GLP-1 (7-37)

The native 31-amino-acid active form of human glucagon-like peptide-1, cleaved from proglucagon in intestinal L cells, whose one-to-two-minute half-life under DPP-4 is the reason every marketed GLP-1 drug is an analog rather than the hormone itself.

Also referenced as: GLP-1(7-37), GLP-1 7-37, Glucagon-like peptide-1 (7-37), Glucagon-like peptide I (7-37), Insulinotropin, Glycine-extended GLP-1, Proglucagon (78-108)

Also appears in: Hormone · Metabolic

COAPublic product evidenceSearch the public certificate ledger for this compoundNo exact compound records are currently indexed under this profile name.→

---

What is GLP-1 (7-37)?

GLP-1 (7-37) is the native, unmodified human incretin hormone: a 31-amino-acid peptide (HAEGTFTSDVSSYLEGQAAKEFIAWLVKGRG, corresponding to proglucagon residues 78–108) liberated by prohormone convertase processing of proglucagon in intestinal L cells. It is one of two naturally occurring bioactive forms of GLP-1. The other, GLP-1 (7-36) amide, is amidated at the C-terminus and is the dominant circulating species in humans; GLP-1 (7-37) is the glycine-extended minor form, and in the one human head-to-head comparison the two were indistinguishable on every endpoint measured.

Tissue matters here, and it is routinely stated wrong. The active truncated forms are an intestinal product. Proglucagon is processed differently in the two tissues that express it: in gel permeation chromatography of human small intestine, all GLP-1 immunoreactivity eluted as the two insulin-releasing peptides, whereas in human pancreas 71% eluted as the major proglucagon fragment, 24% as GLP-1 (1-36) amide and 5% as GLP-1 (1-37) — that is, essentially none of it as the active hormone (Orskov et al., Diabetes, 1994;43(4):535–539). The pancreatic alpha cell makes glucagon and leaves GLP-1 in its N-terminally extended, inactive state; the L cell makes the hormone this page is about.

The numbering is a frequent source of confusion and is worth understanding before reading anything else about this molecule. Proglucagon processing first yields GLP-1 (1-37), which is biologically inert; removing the first six residues produces the active hormone, which by convention keeps the original numbering and therefore begins at His7. The residue positions cited throughout the GLP-1 drug literature — Ala8, Lys26, Lys34 — all refer to this scheme.

It appears in the research-peptide market for two reasons, and only one of them is about the molecule. The legitimate reason is that it is the parent compound and reference standard for an entire drug class, so it is stocked as an analytical and pharmacological control. The other is that “GLP-1” is now one of the most recognizable names in medicine, and listings carrying that name frequently contain an analog rather than the native hormone.

