Reprogramming chick RPE progeny cells to differentiate towards retinal neurons by ash1

Chick embryos

Fertilized, pathogen-free White Leghorn chicken eggs were purchased from Spafas (Preston, CT) and incubated in a Petersime egg incubator (Gettysburg, OH). The care and use of animals adhered to the procedures and policies comparable to those published by the USA Public Health Service (Public Health Service Police on Humane Care and Use of Animals) and set by the Institutional Animal Use and Care Committee at the University of Alabama at Birmingham.

Replication Competent Avian Splice retroviruses

Replication Competent Avian Splice (RCAS) retrovirus [38] expressing ash1 (RCAS-ash1) was generated as described [37]. During cloning process, a recombinant (rb) form of Ash1, Ash1ΔC_rb, was generated. It lacked the C-terminal 33 amino acids of the wild-type Ash1 and contained (at the C-terminus) 19 extra residues derived from the shuttle vector Cla12Nco. RCAS-ash1ΔC_rb was included in the study to determine whether the C-terminus of Ash1 was dispensable for its function. All recombinant RCAS constructs were sequenced for verification and for ruling out possible changes including frameshift or nonsense mutations. Sequencing was performed at the DNA Sequencing Service of the Genomics Core Facility at the University of Alabama at Birmingham. RCAS-green fluorescent protein (GFP) was produced as previously described [14].

RPE cell culture

E6 chick eyes were enucleated, and tissues outside the PRE were removed. A circular incision in the ciliary region was made to get rid of the anterior portion of the eye, producing eye cup consisting of the RPE, the retina, and the vitreous. The RPE was then peeled off from the neural retina. RPE tissues were dissected from day 6 chick embryos (E6) as previously described [14]. Cells in RPE tissues were dissociated by trypsin/EDTA and gentle trituration. Fully dissociated RPE cells were resuspended in Gibco medium 199 (Invitrogen, Carlsbad, CA) plus 10% fetal bovine serum. The cells were then seeded at low density (approximately 1×10⁵ cells) in a 35 mm dish and cultured at 37 °C under 5% CO₂ with Knockout Dulbecco's Modified Eagle Medium (D-MEM) supplemented with 20% serum replacement (Invitrogen). After 3–4 days in culture, when 50% confluency was reached, 10–15 μl of retrovirus stock (2–5×10⁸ pfu/ml) was added to each 35 mm dish. After an additional 8–12 days in culture, cells were directly subjected to Ca²⁺ imaging or fixed with ice-cold 4% paraformaldehyde in phosphate-buffered saline (PBS) for analyses with immunocytochemistry and in situ hybridization.

To facilitate visualizing the morphologies of reprogrammed cells, we transfected RPE cell cultures 6 days after the administration of RCAS-ash1ΔC_rb. Transfection was done with recombinant DNA of adeno-associated virus expressing GFP (AAV-GFP) driven by a chimerical promoter of Cytomegalovirus (CMV)-actin-β globin. Cellular morphologies were imaged with a blue filter under a 20X or 40X objective attached to a Nikon TE300 (Melville, NY) inverted microscope 2–4 days after AAV-GFP DNA transfection.

The calcium indicator fluo-4 AM [39] (Molecular Probes, Eugene, OR) was also used to visualize the morphologies of reprogrammed cells. Cells in RPE cell cultures infected with RCAS-ash1 or RCAS were incubated with fluo-4 AM as described in detail late under “Calcium imaging” [17]. Cellular morphologies were imaged under a 40X objective attached to a Nikon TE300 inverted microscope with a blue filter.

Retinal cell culture

Retinas were dissected from day 16 chick embryos. Retinal cells were dissociated with 0.05% trypsin and 0.53 mM EDTA (Invitrogen) and mechanical trituration. Dissociated retinal cells were cultured in polyornithine-treated 35 mm dishes with medium 199 (Invitrogen) plus 10% fetal bovine serum at 37 °C under 5% CO₂ for 5 days. Culture medium was changed every other day, and the cells were subjected to Ca²⁺ imaging in the same way as those in RPE cell cultures.

