UBC Faculty Research and Publications

Nr2e1 regulates retinal lamination and the development of Müller glia, S-cones, and glycineric amacrine… Corso-Díaz, Ximena; Simpson, Elizabeth M Jun 20, 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13041_2015_Article_126.pdf [ 4.14MB ]
JSON: 52383-1.0228470.json
JSON-LD: 52383-1.0228470-ld.json
RDF/XML (Pretty): 52383-1.0228470-rdf.xml
RDF/JSON: 52383-1.0228470-rdf.json
Turtle: 52383-1.0228470-turtle.txt
N-Triples: 52383-1.0228470-rdf-ntriples.txt
Original Record: 52383-1.0228470-source.json
Full Text

Full Text

RESEARCH Open AccessNr2e1 regulates retinal lamination and thedevelopment of Müller glia, S-cones, andglycineric amacrine cells during retinogenesisXimena Corso-Díaz1,2 and Elizabeth M. Simpson1,2,3,4*AbstractBackground: Nr2e1 is a nuclear receptor crucial for neural stem cell proliferation and maintenance. In the retina,lack of Nr2e1 results in premature neurogenesis, aberrant blood vessel formation and dystrophy. However, thespecific role of Nr2e1 in the development of different retinal cell types and its cell-autonomous and non-cellautonomous function(s) during eye development are poorly understood.Results: Here, we studied the retinas of P7 and P21 Nr2e1frc/frc mice and Nr2e1+/+ ↔ Nr2e1frc/frc chimeras.We hypothesized that Nr2e1 differentially regulates the development of various retinal cell types, and thus thecellular composition of Nr2e1frc/frc retinas does not simply reflect an overrepresentation of cells born early andunderrepresentation of cells born later as a consequence of premature neurogenesis. In agreement with ourhypothesis, lack of Nr2e1 resulted in increased numbers of glycinergic amacrine cells with no apparent increase inother amacrine sub-types, normal numbers of Müller glia, the last cell-type to be generated, and increased numbersof Nr2e1frc/frc S-cones in chimeras. Furthermore, Nr2e1frc/frc Müller glia were mispositioned in the retina andmisexpressed the ganglion cell-specific transcription factor Brn3a. Nr2e1frc/frc retinas also displayed laminationdefects including an ectopic neuropil forming an additional inner plexiform layer. In chimeric mice, retinal thicknesswas rescued by 34 % of wild-type cells and Nr2e1frc/frc dystrophy-related phenotypes were no longer evident. However,the formation of an ectopic neuropil, misexpression of Brn3a in Müller glia, and abnormal cell numbers in the innerand outer nuclear layers at P7 were not rescued by wild-type cells.Conclusions: Together, these results show that Nr2e1, in addition to having a role in preventing premature cell cycleexit, participates in several other developmental processes during retinogenesis including neurite organization in theinner retina and development of glycinergic amacrine cells, S-cones, and Müller glia. Nr2e1 also regulates variousaspects of Müller glia differentiation cell-autonomously. However, Nr2e1 does not have a cell-autonomous role inpreventing retinal dystrophy. Thus, Nr2e1 regulates processes involved in neurite development and terminal retinalcell differentiation.Keywords: Nr2e1, Amacrines, S-cones, Müller glia, Brn3a, Chimera, Ectopic plexiform layer* Correspondence: simpson@cmmt.ubc.ca1Centre for Molecular Medicine and Therapeutics at the Child and FamilyResearch Institute, University of British Columbia, 950 W 28 Ave, VancouverV5Z 4H4BC, Canada2Genetics Graduate Program, University of British Columbia, Vancouver V6T1Z2BC, CanadaFull list of author information is available at the end of the article© 2015 Corso-Díaz and Simpson. This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly credited. The Creative Commons Public DomainDedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article,unless otherwise stated.Corso-Díaz and Simpson Molecular Brain  (2015) 8:37 DOI 10.1186/s13041-015-0126-xBackgroundDuring retinal development, six neuronal cell types andone type of glia are generated in a conserved histogenicyet overlapping order [1]. Ganglion neurons are thefirst cells to be generated followed by amacrine, hori-zontal, cones, and rods during the embryonic period,and bipolar and Müller glia during the postnatal period[1]. These cell types are organized into three nuclearlayers and generate two plexiform layer neuropilswhere most of the synapses are confined. Retinal pro-genitor cell (RPC) proliferation and fate choice are reg-ulated by intrinsic mechanisms [2–4] and extracellularsignaling from differentiated cells [5–7].During retinogenesis, the cell-cycle length increasesover time in progenitor cells [8]. In various vertebrates,including mice, premature cell cycle exit in the retinaincreases the number of cells that are born early anddecreases the number of cells born towards the end ofretinal histogenesis [9–12]. RPCs are heterogeneous ingene expression [13] and express transcription factorsand cell cycle regulators that play dual roles in control-ling RPC cell cycle and fate [12, 14, 15].Nr2e1 is a conserved orphan nuclear receptor thatregulates neural stem cell proliferation and mainten-ance [16]. Nr2e1-null mice have smaller brains andretinas, are blind, and highly aggressive [17]. Nr2e1 ismostly a repressor that operates in a cell-autonomousfashion, but it can also activate gene transcription [18,19] and act non-cell-autonomously through the Wntpathway to regulate neural stem cell proliferation andself-renewal [19]. Lack of Nr2e1 results in prematurecell cycle exit during corticogenesis and reduced thick-ness of superficial cortical layers due to a depletion ofthe neural stem cell pool [16]. Lack of Nr2e1 in theretina results in precocious neurogenesis, impairedblood vessel development [20], and progressive dystrophy[21, 22]. This complex phenotype poses a challenge tounderstanding the role of Nr2e1 in specific retinal cellpopulations.Chimeras provide valuable information regarding theautonomous and non-autonomous cellular consequencesof gene mutations, the development of different cell-types and their interaction through cell-signaling, as wellas the nature of tissue-tissue interactions in vivo [23].To better understand the role(s) of Nr2e1 in retinaldevelopment, we studied the cellular composition andmorphology of Nr2e1frc/frc, and Nr2e1+/+↔Nr2e1frc/frcchimeric mouse retinas. We found that dystrophy-relatedphenotypes in Nr2e1frc/frc retinas are not generated cell-autonomously. In addition, we found that lack of Nr2e1results in an ectopic plexiform layer in the inner retina,aberrant development of Müller glia and a bias towardsthe generation of glycinergic amacrine cells, S-cones andMüller glia.ResultsTo get insight into the cell autonomy of Nr2e1 duringretinogenesis we used Nr2e1frc/frc and chimeric micecomprised of both Nr2e1frc/frc and wild-type cells. Westudied abnormal phenotypes previously reported to bepresent in Nr2e1 null retinas, such as reduced retinalthickness and blood vessel numbers. We later focusedon the role of Nr2e1 in cell type development by studyingthe numbers and localization of different cell types.Expression of EGFP and β-galactosidase in mousechimerasTo better understand the cell-autonomous and non-cell autonomous roles of Nr2e1 during retinogenesis,we made chimeric mice comprised of Nr2e1+/+ andNr2e1frc/frc cells, herein referred as Wt↔frc chimeras.Experimental and control chimeric mice were made byblastocyst injection of Nr2e1+/+ or Nr2e1frc/frc embryonicstem cells (ESCs) harboring a ubiquitous-expressing EGFPtransgene (Additional file 1: Figure S1A and B). Incontrast, host blastocyst contained the lacZ gene underthe control of ROSA26 promoter (R26lacZ) thus expressingβ-galactosidase (β-gal) (Additional file 1: Figure S1B). Inthis way ESC-derived cells could be identified by the greenepifluorescence of EGFP and blastocyst-derived cells bythe enzymatic product of β-gal.We used two different embryonic stem cell lines pergenotype to control for possible cell-line specific traits thatcould affect the phenotype of the mice. Four Wt↔Wt andfour Wt↔frc chimeras were studied at P7. Nine Wt↔Wtand ten Wt↔frc chimeras were studied at P21. Eyes fromthese chimeras were subjected to funduscopy and col-lected for cryosectioning. First, we determined that theEGFP and β-gal markers were expressed appropriately inthe chimeras. We assessed the expression of β-gal by itsenzymatic activity and could clearly observe the blue pre-cipitate formed by the hydrolysis of X-gal in perinuclearregions (Additional file 1: Figure S1C). Importantly, thisenzymatic reaction did not interfere with the EGFP epi-fluorescence and both markers were expressed in mutuallyexclusive regions of the chimeric retinas (Additional file 1:Figure S1D). We assessed the percentage of chimerism bymeasuring the area displaying EGFP epifluorescence inthe ONL plus INL of each retina and comparing it to thetotal ONL plus INL area. We excluded the IPL and GCLto decrease the interfering signal recovered from neuralprocesses. Thus, we were able to use these two markersreliably as indicators of the origin of the different celltypes; ESC or host blastocyst.Nr2e1frc/frc reduced retinal thickness and blood vesselnumbers were rescued in Wt↔frc chimerasNext, we assessed whether the reduced retinal thicknessand blood vessel numbers of Nr2e1frc/frc retinas could beCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 2 of 21rescued by wild-type cells and how many wild-type cellswould be needed to achieve rescue. Detailed informationfor each chimera is given in Table 1.We took fundus images of the chimeric eyes andmanually counted the numbers of blood vessels in eachchimera. We found a very poor correlation (R2 = 0.2) be-tween the numbers of blood vessels and the percentageof chimerism in Wt↔frc eyes (Fig. 1a). Interestingly,only two chimeric eyes with the lowest percentage ofchimerism (39 % and 45 %) had normal blood vesselnumbers (10 and 9 blood vessels, respectively; Z-scorehigher than −3), suggesting that a contribution ≥55 %wild-type cells was necessary to rescue normal bloodvessel development (Fig. 1a).We measured the retinal thickness in a central sectioncontaining the optic nerve and two other adjacent sec-tions 60 μm away in both directions. We found that thepercentage of chimerism did not affect the retinal thick-ness of Wt↔Wt chimeras (Fig. 1b) but it affected thethickness of Wt↔frc chimeras (R2 = 0.67). In this case, a66 % chimeric retina had a retinal thickness comparableto wild type (146.33 ± 5.78 μm, Z-score greater than −3)suggesting that ≥34 % wild-type cells were needed torescue it (Fig. 1b). Only the two highest percentagechimeric retinas (67 % and 86 %) were not rescued bywild-type cells having thinner than wild-type retinas(127.5 ± 7.5 and 108.33 ± 4.81 μm, respectively, Z-scorelower than −3).In summary, a high contribution of wild-type cells(≥55 %) was needed to achieve rescue of blood vesselswhereas retinal thickness was restored by fewer wild-typecells (≥34 %). The rescue of retinal thickness suggests thatcell loss is not regulated cell autonomously by Nr2e1. Wethen studied whether other retinal abnormalities occurringTable 1 Characteristics of the chimerasID Age Chimerism % ESC line Genotype Blood vessel # Retinal thickness (μM)7586 P7 48 mEMS4922 Nr2e1frc/frc na 213.63 ± 9.367585 P7 58 mEMS4922 Nr2e1frc/frc na 234.90 ± 13.847577 P7 62 mEMS4914 Nr2e1frc/frc na 205.67 ± 10.097584 P7 70 mEMS4922 Nr2e1frc/frc na 376.15 ± 38.777582 P21 39 mEMS4914 Nr2e1frc/frc 10 183.67 ± 2.337588 P21 42 mEMS4914 Nr2e1frc/frc 4 197.67 ± 5.617599 P21 45 mEMS4922 Nr2e1frc/frc 9 146.33 ± 5.787597 P21 54 mEMS4922 Nr2e1frc/frc nd 160.00 ± 0.007587 P21 61 mEMS4914 Nr2e1frc/frc 7 170.00 ± 7.647583 P21 64 mEMS4914 Nr2e1frc/frc 1 160.00 ± 5.777600 P21 66 mEMS4922 Nr2e1frc/frc 5 149.33 ± 4.707604 P21 67 mEMS4922 Nr2e1frc/frc 5 127.50 ± 7.507589 P21 77 mEMS4914 