Objective
We propose a novel strategy to monitor the immune system and predict disease progression that avoids invasive biopsies, enables characterization of the immune response, and creates a therapeutic window for personalized immune suppression that extends β-cell function. The foundation of this technology began in cancer metastasis, in which we reported that subcutaneous scaffold implants sample immune cells and identify immune cell phenotypes within tissues, with the immune cells recruiting metastatic cancer cells prior to colonization of other sites. We have extended this technology for monitoring immune responses to predict immune injury of heart allografts, and for early identification of immune responses prior to symptom onset in mouse models of MS. Here, we propose to translate the diagnostic potential of this immunosurveillance technology to monitor for altered immune responses denoting autoimmune disease progression prior to hyperglycemia in the NOD mouse model. We will employ immune cell-capturing scaffolds to identify predictive biomarkers of autoimmune disease progression, as well as to investigate the T cell populations at the scaffold in comparison to the native pancreas in order to identify new targets for immunotherapy treatment.
Background Rationale
Type 1 diabetes (T1D) affects an estimated 1.25 million people in the US, and as with many immune disorders, the diagnosis of T1D occurs too late for prevention. The lifelong management of T1D with exogenous insulin, as well as addressing the severe vascular complications (e.g. coronary artery disease, stroke, retinopathy, neuropathy, nephropathy) place a tremendous burden of care and expense on patients without a true cure. Delaying or preventing the onset of T1D before irreversible β-cell destruction occurs represents an unprecedented clinical opportunity – which requires i) improvements in detecting immune dysfunction, ii) identifying the dysregulated cells or pathways, and then iii) applying an appropriate immune therapy. The ability to reliably perform these 3 steps could fundamentally change treatment, and in the case of T1D, enable primary prevention of β-cell destruction in the pancreas.
Currently, at-risk populations are identified based on genetic screens, and first degree relatives. The major monitoring approach in these populations has been measuring autoantibodies in serum, with positive autoantibody measurements denoting increased risk of developing T1D within 10 years. While these autoantibodies are useful in monitoring for risk of disease, these tests do not reflect active disease, and do not give an actionable readout for preventative interventions. As a result, additional techniques are needed to clearly distinguish the stages of disease and indicate that the disease may progress to a subsequent stage. The use of scaffolds to detect immune dynamics associated with disease initiation and progression, will act to fill the gap in monitoring, which can then be employed to administer a therapy prior to islet destruction.
Description of Project
Type 1 diabetes (T1D) affects an estimated 1.25 million people in the US, and as with many immune disorders, the diagnosis of T1D occurs too late for prevention. Patients are most often diagnosed when they present acutely ill due to prolonged hyperglycemia or diabetic ketoacidosis, which is a result of the destruction of the insulin-producing pancreatic β-cells. This irreversible loss of β-cells necessitates managing the disease with lifelong use of insulin and addressing the severe macrovascular and microvascular complications (e.g., coronary artery disease, stroke, retinopathy, neuropathy, nephropathy) at tremendous expense rather than a true cure. Teplizumab has emerged as an immunotherapy with the potential to delay onset of T1D, yet this development has also indicated a need to identify disease prior to β-cell destruction. The limitations of this therapy have motivated our long-term goal of developing a surrogate tissue in the subcutaneous space that can monitor immune responses in tissues and identify active, rather than risk of, disease prior to β-cell destruction. The prognostic capabilities of this surrogate tissue would enable monitoring for disease progression in at risk individuals, leading to more individualized care.
Anticipated Outcome
Our preliminary data identify the scaffold-based immune dynamics provide a specific molecular signature correlating with disease in models other than diabetes (e.g. multiple sclerosis, transplant rejection), which will now be applied to autoimmune diabetes progression prior to substantial destruction of β-cells. We anticipate that analysis of the scaffolds can distinguish mice that will progress to hyperglycemia (stage 3 disease) from those that will remain euglycemic. We anticipate that addition of insulin to the scaffold will enrich the autoreactive T cell population at the scaffold, and that the signatures for both antigen loaded, and blank scaffolds will be able to identify disease. We anticipate that we will be able to distinguish healthy mice from mice with insulitis or diabetes, as well as between mice insulitis and diabetes. Also, we anticipate that our computational methods will lead to novel methods for characterization of immune responses in graft rejection. One consideration for machine learning is the potential for overfitting, which we address through appropriate controls and cross-validation methods.
An additional outcome will be the cellular and molecular composition in the scaffold as a measure of disease progression. We anticipate that myeloid cells in healthy mice will have limited capacity for activating or recruiting T cells, which will change with the transition to T1D. We anticipate gene expression sequencing will provide a more comprehensive and unbiased strategy relative to other experimental and clinical approaches. One alternative would be to isolate specific populations for sequencing and signature development, such as monocytes or T cells. The analysis of T cells could identify antigen specificity and T cell state (e.g., activated vs. exhausted), and could be done in addition to the signature based on the total cell population. Collectively, we anticipate that molecular changes in the immune response at the IN will predict T1D onset and progression.
Relevance to T1D
Type 1 diabetes (T1D) affects an estimated 1.25 million people in the US, and the most common treatment is life-long exogenous insulin. Although insulin therapy has been successful, hypoglycemic events and vascular complications persist, indicating a need for preventative treatments. The major monitoring approach used clinically has been measuring autoantibodies in patient serum, but these tests only denote a risk status, and are not an indicator of active disease. The immunotherapy teplizumab has had successful clinical results, yet depends on administration during recent onset disease, for which there are currently no pre-symptomatic identifiers. Additional techniques are necessary to clearly distinguish the stage of T1D and indicate that disease may progress to a subsequent stage. A novel assay is needed to identify systemic immune activation that precedes β-cell destruction, which could be applied to identify the time at which personalized immunotherapy is administered. Patients with positive autoantibody tests may represent a patient population at significant risk for which a scaffold could monitor immune processes that could guide the administration of therapies such as teplizumab.