Figure 1.
Dendritic Morphology of Retinal Microglia.
A. Retinal section from an adult mouse (2–3 months of age) showing the laminar distribution of retinal microglia (green) in the inner half of the retina, including the ganglion cell layer (GCL), inner plexiform layer (IPL), inner nuclear layer (INL), and outer plexiform layer (OPL). B. Dendritic arbors of individual retinal microglia, as seen in a confocal image of a flat-mounted retinal explant captured in the horizontal plane of the retina, have symmetrically oriented processes that do not overlap with those in neighboring cells. C. Analysis of the dendritic morphology of microglia in ex vivo retinal explants from 2-dimensional image projections of confocal images. Three morphological parameters were analyzed: 1) Dendritic tree area (area circumscribed by the polygonal object defined by connecting the outer points of the dendritic ramified arbor, shadowed), 2) Total dendritic length (sum of all dendritic segments identified in a skeletonization of the arbor), and 3) Total branch point number (sum of branch points identified in a skeletonized rendition of the arbor, points in blue).
Figure 2.
Endogenous ionotropic glutamatergic neurotransmission positively regulate dendritic morphology and process dynamics of retinal microglia.
A. Confocal images of a representative retinal microglia cell under control conditions (left), in the presence of NBQX (10µM), a glutamate receptor antagonist of AMPA- and kainate-gated channels (middle), and during washout of NBQX (right). Microglial dendritic morphology is significantly reduced in size and complexity in the presence of ionotropic glutamatergic blockade with NBQX, and is partially reversed with washout with Ringer's solution. Scale bar = 20µm. B,C Quantification of morphological parameters demonstrate that all three morphological parameters are significantly decreased from baseline values in the presence of NBQX (10µM; n = 42 cells), as well as GYKI-52466 (100 µM; n = 48 cells), an antagonist of AMPA-gated channels. D. Similar recordings were also performed with the application of APV (100 µM; n = 29 cells), an antagonist of ionotropic NMDA-gated glutamatergic channels. All 3 morphological parameters were also significantly reduced, albeit to a smaller extent than seen with NBQX. E. Overall process velocities were also calculated from the above recordings, with NBQX (n = 44 cells), GYKI (n = 52 cells) and APV (n = 26 cells) applications significantly decreasing baseline process motility, with NBQX exerting a larger decrement compared to GYKI or APV. All significant changes (p<0.05) from baseline are indicated by an asterisk (*).
Figure 3.
Exogenous ionotropic glutamatergic neurotransmission increases retinal microglia dendritic morphology and process dynamics.
A. Confocal images of a representative retinal microglia cell ex vivo under control conditions (left), with the bath application of AMPA (100 µM)(middle), and during washout (right). Microglial dendritic morphology is significantly increased in size and complexity during stimulation with AMPA with partial recovery during washout. Scale bar = 20µm. B–C. Quantification of morphological parameters demonstrate that all three morphological parameters of dendritic tree area, total dendritic length, and total number of branch points are significantly increased in response to AMPA (100 µM; n = 33 cells) and kainate (100 µM; n = 42 cells) application. D. Microglia morphology however was relatively stable with the application of NMDA (100 µM; n = 54 cells) in the presence of glycine (10 µM). E. Overall, process velocities were increased in response to AMPA and kainate, but remained unchanged when NMDA was applied. All significant changes (p<0.05) from baseline are indicated by an asterisk (*).
Figure 4.
Ionotropic GABAergic neurotransmission decreases retinal microglia dendritic morphology and process dynamics.
A. Confocal images of a representative retinal microglia cell under control conditions (left), in the presence of bicuculline (150 µM)(middle), an antagonist of ionotropic GABAA receptors, and during washout (right). Microglial dendritic morphology is slightly increased in size during stimulation with bicuculline, with partial recovery during washout. Scale bar = 20µm. B. Dendritic tree area and total dendritic length was slightly but significantly increased in the presence of bicuculline, while branch point number was unchanged (n = 53 cells). C. Conversely, application of GABA (1mM; n = 23 cells) exerted slight increases in total dendritic length and branch point number, but did not change dendritic tree area significantly. D. Overall process velocity was increased with GABAergic blockade with bicuculline and decreased with GABA application. All significant changes (p<0.05) are indicated by an asterisk (*).
Figure 5.
Extracellular ATP increases dendritic morphology and process dynamics of retinal microglia.
A. Confocal images of a representative retinal microglia cell under control conditions (left), with the bath application of ATP (1mM) (middle), and during washout (right). Microglial dendritic morphology is significantly increased in size and complexity with ATP. Scale bar = 20µm. B. All morphological parameters were significantly increased in the presence of ATP, (n = 58 cells). C–D. Conversely, morphological parameters were decreased in the presence of suramin (100 µM; n = 25 cells), broad-spectrum antagonist of P2 receptors, and also, to a lesser extent, with apyrase (10U/ml; n = 38 cells), which catalyzes the hydrolysis of extracellular ATP. E. Overall process velocity was increased with ATP application and decreased in the presence of suramin and apyrase. All significant changes (p<0.05) from baseline are indicated by an asterisk (*).