Nr2e1frc/frc 1 Irregular morphology7598 P21 86 mEMS4922 Nr2e1frc/frc 7 108.33 ± 4.817578 P7 29 mEMS4919 Nr2e1+/+ na 204.43 ± 18.807580 P7 46 mEMS4919 Nr2e1+/+ na 259.95 ± 42.377579 P7 51 mEMS4919 Nr2e1+/+ na 196.42 ± 4.997581 P7 71 mEMS4919 Nr2e1+/+ na 180.5 ± 8.597612 P21 0.3 mEMS4926 Nr2e1+/+ 11 185.00 ± 4.937611 P21 10 mEMS4926 Nr2e1+/+ 11 190.33 ± 14.907610 P21 11 mEMS4926 Nr2e1+/+ 11 159.33 ± 6.337593 P21 18 mEMS4919 Nr2e1+/+ 11 178.00 ± 8.007592 P21 45 mEMS4919 Nr2e1+/+ 10 193.67 ± 5.847595 P21 50 mEMS4919 Nr2e1+/+ 13 172.67 ± 11.857591 P21 54 mEMS4919 Nr2e1+/+ 11 Irregular morphology7594 P21 58 mEMS4919 Nr2e1+/+ 10 190 ± 0.007590 P21 92 mEMS4919 Nr2e1+/+ 9 204.33 ± 7.17na, not applicable; nd, not doneCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 3 of 21in Nr2e1 null mice including gliosis and gross structuraldefects could also be rescued by wild-type cells in chimericretinas.Nr2e1frc/frc Müller glia misexpression of GFAP and retinalstructural defects were rescued in Wt↔frc chimerasAs previously reported [21], we found that centralMüller glia of Nr2e1frc/frc retinas express GFAP (Fig. 2a),a neurofilament protein that is normally expressed onlyin astrocytes and reactive Müller glia [24]. We alsoobserved that, in addition to its thinner size, Nr2e1frc/frcretinas present structural defects including regions ofINL and ONL that overlap with an apparent loss of theOPL (Fig. 2b). We observed that these defects occurred2.43 ± 1.36 times per mm in the mutant retinas.We found that none of the chimeric retinas, eventhose with a very high contribution of Nr2e1frc/frc cells,express GFAP in central Müller glia, suggesting thatwild-type cells rescue this defect (Fig. 2c). We alsoobserved that the retina of high percentage chimeraswas still very thin but did not have the structural defectsseen in Nr2e1frc/frc retinas (Fig. 2c).In summary, misexpression of GFAP and gross structuraldefects do not emerge cell-autonomously in Nr2e1frc/frcretinas and are likely a consequence of other defects suchas abnormal cell type development. We then focused onstudying the numbers of each retinal cell type in P7Nr2e1frc/frc and chimeric retinas.Lack of Nr2e1 resulted in increased numbers ofglycinergic amacrine cells and S-cones, and normalnumbers of Müller gliaAlthough the role of Nr2e1 in regulating retinal thick-ness has been assessed [21, 22], the numbers of differentretinal cell types in Nr2e1-null retinas has not yet beenstudied. Similarly, the presence of each retinal cell typehas previously only been assessed in adult eyes afterapoptosis has already severely altered the structure andcomposition of the retina [21, 22]. To gain better insightinto the cell numbers generated in Nr2e1frc/frc mouseretinas, we studied the retina at P7. At this time pointmost cells have already been differentiated and less than5 % of bipolar and Müller glia are still being generatedin the rat retina [25]. At P7, cells have not been exposedto daily visual activity and are less prone to apoptosis inNr2e1 null retinas [22]. To quantitatively assess differencesin cell numbers, we manually counted all cell types in theretinas of wild-type and Nr2e1frc/frc mice at P7. Fivesections throughout the eyes were analyzed.To better understand the interplay between wild-typeand Nr2e1frc/frc cells during the development of chimericretinas, and to better assess clonal composition, welooked at the density of different retinal cell types be-longing to each genotype in Wt↔frc chimeras at P7.The number of each EGFP positive or EGFP negative celltype was divided over the EGFP-positive or EGFP-negativearea, respectively.Previous studies have shown a reduced outer nuclearlayer (ONL) but have conflicting results regarding thethickness of the inner nuclear layer (INL) in P4-P7Nr2e1 mutant retinas [21, 22]. Miyawaki et al., reporteda thicker INL [21] whereas Zhang et al., reported athinner INL [22]. We observed, in agreement with areduced ONL, a 33 % reduction in rods in Nr2e1frc/frcretinas when compared to wild type (Additional file 2:Figure S2A and C). We found no difference in INLthickness between wild-type and Nr2e1frc/frc P7 centralretinas measured in sections taken through the opticnerve (Wt = 85.9 ± 0.17; frc = 85.86 ± 3.27; n = 3 for Wt;n = 3 for frc).Figure 1 Reduced retinal thickness and blood vessel numbers ofNr2e1frc/frc retinas were rescued in Wt↔frc chimeras. a The bloodvessel number of each chimera was assessed by funduscopy. Scatterplot showed normal blood vessel numbers in chimeric retinascontaining up to 45 % Nr2e1frc/frc cells. n= 9 for Wt↔Wt, n= 9 forWt↔frc. b Retinal thickness in each chimera was measured in 3sections; a central section containing the optic nerve, and two sections60 μm away in both directions. Scatter plot showed normal retinalthickness in chimeric retinas containing up to 66 % Nr2e1frc/frc cells. n= 8for Wt↔Wt, n= 9 for Wt↔frc; *, Z-score≤−3; error bars represent SEMCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 4 of 21We also found a 27 % reduction in bipolar cells inNr2e1frc/frc retinas when compared to wild type (Additionalfile 2: Figure S2B and C). A reduction in the numbers ofrods and bipolar cells is expected given that they differenti-ate at later time-points in the mouse retina [26]. However,since the thickness of the Nr2e1frc/frc INL is not overallreduced at P7, other cell types must be overrepresented inthese retinas.In Wt↔frc chimeras, Nr2e1frc/frc rods (Additional file2: Figure S2D and F) and bipolar cells (Additional file 2:Figure S2E and G) were 50 % and 53 % less abundant,respectively, compared to wild type. This suggests thatwild-type cells cannot rescue the numbers of Nr2e1frc/frcrods and bipolar cells that fail to differentiate.We also observed a reduction of Brn3a-positiveganglion cells in Nr2e1frc/frc retinas, concomitantly withan ectopic expression of this marker in the ventral retina(Fig. 3a). This ectopic expression was seen in cell somaslocated in the INL and in cytoplasmic projections withtermini that resemble Müller glia end-feet in the GCL(Fig. 3a and b). Müller glia end-feet normally have acobblestone pattern and enclose the ganglion cell soma[27]. Quantification of cell numbers in the GCL revealedthat Nr2e1frc/frc retinas have a 63 % decrease in Brn3apositive ganglion cells (Fig. 3c). However, this lowernumbers were rescued in Wt↔frc chimeric retinas(Fig. 3d and e). Furthermore, we also observed manyBrn3a positive cells that aberrantly localized in the IPLand adjacent INL of chimeric retinas (Fig. 3d).In contrast to the reduction in rods, bipolar, andganglion cells, Pax-6 and syntaxin-1A positive amacrinecells were increased in numbers in the INL of Nr2e1frc/frcretinas, representing the majority of cells in this layer(Fig. 4a and b). Quantification of amacrine cells in theGCL (assessed as Pax-6 positive minus Brn3a positivecells) revealed a 45 % increase in mutant amacrine cellscompared to wild type (Fig. 4h). To better characterizethese cells, we stained the retina with antibodies for pro-teins expressed in subpopulations of amacrine cells. Inter-estingly, the excess Pax-6 and syntaxin-1A positive cellswere negative for the GABAergic amacrine-cell markersGABA, calretinin, and Islet-1/2 but positive for glycinetransporter 1 (GlyT1), a marker of glycinergic amacrines(Fig. 4d-g). This suggests that Nr2e1frc/frc retinas generatedpreferentially an excess of glycinergic amacrine cells.In the INL of Wt↔frc chimeric retinas, Nr2e1frc/frcamacrine cells were 49 % more abundant than wildtype(Fig. 5a and b). In agreement with this increase, there wasan overrepresentation of mutant GLYT1 positive cells inchimeras (Additional file 3: Figure S3). In contrast, thenumbers of amacrine cells in the GCL of Wt↔frcchimeric retinas were comparable to wild type (Fig. 5c),suggesting that wild-type cells prevented the migration orsurvival of excess amacrine cells in the GCL.Figure 2 Nr2e1frc/frc-Müller-glia misexpression of GFAP and retinal structural defects were rescued in Wt↔frc chimeras. Transverse retinal sections fromP21 mice were immunostained for GFAP (red). a GFAP was observed in the GCL of wild-type and Nr2e1frc/frc retinas but also in the processes of Müllerglia in Nr2e1frc/frc retinas (arrow). b Nr2e1frc/frc retinas had abnormal INL intrusions into the ONL (asterisk). c A Wt↔frc chimera with 86 % Nr2e1frc/frc cells(EGFP positive, green) showed absence of GFAP expression in Müller cells and absence of INL intrusions into the ONL. GCL, ganglion cell layer; INL,inner nuclear layer; ONL, outer nuclear layer; Hoechst, nuclear counterstain (blue); scale bar = 50 μm. n = 9 for Wt↔Wt, n = 10 for Wt↔frcCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 5 of 21Figure 3 Nr2e1frc/frc P7 retinas had reduced numbers of ganglion cells that were rescued in Wt↔frc chimeras. Transverse retinal sections from P7Nr2e1+/+, Nr2e1frc/frc, and chimeric mice were immunostained for Brn3a (ganglion cells). a Nr2e1frc/frc retinas had less Brn3a positive cells (green)in theGCL and misexpressed Brn3a in cells ofthe ventral INL that resembled Müller glia b Magnification of the rectangle in A showing Brn3anuclear stainingin some GCL cell nuclei (arrow heads) and cytoplasmic projectionsresembling Müller glia end-feet c Ganglion cells were counted throughout fivesections across the retina of Nr2e1+/+ and Nr2e1frc/frc mice. Numbers were normalized to retinal length and expressed as percentages of Nr2e1+/+ cellnumbers. Reduced numbers of ganglion cells were observed in Nr2e1-mutant retinas compared to wild type (63 % decrease). d In Wt↔Wt and Wt↔frcchimeras, the density of EGFP positive (green) ganglion cells (arrows), appeared similar to the density of EGFP negative ganglion cells. Representativeimages of a 71 % Wt↔Wt and a 58 % Wt↔frc chimeric retina are shown. Brn3a positive cells were mislocalized in the IPL (arrowheads) and INL ofWt↔frc retinas. e Quantification of the density of ganglion cells that were derived from host blastocyst or ESCs in chimeras was assessed bycounting single-labeled cells (Brn3a positive and EGFP negative) or double-labeled cells (Brn3a positive and EGFP positive) and dividing themby the EGFP negative or EGFP positive retinal area (ONL + INL), respectively. The density of Nr2e1frc/frc ganglion cells (EGFP positive) was similar tothe density of wild-type ganglion cells (EGFP negative) in Wt↔frc chimeras. n= 3 for Nr2e1+/+, n= 3 for Nr2e1frc/frc, n= 3 for Wt↔Wt, n= 3 for Wt↔frc;*,P≤ 0.05; ns, not significant; error bars represent SEM. GCL, ganglion cell layer; Hoechst, nuclear counterstain (blue); INL, inner nuclear layer; neg., negative;ONL, outer nuclear layer; pos., positive; scale bar = 50 μmCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 6 of 21In P7 Nr2e1frc/frc retinas, the densities of horizontal,cones, and Müller glial cells were similar to wild-type(Fig. 6a-d). Interestingly, these densities were also normal atP21 (Fig. 6e) suggesting that horizontal, cones, and Müllerglial cells were spared from the excessive apoptosis that oc-curs during the postnatal period in Nr2e1frc/frc retinas [22].Similarly, there was no difference in the number ofS-cones labeled with S opsin between P7 Nr2e1frc/frc andNr2e1+/+ retinas (Fig. 7a and b). However, in P7 Wt↔frcchimeras the difference between Nr2e1frc/frc and Nr2e1+/+S-cone numbers was very obvious (Fig. 7c, arrow). Quanti-fication of S Opsin positive cells in P7 Wt↔frc revealed aFigure 4 Nr2e1frc/frc P7 retinas had increased numbers of glycinergic amacrine cells. Transverse retinal sections from P7 Nr2e1+/+ and Nr2e1frc/frcmice were immunostained for the pan-amacrine markers Pax-6 and syntaxin-1A, and the subclass markers Islet-1/2, GABA, calretinin, and glycine transporter 1(GlyT1). a Retinal section showing amacrine cells immunostained with Pax-6 (red) and syntaxin-1A (green). Nr2e1frc/frc retinas had more Pax-6 and syntaxin-1Apositive cells than wild-type retinas, and these cells occupied the majority of the INL. An ectopic plexiform layer (EPL) is also indicated (open arrow). Magnificationof the Pax-6 (red) and syntaxin-1A (green) staining is shown for a region of b Nr2e1+/+ or c Nr2e1frc/frc retina (rectangle). d-f Neither (d) Islet-1/2 positive(green) (arrows), (e) GABA positive (red), or (f) calretinin positive amacrine cells appeared increased in Nr2e1-mutant retinas. g In contrast, GlyT1positive amacrine cells appeared increased in the INL of Nr2e1-mutant retinas. EPL (open arrow). h Amacrine cells were counted in the GCL throughoutfive sections across the retina of Nr2e1+/+ and Nr2e1frc/frc mice. Numbers were normalized to retinal length and expressed as percentages of Nr2e1+/+cell numbers. Increased numbers of amacrine cells (assessed as Pax6 positive minus Brn3a positive) were observed in the GCL of Nr2e1-mutant retinascompared to wild type (45 % increase). n = 3 for Nr2e1+/+, n = 3 for Nr2e1frc/frc; *, P≤ 0.05; error bars represent SEM. EPL, ectopic plexiform layer;GCL, ganglion cell layer; Hoechst, nuclear counterstain (blue); INL, inner nuclear layer; scale bar = 50 μm in all images except for B where it represents 12.5 μmCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 7 of 2148 % increase in Nr2e1frc/frc S-cones compared to wild type(Fig. 7c and d).In summary, P7 Nr2e1frc/frc retinas have a marked reduc-tion of ganglion, bipolar, and rod cells, increased numbersof glycinergic amacrine cells, and normal numbers of cones,horizontal, and Müller glial cells. In Wt↔frc P7 chimericretinas, an increase in Nr2e1frc/frc S-cone numbers was alsoevident. In the same chimeras, wild-type cells did not res-cue Nr2e1frc/frc abnormal cell numbers with the exceptionof ganglion and amacrine cells in the GCL. We furtherstudied whether these cellular defects were accompanied byabnormalities in retinal lamination as explained below.Nr2e1frc/frc retinas displayed an ectopic plexiform layerand a disorganized inner plexiform layer, which were notrescued by wild-type cells in Wt↔frc retinasWe observed that Nr2e1frc/frc retinas displayed an ectopicplexiform layer (EPL) in the inner nuclear layer (INL) evidentat P7 (Fig. 8a and b). Looking back, we can see that the EPLwas labeled with the amacrine markers syntaxin-1A andGlyT1 throughout the retina (Fig. 4a and g, open arrow).To evaluate whether bipolar cells that normally establishsynaptic connections with retinal ganglion cells (RGCs)and amacrine cells in the inner plexiform layer (IPL) [28]also had terminals in the EPL, we stained retinas forPKC-alpha expressed in ON bipolars. We found thatPKC-alpha positive bipolar cells had mislocalized axonbranchings in the EPL (Fig. 8c and d, solid arrow).We then asked if the lamination defects generated byNr2e1 mutant cells could be rescued by extracellular signalsfrom wild-type cells in Wt↔frc chimeras. At P7 the EPLwas present in all four chimeras, and regions of wild-typecells failed to form this layer, showing that wild-type cellscannot rescue this lamination defect or be influenced bymutant cells to form this ectopic layer (Fig. 8e).Furthermore, the IPL of Wt↔frc chimeras was thickerthroughout the whole retina and less defined than that ofWt↔Wt chimeras suggesting that the development of theIPL is also affected by the lack of Nr2e1 (Fig. 8e, brackets).To better characterize the EPL, we stained Nr2e1frc/frcP7 retinas with antibodies that normally recognize theinner plexiform layer (IPL) and found that unlikeFigure 5 Increased numbers of Nr2e1frc/frc amacrine cells were not rescued in P7 Wt↔frc chimeras. Chimeric eyes were immunostained for thepan-amacrine marker Pax-6. a In Wt↔Wt chimeras, the density of EGFP positive (green) amacrine cells appeared similar to the density of EGFPnegative amacrine cells. In Wt↔frc chimeras, the density of amacrine cells that were EGFP positive (green) appeared higher than the density ofEGFP negative amacrine cells. Representative images of a 29 % Wt↔Wt and a 58 % Wt↔frc chimeric retina are shown. The arrow shows a regionwith high numbers of mutant cells (EGFP positive) where amacrine cells (Pax-6 positive) are overrepresented. b,c Quantification of the density ofamacrine cells that were derived from host blastocyst or ESCs in chimeras was assessed by counting single-labeled cells (Pax-6 positive and EGFPnegative) or double-labeled cells (Pax-6 positive and EGFP positive) and dividing them by the EGFP negative or EGFP positive retinal area (ONL + INL),respectively. Cell counts were performed in both the INL and the GCL. In the INL, large and round nuclei adjacent to the OPL, presumably belonging tohorizontal cells, were excluded from counting. Amacrine numbers in the GCL were obtained by subtracting Brn3a positive ganglion cells from the totalPax-6 positive cells in this layer. In Wt↔frc chimeras, the density of Nr2e1frc/frc amacrine cells (EGFP positive) was (b) higher (49 % increase) than the densityof wild-type amacrine cells in the INL but was (c) similar to wild type in the GCL. n= 3 for Wt↔Wt, n= 3 for Wt↔frc; *, P≤ 0.05; ns, not significant; errorbars represent SEM; GCL, ganglion cell layer; Hoechst, nuclear counterstain (blue); INL, inner nuclear layer; neg., negative; pos., positive; scale bar = 50 μmCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 8 of 21syntaxin-1A and GlyT1, calbindin and mGluR1 were oc-casionally but not always found in this ectopic layer(Additional file 4: Figure S4A-D).Moreover, some calbindin and mGluR1 positive cellswere observed in proximity to the EPL of Nr2e1frc/frcretinas (Additional file 4: Figure S4B and D, asterisks)and Pax-6-positive cells were found between mGluR1positive sublaminae (Additional file 4: Figure S4D,arrowhead) revealing lamination defects.We also looked for the EPL in P21 retinas with anti-bodies against syntaxin-1A and GlyT1. We found thatsyntaxin-1A labeled the EPL in some regions of the retinadespite extensive retinal dystrophy (Fig. 9a and b). Wecould not detect the EPL with GlyT1 but observed manyGlyT1 positive interplexiform amacrine cells extendingneurites towards the OPL in Nr2e1frc/frc retinas (Fig. 9cand d). In P21 Wt↔frc chimeras there were manyNr2e1frc/frc GlyT1 positive interplexiform amacrine cellssuggesting that this phenotype is not corrected by wild-type cells (Fig. 9e-g).Taken together, these results suggested a role of Nr2e1in constraining the neurites of INL neurons to an orga-nized IPL. P21 chimeric retinas also revealed a rescue ofamacrine cell numbers by wild-type cells that we furtherstudied.Nr2e1frc/frc amacrine cell loss is rescued by wild-type cellsin Wt↔frc retinasWe observed few syntaxin-1A and GlyT1 positive cells inP21 Nr2e1frc/frc retinas suggesting extensive cell loss ofamacrine cells as expected from dystrophy (Fig. 9a-d).However, in P21 Wt↔frc chimeras the numbers ofNr2e1frc/frc GlyT1 positive amacrine cells appeared normal(Fig. 9h). Quantification of syntaxin-1A cell numbers in theINL of chimeras revealed normal numbers of Nr2e1frc/frcamacrine cells (Additional file 5: Figure S5), suggesting thatwild-type cells prevented the excessive loss of mutantamacrine cells during the postnatal period and thatamacrine cell death is not a primary consequence ofNr2e1 loss.The primary roles of Nr2e1 in the retina are then theregulation of RPC cell cycle, cell type numbersand innerneurite development. Nr2e1 remains to be expressed in theadult Müller glia and thus it may also have specific primaryroles in this cell type. Therefore, we further characterizedMüller glia with specific markers as shown below.Nr2e1frc/frc Müller glia were aberrantly positioned in the innernuclear layer and cell-autonomously misexpressed Brn3aNr2e1 is expressed in the Müller glia of adult mice [29].To better characterize the expression of Nr2e1 in Müllerglia during postnatal development, we stained P3, P7,P14, and P21 retinas with antibodies against β-gal andSOX-2 in mice expressing β-gal under the control of thehuman NR2E1 promoter (NR2E1-lacZ) [29]. We madeuse of this mouse strain due to the lack of a reliablecommercial antibody for Nr2e1. Previous work by othershas shown that SOX-2 is expressed in progenitors and asubpopulation of amacrine cells starting at P0. By P7,Figure 6 Nr2e1frc/frc retinas had normal numbers of horizontal, cone,and Müller glia cells. Transverse retinal sections from P7 and P21 Nr2e1+/+and Nr2e1frc/frc mice were immunostained for calbindin (horizontal cells,adjacent to the ONL), arrestin (cones), and SOX-9 (Müller glia). a-c P7 retinalsections showing (a) horizontal cells in green, (b) cones in red, and (c)Müller glia in green. d,e Each retinal cell type was counted throughout fivesections across the retina of (d) P7 or (e) P21 Nr2e1+/+ and Nr2e1frc/frc mice.Numbers were normalized to retinal length and expressed as percentagesof Nr2e1+/+ cell numbers. Normal numbers of horizontal, cone, and Müllerglia cells were observed in Nr2e1-mutant retinas compared to wild type atboth time-points. n=3 for Nr2e1+/+, n=3 for Nr2e1frc/frc; *, P≤ 0.05; ns, notsignificant; error bars represent SEM. GCL, ganglion cell layer; INL, innernuclear layer; ONL, outer nuclear layer; scale bar = 50 μmCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 9 of 21SOX-2 expression becomes mainly restricted to theinner nuclear layer in a subpopulation of amacrine cellsand in Müller glia. This expression in amacrine cells andMüller glia continues through adulthood [30]. We ob-served co-localization of SOX-2 and β-gal at all time-points in cells that extend apical and basal processesindicative of Müller glia (Additional file 6: Figure S6A).This suggests sustained expression of Nr2e1 in Müllerglia throughout postnatal development. Furthermore, weobserved SOX-2 positive cells in Nr2e1frc/frc retinas at alltime-points (Additional file 6: Figure S6B), further show-ing that Nr2e1frc/frc Müller glia numbers are comparableto wild type.Several defects in the Müller glia of Nr2e1frc/frcretinas were noticed suggesting a cell-autonomousrole of Nr2e1 in these cells. For example, while wild-type Müller glia (SOX-2 positive) are positioned inthe middle of the INL below bipolar cells, Nr2e1frc/frcMüller glial somas are located adjacent to the OPLintermingling with bipolar cells (Fig. 10a-b). To ascer-tain whether this defect could be rescued by wild-type cells, we stained P7 chimeras for SOX-2. Wefound that, in Wt↔frc chimeras, the soma ofNr2e1frc/frc Müller glia was still mislocalized whereaswild-type Müller glia localized correctly within the INL(Fig. 10c-e).Figure 7 Nr2e1frc/frc S-cones were overrepresented in Wt↔frc chimeras. Transverse retinal sections from P7 Nr2e1+/+, Nr2e1frc/frc, and chimeric micewere immunostained for S opsin (S-cones). a Retinal sections showing S-cones labeled with S Opsin (red). b To assess differences in S-cones numbersbetween wild-type and Nr2e1-mutant retinas, S opsin positive cells were counted throughout five sections across the retina of P7 Nr2e1+/+ and Nr2e1frc/frc mice. Numbers were normalized to retinal length and expressed as percentages of Nr2e1+/+ cell numbers. There was no significant differencebetween the numbers of S opsin positive cells in wild-type and Nr2e1-mutant retinas. c Retinal sections of chimeras showing S opsin in red. The densityof Nr2e1frc/frc S opsin positive cells appeared higher in Nr2e1-mutant regions (EGFP positive) compared to wild-type regions. The arrows show a regionwith Nr2e1-mutant cells and a high density of S-cones. Representative images of a 50 % Wt↔Wt and a 39 % Wt↔frc chimeric retina are shown. dQuantification of the density of S-cone cells that were derived from host blastocyst or ESCs in chimeras was