RESEARCH ARTICLE

c-Kit+ cells isolated from human fetal retinas represent a newpopulation of retinal progenitor cellsPeng-Yi Zhou1, Guang-Hua Peng1,2,*, Haiwei Xu3,4 and Zheng Qin Yin3,4,*

ABSTRACTDefinitive surface markers for retinal progenitor cells (RPCs) are stilllacking. Therefore, we sorted c-Kit+ and stage-specific embryonicantigen-4− (SSEA4−) retinal cells for further biological characterization.RPCswere isolated fromhuman fetal retinas (gestational age of 12–14weeks). c-Kit+/SSEA4− RPCs were sorted by fluorescence-activatedcell sorting, and their proliferation and differentiation capabilities wereevaluated by using immunocytochemistry and flow cytometry. Theeffectiveness and safety were assessed following injection of c-Kit+/SSEA4− cells into the subretina of Royal College of Surgeons (RCS)rats. c-Kit+ cells were found in the inner part of the fetal retina. Sortedc-Kit+/SSEA4− cells expressed retinal stem cell markers. Our resultsclearly demonstrate the proliferative potential of these cells. Moreover,c-Kit+/SSEA4− cells differentiated into retinal cells that expressedmarkers of photoreceptor cells, ganglion cells and glial cells. Thesecells survived for at least 3 months after transplantation into the hostsubretinal space. Teratomas were not observed in the c-Kit+/SSEA4−-cell group. Thus, c-Kit can be used as a surface marker for RPCs, andc-Kit+/SSEA4− RPCs exhibit the ability to self-renew and differentiateinto retinal cells.

KEY WORDS: Fetal retinal progenitor cells, c-Kit, Transplantation,Retinal degeneration

INTRODUCTIONPhotoreceptor degeneration occurs as a result of disorders affectingeither the photoreceptors themselves or the associated retinalpigment epithelium (RPE) cells. This disease is a common cause ofblindness and severely affects an individual’s quality of life (Pinillaet al., 2004). Photoreceptor degeneration is difficult to treat becausethe pathological process involves the apoptosis of photoreceptors orRPE cells, and photoreceptor cells cannot regenerate or self-repair.AlthoughRPE cells have the capacity to proliferate in vivo and in vitro(Chiba, 2014; Stanzel et al., 2014), it is difficult for damaged RPEcells to repair themselves (Chiba, 2014). Currently, there are severaltreatment methods available, including gene therapy, transplantationtherapy, drug therapy and artificial vision prostheses. Rescuing orregenerating photoreceptor cells or RPE in individuals with retinaldegeneration is the key to treatment. Recently, stem-cell-based celltherapy has become a hot topic. Schwartz and colleagues havereported the safety and tolerability of human embryonic stem cell(ESC)-derived RPE cells for the treatment of dry age-related macular

degeneration (AMD)andStargardt’s disease, and that report is the firstdescription of the transplantation of human-ESC-derived cells intohumans for the treatment of such diseases (Schwartz et al., 2012).Moreover, the authors observed no adverse proliferation or systemicsafety issues related to the transplanted cells, and the best-correctedvisual acuity was improved in some individuals (Schwartz et al.,2015). Thus, this human-ESC-based cell therapy in the treatment ofblindness-causing diseases is of great significance because it providesa promising foundation for the use of cell-therapy intervention innumerous diseases (Sowden, 2014).

Tissue-specific retinal progenitor cells (RPCs) are an ideal sourceof differentiated cells with low tumor risk (Kajstura et al., 2011).Currently, most RPCs that are used for transplantation are derivedfrom ESCs or are isolated from fetal tissues, but ESCs are difficult todifferentiate in vitro, are not easily purified and might contain avariety of cell types or cells at different development stages. Thecontinuous development of flow cytometry techniques to assess cellsurface antigens has provided biomarkers that can be used to obtainhighly purified tissue-specific RPCs. To date, several cell markershave proven to be suitable for the specific identification, isolationand enrichment of RPCs (Carter et al., 2009; Koso et al., 2009);these markers are invaluable for RPC research.

Stemcell factor receptor c-Kit (CD117), a progenitor cellmarker, isa recognized antigen that is located on the cell surface which plays animportant role in the survival and proliferation, as well as theprevention of apoptosis, of hematopoietic stem cells (HSCs) and lungcancer cells (Lennartsson and Ronnstrand, 2012). Koso andcolleagues have identified c-Kit as an RPC marker in the mouseretina and have demonstrated a dramatic change in the expressionprofiles of the cell surface antigens c-Kit and stage-specificembryonic antigens (SSEAs) on RPCs during development (Kosoet al., 2007). Hasegawa et al. have demonstrated that the humanembryonic retina has a pool of c-Kit+ cells; however, the authors didnot further culture or characterize them (Hasegawa et al., 2008).Stage-specific embryonic antigen-4 (SSEA-4), a human-ESC-associated antigen (Wright and Andrews, 2009), has previouslybeen used as amarker to distinguish primitive ESCs (Kawanabe et al.,2012). A subpopulation of c-Kit+ cells expressing SSEA1 showshigher proliferative potential than c-Kit+/SSEA1− cells (Koso et al.,2007). However, in contrast to mouse ESCs, human ESCs lackSSEA1 and express SSEA4 (Wright and Andrews, 2009). Therefore,to reduce the risk of tumorigenicity, we used SSEA4 as a surfacemarker in fluorescence-activated cell sorting (FACS) analyses toexclude cells with high proliferative potential that had been isolatedfrom the retina of human fetuses.

