Creating Glucose-Responsive Insulin: The Crowdsourcing Approach

From its isolation and first use in 1922, insulin has been called one of the few miracle drugs, saving diabetics from the early death of an untreatable disease. Any diabetic nowadays, however, will tell you– insulin treatment for glucose control is a tricky business. As Dr. Bruce Buckingham said at the recent TEDxDelMar conference, insulin has “a very narrow therapeutic margin.” In other words, it’s easy to give too little– resulting in hyperglycemia– or too much– resulting in hypoglycemia. Titrating insulin dosage throughout the day is a difficult and time consuming process for diabetics, and the current system results in many dangerous excursions outside the ranges of normal blood glucose levels.

What can we do about this? Two prominent answers are: replace the body’s beta cells, as they are better glucose regulators than any human will ever be; or, develop an artificial pancreas that can measure glucose levels and algorithmically dose accordingly.

And then there’s a third option, one that could provide an alternate route to success: make an insulin which works only when there’s glucose circulating that needs to be taken up by cells. If there’s no excess glucose, the insulin is inert, and patients don’t become hypoglycemic. If there is too much glucose in the bloodstream, the insulin would somehow be activated, lowering glucose levels and avoiding hyperglycemia.

Sounds cool, right? There’s only one problem: we don’t know how to make that kind of insulin yet. A number of approaches have been tried, including, for example, systems that rely on glucose and pH-sensitive hydrogels, but so far nothing has proved to be a reliable, usable way of delivering insulin only in response to glucose stimulation.

How hard can it be? Well, that’s where you come in. Really. The JDRF has put together an open contest, asking for written proposals for a way of developing glucose-responsive insulin. They are running the challenge through InnoCentive, a platform for crowdsourcing solutions to complex problems. The JDRF is offering up to $100,000 for the proposal or proposals they think have the most promise.  The deadline is November 9th– so get cracking!

To find out a little bit more about the contest, we spoke to the director of the JDRF’s Insulin Initiative, Dr. Sanjoy Dutta:

What is glucose-responsive insulin? Why would it be useful for diabetics?
Glucose-responsive insulin (GRI) is a novel form of insulin that would basically work only when the body needs it. It would deliver the precise amount of insulin in response to circulating glucose levels, to control blood glucose levels with once-daily or less frequent dosing in people who have insulin-dependent diabetes. Compared to the current administration of insulin multiple times or continuously throughout a day, glucose-responsive insulin would not need to be calibrated with carbs or blood glucose testing.

This type of insulin holds the potential to transform the lives of the hundreds of millions of people with diabetes who need insulin to live. It would improve glucose control, decrease or eliminate the need to test or monitor blood glucose levels, improve the quality of life for people with diabetes, and reduce their chances of both short- and long-term complications.

How much would glucose-responsive insulin be able to accomplish? That is, if we had glucose-responsive insulin, would we still need an artificial pancreas? Continuous glucose monitors?
A GRI, in its truest form, would be game-changing in the way insulin-dependent diabetes (both types 1 & 2) is treated. It would be an alternative for those who may not want to use devices such as pumps, glucose meters, CGMs, and even an artificial pancreas system. We envision it to be a once-a-day (or less frequent) drug, with occasional (maybe once-a-day) glucose measurement. However, it is important to understand that no drug, including a GRI, is going to work for all individuals with insulin-dependent diabetes. Therefore, there will continue to be a need to discover and develop other treatments, such as artificial pancreas systems, to benefit all individuals with type 1 diabetes.

Why hasn’t this been done yet? What are the challenges in creating an insulin that acts only in the presence of glucose? In other words, what makes this a hard challenge?

Designing a drug that will be device-free and deliver in response to circulating levels of glucose (or any other chemical/protein) is a very complex, hence challenging, problem. We believe it will require the concerted effort of diabetes and non-diabetes researchers, such as pharmacologists, toxicologists, biochemical engineers, and polymer scientists, just to name a few. Novel concepts and initial attempts in this field have either failed or languished due to the lack of such coordinated teams.

