The most famous of diabetes hormones may be insulin, but at the American Diabetes Association’s Scientific Sessions this year in Chicago, the hot topic was a different hormone: glucagon. From the talks to the exhibit hall to the after-parties, the buzz went glucagon glucagon glucagon.
Why? Isn’t glucagon just a powder in a kit used to rescue diabetics passed out from hypoglycemia? Up until recently, glucagon was only discussed as a rarely used rescue treatment for diabetics. However, glucagon is now gaining attention as the yin to insulin’s yang, and a hormone that could be beneficial for daily glucose management, and that may be necessary for a closed-loop artificial pancreas system.
What is Glucagon?
But before we jump to the question of why glucagon, we should start with what glucagon is . Like insulin, glucagon is a hormone produced by the cells of the pancreas. Insulin is produced by the beta cells, and glucagon is produced by the nearby alpha cells. Whereas insulin’s net effect is to allow cells of the body to take glucose out of the blood, lowering blood glucose levels, glucagon has the opposite effect. Glucagon’s main site of activity is the liver, as it induces the liver to convert stored glycogen into glucose and release the glucose into the bloodstream, thereby increasing blood glucose levels.
In a normal, non-diabetic person, increased sugar intake causes the beta cells to release insulin. The blood vessels supplying the beta cells carry that insulin as they pass the alpha cells. The alpha cells interpret signals passed from the upstream beta cells, and repress their glucagon production while the beta cells are working to lower blood sugar levels. However, as blood sugar levels drop, and the beta cells stop releasing insulin, the alpha cells are free to produce glucagon. The same blood vessels carry that glucagon directly into the portal vein and into the liver, where glucose is eventually released. This cycle allows the body to maintain blood glucose levels within a tight range — insulin drives levels down, and glucagon keeps them up, alternating activity to maintain consistent blood glucose levels.
In people with type 1 diabetes, the beta cells are gone, but the alpha cells still produce glucagon. It is not precisely clear what happens to glucagon production in type 1 diabetics, but some studies imply that overall glucagon levels are higher, and further that the alpha cells are no longer signal responsive — that is, hypoglycemia does not induce an increase in glucagon production as it does in non-diabetics.
In sum, then, insulin makes blood sugar go down, and glucagon makes it go up. In the world of type 1 diabetes, glucagon is like non-caloric glucose tablets, or the magical insulin-eraser we all wish we had.
And hence the brilliance of using glucagon in daily therapy. I don’t know about most diabetics, but I, for one, end up eating half my daily calories in corrective carbohydrates as I try to carefully titrate my insulin doses, and it sucks. What if instead of taking in sugar every time I were going low, I could use glucagon to release stored glycogen in my liver, just like a non-diabetic would? Or, better yet, what if my artificial pancreas were able to use glucagon in concert with insulin, alternating between the two, to mimic what the natural pancreas does to control glucose levels?
Glucagon and the Artificial Pancreas
The problem is that glucagon is a finicky peptide, even more so than insulin. The glucagon in rescue kits is a powder, because glucagon does not stay dissolved in saline. In order to have glucagon to use in daily treatment or pump systems, we need it in liquid form. Up until now, no one has been able to create a stable, soluble glucagon — arguably because there is no reason for pharmaceutical companies or academics to invest so much time and effort into solubilizing a hormone that is only used in rescue kits.
Several companies, including Enject and, more recently, Biodel, are trying to create devices that make it easy to mix the powder and saline on-demand. This solves the problem of the hard-to-use rescue kits, but does not solve the glucagon problem for pumps or frequently used, glucose-tab-sized micro-doses of glucagon. However, all the excitement around glucagon in the artificial pancreas has led the JDRF to enter into partnerships with two new companies trying to make soluble glucagon. Xeris Pharmaceuticals is trying to create a liquid glucagon by suspending the peptide in a low-concentration solvent called dimethyl sulfoxide (DMSO), and LATITUDE Pharmaceuticals is trying to create an emulsion to stabilize glucagon.
