Objective
We believe that the computerised Automated Oral Minimal Model (AOMM) represents a major advance in accurate monitoring of how many insulin making cells are left in early stage T1D. This test should involve little or no additional burden for subjects. We propose that it could be used to rapidly select treatments which are effective at preserving insulin making cells in early stage T1D. If so, this should accelerate the development of more drugs that can delay or prevent individuals requiring insulin.
To test this, our objectives are to:
(1) Use existing datasets on more 7000 individuals with early stage (preclinical T1D) to see if the AOMM is a better way of monitoring who is losing their insulin making most rapidly.
(2) Collect additional early samples (at 10 and 20 minutes) in the standard oral glucose tolerance test in individuals with prediabetes being monitored by TrialNet and INNODIA to see if this improves the performance of the AOMM.
(3) Conduct a novel state-of-the-art antigen immune cell assay alongside the AOMM measurements to determine whether AOMM is a sensitive marker of the immune process which destroys the insulin making cells in early stage T1D.
Background Rationale
The recent licensing of teplizumab (TZield) in the USA for the delay or prevention of type 1 diabetes (T1D) is a major landmark in T1D research. For the first time we are now able to change the disease course in T1D and slow the destruction of the insulin making cells. The next step is to find other treatments which either alone or combined with teplizumab slow the disease process more and more until eventually insulin is not needed at all. From studies in individuals with recently diagnosed (“stage 3” T1D), we already know of 7 other treatments besides teplizumab which can slow the disease process. We now need to test these earlier in the disease process, to see if they too can delay the need for insulin.
However, the TrialNet study – TN10 – that led to the licensing of teplizumab took nearly 10 years to complete. To accelerate progress, we need a way of detecting what is happening to the insulin making cells more rapidly.
The oral glucose tolerance test (OGTT) involves a drink of sugar followed by blood sampling for 2 hours at 5 times (0, 30, 60, 90 and 120 mins) and is a standard test used to see if individuals have developed diabetes. Until now, researchers have routinely measured the amount of insulin the body releases during this test to estimate the number of insulin making cells remaining. However, in the early stages of diabetes the standard way of analysing this test does not seem to work well. We have initial evidence that an advanced computer modelling analysis that matches the amount of insulin released to the level of glucose every 30 minutes during this test – the so-called “Oral Minimal Model” - more accurately reflects how many insulin making cells are remaining. In addition, we have found that testing the levels at two extra time points early in the test (10 and 20 minutes after the start) improves its accuracy significantly without increasing the burden for people being tested. Recently, the software company Nanomath has developed a programme that can calculate the “Oral Minimal Model” rapidly on large datasets. We refer to this as the Automated Oral Minimal Model (AOMM).
In this study we will test whether the Nanomath Automated Oral Minimal Model with the additional two early samples is a major advance on the way we currently obtain results from the oral glucose tolerance test. If so, using this new method should allow us to test new drugs to delay the need for insulin much more rapidly.
Description of Project
In November 2022, the first drug – teplizumab (TZield) - was licensed in the USA for the delay or prevention of type 1 diabetes (T1D). For the first time we are now able to change the disease course in T1D and slow the destruction of the insulin making cells. The next step is to find other treatments which either alone or combined with teplizumab slow the disease process more and more until eventually insulin is not needed at all. From studies in individuals with recently diagnosed (“stage 3” T1D), we already know of 7 other treatments besides teplizumab which can slow the disease process. We now need to test these earlier in the disease process, to see if they too can delay the need for insulin.
However, the TrialNet study – TN10 – that led to the licensing of teplizumab took nearly 10 years to complete. To accelerate progress, we need a way of detecting what is happening to the insulin making cells years before so many are lost that insulin is required.