assessed by counting single-labeled cells(S opsin positive and EGFP negative) or double-labeled cells (S opsin positive and EGFP positive) and dividing them by the EGFP negative or EGFPpositive retinal area (ONL + INL), respectively. In Wt↔Wt chimeras, the density of EGFP positive and EGFP negative S-cone cells was similar while inWt↔frc chimeras the density of Nr2e1frc/frc S-cone cells (EGFP positive) was higher (48 % increase) than the density of wild-type S-cone cells (EGFPnegative). n = 3 for Nr2e1+/+, n = 3 for Nr2e1frc/frc, n = 3 for Wt↔Wt, n = 3 for Wt↔frc; *, P≤ 0.05; neg., negative;ns, not significant; error bars representSEM. GCL, ganglion cell layer; Hoechst, nuclear counterstain (blue); INL, inner nuclear layer; ONL, outer nuclear layer; pos., positive; scale bar = 50 μmCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 10 of 21In addition, Nr2e1frc/frc Müller glia misexpress thetranscription factor Brn3a, which is normally expressedin sensory neurons including ganglion cells [31] (Fig. 11aand b). Interestingly, misexpression of this marker wasonly present in the ventral retina and was also mislocal-ized to the cell soma and processes instead of beingrestricted to the nuclear compartment (Fig. 11). Thestaining of Brn3a in the ventral retina appears diffuseand non-specific at first but careful examination sug-gested cytoplasmic expression in Müller glial projec-tions, which span the whole retina and branch into theplexiform layers. To further confirm that the misexpres-sion of Brn3a occurs in Müller glia, we also stained ret-inas for the Müller glia marker vimentin and observedits co-localization with Brn3a at both P7 and P21(Fig. 11c-f and Additional file 7: Figure S7). Importantly,in P7 Wt↔frc chimeras the misexpression of Brn3a inMüller glia was evident in mutant but not in wild-typecells (Fig. 11g).Overall, these results suggested that Nr2e1 regulatesthe maturaion of Müller glia cell-autonomously.DiscussionThe phenotype of the Nr2e1-null eye is very complexinvolving various cell types and tissues, thus makingthe identification of the primary function of Nr2e1 dur-ing eye development challenging. In this study we usedchimeras to clarify the cell-autonomous roles of Nr2e1during retinogenesis, and described novel phenotypesof the Nr2e1frc/frc mice in the P7 retina before extensivedystrophy has occurred. Our results suggest that Nr2e1regulates the development of specific retinal cell-typesand the organization of inner retinal neurites.Nr2e1 does not prevent retinal dystrophy cell-autonomouslyThe thickness of the P21-P28 Nr2e1 mutant retina ishighly reduced compared to wild type due to progressivepostnatal cellular loss [21, 22]. Our findings suggest thatwild-type cells can rescue cell loss in Wt↔frc chimeras.First, we showed that in P21 Wt↔frc chimeras, the re-duced Nr2e1frc/frc retinal thickness could be rescued by≥34 % wild-type cells. Second, we observed that postna-tal amacrine cell loss is rescued in P21 Wt↔frc chimericFigure 8 Nr2e1frc/frc retinas displayed an ectopic plexiform layer that was not rescued in Wt↔frc chimeras. Transverse retinal sections from P7Nr2e1+/+, Nr2e1frc/frc, and chimeric mice were immunostained for PKC-alpha or syntaxin-1A. a A region devoid of Hoechst-stained nuclei indicatedan ectopic plexiform layer (EPL) in Nr2e1frc/frc retinas (open arrow). b Magnification of the box in A showing the EPL (open arrow). c PKC-alpha positivebipolar cells (green) extended processes into the IPL in wild-type retinas but also to the EPL (open arrow) in Nr2e1-mutant retinas. d Magnification ofthe box in G showing bipolar cell processes (solid arrow) in the EPL (open arrow). e Chimeras showing the distribution of syntaxin-1A (red) in the INL.Representative images of a 51 % Wt↔Wt and a 58 % Wt↔frc chimeric retina are shown. Note that syntaxin-1A positive EPL (open arrow) was seen inWt↔frc chimeras in regions enriched with Nr2e1-mutant cells (EGFP positive, green). The solid arrows show a region with predominantly wild-type cells(EGFP negative) where the EPL is absent. Also note that the IPL (brackets) from Wt↔frc retina is thicker and more disorganized compared to Wt↔Wt.n = 3 for Nr2e1+/+, n = 3 for Nr2e1frc/frc, n = 4 for Wt↔Wt, n = 4 for Wt↔frc; EPL, ectopic plexiform layer; GCL, ganglion cell layer; Hoechst, nuclearcounterstain (blue); INL, inner nuclear layer; ONL, outer nuclear layer; OPL, outer plexiform layer; scale bar = 50 μmCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 11 of 21retinas. These retinas appeared to have normal insteadof reduced numbers of Nr2e1frc/frc glycinergic amacrinesand had normal numbers of Nr2e1frc/frc total amacrinecells. Third, chimeras still had low numbers of mutantrods and bipolar cells that do not account for the ob-served rescue of retinal thickness. Fourth, we showedthat low numbers of ganglion cells were rescued in P7Wt↔frc chimeras. According to a previous study, in-creased apoptosis in the ganglion cell layer occurs inNr2e1 mutant retinas as early as P0 [21]. Thus, excessivecell-death may be the cause of ganglion cell reduction inP7 Nr2e1frc/frc retinas that should be otherwise over-represented as a consequence of premature neurogen-esis. Therefore, the rescue of mutant ganglion cellnumbers that we observed in chimeras is likely due to arestoration of ganglion cell survival. In conclusion,Nr2e1 does not have a cell-autonomous role in prevent-ing postnatal cell death as wild-type cells are able torescue cell loss.Moreover, structural and gliosis-related phenotypeswere rescued in P21 Wt↔frc chimeras. In Nr2e1frc/frcretinas, some cells in the INL protrude into the ONL,and cells in the ONL migrate into the subretinal space[22]. In Wt↔frc chimeras, we do not see overlapping ofINL and ONL cells or any other structural defect evenin a retina with 86 % Nr2e1frc/frc cells, suggesting thatthese defects are not regulated cell-autonomously byNr2e1. Interestingly, expression of GFAP in Müller glia,which is typical during gliosis but also in Nr2e1frc/frcretinas [21], was absent in Nr2e1frc/frc Müller glia ofFigure 9 P21 retinas show lamination defects and loss of amacrine cells. Transverse retinal sections from P21 Nr2e1+/+, Nr2e1frc/frc, and chimericmice were immunostained for syntaxin-1A or GlyT1. a In wild-type retinas, syntaxin-1A positive cells (green) were located in the INL with processes thatprojected into the IPL. In P21 Nr2e1-mutant retinas, syntaxin-1A positive processes projected to both the IPL and the EPL (open arrow). Note a decreasein syntaxin-1A positive cell bodies. b Magnification of the box in A showing the EPL (open arrow) of Nr2e1-mutant retinas positive for syntaxin-1A (solidarrow). c The glycinergic marker GlyT1 (red) revealed a reduction in the number of this subtype of amacrine cells in Nr2e1frc/frc retinas. d Magnificationof the box in C showing the presence of many interplexiform amacrine cells extending neurites to the OPL in Nr2e1frc/frc retinas (solid arrow) indicativeof additional lamination defects. In P21 chimeras, (e) the presence of many GlyT1 positive cells (red) in both Wt↔Wt and Wt↔frc is evident. Representativeimages of a 46 % Wt↔Wt and a 62 % Wt↔frc chimeric retina are shown. f,g Magnification of the boxes in E showing (f) not detectable interplexiform cellsin the 46 % Wt↔Wt chimera and (g) the presence of strongly labeled interplexiform cells in the 62 % Wt↔frc chimera (asterisks). h Magnification ofdotted boxes in E showing Nr2e1-mutant cells (EGFP positive, green) that are also GlyT1 positive (red) in the INL (exemplified by arrows). n= 3 for Nr2e1+/+,n= 3 for Nr2e1frc/frc, n= 3 for P21 Wt↔Wt, n= 3 for P21 Wt↔frc; EPL, ectopic plexiform layer; GCL, ganglion cell layer; Hoechst, nuclear counterstain (blue);INL, inner nuclear layer; ONL, outer nuclear layer; OPL, outer plexiform layer; scale bar in A to E = 50 μm; scale bar in F = 7 μm; scale bar in H = 9 μmCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 12 of 21chimeric retinas in spite of GFAP being a direct targetof Nr2e1 [32]. This further supports the suggestion thatdystrophy-related phenotypes are not regulated cell-autonomously by Nr2e1.Furthermore, we observed that blood vessel numbersdo not influence retinal thickness. Nr2e1frc/frc reducedblood vessel number was rescued by ≥55 % wild-type cells.The high number of cells needed to rescue the blood ves-sel numbers indicates that this defect is not corrected byrescuing the mutant cells and is likely cell-autonomous.Nr2e1 is expressed in proangiogenic astrocytes [20] andthus may act cell-autonomously in these cells to regulateretinal blood vessel development. In addition, the contri-bution of mutant cells to the astrocyte lineage could havebeen different than that to the retinal lineage, as astrocytesand retinal progenitors come from different cell popula-tions. Regardless of the mechanism, it appears that theblood vessel and retinal thickness phenotypes of Nr2e1frc/frcretinas are not influenced by each other and thus retinaldystrophy may not be influenced by abnormal blood vesseldevelopment. In addition, Nr2e1frc/frc defects such asabnormal cellular proportions and lamination were neverobserved in wild-type regions of chimeric retinas suggest-ing that vascular abnormalities are not the primary causeof retinal defects in the mutant retinas.These results taken together suggest that cell death,abnormal retinal structure, and gliosis, all related todystrophy in Nr2e1 mutant retinas, are not regulatedcell-autonomously and can be rescued by wild-type cells.However, blood vessel abnormalities may not be a cru-cial factor in influencing dystrophy in Nr2e1 mutantretinas. What are then the primary roles of Nr2e1 duringretinogenesis? We found that in addition to the previ-ously described role in cell-cycle regulation, Nr2e1 regu-lates the development of specific retinal cell types asdiscussed below.Nr2e1 regulates the development of specific retinal celltypesIn the retina, Nr2e1 is thought to regulate cell numbersthrough shortening cell-cycle length and thus delayingneurogenesis [22]. We found that at P7, the density ofsome cell-types in Nr2e1frc/frc retinas does not corres-pond to the proportions expected from premature cell-cycle exit. Importantly, the same cellular proportionswere seen in Nr2e1frc/frc regions of chimeras suggestingthey emerge intrinsically from each clone of RPCs andare not secondary to eye volume differences.We found that Nr2e1frc/frc retinas are biased to gener-ate Müller glia. The reduction in rods and bipolar cellnumbers in Nr2e1frc/frc retinascould be explained byunderproduction of late-born cell types due to prema-ture neurogenesis. Müller glia is the last retinal cell typeto be generated [1] and thus premature neurogenesisFigure 10 Nr2e1frc/frc Müller glia cells were aberrantly positioned in the INL. Transverse retinal sections from P7 Nr2e1+/+, Nr2e1frc/frc, and chimericmice were immunostained for SOX-2 and Islet-1/2. a Nr2e1+/+ SOX-2 positive Müller-glia somas (red, bracket) were localized between cholinergicamacrines in the lower INL and ON bipolars in the upper INL, the latter two cell types both labeled with Islet-1/2 (green). Müller-glia somas localizedclose to the middle of the INL and away from the OPL. b Nr2e1frc/frc Müller-glia somas (red, bracket) intermingled with ON bipolars (green) and localizedadjacent to the OPL. c In Wt↔frc chimeras, the mispositioning of the Müller-glia soma was seen only in Nr2e1-mutant cells (EGFP positive, green),which were located closer to the OPL compared to wild-type cells (EGFP negative). Representative images of a 51 % Wt↔Wt and a 58 %Wt↔frc chimeric retina are shown. d,e Magnification of the boxes in (c) showing a region from the (d) Wt↔Wt chimera depicting Müller glialocated close to the middle of the INL, and from the (e) Wt↔frc chimera depicting mutant Müller glia located adjacent to the OPL. The dottedlines represent an imaginary boundary above which most wild-type cells localize. Note that in a Wt↔frc chimera, most Nr2e1-mutant cells (EGFPpositive) localize under this line close to the OPL and most wild-type cells (EGFP negative) localize above this line and away from the OPL. n = 3 forNr2e1+/+, n = 3 for Nr2e1frc/frc, n = 4 for Wt↔Wt, n = 4 for Wt↔frc; EPL, Ectopic plexiform layer; GCL, ganglion cell layer; INL, inner nuclear layer; OPL,outer plexiform layer; Hoechst, nuclear counterstain (blue); scale bar = 50 μmCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 13 of 21would also predict reduced numbers. The fact that theirnumbers are not reduced in Nr2e1frc/frc mice suggests abias towards Müller glia differentiation in the absence ofNr2e1 regulation. In adult rat hippocampal-progenitors,Nr2e1 represses glial differentiation and activates theneuronal lineage [18]. Nr2e1 also directly binds to the pro-moter of glial-specific genes [32]. All of this taken togethersuggests a role for Nr2e1 in repressing the Müller gliallineage during development.We also found an expanded population of glycinergicamacrine cells in the INL of Nr2e1frc/frc retinas and in-creased numbers of Nr2e1 mutant amacrine cells com-pared to wild type in chimeras. Amacrine cells are mainlysubdivided in two groups: GABAergic and glycinergic [33].Amacrine cell overproduction in Nr2e1frc/frc retinas couldbe a consequence of precocious neurogenesis. However, abias towards a particular subtype was not expected, espe-cially when glycinergic amacrines are generated after theGABAergic subtype [33] making it more likely thatGABAergic amacrines would be overrepresented. Thissuggests that Nr2e1 may have a role in regulating amacrinesubtype development during mouse retinogenesis and thata switch in cell fate may have occurred in favor of glyciner-gic amacrine cells at the expense of other cell types thatare reduced in number such as bipolar cells and rods.We also observed that Nr2e1 regulates most retinalcell numbers cell-autonomously. To assess how Nr2e1regulates extrinsic and intrinsic signals that affect RPCproliferation and differentiation in vivo, we evaluated theinterplay between wild-type and Nr2e1frc/frc cells inWt↔frc chimeras. Intriguingly, the cellular compositionof the GCL, but not the INL or ONL, was rescued byextra-cellular signals from wild-type cells. The GCL iscomprised of ganglion and amacrine cells. The rescue ofganglion cell numbers in this layer is likely due to pre-vention of cell-death, as explained above. The fact thatNr2e1frc/frc amacrine cell numbers were normal insteadof overabundant in the GCL of chimeras suggests a roleof extracellular signals in regulating their migration orsurvival in this layer rather than in preventing theirFigure 11 Nr2e1frc/frc Müller glia misexpressed Brn3a cell-autonomously. Transverse retinal sections from P7 and P21 Nr2e1+/+ and Nr2e1frc/frc miceand P7 chimeras were immunostained for SOX-2, Brn3a, and Vimentin. a In P7 wild-type retinas, ganglion cells labeled with Brn3a (green) appeared as adifferent population from Müller glia labeled with SOX-2 (red, bracket). b In Nr2e1frc/frc retinas, the marker Brn3a (green) was misexpressed in Müller-gliasomas (red, bracket), and in processes that branch in the EPL and IPL, with termini in the GCL that resemble end-feet. c-f Immunostaining for the Müller-glia markers vimentin (red) and Brn3a (green) in P7 and P21 retinas. In P7 retinas, these markers did not co-express in (c) Nr2e1+/+ retinas but they did in(d) Nr2e1frc/frc retinas. This also occurred in P21 retinas where the markers did not co-express in (e) Nr2e1+/+ retinas but they did in (f) Nr2e1frc/frc retinas. d,fInset boxes show a magnified view of the vimentin and Brn3a co-localization in Nr2e1frc/frc Müller-glial processes. g In P7 Wt↔frc chimeras, misexpressionof Brn3a (red) in Müller glia was seen in Nr2e1-mutant cells (EGFP positive, green) throughout the cell soma and processes, however, this was not seen inwild-type cells (EGFP negative) where Brn3a was only expressed in ganglion cells. Representative images of a 71 % Wt↔Wt and a 58 % Wt↔frc chimericretina are shown. n= 3 for Nr2e1+/+, n= 3 for Nr2e1frc/frc, n= 4 for Wt↔Wt, n= 4 for Wt↔frc; EPL, Ectopic plexiform layer; GCL, ganglion cell layer; INL,inner nuclear layer; OPL, outer plexiform layer; Hoechst, nuclear counterstain (blue); scale bar = 50 μmCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 14 of 21generation because amacrines in the INL were still over-abundant. Therefore, we think that wild-type cells did notaffect the proportions of cells generated from RPCs. Thissuggests that RPC proliferation and differentiation doesnot depend on non-cell-autonomous roles of Nr2e1 as isthe case of adult neural stem cell proliferation [19]. Inother words, the role of Nr2e1 in differentiation and cell-cycle regulation is likely cell-autonomous in the retina.Furthermore, chimeras revealed a role of Nr2e1 inS-cone development. In Nr2e1frc/frc retinas, total conesand S-cones numbers were normal but in Wt↔frc chi-meras Nr2e1frc/frc S-cones were more abundant thanwild-type S-cones. The late stage of development of S-cones, which includes S-Opsin expression, can be regu-lated by extracellular signaling such as the one mediatedby thyroid hormone [34]. It is then possible that Nr2e1regulates extrinsic signals necessary for the developmentand/or survival of S-cones. However, we cannot rule outthat a bigger sample size would yield a significant differencein S-cones between wild-type and mutant retinas given thatthe quantified difference has a p-value of only 0.1. In agree-ment with this, a previous study reported increased S opsintranscript levels in P14 Nr2e1 mutant retinas [22]. How-ever, without information regarding protein levels and cellnumbers, we cannot know with certainty that the numbersof S-cones were higher in those mutant retinas. Finally, thefact that Nr2e3, the closest relative of Nr2e1, regulatesphotoreceptor development and is mutated in enhancedS-cone syndrome in humans [35] further supports thehypothesis that Nr2e1 has a role in S-cone development.Interestingly, in Nr2e1frc/frc retinas, the densities of hori-zontal cells, cones, and Müller glia remained comparableto wild type from P7 through P21 when retinogenesis haslong been completed. This indicates that these cells areFigure 12 Model depicting the composition and organization of Nr2e1+/+ and Nr2e1frc/frc retinal cells at P7. a Wild-type retina with 3 nuclearlayers and 2 plexiform layers. b Nr2e1frc/frc retinas contain reduced numbers of ganglion, bipolar, and rod cells, but increased numbers of glycinergicamacrine cells. Disorganization of the IPL is evident, as well as the presence of an ectopic plexiform layer (EPL) in the INL. Müller glia misexpress GFAPand Brn3a, and have somas localized closer to the OPL intermingled with bipolar cells. c A predominantly Nr2e1frc/frc clone of cells in a Wt↔frc chimerahas reduced numbers of bipolar and rods, but increased number of amacrine cells. Ganglion cell numbers are restored to wild type. There is also an in-crease in S-cone numbers compared to a wild-type clone. Disorganization of the IPL, and an EPL, are still evident. Müller glia misexpress Brn3a but notGFAP, and have mispositioned somas. Ganglion cells are ectopically positioned in the IPL and INL. EPL, ectopic plexiform layer; GCL, ganglion cell layer;INL, inner nuclear layer; IPL, inner plexiform layer; Neg., negative; ONL, outer nuclear layer; OPL, outer plexiform layer; Pos., positiveCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 15 of 21spared from the excessive cell-death seen in Nr2e1frc/frcretinas during the postnatal period. This also suggests thatthe majority of cell death observed between P7 and P21 inNr2e1frc/frc retinas may be due to the loss of amacrine cellsthat make most of the INL at P7.Nr2e1 then has an important role in the development ofcertain cell-types during retinogenesis perhaps throughthe regulation of genetic networks that are also involvedin cell-fate. During retinogenesis, transcription factorssuch as Prox1 [12], P57(Kip2) [14] and Math5 [15] arereused to perform multiple functions including cell prolif-eration and cell fate. Interestingly, Nr2e1 remains to beexpressed in Müller glia but the specific roles of Nr2e1 inthis cell type remains elusive. We further studied the ex-pression of Nr2e1 during Müller glia development and thelocalization of Müller glia within the retina of Nr2e1frc/frcand chimeric retinas as discussed below.Nr2e1 regulates Müller glia differentiation cell-autonomouslyWe observed that Nr2e1 is expressed in Müller gliathroughout postnatal development and that Nr2e1frc/frcMüller glia misexpress Brn3a and position their somaectopically in the upper INL. Brn3a is a transcriptionfactor normally expressed only in ganglion cell precur-sors and mature ganglion cells in the retina [36]. Brn3aectopic expression was likely specific in Müller glia as 1)it was present only in the ventral retina; 2) it co-localized with Müller glial markers; and 3) it was presentonly in Nr2e1frc/frc cells of chimeras. Thus, Nr2e1 maybe required to maintain a differentiated state in Müllerglia by directly or indirectly repressing Brn3a. Interest-ingly, Brn3a misexpression in Müller glia was also cyto-plasmic and not exclusively nuclear. However, this is notsurprising since various studies have reported cytoplasmiclocalization of other transcription factors including Crx[37], Hes [38], Pax-6, SOX-2 and Chx10 [39] in culturedMüller glia. This suggests that nuclear trafficking mecha-nisms of various transcription factors may work differentlyin Müller glia under atypical conditions. Importantly,Brn3a misexpression in the ventral retina suggests thatNr2e1 may have a role during dorso-ventral patterning inthe retina as it does in the brain [40].Both, the Brn3a misexpression and ectopic soma posi-tioning of Müller glia were present only in Nr2e1frc/frc cellsof Wt↔frc chimeras. The fact that Nr2e1 is expressed inMüller glia and that wild-type Müller glia are unaffected byNr2e1frc/frc cells, suggests that Nr2e1 may have a cell-autonomous role in the maturation of Müller glia by regu-lating their transcriptional profile and localization withinthe retina. Furthermore, Müller glia mislocalization andmisexpression of Brn3a are not a consequence of dys-trophy since dystrophy is rescued by wild-type cells in chi-meras but these Müller glia defects are not. Nr2e1 nullMüller glia were shown by others to have thinner pro-cesses compared to wild-type [21]. Our results suggestthat this phenotype is likely cell-autonomous and not aconsequence of other cellular defects in Nr2e1 mutant ret-inas. However, this was not determined in that study.The role of Nr2e1 in regulating the function of Müllerglia or any other cell-type was not assessed in this studyand deserves further investigation. However, defects inneuropil development were observed in Nr2e1 mutantretinas, which suggest major defects in retinal connectiv-ity and are discussed below.Nr2e1 regulates retinal lamination by preventingdisorganization of inner retinal neuritesNr2e1frc/frc retinas have a disorganized IPL, an ectopicplexiform layer (EPL) within the INL, and abundantinterplexiform amacrine cells. Together these observa-tions suggest that Nr2e1 is involved in a cellular mech-anism that organizes and constrains the neurites ofinner retinal neurons to the IPL. Interestingly, the EPLcan attract normal pre-synaptic partners, as bipolarcells arborized within it.In Wt↔frc chimeras, wild-type regions did not havean EPL, however an EPL was evident in mutant regions.This suggests that a long-range extracellular