Light adaptation does not prevent early retinal abnormalities in diabetic rats

Abstract

The aetiology of diabetic retinopathy (DR), the leading cause of blindness in the developed world, remains controversial. One hypothesis holds that retinal hypoxia, exacerbated by the high O2 consumption of rod photoreceptors in the dark, is a primary cause of DR. Based on this prediction we investigated whether early retinal abnormalities in streptozotocin-induced diabetic rats are alleviated by preventing the rods from dark adapting. Diabetic rats and their non-diabetic littermates were housed in a 12:12 hour light-dim light photocycle (30 lux during the day and 3 lux at night). Progression of early retinal abnormalities in diabetic rats was assessed by monitoring the ERG b-wave and oscillatory potentials, Müller cell reactive gliosis, and neuronal cell death, as assayed by TUNEL staining and retinal thickness at 6 and 12 weeks after diabetes induction. Maintaining diabetic animals in a dim-adapting light did not slow the progression of these neuronal and glial changes when compared to diabetic rats maintained in a standard 12:12 hour light-dark photocycle (30 lux during the day and 0 lux at night). Our results indicate that neuronal and glial abnormalities in early stages of diabetes are not exacerbated by rod photoreceptor O2 consumption in the dark.

Figure 1. Effect of background illumination on the ERG in control (non-diabetic) rats housed under standard lighting conditions.

(a) ERG responses to flash stimuli of 1.50 log cd.s/m² recorded under dark-adapted (0 lux background) and dim- and light-adapted conditions (background light intensities of 0.3, 3 and 30 lux). (b,c) Intensity-response relations for the a-wave (b) and the b-wave (c) under different background illuminance conditions. N = 5 rats. Background light levels as low as 0.3 lux light adapt the retina. ***P < 0.001 for multiple comparisons between ‘0 lux’ and background light intensities of 0.3, 3 and 30 lux. Error bars here and in other figures denote ± s.e.m.

Figure 2. Electroretinograms from control and diabetic rats.

Left panels show representative ERG waveforms from rats diabetic for 12 weeks (red traces) and aged-matched controls (black traces) housed under standard lighting conditions. Right panels show summary intensity-response relations. (a,b) Light-adapted retinas. Responses in (a) are to a flash at 2.17 log cd.s/m². b-Wave amplitudes are shown in (b). (c,d) Dark-adapted retinas. Responses in (c) are to a flash at 1.50 log cd.s/m². b-Wave amplitudes are shown in (d). (e,f) Oscillatory potentials. Responses in (e) are to a flash at 1.50 log cd.s/m². The sum of the oscillatory potentials time-to-peak (TTP) for OP₁ through OP₄ are shown in (f) In the right panels, rats diabetic for 6 and 12 weeks (Db) and aged-matched controls (Ctrl), housed under standard light conditions (0 lux, continuous traces) or dim-adapted at night conditions (3 lux, dashed traces) are shown. Diabetic rats had reduced b-wave amplitudes and increased latency of OPs at both 6 and 12 weeks of diabetes, compared to controls. Housing conditions (standard vs. dim-adapting) did not affect b-wave amplitude or oscillatory potential delay. Numbers in parentheses indicate number of rats. §, not significant for comparison among diabetic groups; *P < 0.05, for comparison between each diabetic group and aged-matched control. ^##P < 0.01 and ^###P < 0.001 for comparison between standard diabetic rats at 6- and 12-weeks. ^xP < 0.05 and ^xxxP < 0.001 for comparison between control groups at 6- and 12-weeks (standard in (b), and dim-adapted in (d)).

Figure 3. Müller cell reactive gliosis.

(a) Immunohistochemical labelling for the Müller cell specific enzyme glutamine synthetase (GS, red) and for GFAP (green) from rats diabetic for 12 weeks (bottom panels) and aged-matched controls (top panels) housed under standard lighting conditions. The nuclear stain DAPI is shown in blue. In control retinas Müller cells showed little labelling for GFAP, with expression restricted to astrocytes (arrows) at the retinal surface and along penetrating vessels. In diabetic rats, Müller cells displayed increased gliosis (GS-positive processes that were also GFAP-positive; arrowheads). For this and Fig. 4, GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, other plexiform layer, ONL, outer nuclear layer; PL, photoreceptor layer. Scale bar: 50 μm. (b) Summary data for Müller cell gliosis (percentage of GS-positive processes in the IPL that were also GFAP positive), at 6 and 12 weeks in peripheral, mid and central retina in control (Ctrl) and diabetic rats (Db) housed under standard (0 lux) or dim-adapted at night (3 lux) conditions. Müller cells in the mid and central retina display increased gliosis at 6 and 12 weeks of diabetes, and in the peripheral retina at 12 weeks. Housing conditions (standard or dim-adapting) did not affect the degree of gliosis at either 6 or 12 weeks. Numbers in parentheses indicate number of rats. ***P < 0.001 compared to aged-matched control.

Figure 4. Cell death in control and diabetic retinas.

(a) TUNEL staining in retinal cross-sections from rats diabetic for 6 weeks and aged-matched controls, housed under standard lighting conditions. DAPI-labelled nuclei (blue) and a TUNEL-positive cell (blue/green; arrow) are shown. Scale bar: 50 μm. (b) Mean number of TUNEL-positive cells from 12-micron-thick retinal sections through the optic disc, normalized to the length of the retinal slice. (c) Total retinal thickness in control (Ctrl) and diabetic rats (Db), housed under standard (0 lux) or dim-adapted at night (3 lux) conditions. Numbers in parentheses indicate number of rats. *P < 0.05, ***P < 0.001 vs age-matched control; ^#P < 0.05 comparison between dim-adapted diabetic rats at 6- and 12-weeks.