Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Nat Cell Biol. Author manuscript; available in PMC 2013 April 1.
Published in final edited form as:
PMCID: PMC3463712

Insm1a-mediated gene repression is essential for the formation and differentiation of Müller glia-derived progenitors in the injured retina


In zebrafish, retinal injury stimulates Müller glia (MG) reprograming; allowing them to generate multipotent progenitors that regenerate damaged cells and restore vision. Recent studies suggest transcriptional repression may underlie these events. To identify these repressors, we compared the transcriptomes of MG and MG-derived progenitors and identified insm1a, a transcriptional repressor exhibiting a biphasic pattern of expression that is essential for retina regeneration. Insm1a was found to suppress ascl1a and its own expression and link injury-dependent ascl1a induction with dickkopf (dkk) suppression, which is necessary for MG dedifferentiation. We also found that Insm1a was responsible for sculpting the zone of injury-responsive MG by suppressing hb-egfa expression. Finally, we provide evidence that Insm1a stimulates progenitor cell cycle exit by suppressing a genetic program driving progenitor proliferation. Our studies identify Insm1a as a key regulator of retina regeneration and provide a mechanistic understanding of how it contributes to multiple phases of this process.

In contrast to mammals where retinal injury results in glial scaring1, zebrafish mount a robust regenerative response that culminates in the restoration of vision2. Müller glia (MG) are largely responsible for this remarkable phenomenon. They accomplish this by injury-induced cellular reprogramming that allows them to generate progenitors that are able to regenerate all retinal cell types39. Although mammalian MG can be coaxed to proliferate, their regenerative capacity is meager and limited1013. Therefore, understanding the mechanisms underlying successful retina regeneration in zebrafish may provide insight for developing strategies to stimulate MG reprogramming and retina regeneration in mammals.

Recent studies have identified a number of gene products and signaling cascades contributing to retina regeneration. These include an Ascl1a/Lin28/let-7 miRNA signaling pathway that supports the conversion of MG into progenitors4; an Ascl1a/Dkk/Wnt signaling pathway that in conjunction with Pax6, Stat3 and Hspd1 regulates MG reprogramming and progenitor proliferation5,6,14,15; a heparin binding epidermal-like growth factor (HB-EGF)-dependent signaling pathway that stimulates MG reprogramming and retina regeneration16; and FGF, Mps1 and galectin-dependent signaling that contribute to photoreceptor regeneration14,17,18.

Although many of the signaling molecules described above are positive effectors, it is becoming increasingly clear that gene repression is essential for retina regeneration. For example, dickkopf (dkk) gene suppression contributes to progenitor formation5, and hb-egf gene repression helps modulate the size of the progenitor population16. Perhaps the most obvious role for repression is the curtailing of cell proliferation programs as progenitors begin to differentiate. Mechanisms underlying this latter repression are of major interest since they prevent uncontrolled growth that may lead to tumor formation and glial scarring. Here we report the remarkable finding that a single transcriptional repressor, Insm1a, contributes to all these repressive events, placing it among the key regulators of retina regeneration.


We probed zebrafish whole genome microarrays (Agilent) to identify transcriptional repressors regulating retina regeneration (GEO accession:GSE36191). Probes were derived from FACS purified MG and MG-derived progenitors from gfap:gfp and 1016 tuba1a:gfp transgenic fish retinas, respectively (Supplementary Fig. S1a)3,19. The advantage of using 1016 tuba1a:gfp fish is that they specifically label MG-derived progenitors, free of contaminating MG and other cell types. This analysis identified over 1200 genes induced greater than 2-fold and over 300 genes suppressed by at least 50%. Of these, we identified transcriptional repressors belonging to her, bcl11, klf, eng and insm1 gene families (Supplementary Fig. S1b). RT-PCR confirmed injury-dependent regulation of these genes and revealed an intriguing biphasic pattern of expression for insm1a (Supplementary Fig. S1c). insm1a mRNA was first detected around 6 hrs post injury (hpi), suppressed at 24 hpi and reappeared by 4 days post injury (dpi). This pattern of expression suggested that Insm1a may play multiple roles during retina regeneration.

Injury-dependent insm1a induction is necessary for retina regeneration

In situ hybridization and immunofluorescence at 4 dpi in 1016 tuba1a:gfp transgenic fish3 showed insm1a was expressed in GFP+ progenitors that also expressed the MG marker, glutamine synthetase (GS) (Fig 1a, Supplementary Fig. S2a). Furthermore, many, but not all of these insm1a+ cells incorporated BrdU (Fig. 1b, c). Importantly, Insm1a knockdown using 2 different lissamine-tagged morpholino-modified antisense oligonucleotides (MOs) electroporated into the retina at the time of injury showed Insm1a was necessary for the generation of GFP+ and BrdU+ MG-derived progenitors (Fig. 1d, e; Supplementary Fig. S2b; Supplementary Table S1). Insm1a-targeting morpholinos prevented insm1a:gfp reporter gene expression in zebrafish embryos (Supplementary Fig. S2c) and suppressed injury-dependent Insm1a protein induction at the injury site (Supplementary Fig. S2d, e), without affecting cell death (Supplementary Fig. S8).

Figure 1
Insm1a is necessary for MG dedifferentiation in the injured retina. (a, b) ISH and IF shows insm1a RNA is expressed in GFP+, GS+ and BrdU+ MG-derived progenitors of 1016 tuba1a:gfp transgenic fish at 4 dpi. Arrows in (a, b) identify co-labeled cells, ...

Although insm1a was localized to MG-derived progenitors at 4 dpi, it exhibited a pan retinal pattern of expression at 6 hpi that became restricted to the injury site by 2 dpi (Fig. 1f–h; Supplementary Fig. S3). This spatial and temporal expression pattern was reminiscent of that previously reported for ascl1a5 and both these RNAs localized to MG-derived progenitors at 2 dpi (Fig. 1g). Interestingly, close inspection of insm1a expression and BrdU incorporation at 2 dpi suggested insm1a often associates with cells flanking BrdU+ progenitors (Fig. 1h, i). Quantification showed ~30% of the progenitors expressed insm1a at 2 dpi which increased to 40% by 4 dpi (Fig. 1j; Supplementary Table S2). These data, along with the observation that Insm1a knockdown blocks progenitor production by over 80% (Fig. 1e; Supplementary Table S1), suggested Insm1a expression in dedifferentiated MG is necessary for them to produce progenitors for retinal repair.