We hypothesized that c-Kit+ cells isolated from fetal retina tissuesrepresent a population of stem cells. The RPCs were evaluated forcell characteristics, including self-renewal capacity, clonogenicity,the ability to differentiate into three types of retinal cells in vitro andthe ability to differentiate into photoreceptors in vivo. Additionally,we injected the cells into the subretinal space of Royal College ofReceived 19 January 2015; Accepted 20 April 2015

1Department of Ophthalmology, The First Affiliated Hospital of ZhengzhouUniversity, Zhengzhou, He’nan 450003, China. 2Department of Ophthalmology,General Hospital of Chinese People’s Liberation Army, Beijing 100853, China.3Southwest Hospital/Southwest Eye Hospital, Third Military Medical University,Chongqing 400038, China. 4Key Lab of Ophthalmology of Chinese People’sLiberation Army, Chongqing 400038, China.

*Authors for correspondence (ghp@zzu.edu.cn; yzhengqin@163.com)

2169

© 2015. Published by The Company of Biologists Ltd | Journal of Cell Science (2015) 128, 2169-2178 doi:10.1242/jcs.169086

Journal

ofCe

llScience

Surgeon (RCS) rats, an animal model in which vision deterioratesdue to RPE dysfunction; these rats serve as a model for a recessivemutation in a receptor tyrosine kinase gene (Mertk) that results in theprogressive death of the photoreceptors (Pinilla et al., 2004). Then,we observed the survival, migration and differentiation of the cells,and determined their role in photoreceptor rescue and its impact onvisual function. Finally, we assessed the safety of transplantation ofc-Kit+/SSEA4− retinal progenitor cells into severe combinedimmune deficiency (SCID) mice. In our study, we found that thetransplantation of c-Kit+/SSEA4− cells derived from human fetalretina tissues could protect the neural retina and preserve the retinalouter nuclear layer in RCS rats. Therefore, this study describesa method that can be used to obtain tissue-specific RPCs thatcould be used to delay the photoreceptor degeneration and topreserve the retinal outer nuclear layer in an animal model of retinaldegeneration.

RESULTSDistribution of c-Kit+ cells in human fetal eyesThe fetal neural retina at 12–13 weeks mainly comprises two layers:the outer neuroblastic layer (ONbL) and the inner neuroblastic layer(INbL). The cells in the ONbL are densely packed, whereas the cellsin the INbL are more loosely packed. The ganglion cell layer can beseen around the optic disc, but not the peripheral retina.c-Kit+ cells were not only distributed in the retina, but also in

the cornea and choroid of the eye from a 13-week-old human fetus.

c-Kit+ cells were localized in the inner retina from the optic nerve tothe ora serrata (supplementary material Fig. S1), althoughmore cellswere located in the peripheral portion than at the posterior pole ofthe retina; c-Kit+ cells were also scattered in the corneosclerallimbus of the cornea (supplementary material Fig. S2) and choroid(Fig. 1A–C).

c-Kit+/SSEA4− cell isolation and cultureWe found c-Kit+ cells in the retinas of eyes from human fetuses.c-Kit+/SSEA4− cells were sorted directly from the fetal retina;however, these cells were difficult to culture. Therefore, we firstcultured retinal cells for 2–3 passages. Then, we collected retinalcells (a minimum of 5×106–10×106 cells) and sorted c-Kit+/SSEA4− cells by using FACS. These cells were collected and platedonto 24-well plates (5×103/cm2) (Fig. 1G). The morphology withinthe adherent population was spindle-shaped (Fig. 1H). By contrast,spheres were formed when the cells were grown in serum-freeproliferation medium (Fig. 1J). The c-Kit epitope remaineddetectable by using immunofluorescence and FACS analyses afterpassaging (Fig. 1I,K).

Characteristics of c-Kit+/SSEA4− cellsCells did not express the embryonic stem cell or tumor cell markersSSEA4 and the multi-drug resistance protein (MDR, also knownas ABCB1) (Fig. 2C,F). By contrast, they expressed the RPCmarkers – including Pax6, Sox2, Rax and nestin – as analysed by

Fig. 1. c-Kit+ cell distribution, isolation and culture. (A) Immunofluorescence staining of C-kit in the human retina. (B,D,E) c-Kit+ cells are located in the retinaand choroid. (C,F) C-kit cells are distributed in the corneoscleral limbus. (D,E) Regions in B at higher magnification. D is the upper box in B, and E is the lower boxin B. (F) Region in C at higher magnification. (G) Isolation of c-Kit+/SSEA4− cells by using FACS in the upper-left quadrant with depletion of ESCs. (H) c-Kit+/SSEA4− cells growing in a monolayer under adherent conditions. (J) Formation of neurospheres under non-adherent conditions after sorting. (I,K) Sorted cellswere positive for the C-kit receptor by (I) immunocytochemistry staining and (K) FACS. OS, ora serrate; INbL, inner neuroblastic layer; ONbL, outer neuroblasticlayer. Scale bars: 200 μm (A), 50 μm (B,C,H,I,J), 10 μm (D–F).

2170

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2169-2178 doi:10.1242/jcs.169086

Journal

ofCe

llScience

using flow cytometry and immunofluorescence. c-Kit+/SSEA4−

cells were found to express Pax6 (91.6±2.7%), Sox2 (95.2±2.0%),Rax (94.1±2.5%) and nestin (98.9±2.8%) (Fig. 3). Because c-Kit+

has also been described as a stem cell marker in other organs [i.e.mesenchymal stem cells (MSCs) and HSCs], we stained the sortedcells with MSC and HSC markers. The cells were negative for boththe HSC markers (CD11b and CD45, also known as ITGAM andPTPRC, respectively) (Fig. 2B,H) and MSC markers (CD29 andCD140b, also known as ITGB1 and PDGFRB, respectively)(Fig. 2E,I).

