Regulation of retinal ganglion cell production by Sonic hedgehog

Abstract

Previous work has shown that production of retinal ganglion cells is in part regulated by inhibitory factors secreted by ganglion cell themselves; however, the identities of these molecules are not known. Recent studies have demonstrated that the signaling molecule Sonic hedgehog (Shh) secreted by differentiated retinal ganglion cells is required to promote the progression of ganglion cell differentiation wave front and to induce its own expression. We present evidence that Shh signals play a role to negatively regulate ganglion cell genesis behind the differentiation wave front. Higher levels of Shh expression are detected behind the wave front as ganglion cells accumulate, while the Patched 1 receptor of Shh is expressed in adjacent retinal progenitor cells. Retroviral-mediated overexpression of Shh results in reduced ganglion cell proportions in vivo and in vitro. Conversely, inhibiting endogenous Shh activity by anti-Shh antibodies leads to an increased production of ganglion cells. Shh signals modulate ganglion cell production within the normal period of ganglion cell genesis in vitro without significantly affecting cell proliferation or cell death. Moreover, Shh signaling affects progenitor cell specification towards the ganglion cell fate during or soon after their last mitotic cycle. Thus, Shh derived from differentiated ganglion cells serves as a negative regulator behind the differentiation wave front to control ganglion cell genesis from the competent progenitor pool. Based on these results and other recent findings, we propose that Shh signals secreted by early-differentiated retinal neurons play dual roles at distinct concentration thresholds to orchestrate the progression of retinal neurogenic wave and the emergence of new neurons.

Fig. 1.

Expression patterns of Shh and Ptc1 during early chick retinal neurogenesis. In situ hybridization of retinal sections using chick Shh (A,C,E,G,I) and Ptc1 (B,D,F,H,J) probes are shown. Developmental stages according to Hamburger and Hamilton (Hamburger and Hamilton, 1951) are indicated on the left. In adjacent sections of the same eye, Shh is expressed as a gradient in the inner retina (A), whereas Ptc1 shows a similar gradient of expression with the highest levels near the optic nerve head (B). Also note the intense Ptc1 signals in the ventral forebrain, which reflects the significantly higher levels of Shh expression in the vicinity (data not shown). Hybridization signals of Shh are localized in the inner portion of the retina between stage 24 and 29 (arrowheads in C,E,G). No Shh signals are detected in the pigmented epithelium or in the periocular mesenchyme. Ptc1 signals are detected in the proliferative zone between stage 24 to 29 complimentary to the inner retina (D,F,H). Nearby sections of a stage 24 retina infected by the Shh virus at stage 10 (I,J) demonstrate that the non-uniform viral-mediated Shh expression results in a broad induction of Ptc1 mRNA transcription in the ventricular zone of the infected retina. fb, forebrain; gc, ganglion cells; on, optic nerve; pe, retinal pigmented epithelium; ret, retina.

Fig. 2.

Effects of perturbing Shh signals on ganglion cell differentiation in vivo. Immunohistochemical staining of central retina sections with anti-neurofilament (NF68) (A,C,E,G) or anti-Islet 1 (Isl-1) (B,D,F,H) antibodies are shown. Compared with the control retina infected by the RCAS virus at stage 10 (A,B), Shh-virus infected retina (C,D) shows a reduction of NF68 and Islet 1-positive cells at stage 24. Shh virus infected retinas also have increased thickness. Arrowheads point to marker-positive cells, possibly differentiating ganglion cells, present near the ventricular surface of Shh virus-infected retina. Compared with retina derived from the control 3C2 hybridoma cell injected eye (E), 5E1 hybridoma cell influenced retinas (G) display a thicker ganglion cell fiber layer stained positive for NF68 at stage 29. In addition, 5E1 hybridoma cell influenced retinas (H) also show increased number of Islet 1-positive cells compared with the non-injected contralateral retina (F) from the same embryo. gc, ganglion cell layer; pe, pigmented epithelium.

Fig. 3.

