Engineering smart insulin, human clinical trials, translational research
by Adam Burrack, PhD
In an exciting paper released recently in the Proceedings of the National Academy of Sciences of the USA, Dr Danny Chou, who was at the time of a post-doctoral researcher at MIT in the laboratory of Daniel Anderson in the department of chemical engineering described the development of a type of “smart” insulin to treat type 1 diabetes. In short, this type of smart insulin is a variation on the chemical structure of human insulin and currently used synthetic insulins which mimics the biological function of the hormone – regulating blood glucose levels – but which can be injected once every 10-14 hours (in mice in this study). It’s worth noting that the JDRF has been funding “smart insulin” research since at least 2003. This breakthrough represents a big “win” for both the tactic of working on smart insulin and the JDRF’s broad strategy to fund many areas of potential cure-based research.
As our readers currently using insulin pumps will know, insulin pump therapy requires continuous infusion of “basal rate” insulin along with meal “bolus” insulin. This combination very closely mimics normal insulin production by the human pancreas. This so-called smart insulin delivery strategy is facilitated by chemical modifications to the insulin that prevent it from being used by cells immediately after injection. The typical half-life of an insulin molecule in the blood is 5-7 minutes. This means that 50% of the insulin you inject for a meal bolus is bound to insulin receptors on the surface of muscle or fat cells – and is promoting production of glucose receptors to move sugar into cells – within a handful of minutes. Insulin gets to where it needs to go quickly via the circulation. The problem has always been that insulin receptor signaling, like making insulin in the pancreas, is a multi-step process. Smart insulin is more efficient in that its chemical modifications prevent it from binding to the insulin receptors, meaning it can stick around for later use, as needed. One of the modifications to this insulin binds a common protein found in blood, so this insulin continues to circulate.
Smart insulin is essentially held in reserve until needed. In theory, using this type of insulin instead of conventional formulations a person using smart insulin would inject the entire insulin dose they may require over the following 10-14 hours, eat and exercise as normal, and the insulin would be released to act as needed. The second chemical modification to this insulin is a chemical compound that is changed by elevated blood sugar levels. This change releases the smart insulin to act on cells with insulin receptors (ie muscle and fat cells).
There are pros and cons to this general approach to insulin therapy. One benefit of the smart insulin approach: a slight over-dose may not promote low blood sugar events (hypoglycemia), since smart insulin is only released to promote glucose uptake by cells when blood sugar levels are elevated. One con is that under-dosing is still possible, meaning that if there is not enough insulin available high blood sugar levels can still result. This means that the basics of insulin pump diabetes management – knowing your basal rate and carbohydrate ratio – would still apply. The primary pro to this method, in my opinion, is the once-every-12-hours bolus.
This exciting new formulation of insulin is predicted to revolutionize diabetes therapy in a press release from MIT and in the popular press. Given the hype surrounding this research breakthrough and the iPS-derived beta cell breakthrough by Doug Melton and colleagues this past fall, I thought it might be appropriate to review the pathway that a new discovery must travel before becoming a clinical therapy that is routinely used in human patients. Long story short, while these are both cutting-edge development in basic research, it will take a lot of proving over several years’ worth of rigorous scientific and regulatory testing before either of these therapies reaches the clinic.
In the United States, the Food and Drug Administration (FDA) is tasked with protecting the public health. The primary task of the FDA is to protect the public from unsafe medical treatments and devices. As much as it may frustrate individuals who have a disease that could be treated by a new breakthrough, the mandate of the FDA is not to accelerate medical innovation or fast-track promising therapies in the absence of appropriately controlled clinical trials.
A typical time frame from drug discovery to FDA approval and market use – as prescribed by physicians – is 15 to 20 years. There are several steps in this process – several different levels of clinical trials – and several “trap doors” through which a candidate drug can fall and exit the pathway. Let’s walk through this process and some potential pros and cons for smart insulin.
Before the product can be produced, there are almost certainly patents under review regarding this technology and those issues/rights will need to be ironclad before any company will license the technology from MIT (where Dr Chou was a post-doc during discovery) to conduct a clinical trial. After determining who owns the technology (MIT will own this technology), the technology needs to be licensed so a company can make pure chemical compound.
Prior to conducting a clinical trial you need a GMP (good manufacturing process)-produced product. This was a research paper, not a techical sheet from a company practiced at making insulin for injection into people. On the plus side, Novo Nordisk and Eli Lily have decades of experience producing synthetic insulin molecules for use in humans. Scaling this type of production process from research application to industrial application presents unforeseen challenges nearly every time. Some products die at this point due to these logistic challenges. Fortunately Novo Nordisk and Eli Lilly have been making insulin for decades, so the infrastructure exists to make this phase quick.
Next we’ll overview the clinical trial steps. Phase 1 clinical trial is just safety, not efficacy, and is at least 2 years. Hopefully smart insulin is safe in people (ie no life-threatening lows due to crashes or diabetic ketoacidosis (DKA) because the insulin didn’t release when it should have). Phase 2a is pilot efficacy and 2b is larger scale. Again, these are both multi-year, safety still matters, and scientists have to publish results and companies need to move products through this pipeline. Count on at least 1-2 years of study plus delays for data review, analysis, and publications. Phase 3 is the big safety/efficacy large–scale trial. This is what people mean when they say ‘clinical trial’. Phase three requires thousands of people, ideally including multiple countries and different genetic backgrounds over an extended (>2 years) period of time. Bad (adverse health events, especially if life-threatening or severe) results here — in even a tiny subset of patients — kill the product. FDA avoids negative health outcomes at all costs (this is a good thing for protection of public safety). On the flip side, this process significantly delays time-to-market. Good results in phase 3 mean the product can be FDA approved. Good efficacy results are required but not sufficient at this stage. Phase 4 is post-approval surveillance. This phase includes continued monitoring, patient sub-set analysis, and further revision of the product. Importantly, the approved product can still be pulled from the shelf if adverse health events resulting from long-term use are discovered.
Smart insulin is an exciting discovery, and may mean we can do one or two shots per day instead of continuous insulin infusion with pumps. However, it is important to keep in mind that smart insulin does not mean the patient can be ignorant of their insulin requirements, including the difference between basal and bolus insulin. Smart insulin will need to go through the FDA approval process, which will probably take years. Importantly, if approved, this treatment does not fix the underlying autoimmunity which caused the disease. Smart insulin represents an improved insulin formulation, not a biologic cure for the T1D.