Characterization of proliferationThe cells exhibited the ability to proliferate and grow in a monolayeron plastic plates in proliferation medium (Fig. 1H), and to formspheres in serum-free proliferation medium (Fig. 1J). Proliferatingcells were identified by the proliferation marker Ki67 (also known asMKI67); the percentage of cells that expressed Ki67 was determinedto be 82.0±3.1% by FACS analysis, which was consistent with theimmunofluorescence staining results (Fig. 4A,B). c-Kit+/SSEA4−

cells were seeded onto 24-well plates at a density of 10,000 cells/well,and the cell numbers were counted 1, 3, 5, 7 and 9 days after seeding(n=3). Cell proliferation reached a maximum and then plateaued afterthe cells were plated; by day 7, the cell number had increased bymorethan 20-fold (Fig. 4D). Additionally, we examined the cell cycledistribution of the c-Kit+ cells and found that 41.13±2.99% of cellswere in the G2 and S phases (Fig. 4C).c-Kit+ cells were cultured under two different conditions –

adherent conditions with proliferation medium and non-adherent

conditions with serum-free medium. When a single c-Kit+ cell(sorted using flow cytometry) was seeded onto a 96-well plate,clones formed approximately 20 days later (Fig. 1H). The cloneswere digested, 500 cells were seeded onto 10-cm dishes and morecolonies formed. When these cells were cultured in serum-freemedium, neurospheres formed after 20–30 days (Fig. 1J). Thespheres were defined as free-floating with a diameter >40 μm. Thepercentage of cells that formed colonies after 30 days was 0.06±0.01%, but the efficiency of colony formation was less than thatobserved under adherent conditions (3.33%±1.53%) (P<0.01). Thisfinding is consistent with the results of a previous report (Rangelet al., 2013).

Characterization of differentiationMultipotency is another property of stem cells. Therefore, weused two differentiation media – photoreceptor differentiationmedium and glia differentiation medium. The in vitro differentiationwas assessed by culturing c-Kit+ cells in differentiation medium for3 weeks using the two types of differentiation medium. Fewdifferentiated cells expressed rhodopsin in the differentiationmedium. By contrast, many cells differentiated into glial cells(66.7±5.8%). Because previous studies have shown that retinoicacid can promote the differentiation of photoreceptor cells in vitro,we changed the medium in order to induce photoreceptordifferentiation, as previously described (Nakano et al., 2012; Liet al., 2013).

Photoreceptor differentiation medium containing retinoic acidwas used to culture c-Kit+/SSEA4− cells for 3 weeks. Then, we

Fig. 2. The phenotype of C-kit cells.(A,D,G) Negative controls. (B,H) Distributionplots of c-Kit+ cells that were negative for HSCmarkers (CD11b and CD45). (E,I) Distributionplots of c-Kit+ cells that were negative formesenchymal stem cell markers (CD29 andCD140b). (C,F) Distribution plots of c-Kit+ cellsthat were negative for ESC markers [SSEA4and MDR (also known as CD243 and ABCB1,respectively)].

2171

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2169-2178 doi:10.1242/jcs.169086

Journal

ofCe

llScience

stained for photoreceptor cell markers (Otx2, Crx, rhodopsin andrecoverin), a ganglion cell marker (Thy1) and a glial cell marker(GFAP) (Fig. 5G–L). Otx2 (49.9±4.1%), Crx (59.9±4.0%),recoverin (68.1±5.1%) and rhodopsin (4.0±0.3%) were expressedin the differentiated cells, and 29.1±5.4% of the cells expressedThy1 (Fig. 5A–F). By contrast, the cells that had been cultured in theinitial differentiation medium only expressed GFAP and Thy1.

Differentiation of grafted cellsThe results of our immunostaining showed that some cells migratedinto the inner retina at 8 and 12 weeks after transplantation, andthat a small number of cells expressed recoverin (Fig. 6H).However, the majority of the cells were located in the subretinalspace. The percentage of transplanted cells that expressed

photoreceptor cell markers at 4, 8 and 12 weeks were 1.01±0.11%, 2.36±0.25% and 5.22±0.14%, respectively.

Outer nuclear layer thicknessTo verify the protective effect of transplanted cells on retinaldegeneration, we measured retinal outer nuclear layer (ONL)thickness. The ONL of the cell-grafted retina was significantlythicker compared to the control and untreated groups at4 weeks (28.43±1.95 μm versus 7.67±1.08 μm and 8.50±1.47 μm,n=3, P<0.01), 8 weeks (23.27±0.85 μmversus 6.61±0.65 μm, 6.83±1.08 μm, n=3, P<0.01) and 12 weeks (19.43±0.84 μm versus 4.17±0.75 μm, 4.80±1.08 μm, n=3, P<0.01). There was no significantdifference in ONL thickness between the control and untreatedgroups (Fig. 6M).

Electroretinogram measurementsElectroretinogram (ERG) analysis was performed, and b-waveswere measured at 4 weeks, 8 weeks and 12 weeks. Significant high-amplitude b-waves were detected in the transplanted group at 4 and8 weeks compared to the control and untreated groups (P<0.01).Recordings at 12 weeks were not significantly different between thesham-surgery and untreated groups (P>0.05) (Fig. 6N–Q).

Teratoma assayThe safety of c-Kit+/SSEA4− cells was tested through subcutaneousinjection into the groin of six SCID mice; human ESCs wereinjected into another six SCID mice as a positive control. After 8weeks, no gross inflammatory reaction was observed in any of theanimals, and teratomas were not observed in the c-Kit+/SSEA4−-cell group. By contrast, teratomas were seen in the human-ESCgroup at 8 weeks post injection (Fig. 7).

DISCUSSIONIn 2011, the FDA approved PhaseI/II clinical trials of celltransplantation therapy for AMD and Stargardt’s disease, which

Fig. 3. Retinal progenitor cell markers were detected by FACS andimmunocytochemistry staining. (A–D) More than 90% of c-Kit+ cellsexpressed RPC-specific markers and the neural stem cell markers nestin, Rax,Sox2 and Pax6, as detected by using FACS (n=3 independent experiments).(E–H) Representative immunofluorescence staining for Nestin, Rax, Sox2 andPax6. Scale bar: 50 μm.