Why has JDRF decided to use the InnoCentive contest platform for this challenge? What does InnoCentive offer that is different from JDRF’s normal approach of funding individual researchers?

JDRF has opened this challenge to the public with the aim of leveraging the knowledge and skills of a new, fresh pool of talent, and to obtain different perspectives on how glucose-responsive insulin treatment could be designed.

This type of crowdsourcing approach usually allows for efficient and timely problem solving, and the proof is in the pudding. InnoCentive’s Challenge Driven Innovation approach has led to numerous healthcare solutions in recent months, including the identification of a biomarker for ALS and a new delivery method for folic acid to women in third world countries. In addition, our glucose-responsive insulin challenge will reach the attention of InnoCentive’s 250,000 registered Solvers. It is important that this challenge reaches the ears of potential Solvers as far and as wide as possible, and that is why JDRF is working to get the word out.

The prize description says that up to four ideas may be selected, with awards of up to $100,000. What will the actual value of a winning proposal be? Will there be a sliding scale of prize values, and, if so, what will the awarded amount be based on?
The submitted applications/ideas will be reviewed by a JDRF-convened panel of experts. The number and amount of prize award(s) will depend on the decision of this panel, and JDRF is open to all possibilities in that respect.

If a good proposal is submitted and selected, what will be the next steps? Who will lead the development, trials, and tests? How long might that take?
If we decide to award a Solver(s) for the first phase of the challenge, the “Ideation Phase,” JDRF will invite the winning Solver(s) to participate—either alone or as part of a larger team of Solvers—in the next phase of the challenge: the “Pre-Clinical Proof-of-Principle Validation Phase.” Phase two will build on successful ideas from phase one, and provide detailed research plans, including pre-clinical proof-of-concept studies in animal model(s) of T1D, timelines, and budget estimates. We are unable to predict the length of time for research at this early a stage.

Should phase two prove successful, JDRF would then begin a third phase, the “Clinical Proof-of-Concept Phase.” In this phase, JDRF will work to outline and then follow through with a projected path forward for the clinical development of a pharmacologically safe and effective glucose-responsive insulin.

Each of the steps in this challenge will only transpire according to the success of its preceding step. That is why it is critical that we receive good ideas in this first “Ideation Phase,” as it will set the scene for the entire challenge.

If a proposal is selected, who owns it? What role will the authors of the proposal have in the next steps of development?

Under the Challenge agreement, JDRF will obtain exclusive rights to the intellectual property of the winning solution.

Winning Solver(s) from the first phase will be invited to participate in the next step(s) of the challenge, either alone or as part of a larger team of Solvers.

In theory, if there were such a glucose-responsive insulin, how would it compare to endogenous insulin secretion? How fast would insulin have to be to be both glucose-responsive and sufficient to maintain normoglycemia?
In theory, a GRI drug should closely mimic endogenous insulin secretion, i.e., deliver the precise amount of insulin in response to the glucose levels/demands of the target organs at all times during the course of a given day. At this time, the speed of a GRI drug is not of concern since it is a “continuum” of insulin release and action based on local glucose concentrations.

In the current conceptions, does the insulin respond to glucose in the bloodstream, or glucose that has diffused into tissues and interstitial fluids?
Currently, insulin delivered by a pump in response to glucose levels measured by a CGM device is based on interstitial fluidic concentrations of glucose, whereas manually injected insulin is based on blood glucose levels measured by fingerstick glucometer devices.

