Publications and posters
Exocrine Proteins Including Trypsin(ogen) as a Key Biomarker in Type 1 Diabetes
It is known that the pancreas is smaller in individuals with type 1 diabetes. This is caused by a cascade of processes affecting pancreatic cells, not only endocrine cells producing insulin but also cells which make digestive proteins called exocrine cells.
This study describes the first hypothesis-free proteomic analysis of over 4000 proteins in identical twins, affected and non-affected with type 1 diabetes. The twins with T1D were compared to their twin without the condition, surprisingly, the results identified not endocrine, but exocrine biomarkers as the top five differences.
Those results were then validated by measuring one of the digestive proteins – trypsin(ogen) in over a thousand well-characterized individuals with T1D. This involved the following participant groups taking part in the BOX study; those identified at increased risk of type 1 diabetes with multiple islet autoantibodies; individuals close to diagnosis; and those who had long-standing T1D. First-degree blood relatives without T1D were used as a control group.
We concluded that trypsin(ogen) levels are decreased before, at, and after diagnosis of T1D. The analysis in twins showed that genetic and non-genetic factors determine type 1 diabetes.
This study shows that biomarkers of pancreatic mass such as exocrine proteins should be included in prediction and progression studies. Diabetes Care. 2023 Jan 26;dc221317. doi: 10.2337/dc22-1317. Online ahead of print.
The longitudinal loss of islet autoantibody responses from diagnosis of type 1 diabetes occurs progressively over follow-up and is determined by low autoantibody titres, early-onset, and genetic variant
The long-running dedication of individuals with type 1 diabetes that participate in BOX has enabled the first and largest study of biomarkers (islet autoantibodies) from diagnosis to long-standing diabetes (greater than 30 years).
Confirming previous studies, islet autoantibodies are lost over time, but autoantibodies to ZnT8 were lost more rapidly than GAD and IA-2 after diagnosis. This study found that people with lower autoantibody levels at diagnosis, people who were diagnosed younger, and who had type 1 diabetes for longer, were more likely to lose their islet autoantibody responses. We also found that genes that influence islet autoantibody responses are different at and after diagnosis. This suggests that the influence of genes changes depending on the duration of diabetes.
Interestingly, we detected high levels of islet autoantibodies in some people over 15 years after type 1 diabetes diagnosis. The reason for this is not known but may help us to understand what happens to the immune response after someone is diagnosed with type 1 diabetes. Improved understanding of this could be important when monitoring whether immunotherapy is working in individuals after diagnosis of type 1 diabetes in clinical trials. Williams et al. Clin Exp Immunol. 2022 Oct 1;uxac087. doi: 10.1093/cei/uxac087. Online ahead of print.
The measurement of autoantibodies to insulin informs diagnosis of diabetes in a childhood population negative for other autoantibodies
It is important that everyone diagnosed with diabetes receives an accurate diagnosis of the type of diabetes they have. This ensures they receive the correct treatment. One important first step is to identify those with autoimmune diabetes as opposed to genetic forms of diabetes. One way of doing this is by measuring levels of islet autoantibodies. There are currently four islet autoantibody tests available but often only two are used in clinical practice. Children that participate in the BOX study who provided serum samples within 2 weeks of their type 1 diabetes diagnosis allowed us to investigate all four islet autoantibody tests for identifying autoimmune diabetes.
Autoantibodies to GAD, IA-2, and ZnT8 are often tested in research studies but are also available in NHS laboratories to aid type 1 diabetes diagnosis. Autoantibodies to insulin are more difficult to measure and must be tested before insulin treatment is given.
In children with a type 1 diabetes diagnosis, testing for GAD, IA-2, and ZnT8 autoantibodies failed to detect type 1 diabetes in approximately 1 in 20 children. When insulin autoantibodies were tested, almost half of the children who had appeared not to have autoimmune diabetes, were proven to be autoimmune. The majority of these children were diagnosed under the age of 5 years. As a single test, insulin autoantibodies identified type 1 diabetes in 91.2% of children diagnosed under 5 years of age.
Our study found that testing for insulin autoantibodies halved the number of children that would have otherwise undergo monogenic diabetes genetic screening. This genetic screening costs approximately 10 times more than islet autoantibody tests. Williams et al. Diabet Med. 2022 Dec;39(12):e14979. doi: 10.1111/dme.14979. Epub 2022 Oct 30.
Age-period-cohort modelling of type 1 diabetes incidence rates among children included in the EURODIAB 25-year follow-up study
Data from BOX families contributed to this study of people with diabetes across Europe. We know that the incidence of type 1 diabetes has been increasing and the aim of this study was to examine children diagnosed between 0-14 years in the 25-year period from 1989-2013 in 19 European centres. A total of 57,487 cases were included. The results showed evidence for similar age-specific type 1 diabetes incidence rates across the EURODIAB population and peaks at a younger age for girls than boys. There was no obvious environmental causes of the increase in incidence. Acta Diabetol. 2022 Oct 7. doi: 10.1007/s00592-022-01977-x. Online ahead of print.
