Insulin Like A String of Pearls: How Degludec Works


Degludec MoleculeI’m going to be upfront about it: the news that Novo Nordisk was seeking FDA approval for Degludec, their most advanced long-acting insulin, did not grab my attention.  Don’t get me wrong—the insulin is impressive in its claims: clinical trial results show that it lasts longer than 24 hours; can be administered at any time during the day; is less likely to cause hypoglycemia, especially at night; and is non-inferior at helping patients achieve their target HbA1c goals. Those are not small potatoes, especially when you consider what a pain it can be to deal with current long-acting insulins that only barely make 24 hours, requiring that patients inflexibly inject their daily dose at exactly the same time each day.

But I wear an insulin pump, and haven’t used any non-rapid-acting insulin since NPH. And I’m frightfully self-interested. So Degludec? Cool, whatever, sure.

But there was one thing about Degludec that caught my attention, and after speaking to Dr. Alan Moses of Novo Nordisk about it, I am now of the opinion that Degludec is awesome.

What was it that caught my attention? Well, after hearing about the extremely long duration of action—on the order of 40 hours!—and the promise of extremely even release, without the peaks in activity that cause unpredictable blood sugars, I couldn’t help but ask: how does it work? What makes it different than other insulins?

The answer is so elegant, so picturesque, that my geeky self can’t help but love Degludec. Before I get to what makes it different, though, we should probably review how normal insulin works.

How Insulin Works

Under normal conditions, insulin is a gene encoded in the DNA contained in each human cell. In every cell except the beta cell, this gene is silenced, but in the beta cells of the pancreas, the insulin gene is transcribed, and turned into a string of 110 amino acids called proinsulin. This long proinsulin chain is then cut up by machinery inside the beta cell, leaving three chains of amino acids, two of which are bound together. When the middle of the three chains is cleaved out, the first and the third are left to form a single unit that we call the insulin protein—a powerful molecule, as we diabetics know, capable of binding to insulin receptors throughout the body and sending crucial metabolic signals.

The beta cell, however, does not generally make single molecules of insulin and push them immediately into the bloodstream. Instead, the beta cell makes these insulin molecules and stores them to be released only when the cell receives the proper signals that glucose levels in the body are increasing, and more insulin is needed. The single insulin molecules, called monomers, are therefore sequestered by the beta cells into vesicles, or little holding tanks, within the cell. At the concentration they are kept within the vesicles, the monomers quickly pair up to form dimers of insulin molecules. Importantly, the solution in the vesicles in which the dimers are kept also contains zinc ions; in the presence of these zinc ions, triplets of insulin dimers come together, and bind to form hexamers. These hexamers, then, contain a total of six insulin molecules, held together by zinc ions.

The insulin hexamers are stable when stored in the beta cell vesicles, protected from degradation or modification, but they are not kept as hexamers forever. The active form of insulin—the form that binds to the insulin receptor and does all the important signaling throughout the body—is the monomer. As a result, when the beta cell is ready to release insulin into the body, the vesicles containing the hexamers are fused with the plasma membrane of the cell, and the vesicle contents are poured out into circulation. The hexamers, no longer in the comfortable zinc bath of the vesicle, begin to deconstruct, and within seconds, have broken down into the monomers that are active throughout the body.

To sum up, then: insulin exists as a singular monomer, but naturally forms dimers, and comes together into hexamers in the presence of zinc. The monomers are crucial to making insulin work in the body, but insulin does not naturally hang out in its monomeric form.

These natural structural tendencies of insulin are both a complication and an advantage when we begin to consider how to make insulin analogues for use in diabetics like me, who no longer have beta cells that can so cleverly manage insulin distribution. The fact that insulin likes to form larger, inactive structures is a complication, because the dimers, tetramers, and hexamers that insulin naturally forms in solution are part of why injected insulin is so delicate and so unpredictable; if the insulin molecules all aggregate before being injected, they have to diffuse and break down into monomers before beginning to work in the body, but this happens at varying rates and with varying efficiency.

However, once we know these structural properties of insulin, we can use them to our advantage. Instead of just isolating natural insulin, we can modify the protein so that it is more likely to stay as a monomer. Rapid acting insulins like Novo Nordisk’s Novolog (insulin aspart) have a mutation in a single amino acid of the insulin chain that changes the structure of the protein. This change makes the dimers and hexamers degrade more rapidly into monomers. Thus, when Novolog is injected, it very quickly separates into monomers, and can by used by the body within minutes instead of hours.

And then there’s the other side of the equation—instead of keeping insulin as a monomer to make it faster, we can keep it as a hexamer to make it slower. This is the theory behind long-acting insulins like Novo Nordisk’s Levemir (insulin detemir), where a fatty acid is attached to one of the amino acids in the insulin chain. This causes the insulin hexamers to form even larger dihexamers, and then further to bind to a protein called albumin when injected under the skin and even when circulating in the blood. This large protein complex is then slowly broken down under the skin, resulting in a slow, gradual release of insulin monomers that act throughout the day.

