The neurovascular unit in early diabetic retinopathy: cellular changes and an emerging role for microglia

Abstract

In the developed world the leading cause of blindness in working age adults is diabetic retinopathy (DR). The risk of vision loss due to diabetes increases with disease duration, and most frequently affects those who have lived with diabetes for one to two decades. For individuals with diabetes, there is a need to develop therapeutic interventions that are capable of halting or delaying degeneration of the retina as early as possible to preserve vision throughout the life of the patient.

DR ultimately leads to vision loss via the death of retinal neurons, which occurs after severe dysregulation of the retinal vasculature. Retinal oedema and/or the proliferation of blood vessels both signal the advancement to a high-risk stage of vision-threatening DR, and our understanding and management of these late stages is comprehensive. However, the initial cellular changes that occur soon after the onset of diabetes and contribute to this end-stage vascular pathology are not completely understood. These include changes to the neurovascular unit: retinal blood flow, neuronal function, and glial / microglial cell reactivity. It is these early signs that are investigated in this thesis, for it is these changes at this early timepoint that hold the most promise for intervention to prevent onset of clinical vasculopathy.

Using an animal model of diabetic retinopathy, the relative contribution of neurons, blood vessels, and glia/microglia to early DR pathology was assessed at two early time-points after diabetes onset. In vivo measurement of retinal blood flow, blood vessel structure, and neuronal function was undertaken; in conjunction with histological analysis of pericytes, Müller glia, astrocytes, and microglia, as well as the interactions of these cells with the vasculature. Additionally, microglia were investigated for their potential to assist in vaso-regulation using ex vivo live cell imaging and single cell population RNA-Seq.

Functional measurements revealed that retinal blood flow and retinal ganglion cell function were both reduced after 4 weeks of hyperglycaemia. In depth analysis of retinal vessel structure showed capillaries to be constricted at this time-point and were also unresponsive to hyperoxic challenge. Neuronal function was further degraded by 12 weeks of diabetes, although there were no signs of cell death. Glial cells showed no difference in structure, or signs of gliosis after 4 weeks of diabetes, and did not alter contact with the vasculature. Microglia were found to be in a resting phenotype and increased their contact with retinal capillaries after 4 weeks of diabetes. The microglia specific fractalkine signalling axis was shown to initiate vasoconstriction on retinal blood vessels, a response which was absent in the diabetic retina. Vaso-active genes were identified in microglial populations, several of which were upregulated in the diabetic retinae.

In summary there are two main findings to report. Firstly, this project expands our understanding of the chronopathology of cell-specific dysfunction in the early stages of DR, and indicates that capillary constriction may be a new biomarker for early stage disease. Secondly, the discovery that microglia contribute to blood flow maintenance opens new avenues for research into vascular diseases, including DR, which may unlock novel therapeutic targets.