Activated but functionally impaired memory Tregs are expanded in slow progressors to type 1 diabetes
This study focuses on the regulation of the immune response in people with multiple islet autoantibodies who progress slowly to diabetes. It follows our previous observation that features of the immune response are different in ‘slow progressors’ within BOX. The study showed that the regulatory cells in slow progressors are more numerous but less functional than those in a control group without islet autoantibodies. Boldison J et al. Diabetologia 2021 Oct 28.doi: 10.1007/s00125-021-05595-0. Online ahead of print.
Four decades of the Bart’s Oxford Study; Improved tests to predict risk of type 1 diabetes
Dedicated to islet autoantibody specialist, Alistair John Kerr Williams (1959-2020), this review uses the timeline of the long running Bart’s Oxford study to chart improvements in immunogenetic strategies to predict type 1 diabetes made possible through careful characterisation of diabetes in participants as well as collection and storage of high quality samples for almost 40 years. Since the advent of genome wide association studies, there has been a move from HLA to cumulative genetic risk scores to aid genetic prediction to T1D. There is particular emphasis on the transition from testing for islet cell antibodies using immunofluorescence to biochemical autoantibodies to insulin, GAD, IA-2, ZnT8 and Tspan7. The advantages of particular epitope specific assays are also discussed in improving prediction of diabetes in the general population. Gillespie KM et al. Diabetic Medicine.2021 Oct16;e14717
Regulatory B Cells: Role in Type 1 Diabetes
This review focuses on how regulatory B cells (that have an anti-inflammatory role which can suppress autoimmunity) are altered in the development of type 1 diabetes, highlighting both frequency and function in human and animal studies. Bolidison J et al. Frontiers in Immunology. 2021 Sep;20
What have Slow Progressors Taught Us About T1D—Mind the Gap!
This review highlights the lack of understanding of the mechanisms underlying adult onset autoimmune diabetes. We know that approximately half of type 1 diabetes is diagnosed in adulthood yet natural history studies have really only focused on childhood onset cases. We know that over 80% of children who have two or more islet autoantibodies circulating in their blood will develop type 1 diabetes by the age of 20 years. We have characterised in detail the remaining 20% (Slow Progressors). BUT if 50% of cases are diagnosed in adulthood how does this arise? How common is seroconversion to islet autoantibody positivity in adulthood? Are the islet autoantibody characteristics the same as those observed in childhood? Are there rapid and slow progressors in adulthood as we have observed in childhood? Gillespie KM et al. Current Diabetes Reports. 2019 Sep;19(10):99
Slow progressors to type 1 diabetes lose islet autoantibodies over time, have few islet antigen-specific CD8 + T cells and exhibit a distinct CD95 hi B cell phenotype
This study of immune cell characteristics in a well characterised cohort of elite slow progressors to type 1 diabetes for the Bart’s Oxford study shows that over time, levels of islet autoantibodies decrease. When analysed at least a decade after the first multiple islet autoantibody detection we observed that unlike individuals diagnosed recently (or some time ago) with type 1 diabetes who had robust CD8 T cell responses, Slow progressors did not. In addition, B cells expressed higher levels of CD95 compared with controls. This marker regulates apoptosis of autoreactive cells suggesting a possible mechanism whereby slow progressors are relatively protected. Hanna SJ et al. Diabetologia.2020 Jun;63(6):1174-1185
Persistence of islet autoantibodies after diagnosis in type 1 diabetes
This review builds on a study that some BOX participants took part in around 5 years ago, to examine the relationship between islet autoantibodies and beta cell function after diagnosis. For many years we have known that islet autoantibodies which are common at diagnosis decline in level and number after diagnosis. Using BOX data combined with other studies we described and highlight gaps in our understanding of the rate of loss of islet autoantibodies after diagnosis in different people (i.e. those diagnosed at a young age, or with particular genetic risk) and how this relates to beta cell function. These data are relevant for those testing islet autoantibodies after diagnosis clinically, to monitoring the effect of immunotherapies after diagnosis, and for understanding the progression of islet autoimmunity after diagnosis. Long AE et al. Diabetic Medicine 2021 Oct 6:e14712. doi: 10.1111/dme.14712 online ahead of print.
Characteristics of slow progression to diabetes in multiple islet autoantibody-positive individuals from five longitudinal cohorts: the SNAIL study
Multiple islet autoimmunity increases risk of diabetes, but not all individuals positive for two or more islet autoantibodies progress to disease within a decade. Major islet autoantibodies recognise insulin (IAA), GAD (GADA), islet antigen-2 (IA-2A) and zinc transporter 8 (ZnT8A).
