Statistics from Altmetric.com
Sticky blood or sticky vessels, or both
There has been considerable interest in the past few years in the early pathological events that lead to vascular occlusion in diabetic retinopathy. The finding of increased leucostasis (leucocytes attached to the endothelial wall) is a common pathological event in both human disease1,2 and in experimental models.3,4 The heightened leucocyte/endothelial interaction induced by hyperglycaemia occurs very early in the diabetic process and, as a result, endothelial dysfunction and subsequent apoptosis occur. Although there is some reserve in terms of endothelial division and proliferation, this becomes exhausted with time, leading to the appearance of acellular capillary tubes, the pathological hallmark of disease.5 These tubes do not support blood flow and retinal ischaemia supervenes. Presumably there is pre-capillary arteriolar thrombosis due to the loss of the anti-thrombogenic endothelial lining, but pathological evidence for this is scant.6 The question arises as to whether this process is mediated by changes in leucocytes (sticky blood), changes in the endothelial surface (sticky vessels), or both.
Supporting the “sticky blood” hypothesis are the findings that leucocytes (particularly polymorphonuclear cells) are less deformable in diabetes,7 that there is upregulation of integrins (ligands for vascular adhesion molecules) on their surface,8 and that they adhere more strongly to cultured endothelial cells in vitro in both static assays and under conditions of flow.9 Furthermore, in vivo treatment with anti-integrin monoclonal antibodies experimentally markedly reduces leucostasis.4 Recent work has shown that the glycosylating enzyme, core 2 GlcNAc transferase, activity is upregulated in leucocytes derived from diabetic patients and that this upregulation positively correlates with the presence of established retinopathy. This enzyme causes post-translational modifications of the O-glycans on the cell membrane, some of which are molecules involved with endothelial interaction, and specific blockade of the phosphorylation of this enzyme abrogated the heightened leucocyte-endothelial interaction in vitro.10,11
Support for abnormalities in the endothelium (sticky vessel) comes from the immunohistochemical demonstration of increased adhesion molecule expression (particularly intercellular adhesion molecule-1, ICAM-1) on the cell surface in response to hyperglycaemia in both human1 and experimental models,3 and that the observed leucostasis in experimental models could be partially prevented by treatment with a monoclonal antibody against ICAM-1. Moreover, ICAM-1 knockout mice made diabetic do not develop the expected retinal vascular changes evident in their wild type counterparts.5 The cause of this heightened expression is not known but may be from the action of endogenous vascular endothelial growth factor (VEGF) acting through a nitric oxide pathway since treatment with aptamers to VEGF reduced retinal leucostasis and blood-retinal barrier breakdown in experimental models.12,13
Abnormalities of both blood and vessel wall probably contribute to the process of vaso-obliteration in diabetes
What occurs as a result of this leucocyte-endothelial interaction is not certain. Under normal physiological conditions, the interaction is necessary to allow the immune system to sample the microvascular environment with the leucocyte rapidly returning to the circulation if the correct sequential expression of adhesion molecules to allow firm tethering to the endothelial surface is not found. In diabetes, the reduced deformability and increased stickiness of the leucocyte may be sufficient to physically block capillary tubes.14 More likely is that the prolonged interaction leads to endothelial dysfunction and apoptosis, eventually leading to complete loss of the capillary lining.15,16
One of the principal hypotheses on which the above scenario is based—namely, that of increased adhesion molecule expression in the diabetic retinal vasculature, is challenged in this issue of the BJO. In a carefully performed immunohistochemical study of human specimens, Hughes et al (p 566) show, in contrast with previous studies, that there is little difference in the expression of ICAM-1 between normal and diabetic retinas. Furthermore, their studies also show diffuse ICAM-1 staining of neural retina; this was increased in the diabetic specimens and correlated with local breakdown of the blood-retinal barrier. The reason for the disparity in their results and those of previous studies is comprehensively discussed and, although the diabetic history of their specimens is not known, the evidence is sufficiently forceful to prompt a re-evaluation of the subject.
It seems likely that abnormalities of both blood and vessel wall contribute to the process of vaso-obliteration in diabetes although the relative importance of either still remains to be determined. Further work examining the role of reactive oxygen species in the promotion of a prothrombotic phenotype in the microvasculature and the contribution of leucocyte-platelet interactions to microthrombus formation needs to be done. Understanding of the pathophysiological basis of disease is fundamental to the formulation of new treatments. Already, considerable interest has been shown in the use of anti-inflammatory drugs in the amelioration of diabetic retinopathy17 and early reports of the use of intravitreal steroids are promising. Certainly, there is now an increasing rationale for the use of specific protein kinase C inhibitors.11,18,19 Current advances in knowledge of the pathology of the disease are likely to throw up further candidates in the near future.
Note in Proof
Sticky blood or sticky vessels, or both