Autoimmune Diabetes: Caught in a NET

The mantra is “no type 1 diabetes without autoimmunity.” Indeed, it is difficult to resist the evidence that autoimmunity plays a key role in the origins of type 1 diabetes (1,2). Recently, some aspects of that certainty have been shaken. As with other autoimmune diseases, type 1 diabetes is believed to arise from the impact of environmental insults in genetically susceptible individuals (3). The disease is β-cell specific, associated with β-cell autoantigens, which are recognized both by T cells and highly disease-predictive autoantibodies (2,4). Adoptive transfer of T cells from type 1 diabetic donors or of a pancreas from a twin donor to their type 1 diabetic twin induces the disease process (1,5). Moreover, genes associated with genetic susceptibility to the disease have a number of immune-associated functions, including decreasing the threshold of T-cell signaling and activation and modifying antigen presentation, each of which implies an adaptive immune effect (3). Quod erat demonstrandum? Well, not quite, because additional changes occur that coexist or antedate the development of diabetes-predictive autoantibodies; namely, changes in innate effector cells (6–8), decreased pancreatic weight (9), and metabolic dysfunction (10–13).

Innate effector cells most likely play a role in causing type 1 diabetes, as macrophages, monocytes, dendritic cells, and neutrophils infiltrate the islets at the onset of the disease (1). Indeed, it has long been recognized that type 1 diabetic pancreatic islets, including islet β-cells, hyperexpress the innate cytokine interferon-α (IFN-α), an antiviral defense (14). In line with that evidence, a transcription signature of 225 genes associated with IFN-α stimulation was detected in peripheral blood mononuclear cells before the development of both autoantibodies and diabetes (6). Furthermore, peripheral blood monocytes and macrophages show altered monocyte mRNA and protein expression, even in subjects at disease risk (7,8). Neutrophils also infiltrate human islets in type 1 diabetes and are an early feature of insulitis in NOD mice, in which cross talk between the islet β-cells, macrophages, and neutrophils promote disease (15). Low neutrophil counts and altered neutrophil function have each been implicated in the origins of human autoimmune diseases, and there is a small reduction in blood neutrophil counts both in type 1 diabetes as well as in subjects at risk for the disease (16). But now, in this issue of Diabetes, comes striking evidence for altered neutrophil function, associated with neutrophil death, in the peripheral blood of type 1 diabetic patients (17).

There are many ways for a cell to die. Broadly, cell death is categorized as programmed, regulated, or accidental (18). In autoimmune diseases, research has focused on apoptosis (programmed target cell death) or death through infections, peroxidation, or necrosis (accidental death). But a neutrophil can die using a regulated pathway in which fibrillary nets, composed of DNA, histones, and granular proteins, are exuded from the cell surface. These neutrophil extracellular traps (NETs) trap microorganisms and limit inflammation (19). But such NETs are also potentially toxic as they can activate other innate effector cells to release inflammatory cytokines, promoting NET formation and exposing the immune response to cellular antigens in a feed-forward inflammatory loop, as shown in systemic lupus erythematosus (20). As potential autoantigens exposed by NETs can undergo enzymatic posttranslational modification, this process, controversially called NETosis as it is not certain that the neutrophils are dead, could be important in promoting loss of self-tolerance and autoimmunity (20) (Fig. 1). The new results from Wang et al. (17) suggest that NETs are involved in type 1 diabetes. In patients with the disease, both the levels and enzyme activity of circulating neutrophil protein serine proteases, elastase, and proteinase 3 were substantially increased and positively associated with the number of coexistent diabetes autoantibodies. By comparison, alpha1-antitrypsin levels, an endogenous inhibitor of neutrophil serine proteinases, were decreased. A parallel study in NOD mice suggested that the same changes could occur well before diabetes develops.

At diagnosis of type 1 diabetes, the pancreas weight is decreased with atrophy of the dorsal region and the exocrine system. This weight reduction likely predates the disease, as it can be found in antibody-positive subjects at postmortem (9). Quite what it means, how early it can be found, and how it might contribute to the disease remains unclear. But it is a striking observation that suggests autoimmunity might not initiate the disease. Extensive metabolic changes also have been described in the prediabetic period, even in genetically susceptible twins and siblings without autoantibodies. These changes include fasting hyperproinsulinemia (10), increased advanced glycation end products (11), and altered lipids (12,13). The increase in fasting proinsulin suggests a dysfunctional and stressed β-cell. Less clear is the origin of increased serum advanced glycation end product levels, though a common environmental effect has been implicated (11). Lipidomic alterations antedate autoantibody conversion and can be detected in cord blood (12,13). At birth, prediabetic infants have reduced serum levels of succinic acid and phosphatidylcholine (PC) and later show reduced levels of triglycerides and phospholipids with increased levels of proinflammatory lysoPC (12). In a validation study of a different cohort, cord-blood PCs were again decreased in children who later developed type 1 diabetes (13). It appears, therefore, that lipidomic changes at the earliest stage predispose to diabetes, and altered levels of PC, a major substrate for oxidative metabolism during tissue repair, implicate increased bioenergetics.

Taken in the round, these observations suggest that cell activation and increased bioenergetic demands developing early in life in a stressed environment predispose to autoimmune diabetes. The new results from Wang et al. (17) regarding NET formation are consistent with a role for innate effector cells, including neutrophils, in that process. It should be acknowledged, however, that the extent to which NETs contribute to the origins of human type 1 diabetes is uncertain. None of which is to deny a major role for adaptive immunity—but in a field dominated by T-cell immunologists, it may be time to look the other way.