<h1> <FONT FACE="Arial,Helvetica,sans-serif"> JDRF Funded Research</h1> <table width = "600" > <tr> <td><FONT SIZE="-1" FACE="Arial,Helvetica,sans-serif"><B>JDRF-Funded Research: Lay Abstract </b></font> </td> </tr> </table> <br> <table cellpadding="0" border="0" width="800" cellspacing="2"> <tr> <td width = "60%" bgcolor = "4978AC" height = "20" align="left"> <font color="FFFFFF" face="Arial,Helvetica,sans-serif" size="2"> <strong>PI Name:</strong> Michael Strano <br> <strong>Institution Name:</strong> Massachusetts Institute of Technology<br> <!--strong>Lab Website:</strong> http://web.mit.edu/<br>--><br> <strong>Project duration:</strong> 01-June-2017 to 31-May-2020<br> <strong>Mechanism:</strong> Strategic Research Agreement (SRA)<br> <strong>Project grant award: </strong>$1,758,119.40*<br> * Total Grant award amount may vary depending on budget adjustments and it is contingent upon research progress and availability of JDRF research funds. <br><br> </font> </td> <td width = "40%" bgcolor = "4978AC" height = "20" align="left"><font color="FFFFFF" face="Arial,Helvetica,sans-serif" size="2"> <strong>Grant Key:</strong> 3-SRA-2017-426-A-N<br> <!--strong>Therapeutic Area:</strong> <br>--> <strong>Grant status:</strong> Complete Grant<br></font> </td> </tr> </table> <table cellpadding="0" border="0" width="800" cellspacing="2"> <tr> <td width="100%" bgcolor = "6699CC" height = "20" align="left"><FONT FACE="Arial,Helvetica,sans-serif" SIZE="2" COLOR="FFFFFF"> <strong>Project Title</strong></font> </td> </tr> <tr> <td width = "100%" bgcolor = "EBEBEB" align="justify"><FONT FACE="Arial,Helvetica,sans-serif" SIZE="2"> Computational and Physiological Modeling for Glucose Responsive Insulin Design</font> </td> </tr> </table> <table cellpadding="0" border="0" width="800" cellspacing="2"> <tr> <td width="100%" bgcolor = "6699CC" height = "20" align="left"><FONT FACE="Arial,Helvetica,sans-serif" SIZE="2" COLOR="FFFFFF"> <strong>Objective</strong> </font></td> </tr> <tr> <td width = "100%" bgcolor = "EBEBEB" align="justify"><FONT FACE="Arial,Helvetica,sans-serif" SIZE="2"> This JDRF project addresses the outstanding need for a framework that provides chemists and biologists guidelines for choosing kinetic parameters, enzyme loadings, crosslinking density, insulin release profile, and other important parameters for optimal glycemic control in mice models and patients. To develop these guidelines we combine mathematical models of GRIs with a pharmacokinetic model of glucose-insulin metabolism to directly couple the GRI parameters with efficacy. This mathematical formulation can allow in vitro tests to rule out unfeasible GRIs and minimize animal studies. This proposal adopts as an experimental test set a class of insulin analogs designed by M. Weiss and colleagues at CWRU that contain as variable features (a) a glucose-responsive “cassette” of A- and B-chain modifications and (b) an albumin-binding moiety known in the field to provide a prolonged circulatory life time. The goal of the Weiss group is to create a collection of insulin analogs whose molecular properties span a range of biochemical parameters (such as albumin- and glucose-binding affinities) to provide a functional test for the mathematical modeling. We emphasize that it will be an advantage to model development if the collection contains a spectrum of unfavorable, suboptimal and optimal settings of the relevant parameters in order to provide a multi-dimensional “map” of the parameter space. Such mapping, distinct from a traditional R&D pipeline, represents a novel feature of this proposal. The objectives will be accomplished through the following Specific Aims: Aim 1 – Development of the State Functional Approach to GRI Design. Aim 2 - Determine optimal parameters for each GRI class by solving the variational problem posed by the convolution of models in Aim 1 with human and animal physiological models. Aim 3 – Biosynthesis of 24 insulin analogs functioning as GRI candidates, fully characterized using the criteria outlined in Aim 2. Aim 4 - The five most promising candidates from the library in Aim 3 will be selected for in-vivo testing and model validation. Overall, these efforts hope to accelerate the development of effective GRI concepts and serve as a model for other therapeutics designed to modulate potency in response to a therapeutic endpoint. </font> </td> </tr> </table> <table cellpadding="0" border="0" width="800" cellspacing="2"> <tr> <td width="100%" bgcolor = "6699CC" height = "20" align="left"><FONT FACE="Arial,Helvetica,sans-serif" SIZE="2" COLOR="FFFFFF"> <strong>Background/Rationale</strong> </font></td> </tr> <tr> <td width = "100%" bgcolor = "EBEBEB" align="justify"><FONT FACE="Arial,Helvetica,sans-serif" SIZE="2"> Insulin is a peptide hormone, produced by the -cells of the pancreas that is responsible for the regulation of glucose homeostasis in the body. In diabetics there are complications in either the self-production or the efficacy of the insulin. Strict glycemic control is necessary for the treatment of all forms of diabetes. This is accomplished primarily by adhering to a strict schedule of insulin administration along with lifestyle changes. However, glycemic control is difficult due to the open-loop nature of the problem and the large patient-to-patient variation in response to insulin treatment. In addition, physicians report that the majority of their patients do not maintain their insulin regiment properly. Further, the disease cannot be treated aggressively due to hypoglycemia risk. Technological developments have been made to attempt to “close the loop” so that insulin dosing is directly controlled by the patient’s blood glucose levels. Continuous glucose monitoring (CGM) in conjunction with an insulin infusion pump has shown significant improvement over injection therapy in