An Ascl1a/Insm1a regulatory loop

Because ascl1a and insm1a exhibited a similar pattern of expression in the injured retina, we wondered if there was a hierarchical relationship. An analysis of injury-dependent ascl1a and insm1a mRNA expression showed insm1a lagged behind that of ascl1a and that insm1a was transiently reduced at 24 hpi (Fig. 2a, b). Ascl1a knockdown suppressed insm1a expression (Fig. 2c–e; Supplementary Fig. S2e), while Insm1a knockdown caused a small increase in insm1a and ascl1a expression (Fig. 2e). Inspection of the insm1a promoter revealed 3 putative Ascl1a binding sites (Fig. 2f) and ChIP assays using myc-Ascl1a overexpressed in zebrafish embryos showed Ascl1a binds these sites (Fig. 2g). Coinjection of zebrafish embryos with an insm1a:gfp-luciferase reporter and increasing concentrations of ascl1a mRNA showed Ascl1a stimulates insm1a promoter activity (Fig. 2h). Although these data are consistent with Ascl1a directly regulating insm1a promoter activity, we were unable to confirm this direct regulation in adult retinas due to the lack of a suitable antibody for its detection.

Figure 2
An Ascl1a/Insm1a regulatory loop. (a) RT-PCR and (b) qPCR were used to assay injury-dependent insm1a, ascl1a and ribosomal protein l-24 gene expression, n=6 biological replicates. (c) RT-PCR and qPCR shows Ascl1a knockdown inhibits insm1a induction. * ...

Because Insm1a knockdown resulted in a small, but significant increase in ascl1a and insm1a expression (Fig. 2e), we suspected that Insm1a may feedback to regulate their promoters. Inspection of ascl1a and insm1a promoter sequences revealed putative Insm1a binding sites (Fig. 2i, k) and ChIP assays confirmed that endogenous Insm1a bound these sites in the injured retina (Fig. 2j, l). Furthermore, coinjection of zebrafish embryos with ascl1a:gfp-luciferase or insm1a:gfp-luciferase reporters harboring wild type or mutant promoters and increasing concentrations of insm1a mRNA showed that these Insm1a binding sites confer Insm1a-dependent regulation (Fig. 2m, n). This led us to speculate that Insm1a itself may be contributing to its transient suppression at 24 hpi (Fig. 2a, b), and Insm1a knockdown confirmed this suspicion (Fig. 2o). Together these experiments reveal an Ascl1a-Insm1a regulatory loop and an Insm1a autoregulatory loop that contributes to the dynamic expression pattern of ascl1a and insm1a during the course of regeneration.

Insm1a mediates dickkopf (dkk) gene repression in the injured retina

We previously demonstrated that the formation of MG-derived progenitors required dkk repression5. Interestingly, here we report an opposing pattern of insm1a and dkk1b gene expression. We found that insm1a was expressed throughout the retina at 6 hpi, but restricted to MG-derived progenitors at 4 dpi, while dkk1b exhibited the opposite pattern of expression (Fig. 3a, b)5. Importantly, Insm1a knockdown prevented injury-dependent dkk suppression (Fig. 3c). Inspection of the dkk1b promoter identified 2 putative Insm1a binding sites that bound endogenous Insm1a in the injured retina (Fig. 3d, e). In addition, zebrafish embryos coinjected with a wild type or mutant dkk1b:gfp-luciferase reporter and increasing concentrations of insm1a mRNA showed that both Insm1a binding sites contribute to dkk1b promoter repression by Insm1a (Fig. 3f). Finally, Insm1a knockdown in zebrafish embryos increased promoter activity from a co-injected dkk1b:gfp-luciferase reporter and also prevented Ascl1a-dependent dkk1b:gfp-luciferase reporter suppression (Fig. 3g). These data suggest Insm1a represses dkk promoter activity in the adult retina and that this repressive mechanism can be recapitulated in zebrafish embryos by either Insm1a or Ascl1a overexpression. Taken together these data identify an Ascl1a-Insm1a-Dkk signaling cascade underlying MG reprogramming in the adult retina.

Figure 3
Insm1a regulates dkk promoter activity. (a) ISH and IF show insm1a and dkk exhibit opposing expression patterns in the uninjured and injured retina. Arrows point to BrdU+ MG-derived progenitors with high levels of insm1a and low levels of dkk1b mRNAs. ...

Insm1a restricts the zone of MG-derived progenitors by suppressing hb-egf expression

We next investigated the significance of insm1a expression around 4–6 dpi when MG-derived progenitors are proliferating and differentiating. To bypass the block in regeneration that would result from Insm1a knockdown at the time of injury, we waited until 4 dpi to electroporate the control or insm1a-MOs (Fig. 4a) and confirmed their effectiveness in knocking down Insm1a at 6 dpi (Supplementary Fig. S4g). This delayed knockdown had no effect on cell death (Supplementary Fig. S8). In contrast to electroporation at the time of injury which blocks progenitor formation (Figs. 1d, e), Insm1a knockdown at 4 dpi greatly expanded the number of progenitors (Fig. 4b–d; Supplementary Fig. S4b). This result was reminiscent of DAPT-dependent Notch inhibition16 and suggested that Insm1a suppression may underlie the expanded zone of progenitors in DAPT-treated retinas. Consistent with this idea, DAPT inhibited insm1a expression at 4 dpi (Fig. 4e). Interestingly, unlike the initial injury response5, progenitor proliferation following Insm1a knockdown at 4 dpi was insensitive to Dkk overexpression (Fig. 4f; Supplementary Fig. S4c–f).

Figure 4
Insm1a regulates the zone of injury-responsive MG. (a) Experimental time line. Control or insm1a-targeting MO was injected into the retina at the time of injury and 4 days later delivered to cells by electroporation; 4 hrs later, fish received an IP injection ...

Since HB-EGF is released locally at the injury site and stimulates MG dedifferentiation16, its suppression by Insm1a would provide a convenient mechanism for restricting the zone of dedifferentiating MG. Indeed, insm1a and hb-egfa are coexpressed in BrdU+ dedifferentiated MG (Fig. 5a), and Insm1a knockdown at 4 dpi stimulated hb-egfa expression (Fig. 5b, c). Furthermore, we identified 2 putative Insm1a binding sites in the hb-egfa promoter and ChIP assays showed endogenous Insm1a in the injured retina bound these sites (Fig. 5d). Finally, HB-EGFa knockdown or inhibition of its receptor with PD158780 prevented progenitor expansion in Insm1a depleted retinas (Fig. 5e, f) without affecting cell death (Supplementary Fig. S8). These results suggest that a Notch-Insm1a-hb-egfa signaling cascade helps define the zone of reprogrammed MG flanking the injury site.

