greater efficacy in glucose-lowering and bodyweight reduction than those only activate GLP-1R. For example, Finan et al. reported a GLP-1R/GIPR dual agonist, RG7697, that showed greater efficacy in improving glucose tolerance than liraglutide in cynomolgus monkeys.⁸ Coskun et al. reported a GLP-1R/GIPR dual agonist, LY3298176, that showed comparable efficacy to semaglutide in improving glucose tolerance in intraperitoneal glucose tolerance tests, but significantly higher efficacy in lowering body weight than semaglutide in a diet-induced obesity mouse model.⁹ Owing to their outstanding efficacies, both RG7697 and LY3298176 have entered into clinical trials.^10–15 These studies highlighted the potential of GLP-1R/GIPR dual agonist to become the next generation antidiabetic drug.

For improved compliance and broader (and earlier) adoption by diabetic patients, it is highly desirable that these efficacious GLP-1R/GIPR dual agonists are made available for oral delivery. However, the oral administration of peptide therapeutics is challenged by their gastrointestinal instability and inefficient uptake into the circulation, leading to generally low oral bioavailability (<1%). Several strategies have been successfully developed in recent years to meet this challenge.^16–17 These include formulation with permeation enhancers,¹⁸ development of the liposome¹⁹ and nanoparticle-based²⁰ delivery systems, peptide chemical modifications²¹ such as cyclization, N-methylation, and lipidation, and conjugation with cell-penetrating peptide motifs. A breakthrough in formulation approach came in October 2019 when the FDA approved oral semaglutide, a long-acting GLP-1R-only agonist for treating type 2 diabetes;²² the permeation enhancer SNAC was formulated with the fatty-diacid modified semaglutide for efficient transcellular stomach absorption.²³ However, low bioavailability of 0.8% after oral administration²⁴ makes daily administration of a semaglutide tablet necessary to avoid wide fluctuations in drug exposure. In addition, it must be taken on an empty stomach, and for 30 minutes after taking oral semaglutide, no other food, drink, or medication can be taken to permit undisturbed absorption. To get around this complex dosing scheme, a multifaceted approach combining amino acid backbone α-methylation and side-chain lipidations with permeation enhancers has afforded a peptide GLP-1 RA clinical candidate, MEDI7219, with bioavailability of ~6% in dogs receiving oral tablets,²⁵ suggesting chemical modifications may be needed for more conventional oral administration of peptide drugs.

Since GLP-1 and GIP bind to their cognate receptors in an α-helical conformation,^26–27 modifications that stabilize peptide α-helix should increase their binding affinity and improve their proteolytic stability. Indeed, numerous studies have shown that stabilizing α-helical conformation via peptide stapling has improved receptor activity and proteolytic stability.^28–32 Recently, we reported biaryl stapling chemistry and its application to several bioactive peptides to improve proteolytic stability, cell permeability, and receptor activity.^33–35 We surmise that biaryl stapling may produce the peptide GLP-1R/GIPR dual agonists with improved physicochemical properties suitable for oral administration. Here, we report the design of a series of biaryl-stapled GLP-1R/GIPR dual agonists and the successful identification of biaryl stapled peptides with sub-nanomolar receptor activity, high proteolytic stability, and excellent glucose-lowering effect in mice.

RESULTS AND DISCUSSION

Design, Synthesis, and Evaluation of Biaryl-Stapled Peptide Dual Agonists.