Figure 6.
Resting retinal microglia lack direct responses to glutamatergic or GABAergic agonists.
A. Representative whole cell recordings (n = 8) from green fluorescent protein (GFP)-positive microglia in retinal slices from CX3CR1+/GFP transgenic mice. Under voltage clamp configuration, cells were held at −70 mV and a series of voltage steps (400 ms, 20 mV steps from −40 mV) were applied. Microglia cells did not display delayed rectified K+ currents in response that would indicate an active state. B. Under current clamp configuration, a series of currents (40 ms; −20, 20, 40, 60, 80 pA) were injected into cells. The large membrane potential responses reflect the high input resistance of microglia cells. C. Representative traces showing whole cell currents in retinal microglia. Microglia were voltage clamped at −20 mV, near the resting membrane potential. Local application of ATP by pressurized puff (1 s, 10 mM, n = 4) elicited a large inward current that rose and decayed quickly over several seconds. GABA (1mM, n = 4), Glutamate (1mM, n = 4), or AMPA (100 mM, n = 3), failed to elicit any response. D. Immunolabelling of AMPA receptors, GluR2 and GluR3 (red) in the retina of a CX3CR1+/GFP mouse. The distribution of GluR2 and GluR3 are visible as discrete puncta in the inner plexiform layer (left) that are located around microglial processes (green) but are not visibly colocalized on microglial processes (expanded inset at higher magnification, right). E. Quantitative analysis of colocalization demonstrates that GluR2/3-positive puncta were not located on microglial processes. Multiple square insets (measuring 23 by 23 pixels) were sampled from the image; each inset was centered on a single GluR2/3-positive punctum in the vicinity of the microglial dendritic process (left). An averaged plot of intensity versus pixel position was plotted for the red channel (GluR2/3-localization, middle) and the green channel (microglial GFP-localization, right). The peak in the red channel (middle) reflects the discrete puncta nature of the Glu2/3 distribution, while the caldera-like pattern in the green channel indicates that the location of microglial processes are anti-correlated with that of GluR2/3, signifying that these AMPA receptors are not colocalized with microglial processes.
Figure 7.
ATP Application Rescues Microglial Morphology and Motility from the Negative Effects of Glutamatergic Blockade.
A. Confocal images of a retinal microglial cell ex vivo under control conditions (left), in the presence of NBQX (10µM) and ATP (1mM) (middle), and during washout with Ringer's solution (right), demonstrating a net increase of dendritic structure, as is typically seen with ATP alone. B. Similar images showing the effect of sequential addition of NBQX (10µM) alone, followed by a combination of NBQX (10µM) and ATP (1mM), demonstrating that the loss of dendritic structure induced glutamatergic blockade can be reversed by concurrent application of ATP. C. Morphologic parameters of dendritic tree area, total dendritic length, and branch point number, were all increased with concurrent NBQX (10µM) and ATP (1mM) application. (n = 36 cells) D. Overall process velocity was also increased with concurrent NBQX (10µM) and ATP (1mM) application, to a similar degree seen with ATP alone (1mM), and in the opposite direction to the changes induced with the application of NBQX alone (10µM). All significant changes (p<0.05) from baseline are indicated by an asterisk (*).
Figure 8.
P2 receptor blockade by suramin neutralizes the positive effects of exogenous AMPA application on microglial morphology and process motility.
A. Confocal images of a retinal microglial cell ex vivo under control conditions (left), in the presence of AMPA (1mM) and Suramin (100µM) (middle), and during washout with ringers (right), demonstrating that the marked increase in microglia dendritic size and complexity is neutralized in the presence of suramin. B. Morphologic parameters of dendritic tree area and total dendritic length were unchanged, and branch point number slightly decreased with concurrent AMPA (1mM) and Suramin (100µM) application; (n = 56 cells). C. Overall process velocity was unchanged with concurrent AMPA (1mM) and Suramin (100µM) application, as compared to the increase induced with the application of AMPA alone (1mM). All significant changes (p<0.05) from baseline are indicated by an asterisk (*).
Figure 9.
Pannexin-1 hemichannel blockade by probenecid decreases dendritic morphology and process dynamics of retinal microglia.
A. Microglial morphologic parameters of dendritic tree area, total dendritic length, and branch point number were significantly decreased by the application of probenecid (1mM), an antagonist of pannexin-1 hemichannels; (n = 46 cells). B. Morphological parameters were similarly decreased with concurrent AMPA (1 mM) and probenecid (5mM) application, indicating that AMPA-induced ATP_release may be mediated significantly by pannexin-1 hemichannels. C. Overall microglial process velocity was similarly decreased with p probenecid (5mM) alone and with concurrent AMPA (1mM) and probenecid (5mM). All significant changes (p<0.05) from baseline are indicated by an asterisk (*).