Figure 5
Insm1a restricts the zone of dedifferentiating MG by suppressing hb-egfa gene expression. (a) Fluorescent ISH and IF shows insm1a and hb-egfa mRNAs are co-expressed in BrdU+ MG-derived progenitors at 4 dpi. Asterisk flanks the injury site. Scale bar: ...

Insm1a stimulates progenitor cell cycle exit by suppressing a genetic program driving cell proliferation

Fish receiving an IP injection of BrdU 3 hrs prior to sacrifice showed that the fraction of proliferating progenitors expressing insm1a increased from ~40% at 4 dpi to ~80% at 6 dpi (Fig. 6a; Supplementary Fig. S5a; Supplementary Table S2). Because progenitor proliferation is decreasing during this time period3, we hypothesized that Insm1a may be associated with cell cycle exit. Our transcriptome analysis of MG and MG-derived progenitors identified 16 cell cycle-related genes that were induced in the injured retina (Supplementary Fig. S5b) and whose expression peaked around 4 dpi (Fig. 6b). We hypothesized that Insm1a may suppress these genes around 5–7 dpi and thereby promote cell cycle exit. Indeed, almost all of the genes examined showed higher expression at 6 dpi when retinas were electroporated with an insm1a-MO compared to a control MO (Fig. 6c, d).

Figure 6
At 4–6 dpi Insm1a inhibits a genetic program driving cell proliferation and stimulates a genetic program driving cell cycle exit. (a) Quantification shows that the number of BrdU+ cells that are also insm1a+ increases from 4–6 dpi. (b) ...

Unique among the cell cycle genes is p57kip2 (cdkn1c), a cyclin kinase inhibitor associated with cell cycle exit20,21 whose expression was detected from ~6 hpi to 8 dpi (Fig. 6e). At 4 dpi, p57kip2 was co-expressed with insm1a in MG-derived progenitors (Fig. 6f). While Insm1a and Ascl1a knockdown from 0–2 dpi had little effect on p57kip2 expression (Fig. 6g), knockdown from 4–6 dpi dramatically suppressed its expression and increased progenitor proliferation (Fig. 6g, h). Thus, Insm1a stimulates progenitor cell cycle exit by inhibiting genes driving proliferation and stimulating expression of the cyclin-cdk inhibitor, p57kip2.

Insm1a most likely mediates p57kip2 induction by suppressing an intervening repressor. One candidate was bcl11b (CTIP2) that inhibit p57kip2 gene expression in SK-N-MC cells22. Interestingly, our microarray data suggested bcl11b was reduced by ~80% in MG-derived progenitors (Supplementary Fig. S1b). Zebrafish harbor 4 different bcl11 genes and all 4 are coordinately repressed in a pan retinal fashion shortly after injury (Fig. 7a, b). Using bcl11a as a representative member of this gene family, we found that its expression returned to non-MG, but remained repressed in MG-derived progenitors at 4 dpi (Fig. 7b). This expression profile was temporally and spatially similar to that of dkk1b (Fig. 3b)5 and suggested a shared mechanism of regulation. Indeed, like dkk (Fig. 3c)5, Insm1a or Ascl1a knockdown relieved injury-dependent bcl11a repression (Fig. 7c).

Figure 7
Insm1a stimulates p57kip2 expression by suppressing bcl11 gene expression. (a) RT-PCR shows the temporal pattern of bcl11 family members following retinal injury. (b) RT-PCR and qPCR using mRNA from FACS purified non-MG and MG show bcl11a mRNA is transiently ...

We next investigated if Insm1a directly regulated bcl11a gene expression. Inspection of the bcl11a promoter identified a single putative Insm1a binding site in its proximal region (Fig. 7d) and ChIP assays showed endogenous Insm1a bound this site in the injured retina (Fig. 7e). To test if this Insm1a binding site was functional we co-injected zebrafish embryos with bcl11a:gfp-luciferase reporters harboring wild type or mutant promoters and insm1a mRNA. This experiment showed that the Insm1a binding site was necessary for Insm1a-dependent promoter repression (Fig. 7f). Furthermore, Insm1a overexpression in zebrafish embryos inhibited bcl11a expression (Fig. 7g) and induced p57kip2 expression (Fig. 7h); while Bcl11a overexpression suppressed endogenous p57kip2 expression (Fig. 7i). These Insm1a-dependent gene expression changes observed in the embryo recapitulate that found in the adult injured retina and suggests Insm1a drives progenitor cell cycle exit, in part, by relieving Bcl11a-dependent repression of p57kip2 gene expression.

Insm1a knockdown transiently suppresses progenitor differentiation

One consequence of Insm1a driven cell cycle exit is progenitor differentiation. Therefore, we investigated if Insm1a knockdown at 4–6 dpi reduced the number of differentiating progenitors. HuC/D is often used as a marker of differentiating neurons. In the retina HuC/D also identifies amacrine and ganglion cells. We investigated if HuC/D could be used to identify differentiating cells in the regenerating retina and whether Insm1a was necessary for their formation. For these experiments fish received control or insm1a-targeting MO at the time of injury, an IP injection of BrdU at 4 dpi, MO electroporation 5 hrs post BrdU injection and then sacrificed at 6 dpi (Fig. 8a). Interestingly, double labeled BrdU+ and HuC/D+ cells migrating towards the outer nuclear layer (ONL) were readily identified in the control MO-treated retina, but not those with Insm1a knocked down (Fig. 8b, c) suggesting Insm1a drives differentiation in the injured retina.

Figure 8
Insm1a knockdown transiently suppresses progenitor differentiation. (a) Experimental time course showing BrdU labeling 4 hrs prior to MO electroporation at 4 dpi and analysis at 6 dpi. (b) HuC/D and BrdU IF shows that in control MO-treated retinas BrdU ...

