Albumin binding half-life extension (revision 7)

Albumin binding half-life extension is a peptide-engineering strategy in which a fatty acid or fatty diacid is covalently attached to a peptide so that the conjugate binds reversibly to circulating serum albumin. The bound fraction is protected from renal filtration and from proteolysis, and acts as a depot from which free peptide is released continuously.^[1]

The strategy is the basis of the long dosing intervals achieved by liraglutide and semaglutide, and it is used in insulin analogues and in several unrelated peptide drugs. Its principal advantage over covalent fusion to a large carrier is reversibility: the free and bound pools remain in equilibrium, so the peptide retains full receptor activity when released, and the peak-to-trough ratio flattens without the pharmacophore being encumbered.^[2]

Design involves three choices — where to attach, what spacer to use, and which fatty acid. All three are constrained by the requirement that the modification lie distal to the receptor-binding region of the peptide, which for the incretin analogues means the C-terminal half.^[1]

Human serum albumin circulates at roughly 600 μM and carries several hydrophobic binding sites, principally the Sudlow sites, which in physiology transport long-chain fatty acids, bilirubin and a wide range of drugs. A peptide bearing a fatty-acid tail occupies one of these sites with an affinity that can be tuned across several orders of magnitude by chain length and by the presence of a terminal carboxylate.^[2]

Albumin itself has a plasma half-life of about nineteen days, maintained by FcRn-mediated recycling that rescues it from lysosomal degradation. A peptide that spends most of its time bound inherits part of this protection: it is too large as a complex to be filtered at the glomerulus, and its proteolytic exposure is reduced.

The equilibrium is what makes the approach work. If binding were irreversible the peptide would never reach its receptor; if it were too weak the depot effect would be negligible. Reported bound fractions for the marketed incretin analogues exceed 99%, meaning free drug is under 1% of total at any moment and the receptor sees a small, steady concentration rather than a peak.^[1]