Abstract
Using the mouse ES cell line with green fluorescent protein knocked-in at the Rx locus (Rx-KI ES cell), we previously showed that photoreceptors can be efficiently obtained in defined culture conditions by enriching Rx-positive retinal progenitor cells. We aimed to explore a protocol applicable for non-Rx-labeled stem cell lines for subsequent enrichment of retinal photoreceptor precursors for transplantation. The Rx-KI ES cell line was differentiated according to the serum-free suspension conditions with serum-free suspension/Dkk1/LeftyA/serum/activin method (SFEB/DLFA) described previously. Enrichment efficacy by negative selection was compared among 20 different lectins and the lectin combination that effectively enriched the Rx-positive cells by selecting the lectin low-binding population was determined. Subsequent differentiation efficiency to photoreceptor precursors and the contamination of Nanog or Oct3/4+ cells in the culture were evaluated between the cell cultures using negative selection with lectins and Rx positive selection. The effect of cytarabine (Ara-C) for minimizing the contamination of undifferentiated cells after the selection was also studied. The combination of the lectins, wheat germ agglutinin (WGA), and Erythrina crista-galli agglutinin (ECA) enabled us to enrich the Rx-positive population by approximately twice the original Rx percentage. The selection also minimized the percentage of Oct3/4+ cells. The lectin-selected cells produced a comparable percentage of Crx/rhodopsin-positive colonies with Rx-positive selection and were differentiated into photoreceptors. The Ara-C treatment on differentiating days 24–26 decreased Nanog and Oct3/4 expression in subsequent cultures. Enrichment of Rx-positive cells using WGA and ECA was comparable to Rx-positive selection, and the method could be applied to achieve efficient photoreceptor differentiation from other ES or iPS cell lines in which the Rx gene is not marked.
Introduction
The primary cause of some retinal degenerative diseases such as retinitis pigmentosa is apoptosis of photoreceptors. In these diseases, the photoreceptors, which are the first neurons in the visual transduction system, selectively die in the early stage, followed by the death of the second and other neurons. One therapeutic strategy for treatment of such diseases would be the reconstruction of the photoreceptor layer when the rest of the neural network is still viable. Recently, retina or retinal cell transplantation has been suggested to be a promising approach; photoreceptor precursors at a specific developmental stage can be efficiently incorporated into the adult retina by subretinal transplantation and these cells can acquire mature photoreceptor properties (3, 11, 21). Also, even with advanced retinal degeneration, Suzuki et al. suggested a possible synapse formation between host and graft cells with the help of chondoroitinase ABC (17). Results of a clinical trial of transplantation of a fetal retinal sheet suggested the possibility of the use of transplantation as the therapeutic strategy for AMD or retinitis pigmentosa (15).
With this progress, the preparation of a substantial number of good quality photoreceptors has become an imperative objective for further improvement in retinal transplantation therapy. However, in some countries such as Japan, law prohibits the use of fetal tissue. Pluripotent stem cells (iPS) can be now induced from human somatic cells (1, 18, 19); therefore, the use of embryonic stem (ES) or iPS cell lines provides an excellent opportunity for universal transplantation therapy.
Rx is a paired-like homeobox gene that is required for vertebrate eye formation (22). We previously reported the efficient generation of photoreceptor precursor cells from Rx-positive retinal progenitor cells derived from ES cell lines of mouse, monkey, and human, as well as mouse iPS (8, 14). These cells express retinal-specific cell markers, and mouse ES-derived cells were incorporated into retinal tissue in coculture, suggesting that differentiated ES or iPS cells would be good graft cell candidates. In order to utilize these cells for clinical application, the graft cell preparation should be sufficiently enriched by photoreceptors and be devoid of unwanted cells. In our mouse ES cell protocol, treatment with the notch inhibitor or gamma-secretase inhibitor N-[N-(3,5-difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester (DAPT) markedly enhanced the differentiation efficacy of photoreceptors among the Rx-positive population, but not in cells of an unselected mass culture (8). Therefore, enrichment of the Rx-positive cell population is a key step in the process of efficient photoreceptor generation. However, cell surface markers for the Rx-positive population have not yet been identified.