We suspected that the excessive proliferation following Insm1a knockdown at 4 dpi may result in MO dilution so that progenitors may ultimately restore Insm1a levels and return to a normal program of regeneration. To investigate this possibility, we followed a similar experimental paradigm as described above except we gave fish daily IP injections of BrdU from 4–7 dpi before sacrificing them at 30 dpi (Fig. 8d). Differential interference contrast (DIC) microscopy showed complete recovery of retinal lamination at the injury site (Fig. 8e). Immunofluorescence detection of BrdU and retinal cell type-specific proteins showed that progenitors migrated to all 3 retinal layers regenerating zpr1+ photoreceptors, PKC+ bipolar cells, GS+ MG, HuC/D+ amacrine and putative ganglion cells, and Zn5+ differentiating ganglion cells (Fig. 8e, f; Supplementary Fig. S6). Interestingly, Insm1a knockdown had little effect on the distribution of differentiated cell types. In contrast, progenitors originally residing in the expanded zone flanking the injury site were generally confined to a single lamina of the INL and predominantly regenerated GS+ MG (Fig. 8e, f; Supplementary Fig. S6; Supplementary Table S3). These data suggested that MG-derived progenitors were eventually able to overcome the cell cycle trap imposed by Insm1a knockdown. However, unlike in previous studies where we found that progenitors residing outside the injury zone could regenerate all cell types at 18 dpi5,16, this study suggests that at later times (30 dpi) MG selectively survive (Fig. 8f). We suspect these are the original reprogrammed MG that reacquired mature MG characteristics. Although we did not investigate the reason why newly regenerated non-MG cells flanking the injury site eventually died off, it may be due to their inability to integrate into functional circuits like those at the injury site.


Insm1 is a transcriptional repressor associated with neuroendocrine tumors23, small cell lung cancer24, and the terminal cell divisions associated with neurogenesis2528. Although Insm1 plays an important role during nervous system development, surprisingly little is known about Insm1’s action in the adult nervous system. Furthermore, the mechanisms controlling insm1 gene induction and underlying Insm1’s action on progenitor differentiation remain poorly understood.

Zebrafish harbor 2 insm1 genes referred to as insm1a and insm1b29. In the adult retina, insm1a is normally expressed in the neurogenic ciliary marginal zone, but can also be found in the central retina of fish with photoreceptor degeneration30. Our studies suggest that Insm1a is a multifaceted transcriptional repressor that plays an essential role in the formation, expansion and differentiation of MG-derived progenitors during retina regeneration. These studies highlight how a single transcription factor can assume diverse roles at different stages of regeneration and point to the cellular environment as a critical factor in determining Insm1a function. Furthermore, our studies revealed signaling mechanisms underlying injury-dependent insm1a gene induction and also identified mechanisms by which Insm1a acts to control MG reprograming, along with the proliferation and differentiation of MG-derived progenitors. These results may have important implications for stimulating retina regeneration in mammals and for preventing uncontrolled MG proliferation in diseased and damaged human retinas. Furthermore our studies suggest mechanisms underlying Insm1 gene function and action in mammals.

Although Insm1 is best known as a transcriptional repressor that is associated with terminal cell division and neuronal differentiation, we uncovered a number of novel roles for Insm1a during retina regeneration. First, Insm1a links ascl1a gene induction with dkk gene repression, which we previously showed was necessary for MG dedifferentiation and retina regeneration5. Second, using a novel protocol of delayed MO electroporation and gene knockdown, we found that Insm1a regulates the zone of injury-responsive MG flanking the injury site. This later function of Insm1a was only observed if MG were allowed to initially dedifferentiate and generate progenitors, suggesting that knockdown of Insm1a in these progenitors stimulated reprogramming of neighboring MG so they could generate additional progenitors. A possible mechanism underlying this effect was suggested by our finding that Insm1a controls hb-egfa gene expression, whose product was recently shown to stimulate MG reprogramming and progenitor formation in the uninjured retina16.

Investigation of the mechanisms initiating injury-dependent Insm1a induction identified Ascl1a, a gene product that is a nodal point for a number of signaling cascades during retina regeneration4,5,31,32. We previously showed that ascl1a gene expression is regulated by HB-EGF and is among the earliest gene inductions following retinal injury16. Here we show Ascl1a is necessary for injury-dependent insm1a induction. Thus, together these studies identify an HB-EGF/Ascl1a/Insm1a/dkk signaling cascade as a critical signaling mechanism underlying the formation of injury-induced MG-derived progenitors. Interestingly, Ascl1 and Insm1 induction are also associated with Dkk repression in certain human lung cancers24,33 and pancreatic endocrine tumors34,35, perhaps suggesting a conserved signaling pathway.

In addition to regulating dkk gene expression, Insm1a also feeds back to inhibit hb-egfa, ascl1a and insm1a promoter activity. This kind of feedback is often associated with oscillations36,37 and appears to underlie the biphasic pattern of insm1a gene expression during retina regeneration and also helps restrict injury-responsive MG to the injury site.

Insm1a knockdown at 4 dpi not only expanded the zone of MG reprogrammed to produce progenitors, but also dramatically increased progenitor proliferation at the expense of differentiation. This action is consistent with Insm1’s role in mammals where it is associated with cell cycle exit. Insm1 appears to stimulate cell cycle exit by sequestering cyclin D1 with its proline-rich cyclin D1 binding domain38. However, this domain is missing in zebrafish Insm1a (Supplementary Fig. S7), suggesting another mechanism of action. Remarkably, we found that Insm1a not only suppressed a gene expression program that drives cell proliferation, but also relieved repression of p57kip2, a cyclin kinase inhibitor that along with p27kip1 drives cell cycle exit during mouse retina development20,39. In the mouse retina, reduced expression of p27kip1 is associated with MG proliferation and reactive gliosis40. Interestingly, our transcriptome analysis of MG and MG-derived progenitors revealed that p27kip1 is constitutively expressed in these two populations. Whether this constitutive expression in the injured zebrafish retina helps prevent a gliotic response and promotes a regenerative one is not known.

In summary, our data suggest Insm1a plays at least 3 important roles during retina regeneration (Fig. 8g, h). First, it contributes to MG reprogramming and the generation of progenitors by inhibiting Dkk expression and releasing the Wnt/β-Catenin pathway from inhibition. Second, it helps sculpt the zone of injury-responsive MG by regulating hb-egfa gene expression. Third, it contributes to the cessation of retina regeneration by stimulating cell differentiation via the suppression of genetic programs driving cell proliferation. Finally, our studies revealed mechanisms by which Insm1a mediates these effects and the signaling pathways underlying insm1a gene regulation in the injured retina. These studies place Insm1a among the key factors underlying retina regeneration and provide novel insight into signaling pathways that may help shift the response of MG in the injured mammalian retina from reactive gliosis towards retinal repair.