To identify a suitable peptide template for designing biaryl-stapled dual agonists, we considered RG7697 for its potent activity toward GLP-1R and GIPR.⁸ RG7697 is a chimera of GLP-1 and GIP peptides and contains a C-terminal exendin-4-derived helix cap and aminoisobutyric acids for helix stabilization and DPP-IV resistance, respectively (Figure 1a). A C16-fatty acid was appended at the C-terminus to permit serum albumin binding for an extended serum half-life. Accordingly, we designed DA1 with the same amino acid sequence as RG7697 but without the fatty-acid modification (Table 1). To identify suitable stapling sites, we took advantage of the recently reported cryo-EM structures of GLP-1R in complex with GLP-1 and a Gs heterotrimer²⁶ and GIPR in complex with GIP and a Gs heterotrimer,²⁷ which show the binding of peptide ligands toward their respective receptors (Figure 1b, 1c). Since the N-termini of these peptides are highly conserved and crucial for receptor activity, we decided to carry out Cys substitutions to the middle region of RG7697. Inspection of the cryo-EM structures revealed a pair of i,i+7 biaryl stapling sites: one between residues 14 and 21, and the other between residues 17 and 24, as these sites are not directly involved in binding to either the transmembrane domains (TMD) or the extracellular domains (ECD) of GLP-1R and GIPR (Figure 1b, 1c).

Figure 1.

Identifying appropriate sites on the peptide for biaryl stapling. (a) Sequences of GLP-1, GIP, exendin-4, and RG7697: X = Aib; K’ = Lys-C₁₆. The conserved residues responsible for receptor binding are highlighted in box. (b) Cryo-EM structure of GLP-1 bound to GLP-1R (PDB code: 5VAI). (c) Cryo-EM structure of GIP_1–42 bound to GIPR (PDB code: 7DTY): residue 31–42 of GIP_1–42 not shown. The selected stapling sites are colored in blue (residue 14 and 21) and red (residue 17 and 24), respectively, on the GLP-1 and GIP_1–42 ligands.

Table 1.

Sequences of the GLP-1R/GIPR peptide dual agonists and their activities^a

		Agonist activity (nM)
name	Sequence	GLP-1R	GIPR
DA1	YXEGTFTSDYSIYLDKQAAXEFVNWLLAGGPSSGAPPPSK	0.02	0.06
DA2	YXEGTFTSDYSIYLDKCAAXEFVCWLLAGGPSSGAPPPS	1.91	1.43
DA2-Bph	YXEGTFTSDYSIYLDKC’AAXEFVC’WLLAGGPSSGAPPPS	0.12	0.11
DA3	YXEGTFTSDYSIYCDKQAAXCFVNWLLAGGPSSGAPPPS	2.68	2.14
DA3-Bph	YXEGTFTSDYSIYC’DKQAAXC’FVNWLLAGGPSSGAPPPS	0.82	1.64
DA4	YXEGTFTSDYSIYLDKCAAXEFVCWLLAG	4.62	3.03
DA4-Bph	YXEGTFTSDYSIYLDKC’AAXEFVC’WLLAG	0.52	0.08
DA4-Bpy	YXEGTFTSDYSIYLDKC”AAXEFVC”WLLAG	0.06	0.04
DA5	YXEGTFTSDYSIYLDKCAAXEFVCWLLA	21.24	29.75
DA5-Bph	YXEGTFTSDYSIYLDKC’AAXEFVC’WLLA	1.82	0.56
DA5-Bpy	YXEGTFTSDYSIYLDKC”AAXEFVC”WLLA	0.14	0.17
DA6	YXEGTFTSDVSIYLDKCAAXEFVCWLLAG	16.22	23.85
DA6-Bph	YXEGTFTSDVSIYLDKC’AAXEFVC’WLLAG	3.04	1.35
DA6-Bpy	YXEGTFTSDVSIYLDKC”AAXEFVC”WLLAG	0.21	0.36

^aAll peptides contain unmodified N-termini and amidated C-termini; X = Aib; C’ = Bph-stapled cysteine; C” = Bpy-stapled cysteine.