maintaining glycemic control, while minimizing the need for patient adherence and lowering hypoglycemia. However, CGM with insulin pumps is expensive, and the implanted components both increase the risk of inflammation and require frequent replacement due to the body‘s immune response. A different strategy for achieving closed-loop insulin administration, and a longstanding goal of diabetes treatment, comes from the expanding field of insulin delivery technologies that are responsive to the glucose concentration of their environment. These glucose-responsive insulins (GRIs) have the ability to sense the local glucose concentration through various chemistries and subsequently trigger insulin release. The end goal of such technologies is to develop a synthetic pancreas that can properly respond to blood glucose levels. The literature to date describes four distinct mechanistic schemes of GRI constructs. The common features of all the GRIs include a glucose-sensing mechanism and a trigger for insulin release, activation of potency, or biological accessibility, where each specific scheme has advantages and disadvantages ranging from biocompatibility to insulin release kinetics. We assert that the large number of different GRIs proposed and demonstrated requires a standardized framework to evaluate the limits and conditions of efficacy as an insulin therapeutic. A common experimental approach in GRI development is to empirically test candidate constructs in vivo within an animal model to evaluate the therapeutic effectiveness in what amounts to a costly and time consuming development process. We envision that a mathematical modeling framework for GRI design could potentially predict GRIs performance before testing in animal models, and narrow the parameter space for desired synthesis. Strikingly, there has been minimal emphasis on mathematical modeling of enzyme-based GRI performance at the macro-scale, for example, and instead, efforts to date have focused on the enzyme reactions themselves or the insulin diffusivity inside various polymer formulations. These models do not directly link the GRI parameters to efficacy, which promises to be the central contribution of the current work.</font> </td> </tr> </table> <table cellpadding="0" border="0" width="800" cellspacing="2"> <tr> <td width="100%" bgcolor = "6699CC" height = "20" align="left"><FONT FACE="Arial,Helvetica,sans-serif" SIZE="2" COLOR="FFFFFF"> <strong>Description of Project</strong> </font></td> </tr> <tr> <td width = "100%" bgcolor = "EBEBEB" align="justify"><FONT FACE="Arial,Helvetica,sans-serif" SIZE="2"> The concept of insulin therapy that is activated or self-dosing in response to a patient’s blood glucose level has been a significant goal of diabetes care. This is idea, called Glucose Response Insulin (GRI) is an administered insulin that turns on or increases its potency, concentration or dosing with changes in a patient’s dynamic blood glucose concentration. In this way, the drug could mimic the normal function of the pancreas. However, drugs that respond in this way to a biomolecule have little precedent in biotechnology. Creating a drug that responds to the concentration of a certain biomolecule is a fundamentally different concept from controlling the drug delivery rate or the duration of drug activity in the body. A GRI is a drug that responds to the patient’s physiological state, in this case glucose. This concept may make therapies like insulin safer by building in potency constraints that prevent hypo-glycemic or hyper-glycemic excursions from healthy blood sugar levels. Because the concept behind a GRI is relatively new, there has been little work on how to systematically design them. In particular, the GRI has to have in its chemical and physical operation knowledge of the desired state for the human body. In contrast, GRI development to date has proceeded by a costly trail-and-error, normally testing directly in rodent or other animal models. It remains an unaddressed, fundamental issue of how a therapeutic can itself encode for a desired physiological human state or range of states. A recent workshop held at JDRF in April 2016 of major stakeholders for GRI highlighted the essential problem that the field has been hampered by a lack of fundamental design principles based on physiological dynamics. This problem is addressed in our current application to JDRF as a collaboration between MIT and Case Western. Our team has developed a novel state-function framework in which to mathematically describe any GRI construct according to its interactions with glucose and the resulting active insulin. Our approach to developing design criteria is to then couple these state functions to existing physiological models of the human or mouse endocrine and circulatory systems. Parametric optimization subject to different performance criteria results in molecular and structural parameters that can provide unprecedented guidance to current and future medicinal chemists and clinicians focused on GRI synthesis. This project will underscore this utility by guiding our own novel GRI biosynthetic efforts as well as their testing and validation within a standard rodent model. Our targets for modeling consist of four major schemes of GRIs in active development within the community, and the theoretical aspect of the project will consider all of them including important variations. The project will produce a set of open-source tools and design criteria that will link critical GRI parameters to performance criteria. Our biosynthetic efforts, however, will focus on just one GRI scheme that consists of an insulin construct activated to potency upon specific binding to glucose. A library of such constructs will be synthesized, tested in-vitro, and down-selected according to our modeling criterion for in-vivo testing. Overall, these