When I exercise vigorously, I notice that very often my continuous glucose monitor, with the sensor inserted in my abdomen, detects a quick decrease in blood glucose. If I use a finger-stick to measure actual blood glucose concentration, I see that I am not in fact hypoglycemic. In my folk-science reasoning, I assume this is a result of differential glucose distribution during exercise, as my body moves energy sources to the muscles that need it most. How might such physiological complexities of glucose distribution affect the action of a glucose-responsive insulin? Would differential diffusion and absorption rates matter?
Again, at the present time, it is hard to address these specific metabolic conditions in the absence of a GRI drug ready for clinical evaluation. However, as discussed earlier, an ideal GRI drug would be able to deliver insulin proportional to local glucose levels/requirements, such as the skeletal muscle during exercise, or fat tissue during extended periods of fasting. Thus, differential diffusion and/or absorption rates would be accounted for in the net amount and delivery of insulin through a GRI.

What have researchers tried thus far to create glucose-responsive insulin? Why didn’t these attempts work?
Based on the published data base, it is our understanding that different research groups have addressed different aspects of the design, synthesis or experimental testing of potential GRI and GRI-like drugs. And there are quite a few interesting and promising concepts in the public domain. However, given the characteristic requirements of a GRI drug and complexity of diabetes, it has been difficult to successfully develop a GRI drug to date.

What sort of proposals is JDRF willing to entertain? Is a modified insulin molecule the most likely route? Some sort of encapsulation strategy? Would a nanotechnology-based proposal be thought reasonable as well?
At the present time, JDRF is absolutely open to and will entertain any/all novel ideas for the design of a GRI drug. We believe all of the above, plus additional approaches, are entirely feasible. The main purpose of the prize is to spur innovative thinking and generate novel ideas/concepts—and we hope some of these can be translated into bona fide GRI drugs.

Are there degrees of glucose-responsiveness? Would a plan for semi-responsive insulin be valuable?
Ideally, a GRI drug should be able to release/deliver insulin proportional to a wide range of ambient blood glucose concentrations, from very low blood glucose levels with minimal/no insulin release to very high blood glucose levels with large amounts of insulin release. This would enable control of blood glucose extremes, to maintain a euglycemic (normal) range and avoid hypo- and hyperglycemia.

Whether the solution is a GRI drug that releases insulin semi-responsive to glucose levels, or a combination of two or more such GRI drugs to cover the gamut of blood glucose values, is yet to be determined.

What are the potential risks? Would the constant influx of insulin made possible by such a drug be a health risk?
Safety is foremost in the design and development of any drug, and hence for GRI drugs as well. Diligent and exhaustive safety pharmacology and toxicology studies will need to be performed in both animal models and in human trials. The specific safety studies required to derisk the GRI drug will depend on the design and ingredients of the drug.

There should not be any constant influx/release of insulin from a GRI drug; the release should be strictly glucose-responsive. This will need to be assessed during the drug development process.

Are there any existing drugs that behave in a similar manner, activating only when a related, but not directly targeted, molecule is present?
Great question. To the best of our knowledge, there are devices available that deliver drugs in response to a target molecule (such as drug-eluting stents); however, we are not aware of non-device approaches to achieve this sophisticated level of drug delivery in response to a circulating signal.

What is your best proposal? How would you make glucose-responsive insulin?

In short, we do not have the answer! Hence the search for ideas through this prize! It would be remiss to restrict one’s thinking to approaches already considered.

However, prior art in this and related fields suggests that natural and/or synthetic glucose-responsive polymers can be utilized to “package” insulin such that the combination can possess the required characteristics.

Alternatively, one can learn from the body’s pancreatic beta cells that synthesize and secrete insulin, and mimic the machinery and/or architecture of these cells to generate glucose-responsive insulin drugs.

Karmel Allison

Karmel was born in Southern California, diagnosed with Type 1 Diabetes at the age of nine, and educated at UC Berkeley. Karmel now lives in San Diego with her husband, where she is loving the sunshine, working in computational biology at the University of California, San Diego, and learning to use the active voice when talking about her diabetes.

Karmel Allison

Karmel was born in Southern California, diagnosed with Type 1 Diabetes at the age of nine, and educated at UC Berkeley. Karmel now lives in San Diego with her husband, where she is loving the sunshine, working in computational biology at the University of California, San Diego, and learning to use the active voice when talking about her diabetes.