Lectins are known to interact with glycogens on the cell surface, and the interaction of each lectin is specific. Stem cell surfaces are glycogen rich and stem cell populations have been characterized using interacting lectin profiles (13, 20). The well-known stem cell markers such as stage-specific embryonic antigen (SSEA)-3 and −4 are in fact globo-series cell surface glycoproteins that were first used to delineate embryological changes in the developing mouse embryo (9, 16). Moreover, the neuronal cell adhesion molecule (NCAM) carrying polysialic acid (PSA-NCAM) has important roles in development and neural plasticity and is sometimes used as a neural progenitor marker (4, 7). These results suggested that Rx-positive retinal progenitors might interact with certain lectins. In the present study, we tested some of the lectins to determine whether they could identify the Rx-positive cell population using the murin ES cell line in which the GFP is knocked-in at the Rx locus.
In considering the clinical application of transplantation therapy, the evaluation of the graft cells should contain the quality control and the safety control. The quality control involves the enrichment of needed cells whereas the safety control involves the elimination of possibly harmful cells. Because the integrated graft photoreceptors upon transplantation were considered to be dominantly postmitotic (11), we also examined if the treatment of a DNA synthesis inhibitor could eliminate the unwanted proliferating cells without affecting the photoreceptors in graft preparation culture. For this purpose, cytarabine (Ara-C), which is often used for neural cell enrichment in the primary cultures, was used to test its effect on our differentiating culture. Thus, in addition to the retinal progenitor enrichment, we also studied if Ara-C could reduce the risk of contamination of the proliferating cells in the cell preparation for the purpose of transplantation.
Materials and Methods
ES Cell Culture for Photoreceptor Differentiation
The mouse reporter ES cell line in which green fluorescent protein (GFP) was knocked-in at the Rx locus of EB5 cells (Rx-KI ES cell) was used, and the cells were differentiated as described previously (8). Briefly, ES cells were cultured under the serum-free suspension/Dkk1/LeftyA/FCS/activin (SFEB/DLFA) condition for 9 days and sorted on differentiating day 10 (d.d. 10). The sorted cells were plated as a pellet of 5 × 104 cells/well on an eight-well slide chamber or a 6-cm dish, which were coated with poly-D-lysine/laminin/fibronectin and 10% FCS. On the next day, the FCS was removed by replacing it with serum-free medium and the gamma-secretase inhibitor DAPT (10 μM) was added thereafter (d.d. 11). From d.d. 17, the medium was changed to retinal culture medium as described previously (8). Briefly, the medium consisted of 66% E-MEM-HEPES (Sigma-Aldrich, St; Louis, MO), 33% HBSS (GIBCO/Invitrogen, Carlsbad, CA), 1% FCS supplemented with N2 supplement (GIBCO), 5.75 mg/ml glucose, 200 μM L-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. aFGF (R&D Systems, Minneapolis, MN) and bFGF (R&D Systems) were added during d.d. 17–24, and taurine (Sigma), Shh (R&D Systems), and retinoic acid (Sigma) were added during d.d. 17–28.
Lectin Binding Analysis and Sorting
Three lectin kits containing a total of 20 biotinylated lectins were used (BK1000, BK2000, BK3000; Vector Laboratories Burlingame, CA). The abbreviations of the lectins used in this study are listed in Table 1. The above-mentioned Rx-KI ES cell line was differentiated as described. On d.d. 9, the cells were dissociated using 0.05% trypsin-EDTA and 2 × 105 cells were incubated with 1 μg of lectins in 100 μl of PBS with 3% BSA for 30 min at room temperature. The cells were washed with PBS/3% BSA twice followed by incubation with 1:200 streptoavidin phycoerythrin PE (BD Pharmingen, San Diego, CA) in PBS/3% BSA. The cells were analyzed for a GFP-positive Rx population versus a PE-labeled lectin staining pattern using a FACS Aria flow cytometer (Beckton Dickinson, Franklin Lakes, NJ). For cell sorting, Erythrina crista-galli agglutinin (ECA)-biotin and wheat germ agglutinin (WGA)-biotin were each added at 10 μg/2 × 106 cells/ml of PBS with 3% BSA.
Abbreviation of the Lectins Used
Immunohistochemistry
Cells were fixed with 4% paraformaldehyde and immunolabeled as described previously (8). The primary antibodies and their working dilutions were mouse anti-2-bromo-5′-deoxyuridine (BrdU) (1:100, Roche, Mannheim, Germany), rat anti-Crx (1:200, made in our laboratory ref), mouse anti-Oct3/4 (1:200, BD Pharmingen), mouse anti-rhodopsin (RET-P1, 1:2000, Sigma). The secondary antibodies used were anti-mouse IgG, anti-rabbit IgG, and anti-rat IgG antibodies conjugated with Cy3 or FITC (1:300, Jackson ImmunoResearch Laboratories West Grove, PA).