Fig. 4. Proliferation of c-Kit+ cells. (A) Flow cytometry and (B)immunofluorescence staining were performed to detect the proliferationcapacity of c-Kit+ cells; more than 80% of C-kit cells expressed Ki67. Scale bar:50 μm. (C) The cell cycle distribution of c-Kit+ cells was examined; a total of41.13±2.99% of cells were in the G2 and S phases. (D) Growth curve of C-kitcells; a total of 10,000 cells/cm2 were plated, and the cell number increasedmore than 20-fold after 7 days. n=3 independent experiments. Dip, diploid.

2172

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2169-2178 doi:10.1242/jcs.169086

Journal

ofCe

llScience

led to immense advances in transplantation treatment of retinaldegeneration (Schwartz et al., 2012). RPC transplantation for thetreatment of retinal degenerative diseases has become the mostpromising therapeutic strategy. c-Kit+ RPCs are derived fromfetal retinas and have advantages that include the potential todifferentiate into retinal cells and a low risk of tumor formation.These factors make c-Kit a good marker for the selectionof candidate cells for transplantation in order to treat retinaldegeneration.Our results showed that cells expressing the c-Kit epitope on

their cell surface were distributed in the eye. c-Kit+/SSEA4−

cells possessed characteristics of self-renewal and the abilityto differentiate into three types of retinal cells. Owing to thedevelopment of flow cytometry technology and the discovery ofnew cell surface antigens, researchers have been able to obtain ahigher purity of stem cells by using FACS isolation. c-Kit defines aregionally and temporally restricted immature subset of RPCs, theexpression of which starts centrally and progresses centrifugally(Koso et al., 2007).At 9–10 fetal weeks, the retina is divided into the ONbL and the

INbL. At 11–13 fetal weeks, the ganglion cell layer (GCL) is severalcell layers thick. At this stage, the outer boundary of the ONbLadjacent to the RPE is bordered by a row of cones, and the ONbL isseparated from the differentiating GCL by a thin inner plexiformlayer. The nerve fiber layer also becomes apparent on the innerboundary of the GCL close to the optic disc. The expression ofrecoverin and of the cone marker S-opsin does not begin until 11–12

weeks, and rod markers are expressed at 15–16 weeks (O’Brienet al., 2003; Hendrickson et al., 2008). In our study, fetal neuralretina at 12–13 weeks mainly comprised two layers – the ONbLand the INbL, which is consistent with previous reports onthe development of the fetal retina (O’Brien et al., 2003;Hendrickson et al., 2008). We found that c-Kit+ cells not onlyexisted in the retina of the fetal eye but that they were also located inthe cornea and choroid. Our results showed that c-Kit cells werelocated in the inner layer of the retina, and this finding is inagreement with the demonstration that the c-Kit ligand SCF is alsoexpressed in the inner retina (Hasegawa et al., 2008). Wesuccessfully isolated and cultured c-Kit+ cells that had beenderived from the human fetal retina, and identified their self-renewal, proliferation and differentiation characteristics. We alsosuccessfully used surface markers to isolate tissue-specific RPCsthat did not express SSEA-4.

It is important to obtain pure RPCs and to exclude other types ofESCs after isolation of eye tissue from the human fetus. The sortingof both the c-Kit+ and SSEA4− surface markers allowed us toobtain purified RPCs. Additionally, the sorted cells were stainedfor stem cell markers (MDR) and then isolated by using flowcytometry to ensure that there was no contamination with ESCs; atotal of 99% of the double-marker-sorted cells did not express theMDR marker.

c-Kit+ cells have been previously reported to be able to surviveunder suspension or adherent growth conditions (Koso et al., 2007;Kajstura et al., 2011; Rangel et al., 2013). The c-Kit cells isolated inthis study could also survive under both conditions. A higherproliferation rate was observed under adherent conditions [inmedium supplemented with fetal bovine serum (FBS)], which isconsistent with the findings of Rangel et al. (Rang et al., 2013). Onepossible reason for this finding is that the cell–cell interactions andthe cell adhesions that are present under adherent conditions areimportant for c-Kit+-cell growth (Rangel et al., 2013). Retinalprogenitor cells grown under adherent conditions have also beenreported to exhibit a greater proliferative potential than cells grownin suspension (Xia et al., 2012).

To ensure that the c-Kit+/SSEA4− cells that had been sorted usingFACS were highly tissue-specific, these cells were further evaluatedfor the human RPC tissue-specific markers nestin, Rax, Pax6 andSox2 by using immunohistochemistry and flow cytometry analyses.More than 90% of the cells expressed the RPC markers Pax6, Sox2,Rax and nestin, indicating that the cells were in an immature retinalcell state (Schmitt et al., 2009). Additionally, cell proliferationpotential was assessed using the Ki67 marker. More than 80% of thec-Kit+/SSEA4− cells expressed this marker, demonstrating that thec-Kit+/SSEA4− cells possessed a high proliferation ability in vitroand represented a highly pure and tissue-specific RPC with a certainproliferative capacity.

Furthermore, we found that c-Kit+/SSEA4− cells could formcolonies under adherent and suspension conditions. This finding isconsistent with the other properties of c-Kit cells that have beenreported by Kajstura et al. and Rangel et al. (Kajstura et al., 2011;Rangel et al., 2013).

In addition to proliferation capability, another characteristic ofprogenitor cells is the capacity to differentiate. In differentiationmedium, c-Kit+/SSEA4− cells could be induced into photoreceptorcells, ganglion cells and glial cells that expressed the correspondingcell-specific markers (Otx2, Crx, recoverin, rhodopsin, Thy1 andGFAP). However, in our study the proportion of cells expressingrhodopsin (4.0±0.3%) was lower than the proportion reported in thestudy by Coles et al. (34.5±9.1%); furthermore, we found that more

Fig. 5. The differentiation capacity of c-Kit+/SSEA4− cells. (A–F) The c-Kit+

cells differentiated into photoreceptor cells, ganglion cells and glial cells, andexpressed cell-specific markers – Otx2, Crx, recoverin, rhodopsin, theganglion cell marker Thy1 and the glial cell marker GFAP. Scale bars: 50 μm.(G–L) Staining of Otx2, Crx, recoverin, rhodopsin, Thy1 and GFAP wasevaluated by using FACS. Green line, isotype.

