Knockout of PARG110 confers resistance to cGMP-induced toxicity in mammalian photoreceptors

Abstract

Hereditary retinal degeneration (RD) relates to a heterogeneous group of blinding human diseases in which the light sensitive neurons of the retina, the photoreceptors, die. RD is currently untreatable and the underlying cellular mechanisms remain poorly understood. However, the activity of the enzyme poly-ADP-ribose polymerase-1 (PARP1) and excessive generation of poly-ADP-ribose (PAR) polymers in photoreceptor nuclei have been shown to be causally involved in RD. The activity of PARP1 is to a large extent governed by its functional antagonist, poly-ADP-glycohydrolase (PARG), which thus also may have a role in RD. To investigate this, we analyzed PARG expression in the retina of wild-type (wt) mice and in the rd1 mouse model for human RD, and detected increased PARG protein in a subset of degenerating rd1 photoreceptors. Knockout (KO) animals lacking the 110 kDa nuclear PARG isoform were furthermore analyzed, and their retinal morphology and function were indistinguishable from wild-type animals. Organotypic wt retinal explants can be experimentally treated to induce rd1-like photoreceptor death, but PARG110 KO retinal explants were unexpectedly highly resistant to such treatment. The resistance was associated with decreased PAR accumulation and low PARP activity, indicating that PARG110 may positively regulate PARP1, an event that therefore is absent in PARG110 KO tissue. Our study demonstrates a causal involvement of PARG110 in the process of photoreceptor degeneration. Contrasting its anticipated role as a functional antagonist, absence of PARG110 correlated with low PARP activity, suggesting that PARG110 and PARP1 act in a positive feedback loop, which is especially active under pathologic conditions. This in turn highlights both PARG110 and PARP1 as potential targets for neuroprotective treatments for RD.

Figure 1

Retinal PARG expression in different genotypes: In wt retina, PARG expression was particularly evident in the NFL and in the perinuclear parts of a subpopulation of amacrine cells and horizontal cells (white arrows), as assessed by co-staining with calbindin (a–c). In PARG110 KO, PARG expression in perinuclear areas of amacrine and horizontal cells (white arrows) was strongly reduced, while PARG levels in the synaptic layers and the NFL appeared to be unaffected (d–f). In rd1 retina, the perinuclear areas of many photoreceptors displayed distinct PARG expression (g–i), in contrast to the wt situation (white arrows indicate horizontal cells). The images shown are representative for observations on at least three different specimens for each genotype

Figure 2

Retinal morphology and function is normal in PARG110 knockout animals. At P30, haematoxylin/eosin staining revealed normal morphology and layering for PARG110 KO retina, when compared to age-matched wt (a and b). OCT in vivo imaging confirmed the histological results (c and d). The quantification of photoreceptor rows in the outer nuclear layer (ONL) and ONL thickness showed no significant differences between wt and PARG110 KO retina at day 11 (P11) and day 30 (P30) after birth (e). In vivo functional analysis with ERG did not detect any differences between wt (black) and PARG110 KO (red) mice under dark-adapted (scotopic; SC) and light-adapted (photopic; PH) conditions (f and g). Statistical evaluation of b-wave amplitudes (box-and-whisker plot, indicating 5, 25, 50, 75, 95 percentiles of the data) and comparison of representative ERG traces are shown in (f) and (g), respectively. Error bars represent S.E.M. GCL, ganglion nuclear layer; INL, inner nuclear layer; IPL, inner plexiform layer; IS, inner segment; ns, not significant; ONL, outer nuclear layer; OPL, outer plexiform layer; OS, outer segment; RPE, retinal pigment epithelium

Figure 3

PARG110 KO photoreceptors are resistant to pharmacologically induced RD: organotypic retinal explants cultures derived from wt and PARG110 KO animals were treated with the PDE6 inhibitor zaprinast to induce an rd1-like retinal degeneration. Untreated wt and PARG110 KO explants exhibited minimal IF for cGMP, while zaprinast-treated specimens showed strongly increased ONL cGMP levels (a–d). PARG expression was seen only in the nerve fiber layer and in the perinuclear parts of a subpopulation of amacrine cells and horizontal cells for untreated wt, untreated PARG110 KO and zaprinast-treated PARG110 KO. However, in zaprinast-treated wt, an increased PARG expression was observed in the ONL and INL (e–h). PAR accumulation, a marker for PARP activity, was found in large amounts only in zaprinast-treated wt ONL (i–l), but rarely in untreated wt or in either treated or untreated PARG110 KO. The TUNEL assay for dying cells showed few positive cells in untreated wt but increased numbers in zaprinast-treated wt (m and n). Zaprinast-treated and untreated PARG110 KO retinal explants showed equally low numbers of TUNEL positive cells (o and p). Bar graphs show the quantification of TUNEL- (q) or PAR- (r) positive ONL cells. Error bars represent S.E.M. INL, inner nuclear layer; ONL, outer nuclear layer. Levels of significance: **P<0.01, ***P<0.001