Reverse-Transcription Polymerase Chain Reaction (RT-PCR)
The undifferentiated ES cells were sorted as lectin-positive or -negative cells. Immediately after the sorting procedure, the RNAs were obtained on d.d. 10, 26, and 33 using Trizol (Invitrogen) according to the manufacturer's protocol. Reverse transcription was performed with the first-strand cDNA synthesis kit (Amersham Biosciences Piscataway, NJ). The primers used are summarized in Table 2.
Primer Sequences
The forward (upper line) and reverse (lower line) primers are listed for each RNA.
Cell Proliferation Assay
One μg/ml of BrdU was added for 24 h between d.d. 14–15, 17–18, and 18–19, and the cells were fixed on d.d. 20, or BrdU was added between d.d 19–20 and fixed on d.d. 21. The cells were immunostained for crx and BrdU.
Statistics
For the statistical analysis, the t-test with Bonferroni correction was applied to multiple comparison. Unpaired t-test was applied to the comparison between with or without DAPT treatment in each group.
Results
Lectin Staining Pattern
All the lectins were tested for positivity against the GFP-positive, Rx-positive populations. No lectin specifically labeled an Rx-positive population, and Rx was dominant in lectin low-binding populations (PElo) with most of the lectins (Fig. 1A). The percentage of lectin low-binding populations (low-PE populations; PElo) and Rx populations among the low PE-populations (Rx/PElo) with each lectin is shown in Figure 1B.
(A) Examples of flow cytometric analysis of GFP (y-axis) versus PE-lectin-binding cells (x-axis) using Rx-KI mES cells. Most of the Rx-positive cells expressing GFP show low binding to either ECA or WGA. The dot profiles were divided into four quadrants and lectin low-binding cells (PElo) and Rx-positive cells were defined as cells in the indicated column as shown in the figure. (B) Lectin low-binding population (PElo) and the subsequent percentage of Rx-positive cells among PElo cells are shown with each tested lectin.
Next, we determined whether a combination of certain lectins produced better results than a single lectin for enriching the Rx populations. We selected 1) the lectins that provided less than 50% of the total cells as negative (PElo < 50%) and 2) the lectins that increased the Rx-positive population to more than 20% among the lectin-negative population (Rx/PElo > 20%). The average percentage of Rx-positive cells before selection was approximately 10%. The selected lectins were RCA (Ricinus communis), LEA (Lycopersicon esculentum), ECA, and WGA. Using RCA decreased the efficiency of cell adhesion to the plates after sorting, and using LEA resulted in a higher percentage of Oct3+ cells after negative sorting (data not shown). We therefore performed the analysis using WGA, ECA, or a combination of the two. A typical cell sorting profile is shown in Fig. 2. All cells with a lectin low-binding population were sorted using a histogram profile as shown in Figure 2.
A representative example of cell sorting based on the flow cytometric analysis using the lectin combination of ECA and WGA. Left: 15.9% of the total cells were Rx positive. Middle: PE fluorescence pattern of lectin-bound cells is shown by histogram. In the selection, lectin low-binding (PElo) cells were determined by the histogram as shown in the figure. Here, 23.5% of the total cells were regarded as PElo cells. Right: Among the PElo cells judged by the histogram, Rx-positive cells contributed 30.7% of the selected population.
Enrichment of Rx-Positive Cells and Remaining Oct3+ Cells
The percentages of Rx-positive cells among the lectin low-binding cell population for three to five independent experiments are summarized in Fig. 3A. The combination of WGA and ECA significantly increased the Rx-positive population, but using LEA on WGA and ECA did not enhance the number of Rx-positive cells. The final Rx percentage after lectin selection (WGA + ECA) depended on the initial Rx-positive percentage before sorting, and the final Rx population was consistently two- to three-fold of the initial Rx-positive population. Next, the sorted cells were plated on a slide chamber and fixed the next day (d.d. 11) for immunostaining with the anti-Oct3 antibody. The lectin selection significantly removed the Oct+ population, although the effect was not complete (Fig. 3B). Considering the enrichment efficiency and the OCT removal effect together, the protocol using the combination of WGA and ECA was used for further analysis.