Animals, heat shock, drugs and retinal injury

Zebrafish were kept at 26–28 °C on a 14h/10h light/dark cycle. 1016 tuba1a:gfp, gfap:gfp, and hsp70:dkk1b-gfp transgenic fish have been previously described3,19,41. Embryos were obtained by natural matings. Heat shock was performed by transferring hsp70:dkk1b-gfp fish to at 36.5 °C water bath for 1 h every 12 h beginning at 3 dpi and ending at 6 dpi. The EGFR inhibitor, PD158780 (Sigma-Aldrich) was used at 10 μM. Drugs were delivered intravitreally through the front of the eye using a Hamilton syringe equipped with a 30 gauge needle. Retinal lesions were performed as previously described3. Briefly, fish were anesthetized in tricaine methane sulfonate and the right eye was gently rotated from its socket and the retina stabbed 4–8 times (once or twice in each quadrant) through the sclera with a 30-gauge needle inserted the length of the bevel. All experiments were repeated a minimum of 3 times.

Florescence-activated cell sorting (FACS) and microarray

RNA was obtained from FACS purified MG and MG-derived progenitors at 4 dpi as previously described5. Briefly uninjured and injured retinas were isolated from gfap:gfp and 1016 tuba1a:gfp transgenic fish. GFP+ MG from gfap:gfp uninjured retinas and GFP+ MG-derived progenitors from 1016 tuba1a:gfp retinas at 4 dpi were isolated by treating retinas with hyaluronidase and trypsin and then sorted on a BC Biosciences FACSViDa 3 laser high speed cell sorter. Four uninjured retinas from gfap:gfp fish yielded 235,000 GFP+ cells, while 30 injured retinas from 1016 tuba1a:gfp fish yielded 140,000 GFP+ cells. Total RNA was isolated using TRIzol (Invitrogen) and underwent one round of amplification to generate probes for screening a Zebrafish 44K microarray (Agilent, #G2519F). Duplicate samples were analyzed. Microarray data have been submitted to GEO (GSE36191).

Primers and Plasmid construction

All primers are listed in Supplementary Table S4. ascl1a, insm1a, bcl11a promoters were amplified from zebrafish genomic DNA using primer pairs XhoAscl1aPro-F and BamAscl1aPro-R (~6kb), Xho-insm1a-Pro-F & Bam-insm1a Pro-R (~3kb) and Xho-bcl11a Pro-F & EcoR1-bcl11a Pro-R (~2kb), respectively. The PCR amplicons were digested and cloned into a pEL luciferase expression vector to create ascl1a:gfp-luciferase, insm1a:gfp-luciferase and bcl11a:gfp-luciferase constructs. The dkk1b:gfp-luciferase construct was described previously5. Site-directed mutagenesis was done as previously described4.

insm1a and bcl11a cDNA were amplified from zebrafish retina RNA at 4 dpi using primer pairs Bam-insm1a-F and Xho-insm1a-R (1.1kb) and Cla-bcl11a-F and Cla-bcl11a-R (2.4kb). The PCR amplicons were cloned into their respective enzyme sites in pCS2+-MT (insm1a) and pCS2+ (bcl11a) plasmid to obtain cmv:myc-insm1a and cmv:bcl11a. The cmv:myc-ascl1a construct was described previously4. All primers used in this study are listed in Supplementary Table S1.

RNA isolation, RT-PCR and qPCR

Total RNA was isolated from control and injured retinas using TRIzol (Invitrogen). Random hexamers and Superscript II reverse transcriptase (Invitrogen) were used to generate cDNA. PCR reactions used Taq polymerase and gene-specific primers (Supplementary Table S4) with previously described cycling conditions4. qPCR was carried out in triplicate with Absolute SYBR Green Fluorescein Master Mix (Thermo Scientific) on an iCycler real-time PCR detection system (BioRad). The ΔΔCt method was used to determine relative expression of mRNAs in control and injured retinas and normalized to ribosomal protein L-24 mRNA levels.

mRNA synthesis, embryo microinjection and ChIP assay

pCS2+ and pCS2+-MT plasmids harboring cDNA inserts were linearized and capped mRNAs were synthesized using the mMESSAGE mMACHINE (Ambion). Single cell zebrafish embryos were injected with ~200 pl of solution generally containing 0.02 pg of Renilla luciferase mRNA (normalization), 2 pg of promoter:gfp-luciferase vector and 0–4 pg of ascl1a, insm1a or bcl11a mRNA. To ensure reproducibility a master mix was made for daily injections and ~200 embryos were injected at the 1–2 cell stage. 24 hrs later, embryos were divided into 3 groups (~65 embryos/group) and lysed for dual luciferase reporter assays (Promega).

ChIP assays to analyze endogenous Insm1a binding to various promoters in the adult injured retina was performed using ~100 adult retinae collected at 4 dpi after dark adaptation. Chromatin was isolated and sonicated as described previously42. The sonicated chromatin was distributed into 3 equal aliquots; 2 were probed with an anti-zebrafish Insm1a antibody (AnaSpec, Fremont, CA. Cat No # 55795-2, 1:250 dilution) and the third served as a control. For ChIP assays in embryos with myc-tagged protein overexpression, embryos were injected with 4 pg of myc-tagged mRNA and ChIP assays performed as previously described42. Primers used for ChIP assays are described in Supplementary Table S4.

Morpholino (MO) electroporation

Lissamine-tagged MOs (Gene Tools, LLC) (~0.5 μl of 0.02–0.5 mM) were introduced at the time of injury using a Hamilton syringe. MO delivery to cells was accomplished by electroporation as previously described31. The control, ascl1a- and hb-egfa-targeting MOs have been previously described16,31. The insm1a-targeting MOs are: insm1a MO1 5′-ATGCCCCCGGCAAATCCGCATCTCA-3′ and insm1a MO2 5′-GCTTGACTAAAAATCCTCTGGGCAT-3′. Because antibodies detecting zebrafish Insm1a are unavailable, we evaluated the efficacy of these MOs in vivo in zebrafish embryos using an indirect assay. For this purpose, we prepared cDNA from injured retinas and used primers MO-Hind-insm1a-F and MO-Bam-insm1a-R (Supplementary Table S4) to amplify a 121 bp 5′ insm1a cDNA fragment that harbored both of the insm1a-MO target sites. This product was appended to the coding N-terminus of GFP in the cmv:gfp expression vector and injected into zebrafish embryos with either lissamine-tagged control (0.5mM) or lissamine-tagged insm1a-targeting MOs (0.25mM) in separate experiments. After 24 hrs, embryos were assayed under fluorescent microscopy for GFP expression (Supplementary Fig. S2c).

