A team at the Institute of Organic Chemistry and Biochemistry in Prague (Selicharová, Jiráček) and the Walter and Eliza Hall Institute in Melbourne (Kirk) published a new insulin analog in Science Advances on Friday ↗ that breaks a usually rigid pharmacology rule. The hormone insulin and the growth factor IGF-1 evolved from a common ancestor and bind different receptors. Insulin binds the insulin receptor in two splice variants, IR-A and IR-B ↗. IGF-1 binds the IGF-1 receptor ↗. The two systems control different things: glucose handling versus growth and tissue repair. Pharma keeps them separate on purpose, because a drug that drifts between them can promote cancer-cell proliferation or mismatch the dose. The new analog, called 1(Ins), changes that picture by hitting all three receptors well after a four-residue edit to the insulin B chain.
The four substitutions are GluB10, D-HisB24, GlyB31, and TyrB32. In plain terms, two natural residues are swapped for chemically related ones, one is replaced with the mirror-image form of histidine (a D-amino acid that proteases struggle to cut), and the B chain is extended by two residues. Native insulin is 51 amino acids in two disulfide-linked chains. The B chain is 30 amino acids; the new analog adds two on the end. Nothing about the change is exotic. The outcome is what is unusual.
1(Ins) binds IGF-1R about 1,000 times more strongly than native insulin. In plain language, native insulin barely touches IGF-1R; the analog binds it almost as well as IGF-1 itself, which is the receptor's natural ligand. It also keeps high affinity for both splice variants of the insulin receptor. The paper uses cryo-electron microscopy ↗ to show how those four substitutions reposition the ligand inside both receptor sites. The hexameric storage form and the dimerization behaviour of insulin are preserved enough for the molecule to behave as a stable drug-like ligand, while the B-chain edits expose contacts the IGF-1 receptor wants to see. Most "promiscuous" insulin analogs designed in the past have been side effects of formulation chemistry. This one is by design.
What it does in cells
The team applied 1(Ins) to neuronal cultures and read out both pathways. The insulin receptor side (Akt-dominant, glucose handling, neuronal survival) and the IGF-1R side (longer-lasting MAPK and survival signaling) both activated robustly. In a parallel set of neuroprotection assays, 1(Ins) outperformed native insulin and matched or exceeded IGF-1, which is the cleaner benchmark in those models. Phosphoproteomic profiling, which reads out the full set of intracellular proteins that flip on after ligand binding, confirmed dual-pathway activation. It also flagged phosphorylation events that neither insulin nor IGF-1 produces on its own. The signaling fingerprint is partly new.
That matters because the field's working assumption was that a dual agonist would look like a weighted sum of its parents. This one writes its own tail.
In vivo work in mice and rats showed effective glucose regulation, consistent with full insulin-receptor agonism. The paper does not include a tumor-relevant assessment, which is the standard worry with any compound that lights up IGF-1R. IGF-1R signaling is implicated in cancer-cell proliferation, and any clinical translation of 1(Ins) will have to clear that hurdle. The authors frame the molecule as a tool first, not a candidate. Metabolic control plus neuroprotection ↗ is the deliberate target profile (brain insulin signaling failure is one mechanistic theory of Alzheimer's, and IGF-1 has been pursued in ALS for two decades). It could equally serve as a research handle to ask what dual IR-plus-IGF-1R pharmacology actually does in tissues that express both. The phosphoproteomic data already says the answer is not "the average of the two."
What this changes for rational design
Dual agonism at related receptors is usually a long campaign. Tirzepatide took years of medicinal chemistry to balance GLP-1 and GIP binding. Retatrutide added glucagon for a third receptor and consumed similar effort. 1(Ins) accomplishes the analogous trick on the insulin and IGF-1 receptor families with a four-residue edit on a clinically familiar scaffold. The B31 and B32 extension is the kind of small change a high-school molecular biology class draws as a cartoon, and it produces a 1,000-fold affinity shift. That is information about the receptors themselves: how much of their selectivity depends on a couple of contacts at the chain terminus, and how much room rational design has on this scaffold.
What is missing. The paper does not report a head-to-head against IGF-1 in a cancer-cell proliferation assay. It does not report half-life or biodistribution beyond acute pharmacology. There is no mention of compatibility with the hexamer-stabilizing zinc and phenol chemistry that injectable insulins rely on. Each of those is what a follow-up paper or a company licensing the scaffold would need to answer. For now, the result is a reagent-grade engineered insulin that pharma chemists should read closely, because it constrains the assumption that selectivity at the insulin-receptor family is hard to undo.