(A) Final Rx-positive selection percentages using the indicated lectin. Rx is the positive population among the total population. *p < 0.001. (B) Sorted cells were replated and fixed the next day as described in Materials and Methods. The cells were immunolabeled with anti-Oct3/4 antibody. Total: unsorted cells; Rx: sorted cells as Rx-positive cells by GFP; Rx–: Rx-negative cells sorted by GFP. *p < 0.001 (C) Lineage marker gene and stem cell expression were tested in undifferentiated ES cells, and lectin-binding (lectin+) and lectin low-binding cells (lectin–) were collected immediately after sorting and RNAs were extracted.
Characterization of Lectin Low-Binding Cells
Marker gene expression was tested in the lectin low-binding population versus the rest of the lectin-binding population using RT-PCR (Fig. 3C). Rx was enriched in the lectin low-binding population, and the RNA expression of the endoderm marker HNF and mesoderm marker Brachiury was higher in the lectin-binding population.
Effect of DAPT on Cell Differentiation After Lectin Selection in Comparison with Rx-Sorted Cells
After flow cytometric selection of lectin low-binding cells, lectin-binding cells, and Rx-positive cells, they were cultured in photoreceptor-differentiating medium and conditions with or without DAPT. On d.d. 28, they were fixed and immunostained for the early photoreceptor markers Crx and rhodopsin. The percentage of Crx/rhodopsin-positive colonies among the total colonies in each well were counted (n = 4) (Fig. 4). The Crx-positive colonies were 100% rhodopsin-positive colonies and no Crx-negative colonies had rhodopsin-positive cells. A comparable percentage of Crx/rhodopsin colonies were induced in the cultures of lectin low-binding cells with that in the Rx-positive selection, whereas there were few Crx-positive colonies in the lectin-binding cell population regardless of DAPT treatment. Although the starting cell number was the same in each well, the total number of colonies in each well was larger in the lectin-positive and Rx-negative population without DAPT treatment, suggesting that these fractions contained more proliferating cells partially responsive to DAPT. Sheets of retinal pigment cells, which also originated from the Rx-positive retinal progenitor cells, appeared in all the wells of the lectin low-binding and Rx-positive cell populations with or without DAPT treatment, but only one small colony was detected in one well in the lectin-binding cell population.
(A) Lectin-sorted cells and Rx-positive cells (GFP sorted) were further cultured as described in Materials and Methods and immunolabeled with Crx and rhodopsin on d.d. 28. Percentages of Crx/rhodopsin-positive colonies are shown (n = 4). *p < 0.05 (B) Lectin-sorted cells and Rx-sorted cells were harvested on d.d. 26 and tested for the expression of retinal marker genes. The results from two independent experiments are shown for d.d. 26. The left lane is the result of lectin-sorted cells immediately after the sorting procedure on d.d. 10, shown for comparison. (C) The stem cell markers Nanog and Oct3/4 were expressed in the lectin and Rx-sorted cell cultures and compared with those of d.d. 26. The left lane is the result of lectin-sorted cells immediately after the sorting procedure on d.d. 10, shown for comparison.
Differentiation and Contamination of Undifferentiated Cells Among Differentiated Cells From Lectin-Sorted Cells
The cells were sorted on d.d. 10, further differentiated in culture, and harvested on d.d. 26 for RNA extraction. The expression of the photoreceptor specific genes Crx, Nrl, and Nanog were tested using RT-PCR (Fig. 4B). For the lectin-sorted cells, the results from two independent experiments are shown. Lectin low-binding cells expressed the photoreceptor marker genes Crx and Nrl on d.d. 26, although the expression level of Nrl seemed lower compared to that in the Rx-sorted cells. We then examined Nanog expression to ascertain whether these cells still contained undifferentiated cells. The left lane shows the result from the cells immediately after the cell selection for a comparison. Although Nanog expression was very low immediately after lectin or Rx selection, a detectable amount of Nanog expression was observed in both the Rx-sorted and lectin-sorted cells on d.d. 26.
