Midget Retinal Ganglion Cell Dendritic and Mitochondrial Degeneration Is an Early Feature of Human Glaucoma
Item in Clipboard
Midget Retinal Ganglion Cell Dendritic and Mitochondrial Degeneration Is an Early Feature of Human Glaucoma
Glaucoma is characterized by the progressive dysfunction and loss of retinal ganglion cells. However, the earliest degenerative events that occur in human glaucoma are relatively unknown. Work in animal models has demonstrated that retinal ganglion cell dendrites remodel and atrophy prior to the loss of the cell soma. Whether this occurs in human glaucoma has yet to be elucidated. Serial block face scanning electron microscopy is well established as a method to determine neuronal connectivity at high resolution but so far has only been performed in normal retina from animal models. To assess the structure-function relationship of early human glaucomatous neurodegeneration, regions of inner retina assessed to have none-to-moderate loss of retinal ganglion cell number were processed using serial block face scanning electron microscopy (
n = 4 normal retinas, n = 4 glaucoma retinas). This allowed detailed 3D reconstruction of retinal ganglion cells and their intracellular components at a nanometre scale. In our datasets, retinal ganglion cell dendrites degenerate early in human glaucoma, with remodelling and redistribution of the mitochondria. We assessed the relationship between visual sensitivity and retinal ganglion cell density and discovered that this only partially conformed to predicted models of structure-function relationships, which may be affected by these early neurodegenerative changes. In this study, human glaucomatous retinal ganglion cells demonstrate compartmentalized degenerative changes as observed in animal models. Importantly, in these models, many of these changes have been demonstrated to be reversible, increasing the likelihood of translation to viable therapies for human glaucoma.
dendrite; electron microscopy; glaucoma; mitochondria; retinal ganglion cell.
© The Author(s) (2019). Published by Oxford University Press on behalf of the Guarantors of Brain.
Retinal ganglion cell loss assessed in relation to visual field deficits. Human donor eyes from glaucoma donors ( n = 4) complete with visual field tests conducted prior to death ( A; upper row) were analysed in comparison to control donor eyes ( n = 6). Whole retinas were imaged by two-photon microscopy, and cell counts made in regions of the retina corresponding to visual field test locations ( A; lower row). Regions were numbered 1–72 beginning superior-temporally, with regions corresponding to visual field test locations highlighted ( purple shading; n = 54 locations). Following imaging, 11 regions were dissected out and processed for SBFSEM ( magenta boxes). ( B) Retinal ganglion cell densities were estimated for each test location from z-stack counts within an en face area of 350 µm 2 and expressed as cells/mm 2. Density plots for each glaucomatous retina ( B; upper row) with corresponding percentage change from average control density at each region ( B; lower row). Region locations are inverted along the superior to inferior axis to correspond to the visual field plots (as the superior retina views the inferior visual field and vice versa). Retinal ganglion cell density change against eccentricity was plotted for control ( C; left panel) and glaucoma eyes ( C; right panel). The relationship between retinal ganglion cell density [expressed as log 10 cells/area of the stimulus (Goldman III)] and visual sensitivity is plotted in D. There was no correlation when retinas were grouped as shown by linear regression (‘Spearman’s rho’, r = 0.23, P < 0.001; D). Individual retinas show high variation among eyes ( D). The ‘hockey stick’ model fitted by Swanson et al. (2004) to a plot of normal visual field sensitivity, corrected to a 34-year old (Heijl et al., 1987) against normal retinal ganglion cell counts (Curcio and Allen, 1990), is superimposed on the data from the current study in D ( red line). Individual retinas show a similar relationship when retinal ganglion cell density is high but a greater than expected drop-off in sensitivity when cell density is low. Retinal ganglion cell abbreviated to RGC in B– D. For A, F = fovea, ON = optic nerve; retinal orientation identified by N = nasal, I = inferior, S = superior, T = temporal. Retinal ganglion cell density scales through low ( purple) to high ( yellow) ( B; upper row). Retinal ganglion cell density change scales through +100% of control average ( purple) to −100% of control average ( yellow) ( B; lower row). Black regions in B represent regions around the optic nerve where cell counts were not taken.
Dendritic loss is a feature of human glaucoma. Retinal samples [ n = 5 from glaucoma (four eyes) and n = 6 from controls (four eyes) corresponding to visual field test locations were dissected and prepared for SBFSEM]. Volumetric EM data were generated ( A) from 79 × 79 × 100 µm tissue cube (19.2 × 19.2 × 100 nm resolution). A representative single slice ( x– y plane) is shown in B (cropped in x), showing the retinal layers analysed. Retinal ganglion cells were identified, and their dendrites were reconstructed within the IPL. A cartoon of a Golgi-stained midget cell is shown for comparison in C alongside a reconstruction from the current SBFSEM data (for more examples see Kolb and Dekorver, 1991). Seven control ( D, upper panel) and five glaucomatous ( D, lower panel) retinal ganglion cells that met inclusion criteria were reconstructed. Dendrites ( yellow) and synapses ( cyan) are shown, with the dendrite origin at the soma indicated ( white arrowheads). Analysis of dendrites demonstrated reduced dendritic branching ( E) and fewer and shorter secondary and tertiary dendrites in glaucoma, indicating the presence of dendritic atrophy ( F and G; primary dendrites n = 5 in glaucoma, n = 7 in controls; secondary dendrites n = 4 in glaucoma, n = 14 in control; tertiary dendrites n = 0 in glaucoma, n = 3 in control). * P < 0.05, NS = non-significant ( P > 0.05). For B, ILM = inner limiting membrane, NFL = nerve fibre layer, GCL = ganglion cell layer, IPL = inner plexiform layer, INL = inner nuclear layer. Scale bar = 5 μm for D.