2173

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2169-2178 doi:10.1242/jcs.169086

Journal

ofCe

llScience

Fig. 6. See the next page for legend.

2174

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2169-2178 doi:10.1242/jcs.169086

Journal

ofCe

llScience

cells differentiated into glial cells (66.7±5.8%) than has beenpreviously reported (19.7±10.6%) (Coles et al., 2004).The transplantation of stem cells into the subretina of rats or mice

that are afflicted by retinal degeneration has been demonstrated toproduce a neuroprotective effect (Tian et al., 2011; Tzameret et al.,2014). In our study, we found that transplantation of c-Kit+/SSEA4−-cells that had beenderived fromhuman fetal retina tissue could protectneural retinas in RCS rats. The RCS rat is characterized by a recessivemutation in theMertk gene, which encodes a receptor tyrosine kinase.This mutation precludes RPE cells from phagocytosing rod outer

segments that have been shed, leading to the progressive death ofphotoreceptor cells (Pinilla et al., 2004). At (P)18 in these rats,numerous morphological changes are already apparent, including thedisturbance of outer segments (Davidorf et al., 1991). After P21, therod contribution to the mixed b-wave starts declining (Pinilla et al.,2004) and, at P22, changes in photoreceptor nuclei are found (Cuencaet al., 2005). We transplanted cells into the subretina at P21, which iswhen photoreceptor degeneration begins (Luo et al., 2014). Ourexperiment showed that grafted c-Kit+/SSEA4− cells improved ERGb-wave amplitude, which represents the electrical function of theretina (Tian et al., 2011), in RCS rats for 2 months aftertransplantation; this finding is consistent with the results of aprevious study by Tian and colleagues (Tian et al., 2011).

It has been noted by Gonzalez-Cordero and colleagues (Gonzalez-Cordero et al., 2013) that reliable electroretinographic responses areonlyachieved inmice following the rescueof 150,000 functioning rods.Our study showed that ONL thickness was maintained for at least 3months after cell transplantation; however, the thickness decreasedover time. This decrease is likely to be due to the fact that there werenot enough functional photoreceptors in the third month aftertransplantation. The grafted cells expressed the photoreceptor markerrecoverin. However, the migrated human RPCs failed to express theganglion cell marker Thy1. It has been previously reported (Luo et al.,2014) that stem cells grafted into the degenerative retina have difficultydifferentiating intoganglion andphotoreceptorcells in vivo; this findingwas consistent with the results from our study.

Finally, the major concern regarding the use of progenitor cellsfor transplantation is tumorigenesis. The results from the teratomaassay showed that the transplantation of human RPCs is safe. Therewas no evidence of tumor formation 8 weeks after human RPCtransplantation into the subretinal space of RCS rats.

In summary, our study demonstrates that c-Kit+/SSEA4− cellsexist in human fetal eyes and that these cells possess the stem cellproperties of self-renewal, colony formation and pluripotentdifferentiation. Although only a small proportion of the engraftedc-Kit+/SSEA4− cells in a rat model of retinal degeneration coulddifferentiate to express the photoreceptor marker, c-Kit+/SSEA4−

cells delayed apoptosis-mediated photoreceptor death or rescuedhost photoreceptor cells for at least 3 months. Therefore, c-Kit+ andSSEA4− might serve as good markers for the selection of candidatecells for transplantation in order to delay retinal degeneration.

MATERIALS AND METHODSCell isolation and cultureEyes from human fetuses with a gestational age ranging from 12 to 14 weekswere obtained from spontaneous abortions at the Southwest Hospital, ThirdMilitary Medical University (Chongqing, China). The Ethics Committee ofSouthwest Hospital specifically approved this study, and it is registeredin the Chinese Clinical Trial Register (ChiCTR; registration number –ChiCTR-TNRC-08000193). The gestational age of each fetus wasdetermined using the last menstruation date and fetal foot length (Merzet al., 2000; Drey et al., 2005). Postmortem times of less than 1 h were usedbecause they do not alter the ability of progenitor cells to survive in culture(Carter et al., 2007).

Cells were isolated from the neuroretinas of human fetal eyes aspreviously described (Coles et al., 2004; Klassen et al., 2004; Aftab et al.,2009; Schmitt et al., 2009; Baranov et al., 2013). Briefly, the eyes wererinsed in cold Hank’s buffered salt solution (HBSS; Hyclone, South Logan,UT). The neuroretina was dissociated into small pieces and enzymaticallydigested with 1 ml of papain (12 units/ml; Worthington, Lakewood, NJ).The digested retinal tissue supernatant was filtered through a 40 μm filter(BD Biosciences, Franklin Lakes, NJ) to obtain single cells. The cells werecentrifuged and re-suspended in proliferation medium that had beensupplemented with FBS. The isolated cells were seeded into 6-well plates at

Fig. 6. Differentiation, migration, ONL protection of c-Kit+ cells in vivo,and ERG measurements after transplantation. (A,E,I) Wild-type rat retinas.(B,F,J) The retinas from sham-surgery rats. (C,G,K) DiI-labeled c-Kit+ cellsmigrated into the inner retina at 4, 8 and 12 weeks after transplantation, andthey expressed the photoreceptor marker recoverin. (D,H,L) Regions of C,G,Kat highermagnification, respectively. (M) TheONL of the cell-grafted retinawassignificantly thicker than that of the control and untreated groups at 4 weeks(P<0.01), 8 weeks (P<0.01) and 12 weeks (P<0.01). There were no significantdifferences in ONL thickness between the control and untreated groups.(N–Q) ERGanalyses were performed and b-waves weremeasured at 4 weeks,8 weeks and 12 weeks. There were significantly high-amplitude b-waves inthe transplanted group at 4 weeks and 8 weeks compared with those of thesham-surgery and untreated groups (P<0.01). However, there was nosignificant difference between the sham-surgery and untreated groups after12 weeks. **P<0.01. Scale bars: 50 μm (A–C,E–G,I–K); 20 μm (D,H,L).GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer.