BrdU labeling, tissue preparation, in situ hybridization (ISH) and immunofluorescence (IF)

BrdU labeling was accomplished by injecting 20 μl of BrdU (20 mM) IP 3 h prior to sacrifice unless otherwise indicated. Some animals received multiple injections of BrdU over multiple days. Fish were overdosed with tricaine methane sulfonate and eyes were dissected, enucleated, fixed and sectioned as previously described3,43. ISH was performed on retinal sections with digoxigenin-labeled cRNA probes (DIG RNA labeling kit, Roche Diagnostics)44. Fluorescent ISH was performed according to the manufacturer’s instructions (Perkin-Elmer). Sense control probes consistently gave no signal above background. IF protocols and antibodies were previously described35,43. Insm1a IF was performed using anti-zebrafish Insm1a antibody (Anaspec, Fermont, CA. Cat No # 55795-2) at 1:100 dilution. For BrdU IF, sections were treated with 2N HCl at 37 °C for 20 min, rinsed in 0.1 sodium borate (pH 8.5) for 10 min and then processed using standard procedures43. For lineage tracing experiments retinal sections from a single eye were distributed across 6 slides. Each slide was first processed for immunofluorescent detection of cell type-specific markers (one marker per slide) and then a 2N HCl epitope retrieval protocol was performed to identify BrdU+ cells3,8. Each slide was used to react with a different cell type-specific marker, while BrdU+ cells were detected on all slides. The total number of BrdU+ cells and the number of co-labeled BrdU+ cells that also stained for a particular cell type marker were quantified on each slide.

Microscopy, cell counts and Statistical Analysis

Slides were examined with a Zeiss Axiophot microscope equipped with Fluorescence optics or an Olympus FluoView FV1000 confocal imaging system. Cell counts were determined by counting fluorescently labeled cells (BrdU+, retinal cell type-specific+ or insm1a+) in retinal sections visualized using fluorescent microscopy. All sections of the retina were examined and at least 3 individuals were used. Data were analyzed for statistical significance using Stat View software (SAS Institute). Statistical comparisons were conducted using a two-tailed unpaired Student’s t test to analyze data from single parameter experiments. For all other experiments an analysis of variance was performed followed by a Bonferroni/Dunn post hoc t test. Error bars represent standard deviation.

Supplementary Material


Supplementary Figure S1:

Injury-dependent regulation of transcriptional repressors. (a) Representative scatter plots used to FACS purify GFP+ cells from gfap:gfp and 1016 tuba:gfp transgenic fish. Cells from wild type (Wt) non-transgenic fish do not fall into the R3 window, whereas GFP+ MG and MG-derived progenitors from transgenic fish sort into this window. (b) Microarray transcriptome analysis of MG and MG-derived progenitors at 4 dpi identifies candidate transcriptional repressors whose expression is regulated by retinal injury. (c) RT-PCR using total retinal RNA shows temporal expression pattern of transcriptional repressors at different times post injury.





Supplementary Figure S2:

insm1a and ascl1a mRNAs co-localize in MG-derived progenitors and Insm1a knockdown prevents MG dedifferentiation. (a) insm1a ISH along with GFP and GS IF show insm1a is expressed in dedifferentiated MG of the injured retina. Asterisk marks the injury site. (b) Control (Ctl) or insm1a-targeting MOs delivered at the time of injury inhibit MG-dedifferentiation as determined by GFP IF in 1016 tuba1a:gfp transgenic zebrafish. Asterisk marks the injury site. (c) Zebrafish embryos were injected with sCMV:gfp expression vectors that had the insm1a MO target sequence appended to the 5′ end of the gfp sequence, along with either a control MO or one of two different insm1a-targeting MOs. Insm1a MOs targeted either the insm1a 5′ UTR or sequences flanking the initiator codon. The number of GFP+ embryos was quantified 24 hrs post injection. (d) IF shows Insm1a protein induction at the injury site at 4 dpi. Asterisk flanks the injury site. (e) Insm1a (green) and BrdU (blue) IF shows that when retinas are electroporated with MOs at the time of injury, insm1a and ascl1a targeting MOs, but not the control MO, knocks down injury-induced Insm1a expression and inhibits cell proliferation when assayed at 4 dpi. Asterisk marks the injury site. Abbreviations: ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. Scale bar: 10 micron in (a, b, d & e); 100 micron in (c).


Supplementary Figure S3:

Injury-dependent insm1a mRNA expression at 6 hpi (a) and 2 dpi (b). Asterisk in (a) marks injury site. ISH (fluorescent detection) was used to assay insm1a at 6 hpi and ISH (peroxidase detection) was used to assay insm1a at 2 dpi. Scale bar: 150 micron.


Supplementary Figure S4:

Insm1a regulates the zone of dedifferentiating MG via an hb-egfa-dependent mechanism. (a) Experimental time line. (b) insm1a ISH shows an expanded zone of insm1a-expressing dedifferentiated MG in the injured and Insm1a knockdown retina compared to the control injured retina. Low and high magnification images are shown. Asterisk marks the injury site. (c) Experimental time line. (d) IF for Dkk-GFP fusion expression after heat shock shows induction in all 3 retinal layers. Lissamine fluorescence shows MO was delivered throughout the retina. Asterisk flanks the injury site. (e, f) BrdU IF shows expansion of the zone of proliferating progenitors following Insm1a knockdown in hsp70:dkk1b-gfp fish is not prevented by Dkk1b overexpression. Asterisk marks the injury site. (g) Top diagram shows experimental time. Insm1a (green) and BrdU (blue) IF shows that delayed electroporation of insm1a targeting MO at 4 dpi, knocks down injury-induced Insm1a expression and stimulates cell proliferation when assayed at 6 dpi. Asterisk marks the injury site. Scale bar: 50 micron in (b, e, f,) left panels; 20 micron in (b, e, f) right panels; 20 micron in (d); 20 micron in (g) upper panels and 10 micron in (g) lower panels.