Mitochondrial morphometry and distribution are altered in human glaucoma. Organelles from five glaucomatous ( n = 4 eyes) and seven control retinal ganglion cells ( n = 4 eyes) were reconstructed in FIJI (ImageJ). Mitochondria (mitos, magenta; identified by the presence of cristae; n = 198 in controls, n = 55 in glaucoma across all cells analysed) and vacuoles ( green, double membranous structures devoid of cristae and electron dense material; n = 273 in controls, n = 326 in glaucoma across all cells analysed) were reconstructed within retinal ganglion cell dendrites ( yellow). Representative EM images for control ( A) and glaucoma ( C) and reconstructions ( B and D) are shown. The number of mitochondria per micrometre of dendrite for individual retinal ganglion cells was reduced in glaucoma ( E), and mitochondria occupied a reduced percentage of dendritic volume in glaucoma compared with controls ( F). Sholl analysis of mitochondria demonstrates that the mitochondrial distribution across dendrites was altered in glaucoma when expressed as a distribution statistic for primary dendrites ( G; ‘Fisher’s exact’ test; for individual dendrites, n displayed in figure), secondary dendrites ( H; ‘Fisher’s exact’ test) but not tertiary dendrites ( I; ‘Fisher’s exact’ test). When expressed as Sholl AUC ( J), mitochondrial distribution is significantly changed in secondary dendrites in glaucoma compared with controls but not for primary and tertiary dendrites. NND analysis demonstrated no significant change in the proximity of mitochondria to one another in primary dendrites ( K; mitochondria n = 39 in controls, n = 38 in glaucoma) but a significant isolation in secondary dendrites [distances to the first, second and third nearest neighbours ( k = 1–3) increased; L; mitochondria n = 151 in controls, n = 17 in glaucoma]. There were no observable mitochondria in tertiary dendrites in retinal ganglion cells from glaucoma eyes ( M). Individual mitochondrial surface reconstructions were generated in Imaris (mitochondria n = 149 in controls, n = 55 in glaucoma), and representative images are shown for control and glaucomatous retinal ganglion cells ( N). Mitochondria demonstrated significantly reduced volumes in glaucoma ( O) and were more spherical compared with controls ( P). Oblate (rounded) and prolate (cigar-shaped) ellipticity was not significantly altered ( Q). Sholl analysis of vacuoles across the dendritic tree as a whole demonstrated a significant change in distribution ( R; ‘Fisher’s exact’ test) but not in the Sholl AUC ( S). The number of vacuoles per micrometre of dendrite was increased in glaucoma ( E), but the percentage of dendritic volume occupied by vacuoles was unchanged ( F). Nearest neighbour distances in vacuoles for the k = 3 nearest neighbours decreased significantly in glaucoma indicating an increased density of vacuoles ( T; vacuoles n = 273 in controls, n = 326 in glaucoma). * P < 0.05, ** P < 0.01, *** P < 0.001, NS = non-significant ( P > 0.05). Scale bar = 5 μm for B and D and 3 μm for N.
Morphology of single ganglion cells in the glaucomatous primate retina.
Invest Ophthalmol Vis Sci. 1998 Nov;39(12):2304-20.
Invest Ophthalmol Vis Sci. 1998.
Inhibition of the classical pathway of the complement cascade prevents early dendritic and synaptic degeneration in glaucoma.
Mol Neurodegener. 2016 Apr 6;11:26. doi: 10.1186/s13024-016-0091-6.
Mol Neurodegener. 2016.
27048300 Free PMC article.
Retinal ganglion cell dendrite pathology and synapse loss: Implications for glaucoma.
Prog Brain Res. 2015;220:199-216. doi: 10.1016/bs.pbr.2015.04.012. Epub 2015 Jun 30.
Prog Brain Res. 2015.
[Aiming for zero blindness].
Nippon Ganka Gakkai Zasshi. 2015 Mar;119(3):168-93; discussion 194.
Nippon Ganka Gakkai Zasshi. 2015.
Ahmad T, Aggarwal K, Pattnaik B, Mukherjee S, Sethi T, Tiwari BK., et al. Computational classification of mitochondrial shapes reflects stress and redox state. Cell Death Dis 2013; 4: e461.
Anantrasirichai N, Achim A, Morgan JE, Erchova I, Nicholson L. SVM-based texture classification in Optical Coherence Tomography. In: 2013 IEEE 10th International Symposium on Biomedical Imaging, 7–11 April 2013.
Artes PH, Iwase A, Ohno Y, Kitazawa Y, Chauhan BC. Properties of perimetric threshold estimates from full threshold, SITA standard, and SITA fast strategies. Invest Ophthalmol Vis Sci 2002; 43: 2654–9.
Balendra SI, Normando EM, Bloom PA, Cordeiro MF. Advances in retinal ganglion cell imaging. Eye 2015; 29: 1260–9.
Ball MJ, Lo P. Granulovacuolar degeneration in the ageing brain and in dementia. J Neuropathol Exp Neurol 1977; 36: 474–87.