Immunocytochemistry

The following monoclonal antibodies were obtained from the Developmental Studies Hybridoma Bank (Iowa University, Iowa City, IA): anti-AP2 (3B5; 1:50; developed by Dr. T. Williams, Yale University, New Haven, CT); anti-islet-1 (1:100; developed by Dr. T. Jessell, Columbia University, New York, NY), anti-LIM (4F2; 1:50; developed by Dr. T. Jessell, Columbia University, New York, NY), anti-vimentin (H5; 1:500; developed by Dr. J. Sanes, Harvard University, Cambridge, MA), and anti-visinin (7G4; 1:500; developed by Dr. C. Cepko, Harvard University, Cambridge, MA). Antibodies obtained from commercial sources included antiviral protein p27 (1:500; Spafas, Preston, CT), polyclonal antibody against calretinin (1:500; Chemicon, Temecula, CA), polyclonal antibody against red opsin (1:200; Chemicon), polyclonal antibody against AP2α (1:200; Santa Cruz Biotechnology, Santa Cruz, CA), and monoclonal antibody against MAP2 (1:200; Sigma, St. Louis, MO). Monoclonal antibody RA4 (1:1,000 dilution) was a gift from Dr. Steven McLoon (University of Minnesota, Minneapolis, MN). Standard procedures for immunocytochemistry were followed.

The number of immunopositive cells was counted under a Nikon TE300 microscope (Melville, NY). In each experiment, 3 dishes were assigned for each marker. From each dish, the number of immunopositive cells was scored from 10 view areas under a 20X objective, and the total number of immunopositive cells per 35 mm dish was calculated by multiplying the average per view area by 200 (the total number of view areas per dish).

In situ hybridization

Briefly, cells in the culture were fixed with ice-cold 4% paraformaldehyde for 30 min before being permealized with 0.25% triton X-100 for 30 min. Hybridization was carried with 40–80 ng of RNA probes overnight at 48–55 °C. In situ hybridization was used to detect the expression of interphotoreceptor retinal binding protein (IRBP), GAD67 (GenBank 4103977), ath3, and chx10, as previously described [15,40].

Calcium imaging

Calcium imaging was used to study the physiologic responses of reprogrammed cells to neurotransmitters, as described by Lamba et al. [11] with minor modifications. Cells in the 35 mm culture dish (either of RPE cell culture or of retinal cell culture) were washed and incubated with medium 199 for 1 h at 37 °C. Then cells were washed with a magnesium-free bath solution that contained 5.3 mM KCl, 115.0 mM NaCl, 3.1 mM CaCl₂, 5.6 mM glucose, and 3.0 mM HEPES, pH 7.2. The bath solution was adapted from that given in Sosa and Gleason [41]. Next, cells were incubated with fluo-4 AM as described [17] in 1 ml of bath solution for 15 min at room temperature. After a brief rinsing and then incubation with the bath solution for 30 min at room temperature, the culture dish was placed under an Olympus BX60 microscope with constant perfusion with the bath solution at a rate of 1 ml per minute at room temperature. Fluorescence images were captured sequentially without time interval with 400 ms-exposure for each image. Captured images were analyzed with WCFI ImageJ software. Briefly, a region of interest covering the soma of a cell was selected in stacked images. A region of the same shape and area was also selected from the background for background subtraction. Changes in fluorescence intensity were expressed as percentage of ΔF/Fo, where F is the averaged integrated optic density (IOD) of an image, Fo is the average of IOD the images before glutamate application, and ΔF=F-Fo. Approximately 3–5 μl of 1 mM glutamate in the bath solution, or 1 mM N-methyl-D-aspartate (NMDA) and 1 mM glycine, was puffed onto the area of interest within 500 ms through a micropipette driven by a PV800 Pneumatic Picopump (World Precision Instruments, Sarasota, FL). The recording chamber was under constant perfusion (around 1 ml/minute) at room temperature. ΔF/Fo was plotted against time using Origin Pro7 (OriginLab, Northampton, MA).

Altered appearance of RPE cell culture infected with RCAS-ash1

Dissociated RPE cells (seeded at a density of approximately 1×10⁵ cells in a 35 mm dish) lose their dark pigment during the initial days in culture, but they become repigmented as the culture approaches confluency [42]. Infection with RCAS-GFP has little or no effect on the repigmentation of the culture (Figure 1A,B; [42]). Infection with RCAS-ash1, however, prevented the RPE culture from repigmentation, except at places where patches of pigmented cells were present (Figure 1A).