Supplmentary Figure S5:

insm1a expression correlates with cell cycle exit. (a) insm1a ISH and BrdU IF show that from 4–7 dpi an increasing percentage of BrdU+ cells co-express insm1a. Arrows point to insm1a+/BrdU+ cells and arrowheads point to insm1a/BrdU+ cells. Asterisk flanks the injury site. (b) Candidate cell cycle genes identified in a microarray-based transcriptome analysis of MG and MG-derived progenitors that may be regulated by Insm1a. Abbreviations: ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. Scale bar: 10 micron.


Supplementary Figure S6:

Progenitor differentiation in Insm1a knockdown retinas. Top diagram is the experimental time line. Below are representative photomicrographs of lineage tracing of progenitors (BrdU+) differentiated at 30 dpi at and located at the injury site or located in a region flanking the injury site. Double IF was performed using retinal cell type specific antibodies and anti-BrdU antibodies. Antibodies were as follows: Zpr1, photoreceptor; PKC, bipolar cell; GS, MG; HuC/D, in INL was used to detect amacrine cells and in the GCL was used to detect putative ganglion cells; Zn5, differentiating ganglion cells. The predominant cell type detected in flanking regions of Insm1a knockdown retinas was MG. Abbreviations: ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer; PKC, protein kinase C. Scale bar: 10 micron.


Supplementary Figure S7:

Alignment of zebrafish (ZF) Insm1a and human (HS) Insm1 sequence. Yellow highlighted region is the proline-rich cyclin D1 binding domain found in human Insm1. Zinc fingers are highlighted in green. Asterisks mark conserved residues contributing to zinc finger.


Supplementary Figure S8:

TUNEL staining for apoptotic cells at different times post retinal injury and following various treatments. (a) Ouabin was used as a positive control for cell death and shows extensive cell death in all 3 retinal layers (arrows). (b–c) Minor cell death in the control and Insm1a MO electroporated retina at 2dpi (b) and 4dpi (c). Asterisk marks the injury site. (d) Cell death at 6 dpi following various treatments. Asterisk marks the injury site. Scale bar: 20 microns. (e) Quantification of TUNEL+ cells at various stages of retina regeneration and under various experimental conditions.


This research was supported by NEI grant RO1 EY 018132 from the NIH. We thank David Hyde (University of Notre Dame) for gfap:gfp transgenic fish, Randal Moon (University of Washington) for hsp70:dkk1b-gfp transgenic fish; the University of Michigan Flow Cytometry Core for cell sorting; Alan Dombrowski (Wayne State) for microarray screen; Robert Thompson (University of Michigan) for assistance with microarray data organization; Mike Uhler and David Turner (University of Michigan) for providing the pEL and pCS2 vectors, respectively; J. Beals (University of Michigan) for help with confocal microscopy, the Goldman lab for helpful comments and suggestions during the course of this research and Randall Karr for fish care.


Author Contributions

D.G. and R.R. conceived the study and designed experiments. R.R. and X-F.Z. performed the experiments. R.R., X-F.Z. and D.G. analyzed the data and wrote the manuscript.

Competing Financial Interests

The authors declare no competing financial interests.