The oral glucose tolerance test (OGTT) involves a drink of sugar followed by blood sampling for 2 hours at 5 times (0, 30, 60, 90 and 120 mins) and is a standard test used to see if individuals have developed diabetes. Until now, researchers have routinely measured the amount of insulin the body makes during this test to estimate the number of insulin making cells remaining. However, this does not seem to change very much until almost all the insulin making cells have been lost. We have initial evidence that an advanced computer modelling analysis that matches the amount of insulin to the level of glucose every 30 minutes during this test – the so-called “Oral Minimal Model” - more accurately reflects how many insulin making cells are remaining. Recently, the software company Nanomath has developed a programme that can calculate the “Oral Minimal Model” rapidly on large datasets. We refer to this as the Automated Oral Minimal Model (AOMM).
In this project we will test whether using the Automated Oral Minimal Model computer calculation and using a 7-point rather than a 5-point test is a better way to monitor the health of the insulin making cells. First, we will use existing data on more than 7000 children and adults who have been monitored previously by the networks TrialNet, INNODIA and TEDDY. We will recalculate the results of the glucose tolerance tests using the Nanomath software and see if this predicts who is going on to develop diabetes better. Secondly, we will ask participants currently being followed in TrialNet and INNODIA when they come for their glucose tolerance test if we can take additional blood at 10 and 20 minutes after drinking the sugar (7 point test). Using the Nanomath software we will study whether the 7-point test is better in monitoring the insulin making cells. Finally, we will store blood samples for a state-of-the-art test of the immune reaction against the insulin making cells and compare this measure of the activity of the process destroying the insulin making cells with the results of the 7 point Nanomath computed model.
At the end of the study, we hope to have identified a better way of using the standard oral glucose tolerance test to measure how many insulin making cells are left in individuals who have not developed diabetes yet. If we are successful, this will allow us to tell within 6 months whether a drug is slowing this process and is likely to be useful at delaying the need for insulin.
Anticipated Outcome
We anticipate that the new test we are developing – the 7 point Nanomath Automated Oral Minimal Model (AOMM) - will be a major improvement in testing how many insulin making cells are remaining in people who are in the early stages of Type 1 diabetes – the stages before insulin is required. In particular, we anticipate that any change in the test level over 6-12 months will indicate reliably whether the disease is progressing and if so, how quickly. The test can then be used work out within 6-12 months whether drugs like teplizumab are working to slow or halt the disease. We can then rapidly calculate which dose of a new drug is best, and which treatments or combinations or treatments are most effective. Furthermore, we can detect if a treatment is no longer working and the person should be switched to a different treatment.
Nanomath have agreed to license their software to all researchers and companies working in the field of T1D. This will mean that the 7 point AOMM could be become the “industry standard” test that everyone uses to quickly and accurately monitor the insulin making cells in pre-diabetes. This will overcome a major obstacle that has until now hindered rapid progress in developing treatments to prevent the need for insulin.
Relevance to T1D
The development of an automated tool for modeling insulin secretion and action within the standard 2 hour glucose tolerance test will enable more accurate assessment of insulin making cell function in early stage T1D without increasing the burden for trial participants. This is key to building on the success begun with teplizumab in delaying the need for insulin.
The automation of the model will ensure that all researchers and drug companies are using the same measure. The way the model works is “traceable”, which means that agencies like the FDA can easily check on the results and potentially approve its use as a “surrogate marker” for licensing drugs more rapidly. As more drugs become available, measuring the change in the insulin making cells accurately allows treatments to be “personalized”. If a person is on one drug but the test shows the insulin making cells are declining, they can be switched to another drug or another drug added in before it is too late. This also allows the development of “adaptive clinical trials”, in which less effective treatments are rapidly dropped and the trial changes to focus on more effective approaches.
Increasing the speed at which we can get answers from clinical trials will mean that we can develop treatments to increase the delay period before insulin is needed more quickly. The new automated tool is designed to be particularly effective in the very earliest stages of T1D, when there are quite a lot of insulin making cells remaining. Up until now, it has been very difficult to test drugs at this early stage, but this is the best stage at which to halt the disease, as it leaves patients with a lot of insulin making capacity reserve.
Taken together, we believe that if all these properties of the new test are confirmed, it will begin the transformation of the focus of care of T1D in the future from insulin therapy to preventing the need for insulin.