Müller glial cell-provided cellular light guidance through the vital guinea-pig retina

Abstract

In vertebrate eyes, images are projected onto an inverted retina where light passes all retinal layers on its way to the photoreceptor cells. Light scattering within this tissue should impair vision. We show that radial glial (Müller) cells in the living retina minimize intraretinal light scatter and conserve the diameter of a beam that hits a single Müller cell endfoot. Thus, light arrives at individual photoreceptors with high intensity. This leads to an optimized signal/noise ratio, which increases visual sensitivity and contrast. Moreover, we show that the ratio between Müller cells and cones-responsible for acute vision-is roughly 1. This suggests that high spatiotemporal resolution may be achieved by each cone receiving its part of the image via its individual Müller cell-light guide.

Figure 1

Setups to study the retinal light path. (a) Visualization of the beam path through a retinal cross section. The fluorescent sample, a retinal slice fixed on a nitrocellulose membrane stained with Mitotracker orange, was placed under a water-immersion objective (60×, NA 0.95) of an upright laser-scanning microscope (LSM 510; Zeiss). Confocal images of the evoked fluorescent light (red) were recorded using a confocal channel of the laser-scanning unit (laser 543 nm, green; dichroic mirrors, bandpass filter, detector). An external laser (laser 532 nm, green) was launched into a single mode fiber. The core of the fiber was placed in the focal plane of the objective in front of the vitread surface to illuminate a single Müller cell endfoot. The scattering of the laser light inside the sample (green) was detected by a second channel of the confocal imaging unit (dichroic mirrors, longpass filter, detector). A micropositioning stage (xy stage) moved the sample perpendicular to the light propagation. (b) Bidirectional imaging of retina whole-mount preparations combined with local light-transmission. The retina was spread (photoreceptor side up) onto the bottom of a chamber on the stage of an inverted laser-scanning microscope (LSM 510; Zeiss). The three-dimensional localization of dye-filled Müller cells that were pointed toward the objective (40×, N.A. 1.2) was recorded by confocal detection. Thereafter, the objective was replaced by one with lower numerical aperture (10×, NA 0.3) allowing the illumination of single Müller cell endfeet with laser light under a physiological angle (∼26°). An upright custom-built unit was used to image the opposite (outer) surface of the retina. The laser light transmitted from the photoreceptor outer segments was collected by an objective (63×, N.A. 0.95) and imaged by a convex lens and a plane mirror onto a charge-coupled device chip of a camera. Additionally, the same upright imaging unit allowed the recording of a transmission image of photoreceptor cells by wide-field illumination of the sample with parallel light (Xenon lamp) from below. (c) Spot-light illumination experiments on retinal whole-mount preparations. The retina was spread onto the bottom of a chamber with the photoreceptors pointing toward the objective (63×, NA 1.2) of an inverted microscope. Müller cell endfeet were illuminated by a laser through a single-mode fiber mounted vertically on top of the microscope stage. The glass fiber was moved by a piezo-actuator in steps of 1 μm over a distance of 100 μm. The transmitted laser light was detected by a camera (Zeiss). The fiber had a distance of ∼10 μm from the surface of the tissue.

Figure 2

Pathway of light through the vital retina. (a) Position of a single mode fiber (left) in front of the retinal surface (right) recorded by transmission microscopy. (b) The divergent light beam emanating from the core of the glass fiber was visualized by inserting the fiber into an agarose gel that caused scattering of the light (λ = 532 nm, green) in all directions and thus allowed light detection orthogonal to the beam. (c) Slice preparation of a living retina, monitored by using a 63×/0.95 water-immersion objective of an upright confocal microscope. The vital dye, Mitotracker orange (red), predominantly stained the Müller cells, the photoreceptor segments, and the plastic membrane (M). The positioning of the fiber core (dotted line at position y₀) between two Müller cell processes caused light scattering in both plexiform layers (IPL, inner plexiform layer; OPL, outer plexiform layer), as demonstrated by the x-line profiles (at position x₀) of the cellular fluorescence (I_FL, red) and the scattered light intensity (I_SL, green) along the IPL (d). The scattered light spots on the membrane (c and f) visualize the light that was transmitted by the retina. Their yellow appearance results from a merge of strong red fluorescence (vital dye sucked up by the membrane) and the green laser light scattering. (e) The x-line profile of I_SL along the membrane (at position x₁ in panel c) showed a rather wide-spread intensity distribution with a low maximum intensity. (f) In a position where the fiber and the Müller cell process are aligned to each other (along the dotted line at position y₀), the scattered beam pattern changed. The almost complete lack of light scattering in the Müller cell processes is represented by a reduced I_SL in the IPL (g) and OPL. This was accompanied by a narrowing light spot on the membrane with an increased maximum intensity (h).

Figure 3

Müller cells bias the transretinal light path. (a–h) Overlay of a fluorescent retinal slice (red) and the scattering of the laser light (green, yellow) applied by a thin glass fiber in front of the retinal surface (position indicated by a dotted line). When the retina was moved in equal steps of 2 μm (every second position is shown) along the optic fiber, the laser scattering in the retina (green) and the scattering of the transmitted light at the membrane (yellow) changed in dependence upon the presence of a Müller cell in the light path. (b and f) If the center of a Müller cell was positioned in front of the fiber core (indicated by MC in front of the retinal surface), the intensity of light scattering in the IPL was reduced. In addition, the transmitted light was confined to a small area and became more intense. (f) These effects were most obvious if a Müller cell was in the focal plane of the objective (indicated by its strong red fluorescence), i.e., at the same z level where the fiber core was placed.