Quantification of ganglion cell production under different Shh levels in vivo and in vitro. Histograms of percentages of NF68 positive cells among total cells are shown. In this figure and subsequent figures (Figs 6, 8), each bar represents mean±s.e.m. The asterisks * and ** indicate P values between 0.01–0.05 and ≤0.01, respectively. Numbers outside the parentheses under the horizontal axis represent numbers of individual trials conducted; numbers within the parentheses indicate the total number of retinas used. (A) In vivo infection at stage 10 or stage 17 with Shh virus results in reduction of NF68-positive cells at stages 24, 28 and 30 compared with control RCAS virus (R) infection. In vivo injection at stage 17 with 5E1 hybridoma cells leads to an increase of NF68-positive cells at stage 30, whereas injection of control 3C2 hybridoma cells has no effects. (B) Retinal explants (center 75%) established at stage 24.5 and cultured under different conditions in vitro for 48 hours contain different proportions of NF68-positive cells. Viral-mediated Shh expression leads to a decrease of NF68-positive cells, whereas addition of 5E1 antibody results in an increase of NF68-positive cells. Exposure to control RCAS virus (R) or to hybridoma culture medium (NS) show the same percentages of marker positive cells as non-treated explants (−).

Fig. 4.

Influence of Shh levels on temporal differentiation of ganglion cells in vitro. Percentages of NF68-positive cells among total cells in retinal explants cultured for 5 days in vitro (DIV) are shown. Retinal explants were established at stage 22 and cultured under the same condition for 12 hours before the DIV0 samples were assayed. Different types of treatments, including infection with Shh virus or control RCAS virus and exposure to the anti-Shh 5E1antibody or the control 3C2 antibody, were initiated at DIV0. After 48 hours (DIV2), explants were sampled at every 24 hours till DIV5. The peak times for RGC production were the same (DIV2), despite different percentages of RGCs under various Shh conditions. Percentages of RGCs show gradual decline during later culture periods. Shh virus infected explants continued to show statistically significant lower percentages of RGCs at DIV3 and DIV4 compared with controls. Detailed data are summarized in Table 1.

Fig. 5.

Effects of Shh on cell proliferation in vivo. Immunohistochemical staining of E6 retinas treated with Shh virus and anti-Shh 5E1 hybridoma cells at stage 17 are shown. Central (A-C) and ventral (D-F) regions of retinas labeled by BrdU for 6 hours in vivo are stained by the anti-BrdU antibody. No effect of perturbing Shh on BrdU incorporation is found in the central retina; however, both elevated and reduced Shh levels result in less BrdU incorporation in the ventral retina near the optic fissure. Central regions of retina stained for the anti-phospho-histone H3 (PH3) show similar patterns of PH3-positive cell distribution (G-I). No differences are detected among non-treated controls, 3C2 hybridoma cell injected, and RCAS virus infected eyes (data not shown). pe, pigmented epithelium.

Fig. 6.

Quantification of cell proliferation and cell numbers in vivo and in vitro. Histograms of percentage of BrdU-positive cells (A-C) and total cell numbers (D) are shown. (A) BrdU labeling was performed in ovo for 3 hours or in vitro for 6 hours using freshly dissected retinas. For retinas infected at stage 10, Shh virus-infected and control RCAS virus (R)-infected retinas show similar percentages of BrdU-positive cells at stage 24 and stage 28. For eyes injected with hybridoma cells at stage 17, anti-Shh 5E1 cell and control 3C2 cell treat retinas display identical levels of BrdU incorporation at stage 30. (B) Retinal explants (center 75%) were established at stage 24.5 and exposed to different Shh conditions for 48 hours with BrdU added for the last 6 hours of culture. Similar levels of BrdU incorporation were found in anti-Shh 5E1 antibody- or Shh virus-treated retinas as in control retinas exposed to the hybridoma medium (NS), the RCAS virus (R), as well as the non-treated retinas (−). (C) Retinal explants were established at stage 22 and cultured in vitro for 12 hours before the DIV0 samples were assayed. Different treatments, including viral infection and antibody addition, were initiated at DIV0. After 48 hours (DIV2), explants were sampled at every 24 hours until DIV5. Explants were cultured in the presence of BrdU for the last 3 hours before dissociation and staining. No statistically significant changes of BrdU incorporation were detected in Shh virus infected or 5E1 antibody treated explants compared to RCAS virus infected or 3C2 antibody treated controls. See Table 1 for details. (D) Whole retinas were cultured as explants at stage 24 for 42.5 hours in vitro. Prior to dissociation at the end of the culture period, retinas were incubated in the presence of BrdU for 12 hours. Total cell numbers were determined based on cell counting. Total numbers of BrdU marker-positive cells were calculated based on percentages of BrdU-positive cells. Control retinas (−), 5E1 antibody-treated and Shh virus-infected retinas display similar numbers of total cells per retina as well as similar numbers of BrdU-labeled cells per retina.