1. Bringmann A, et al. Cellular signaling and factors involved in Muller cell gliosis: neuroprotective and detrimental effects. Prog Retin Eye Res. 2009;28:423–51. [PubMed]
2. Sherpa T, et al. Ganglion cell regeneration following whole-retina destruction in zebrafish. Dev Neurobiol. 2008;68:166–81. [PMC free article] [PubMed]
3. Fausett BV, Goldman D. A role for alpha1 tubulin-expressing Muller glia in regeneration of the injured zebrafish retina. J Neurosci. 2006;26:6303–13. [PubMed]
4. Ramachandran R, Fausett BV, Goldman D. Ascl1a regulates Muller glia dedifferentiation and retinal regeneration through a Lin-28-dependent, let-7 microRNA signalling pathway. Nat Cell Biol. 2010a;12:1101–7. [PMC free article] [PubMed]
5. Ramachandran R, Zhao XF, Goldman D. Ascl1a/Dkk/{beta}-catenin signaling pathway is necessary and glycogen synthase kinase-3{beta} inhibition is sufficient for zebrafish retina regeneration. Proc Natl Acad Sci U S A. 2011;108:15858–15863. [PubMed]
6. Thummel R, et al. Pax6a and Pax6b are required at different points in neuronal progenitor cell proliferation during zebrafish photoreceptor regeneration. Exp Eye Res. 2010;90:572–82. [PMC free article] [PubMed]
7. Thummel R, Kassen SC, Montgomery JE, Enright JM, Hyde DR. Inhibition of Muller glial cell division blocks regeneration of the light-damaged zebrafish retina. Dev Neurobiol. 2008;68:392–408. [PMC free article] [PubMed]
8. Ramachandran R, Reifler A, Parent JM, Goldman D. Conditional gene expression and lineage tracing of tuba1a expressing cells during zebrafish development and retina regeneration. J Comp Neurol. 2010;518:4196–4212. [PMC free article] [PubMed]
9. Fimbel SM, Montgomery JE, Burket CT, Hyde DR. Regeneration of inner retinal neurons after intravitreal injection of ouabain in zebrafish. J Neurosci. 2007;27:1712–24. [PubMed]
10. Karl MO, et al. Stimulation of neural regeneration in the mouse retina. Proc Natl Acad Sci U S A. 2008;105:19508–13. [PubMed]
11. Osakada F, et al. Wnt signaling promotes regeneration in the retina of adult mammals. J Neurosci. 2007;27:4210–9. [PubMed]
12. Takeda M, et al. alpha-Aminoadipate induces progenitor cell properties of Muller glia in adult mice. Invest Ophthalmol Vis Sci. 2008;49:1142–50. [PMC free article] [PubMed]
13. Wan J, et al. Preferential regeneration of photoreceptor from Muller glia after retinal degeneration in adult rat. Vision Res. 2008;48:223–34. [PubMed]
14. Qin Z, Barthel LK, Raymond PA. Genetic evidence for shared mechanisms of epimorphic regeneration in zebrafish. Proc Natl Acad Sci U S A. 2009;106:9310–5. [PubMed]
15. Kassen SC, et al. CNTF induces photoreceptor neuroprotection and Muller glial cell proliferation through two different signaling pathways in the adult zebrafish retina. Exp Eye Res. 2009;88:1051–64. [PubMed]
16. Wan J, Ramachandran R, Goldman D. HB-EGF is necessary and sufficient for Müller glia dedifferentiation and retina regeneration. Dev Cell. 2012;22:334–347. [PMC free article] [PubMed]
17. Craig SE, et al. The zebrafish galectin Drgal1-l2 is expressed by proliferating Muller glia and photoreceptor progenitors and regulates the regeneration of rod photoreceptors. Invest Ophthalmol Vis Sci. 2010;51:3244–52. [PMC free article] [PubMed]
18. Hochmann S, et al. Fgf Signaling is Required for Photoreceptor Maintenance in the Adult Zebrafish Retina. PLoS One. 2012;7:e30365. [PMC free article] [PubMed]
19. Kassen SC, et al. Time course analysis of gene expression during light-induced photoreceptor cell death and regeneration in albino zebrafish. Dev Neurobiol. 2007;67:1009–31. [PubMed]
20. Dyer MA, Cepko CL. p27Kip1 and p57Kip2 regulate proliferation in distinct retinal progenitor cell populations. J Neurosci. 2001;21:4259–71. [PubMed]
21. Shkumatava A, Neumann CJ. Shh directs cell-cycle exit by activating p57Kip2 in the zebrafish retina. EMBO Rep. 2005;6:563–9. [PubMed]
22. Topark-Ngarm A, et al. CTIP2 associates with the NuRD complex on the promoter of p57KIP2, a newly identified CTIP2 target gene. J Biol Chem. 2006;281:32272–83. [PMC free article] [PubMed]
23. Lan MS, Russell EK, Lu J, Johnson BE, Notkins AL. IA-1, a new marker for neuroendocrine differentiation in human lung cancer cell lines. Cancer research. 1993;53:4169–71. [PubMed]
24. Pedersen N, et al. Transcriptional gene expression profiling of small cell lung cancer cells. Cancer research. 2003;63:1943–53. [PubMed]
25. Farkas LM, et al. Insulinoma-associated 1 has a panneurogenic role and promotes the generation and expansion of basal progenitors in the developing mouse neocortex. Neuron. 2008;60:40–55. [PubMed]
26. Rosenbaum JN, Duggan A, Garcia-Anoveros J. Insm1 promotes the transition of olfactory progenitors from apical and proliferative to basal, terminally dividing and neuronogenic. Neural Dev. 2011;6:6. [PMC free article] [PubMed]
27. Jacob J, et al. Insm1 (IA-1) is an essential component of the regulatory network that specifies monoaminergic neuronal phenotypes in the vertebrate hindbrain. Development. 2009;136:2477–85. [PubMed]
28. Wildner H, Gierl MS, Strehle M, Pla P, Birchmeier C. Insm1 (IA-1) is a crucial component of the transcriptional network that controls differentiation of the sympatho-adrenal lineage. Development. 2008;135:473–81. [PubMed]
29. Lukowski CM, Ritzel RG, Waskiewicz AJ. Expression of two insm1-like genes in the developing zebrafish nervous system. Gene expression patterns: GEP. 2006;6:711–8. [PubMed]
30. Morris AC, Forbes-Osborne MA, Pillai LS, Fadool JM. Microarray analysis of XOPS-mCFP zebrafish retina identifies genes associated with rod photoreceptor degeneration and regeneration. Investigative ophthalmology & visualscience. 2011;52:2255–66. [PMC free article] [PubMed]
31. Fausett BV, Gumerson JD, Goldman D. The proneural basic helix-loop-helix gene ascl1a is required for retina regeneration. J Neurosci. 2008;28:1109–17. [PMC free article] [PubMed]
32. Powell C, Elsaeidi F, Goldman D. Injury-dependent Muller glia and ganglion cell reprogramming during tissue regeneration requires Apobec2a and Apobec2b. The Journal of neuroscience: the official journal of the Society for Neuroscience. 2012;32:1096–109. [PMC free article] [PubMed]
33. Osada H, et al. Roles of achaete-scute homologue 1 in DKK1 and E-cadherin repression and neuroendocrine differentiation in lung cancer. Cancer Res. 2008;68:1647–55. [PubMed]
34. Johansson TA, Westin G, Skogseid B. Identification of Achaete-scute complex-like 1 (ASCL1) target genes and evaluation of DKK1 and TPH1 expression in pancreatic endocrine tumours. BMC cancer. 2009;9:321. [PMC free article] [PubMed]
35. Lowe AW, et al. Gene expression patterns in pancreatic tumors, cells and tissues. PLoS One. 2007;2:e323. [PMC free article] [PubMed]
36. Tiana G, Krishna S, Pigolotti S, Jensen MH, Sneppen K. Oscillations and temporal signalling in cells. Physical biology. 2007;4:R1–17. [PubMed]
37. Momiji H, Monk NA. Oscillatory expression of Hes family transcription factors: insights from mathematical modelling. Advances in experimental medicine and biology. 2008;641:72–87. [PubMed]
38. Zhang T, Liu WD, Saunee NA, Breslin MB, Lan MS. Zinc finger transcription factor INSM1 interrupts cyclin D1 and CDK4 binding and induces cell cycle arrest. The Journal of biological chemistry. 2009;284:5574–81. [PMC free article] [PubMed]
39. Dyer MA, Cepko CL. p57(Kip2) regulates progenitor cell proliferation and amacrine interneuron development in the mouse retina. Development. 2000;127:3593–605. [PubMed]
40. Dyer MA, Cepko CL. Control of Muller glial cell proliferation and activation following retinal injury. Nature neuroscience. 2000;3:873–80. [PubMed]
41. Stoick-Cooper CL, et al. Distinct Wnt signaling pathways have opposing roles in appendage regeneration. Development. 2007;134:479–89. [PubMed]
42. Lindeman LC, Vogt-Kielland LT, Alestrom P, Collas P. Fish’n ChIPs: chromatin immunoprecipitation in the zebrafish embryo. Methods Mol Biol. 2009;567:75–86. [PubMed]
43. Senut MC, Gulati-Leekha A, Goldman D. An element in the alpha1-tubulin promoter is necessary for retinal expression during optic nerve regeneration but not after eye injury in the adult zebrafish. J Neurosci. 2004;24:7663–73. [PubMed]
44. Barthel LK, Raymond PA. In situ hybridization studies of retinal neurons. Methods Enzymol. 2000;316:579–90. [PubMed]