Here we describe the baseline characteristics of a unique cohort of 'slow progressors' (n = 132) who were positive for multiple islet autoantibodies (IAA, GADA, IA-2A or ZnT8A) but did not progress to diabetes within 10 years. Long AE et al. Diabetologia. 2018 Jun;61(6):1484-1490
A quarter of patients with type 1 diabetes have co-existing non-islet autoimmunity; the findings of a UK population-based family study
Individuals with type 1 diabetes (T1D) are at increased risk of coeliac disease (CD), autoimmune thyroiditis and autoimmune gastritis, but the absolute risks are unclear. The aim of this study was to investigate the prevalence of autoantibodies to tissue transglutaminase (TGA), thyroid peroxidase (TPOA), and gastric H+ /K+ -ATPase (ATPA) and their genetic associations in a well-characterised population-based cohort of individuals with T1D from the Bart's-Oxford family study for whom islet autoantibody prevalence data were already available. Kozhakhmetova A et al. Clinical & Experimental Immunology. 2018 Jun;192(3):251-258
A novel LIPS assay for insulin autoantibodies
Insulin autoantibodies (IAA) are often the first marker of autoimmunity detected in children in the preclinical phase of type 1 diabetes (T1D). Currently, the vast majority of laboratories adopt the radiobinding micro-assay (RBA) for measuring IAA. Our aim was to replace RBA with a novel non-radioactive IAA Luciferase Immuno Precipitation System (LIPS) assay with improved performance. Liberati D, Wyatt RC et al. Diabetologia 2018 Mar; 55(3):263-270
Beta cell function and ongoing autoimmunity in long-standing, childhood onset type 1 diabetes
This study aimed to determine the frequency of residual beta cell function in individuals with long-standing type 1 diabetes who were recruited at diagnosis, and relate this to baseline and current islet autoantibody profile. Williams GM at al. Diabetologia 2016 Dec;59(12):2722-2726
Attenuated humoral responses in HLA-A*24-positive individuals at risk of type 1 diabetes
In collaboration with a research group at the University of Exeter, we have shown that a proportion of participants who had type 1 diabetes between 12-28 years ago still have some insulin secreting cells. This is exciting because it suggests that the immune system is no longer attacking insulin producing beta cells and/or that there is continuing renewal of insulin-producing beta cells, combined with some regulation or alteration of the autoimmune response.
Learning more about these changes could help preserve more beta cells and perhaps make blood glucose a bit easier to control. Ye J et al. Diabetologia 2015 Oct;58(10):2284-7
Improved laboratory tests to identify ‘who is at risk of type 1 diabetes in the future’
With specialist colleagues in Italy and Germany, we recently developed a new test to measure antibodies to glutamic acid decarboxyase (GAD).
This test has proved to be more accurate at identifying who will develop type 1 diabetes in the future.
This is important when identifying participants to include in clinical trials of medicines to prevent or delay the onset of type 1 diabetes. Williams at al. Diabetes 2015 64:3247-52
Fewer ‘high risk’ genes needed today to develop type 1 diabetes and altered islet autoantibody patterns at diagnosis
We have compared the genes (DNA) from people with type 1 diabetes in the BOX study with genes from people who had developed type 1 diabetes more than 50 years ago, when diabetes was much less common.
We showed that a ‘lower dose’ of the diabetes-related genes is required to develop type 1 diabetes today than in the first half of the twentieth century.
The results show that the current increase in type 1 diabetes is due to exposure to changing environmental factors. Analysis of genetic material from our more recent mouth brush collection suggests that the trend for people developing diabetes with a ‘lower dose’ of diabetes-related genes is continuing. Gillespie et al. Lancet 2004 364:1699-700
We have also found that the pattern of islet autoantibodies in BOX patients at diagnosis changed between 1985 and 2002, with more antibodies targeting the IA-2 and ZnT8 self-proteins. This suggests that the autoimmune response is also changing over time. Long et al. Diabetes 2012 61:683-686
Type 1 diabetes diagnosed early in life is important for relatives in BOX
In a group of children who have high risk genetic markers HLA DR3-DQ2/DR4-DQ8, we found that their chance of developing type 1 diabetes in the future was further increased if their sibling (with type 1 diabetes) had been diagnosed early in life. Gillespie et al. Diabetes 2014. 63:1041-7
Some islet autoantibodies are less common in people at diagnosis of type 1 diabetes with a high risk gene HLA A*24
This is an unexpected result that has also been demonstrated in people at risk of type 1 diabetes and people with longstanding diabetes.
We think that some islet autoantibodies are found less often in people with HLA A*24 because this gene variant occurs in people who develop type 1 diabetes early in life.
This may mean a more aggressive immune response kills beta cells more quickly without time for some islet autoantibodies to be generated. Long et al. Diabetes 2013 62: 2067-71