molecule isnot involved in inducing the formation of the EPL. Onthe other hand, wild-type columns had a disorganizedIPL, suggesting that Nr2e1 mutant cells can misdirectthe neurites of wild-type cells. Wt↔frc chimeras alsodemonstrate that abnormal ganglion cell numbers per seare not the cause of IPL disorganization since they arerestored in these chimeras.The regulation of laminar targeting specificity in the ret-ina depends on repulsive and attractive interactions medi-ated mostly by multiple short-range molecules [41–44]. Adisrupted patterning of the IPL and disorganized amacrinebranching is seen in Pten−/− and Dscam−/− mice [45, 46].Pten−/− mouse retinas have an expanded IPL with scat-tered cell bodies of abnormal morphology and clusters ofdendrites in the IPL [45, 47]. Since Pten is a direct targetof Nr2e1, its abnormal levels in Nr2e1frc/frc retina couldcontribute to aberrant neurite organization. Pten remainsexpressed in the INL, IPL and GCL at P7 [45]. WhetherNr2e1 is expressed in immature amacrine cells duringdevelopment deserves further investigation.Interestingly, a clear single ectopic layer in the INLhasbeen shown only in mice lacking the transmembrane re-pellents Sema5A and Sema5B [43] and the atypical cad-herin Fat3 [41]. Fat3−/− amacrine cells develop a seconddendritic tree that projects away from the IPL and formsan additional synaptic layer in the INL [41]. Thus, ourwork suggests the interesting hypothesis that the Sema5or Fat3 pathways may be affected in Nr2e1frc/frc retinas.Corso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 16 of 21The mechanisms by which signaling molecules orches-trate the development of a normal neuropil in the nervoussystem are still vastly unknown. Understanding how thelack of a transcription factor, such as Nr2e1, allows the for-mation of an ectopic neuropil in the retina would help toshed light into the cellular and molecular mechanismsleading to this process.In summary, we found that Nr2e1 has numerous rolesbeyond preventing premature cell cycle exit in RPCs dur-ing retinal development including neurite organizationand terminal cell differentiation. Our results show thatNr2e1 regulates the organization of inner retina neuritesto the IPL and the development of neuronal subtypes,such as S-cones and glycineric amacrines as well as Müllerglia. In addition, Nr2e1 regulates Müller glia differenti-ation cell-autonomously. Moreover, we have shown thatprevention of retinal dystrophy is not regulated cell-autonomously by Nr2e1. A model summarizing the de-fects in cell numbers and lamination of the P7 Nr2e1frc/frcand chimeric retina is depicted in Fig. 12.Materials and methodsMouse strains, husbandry, and breedingNr2e1frc mice harbor a spontaneous deletion of the entirecoding region of Nr2e1 with no transcripts detected in thebrain. These mice are highly aggressive and display hypo-plastic cerebrum, olfactory lobes and retina [17]. Mice forstudy were generated by crossing Nr2e1frc/+ heterozygotefemales, which were congenic (>N30) on the C57BL/6 J(B6) (JAX stock # 000664) background, to Nr2e1frc/+ het-erozygote males, which were congenic (>N30) on the129S1/SvImJ (129) (JAX stock # 002448) background, thusgenerating Nr2e1+/+ control and Nr2e1frc/frc mutant litter-mates on the hybrid B6129F1 background [48]. Malesharboring the NR2E1-lacZ reporter BAC (bEMS86) wereused to study NR2E1 expression in the postnatal retina;strain B6.129P2(Cg)-Hprttm73(Ple142-lacZ)Ems, abbreviatedhere as NR2E1-lacZ (MMRRC stock # 032962-JAX) [29].Males harboring a random insertion of multiple-linkedcopies of the enhanced green fluorescent protein (EGFP)under an ubiquitous promoter were used for embryonicstem cell (ESC) generation; strain B6-Tg(CAG-EGF-P)1Osb/J, abbreviated here as B6-EGFP (N > 10; JAXStock # 003291) [49]. Mice containing the lacZ transgeneat the ROSA26 locus were used as a source of host blasto-cysts; strain 129S-Gt(ROSA)26Sor/J (JAX Stock # 002292)[50], backcrossed to B6 (N2-3) and abbreviated here as B6-R26lacZ. Eye studies were carried out on mice of either sex.Mice were kept in a pathogen-free animal facility atthe Centre for Molecular Medicine and Therapeutics,University of British Columbia (Vancouver, BC, Canada)on a 6 am to 8 pm light cycle with, 20 ± 2 °C, 50 % ± 5 %relative humidity, and food and water ad libitum. Allprocedures involving animals were in accordance withthe Canadian Council on Animal Care (CCAC) andUBC Animal Care Committee (ACC) (Protocol numbersA11-0370 and A11-0081).Generation of embryonic stem cells (ESCs)To obtain ESCs that contained both the Nr2e1frc alleleand the EGFP transgene as a marker, a two generationcross was performed. First, B6-Nr2e1frc/+ mice werecrossed to B6-EGFP/0 mice to obtain B6-Nr2e1frc/+EGFP/0 females. Then, these females were crossed to129-Nr2e1frc/+ males to obtain B6129F1-Nr2e1frc/frcEGFP/0 blastocysts. ESC lines were derived as previ-ously described [51]. Briefly, blastocysts at 3.5 days postfertilization (dpf ) were cultured in KSOM + AA media(Caisson Laboratories, North Logan, UT) under oil at37 °C for 3–5 h. Then, each blastocyst was transferredto a single (1 x) 96-well containing mitomycin-C-inactivated mouse embryonic fibroblasts (MEFs) andcultured in KSR-ESC media (Knockout™ D-MEM with2 mM l-glutamine, 0.1 mM MEM nonessential aminoacid solution, 16 % Knockout™ Serum Replacement (allfrom Life Technologies Inc., Burlington, ON), and 1000U/ml ESGRO® (LIF) (Millipore, Temecula, CA) . Oncethe blastocyst hatched and had a well-defined inner cellmass, it was trypsinized and transferred to 1 x 24-wellcontaining inactivated MEFs. Cells were replated 1:1 ifnecessary. When cells reached confluence, they weresplit into 3 x 24-wells in 100 % ESC media. All 3 wellswere combined at confluence and frozen in 3 separatevials until needed.Generation of chimerasChimeras were derived by microinjection of ESC into hostblastocysts as previously described [51]. Either B6129F1-Nr2e1+/+ EGFP/0 or B6129F1-Nr2e1frc/frc EGFP/0 ESCswere microinjected into B6-R26lacZ/+ host blastocysts. Thegenerated chimeras were thus comprised of blastocyst-derived cells that were wild-type for Nr2e1 and harboredthe R26lacZ transgene and ESC-derived cells that were ei-ther wild-type or mutant for Nr2e1, and harbored theEGFP transgene. A chimera is defined by the genotype ofthe blastocyst-derived and ESC-derived cells (blastocys-t↔ESC): (Nr2e1+/+, R26lacZ/+) ↔ (Nr2e1+/+, EGFP/0) or(Nr2e1+/+, R26lacZ/+) ↔ (Nr2e1frc/frc, EGFP/0); and abbre-viated as Wt↔+/+ or Wt↔frc/frc, respectively. TwoNr2e1+/+ (mEMS4919 and mEMS4926) and two Nr2e1frc/frc(mEMS4914 and mEMS4922) ESC lines were used to gen-erate chimeras. After injection, blastocysts were implantedinto the uterine horns of B6 2.5 day pseudopregnant fe-males. Four Wt↔+/+ and four Wt↔frc/frc P7 chimericeyes were included in the study. Nine Wt↔+/+ and tenWt↔frc/frc P21 adult chimeric eyes were also included.Corso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 17 of 21Assessment of chimerismChimerism was initially assessed by coat color. Then, inslides prepared of the retina, chimerism was assessed bymeasuring the area of EGFP epifluorescence signal in theinner nuclear layer (INL) plus outer nuclear layer (ONL)using the software ImageJ [52]. The percentage of this fluor-escence was assessed over the total INL plus ONL area.HistologyEyes were fixed by intracardial perfusions performed onavertin-anesthetized mice with 4 % paraformaldehyde(PFA) in phosphate buffered saline (PBS). Eyes were thenpost-fixed in 4 % PFA for 30 min prior to cryoprotection in25 % sucrose-PBS overnight. Subsequently, eyes wereembedded in optimal cutting temperature (OCT) medium,cryosectioned at 12 μm and mounted on SuperFrost Plusslides (Thermo Fisher Scientific, Waltham, MA, USA). Forβ-galactosidase (β-gal) immunohistochemistry, staining wasperformed using X-gal substrate (5-bromo-4-chloro-3-indolyl-beta-D-galacto-pyranoside) for 18 h at 37 °C. Toevaluate retinal thickness, retinal sections were subjected tohematoxylin and eosin staining. Briefly, tissue was incu-bated in hematoxylin for 5 min, washed in tap water andincubated in 1 % lithium carbonate solution for 30 sec.After washing in tap water again, the tissue was incubatedin acid alcohol (1 %) for 5 s followed by another tap waterwash and incubation in eosin Y solution for 5 min. After afinal tap water wash, tissue was dehydrated in a gradient ofethanol, and xylene before mounting for microscopy. Forimmunofluorescence, at least three Nr2e1+/+ and threeNr2e1frc/frc retinal sections per epitope were incubated inblocking solution (5 % bovine serum albumin, 0.3 % TritonX-100 in PBS) for 1 h. Subsequently, sections were incu-bated in primary antibody in blocking solution at 4 °C over-night. After three washes of 10 min in PBS, sections wereincubated in secondary antibody in blocking solution withthe DNA dye Hoechst-33342 for 1 h. Sections were washed3 times in PBS and mounted in ProLong® Gold AntifadeReagent (Life Technologies Inc., Carlsbad, CA, USA). Atleast three transverse retinal sections were immunostainedfor each antibody using at least three eyes per genotype forimaging analysis.AntibodiesTo label ganglion cells, we stained the retinas with anti-bodies to Brn3a, which is expressed in approximately 80 %of these cells [31]. For amacrines we used antibodies tothe pan-amacrine markers Pax-6, and syntaxin-1A, whichlabel amacrine cells and horizontal cells in the inner nu-clear layer (INL) and amacrine and ganglion cells in theganglion cell layer (GCL) [53, 54]. To visualize horizontalcells we used antibodies to calbindin, which is expressedin horizontal cells and a subpopulation of amacrine cells[55]. We used for: cones, cone-arrestin (arrestin-C) andS-Opsin antibodies that at P7 label the cell body, processesand outer segment of cones, which are located throughoutthe outer nuclear layer [56]; rods, rhodopsin antibody; bi-polar cells, CHX10 antibody [53]; and Müller glia, Sox9antibody [57]. The following primary antibodies wereused: mouse anti-Brn3a 14A6 from Santa Cruz Biotech-nology, Inc. (Dallas, TX, USA); mouse anti-Pax-6, anti-ISL1/2; rabbit anti-Pax-6 from Coavance (Princeton, NJ,USA); mouse anti-calbindin and anti-syntaxin fromSigma-Aldrich (St. Louis, MO, USA); mouse anti-Rhodopsin (ID4) from Dr. R.S. Molday, University ofBritish Columbia (Vancouver, BC, Canada); sheep anti-Chx10 from Exalpha Biologicals, Inc. (Shirley, MA, USA);rabbit anti-cone-arrestin, rabbit anti-Sox9, rabbit anti-S-opsin, rabbit anti-mGluR1, goat anti-calretinin, chickenanti-vimentin, and guinea pig anti-GABA from Millipore(Billerica, MA, USA); mouse anti-PKCalpha, chicken anti-β-galactosidase, and rabbit anti-SOX-2 and anti-GlyT1from Abcam (Cambridge, England); and mouse anti-GFAPfrom New England Biolabs (Ipswich, MA, USA).Imaging and cell countingFor dual visualization of X-gal blue precipitate and EGFPfluorescence in chimeras, images were taken with theOlympus BX61 motorized microscope using both DPController and In Vivo software (Olympus Corporationof the Americas, Center Valley, PA, USA). All remainingfluorescent images were taken with a Leica TCS SP5 IIconfocal microscope (Leica Microsystems, Wetzlar,Germany). To assess the numbers of each cell type inthe retina, images were taken at 200 times magnificationthroughout a 12 μM retinal section and tiled. At leastfive equidistant transverse retinal sections throughoutthe entire eye were imaged: two images were close tothe edge; two close to the middle half; and one centralthat included the optic nerve. One eye, from three differ-ent mice of each genotype, was included. Labeled cellswere manually counted using the Cell Counter tool ofImageJ. Images were taken with a scan size of at least1024 X 1024 pixels using 200 times magnification andvisualized