Comparative evaluation of methods for estimating retinal ganglion cell loss in retinal sections and wholemounts.

Ben Mead,Ann Logan,Wendy Leadbeater,Ben A Scheven,Adam Thompson,Martin Berry,Tudor C Badea

doi:10.1371/journal.pone.0110612

Abstract

To investigate the reliability of different methods of quantifying retinal ganglion cells (RGCs) in rat retinal sections and wholemounts from eyes with either intact optic nerves or those axotomised after optic nerve crush (ONC). Adult rats received a unilateral ONC and after 21 days the numbers of Brn3a+, βIII-tubulin+ and Islet-1+ RGCs were quantified in either retinal radial sections or wholemounts in which FluoroGold (FG) was injected 48 h before harvesting. Phenotypic antibody markers were used to distinguish RGCs from astrocytes, macrophages/microglia and amacrine cells. In wholemounted retinae, counts of FG+ and Brn3a+ RGCs were of similar magnitude in eyes with intact optic nerves and were similarly reduced after ONC. Larger differences in RGC number were detected between intact and ONC groups when images were taken closer to the optic nerve head. In radial sections, Brn3a did not stain astrocytes, macrophages/microglia or amacrine cells, whereas βIII-tubulin and Islet-1 did localize to amacrine cells as well as RGCs. The numbers of βIII-tubulin+ RGCs was greater than Brn3a+ RGCs, both in retinae from eyes with intact optic nerves and eyes 21 days after ONC. Islet-1 staining also overestimated the number of RGCs compared to Brn3a, but only after ONC. Estimates of RGC loss were similar in Brn3a-stained radial retinal sections compared to both Brn3a-stained wholemounts and retinal wholemounts in which RGCs were backfilled with FG, with sections having the added advantage of reducing experimental animal usage.

Highlights

Retinal ganglion cells (RGCs) populate the ganglion cell layer (GCL) of the retina, in some species a few are found in the inner nuclear layer (INL; displaced RGCs) [1]
Accurate estimates of the total number of RGCs are compromised by counting errors engendered by: (1), the presence of astrocytes and displaced amacrine cells, estimated to contribute up to 50% of cells in the GCL [5,6] and which co-localize with many markers used to identify RGCs [7,8,9]; (2), RGCs displaced into the INL [1]; and (3), a progressive decrement in RGC density from the centre of the retina to the periphery [10]
In Group 2, bIII-tubulin, Islet-1- and Brn3a-stained RGCs were counted in radial sections of the retinae through the optic disk and double stained for astrocytes using glial fibrillary acidic protein (GFAP) amacrine cells using syntaxin-1 [29,30], and macrophages/ microglia using ED1 (Table 2)

Summary

Introduction

Retinal ganglion cells (RGCs) populate the ganglion cell layer (GCL) of the retina, in some species a few are found in the inner nuclear layer (INL; displaced RGCs) [1]. Accurate estimates of the total number of RGCs are compromised by counting errors engendered by: (1), the presence of astrocytes and displaced amacrine cells, estimated to contribute up to 50% of cells in the GCL [5,6] and which co-localize with many markers used to identify RGCs [7,8,9]; (2), RGCs displaced into the INL [1]; and (3), a progressive decrement in RGC density from the centre of the retina to the periphery [10]. Total absolute counts of RGC number is possible [10] yet may be erroneous if calculated from samples in which the spatial differences in RGC density are not taken into account when correcting for total retinal area, or if the retina is not dissected accurately. In RGC viability studies it is generally accepted that retinal wholemount and sectional quantitative data give acceptable estimates of percentage differences in RGC numbers between treatment and control groups

Methods

Results

Conclusion