In addition, cultures infected with RCAS-ash1 also contained cell clusters noticeable under a microscope with phase contrast or Hoffman modulation optics (Figure 1C). The clusters of cells were irregularly shaped and contained cells with a round cell body and often with processes, reminiscent of neural clusters. No such cell clusters were present in the control culture, in which hexagonal cells were organized in a cobblestone fashion (Figure 1B).

Induction of neural genes by ash1

To determine whether ash1 reprogrammed RPE progeny cells to differentiate toward retinal neurons, RPE cell cultures infected with RCAS-ash1 were examined for the expression of genes/markers that are normally expressed in various types of retinal neurons. These markers included calretinin, expressed in horizontal and ganglion cells and some amacrine cells [43,44]; general neural marker MAP2 present in all retinal neurons except bipolar cells [45]; ganglion cell markers RA4 [46] and Islet-1 [47], which also labels bipolar cells at late developmental stages, amacrine markers AP2α [48] and GAD67 [49]; bipolar cell markers ath3 and chx10 [50]; horizontal cell marker LIM [51]; photoreceptor markers red opsin, IRBP [52,53], and visinin [54], whose immunoreactivity is also detected weakly in developing amacrine cells in the chick retina [55,56].

Overall, calretinin was the most expressed among the markers tested. While infection by RCAS-GFP did not induce calretinin expression (Figure 2A), numerous calretinin⁺ cells were detected in cultures infected with RCAS-ash1 (Figure 2B,C). Because calretinin⁺ cells were frequently localized in clusters (Figure 2B,C, arrows), calculating their exact number was impractical; nonetheless, we estimate that as many as 20% of the cells present in the culture were calretinin⁺ (>30,000 in a 35 mm dish; Table 1). Long, thin processes typical of neurons emanated from and webbed among the calretinin⁺ cells. Neuron-like processes were also observed on MAP2⁺ cells from RPE cultures infected with RCAS-ash1ΔC_rb (Figure 2D,E).

Immunocytochemistry and in situ hybridization of RPE cell cultures infected with RCAS-ash1 showed the following: a substantial numbers of cells immunopositive for RA4 (>4,000 in a 35 mm dish; Table 1) or visinin (>11,000; Figure 2G); a small number of cells expressing IRBP (>1,000); a few expressing GAD67 (>100) or red opsin (>90); and none positive for AP2α, ath3, or chx10 (Table 1). None of the genes/markers were expressed in the control infected with RCAS-GFP (Figure 2A) [14,16], with the exception of RA4, which typically is detected in hundreds of cells in the control [14]. We also examined the expression of neural markers in RPE cell cultures infected with RCAS-ash1ΔC_rb, which expressed a truncated Ash1 lacking the C-terminal sequence and with a short sequence acquired from the cloning vector (see Methods). Chick embryos infected with RCAS-ash1ΔC_rb exhibited all, but more pronounced, phenotypes observed with RCAS-ash1 [37]. In RPE cell culture, infection with RCAS-ash1ΔC_rb produced similar results as infection with RCAS-ash1, providing further support to the C-terminal sequence being nonessential to Ash1’s function [37].

The prevalence of calretinin⁺ cells along with ash1 promoting amacrine overexpression in the retina [37] raised a question of whether some of the visinin⁺ cells were amacrine-like, as differentiating amacrine cells in the retina transiently express of photoreceptor markers [55,56]. Double-labeling showed the presence of visinin⁺/calretinin⁺ cells (Figure 2F-H, arrows). Double-labeled cells accounted for an estimated 75% of visinin⁺ cells (in 5 view areas under a 40X objective) and 5%–10% of calretinin⁺ cells (in the same areas).

Morphologies of reprogrammed cells

The calcium indicator fluo-4 AM selectively identifies reprogrammed, neuron-like cells in an RPE cell culture transduced with proneural gene neuroD [17]. Thus, we used fluo-4 AM coupled with fluorescence photomicrography to directly visualize the cellular morphologies of reprogrammed cells in RPE cultures infected with RCAS-ash1. Cells in the control cultures infected with RCAS were “invisible” under our experimental conditions (Figure 3A,B), as their intracellular Ca²⁺ concentrations were below the level of detection with a 2 s exposure in image-capture. Under the same experimental conditions, reprogrammed cultures (i.e., infected with RCAS-ash1) contained a large number of fluoresced cells (~15%), with the rest of the cells (presumably not being reprogrammed) remaining “invisible.” Like individual, calretinin⁺ cells with adornment by elaborate processes (Figure 3C), most of the fluoresced cells emanated multiple processes from their soma (Figure 3D), a morphological characteristic of horizontal, amacrine, or ganglion cells of the retina.