at 18 X 18 cms for accurate counting. Theretinal marker channels and blue channel (Hoeschststained-nuclei) images were superimposed followed byalternative switching between the channels to make sureindividual cells were counted and to assess co-localization.Using the Cell Counter Tool a dot was placed on each cellcounted to avoid double counting or miscounting of cells.Cell numbers were normalized to retinal length (μm). Thislength was assessed by drawing a line from edge to edgeof retinal tissue along the nuclear layer where the cell-typeto be counted was located. To count cells in chimeric ret-inas, marker-positive and EGFP-negative singled-labeledcells or marker-positive and EGFP-positive double-labeledcells were counted throughout a central retinal section.Corso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 18 of 21Cell numbers were normalized to retinal area. The EGFP-positive area was assessed by measuring the area of EGFPepifluorescence signal in the INL plus ONL using ImageJ.The EGFP-negative area was assessed by subtracting theEGFP positive area from the total INL plus ONL area. Celldensity was expressed as the number of single-labeled cellsover the EGFP-negative area or double-labeled cells overthe EGFP-positive area. One eye, from three different chi-meras for each ESC genotype, was included; totaling 6eyes. Only chimeras with a contribution of >46 % mutantcells were counted. Quantification of retinal structural de-fects where INL and ONL cells overlap was performed in3 eyes per genotype and 5 retinal sections as explainedabove for retina cell number quantification. A structuraldefect was counted when INL and ONL cells where insuch proximity to each other that the OPL was no longerevident.FunduscopyTo assess the number of retinal blood vessels, funduscopywas performed as previously described [29, 58]. Briefly,eyes were dilated with 1 % atropine in PBS and photo-graphed after 30 min. Animals were manually restrainedwithout sedation.Image processingImages were processed using ImageJ and Adobe Photo-shop (Adobe Systems Incorporated, San Jose, CA, USA).Brightness, contrast, and scaling adjustments were per-formed as necessary.Statistical analysesStatistical analyses were performed using Microsoft Excel(Microsoft Corporation, Redmond, Washington, USA)and XLSTAT (Addinsoft, New York, NY, USA). Student’st-test was used to compare total cell numbers using rawdata. In some cases, data was graphed as percentage ofwild-type cells. Results were considered significance with aP-value ≤ 0.05. The standard error of the mean (SEM) wasindicated with error bars. Z-scores were calculated as(retinal thickness or blood vessel number of Wt↔frcchimeras) minus ((mean of Wt↔Wt chimeras) divided by(standard deviation of Wt↔Wt chimeras)).Additional filesAdditional file 1: Figure S1. Labeling of blastocyst-derived and ESC-derived cells was mutually exclusive in chimeras. (A) Embryonic stem cell(ESC) lines carrying wild-type and Nr2e1-mutant alleles plus the ubiquitouslyexpressed EGFP transgene, were generated from E3.5 blastocysts derivedfrom a two generation cross; first between B6-Nr2e1frc/+ and B6-Tg(CAG-EGF-P)1Osb/J/0 (abbreviated here as B6-EGFP/0) mice, and then B6-Nr2e1frc/+EGFP/0 and 129-Nr2e1frc/+ mice. (B) ESCs were injected into B6-Gt(ROSA)26-Sor/J/+ (abbreviated here as B6-R26lacZ/+) E3.5 blastocysts containing theubiquitously expressed lacZ gene, which encodes the enzyme β-galactosidase(β-gal). The resulting chimeras, denoted as blastocyst↔ESC [Wt↔Wt orWt↔frc], contained wild-type (Wt) blastocyst-derived cells expressing β-galand ESC-derived cells expressing EGFP. The latter cells were either wild-type ormutant (frc) for Nr2e1. (C) To demonstrate that β-gal and EGFP werecell-specific markers, retinal sections from B6-R26lacZ/+ and B6-EGFP/0control mice at P7 and P21 were incubated with X-gal. The distributionof X-gal product (blue) and EGFP epifluorescence signal (green) isshown. The EGFP signal was not affected by the X-gal reaction. (D) Todemonstrate that β-gal and EGFP were cell-specific markers in chimeras,Wt↔Wt and Wt↔frc chimeric retinal sections from P7 and P21 mice weretreated as in C. The labeling of blastocyst-derived (β-gal positive, blue) andESC-derived (EGFP positive, green) cells was mutually exclusive. n = 4 for P7Wt↔Wt, n = 9 for P21 Wt↔Wt, n = 4 for P7 Wt↔Wt, n = 10 for P21 Wt↔frc.Additional file 2: Figure S2. Nr2e1frc/frc P7 retinas had reducednumbers of rod and bipolar cells that were not rescued in Wt↔frc chimeras.Transverse retinal sections from P7 Nr2e1+/+, Nr2e1frc/frc, and chimeric micewere immunostained for rhodopsin (rods) and CHX10 (bipolars). (A) rods(green) .(B) bipolar cells (red). (C) Each retinal cell type was countedthroughout the retina as described in methods.. Reduced numbers of bothrods and bipolar cells were observed in Nr2e1-mutant retinas compared towild type (34 % and 27 % decrease, respectively). (D, E) The density of (D)rod and (E) bipolar cells (red) that were EGFP positive (green) appearedsimilar to the density of the corresponding EGFP negative cells in Wt↔Wtchimeras but looked lower in Wt↔frc chimeras. Representative images of a46 % and a 71 % Wt↔Wt, and 62 % and 58 % Wt↔frc chimeric retina areshown for rods and bipolar cells, respectively. The arrows in E show a regioncomprised mostly of wild-type cells (EGFP negative) with a higher density ofbipolar cells. (F, G) Quantification of the density of rod and bipolar cells thatwere derived from host blastocyst or ESCs in chimeras was assessed bycounting single-labeled cells (marker positive and EGFP negative) or double-labeled cells (marker positive and EGFP positive) and dividing each by theEGFP negative or EGFP positive retinal area (ONL + INL), respectively. Thedensity of Nr2e1frc/frc EGFP positive (F) rods and (G) bipolar cells was signifi-cantly lower (50 % and 53 % decrease, respectively) compared to the densityof the corresponding EGFP negative cells in Wt↔frc chimeras. n = 3 forNr2e1+/+, n = 3 for Nr2e1frc/frc, n = 3 for Wt↔Wt, n = 3 for Wt↔frc; *, P≤ 0.05;ns, not significant; error bars represent SEM. GCL, ganglion cell layer;Hoechst, nuclear counterstain (blue); INL, inner nuclear layer; neg., negative;ONL, outer nuclear layer; pos., positive; scale bar = 50 μm.Additional file 3: Figure S3 Increased numbers of Nr2e1frc/frc GlyT1positive cells were not rescued in P7 Wt↔frc chimeras. Transverse retinalsections from P7 chimeric mice were immunostained for the amacrinemarker GlyT1. In Wt↔Wt chimeras, the density of GlyT1 amacrine cells thatwere EGFP positive (green) appeared similar to the density of EGFP negativeamacrine cells. In Wt↔frc chimeras, the density of GlyT1 amacrine cells thatwere EGFP positive (green) appeared higher than the density of EGFPnegative amacrine cells. Representative images of a 51 % Wt↔Wt and a62 % Wt↔frc chimeric retina are shown. The arrow shows a region withhigh numbers of mutant cells (EGFP positive) where GlyT1 positive amacrinecells are overrepresented. GCL, ganglion cell layer; Hoechst, nuclearcounterstain (blue); INL, inner nuclear layer; scale bar = 50 μm.Additional file 4: Figure S4. The ectopic inner plexiform layer ofNr2e1frc/frc retinas harbors neurites from various amacrine populations.Transverse retinal sections from P7 Nr2e1+/+ and Nr2e1frc/frc mice wereimmunostained for calbindin and mGluR1. (A) Calbindin-positive amacrinecells (green) extended processes in the IPL of Nr2e1+/+ and Nr2e1frc/frc retinas.In Nr2e1frc/frc retinas, some calbindin-positive cells also extended processesinto the EPL (open arrow). (B) Magnification of the box in A showing thoseoccasional ectopic calbindin-positive cell bodies (asterisks) that extendedprocesses into the EPL (open arrow). (C) In wild-type retinas, mGluR1 (red)was expressed in some INL amacrine cell bodies immunostained for Pax-6(green). In Nr2e1-mutant retinas, mGluR1 was also expressed in the ectopicEPL (open arrow). (D) Magnification of the box in E showing the expressionof mGluR1 (solid arrow) in the Nr2e1-mutant EPL (open arrow) andoccasional mGluR1 positive cell bodies that co-label with Pax-6 (asterisk)and extend processes into the EPL. The disorganized IPL of Nr2e1frc/frc retinasalso harbored occasional amacrine cells bodies located between IPLsublamina (arrowhead). n = 3 for Nr2e1+/+, n = 3 for Nr2e1frc/frc; EPL, ectopicplexiform layer; GCL, ganglion cell layer; Hoechst, nuclear counterstain (blue);Corso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 19 of 21INL, inner nuclear layer; ONL, outer nuclear layer; OPL, outer plexiform layer;scale bar in A = 50 μm; scale bar in B = 22 μm.Additional file 5: Figure S5. Nr2e1frc/frc amacrine cell numbers arenormal in P21 Wt↔frc chimeras. Quantification of the density of syntaxin-1Apositive cells that were derived from host blastocyst or ESCs in P21 chimeraswas assessed by counting single-labeled cells (Syntaxin-1A positive andEGFP negative) or double-labeled cells (Syntaxin-1A positive and EGFPpositive) and dividing them by the EGFP negative or EGFP positive retinalarea (ONL + INL), respectively. In Wt↔ frc chimeras, the density of EGFPpositive and EGFP negative syntaxin-1A cells was similar. n = 3 for Wt↔Wt,n = 3 for Wt↔frc; *, P≤ 0.05; ns, not significant; error bars represent SEM;neg., negative; pos., positive.Additional file 6: Figure S6. Nr2e1 was expressed in Müller gliathroughout postnatal development. Transverse retinal sections from P3, P7,P14, and P21 retinas from Nr2e1+/+ mice containing a human NR2E1-drivenlacZ reporter (NR2E1-lacZ) and Nr2e1frc/frc mice were immunostained forSOX-2 (red). This marker recognizes retinal precursor cells at earlier time-points, and an amacrine subpopulation and Müller glia at later time-points.(A) Sections from NR2E1-lacZ retinas were immunostained for β-galactosidase(β-gal). Co-localization of SOX-2 with β-gal in progenitors and Müller glia wasobserved at all time-points. Merge 1 shows SOX-2 (red) and β-gal (green)staining. Merge 2 shows Hoechst nuclear staining (blue) and SOX-2 staining(red). (B) In Nr2e1frc/frc retinas stained as in Merge 2, Müller glial densityappeared comparable to wild type. n= 3 for Nr2e1+/+, n= 3 for Nr2e1frc/frc;GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer;scale bar = 50 μm.Additional file 7: Figure S7. Nr2e1frc/frc Müller glia express Brn3athroughout the cell body. Magnification of a region of figure 11D(rectangle) showing a Müller cell displaying a typical irregular soma labeledwith Brn3a and, as expected, not labeled with Vimentin. Distal processestowards the OPL, a stalk towards the GCL and projections that normallysurround ganglion cell soma (end-feet) are labeled with both Brn3a andVimentin. Müller glia lateral projections in the EPL are also evident withBrn3a but are not labeled with Vimentin. EPL = ectopic plexiform layer; GCL,ganglion cell layer; INL, inner nuclear layer; IPL = inner plexiform layer; ONL,outer nuclear layer; OPL = outer nuclear layer; scale bar = 50 μm.Competing interestsThe authors declare that they have no competing interests.Authors’ contributionsEMS conceived the study. EMS and XCD contributed to experimental design.XCD performed experiments and data analysis. XCD drafted the manuscriptwith help from EMS. Both authors read and approved the final manuscript.AcknowledgmentsThis work was supported by a Sharon Stewart Aniridia Research Award[grant number 20R64586] and Sharon Stewart Aniridia Trust [grant number20R67715] to EMS, and a CFRI Partnership Graduate Studentship award toXCD. We are very grateful to: the Microinjectionists Sonia F. Black, Tom W.Johnson and Tess C. Lengyell for their role in making the chimeric mice aspart of the Mouse Animal Production Service at the Centre for MolecularMedicine and Therapeutics, Vancouver, BC, Canada; Dr. Robert S. Molday,University of British Columbia, Vancouver, BC, Canada, for providing us withmouse anti-Rhodopsin (ID4) antibody; and Dr. Dan Goldowitz, University ofBritish Columbia, Vancouver, BC, Canada, for his input during manuscriptpreparation.Author details1Centre for Molecular Medicine and Therapeutics at the Child and FamilyResearch Institute, University of British Columbia, 950 W 28 Ave, VancouverV5Z 4H4BC, Canada. 