Use of Cytarabine (Ara-C) on Culture Cells
Lectin low-binding cells were expected to contain various types of nonneuronal proliferating cells including glia. We previously observed a decrease in overall proliferation of the culture cells as a mass after DAPT treatment, and because photoreceptor cells also exit the cell cycle around the time of Crx or Nrl expression, we considered that the desired photoreceptor-fated cells had also exited the cell cycle soon after DAPT treatment. Thus, we expected to enrich photoreceptor cells by eliminating other proliferating cells, including Nanog-positive cells, by adding the selective DNA synthesis inhibitor Ara-C following DAPT treatment. In order to decide the appropriate timing of Ara-C application, we first analyzed the time course for proliferation of the “late” Crx-positive cells by BrdU incorporation and subsequent Crx immunostaining. The overall BrdU-positive population decreased as a mass with time after DAPT treatment (Fig. 5A). Then, the proliferating state of “late” Crx-positive cells was analyzed by applying BrdU at different time points and fixing cells on d.d. 19 or 20, followed by BrdU and Crx immunolabeling. The proliferating percentage of “late” Crx-positive cells tended to decrease for the first several days after DAPT treatment, but it remained almost constant thereafter, and the percentage varied from colony to colony (Fig. 5B, C). Although some Crx-positive colonies had completely exited the cell cycle by d.d. 17, showing Crx positivity with no BrdU incorporation on d.d. 20, other colonies still contained many proliferating cells among Crx-positive cells. This indicated that the timing of the cell cycle exit of some of the photoreceptor precursors was even later than that of the surrounding cells, as judged by the proliferation status of cells as a mass and the “late” Crx-positive cells. A wide range of percentages of proliferating cells among “late” Crx-positive cells at each time point indicated that 1) some of the cells continued to divide until they became positive for Crx after DAPT treatment, or (2) some of the Crx-positive cells continued to proliferate as photoreceptor precursors. Because the application of Ara-C soon after DAPT treatment resulted in only a few Crx-positive cells and a small number of overall cells (data not shown), we decided to apply Ara-C later than d.d. 20 to obtain a substantial number of Crx-positive cells.
(A) Percentage of BrdU-positive cells among the total cells at each d.d. (n = 5). (B) Percentage of BrdU-positive cells among Crx-positive cells (double positive/Crx positive). BrdU were added on indicated d.d., the cells were fixed on d.d. 19 or 20, and immunolabeled with BrdU and Crx (n = 8–10). (C) Immunolabeled cells at different time points of BrdU treatment. On d.d. 14–15, most of the Crx-positive cells (red) were positive for BrdU (arrows). On d.d. 17–18, some colonies totally lacked BrdU-positive cells indicating that Crx-positive cells on d.d. 20 were no longer dividing between d.d. 17 and 18 (arrowheads), whereas in other colonies, many of the Crx-positive cells on d.d. 20 were still dividing on d.d. 17–18. On d.d. 18–19, some colonies consisted of Crx-positive cells that were no longer dividing and cells that were still dividing. Red: Crx, green: BrdU. Scale bars: 20 μm.
Treatment with Ara-C for Transplantation
To efficiently and maximally obtain Crx-positive cells in our culture protocol, we decided to add Ara-C between d.d 21 and 24, which was close to the time of graft cell preparation. We then evaluated if Ara-C-treated cells could in fact differentiate into mature photoreceptor cells by immunolabeling rhodopsin, a mature photoreceptor marker.
The cells cultured with Ara-C treatment were harvested once on day 26 and replated, and further cultured for 7 days. Both the lectin-sorted and Rx-sorted cells expressed Crx and rhodopsin (Fig. 6A). The percentage of Crx/rhodopsin-positive colonies was determined on d.d. 26 (n = 4) (Fig. 6B). Ara-C treatment did not significantly increase the number of positive colonies in either the Rx-positive or lectin-negative selection.
(A) Differentiation into photoreceptor cells after either lectin selection (upper two photos) or Rx selection (lower two photos) with Ara-C treatment. Green: rhodopsin, red: Crx. (B) Ara-C was added on d.d. 21–24, the cells were fixed on d.d. 28, and the colonies positive for Crx/rhodopsin were counted against the total colonies (n = 4). Scale bars: 20 μm.
These cells were then tested for the contamination of Oct3 and Nanog expressing cells by RT-PCR (Fig. 4C). By d.d. 38, the gene expression levels were again reduced almost to comparable levels of immediately after the cell selection of d.d. 10, indicating that Ara-C may have inhibited the growth of undifferentiated cells.