Fig. 7. Teratoma assayof c-Kit+/SSEA4− cells inSCIDmice. (A,C) Teratomaswerenot observed in thec-Kit+/SSEA4−-cell group. (B,D)Teratomaformationwasdetected in one out of three mice at 8 weeks in the human-ESC group. (E) Theproportion of c-Kit+ cells and human-ESCs in the teratoma assay. (F) A teratomaderived from human ESCs. Scale bar: 1 cm.

2175

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2169-2178 doi:10.1242/jcs.169086

Journal

ofCe

llScience

a density of 5×105 cells/well. The cells were cultured at 37°C under 5%CO2,and the medium was changed every 3 days.

The following antibodies were used – allophycocyanin (APC)-conjugatedanti-human c-Kit antibody (Biolegend, San Diego, CA), fluoresceinisothiocyanate (FITC)-conjugated anti-human SSEA-4 antibody (BDBiosciences), phycoerythrin (PE)-conjugated anti-human CD29 antibody(Biolegend), PE-conjugated anti-human multidrug resistance protein (MDR)antibody (number 348605, Biolegend), APC-CY7-conjugated anti-humanCD45 antibody (Biolegend), FITC-conjugated anti-human CD11b antibody(BD Biosciences), and PE-conjugated anti-human CD140b antibody(Biolegend).

The proliferation medium included Dulbecco’s modified Eagle mediumwith nutrient mixture F-12 (DMEM/F-12; Hyclone) that had beensupplemented with 20 ng/ml fibroblast growth factor-basic (bFGF,PeproTech, Rocky Hill, NJ); 20 ng/ml epidermal growth factor (EGF,PeproTech); 1×insulin, transferrin and selenium (ITS; GIBCO, Carlsbad,CA); 1×penicillin-streptomycin (GIBCO); and 10% FBS (GIBCO).

ImmunocytochemistryImmunocytochemistry was performed as previously described (Tian et al.,2011; Pearson et al., 2012). Briefly, rats were euthanizedwith an overdose ofanesthesia, and the eyes were enucleated and fixed in 4% paraformaldehyde(0.01 M, pH 7.4). For human fetuses, the eyes were enucleated and fixed in2% paraformaldehyde (0.01 M, pH 7.4) (Hendrickson et al., 2008). The eyecups were immersed in a graded series of sucrose solutions overnight,embedded in an optimal cutting temperature compound and then sectionedon a cryostat (Leica CM190, Wetzlar, Germany). Frozen tissues sectionswere cut as 10-μm-thick transverse sections.

Immunocytochemistry was performed using our previously describedmethods (Duan et al., 2013). Briefly, the cells or slides were fixed andpermeabilized, then blocked in 10% serum for 30 min. The cells wereincubated with primary antibody at 37°C for 2 h, washed with PBS,incubated with secondary antibody, washed with PBS, incubated with 6-diamidino-2-phenylindole (DAPI) for 10 min at room temperature and thenwashed with PBS.

The following primary antibodies were used – anti-c-Kit antibody (1:100,R&D Systems, Minneapolis, MN), anti-Pax6 (1:200, Abcam, Cambridge,MA), anti-Sox2 (1:100, Abcam), anti-rhodopsin (1:200, Abcam), anti-Otx2(1:200, Abcam), anti-Rax (1:100, Abcam), anti-nestin (1:200, Sigma, StLouis, MO), anti-Ki67 (1:300, Sigma), anti-GFAP (1:300, Abcam), anti-Thy1 (1:100, BDBiosciences) and anti-recoverin (1:10,000,Millipore). Thefollowing secondary antibodies were used – Cy3-conjugated antibody(1:800, Beyotime, NanTong, JiangShu, China) and FITC-conjugatedantibody (1:200, Beyotime). Additionally, the cells were stained withDAPI (1:10, Beyotime). The images were captured using an Olympus OP70microscope (Olympus Microscopy, Japan) or Leica TCS SP50 confocalmicroscope (Leica Microsystems, Wetzlar, Germany).

Fluorescence-activated cell sortingThecellswere digestedwithHyQtase for 5min, followedby the additionof3mlof wash buffer (Biolegend) and centrifugation at 400 g for 5 min at 4°C. Then,the cells were re-suspended in Stain Buffer (Biolegend), and 2 µl of Fc block(Biolegend) was added to each tube. The cells were incubated for 15min at 4°Cand then incubated with fluorochrome antibody conjugates for 20–30 min on ashaker at 4°C. The cells werewashed with wash buffer and centrifuged at 200 gfor 5min at 4°C. The supernatant was removed, 300 μl of PBSwas added to thecells, and the suspension was transferred to a standard flow cytometry tube forFACS analysis (BD Biosciences, AriaII, Franklin Lakes, NJ).

Expansion of c-Kit+ cellsSorted cells were plated at a density of 3×103–5×103/cm2 and cultured inproliferation medium DMEM/F12 that had been supplemented with20% FBS, 20 ng/ml bFGF, 20 ng/ml EGF and 1×ITS for 3 days. Then,the mediumwas changed, and the FBS concentration was decreased to 10%.The medium was changed every 3 days, and the cells were passaged untilthey reached 80% confluence (every 3–4 days) at a constant seeding densityof 10×103–13×103 cells/cm2.

Cell differentiationAnalysis of cell differentiation was performed using previously describedmethods (Coles et al., 2004; Li et al., 2013). Briefly, 5×103 cells/well ofRPCs that had been passaged three times were seeded into 24-well plates onglass coverslips pre-coated with 0.015 mg/ml poly-lysine (Sigma) in gliadifferentiation medium. For photoreceptor and other retinal celldifferentiation, 5×103 cells were seeded into 24-well plates on coverslipspre-coated with poly-L-lysine in photoreceptor differentiation medium; thedifferentiation medium was changed every 4 days.