Fig. 7.

Effects of Shh on cell death. TUNEL staining of eye sections treated in vivo (A-D) and retinal explant sections treated in vitro (E-G) are shown. Arrowheads point to a subset of TUNEL-positive cells. The right eyes were infected with the Shh virus (B) or injected with 5E1 hybridoma cells (D) at stage 17, and the corresponding left eyes (A,C) of the same embryos are used as stage matched controls. Similar regions of the two retinas show that the Shh virus-infected (B) retina has similar level of cell death at stage 30 (E6) as the non-infected left retina (A). The viral infection was limited to the right eye as confirmed with anti-viral GAG antibody staining (data not shown). Similar regions of two retinas show that the 5E1 cell influenced retina (D) contained slightly more apoptotic cells at stage 27 (E5) than the non-injected left retina (C). Control 3C2 hybridoma cell injected eyes showed similar levels of cell death as the non-treated retinas (data not shown). Representative sections of Stage 24 retinal explants cultured for 48 hours in vitro in the presence of 5E1 antibodies (F), Shh virus (G) or without treatment (E, cont) show similar levels of apoptosis. Arrows point to pigmented epithelium. h, hybridoma cells; le, lens; pe, pigmented epithelium; ret, retina.

Fig. 8.

Shh levels affect ganglion cell fate specification in vitro. Histograms show percentages of double marker-positive cells among BrdU-positive cells. Retinal explants (center 75%) were established at stage 23.5 and cultured for 40 hours under conditions without any treatment (−), or exposed to Shh viruses or 5E1 antibodies from the beginning of the culture period. The explants were then labeled with BrdU for 2.5 hours, followed by washes and further incubation in fresh medium for either 6 or 12 hours. At 6 hours after the BrdU labeling, retinal explants show similar percentages of phosphohistone H3 (PH3, a M phase marker) and BrdU double-positive cells among all BrdU-labeled cells. However, at 6 hours post BrdU labeling, Shh virus-infected and 5E1 antibody-treated explants display decreased and increased percentages of neurofilament (NF145) and BrdU double-positive cells among all BrdU-labeled cells, respectively. By 12 hours post BrdU labeling, β-tubulin (TUB) and BrdU double-positive cells emerged. Shh viral infection causes a decrease, and 5E1 antibody treatment causes an increase of β-tubulin and BrdU double-positive population among BrdU-positive cells.

Fig. 9.

Proposed dual roles for Shh during early retinal neurogenesis. A schematic cartoon depicts the early retina after RGC differentiation has begun. Shh is secreted by differentiated RGCs (red) located in the inner retina. Ahead of the RGC differentiation wave front naïve retinal progenitor cells (gray) are exposed to low levels of Shh signals, because they are farther away from the Shh-expressing cells. Low levels of Shh signal may be necessary to induce naïve progenitor cells to become competent for differentiation, and some eventually become Shh-producing RGCs. Behind the RGC differentiation wave front, progenitor cells (dark green) residing in the ventricular zone (VZ) are likely to have entered a competent state to be specified as RGCs and may contain activated MAPK. Higher levels of Shh are present behind the RGC wave front due to the accumulation of differentiated RGCs. Shh signaling negatively affects RGC fate specification of competent progenitor cells during or soon after M phase (yellow) of the mitotic cycle, and/or influence the further differentiation of nascent RGCs (orange) migrating towards the ganglion cell layer (GC). This model is consistent with data reported by Neumann and Nuesslein-Volhard (Neumann and Nuesslein-Volhard, 2000).