Discussion
The selection of some specific type of cells using lectins has been rather rare compared to those using antibodies, partly because the affinity of lectins to sugar epitopes is generally weaker compared to that of antibodies to target antigens. In our protocol, incubation of the cells with the lectins at room temperature for 30 min followed by gentle wash resulted in consistent results. In fact, there are some preceding examples of cell selection using lectins, such as purification of photoreceptor cells using peanut agglutinin (PNA), microvessel endothelial cells with Ulex europaeus agglutinin-1 (UEA-1), and hematopoietic stem cells using a fucose-binding lectin (2, 6, 12). Because the glycoproteins on the cell surface were reported to change during the development and some are in fact used as markers for stage-specific embryonic antigens (5, 9, 10), we tested to see if some lectins could be used to identify retinal progenitor cells or Rx-positive cells among other differentiating cells from mouse ES cells. The lectins used in our protocol, WGA and ECA, have known sugar specificities; WGA is specific for sugar epitopes of β-N-glucosamin (GlcNAc) and sialic acid, and ECA is specific for α- and β-galactose (Gal) and α- and β-N-acetylgalactosamine (GalNAc), respectively.
Nash et al. studied the lectin affinity for mES cells and found that α-GalNAc was tightly associated with the pluripotent state of mES cells and was lost earlier than more popular pluripotent marker SSEA-1 upon differentiation, thereby suggesting that the lectin Dolichos biflours agglutinin (DBA), which binds to α-GalNAc, could be used as a good marker for mES cells (13). Additionally, in their study, ECA and WGA were also tested and they bound not only mES cells but also primitive ectoderm, suggesting that these lectins could recognize very early differentiating cells as well as mES cells of most undifferentiated state. This longer affinity along differentiation may have contributed for more elimination of Rx-negative cells in our experiment. The detailed characteristic of surface sugar epitopes on Rx-positive cells is yet to be determined. In the present study, WGA were especially good for eliminating Oct3/4-negative cells, indicating that either β-GlcNAc or sialic acid might be also rich on these cell surfaces. Detailed analysis may be needed to further identify the surface sugar characteristics of the cells at each differentiating stage. Presently, we do not have a good marker antibody for retinal progenitor cells, or Rx-positive cells, but our present protocol would reasonably substitute for the Rx enrichment procedure for any stem cell-derived cells without the use of the Rx knocked-in gene marker.
Photoreceptor cells equivalent to mouse postnatal days 3–7 were successfully incorporated into adult host retina. This developmental time period is roughly equivalent to d.d. 20–30 in our mES-differentiating culture as judged by immunolabeling and RT-PCR. Considering the clinical application of transplantation, graft cells should be maximally enriched to produce photoreceptors with or without some other retinal cells and undifferentiated, proliferating cells should be eliminated. Our results showed that lectin selection was not substantially inferior to Rx-positive selection. Lectin selection induced photoreceptors at comparable efficiency with Rx-positive selection and could also minimize the contamination of undifferentiated cells. Although there were very few Oct3/4 undifferentiated cells immediately after Rx-positive selection, they increased about 2 weeks after the sorting, indicating that residual undifferentiated cells might have proliferated. The use of Ara-C decreased the subsequent Oct3/4 and Nanog gene expression.
Even though the present protocol using lectins resulted in an equivalent efficiency for obtaining photoreceptor cells to Rx-positive selection, and the present method could be more widely applicable, there is still room for improvement in both the photoreceptor enrichment and the deletion of undifferentiated cells. One option may be to use additional cell surface markers of the appropriate graft stage or to further refine the culture protocol.
So far, no surface marker specific to early photoreceptors around the developmental stage of mouse p3–p7 is known. Nevertheless, Rx or lectin selection in the early preparatory stage is still important for the following reasons: 1) Selecting a certain neural lineage at some earlier point would enhance the generation of subsequent photoreceptors, and 2) if further final purification was needed, there would be more choices for purification method if the culture cells were of limited diversity.
As to the first point, we found no effect of DAPT with the unsorted culture cells (8); the cells should be enriched for certain cell types either by Rx or lectin for DAPT to effectively enhance photoreceptor differentiation. It is quite possible that any beneficial factor like DAPT would work more effectively, or at least stably, in the culture of more uniform cell types or environment. As to the second point, identifying a “photoreceptor-specific marker” among a limited lineage should be by no means easier than doing so among cell types of diverse lineage. For instance, PNA could work as cone-specific photoreceptor marker among retinal cells.
Conclusion
This protocol used a combination of the two lectins WGA and ECA for enrichment of retinal progenitor cells and can be easily applied to ES- or iPS-derived cells of any animal species, although some modification in the choice of lectins might be needed depending on the species because sugar chains may differ among the same cell types of different species. The efficiency and exclusion of undifferentiated cells must be improved, and we are currently seeking additional methods for preparation of higher quality graft cells.
Footnotes
Acknowledgment
This work was supported by a Grant from MEXT.