Deciphering the Connection Between Microvascular Damage and Neurodegeneration in Early Diabetic Retinopathy

Abstract

Diabetic retinopathy (DR), a common diabetes complication leading to vision loss, presents early clinical signs linked to retinal vasculature damage, affecting the neural retina at advanced stages. However, vascular changes and potential effects on neural cells before clinical diagnosis of DR are less well understood. To study the earliest stages of DR, we performed histological phenotyping and quantitative analysis on postmortem retinas from 10 donors with diabetes and without signs of DR (e.g., microaneurysms, hemorrhages), plus three control eyes and one donor eye with DR. We focused on capillary loss in the deeper vascular plexus (DVP) and superficial vascular plexus (SVP), and on neural retina effects. The eye with advanced DR had profound vascular and neural damage, whereas those of the 10 randomly selected donors with diabetes appeared superficially normal. The SVP was indistinguishable from those of the control eyes. In contrast, more than half of the retinas from donors with diabetes had capillary dropout in the DVP and increased capillary diameter. However, we could not detect any localized neural cell loss in the vicinity of dropout capillaries. Instead, we observed a subtle pan-retinal loss of inner nuclear layer cells in all diabetes cases (P < 0.05), independent of microvascular damage. In conclusion, our findings demonstrate a novel histological biomarker for early-stage diabetes-related damage in the human postmortem retina; the biomarker is common in people with diabetes before clinical DR diagnosis. Furthermore, the mismatch between capillary dropout and neural loss leads us to question the notion of microvascular loss directly causing neurodegeneration at the earliest stages of DR, so diabetes may affect the two readouts independently.

Figure 1

Endothelial cells in retinal whole mount. A: Example of a retinal whole mount stained with UEA revealing endothelial cells (optic disc and sclera have been removed). B and C: Zoomed-in images of boxes shown in A. Capillary-free regions can be seen in the vicinity of arteries (arrow in B and C), which is normal. In contrast, localized small capillary-free patches in the periphery (arrowheads in A and C) are indicative of retinal vasculature pathology. FAZ, foveal avascular zone. Scale bar = 1 mm.

Figure 2

Immunostaining showing ghost vessels in the retina from a donor with diabetes. A–C: Immunohistochemistry from a region with acellular capillaries (arrowheads) showing collagen IV–stained basement membrane (green in A and B) and UEA-labeled endothelial cells (red in A and C). Arrowheads indicate ghost vessels, where endothelial cells have disappeared and only basement membrane remains. Scale bar = 50 μm.

Figure 3

Vessel loss at different retinal vascular plexuses. A: Total capillary loss in control eyes, subgroups of diabetes, and DR eyes. Retinae from people with diabetes presented a mildly, but statistically significant, higher incidence of capillary loss than control eyes. DR retina presented an overwhelmingly higher capillary dropout incidence than any other groups. The dashed line shows average capillary loss incidence in the diabetes group as a whole (3.82%). At least 13,500 vessels were assessed for each donor. B: Capillaries became narrower after losing endothelial cells. A slight increase in the size of normal capillaries was seen in the DDO subgroup, but the size of acellular capillaries was consistent among non-DR groups. At least 15 capillaries were measured from each donor; ≥52 capillaries were measured for each group. C: Intragroup differences in vessel loss of SVP and DVP. Vessel loss from each donor at SVP was never higher than that of DVP in all cases examined. On average, only the DDO subgroup and DR group had higher incidence of vessel loss. (At least 9,000 vessels were examined for each plexus for each donor.) D: Intergroup differences in vessel loss of SVP and DVP. The control group and DNDO subgroup showed no interplexus difference, which was significant in the DDO subgroup and DR, with DVP having a considerably elevated proportion of vessel loss. E: A linear regression model plotting vessel loss in the DVP against SVP in all donors reveals a strong linear relationship (R² = 0.972; P < 0.0001). The control group and the DNDO subgroup formed a clearly separate cluster from the DDO subgroup. DR locates at a distinct region that is farther away for the other groups. Control group, n = 3; DNDO subgroup, n = 4; DDO subgroup, n = 6, and DR, n = 1. Results are presented as mean ± SD. Box and whisker plots show the mean ± SD. Statistical significance was tested by one-way ANOVA with post hoc Dunnett’s analysis (A, B, D) or two-tailed Student t test (C). *P < 0.05, **P < 0.005, ***P < 0.001. BM, basement membrane.

Figure 4

Gliovascular units in diabetic retina. A–D: Immunohistochemistry using antibodies against widely recognized glial marker GFAP (A), glutamine synthetase (GS) (B), CRALBP (C), and AQP4 (D) from consecutive sections of eyes from donors from the DDO subgroup (green), counterstained against collagen IV (white) and UEA (red). Merged images of the three channels are in the first and last (MERGE) column. Arrowheads point at acellular capillaries and arrows point to normal capillaries. Glial interface formed by retinal astrocytes was intact regardless of vascular dropout (A). GS and CRALBP are highly expressed in Müller cell endfeet at the inner limiting membrane and the outer limiting membrane, and less involved in the gliovascular interface (B and C). AQP4 expression is especially concentrated around glial cell endfeet around vessels, which remained present around acellular capillaries (D, arrows), but staining intensity was notably reduced (D, arrowheads). Scale bar = 50 μm in images at lower magnification and 25 μm in zoom-in images.

Figure 5

DVP capillary zone of influence. A: The NND was calculated for each donor and then normalized to the height of the nuclei of RPE cells (>10 RPE cells were measured per donor and the average value was used); thus, results express fold change to the reference value. No difference was found between any two groups, meaning that the NND is consistent across groups. B: Frequency plot of the NND reveals an overlapping distribution pattern in the control group and two diabetes subgroups, and a lower peak of the DR NND distribution was noticed. The dashed line shows the overall normalized mean NND of 4.65. C: The RI, calculated as mean ± SD, tends to decrease with increased severity of capillary loss. No difference was found among the control group and subgroups with diabetes, whereas RI of DR was significantly different from each other group. D: The NND was plotted against the distance from the optic disc–fovea axis along the superior-inferior axis. No linear relationship could be found in any group (DNDO subgroup and DR not shown). Control group, n = 3; DNDO subgroup, n = 4; DDO subgroup, n = 5; and DR, n = 1. Results are presented as mean ± SD. *P < 0.05. Statistical significance was tested by one-way ANOVA with Tukey post hoc comparison (A, C). P(x), probability distribution of normalized NND.

Figure 6

Identifying overall and subpopulation cell loss in the INL. A, C, and E: Confocal microscopy image showing the measurements of INL cell nuclei within the zone of influence of DVP capillaries (circle, A). Examples of normal capillaries (magenta counting marks) and nonperfused capillaries (green counting marks) are shown. Horizontal (C) and bipolar cells (E) were visualized using antibodies against PV and PKCα, respectively. B, D, and F: Quantification of total cells (B) and interneurons (D and F) within deeper capillary zones of influence. Normal capillaries are shown by empty boxes; nonperfused capillaries are shown by filled boxes. Control group, n = 3; DNDO subgroup, n = 4; DDO subgroup, n = 5; and DR, n = 1. Box represents 25th, median, and 75th quartiles; + indicates the mean value. Whiskers represent maximum and minimum. *P < 0.05, ***P < 0.0001. Statistical significance was tested by unpaired two-tailed Student test (intragroup differences) or one-way ANOVA with Dunnett post hoc comparison with control (intergroup differences). Scale bar = 25 μm.