Accordingly, we designed two peptide-based dual agonists, DA-2 and −3, by substituting cysteines at positions 14/21 and 17/24, respectively. We obtained the corresponding stapled peptides after performing the cysteine-mediated stapling reaction with 4, 4’-bis(bromomethyl)biphenyl (Bph). Using a cell-based luciferase reporter assay, we determined the EC₅₀ values of DA1 to be 0.02 nM and 0.06 nM toward GLP-1R and GIPR, respectively. Compared with DA1, the dicysteine-substituted DA2 and DA3 showed more than 90 and 20-fold reduced activity toward GLP-1R and GIPR, respectively (Table 1; SI, Figure S5). To our delight, stapling of DA2 by Bph led to 10-fold increases in receptor activity; indeed, DA2-Bph was only 6- and 2-fold less potent than DA1 in activating GLP-1R and GIPR, respectively. On the other hand, the Bph-stapling with DA3 led to a moderate increase in receptor activity (3-fold toward GLP-1R; < 2-fold toward GIPR). Therefore, we selected DA2-Bph for subsequent further optimization.

The C-terminal exendin-4 sequence adopts a Trp-cage structure³⁶ to stabilize the peptide helix. Since biaryl stapling was known to increase the helical propensity of peptides, we hypothesize that we can remove the Trp-cage to simplify the peptide sequence without losing receptor activity. Thus, we designed a truncated DA2 peptide with 29 residues, DA4, and obtained the stapled DA4-Bph by following the same stapling procedure. To our satisfaction, DA4-Bph displayed no change in GIPR activity and only 4-fold reduction in GLP-1R activity compared to DA2-Bph, suggesting the Trp-cage mediated helix stabilization is recapitulated with Bph-stapling. Furthermore, stapling DA4 with analogous 4, 4’-bis-bromomethyl-bipyridyl (Bpy) led to DA4-Bpy with greater potency; DA4-Bpy showed 9-fold higher activity toward GLP-1R and 2-fold higher activity toward GIPR than DA4-Bph (Table 1). This remarkable improvement was also observed previously with the Bpy-stapled oxyntomodulin analog.³⁵ Further truncation at the C-terminus was not tolerated as removal of Gly-29 led to substantial reductions in receptor activity for DA5 and its stapled analog DA5-Bph (Table 1). To probe whether we can further improve GLP-1R agonist activity, we replaced Tyr-10—a native residue in GIP—with Val-10 found in GLP-1 and obtained DA6 and its stapled analogs. However, all DA6 analogs showed significantly reduced activity toward GLP-1R and GIPR than their DA4 counterparts (Table 1). Therefore, we selected DA4 as the lead peptide dual agonist in subsequent studies.

Design and Synthesis of Lipid-Containing Biaryl-Stapled Peptide Dual Agonists.

Acylation of peptide hormones with a long lipid chain has been shown to increase peptide’s in vitro intestinal permeability.^37–38 To further enhance oral absorption, we appended a fatty acid chain, a motif used in the design of oral drug candidate MEDI7219,²⁵ to the biaryl stapler to simplify the synthesis. Since the long-chain fatty diacid attachment has significantly reduced the stapled oxyntomodulin analogs’ receptor activity in our previous study,³⁹ we appended the fatty diacid to the biaryl stapler via a labile ester linkage. We envision that serum esterase will hydrolyze the ester linkage once the lipid-modified stapled peptides get absorbed into the systemic circulation from the gastrointestinal tract, generating the potent stapled peptides carrying a hydroxyl group on the biaryl stapler. Accordingly, we synthesized four fatty diacid-modified Bph/Bpp staplers with a fatty diacid chain length of 10 or 18, along with the protected Bph-OTBDPS and Bpp-OTBDMS as controls (Figure 2a; SI, Schemes S1–S4). We prepared a series of biaryl-stapled DA4 analogs using these new staplers under the same reaction conditions (Figure 2b). Because C10- and C18-modified biaryl staplers are unsymmetrical, we expect to obtain a pair of regioisomers from the stapling reaction. Indeed, while only one product peak was observed for DA4-Bph-C10 in HPLC (SI, Figure S1), two product peaks with identical masses were observed for DA4-Bph-C18, DA4-Bpp-C10, and DA4-Bpp-C18 (SI, Figures S2–S4).

Figure 2.

(a) Structures of four new biaryl staples containing a labile ester bond. (b) Scheme for the synthesis of the fatty diacid-modified biaryl-stapled peptides.