In anticipation of stable, liquid glucagon, one of the newest entrants into the insulin pump space, Tandem Diabetes Care has jumped ahead of the crowd and created a two-chamber infusion pump capable of holding and injecting both insulin and a secondary hormone, which they expect will be glucagon. This tandem Tandem pump is already being tested in Dr. Ed Damiano’s clinical trial of a dual-hormone bionic pancreas. Dr. Damiano and his team in Boston presented data at the Scientific Sessions demonstrating the relative success of their approach to the bionic pancreas. In imitation of the real pancreas, the bionic pancreas used the Tandem platform to deliver small doses of insulin and glucagon, alternating in response to the movement of the patient’s blood sugar as measured by the Dexcom G4 continuous glucose monitor (CGM). When the dosing curves were plotted, you could see insulin in, then later glucagon to correct if the insulin was too much. The glucagon was mixed and swapped out frequently during the trial, making it impractical for real-life use at this stage, and the base dose was 40 micrograms, which is 4% of the one milligram dose contained in a typical glucagon rescue kit. (For more on the bionic pancreas clinical trial, I highly recommend Kelly Close’s fantastic personal account over at DiaTribe.)
The patient glucose plot lines from Dr. Damiano’s bionic pancreas trials were impressive — but still imperfect. The post-meal glucose peaks were still far higher than a typical non-diabetic’s, and while glucagon helped dampen lows, it did not remove them entirely. Plus, even assuming a stable liquid glucagon, there are downsides to a two hormone pump — more medication to buy and manage, extra tubing, and a second infusion set.
The many ticks in the “Cons” column for glucagon meant not everyone was singing its praises at the Scientific Sessions. I spoke to Dr. John Mastrototaro, the head of Research and Technology at Medtronic Diabetes about Medtronic’s view on glucagon. Medtronic is also working towards an artificial pancreas system, but unlike Tandem and Dr. Damiano, Medtronic does not think glucagon is a critical component of the artificial pancreas. According to Dr. Mastrototaro, suspending insulin delivery will not be enough to prevent all lows in a closed-loop system, but that doesn’t mean that the on-again, off-again glucagon delivery of Dr. Damiano’s bionic pancreas, or a natural pancreas, is the best way to achieve control. Instead of hitting the gas (insulin), then the brake (glucagon), then the gas, and so on to keep blood glucose values in range, Medtronic is taking the approach of being much slower on the gas — that is, being more conservative initially with insulin boluses, aiming to avoid the need for correction with glucagon. Instead of bolusing aggressively with rising blood sugar, Medtronic aims to treat to range, getting glucose values below, for example, 200 mg/dL, and then letting off the gas. Glucagon may be necessary in extreme cases, but not throughout the day– “Ideally, [the artificial pancreas] should only be hitting that point once every day,” if that, said Dr. Mastrototaro.
Why this resistance to glucagon within Medtronic Diabetes? “It comes back to traditional control systems,” Dr. Mastrototaro explained. Imagine a thermostat; you could set it to heat to 72ºF and to cool to 72ºF, and that would maintain the temperature right around 72ºF. However, the system would be very inefficient — heating, cooling, heating, cooling, over and over. That, according to Dr. Mastrototaro, is what the bionic pancreas approach is like. It is preferable, he believes, to instead engineer a more stable solution, perhaps with a slightly wider temperature range, but with less thrash overall. Notably, this is a nice example of how company philosophy dictates approach; Dr. Damiano’s approach mimics the natural pancreas, which is very robust but also very inefficient, creating and destroying proteins constantly so that it can respond to even minor changes in conditions. The Medtronic approach does not aim to mimic a pancreas; the ideal solution is not necessarily the most biologically accurate one. After all, as the classic engineer’s reasoning goes, planes don’t fly by flapping their wings.
As a result of this foundation in control system engineering, Medtronic is actively researching alternative ways of improving closed loop control algorithms that don’t rely on glucagon. “If we can avoid needing glucagon at the end of the day, then that’s better for the patient,” as no glucagon means fewer cartridges, simpler pumps, less medicine to keep track of, and one less component of the artificial pancreas that will have to go through an FDA approval process. Thus, while Medtronic is actively looking into how glucagon fits into their closed loop equation, they are also engaging in partnerships to find faster insulins, which they hope will reduce the stacking and mistiming of boluses that can result in hypoglycemia.
Who is right — Dr. Damiano or Dr. Mastrototaro? Is glucagon a necessary component of the artificial pancreas? Only time will tell, but in the meantime, I suggest getting to know glucagon a bit better– we’ll be hearing a lot about it in coming year!
 Taborsky GJ Jr. The physiology of glucagon. J Diabetes Sci Technol. 2010 Nov