2Genetics Graduate Program, University of BritishColumbia, Vancouver V6T 1Z2BC, Canada. 3Department of Medical Genetics,University of British Columbia, Vancouver V6T 1Z3BC, Canada. 4Departmentof Psychiatry, University of British Columbia, Vancouver V6T 2A1BC, Canada.Received: 1 March 2015 Accepted: 23 May 2015References1. Livesey FJ, Cepko CL. Vertebrate neural cell-fate determination: lessons fromthe retina. Nat Rev Neurosci. 2001;2(2):109–18.2. Cayouette M, Barres BA, Raff M. Importance of intrinsic mechanisms in cellfate decisions in the developing rat retina. Neuron. 2003;40(5):897–904.3. He J, Zhang G, Almeida AD, Cayouette M, Simons BD, Harris WA. Howvariable clones build an invariant retina. Neuron. 2012;75(5):786–98.4. Watanabe T, Raff MC. Rod photoreceptor development in vitro: intrinsicproperties of proliferating neuroepithelial cells change as developmentproceeds in the rat retina. Neuron. 1990;4(3):461–7.5. Hashimoto T, Zhang XM, Chen BY, Yang XJ. VEGF activates divergentintracellular signaling components to regulate retinal progenitor cellproliferation and neuronal differentiation. Development. 2006;133(11):2201–10.6. Yu C, Mazerolle CJ, Thurig S, Wang Y, Pacal M, Bremner R, et al. Direct andindirect effects of hedgehog pathway activation in the mammalian retina.Mol Cell Neurosci. 2006;32(3):274–82.7. Ma L, Cantrup R, Varrault A, Colak D, Klenin N, Gotz M, et al. Zac1 functionsthrough TGFbetaII to negatively regulate cell number in the developingretina. Neural Dev. 2007;2:11.8. Alexiades MR, Cepko C. Quantitative analysis of proliferation and cell cyclelength during development of the rat retina. Dev Dyn. 1996;205(3):293–307.9. Dyer MA, Cepko CL. p27Kip1 and p57Kip2 regulate proliferation in distinctretinal progenitor cell populations. J Neurosci. 2001;21(12):4259–71.10. Ohnuma S, Hopper S, Wang KC, Philpott A, Harris WA. Co-ordinating retinalhistogenesis: early cell cycle exit enhances early cell fate determination inthe Xenopus retina. Development. 2002;129(10):2435–46.11. Casarosa S, Amato MA, Andreazzoli M, Gestri G, Barsacchi G, Cremisi F. Xrx1controls proliferation and multipotency of retinal progenitors. Mol CellNeurosci. 2003;22(1):25–36.12. Dyer MA, Livesey FJ, Cepko CL, Oliver G. Prox1 function controls progenitorcell proliferation and horizontal cell genesis in the mammalian retina. NatGenet. 2003;34(1):53–8.13. Trimarchi JM, Stadler MB, Cepko CL. Individual retinal progenitor cells displayextensive heterogeneity of gene expression. PLoS One. 2008;3(2):e1588.14. Dyer MA, Cepko CL. p57(Kip2) regulates progenitor cell proliferation andamacrine interneuron development in the mouse retina. Development.2000;127(16):3593–605.15. Le TT, Wroblewski E, Patel S, Riesenberg AN, Brown NL. Math5 is requiredfor both early retinal neuron differentiation and cell cycle progression. DevBiol. 2006;295(2):764–78.16. Li W, Sun G, Yang S, Qu Q, Nakashima K, Shi Y. Nuclear receptor TLXregulates cell cycle progression in neural stem cells of the developing brain.Mol Endocrinol. 2008;22(1):56–64.17. Young KA, Berry ML, Mahaffey CL, Saionz JR, Hawes NL, Chang B, et al.Fierce: a new mouse deletion of Nr2e1; violent behaviour and ocularabnormalities are background-dependent. Behav Brain Res. 2002;132(2):145–58.18. Elmi M, Matsumoto Y, Zeng ZJ, Lakshminarasimhan P, Yang W, Uemura A,et al. TLX activates MASH1 for induction of neuronal lineage commitmentof adult hippocampal neuroprogenitors. Mol Cell Neurosci. 2010;45(2):121–31.19. Qu Q, Sun G, Li W, Yang S, Ye P, Zhao C, et al. Orphan nuclear receptor TLXactivates Wnt/beta-catenin signalling to stimulate neural stem cellproliferation and self-renewal. Nat Cell Biol. 2010;12(1):31–40. sup pp 31–39.20. Uemura A, Kusuhara S, Wiegand SJ, Yu RT, Nishikawa S. Tlx acts as aproangiogenic switch by regulating extracellular assembly of fibronectinmatrices in retinal astrocytes. J Clin Invest. 2006;116(2):369–77.21. Miyawaki T, Uemura A, Dezawa M, Yu RT, Ide C, Nishikawa S, et al. Tlx, anorphan nuclear receptor, regulates cell numbers and astrocyte developmentin the developing retina. J Neurosci. 2004;24(37):8124–34.22. Zhang CL, Zou Y, Yu RT, Gage FH, Evans RM. Nuclear receptor TLX preventsretinal dystrophy and recruits the corepressor atrophin1. Genes Dev.2006;20(10):1308–20.23. Collinson JM, Hill RE, West JD. Analysis of mouse eye development withchimeras and mosaics. Int J Dev Biol. 2004;48(8–9):793–804.24. Ekstrom P, Sanyal S, Narfstrom K, Chader GJ, van Veen T. Accumulation ofglial fibrillary acidic protein in Muller radial glia during retinal degeneration.Invest Ophthalmol Vis Sci. 1988;29(9):1363–71.25. Rapaport DH, Wong LL, Wood ED, Yasumura D, LaVail MM. Timing andtopography of cell genesis in the rat retina. J Comp Neurol. 2004;474(2):304–24.26. Cepko CL, Austin CP, Yang X, Alexiades M, Ezzeddine D. Cell fatedetermination in the vertebrate retina. Proc Natl Acad Sci U S A.1996;93(2):589–95.Corso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 20 of 2127. Franze K, Grosche J, Skatchkov SN, Schinkinger S, Foja C, Schild D, et al.Muller cells are living optical fibers in the vertebrate retina. Proc Natl AcadSci U S A. 2007;104(20):8287–92.28. Dubin MW. The inner plexiform layer of the vertebrate retina: a quantitativeand comparative electron microscopic analysis. J Comp Neurol.1970;140(4):479–505.29. Schmouth JF, Banks KG, Mathelier A, Gregory-Evans CY, Castellarin M, HoltRA, et al. Retina restored and brain abnormalities ameliorated by single-copyknock-in of human NR2E1 in null mice. Mol Cell Biol. 2012;32(7):1296–311.30. Surzenko N, Crowl T, Bachleda A, Langer L, Pevny L. SOX2 maintains thequiescent progenitor cell state of postnatal retinal Muller glia. Development.2013;140(7):1445–56.31. Badea TC, Cahill H, Ecker J, Hattar S, Nathans J. Distinct roles of transcriptionfactors brn3a and brn3b in controlling the development, morphology, andfunction of retinal ganglion cells. Neuron. 2009;61(6):852–64.32. Shi Y, Chichung Lie D, Taupin P, Nakashima K, Ray J, Yu RT, et al. Expressionand function of orphan nuclear receptor TLX in adult neural stem cells.Nature. 2004;427(6969):78–83.33. Voinescu PE, Kay JN, Sanes JR. Birthdays of retinal amacrine cell subtypesare systematically related to their molecular identity and soma position. JComp Neurol. 2009;517(5):737–50.34. Roberts MR, Srinivas M, Forrest D, Morreale de Escobar G, Reh TA. Makingthe gradient: thyroid hormone regulates cone opsin expression in thedeveloping mouse retina. Proc Natl Acad Sci U S A. 2006;103(16):6218–23.35. Haider NB, Jacobson SG, Cideciyan AV, Swiderski R, Streb LM, Searby C, et al.Mutation of a nuclear receptor gene, NR2E3, causes enhanced S conesyndrome, a disorder of retinal cell fate. Nat Genet. 2000;24(2):127–31.36. Badea TC, Williams J, Smallwood P, Shi M, Motajo O, Nathans J.Combinatorial expression of Brn3 transcription factors in somatosensoryneurons: genetic and morphologic analysis. J Neurosci. 2012;32(3):995–1007.37. Phillips MJ, Otteson DC. Differential expression of neuronal genes in Mullerglia in two- and three-dimensional cultures. Invest Ophthalmol Vis Sci.2011;52(3):1439–49.38. Nickerson PE, Da Silva N, Myers T, Stevens K, Clarke DB. Neural progenitorpotential in cultured Muller glia: effects of passaging and exogenousgrowth factor exposure. Brain Res. 2008;1230:1–12.39. Lawrence JM, Singhal S, Bhatia B, Keegan DJ, Reh TA, Luthert PJ, et al. MIO-M1cells and similar muller glial cell lines derived from adult human retina exhibitneural stem cell characteristics. Stem Cells. 2007;25(8):2033–43.40. Stenman J, Yu RT, Evans RM, Campbell K. Tlx and Pax6 co-operate geneticallyto establish the pallio-subpallial boundary in the embryonic mousetelencephalon. Development. 2003;130(6):1113–22.41. Deans MR, Krol A, Abraira VE, Copley CO, Tucker AF, Goodrich LV. Control ofneuronal morphology by the atypical cadherin Fat3. Neuron.2011;71(5):820–32.42. Masai I, Lele Z, Yamaguchi M, Komori A, Nakata A, Nishiwaki Y, et al. N-cadherinmediates retinal lamination, maintenance of forebrain compartments andpatterning of retinal neurites. Development. 2003;130(11):2479–94.43. Matsuoka RL, Chivatakarn O, Badea TC, Samuels IS, Cahill H, Katayama K, etal. Class 5 transmembrane semaphorins control selective Mammalian retinallamination and function. Neuron. 2011;71(3):460–73.44. Yamagata M, Sanes JR. Expanding the Ig superfamily code for laminar specificityin retina: expression and role of contactins. J Neurosci. 2012;32(41):14402–14.45. Cantrup R, Dixit R, Palmesino E, Bonfield S, Shaker T, Tachibana N, et al. Cell-typespecific roles for PTEN in establishing a functional retinal architecture. PLoS One.2012;7(3), e32795.46. Fuerst PG, Bruce F, Rounds RP, Erskine L, Burgess RW. Cell autonomy ofDSCAM function in retinal development. Dev Biol. 2012;361(2):326–37.47. Sakagami K, Chen B, Nusinowitz S, Wu H, Yang XJ. PTEN regulates retinalinterneuron morphogenesis and synaptic layer formation. Mol Cell Neurosci.2012;49(2):171–83.48. Silva AJ, Simpson EM, Wolfer DP, et al. Mutant mice and neuroscience:recommendations concerning genetic background. Banbury Conference ongenetic background in mice. Neuron. 1997;19(4):755–9.49. Okabe M, Ikawa M, Kominami K, Nakanishi T, Nishimune Y. Green mice’ as asource of ubiquitous green cells. FEBS Lett. 1997;407(3):313–9.50. Zambrowicz BP, Imamoto A, Fiering S, Herzenberg LA, Kerr WG, SorianoP. Disruption of overlapping transcripts in the ROSA beta geo 26 genetrap strain leads to widespread expression of beta-galactosidase inmouse embryos and hematopoietic cells. Proc Natl Acad Sci U S A.1997;94(8):3789–94.51. Yang GS, Banks KG, Bonaguro RJ, Wilson G, Dreolini L, de Leeuw CN, et al.Next generation tools for high-throughput promoter and expressionanalysis employing single-copy knock-ins at the Hprt1 locus. Genomics.2009;93(3):196–204.52. Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years ofimage analysis. Nat Methods. 2012;9(7):671–5.53. de Melo J, Qiu X, Du G, Cristante L, Eisenstat DD. Dlx1, Dlx2, Pax6, Brn3b,and Chx10 homeobox gene expression defines the retinal ganglion andinner nuclear layers of the developing and adult mouse retina. J CompNeurol. 2003;461(2):187–204.54. Inoue A, Akagawa K. Neuron specific expression of a membrane protein,HPC-1: tissue distribution, and cellular and subcellular localization ofimmunoreactivity and mRNA. Brain Res Mol Brain Res. 1993;19(1–2):121–8.55. Li S, Mo Z, Yang X, Price SM, Shen MM, Xiang M. Foxn4 controls the genesisof amacrine and horizontal cells by retinal progenitors. Neuron.2004;43(6):795–807.56. Glaschke A, Glosmann M, Peichl L. Developmental changes of cone opsinexpression but not retinal morphology in the hypothyroid Pax8 knockoutmouse. Invest Ophthalmol Vis Sci. 2010;51(3):1719–27.57. Poche RA, Furuta Y, Chaboissier MC, Schedl A, Behringer RR. Sox9 isexpressed in mouse multipotent retinal progenitor cells and functions inMuller glial cell development. J Comp Neurol. 2008;510(3):237–50.58. Abrahams BS, Kwok MC, Trinh E, Budaghzadeh S, Hossain SM, Simpson EM.Pathological aggression in “fierce” mice corrected by human nuclearreceptor 2E1. J Neurosci. 2005;25(27):6263–70.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at www.biomedcentral.com/submitCorso-Díaz and Simpson Molecular Brain  (2015) 8:37 Page 21 of 21


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items