Pumping Insulin

Shares

The closed loop system is the Holy Grail sought after as an engineering solution for type 1 diabetes. Unlike glucose sensors which need quite a bit of work to take over any decision making (see my last post for details), pump technology is ready. Indeed many diabetics, both type 1 and type 2, now use an insulin pump.

The pump is basically a microcomputer that drives a tiny motor that pumps insulin from a reservoir into the interstitial space in the abdominal area. The computer is fast. The pump mechanism is quite small (the whole device is a little larger than a cell phone) and can be worn externally. Insulin delivered from the pump is available almost immediately.

If the insulin inside the pump were normal human insulin this would not work. Instead, the insulin delivered by the pump is a product of genetic engineering resulting in something we call “rapid acting insulin”. Normal insulin likes to fit together to form hexamers. When it is released from the pancreas or when it is injected, it is in this form. The insulin receptor will only recognize a monomer, however, so this hexamer has to dissociate (dissolve) into monomers to do its job. This is especially a problem with injected insulin. The insulin hexamers do not like to move around in the interstitial space and take much longer to dissociate into monomers than pancreatic insulin. This should not be surprising. The blood is a rough and tumble place and the hexamers quickly fall apart into monomers in the turbulence. In comparison, the interstitial fluid is a calm place (unless you like doing crunches right after your shot). The result is that you can expect to wait at least 30 min or longer to realize the effect of an injection of normal insulin.

The work of understanding the structure of insulin to the point where we could design a protein that retained the receptor binding and activation function but lost the self association function spanned 30 years and the entire field of molecular biology. Insulin was among the first proteins for which the amino acid sequence was determined. Its processing and folding were intensively studied leading to the eventual production of cloned human insulin which then replaced animal insulins. Surprisingly, in the end, only a couple of amino acids needed to be changed to create a protein that no longer self associated but still retained function. The utility of rapid acting insulin was appreciated long before pump technology was available and so various pharmaceutical companies have marketed different insulin mutants that pretty much do the same thing. Eli-Lilly was the first to market with Humalog followed by Sanofi-Aventis with Apidra and Novo-Nordisk with Novolog. All of these mutants have similar insulin receptor binding domains but each has its own patented mutation that keeps it from self association. This was an essential advance necessary for insulin pumps to work. (For more information on the development of rapid acting insulins click here).

Clinical trials have shown that the pump can provide significantly improved glycemic control for diabetes patients but, importantly, only for those that can put in the effort to work the thing. What we have here is an open loop in which it is our own decision making that closes the circuit. What doctors have found is that patients must be committed to blood glucose testing to see benefit from the pump. For example, patients who will test themselves at least 4 times per day were found, as a population, to have a failure rate of 7% in using the pump. In comparison, the population that will not test that often has a failure rate of 18%. Patients at the Barbara Davis Center have described their introduction to the pump as “getting diabetes all over again”.

Since we are pretty complex, it should not be surprising that the control options for the insulin pump are also complex. Fundamentally, there are two types of injection. Since we need insulin all the time at some low level we have what is called “basal injections”. Then when there is a glucose load at meals and during snacks we need more insulin. The pump supplies that insulin as a “bolus injection”. The basal rate of infusion is set by doctors and is generally not adjusted by the patient. It is the bolus that requires constant effort. One can approximate the amount of insulin needed by counting carbs. This requires some effort however and it would be quite useful if companies would provide that number for processed foods.

Closing the loop will happen in small increments. The first steps will involve better ways to assess insulin needs cognitively along with better technologies to sense blood glucose. Currently studies are considering whether constant glucose monitors are capable of sensing a hypoglycemic state and shutting down the pump in response to that state. We will see how it goes.

4
Leave a Reply

avatar
3000
4 Comment threads
0 Thread replies
0 Followers
 
Most reacted comment
Hottest comment thread
3 Comment authors
Robert ScheinmanPaul SorensenCatherine Price Recent comment authors
  Subscribe  
newest oldest most voted
Notify of
Paul Sorensen
Paul Sorensen

Robert

My brother and I are both type 1.  We find it shocking that there hasn’t been a generic  insulin option.  Does the genetic engineering aspect make this impossible? 

Catherine Price

Great post — and it made me think of a question I’ve always wondered about: can you get rapid-acting insulin (like humalog) to act more quickly if you exercise the area of the body into which you inject it? (In the case you mentioned, doing crunches after giving yourself a shot in your stomach?) I do that myself when I feel like the insulin is not acting quickly enough, and it seems to work — but is that true? Or is it that any form of exercise will help you make better use of the insulin on board?

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