efforts hope to accelerate the development of effective GRI concepts and serve as a model for other therapeutics designed to modulate potency in response to a therapeutic endpoint. </font> </td> </tr> </table> <table cellpadding="0" border="0" width="800" cellspacing="2"> <tr> <td width="100%" bgcolor = "6699CC" height = "20" align="left"><FONT FACE="Arial,Helvetica,sans-serif" SIZE="2" COLOR="FFFFFF"> <strong>Anticipated Outcome</strong> </font></td> </tr> <tr> <td width = "100%" bgcolor = "EBEBEB" align="justify"> <FONT FACE="Arial,Helvetica,sans-serif" SIZE="2"> 1) Development of the State Functional Approach to GRI Design. For each class of GRI, develop reduced physiochemical models that describe their operation under relevant physiological ranges of glucose concentration and target insulin delivery. This aim will include an extensive evaluation of the design parameters available to the medicinal chemist and their relationship to GRI performance. Compile estimates from GRI examples in the literature for these parameters. Estimate the ex-vivo performance of the GRI to given glucose input waveforms. We will also evaluate important rate constants of the GRIs developed in Aim 3 using the physiochemical models, based on the glucose outputs and the parameters used in their development. Experimental assays such as Isothermal Titration Calorimetry will be utilized to provide model free binding isotherm data directly into the algorithm for precision and accuracy. 2) Determine optimal parameters for each GRI class by solving the variational problem posed by the convolution of physiochemical models in Aim 1 with the physiological compartment models. This aim will involve extensive evaluation of physiological compartment models for humans as well as mice, since the majority of GRI testing is conducting in murine models. After determining predictive compartment models for both human and mice, the GRI physiochemical models will be interfaced with the compartment models to optimize a range of design criteria for each scheme of GRI. Use of in-vivo spectroscopic imaging techniques will be utilized to validate the physiological compartment models of rodents, which are necessary for GRI validation. 3) Biosynthesis of 24 insulin analogs functioning as GRI candidates, fully characterized using the criteria outlined in Aim 2. Isothermal Titration Calorimetry (ITC), Western blot and stopped flow spectroscopy will be employed to measure binding kinetics and thermodynamics to be used as inputs to the State Functional Model of GRI design introduced in Aim 1. 4) The five most promising candidates from the library in Aim 3 will be selected for in-vivo testing. Dynamic glucose measurements in the rat model, as well as insulin concentrations will be compared to model predictions from Aim 1. Our overall goal is to accelerate the development of effective GRI concepts. Additionally, these efforts will serve to instruct in the synthesis of other therapeutics designed to modulate potency in response to a therapeutic endpoint. </font> </td> </tr> </table> <table cellpadding="0" border="0" width="800" cellspacing="2"> <tr> <td width="100%" bgcolor = "6699CC" height = "20" align="left"><FONT FACE="Arial,Helvetica,sans-serif" SIZE="2" COLOR="FFFFFF"> <strong>Relevance to Type I Diabetes </strong> </font></td> </tr> <tr> <td width = "100%" bgcolor = "EBEBEB" align="justify"><FONT FACE="Arial,Helvetica,sans-serif" SIZE="2"> Insulin is a peptide hormone, produced by the Beta-cells of the pancreas that is responsible for the regulation of glucose homeostasis in the body. In diabetics there are complications in either the self-production or the efficacy of the insulin. The majority of diabetic cases fall under two main types. Type 1 diabetes results from the autoimmune destruction of the pancreatic -cells, resulting in a complete lack of insulin production, and patients who depend on exogenous insulin for survival. Type 2 diabetes results from abnormal insulin production and insulin resistance, and patients may require exogenous insulin for blood glucose control if diet and exercise are insufficient. Strict glycemic control is necessary for the treatment of all forms of diabetes. This is accomplished primarily by adhering to a strict schedule of insulin administration along with lifestyle changes. However, glycemic control is difficult due to the open-loop nature of the problem and the large patient-to-patient variation in response to insulin treatment. A different strategy for achieving closed-loop insulin administration, and a longstanding goal of DM treatment, comes from the expanding field of insulin delivery technologies that are responsive to the glucose concentration of their environment. These glucose-responsive insulins (GRIs) have the ability to sense the local glucose concentration through various chemistries and subsequently trigger insulin release. The end goal of such technologies is to develop a synthetic pancreas that can properly respond to blood glucose levels. The literature to date describes four distinct mechanistic cases of GRI constructs. The common features of all the GRIs include a glucose-sensing mechanism and a trigger for insulin release, activation of potency, or biological accessibility, where each specific scheme has advantages and disadvantages ranging from biocompatibility to insulin release kinetics. This project will develop the first theoretical design framework for GRIs and validate it experimentally for at least one class of GRI schemes. We assert that the large number of different GRIs proposed and demonstrated require a standardized framework to evaluate the limits and conditions of efficacy as an insulin therapeutic. Our work will accelerate the development of effective GRIs by providing an intellectual framework for their design.</font> </td> </tr> </table>