Ester Stability in Serum for Lipid-Containing Biaryl-Stapled Peptide Dual Agonists.

To assess how fast ester hydrolysis proceeds in serum upon oral absorption, we incubated the fatty diacid-modified DA4 analogs with fresh mouse serum and monitored the bond cleavage by LC-MS (Figure 3a). The Bph-stapled analogs showed longer half-lives (t_1/2 = 11 hours) than the Bpp-stapled ones (Figure 3b). Moreover, the C18-modified DA4-Bpp showed a longer half-life (t_1/2 = 7.5 hours) than the C10-modified one (t_1/2 = 2.5 hours), presumably because the longer chain fatty diacid promotes tighter association with serum albumin protein, which protects the lipid-modified stapled peptide from the esterase. Since the C10/C18-modified Bph staplers are more hydrophobic, they bind to serum albumin protein tighter and exhibit longer half-lives than the corresponding Bpp analogs.

Figure 3.

Assessment of ester hydrolysis in mouse serum for the fatty diacid modified stapled peptides. (a) Reaction scheme. (b) Time courses of the ester hydrolysis as monitored by LC-MS. The total ion counts of the ester starting materials and the hydrolyzed product were used in quantification.

Receptor Activities of Lipid-Containing Biaryl-Stapled Peptide Dual Agonists.

To probe whether the ester hydrolyzed products remain active, we assessed their receptor activities using the luciferase reporter assay (Table 2; SI, Figure S6). Compared to DA4-Bph, DA4-Bph-OH showed roughly 2-fold lower GLP-1R activity and 5-fold lower GIPR activity. Interestingly, DA4-Bpp-OH showed 1.5-fold higher GLP-1R activity and 2-fold higher GIPR activity than DA-Bph-OH, suggesting that isosteric nitrogen replacement on the biphenyl ring leads to more robust interactions with both receptors. However, whereas the fatty diacid-modified DA4-Bph peptides showed reduced receptor activity compared to DA4-Bph, two C10-modified DA4-Bpp analogs unexpectedly showed higher receptor activity than DA4-Bpp-OH (Table 2), indicating that the C10-fatty diacid attachment at the biaryl ring enhances interactions with the two receptors. In contrast, the two C18-modified DA4-Bpp analogs showed pronounced loss of GLP-1R activity but a modest loss of GIPR activity (Table 2), suggesting that the GLP-1R and GIPR may have unique dispositions of hydrophobic patches within their respective ligand-binding pockets.

Table 2.

Agonist activities of the fatty diacid-modified stapled DA4 analogs

		Agonist activity (nM)
Name	Stapler structure	GLP-1R	GIPR
DA4-Bph-OH a		0.98	0.42
DA4-Bpp-OH b		0.64	0.21
DA4-Bph-C10 c		1.45	0.15
DA4-Bph-C18-isomer1		25.43	2.13
DA4-Bph-C18-isomer2		14.31	1.80
DA4-Bpp-C10-isomer1		0.14	0.07
DA4-Bpp-C10-isomer2		0.25	0.11
DA4-Bpp-C18-isomer1		2.04	0.48
DA4-Bpp-C18-isomer2		2.29	0.70

^{a, b, c}DA4-Bph-OH, DA4-Bpp-OH, and DA4-Bph-C10 were used as a mixture of regioisomers.

Proteolytic Stabilities of Biaryl-Stapled Peptide Dual Agonists.

