George Eisenbarth, MD, PhD: Autoantibodies, insulin as the primary T cell target in T1D onset, and the Barbara Davis Center for Childhood Diabetes in Denver
By Adam Burrack, PhD
Relevant content:
New Eng J Med 313(14): 893-894 1985,
New Eng J Med 314(21): 1360-1368 1986,
Nature 435(7039): 220-223 2005,
Journal of Immunology 187(11): 5921-5930
In previous research reviews I’ve described T cells of the immune system as the cell type responsible for destroying the insulin-producing beta cells in the pancreas. This edition of the SMRR will focus on the primary target the T cells are seeking to destroy, a peptide derived from the insulin molecule itself, and how we’ve come to know that insulin is the essential target of this immune response. Central to this story is the career and scientific legacy of Dr George Eisenbarth, MD, PhD, former executive director of the Barbara Davis Center for Childhood Diabetes in Denver, Colorado.
In the mid 1980’s two fundamental pieces of the puzzle of why and how type 1 diabetes (T1D) occurs were hotly debated. First, clinicians needed a better diagnostic tool for T1D than severely elevated blood sugar levels – ideally a predictive tool. Physicians and scientists recognized that it would be highly preferable to develop a predictive test to screen individuals likely to develop type 1 diabetes – and to intervene before disease onset – than to react to established disease. Second, to develop intervention approaches, it was unclear at the time what the primary target of T cells causing the destruction of beta cells was. The two main candidates for this target were peptides derived from either the insulin molecule itself, or from the glutamic acid decarboxylase (GAD65) protein, which is also expressed in neurons. Dr Eisenbarth, who was then at the Joslin Diabetes Institute in Boston, investigated these critical question using cutting-edge techniques.
In a seminal paper in the New England Journal of Medicine in 1985 Dr Eisenbarth characterized the presence of antibodies against the insulin hormone in the blood of patients with T1D. Since these antibodies were specific for a protein made by their own bodies – something the immune system in healthy individuals does not generally do – these antibodies were named “auto antibodies”. This technique was innovative because it established a marker of impending disease and used only peripheral blood. Obtaining a biopsy from the pancreas is an invasive procedure for diagnostic purposes, is open to interpretation of the pathologist analyzing the sample, and would be a dangerous procedure to repeat multiple times on the same at-risk individual. In contrast, testing for the presence of antibodies in the blood is highly specific, is straight-forward to perform in the laboratory, and is repeatable over time on samples taken from the same at-risk person. Importantly, the predictive value of insulin auto antibodies was soon recognized: individuals at risk of developing T1D were found to have detectable levels of antibodies specific for insulin-derived peptides months to years before T1D onset, potentially opening a therapeutic window to stop diabetes onset. Subsequent studies found that at-risk individuals harbored antibodies specific for several proteins present on or within the secretory granules released by beta cells, which contain insulin. The presence one of type of “auto antibody” was found to correlate poorly with later disease onset, whereas the presence of two or three (or even four) auto antibodies correlated much better with impending disease. A key follow-up from these original studies is that auto antibodies specific for the insulin molecule often appear first, enhancing the predictive power of this test. These discoveries facilitated development of the so-called “Eisenbarth curve,” which is pictured below, in a 1986 New England Journal of Medicine paper by Dr Eisenbarth, which remains the model of how T1D develops in at-risk individuals (figure 1).
These discoveries led directly to Dr Eisenbarth’s recruitment to Denver, Colorado, to the position of executive director of the Barbara Davis Center for Childhood Diabetes (BDC). In part through Dr Eisenbarth’s leadership, the BDC has consistently been recognized as the foremost center for the clinical care and basic research of type 1 diabetes. One highlight of the clinical research Dr Eisenbarth was involved with at the BDC is the establishment of the DAISY and TEDDY studies, which track T1D onset in several locations throughout the world. As I’ve mentioned in previous blogs, it has long been recognized that genetics of antigen presentation to T cells and T cell activation are not the only variable at play in T1D onset. The primary aim of the large-scale DAISY and TEDDY studies has been to determine environmental risk factors which promote T1D onset in genetically at-risk individuals or are associated with protection from T1D in at-risk relatives of individuals with T1D. In addition to these ambitious goals, DAISY and TEDDY have contributed to the body of evidence demonstrating that the incidence of T1D is increasing throughout world. Data from these epidemiologic studies implies a key role for the environment, and in particular the diet in the development of T1D.
In his role as director of his own laboratory at the BDC Dr Eisenbarth embarked upon a research program attempting to determine the primary auto antigen in T1D onset in the NOD mouse model of disease. As one might expect, these efforts focused on the insulin molecule and Dr Eisenbarth was also extremely successful in basic research. In a 2005 manuscript in the journal Nature, Dr Eisenbarth’s post-doctoral researcher Maki Nakayama, genetically engineered NOD mice to have one specific mutation in the B chain of the insulin molecule. This slightly modified insulin remained fully functional in its role of promoting nutrient metabolism in muscle and fat cells. Strikingly, these mice did not develop autoimmune diabetes – to any degree. The mutation in the insulin molecule was in the B chain, at the 16th amino acid. This result strongly suggested that peptides derived from sections of the insulin B chain including this mutation were critical to promoting the T cell response that went on to destroy the majority of insulin-producing beta cells. Subsequent work has determined the peptide, amino acids 10-23 of the B chain. Future blog posts will describe on-going work at the BDC and elsewhere, which are attempting to re-establish ‘tolerance’ of the immune system to insulin.
As both an MD and a PhD, it was a focus of Dr Eisenbarth’s to translate basic knowledge of how diseases work to action in the clinic. Another of Dr Eisenbarth’s trainees, an endocrinologist named Aaron Michels, pursued a strategy to physically block the interaction of the insulin B chain-derived peptide with the antigen-presenting molecule in the mouse model of T1D. This strategy was focused on a small molecule, which statistical modeling software predicted would physically prevent the insulin B chain-derived peptide from ‘docking’ in the MHC binding groove. If the peptide could not bind in the MHC groove, there would be no way for autoreactive T cells to become activated, preventing attack of the beta cells. This method achieved some success in the NOD mouse, delaying diabetes development when administered continuously. Dr Michels is moving this concept forward and is conducting an on-going clinical trial aimed at inhibiting insulin-specific autoreactive CD4 T cell activation in new-onset T1D patients.
Sadly, Dr Eisenbarth lost his own battle with pancreatic cancer in December of 2012. Following Dr Eisenbarth’s death, the Juvenile Diabetes Research Foundation named an endowed chair in clinical immunology named after him. In a fitting gesture, the BDC and the JDRF awarded this endowed chair to Aaron Michels, MD, George’s trainee and a physician-scientist at the BDC.