Here we provide evidence that rod photoreceptor homeostasis and regeneration in adult zebrafish is Fgf-dependent but cone photoreceptors do not have the same dependence. First, we show that intravitreal injections of FGF-2 induced rod precursor proliferation and neuroprotection during intense light damage to photoreceptors. Additionally, using the dominant-negative Tg(hsp70:dn-fgfr1) transgenic line, we found that Fgf signaling was required for homeostasis of rod, but not cone, photoreceptors. Finally, we found that Fgf signaling also differentially affected the regeneration of rod photoreceptors in the light-damaged retina.
Fgf signaling has been reported to play a key role in promoting neural retina development and regeneration (Cai et al., 2010
; Guillemot and Cepko, 1992
; Martinez-Morales et al., 2005
; Park and Hollenberg, 1989
; Pittack et al., 1997
; Zhao et al., 2001
) and in maintaining mammalian photoreceptor homeostasis (Stone et al., 1999
). We have shown that intraocular injections of FGF-2 induce rod precursor proliferation, although it is unclear whether these new cells properly differentiate into functional rod photoreceptors and innervate second-order inner neurons. Alternatively, it is possible that the newly generated rods do not function properly or are not needed and therefore targeted for cell death. Future studies combining cell tracing analysis and inducible Fgf-2 transgenes would help elucidate these possible outcomes.
Fgf signaling has been reported to promote neuroprotection (Cai et al., 2011
; Faktorovich et al., 1990
; Harada et al., 2000
; LaVail et al., 1991
; Zhang et al., 2003
), and FGF-2 protects photoreceptors from intense light damage in adult rodents (Faktorovich et al., 1992
; LaVail et al., 1991
). In addition, FGF-2 has been shown to promote photoreceptor survival in adult porcine and chick cell cultures (Frohns et al., 2009
; Traverso et al., 2003
). Our data strongly suggest a conserved neuroprotective function of Fgf signaling for rod photoreceptors in the zebrafish retina.
Of the four Fgf receptor subtypes, only fgfr1
is expressed in the adult zebrafish retina prior to light damage, and it is expressed in both rod and cone photoreceptors (), which is consistent with a previous report in the adult goldfish retina describing Fgfr1 immunolocalization in both rod and cone photoreceptors (Raymond et al., 1992
). Even though fgfr1
is expressed in both rod and cone photoreceptors (), we found that the dominant-negative hsp70:dn-fgfr1
transgene significantly repressed rod photoreceptor regeneration without affecting regeneration of red-green double cone photoreceptors (). As zebrafish have more than twenty Fgfs, it is possible that rod and cone photoreceptors respond differently to the presence of the different Fgf ligands. Alternatively, the differential response of rods and cones may be mediated through downstream targets. Recently it was reported that the loss of the protein tyrosine phosphatase Shp2, a known downstream target of Fgf signaling (Cai et al., 2010
), resulted in severe photoreceptor degeneration of both rod and cone photoreceptors in mice (Cai et al., 2011
). Future studies aimed at manipulating individual components of Fgf signaling in a cell-specific manner will be needed to answer these questions.
We found that endogenous fgfr1
is not expressed in the Müller glia of the zebrafish retina prior to light treatment () and that loss of Fgf signaling during retinal regeneration did not affect Müller glia proliferation or migration of INL retinal progenitors to the ONL (). This is in contrast to previous reports on isolated adult bovine Müller glial cells, and exogenous overexpression of FGF-2 in the adult chick retina, in which FGF-2 was shown to be mitogenic to Müller glial cells (Fischer et al., 2002
; Mascarelli et al., 1991
Our data suggest that in the adult zebrafish retina, Fgf signaling is not required for the early stages of retinal progenitor amplification or migration, but instead is involved at a later point when the progenitors are differentiating into rods and cones. However, the cellular mechanism that underlies the rod regeneration defect is not known. One possibility is that Fgf signaling is required for neuronal progenitors to commit to the rod photoreceptor cell fate but not cone photoreceptors. If this were the case, we would expect to observe a greater number of cone photoreceptors in the Tg(hsp70:dn-fgfr1) retinas following regeneration, but we did not. A second possibility is that inhibition of Fgf signaling affects proliferation of rod precursors in the ONL that give rise to differentiated rods. However, PCNA-positive rod precursors were observed in the ONL of transgenic retinas prior to light treatment () and at 2 days of light treatment (), suggesting this hypothesis is also not valid.
Homeostasis data provided evidence in support of a third explanation for the observed requirement for Fgf signaling in rod photoreceptor regeneration: a trophic effect on newly-differentiated rod photoreceptors. When Tg(hsp70:dn-fgfr1)
fish and wild-type siblings were heat-shocked daily without light lesion for 10 days, degeneration of rod outer segments and apoptosis of rod photoreceptors was observed in the transgenic retinas, but not in the wild-type retinas (). In the heat-shocked Tg(hsp70:dn-fgfr
) retinas, in response to loss of rods, proliferation of rod precursors in the outer nuclear layer was up-regulated (). Under normal Fgf signaling conditions, these proliferating rod precursors would replenish lost rod photoreceptors (Thummel et al., 2010
). However, when sustained inhibition of Fgf signaling for 60 days led to damage and loss of rods, proliferating rod precursors were unable to replenish them. Instead, the thickness of the outer nuclear layer was significantly reduced, reflecting a reduction in number of rods (). These data support the hypothesis that Fgf signaling is required for the survival of existing and newly-differentiated rod photoreceptors. Future studies that combine rod precursor-specific promoters and conditional inhibition of Fgf signaling may better clarify the cellular mechanism that underlies this trophic effect on rod photoreceptors.
In conclusion, this study adds to the growing body of work implicating Fgf signaling as an essential component for tissue regeneration, and is the first, to our knowledge, to provide evidence that Fgf signaling is required for regeneration of photoreceptors in adult retina.