To probe whether biaryl stapling increases proteolytic stability of the stapled peptides in the gastrointestinal tract, we selected those with large quantities and incubated them with some common digestive enzymes such as pepsin, trypsin, and chymotrypsin. For pepsin, while the unstapled DA1 and semaglutide control were quickly degraded (t_1/2 < 2 min), all stapled peptides showed substantially higher stability (Figure 4a). Among them, DA4-Bpp-C18-isomer2 exhibited a half-life of close to 3 hours, significantly longer than the C18-fatty diacid modified semaglutide (Figure 4a), indicating that biaryl stapling is primarily responsible for the enhanced stability. For trypsin, all DA4 analogs remained intact after 8-hour incubation (Figure 4b), attributable to the lack of basic residues in the sequences. Finally, for chymotrypsin, whereas DA4-Bpy and DA4-Bpp-C10-isomer2 exhibited resistance similar to DA1 and semaglutide, all other stapled peptides showed significantly greater proteolytic stability (t_1/2 = 59 – 610 min). Notably, two C18-modified stapled peptides show 50- and 20-fold improvements over DA1 (Figure 4c), presumably due to the addition of the long C18 chain that presents a steric hindrance to the protease. Interestingly, the Bph-stapled analogs displayed higher stability than their Bpy/Bpp counterparts, which can be attributed to the propensity of the Bph-stapled peptides to form the oligomers through hydrophobic interactions. Moreover, the stapled peptides displayed excellent serum stability with half-lives >24 hours, substantially longer than the linear peptide DA1 but similar to semaglutide, which is known to have a very long half-life due to its tight binding to serum albumin protein⁴⁰ (Figure 4d).

Figure 4.

Stability of selected peptide analogs in the presence of (a) pepsin; (b) trypsin; (c) chymotrypsin; and (d) fresh mouse serum.

Glucose-Lowering Effect and Pharmacokinetics of Biaryl-Stapled Peptide Dual Agonists.

To evaluate the in vivo efficacy of the stapled dual agonists, we selected the four most potent stapled peptides that we had sufficient quantities in hand: DA4-Bph, DA4-Bpy, DA4-Bpp-C10-isomer2, and DA4-Bpp-C18-isomer2, and put them through an oral glucose tolerance test (OGTT) in mice. Administration of DA4-Bpy and DA4-Bpp-C10-isomer2 through subcutaneous injection 1 hour before glucose challenge led to glucose reductions of 43% and 47%, respectively, based on the area under the curve (AUC), comparable to semaglutide—a potent GLP-1R-only agonist (Figure 5a and 5b). The OGTT results are consistent with these two stapled peptides’ superb in vitro receptor activities (Tables 1 and 2). On the other hand, although DA4-Bph exhibited higher receptor activity than DA4-Bpp-C18-isomer2 (4-fold higher toward GLP-1R; 9-fold higher toward GIPR), subcutaneous injection of DA4-Bph led to lower glucose reduction (24%) than DA4-Bpp-C18-isomer2 (32%). To gain insights into the origin of the observed efficacies, we performed a single time point pharmacokinetic (PK) study in which the mouse plasma was collected at the end of the OGTT (3 h post-injection) and the plasma peptide concentrations were determined using a cell-based luciferase reporter assay (Figure 5c). The C18-fatty diacid-modified semaglutide displayed the highest plasma concentration (863 nM), indicating excellent PK exposure. The two most active stapled peptides, DA4-Bpy and DA4-Bpp-C10-isomer2, showed good PK exposure with plasma concentrations of 557 nM and 233 nM, respectively, and thus high OGTT efficacy. In contrast, DA4-Bph gave the lowest PK exposure (94 nM), explaining its low OGTT efficacy despite excellent receptor activity. DA4-Bpp-C18-isomer2 showed excellent PK exposure (561 nM); however, due to its moderate receptor activity (Table 2), it displayed a modest OGTT efficacy. Because of the short duration of this test (< 3 h), we do not expect the ester stability to have a significant effect on the PK exposure of the two fatty diacid-modified stapled peptides. Notably, the GLP-1R-only agonist semaglutide gave a comparable glucose-lowering efficacy in OGTT, presumably due to its prolonged PK exposure (863 nM at 3 h) and optimized GLP-1R receptor activity. The two stapled peptide dual agonists, DA4-Bpy and DA4-Bpp-C10-isomer2, gave the highest efficacy in OGTT, likely due to a combination of excellent receptor activity and good PK exposure.

Figure 5.

