Part 2 of 3 | Improved Pharma peptide crystallization series by Stephen R. Byrn. Ph.D.
Insulin and GLP-1 receptor agonists belong in the same conversation about peptide medicines, but their crystallization histories are very different. Insulin has a long record of characterized crystal structures and pharmaceutical use of crystals. For widely used GLP-1 peptides, public evidence about crystals of the isolated, final drug molecule is more limited and needs to be evaluated compound by compound. This difference makes insulin a valuable starting point, while keeping GLP-1 crystallization an active research opportunity.
Insulin crystals as pharmaceutical materials
The organization of insulin into zinc coordinated hexamers helped make crystalline formulations possible. Crystal form, particle size, and formulation environment could influence how quickly material dissolved after subcutaneous injection. Historically, prolonged action insulin preparations showed that a peptide crystal could be an engineered drug delivery material, not just a convenient analytical specimen [1].
Insulin also teaches diversity. There is no single “insulin crystal.” Human insulin has been crystallized in different crystal systems, and laboratory studies have identified forms that change with pH and other components of the crystallization medium. In one study, changing conditions involving urea or thiocyanate and pH yielded orthorhombic and monoclinic human insulin crystal forms with different hexamer packing [2]. These forms are scientifically informative; their existence alone does not mean they were used in a marketed formulation.
Small sequence changes also matter. Human, porcine, and bovine insulins differ at only a few amino acid positions, and their formulations need to be assessed individually. Human insulin is more soluble than either bovine or porcine insulin. Figure 1 illustrates differences between the diffraction patterns of bovine, porcine, and human insulin obtained using synchrotron diffraction. Such examples show why high-quality solid state analytical methods are essential for distinguishing closely related crystalline peptide forms and for supporting pharmaceutical development, quality control, and intellectual-property claims. Interestingly, despite the differences in solubility, a study showed that human ultralente insulin could match the action of bovine ultralente insulin in a human clinical study for the measures examined [3]. These investigations illustrate the importance of understanding the sequence, lattice, crystal surface, and formulation of peptides.
What changes with GLP-1 peptides?
GLP-1 based medicines are chemically diverse. A change in sequence, lipid attachment, molecular size, or excipient environment can substantially change self-association and solubility. Liraglutide, semaglutide, exenatide, and tirzepatide should therefore be treated as separate solid form questions. Finding a crystal of a precursor, a receptor complex, or a different derivative would not establish that the intended drug substance crystallizes under useful manufacturing conditions.
The marketed presentation poses a second distinction. A drug sold as a solution need not contain crystalline peptide in the final container. Crystallization could nevertheless be useful upstream for purification, isolation, stability, or a future dosage form. Conversely, labeling a recovered solid “crystalline” in a process description does not identify its polymorph, hydrate state, or unit cell. Diffraction and composition measurements are needed to support those claims.
The liraglutide patent illustrates the evidence question
A Chinese manufacturing patent, CN113621046A, describes on column modification of the Arg34-GLP-1 precursor, followed by purification and an isolation step that the applicants call crystallization of liraglutide. The modification attaches the lipid containing group required for liraglutide [4].
The patent describes a crystallization medium containing a basic amino acid such as arginine, phenol, and polyethylene glycol, followed by pH adjustment and cooling. It reports favorable stability results for the material it calls crystalline. Those claims make the process interesting. The patent text we reviewed, however, does not provide an XRPD fingerprint or a solved crystal structure for the final isolated material.
A research program that learns from both cases
For a GLP-1 peptide, the first task is to establish which molecular species is actually being screened and how its identity and purity are measured. A small screen can then vary pH, counterion, solvent, water activity, concentration, temperature, and seeding. Micrographs help find candidate particles; XRPD distinguishes crystals from amorphous precipitate; chromatography and mass spectrometry verify the molecule. A promising form should be assessed for water content, stability, redissolution, and process reproducibility before claims about its pharmaceutical value are made.
Insulin shows what a mature crystalline peptide program can achieve. GLP-1 peptides ask a more exploratory question: which exact molecules and conditions produce well characterized, useful crystals? Answering that question requires careful experiments rather than an assumption that all peptide medicines will behave like insulin. The next blog considers other peptides for which the evidence ranges from published crystal structures to detailed solid form patents.
Selected references
[1] Owens DR. Insulin preparations with prolonged effect. Diabetes Technol Ther. 2011;13 Suppl 1:S5-S14. https://doi.org/10.1089/dia.2011.0068
[2] Norrman M et al. Crystallographic characterization of two novel crystal forms of human insulin induced by chaotropic agents and a shift in pH. BMC Struct Biol. 2007;7:83. https://pubmed.ncbi.nlm.nih.gov/18093308/
[3] Holman RR et al. Human ultralente insulin. BMJ. 1984;288:665-668.
[4] CN113621046A. Use and method of polymer anion exchange filler in preparation of liraglutide. https://patents.google.com/patent/CN113621046A/en