Figure 4

Correlation between intraretinal pathway and beam divergence. (a) Overlay of the cellular fluorescence (red) and the beam scattering (green) at two consecutive fiber positions, y₀ and y₁ (dotted lines). At both positions, the core of the fiber was placed in front of the cell axis of one of two adjacent Müller cells. The scattering light spot on the plastic membrane formed two distinct areas (indicated by the black circles) with minimal overlap. (b) The line profile of the scattered light intensity I_SL at position x₁ shows two peaks with nearly the same curve progression and maximum intensity. (c) A line profile of the cellular fluorescence intensity I_FL along the IPL (at position x₀) was used to localize the y position (and, indirectly, the z position) of the inner stem process of Müller cells in the retinal tissue (red). The intensity peaks 1–8 (red) each represent a distinct Müller cell process in the focal plane of the objective where the core of the fiber was also placed. In contrast, several fluorescence minima (1′–6′) represent interjacent retinal tissue compartments devoid of a Müller cell process. All other (intermediate) fluorescence intensities could not be clearly assigned to one of these two cases. The width w of the light spots at the plastic membrane (blue curve in panel c) indicates the beam divergence on its course through the retina; it was estimated as the y width of a two-dimensional Gauss fit of the transmitted light spot. Comparing the cellular fluorescence intensities (red) with the width values (blue) shows that w was low at the fluorescence peaks (1–8) and high at the fluorescence minima (1′–6′). (d) The width w of the well-defined data points (1–8, 1′–6′, red dots) was plotted against the corresponding fluorescence values of the line profile in panel c. The resulting correlation coefficient was r = −0.9. (e) Schemata of a straight Müller cell, the axis of which does not change in the y and z directions. The beam axis and the endfoot axis as well as the resulting spot center are at the same y position. (f) A “bended” Müller cell causes a displacement d between beam axis and spot center. (g) A displacement also occurs if the beam does not hit the Müller cell axis. Only oblique light rays pass the cell.

Figure 5

Coupling efficiency of a light-guiding Müller cell. (a–f) A perfectly straight Müller cell was chosen for the experiment; it was moved in 2-μm steps perpendicular to the divergent light beam to investigate the effects for different coupling conditions into a Müller cell (illustrated in right schemata). (a and f) When the laser light illuminates a region between two Müller cells, no light is guided. The situation corresponds to that in Fig. 2c. (b and e) If the retina was moved to a region close to the center of a Müller cell, only oblique light rays enter the Müller cell. The transmitted spot intensity increased and the spot was displaced (b, below the beam axis; e, above the beam axis). (c and d) For ideal coupling conditions, the laser beam hits the center of a Müller cell. The spot intensity became brightest and no displacement was observed.

Figure 6

Double-sided imaging of retina whole-mount preparations combined with local light-transmission. (a) The x-z view of a confocal stack of fluorescent images of retinal tissue stained with Mitotracker green, acquired using a 40×/1.2 W objective. Shown is a maximum y projection across seven pixels (corresponding to 2.4 μm) around the line indicated in panel b. The endfeet (top) and processes of the Müller cells are clearly visible. (Green triangles) Positions at which the laser beam was focused onto the tissue; (white line) z position of the image in panel b. (b) Confocal x-y image of the same tissue shown in panel a, acquired with a 10×/0.3 objective. This image was used to position the laser beam (positions indicated by green circles). The transmission images corresponding to the two larger spots are shown in panels c and d. (c and d) Transmission images acquired with the upright imaging unit while the laser focus was either placed in the periphery at Müller cell endfoot (c) or on top of a Müller cell process (d). The scale is identical to the scale used for panels a and b; both images show the same field of view. (e and f) Transmission images shown in panels c and d (green) superimposed with an overview image (red) acquired with the same camera during wide-field transmission-mode illumination of the retina's inner surface.

Figure 7

Local light-transmission experiments on retinal whole-mount preparations. (a) Illumination of distinct groups of photoreceptor cells by a light beam directed to the vitread surface of a retina (compare to Fig. 1c). While the fiber was moved in equal steps (1 μm, every second position shown) into the y direction (the y position of the fiber is indicated by green bars at the left side), the field of illuminated photoreceptor cells changes in a discontinuous manner. Scale bar 5 μm. This is quantitatively analyzed in panel b, where the cumulative distance of the field center (center of gravity of the intensities), plotted against the y position of the fiber tip, shows discrete steps.

Figure 8

Biased light propagation through the retina. Laser light emanating from a glass fiber was projected onto the retina from the vitread side (compare to Fig. 1c). (Green square) The y position of the fiber that was moved in equal steps of 1 μm across the retinal surface (every second position shown). (Red) Resulting light pattern at the receptor side of the retina. Unlike the fiber, the pattern of illuminated photoreceptor cells moves rather irregularly; large jumps alternate with sequences of stationarity. In some positions, more than one photoreceptor group was illuminated. (Blue square) Center-of-intensity calculated for each frame after background substraction. The locations of these points were used for the plot of cumulative distances in Fig. 7b. Scale bar 5 μm.

Figure 9

Quantitative (immuno-) histochemistry of Müller cells and cone photoreceptor cells. (a) Confocal image of a retinal whole-mount preparation at the level of the OPL. In this layer the outer processes of Müller cells (labeled by antibodies directed to the intermediate filament vimentin, red) and the cone pedicles (labeled by cone-specific fluorescent peanut agglutinin, PNA, green) are visible. Scale bar, 20 μm. The magnified inlet (b) demonstrates a spatial colocalization of both structures. (c) Slice preparation of a retina that shows the endfoot funneling into the Müller cell process that finally forms thin branches enveloping a soma of a cone cell. This cell was identified as a cone because of its PNA-labeled inner segment. Scale bar 20 μm. (Ganglion cell layer (GCL), inner plexiform layer (IPL), outer plexiform layer (OPL), outer nuclear layer (ONL), photoreceptor segment layer (PRS).)