Oral glucose tolerance test of the stapled DA4 analogs. C57BL/6 male (age 10 weeks; n = 4) were injected with the peptide at a dosage of 0.4 mpk subcutaneously 1 hour before the glucose challenge. (a) Mouse blood glucose concentrations measured by a glucometer at 0, 15, 30, 60, and 90 minutes. (b) Bar graph showing total glucose amounts in the treated mice by measuring area under the curve (AUC) over the monitoring period. ****P < 0.0001 vs. vehicle control; ***P < 0.001 vs. vehicle control using Ordinary One-Way ANOVA analysis in GraphPad Prism 9.2. (c) Peptide plasma concentration at 3 hours post-injection measured by the cell-based luciferase reporter assay.

CONCLUSIONS

In summary, we have designed and synthesized a series of biaryl-stapled peptides that showed potent GLP-1R/GIPR dual agonist activities. Compared to the unmodified dual agonist DA1 and linear fatty diacid-modified semaglutide, the stapled peptides generally showed improved proteolytic stability against common digestive enzymes found in the gastrointestinal tract and fresh mouse serum. In an oral glucose tolerance test in mice, both DA4-Bpy and DA4-Bpp-C10-isomer2 showed high levels of efficacy comparable to semaglutide. Subsequent studies indicated that these two stapled peptides displayed excellent pharmacokinetics with the subcutaneous injection. Our studies demonstrate that highly potent and proteolytically stable GLP-1R/GIPR dual agonists can be obtained readily through biaryl stapling of linear bioactive peptides carrying two cysteines at i,i+7 positions. We further tuned the stapled peptides’ physicochemical properties by attaching a long-chain fatty diacid to the biaryl stapler without altering the peptide sequence. While we continue to optimize the PK/PD properties of the stapled peptide-based GLP-1R/GIPR dual agonists, we expect this biaryl stapling strategy should facilitate the design and discovery of chemically modified peptide hormones suitable for oral administration.

METHODS

Synthesis of the stapled peptide dual agonists.

All peptides were purchased from either LabNetwork or GenScript USA and used directly in biaryl stapling reactions. 4, 4’-Bis-bromomethyl-biphenyl (Bph) was purchased from TCI America. 4, 4’-Bis-bromomethyl-bipyridyl (Bpy) was synthesized by following a published procedure.⁴¹ Other biaryl staplers were synthesized according to the method described in the Supporting Information. The stapling reactions were carried out by incubating 2 mg cysteine-containing peptides with 1.2 equivalent biaryl stapler in 600 μL acetonitrile/30 mM NH₄HCO₃ solution (1:1, pH 8.5). The mixture was stirred at room temperature, and the reaction progress was monitored by analytical HPLC or LC-MS. After more than 95% peptide was consumed, the product was directly purified by preparative HPLC. For the synthesis of DA4-Bph-OH and DA4-Bpp-OH, DA4 was first reacted with Bph-OTBDPS and Bpp-OTBDMS, respectively, under the condition described above. When over 95% DA4 was consumed, 200 μL of 100 mM KF solution was added to the product mixture to remove the TBDPS or TBDMS protecting group. After the deprotection was complete based on analytical HPLC, the mixture was subjected to preparative HPLC.

In vitro receptor activation reporter assays.

HEK293-GLP-1R-CRE and HEK293-GIPR-CRE cells were seeded in a white 384-well plate at a density of 5,000 cells per well and cultured for 24 hours in DMEM with 10% FBS at 37°C with 5% CO2. Cells were treated with different peptides at varying concentrations. After 24 hours, 10 μL of Bright-Glo reagent (Promega) was added to each well and luminescence was determined using an Envision multilabel plate reader (PerkinElmer). The EC₅₀ of each peptide was calculated using GraphPad Prism 9.2 software.

Proteolytic stability against pepsin.