How it works

• Glucose-dependent insulin secretion through the GLP-1 receptor. In the isolated perfused rat pancreas, GLP-1 (7-37) stimulated insulin release at concentrations as low as 5 × 10⁻¹¹ M, producing 3- to 10-fold increases over basal in the presence of 6.6 mM glucose, while the unprocessed GLP-1 (1-37) had no effect even at 5 × 10⁻⁷ M — a separation of at least 10,000-fold that established the N-terminal truncation as the activating step (Mojsov et al., Journal of Clinical Investigation, 1987;79(2):616–619). Receptor engagement raises intracellular cAMP, and the insulinotropic effect is conditional on ambient glucose, which is why native GLP-1 does not drive hypoglycemia the way insulin does.
• Incretin action that survives type 2 diabetes. Under hyperglycemic clamp in nine patients with type 2 diabetes and nine matched controls, GLP-1 (7-36 amide) retained most of its insulinotropic effect — diabetic patients reached 71% of the C-peptide increments seen in normal subjects, a non-significant difference — while the maximum effect of the other major incretin, GIP, was 54% lower in the diabetic group. That contrast is the observation that redirected the field toward GLP-1 rather than GIP as a therapeutic target (Nauck et al., Journal of Clinical Investigation, 1993;91(1):301–307).
• Glucagon suppression. In the same clamp study, GLP-1 (7-36 amide) lowered glucagon concentrations in both diabetic and normal subjects (p < 0.05), an effect GIP did not produce (Nauck et al., 1993, above). Glucagon fell during physiological GLP-1 infusion in healthy volunteers as well (Kreymann et al., Lancet, 1987;2(8571):1300–1304).
• Delayed gastric emptying and reduced appetite. Gastric emptying was inhibited during six weeks of continuous subcutaneous GLP-1 (Zander et al., 2002, below). A separate 48-hour continuous subcutaneous infusion in six patients with type 2 diabetes lowered fasting and 24-hour mean plasma glucose, decreased hunger and prospective food intake, and increased satiety — though glucagon was unchanged in that particular protocol (Toft-Nielsen et al., Diabetes Care, 1999;22(7):1137–1143). These are the same effect arms that the modern analogs are prescribed for.
• Immediate N-terminal cleavage by DPP-4 — the defining property. Dipeptidyl peptidase-4 removes the first two residues (His-Ala) to yield GLP-1 (9-37) or GLP-1 (9-36) amide, which are effectively devoid of insulinotropic activity at physiological concentrations and behave as weak receptor antagonists in some systems. Thirty minutes after a subcutaneous dose of GLP-1 (7-36 amide) in diabetic patients, the truncated metabolite accounted for 88.5 ± 1.9% of the increase in plasma immunoreactivity; during intravenous infusion, intact GLP-1 represented only about 20–25% of the increase in total circulating immunoreactivity (Deacon et al., Diabetes, 1995;44(9):1126–1131). Neutral endopeptidase 24.11 is a second route, and may account for up to half of the GLP-1 entering the circulation (Plamboeck et al., Diabetologia, 2005;48(9):1882–1890); the kidney contributes substantial additional extraction (Deacon et al., 1996, below).
• The metabolite is not simply inert. GLP-1 (9-36), long assumed to be a dead end, produced cardioprotective and vasodilatory effects in isolated hearts and arteries from mice lacking the GLP-1 receptor entirely, through a nitric oxide synthase–requiring, receptor-independent mechanism (Ban et al., Circulation, 2008;117(18):2340–2350). This complicates any attempt to read native-GLP-1 results as pure receptor pharmacology.

Research status

Native GLP-1 has a large and high-quality human research record, but it is a physiology record, not a drug-development one. No regulatory authority anywhere has approved GLP-1 (7-37), and no Phase 3 program has ever been registered for it as a therapeutic.

Discovery and characterization. The work showing that preproglucagon is processed differently in pancreas and intestine — the finding that made a distinct intestinal GLP-1 product possible — appeared in 1986 (Mojsov et al., Journal of Biological Chemistry, 1986;261(25):11880–11889), and the insulinotropic activity of GLP-1 (7-37) was demonstrated the following year (Mojsov et al., 1987, above). Two other groups reported the same physiology in the same year, though for the amidated form rather than 7-37: Holst and colleagues isolated proglucagon 78-107 — the 7-36 fragment — from pig intestinal mucosa and showed it stimulated insulin secretion from the perfused pig pancreas (FEBS Letters, 1987;211(2):169–174), and Kreymann and colleagues infused GLP-1 (7-36) amide into seven human volunteers at postprandial concentrations, establishing it as a physiological incretin more powerful than GIP (Lancet, 1987;2(8571):1300–1304).

Where 7-37 sits relative to 7-36 amide. Gel permeation chromatography of human small intestine found 80% of GLP-1 immunoreactivity as GLP-1 (7-36) amide and 20% as GLP-1 (7-37). In fasting plasma the two are comparable (7 ± 1 and 6 ± 1 pM), but after a breakfast meal the amidated form rose to 41 ± 5 pM while the glycine-extended form rose only to 10 ± 1 pM — so 7-37 is specifically the minor postprandial species (Orskov et al., 1994, above). The direct head-to-head comparison came later: eight type 2 diabetic patients received four-hour intravenous infusions of each form in randomized order, and both normalized fasting hyperglycemia with no significant difference on any endpoint — glucose, insulin, and C-peptide all returned p = 0.99 (Nauck et al., Diabetic Medicine, 1998;15(11):937–945). Eight patients cannot establish equivalence to a regulatory standard, but nothing in that comparison distinguished the two forms, and most human “native GLP-1” studies used 7-36 amide.

