Purchase this article with an account.
Ishani Barai, Benjamin Davis, Li Guo, Nivedita Ravindran, Jessie Van houcke, Lieve K M Moons, Lies De Groef, M Francesca Cordeiro; Characterising retinal microglia: A novel three-dimensional spatial statistics approach. Invest. Ophthalmol. Vis. Sci. 2019;60(9):4009.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Retinal microglia have a significant role in the pathology of retinal disease. Our previous work (Davis et al, 2017) demonstrated a novel technique of using two-dimensional spatial statistics to enable detailed extraction of information from whole-retinal histology. However, the retina is a three-dimensional (3D) tissue with microglia residing in the ganglion cell layer (GL), inner plexiform layer (IPL) and outer plexiform layer (OPL). This study aimed to extend the previous approach to extract information from the retinal layer in which each microglia resided.
To visualize retinal microglia, Tg(coro1a:eGFP) reporter fishwere used. Whole-retinal mounts were harvested 7 days post optic nerve crush (n=9) (ONC) and age matched to naïve retina (n=6). Image analysis was performed, adapting our previously published algorithm (ImageJ), to extract microglia centres of mass (x,y,z). A second script (R programming) corrected z-positions, and K-means clustering automatically segmented the microglial population into the three retinal layers of GL, IPL and OPL. Cell percentage (%), cell density, nearest neighbour distance (NND) and regularity index (RI) were measured in each layer. Matched two-way ANOVA with Bonferroni’s multiple comparisons test (GraphPad Prism) was used to assess difference of the parameters between naïve and injury states (p< 0.05).
The following statistically significant changes were detected in the zebrafish injury state: i) increase in cell % in the IPL and decrease in cell % in the OPL; ii) increase in cell density in the GL and IPL and iii) decrease in NND within the IPL. These preliminary data suggest proliferation of microglia cells in the IPL with the population becoming densely packed, possibly due to activation of microglia near the optic nerve following compromise and subsequent re-innervation. Further time-points post-injury need to be assessed.
We have shown that automated evaluation of the entire microglial population in distinct retinal layers can be used to successfully segment 3D information from Coro1a-labelled whole-retina mounts of zebrafish. We have also demonstrated utility of this technique in rodent ocular hypertension model. This underscores the ability to extract this information in naïve/injury models from multiple species, and the potential of this analysis to provide novel insights into the mechanisms underlying retinal disease processes.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.
This PDF is available to Subscribers Only