Semaglutide was purchased from Peptides International and used directly in the assay. The peptide was dissolved in 10 mM HCl solution containing 1% DMSO to afford a final concentration of 30 μM. The pepsin stock solution was prepared by dissolving pepsin (Promega) in 10 mM HCl solution. The pepsin stock solution was added to the peptide solution to afford a final concentration of 3 μg/ml. The mixture was incubated at 37 °C with shaking. Aliquots (5 μL) of the solution were taken and mixed with 200 μL acetonitrile/water (1:1) at specific time points, and the mixture was analyzed by LC-MS. The intact peptide in the mixture was quantified by the integration of the peptide ion counts.

Proteolytic stability against trypsin.

The peptides were dissolved in PBS solution (10 mM Na₂HPO₄, 1.8 mM KH₂PO₄, 137 mM NaCl, 2.7 mM KCl, pH 7.4) containing 0.025% Tween-20 and 1% DMSO to afford a final concentration of 30 μM. The trypsin stock solution was prepared by dissolving trypsin (Sigma-Aldrich) in PBS solution. The resulting trypsin stock solution was added to the peptide solution to afford a final concentration of 0.5 μg/mL. The mixture was incubated at 37°C with shaking. Aliquots (5 μL) of the peptide solution were taken and quenched by adding 200 μL acetonitrile/water (1:1) at specific time points. The mixture was then analyzed by LC-MS. The intact peptide in the mixture was quantified by the integration of the peptide ion counts.

Proteolytic stability against chymotrypsin.

The peptide was dissolved in PBS solution (10 mM Na₂HPO₄, 1.8 mM KH₂PO₄, 137 mM NaCl, 2.7 mM KCl, pH 7.4) containing 0.025% Tween-20 and 1% DMSO to afford a final concentration of 30 μM. The chymotrypsin stock solution was prepared by dissolving chymotrypsin (Sigma-Aldrich) in PBS solution. The chymotrypsin stock solution was added to the peptide solution to afford a final concentration of 0.5 μg/mL. The mixture was incubated at 37°C with shaking. Aliquots (5 μL) of the peptide solution were taken and quenched by adding 200 μL acetonitrile/water (1:1) at specific time points. The intact peptide in the mixture was analyzed by LC-MS as described above.

Peptide stability in mouse serum.

To a 1-μL peptide stock solution (3 mM in DMSO) was added to 99 μL fresh mouse serum (Invitrogen), and the mixture was incubated at 37 °C with shaking. At a specific time point, aliquots (3 μL) of the solution were taken and mixed with 200 μL ice-cold acetonitrile/water (1:1). The mixture was centrifuged at 13,500 rpm for 10 min. The supernatant was collected and analyzed by LC-MS as described above.

Oral glucose tolerance test and In vivo pharmacokinetics.

Animal care and experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Calibr at Scripps Research and strictly followed the NIH guidelines for humane treatment of animals. C57BL/6 male mice (age 10 weeks) were fasted overnight and then administered with peptides through s.c. injection at a dosage of 0.4 mpk. After 1 hour, mice were orally administered with 2 g of glucose solution per kg body weight and their tail blood glucose levels were measured before (0 min) and after glucose challenge for 90 min. Food was provided to mice 2 hour post glucose administration. Blood was extracted into heparinized tubes and centrifuged at 3,000 g for 15 min. The resulting supernatant plasma was stored at −80°C. Peptide concentration in the plasma was determined by in vitro GLP-1R activation reporter assay.

Plasma peptide concentration determination.

HEK 293 cells overexpressing GLP-1R and CRE-Luc reporter were treated with plasma samples at sequential diutions (5-point dose response, starting from 1:20 dilution of each plasma sample), incubated for 16 hours in DMEM with 10% FBS at 37°C with 5% CO2, and the firefly luciferase activity was measured as described previously. The injected peptide were used to obtain standard curves and parameters for Bottom, Top, EC₅₀, and Hill Slope. Relative luciferase unit (RLU) for each plasma sample was used to calculate the peptide concentrations in plasma (nmol/L) using parameters derived from the standard curve (RLU = Bottom + (Top-Bottom) / (1 + 10^((LogEC50-Conc.)*Hill Slope)).