In This Issue of Diabetes

A drug commonly used to treat asthma can inhibit early diabetic retinopathy in mice with diabetes, according to Bapputty et al. (p. 2004). As a result, they suggest that as the drug is already approved for use in humans, the animal study findings can be rapidly translated to humans for evaluation as a potential treatment for diabetic retinopathy. Using a series of experiments in mice, the authors compared the effects of montelukast on diabetic retinopathy progression in a diabetes mouse model and also control mice without diabetes. They then assessed various diabetes-related retinal pathology measures at time points up to 9 months. They found that diabetes did result in increased retinal vascular permeability compared with controls and that montelukast administered in drinking water prevented such an increase at 3 months. Notably, blood glucose levels remained elevated in the mice with diabetes for up to 9 months. At 9 months, the authors found that diabetes resulted in capillary degeneration and retinal ganglion cell loss. Montelukast meanwhile resulted in reductions in retinal capillary damage, superoxide generation, leukocyte adherence, and leukotriene generation that were all elevated because of diabetes. Long-term studies of retinal microstructure also exhibited a threefold increase in capillary degeneration in untreated mice with diabetes but showed that montelukast inhibited such damage whether administered immediately after induction of diabetes as a prevention approach or after 4.5 months (mid-diabetes duration) as a treatment approach. Commenting more widely, author Rose Gubitosi-Klug told Diabetes: “The historic safety profile of montelukast use by humans and the simplicity of once daily oral dosing make these animal model observations compelling for confirmation in human studies. Moreover, montelukast works on the underlying biological cascades that precede the development of retinal lesions, which would give physicians the opportunity to intervene earlier in at-risk patients. Hopefully, montelukast may offer critical adjunctive therapy to protect the retina, while efforts of patients and clinicians persist to improve glycemic control.”

Bapputty et al. Montelukast prevents early diabetic retinopathy in mice. Diabetes 2019;68:2004–2015

Delays in continuous glucose monitoring (CGM) device readings are likely due to fibrotic encapsulation of the implanted sensor rather than any delay in glucose entering interstitial fluids according to McClatchey et al. (p. 1892). The findings are the result of an investigation that set out to determine the exact reason for delays in CGM readings and centers on glucose measurements in mice. In initial in vitro experiments, they found that time to half-maximum CGM response was ∼35 seconds following exposure to glucose. However, in the experiments in the mice, they found that glucose responses took ∼24 min to register, as opposed to 2 min when blood glucose was measured directly. Using a fluorescent tracer version of glucose, they found much the same pattern—CGM measurements were delayed by over 20 min, whereas it took ∼6 min to appear in adipose tissue. Additional experiments involving histological analyses showed that substantial amounts of tissue material were deposited on the implanted CGM sensors and that diffusion of glucose into those tissues was substantially delayed relative to the transfer of glucose into interstitial fluid. Commenting further, author P. Mason McClatchey said: “As a long-time CGM user, I have often been frustrated by CGM delays. Fifteen to twenty minutes can make a big difference in blood glucose management. I mostly took the delay for granted though, because differences between blood and interstitial fluid seemed like a sensible explanation. That all changed when I started using a fluorescent glucose analog to study insulin resistance and saw glucose reaching interstitial fluid within seconds. So, I found a team of qualified collaborators at Vanderbilt, and we designed a series of experiments to pinpoint the source of CGM delays. The results were surprisingly clear-cut. Fibrous tissue formed around the sensor and was by far the biggest barrier to timely glucose detection. If anyone can develop a non-fibrotic CGM, we expect that this innovation would largely address CGM delays, and I for one would love to use it.”

McClatchey et al. Fibrotic encapsulation is the dominant source of continuous glucose monitor delays. Diabetes 2019;68:1892–1901

There is an association between cardiovascular autonomic neuropathy (CAN) in type 1 diabetes and intermediates of central carbon metabolism, according to Mathew et al. (p. 2035). Compared with healthy control subjects, the authors found that individuals with type 1 diabetes had perturbations in levels of several amino acids as well as tricarboxylic acid (TCA) cycle metabolites that are associated with a measure of worsening CAN. The findings come from a study involving 47 individuals with type 1 diabetes and 10 healthy age-matched control subjects who were subjected to a series of measures for CAN and targeted plasma metabolomics profiling at baseline and after 3 years. They found that compared with control subjects, individuals with type 1 diabetes had higher levels of several amino acids, including threonine, serine, and proline, and lower levels of fumarate from the TCA cycle. They also found that higher baseline levels of fumarate and lower levels of asparagine and glutamine correlated with a baseline measure of CAN. Glutamine and ornithine levels at baseline also reportedly correlated with the same measure at 3 years after adjusting for a series of factors such as baseline HbA_1c, blood glucose, BMI, and cholesterol, among others. They go on to discuss numerous mechanisms that might explain how the identified amino acids and TCA cycle metabolites affect CAN, before concluding that “significant changes in the anaplerotic flux into the TCA cycle could be the critical defect underlying CAN progression.” Quite correctly, they also point out that much larger trials with adequate power will be needed to confirm their observations. Commenting more widely, the corresponding authors told Diabetes: “Our study is the first to explore nutrient metabolic perturbations associated with type 1 diabetes and CAN progression. We plan to systematically study these metabolic pathways in model systems as therapeutic targets and as potential biomarkers in larger patient cohorts.”

Mathew et al. Impaired amino acid and TCA metabolism and cardiovascular autonomic neuropathy progression in type 1 diabetes. Diabetes 2019;68:2035–2044

A gene signature may exist that predicts the onset of type 1 diabetes prior to the appearance of any autoantibodies, according to Kallionpää et al. (p. 2024). The authors suggest that with further evaluation, the patterns they detected could be used for the early identification of children who are likely to develop the disease later on. Currently, the only way to identify individuals at risk is to test for autoantibodies indicative of an active autoimmune reaction, i.e., when immune tolerance is already broken. The study centers on mRNA sequencing of blood samples collected from seven young children, all of whom progressed to β-cell autoimmunity, and matched control subjects. The analysis included samples that were either fractionated to give CD4⁺ and CD8⁺ T cells or CD4^–CD8^– cells, or unfractionated peripheral blood mononuclear cell samples. All were collected longitudinally. They found numerous transcripts that were differentially expressed between the case and the control subjects, and by using various techniques, they identified gene expression clusters associated with type 1 diabetes autoimmunity. They also specifically identified interleukin 32 (IL32) as upregulated before the appearance of autoantibodies. Moving to single-cell RNA sequencing studies, they found that high levels of IL32 came from activated T cells and natural killer cells. They also found that IL32 could be induced via Coxsackie B virus in pancreatic islets and by cytokines in a β-cell line. They stress that a limitation of the study is that they only included seven cases and that the results will need to be validated and expanded in a larger cohort. Nevertheless, they suggest the outcomes can serve as a starting point for further studies, particularly with respect to whether IL32 can serve as a biomarker for type 1 diabetes. Commenting further, author Riitta Lahesmaa stated: “I hope that in the future we will be able to identify children who would rapidly progress to type 1 diabetes for preventive intervention or therapy.”

Kallionpää et al. Early detection of peripheral blood cell signature in children developing β-cell autoimmunity at a young age. Diabetes 2019;68:2024–2034