Pharmacokinetics — the finding that ended the program. In anesthetized pigs, the plasma half-life of biologically intact GLP-1 measured by an ELISA specific for the active peptide was 1.5 ± 0.4 minutes, against 4.1–4.5 minutes for total immunoreactivity measured by assays that cannot distinguish the inactive metabolite (Deacon et al., American Journal of Physiology, 1996;271(3 Pt 1):E458–E464). In humans, clearance of intact GLP-1 runs roughly 4–9 liters per minute and is no different in obese type 2 diabetes than in matched healthy subjects, meaning the low endogenous GLP-1 levels seen in diabetes reflect impaired secretion rather than faster destruction (Vilsbøll et al., Journal of Clinical Endocrinology & Metabolism, 2003;88(1):220–224). A peptide with a 1–2 minute half-life cannot be a once-daily or once-weekly drug, and that single fact is the origin of every analog in the class.

Continuous infusion works, which is the point. Because bolus dosing is futile, the therapeutic proof-of-concept had to be run as an infusion. In a six-week parallel-group pilot, 20 type 2 diabetic patients were assigned — alternately rather than randomly, a real limitation — to continuous subcutaneous GLP-1 or saline, and analyzed per protocol as 10 versus 9. In the GLP-1 group HbA1c fell 1.3% (p = 0.003), fasting and 8-hour mean plasma glucose fell 4.3 and 5.5 mmol/L (p < 0.0001), gastric emptying was inhibited, body weight decreased 1.9 kg, appetite was reduced, and both insulin sensitivity and beta-cell function improved (Zander et al., Lancet, 2002;359(9309):824–830). This result validated the target, not the molecule — the delivery requirement made it commercially unworkable, and development moved to DPP-4-resistant analogs and to DPP-4 inhibitors instead.

Cardiovascular work, and where it stalled. Small studies infusing native GLP-1 reported striking cardiac results in patients with badly impaired ventricles: left ventricular ejection fraction rose from 29 ± 2% to 39 ± 2% over a 72-hour infusion in 10 patients after acute myocardial infarction with severe systolic dysfunction following successful angioplasty (Nikolaidis et al., Circulation, 2004;109(8):962–965), and from 21 ± 3% to 27 ± 3% over a five-week infusion in 12 patients with class III/IV heart failure (Sokos et al., Journal of Cardiac Failure, 2006;12(9):694–699). Both compared against concurrent control groups (11 and 9 patients respectively) rather than randomized, blinded comparators, and both were single-center studies of roughly a dozen patients. When the same group ran a randomized study — 20 patients receiving GLP-1 infusion or open-label standard therapy around coronary artery bypass grafting — there were no differences in preoperative, postoperative, or 7-day ejection fraction (61 ± 4% vs. 59 ± 3%) or in cardiac index at 18 hours; the benefits that survived were better glycemic control, fewer arrhythmias requiring treatment, and reduced inotrope and insulin requirements (Sokos et al., American Journal of Cardiology, 2007;100(5):824–829). That null result is worth reading carefully rather than as a refutation: those patients were enrolled with preserved left ventricular function and a baseline ejection fraction near 60%, leaving no room for the improvement measured in the earlier cohorts at 21–29%. It shows the earlier finding was never tested head-on, not that it failed. A randomized, triple-masked, placebo-controlled Phase 2 study of GLP-1 (7-36 amide) on myocardial function after CABG (NCT00966654, Johns Hopkins) was terminated after 12 participants, with the registry recording the reason as the principal investigator leaving the institution.

