NNC0363-1063: first-in-human glucose-sensitive insulin in type 1 diabetes
Novo Nordisk's glucose-sensitive insulin (GSI) candidate, tested in healthy participants and in people with type 1 diabetes. Three parts: single ascending dose, a proof-of-principle part, and a multiple-dose meal test. This is the study that took the glucose-responsive-insulin class from animals into people with T1D. It completed in March 2026 and no results have been posted, so no human glucose-response curve is public.
Primary endpoints
- Part 1 (single ascending dose): number of adverse events
- Part 2 (proof of principle): apparent serum clearance of NNC0363-1063 after a single dose
- Part 3 (meal test): area under the plasma glucose concentration-time curve at steady state
Results so far
None posted. The study completed on 18 March 2026 with 117 participants enrolled; ClinicalTrials.gov shows no results and Novo Nordisk has not published a human glucose-response curve for this molecule.
The full picture
This is the study that makes "smart insulin has reached humans" a true statement rather than an aspiration — and it is also a good illustration of how little that statement currently buys.
What was tested. NNC0363-1063 is Novo Nordisk's clinical glucose-sensitive insulin. It is not NNC2215, the molecular switch published in Nature in October 2024 — that compound remains animal-only. The two are frequently conflated because both are Novo's and both are "smart insulin"; they are different molecules at very different stages.
Design. Three parts, escalating in ambition: a single-ascending-dose part in healthy adults and then in adults with T1D; a proof-of-principle part measuring how quickly the drug clears; and a multiple-dose meal-test part measuring the glucose curve after eating. Comparators were placebo, insulin degludec and insulin detemir. A meal test is the first design here that could in principle show glucose-responsiveness doing something useful.
What we do not know. Everything that matters. The study completed in March 2026 and no results are posted. There is no published human glucose-response curve, no potency figure against human insulin, and no outpatient data. That gap is precisely the one that killed MK-2640: Merck's candidate looked excellent in dogs and minipigs and then failed to demonstrate glucose-dependent clearance in people with T1D.
Until a response curve is published, the honest summary is that the class has entered humans and has not yet been shown to work in them.