The glia differentiation medium comprised DMEM/F12 (Hyclone) thathad been supplemented with 10 ng/ml bFGF (PeproTech), 1× penicillin-streptomycin (GIBCO), 1% FBS (GIBCO) and 2 μg/ml heparin (Sigma).

The photoreceptor differentiation medium comprised DMEM/F12(Hyclone) that had been supplemented with 10 ng/ml bFGF (PeproTech),1×penicillin-streptomycin (GIBCO), 500 nM retinoic acid (Sigma) and 2%B27 (GIBCO).

Cell proliferation curveThe cell proliferation assay was performed as previously described (Aftabet al., 2009). Briefly, RPCs that had been passaged three times were seededinto 15 wells of a 24-well plate at a density of 10,000 cells/well. The cellnumber was counted in three wells at 1, 3, 5, 7 and 9 days; the experimentwas repeated three times. The average number of cells was used to generatethe cell proliferation curve.

Animal feedingThe animals were treated as described under a protocol approved by theInstitutional Animal Care and Use Committee of the Third Military MedicalUniversity in accordance with the National Institutes of Health guidelinesfor the care and use of laboratory animals, and with the Use of Animals inOphthalmic and Visual Research (ARVO) statement. Mice and rats were fedand housed under a 12 h light–dark cycle. The drinking water of the ratscontained cyclosporine A (210 mg/l) from 1 day before transplantation untilthey were euthanized (Lu et al., 2010).

Cell transplantationRCS rats (males and females, 2–3 weeks old) were used for celltransplantation. Rats with congenital microphthalmia, dysplasia ofextremities or congenital cataracts were excluded from the study.

The RCS rats were randomly divided into two groups: the cell graftedgroup (n=9), in which the rats received a subretinal injection of 3 μl of a c-Kit+/SSEA4− cell suspension (cell concentration, 2×105 cells/μl) and thePBS group (n=9), in which the rats received a subretinal injection of 3 μl ofHBSS. In both groups, the right eyes (oculus dexter, OD) received celltransplants (cell grafted group) or HBSS (group), and the left eyes (oculussinister, OS) were untreated. The left eyes of the cell-grafted group served asthe untreated group.

Transplantation methods were performed as previously described (Tianet al., 2011).All ratswere anesthetizedwith a single intra-peritoneal injectionof 4% chloral hydrate (0.8 ml/100 g of body weight). The pupils weredilated with 1% tropicamide. A Hamilton syringe (29 gauge; Hamilton,Reno, NV) containing the DiI-labeled cell suspension was injected intothe subretinal space. DiI was obtained from Invitrogen (Grand Island, NY).

ElectroretinogramThe ERG techniques used were performed as previously described (Tianet al., 2011). The animals were tested at 4 weeks, 8 weeks and 12 weeks. Allrats were dark-adapted overnight. The anesthesia method used was asdescribed in ‘Cell transplantation’. A Flash-ERG recording electrode,comprising a small silver ring, was positioned on the corneal surface with adrop of methyl cellulose and used to record responses (Roland system,Wiesbaden, Germany). Each ERG response represents the average of threeflashes. For all Flash-ERG recordings, the b-wave amplitude was measuredfrom the a-wave trough baseline to the peak of the b-wave, and b-wavelatency was measured from the onset of the stimulus to the b-wave peak.ERG b-waves were generated with flashes of white light at intensities of−0.3 cds−1 m−2 and 3.0 cds−1 m−2.

2176

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2169-2178 doi:10.1242/jcs.169086

Journal

ofCe

llScience

Teratoma assayc-Kit+/SSEA4− cells (1×107 cells/100μl) were injected into the groin in sixSCID mice, and the animals were observed for 8 weeks to detect possibletumor formation; human ESCs were injected into six SCID mice as apositive control. The animals were anesthetized and examined by apathologist to identify microscopic pathological changes and evidence oftumor formation. The human ESC line H-1 (WA-01; Ma et al., 2014) waskindly provided by Yue Huang (School of Basic Medicine, Peking UnionMedical College, Beijing, China).

Analysis of the thickness of the outer nuclear layerThree areas of retinal ONL thickness were examined in the transplanted area(but not in the area that contained the layers of transplanted cells) in the treatedand untreated groups, and in the sham-surgery group. The thickness of theONL was evaluated in DAPI-stained sections in three areas along the graftedhalf of the retina. The ONL thickness was measured using Image-Pro Expresssoftware.

Cell counts and analysisThe number of DiI and recoverin double-positive cells in each image wascounted at three locations in three areas along the grafted half of the retinafrom the retinal margin to the posterior pole from three rats for statisticalanalysis. Every fifth section was counted (50 µm) to avoid counting the samecell in more than one section; the cells were counted at a ×400 magnification(Wan et al., 2007; Xu et al., 2013).

Statistical analysisStatistical analyses were performed using SPSS for Windows Version 13.0.Data are described as mean±standard error. Statistical comparisons weremade using either Student’s two-tailed t-test or analysis of variance.Differences were considered to be statistically significant at P<0.05.

AcknowledgementsThe authors thank Dr. Xiaoli Liu (Southwest Hospital/Southwest Eye Hospital, ThirdMilitary Medical University, Chongqing, China; Division of Pulmonary and CriticalCareMedicine; Departments of Pediatric NewbornMedicine, Brigham andWomen’sHospital and Harvard Medical School, USA) for technical support, and thankYuxiao Zeng, Qiyou Li and Chuanhuang Weng for assistance with celltransplantation and ERG analyses.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsPeng-YiZhoucontributed toconceptionanddesign,datacollectionandanalysis,aswellas writing and revision of the manuscript. Guang-Hua Peng contributed to conceptionand design, data analysis and interpretation, and the provision of studymaterial, aswellas to writing and revision of the manuscript. Haiwei Xu contributed to experimentaldesign, dataanalysis and interpretation,aswell as revisionof themanuscript. ZhengQinYin contributed to conception and design of the study, and provided study material.