Still with me? If so, I hope you appreciate that our ability to modify and manipulate insulins in this manner is pretty neat. The increasing efficiency and usability of long-acting insulins—from the caveman NPH to the more modern Levemirs—is a direct result of our ability to manipulate the molecule in new ways.

What Makes Degludec Different

And so we arrive at Degludec, which is one more step along the manipulation path. The theory behind Degludec is very similar to that behind Levemir; remove the last amino acid in the insulin chain and stick a fatty acid on instead. The fatty acid is longer, and comes with a spacer, but to a non-chemist, that’s six of one, half a dozen of the other, right?

Except that on a molecular level, these minor differences create very different structural tendencies. So Degludec, like Levemir, exists as a dihexamer in solution—it’s kept with zinc and a chemical called phenol, which helps to maintain loose hexamers that pair up. Upon injection, however, Degludec doesn’t just bind to another protein like albumin, and doesn’t stay a dihexamer. The phenol dissipates, and the dihexamers begin to associate with each other inside the body.

And this is where we move from neat to incredible: according to studies Novo Nordisk has done to observe the structure of Degludec after injection, the dihexamers begin to stack up to form multihexamers. But we’re not just talking four or even ten hexamers together. The hexamers stack up by the thousands, forming long chains. These chains are the width of a single hexamer, but the length of a thousand—“like a string of pearls,” Dr. Moses notes.

Can you see it? The chemical magic of it all? Minor changes that the human eye cannot see, and all of the sudden we have insulin strung together like pearls under the skin. These long, stable structures diffuse throughout the body, ever so slowly losing their zinc ions and breaking down into smaller structures, resulting in a slow release of active insulin monomers that takes some 40 hours. And these chains are so regular and so tight that the release is remarkably even, without any significant peaks that would cause unpredictable blood sugars.

In other words: atoms and subatomic particles bounce around, by the pull of electrical charge forming molecules that we modify to bend slightly differently, and they circulate silently under the skin, an ordered procession, releasing insulin so regularly that – for instance – a man can decide on Saturday morning to stay in bed an extra hour, put off taking his insulin, so that he can watch his wife breathe and sleep and dream for that hour, forgetting for those precious moments that he has diabetes at all.

Molecular poetry. I don’t care if I never use it—Degludec is awesome.

For more information about the structure of insulin, I suggest:

Dunn MF. Zinc-ligand interactions modulate assembly and stability of the insulin hexamer — a review. Biometals. 2005 Aug;18(4):295-303.

Pittman I, Philipson L, Steiner D. Chapter 1: Insulin Biosynthesis, Secretion, Structure, and Structure-Activity Relationships.

Owens DR. Insulin preparations with prolonged effect. Diabetes Technol Ther. 2011 Jun;13 Suppl 1:S5-14.

Leave a Reply

4 Comment threads
0 Thread replies
Most reacted comment
Hottest comment thread
4 Comment authors
BeccaGeraldinekaterinasScott Strumello Recent comment authors
newest oldest most voted
Notify of

I agree with Geraldine plus if someone not fully understands chemistry the breakthrough with Degludec, doesn’t really know what this article is about. I’m also a type 1 and 40 hours in comparison to 24 hours or so, makes a major difference. I guess research is finally working for a betterment for type 1 diabetics. With Lantus, one has to remain on the treadmill to avoid the weight gain. I do like to workout anyhow.  Glory . . . .


I think you have missed the point.  I have type 1 and being able to sleep late or stay up later without an injection is huge.  Insulin does not control my schedule.  These companies are the ones funding diabetes research.  Just look and see how much Novo Nordisk has spent. Also, less hypo and if anything like Levemir, will have less weight gain.  I know I gained 20lbs on Lantus.


Sorry not impressed! I understand that this can be a prized concept for a scientist, but it really does’t offer anything new to my kid or me. One less injection per day isn’t impressive and I haven’t seen any significunt/ not able to deal with peaks on levemir or lantus.  They have to come up with something that really make a difference. So I am glad that the patent of Lantus expires and we could have cheaper insulins doing almost exactly the same thing, and hopefully these companies will loose money and maybe then they will be motivated to do something that will actually change our lives… Read more »

Scott Strumello

I’m not naive; basal insulin is a big business (although not nearly as big as prandial), and frankly, Novo failed abysmally with Levemir, whose market share trails far behind Lantus.  Having tried Levemir, I can say that its the perfect 12 hour basal insulin, but it’s nowhere near the 24 hours it claims to be (which is good, whoever said everyone needs 24 hours is obviously not speaking from personal experience) The only challenge will be whether this gee-whiz chemistry is sufficient to overcome a new reality: the complex healthcare bill also authorized follow-on (or biosimilars), and the last patent… Read more »

Copyright © 2009-2018 Diabetes Media Foundation, All Rights Reserved.
ASweetLife™ is a trademark of the Diabetes Media Foundation, All Rights Reserved.