Dipeptidyl peptidase-4 (revision 23)

Dipeptidyl peptidase-4 (DPP-4), also known as CD26, is a serine exopeptidase that removes the N-terminal two residues from peptides presenting proline or alanine in the second position. Both incretin hormones — Glucagon-like peptide-1 and Glucose-dependent insulinotropic polypeptide — carry alanine at position 2 and are therefore inactivated within minutes of secretion.^[1]

The enzyme exists as a type II membrane protein on endothelial cells, epithelial cells and lymphocytes, and as a catalytically active soluble form in plasma. Cleavage of GLP-1 begins in the capillaries of the intestinal lamina propria, before the hormone reaches the portal vein, so that only a minority of secreted GLP-1 arrives at the systemic circulation intact.^[2]

DPP-4 defines the design problem for the entire GLP-1 receptor agonist class, and also constitutes a drug target in its own right. Inhibitors of the enzyme raise endogenous incretin concentrations two- to three-fold and lower glycated haemoglobin by roughly 0.5–0.8 percentage points, an effect substantially smaller than that of receptor agonists because the ceiling is set by physiological secretion.^[1]

DPP-4 cleaves the dipeptide His-Ala from the N-terminus of GLP-1(7–36)amide to yield GLP-1(9–36)amide. The truncated product binds the GLP-1 receptor with roughly two orders of magnitude lower affinity and behaves as a weak antagonist at pharmacological concentrations; whether it has independent cardiovascular activity remains an open question in the literature.^[2]

The substrate range is wide. Substance P, neuropeptide Y, several chemokines and the incretins are all cleaved, and the enzyme also has non-catalytic functions as a co-stimulatory molecule on T cells and as a binding partner for adenosine deaminase. The breadth of this range is the reason inhibitor development required careful selectivity against the related enzymes DPP-8 and DPP-9.^[1]

Kinetically the enzyme is efficient rather than abundant: plasma DPP-4 activity is sufficient to halve circulating intact GLP-1 in about one minute, which is why the intact fraction must be measured with an assay that distinguishes GLP-1(7–36) from GLP-1(9–36) if the measurement is to mean anything.

Every therapeutic peptide agonist in this class solves the DPP-4 problem, and the solutions are few.

Substitution alone is not sufficient for a long dosing interval — a DPP-4-resistant peptide is still cleared renally within hours — so the resistant residue is almost always combined with an albumin-binding or fusion strategy. Semaglutide combines both, and adds a substitution at position 34 to prevent acylation at the wrong lysine.^[3]

Sitagliptin, vildagliptin, saxagliptin, linagliptin and alogliptin inhibit the enzyme competitively or, in some cases, through a covalent but reversible interaction with the catalytic serine. They raise intact GLP-1 and GIP concentrations into the low postprandial physiological range rather than the supraphysiological range achieved by injected agonists.^[1]

The clinical consequences follow from that ceiling. Glycated-haemoglobin reduction is modest, weight is unchanged rather than reduced, and gastrointestinal adverse effects are uncommon — the same mechanism that limits efficacy also limits toxicity. Cardiovascular outcome trials of the class have been neutral for major adverse cardiovascular events, with a signal for heart-failure hospitalisation reported for one member.^[1]

Because the inhibitors act on endogenous hormone, they are not interchangeable with receptor agonists and combining the two produces little additional benefit — receptor occupancy is already near-maximal under agonist therapy.

DPP-4 matters to anyone measuring incretin concentrations. Blood drawn without a DPP-4 inhibitor in the tube continues to degrade GLP-1 ex vivo, and comparisons between studies that did and did not add an inhibitor at collection are not meaningful. Total-GLP-1 assays that detect both intact and truncated forms give systematically higher and less interpretable numbers than intact-specific assays.^[2]

The same consideration applies to stability work on the therapeutic peptides themselves. A DPP-4-resistant analogue is resistant to that specific enzyme, not to proteolysis in general, and stability in plasma is not evidence of stability in a reconstituted vial, where deamidation and aggregation rather than proteolysis are the degradation routes that matter.^[4]

• Glucagon-like peptide-1
• Glucose-dependent insulinotropic polypeptide
• GLP-1 receptor agonist
• Albumin binding half-life extension
• Enteroendocrine L cell