The living legacy is in the analogs. The human-sequence GLP-1 drugs are deliberate engineering exercises against this backbone, and their labels describe them in exactly those terms. Liraglutide is labeled as 97% homologous to native human GLP-1 (7-37), substituting arginine for lysine at position 34 and attaching a C-16 palmitic acid with a glutamic acid spacer at the remaining Lys26. Semaglutide’s label describes modification at position 26 with a hydrophilic spacer and a C18 fatty di-acid, stabilization at position 8 against DPP-4 degradation, and a minor change at position 34 — two substitutions against the native 31-mer, which is where the commonly quoted 94% homology figure comes from, though the current US label does not itself state a percentage. Dulaglutide’s GLP-1 portion is labeled 90% homologous to native human GLP-1 (7-37) and is fused to the Fc portion of a modified human IgG4 heavy chain. Each of those modifications exists to defeat the enzymatic vulnerability described above. Exenatide took the other available route, deriving instead from exendin-4, a 39-residue peptide originally identified in the lizard Heloderma suspectum that shares only about half its sequence with human GLP-1 but carries glycine rather than alanine at the second position and is therefore naturally DPP-4 resistant.

Common dosage forms

• Lyophilized powder in vials, typically at small milligram scale, sold as a research-use-only reagent for reconstitution.
• Analytical and pharmacological reference standards from life-science catalogs, supplied in microgram-to-milligram quantities as lyophilized solid or in solution — the format that matches the compound’s actual role as the parent-molecule control in receptor and assay work.
• Salt forms vary between suppliers; synthetic peptides of this type are commonly supplied as trifluoroacetate or acetate salts, which affects the net peptide content per stated milligram.
• No format corresponds to the human research protocols. Every human study of native GLP-1 delivered it as a continuous intravenous or subcutaneous infusion measured in pmol/kg/min, using pumps, because of the half-life. A single-dose vial has no counterpart in the clinical literature.

This section describes formats only and is not dosing guidance.

Key considerations

• “GLP-1” on a product listing is usually not this peptide. Research-market listings labeled “GLP-1,” and abbreviated variants that append a letter for the analog they actually contain, most often hold semaglutide, tirzepatide, or another engineered agonist rather than the native hormone. The class name has become a category label, so the specific designation — the parenthetical fragment numbering — is the only reliable identifier. Confirming which molecule a certificate of analysis actually characterizes matters more here than for almost any other compound.
• At least five distinct molecules share the “GLP-1” name. GLP-1 (1-37) and GLP-1 (1-36) amide are the inactive N-terminally extended precursors; GLP-1 (7-37) and GLP-1 (7-36) amide are the two active forms; GLP-1 (9-36) amide and GLP-1 (9-37) are the DPP-4 cleavage products. Only the middle pair activates the receptor, and the gap between the inert precursor and the active hormone is at least four orders of magnitude in potency. A citation, a purity claim, or a mass-spectrometry result attached to a bare “GLP-1” label does not specify which of these was measured — and a 31-mer and a 37-mer are trivially distinguishable by mass if anyone bothers to look.
• The pharmacokinetics are the entire reason the analogs exist. With a 1–2 minute half-life for the intact peptide and the great majority of an administered dose converted to inactive metabolite within half an hour, native GLP-1 is not a substitute for a GLP-1 receptor agonist and does not reproduce their exposure profile. Framing that presents the native hormone as a “natural” version of an approved drug inverts the actual pharmacology.
• No approval, and no pivotal trial. GLP-1 (7-37) is not approved by FDA or any other regulator for any indication. The human evidence base consists of physiology studies and small proof-of-concept infusions — the six-week Lancet study enrolled 20 patients, the cardiac studies roughly a dozen each — and the single randomized cardiac-function comparison was run in patients whose ejection fraction was already normal, so it neither confirmed nor refuted the earlier signal. FDA has separately warned about unapproved GLP-1 products, including unapproved salt forms and products marketed outside the approved supply chain.
• Tolerability data come from short, titrated infusions, not from open-ended use. The six-week infusion study reported no important side effects and the 48-hour study recorded no gastrointestinal side effects at all (n = 6; one brief cutaneous reaction at the infusion site), but the cardiac infusion studies did report transient gastrointestinal effects and, in the heart-failure cohort, occasional hypoglycemia. Nausea, vomiting, and delayed gastric emptying are established dose-related effects across the GLP-1 class, and native GLP-1 measurably inhibits gastric emptying at therapeutic exposures. There is no long-term safety dataset for this peptide because no long-term human program was ever run.