The morphology of reprogrammed cells was also examined with AAV-GFP DNA transduction. As shown in Figure 3E, reprogrammed cells in an RPE culture infected with RCAS-ash1ΔC_rb displayed cellular processes remarkably complex and closely resembling those possessed by retinal horizontal, amacrine, or ganglion cells.

Because Műller glia exhibit spindle-like processes that may appear similar to neural process and its Ca²⁺ concentration may be sufficiently high to be visible with fluo-4 AM, the RPE cell cultures were examined for the presence of Müller glia-like cells by immunostaining for vimentin, which is detected in Müller glial processes [57]. Because cultured RPE cells also express vimentin and other commonly used Műller glial markers [58,59], vimentin⁺ cells with spindle-like processes were the observation target. In an RPE culture infected with RCAS-GFP, antivimentin staining was weak and observed in essentially all cells with morphologies of culture RPE cells (Figure 4A,B). Similar to the control, cultures infected with RCAS-ash1 contained vimentin+ cells that exhibited morphologies typical of culture RPE cells, but lacked vimentin⁺ cells that displayed spindle-like processes (Figure 4C,D). No particular antivimentin immunostaining was observed at places occupied by cells with compacted cell bodies (Figure 4C,D, arrows). Double-labeling for vimentin and calretinin showed the absence of Műller glia-like, vimentin⁺ cells (Figure 4E,F,H) in a culture containing reprogrammed cells that expressed calretinin (Figure 4G,I). The absence of Müller glia-like, vimentin⁺ cells is consistent with published reports that ash1 favors neurogenesis against gliogenesis [29-32].

Response to glutamate and NMDA

The molecular and morphological resemblance of the reprogrammed cells to retinal neurons prompted a further analysis of whether the reprogrammed cells developed physiologic properties typical of retinal neurons. The cells were assayed for functional glutamate receptors. Glutamate is a common neurotransmitter in the retina, and its receptors are either ionotropic or metabotropic [60-62]. Ionotropic receptors include NMDA and non-NMDA receptors, varying in their selective agonists. NMDA receptors are permeable to Ca²⁺, and their activation results in an increase in intracellular Ca²⁺ concentration [62]. By comparison, most of the metabotropic receptors cause a decrease in intracellular Ca²⁺ concentration when they are activated [62].

The development of functional glutamate receptors was assayed by examining Ca²⁺ concentration before and after the administration of glutamate or NDMA along with glycine, which acts as a co-agonist of the NMDA receptor and potentiates the NMDA-triggered response [11,63]. Ca²⁺ concentration was monitored using fluo-4 AM coupled with fluorescence microscopy. We found that reprogrammed cells, particularly those with neural morphologies, responded to the administration of NMDA and glycine by increasing their Ca²⁺ concentrations (Figure 5A). Increases in fluorescence intensity were detected a few seconds after the application of the neurotransmitters, indicative of a Ca²⁺ influx through calcium channels upon receptor activation. Fluorescence intensity peaked usually within15 s, quickly declined, and eventually reached pre-activation level (Figure 5). The response was similar to that observed with retinal cells in culture. When the changes in fluorescence intensity (as ΔF/Fo) were plotted over time, the responses to the neurotransmitters by reprogrammed cells (Figure 5B,C) resembling the responses by retinal neurons (Figure 5D) was evident. Notably, the degree of response by different cells varied (Figure 5B,C), likely due to varied extents of neural differentiation and varied physiologic conditions. Additionally, approximately 20% of the 86 cells subjected to the analysis showed responses. Cells in the control culture (infected with RCAS) showed no response (Figure 5E).

Reprogrammed cells also responded to 1 mM exogenous glutamate (Figure 6A,B). The response curves of the reprogrammed cells were similar to those of retinal cells. Like responses to NDMA+glycine, the extent of responses to glutamate varied among the cells. Varied responses were also observed with retinal neurons (Figure 6C,D). No similar responses were detected with cells in the control culture infected with RCAS (Figure 6E,F).