FundingThis work was supported by the National Key Basic Research Program of China[grant 2013CB967001 to Guang-Hua Peng and grant 2013CB967002 to Zheng QinYin]; and the National Natural Science Foundation of China [grant 31271400 toGuang-Hua Peng].

Supplementary materialSupplementary material available online athttp://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.169086/-/DC1

ReferencesAftab, U., Jiang, C., Tucker, B., Kim, J. Y., Klassen, H., Miljan, E., Sinden, J. andYoung, M. (2009). Growth kinetics and transplantation of human retinal progenitorcells. Exp. Eye Res. 89, 301-310.

Baranov, P., Regatieri, C., Melo, G., Clissold, H. and Young, M. (2013). Syntheticpeptide-acrylate surface for self-renewal of human retinal progenitor cells. TissueEng. Part C Methods 19, 265-270.

Carter, D. A., Mayer, E. J. and Dick, A. D. (2007). The effect of postmortem time,donor age and sex on the generation of neurospheres from adult human retina.Br. J. Ophthalmol. 91, 1216-1218.

Carter, D. A., Dick, A. D. and Mayer, E. J. (2009). CD133+ adult human retinal cellsremain undifferentiated in Leukaemia Inhibitory Factor (LIF).BMCOphthalmol. 9, 1.

Chiba, C. (2014). The retinal pigment epithelium: an important player of retinaldisorders and regeneration. Exp. Eye Res. 123, 107-114.

Coles, B. L. K., Angenieux, B., Inoue, T., Del Rio-Tsonis, K., Spence, J. R.,McInnes, R. R., Arsenijevic, Y. and van der Kooy, D. (2004). Facile isolationand the characterization of human retinal stem cells. Proc. Natl. Acad. Sci. USA101, 15772-15777.

Cuenca, N., Pinilla, I., Sauve, Y. and Lund, R. (2005). Early changes in synapticconnectivity following progressive photoreceptor degeneration in RCS rats.Eur. J. Neurosci. 22, 1057-1072.

Davidorf, F. H., Mendlovic, D. B., Bowyer, D. W., Gresak, P. M., Foreman, B. C.,Werling, K. T. and Chambers, R. B. (1991). Pathogenesis of retinal dystrophy inthe Royal College of Surgeons rat. Ann. Ophthalmol. 23, 87-94.

Drey, E. A., Kang, M. S., McFarland, W. and Darney, P. D. (2005). Improving theaccuracy of fetal foot length to confirm gestational duration.Obstet. Gynecol. 105,773-778.

Duan, P., Xu, H., Zeng, Y., Wang, Y. and Yin, Z. Q. (2013). Human bone marrowstromal cells can differentiate to a retinal pigment epithelial phenotype when co-cultured with pig retinal pigment epithelium using a transwell system.Cell. Physiol.Biochem. 31, 601-613.

Gonzalez-Cordero, A., West, E. L., Pearson, R. A., Duran, Y., Carvalho, L. S.,Chu, C. J., Naeem, A., Blackford, S. J. I., Georgiadis, A., Lakowski, J. et al.(2013). Photoreceptor precursors derived from three-dimensional embryonic stemcell cultures integrate and mature within adult degenerate retina. Nat. Biotechnol.31, 741-747.

Hasegawa, T., McLeod, D. S., Prow, T., Merges, C., Grebe, R. and Lutty, G. A.(2008). Vascular precursors in developing human retina. Invest. Ophthalmol. Vis.Sci. 49, 2178-2192.

Hendrickson, A., Bumsted-O’Brien, K., Natoli, R., Ramamurthy, V., Possin, D.and Provis, J. (2008). Rod photoreceptor differentiation in fetal and infant humanretina. Exp. Eye Res. 87, 415-426.

Kajstura, J., Rota, M., Hall, S. R., Hosoda, T., D’Amario, D., Sanada, F., Zheng,H., Ogorek, B., Rondon-Clavo, C., Ferreira-Martins, J. et al. (2011). Evidencefor human lung stem cells. N. Engl. J. Med. 364, 1795-1806.

Kawanabe, N., Murata, S., Fukushima, H., Ishihara, Y., Yanagita, T., Yanagita, E.,Ono, M., Kurosaka, H., Kamioka, H., Itoh, T. et al. (2012). Stage-specificembryonic antigen-4 identifies human dental pulp stem cells. Exp. Cell Res. 318,453-463.

Klassen, H., Ziaeian, B., Kirov, I. I., Young, M. J. and Schwartz, P. H.(2004). Isolation of retinal progenitor cells from post-mortem human tissueand comparison with autologous brain progenitors. J. Neurosci. Res. 77, 334-343.

Koso, H., Satoh, S. and Watanabe, S. (2007). c-kit marks late retinal progenitorcells and regulates their differentiation in developing mouse retina. Dev. Biol. 301,141-154.

Koso, H., Minami, C., Tabata, Y., Inoue, M., Sasaki, E., Satoh, S. and Watanabe,S. (2009). CD73, a novel cell surface antigen that characterizes retinalphotoreceptor precursor cells. Invest. Ophthalmol. Vis. Sci. 50, 5411-5418.

Lennartsson, J. and Ronnstrand, L. (2012). Stem cell factor receptor/c-Kit: frombasic science to clinical implications. Physiol. Rev. 92, 1619-1649.

Li, T., Lewallen, M., Chen, S., Yu, W., Zhang, N. and Xie, T. (2013). Multipotentstem cells isolated from the adult mouse retina are capable of producing functionalphotoreceptor cells. Cell Res. 23, 788-802.

Lu, B., Wang, S., Girman, S., McGill, T., Ragaglia, V. and Lund, R. (2010). Humanadult bone marrow-derived somatic cells rescue vision in a rodent model of retinaldegeneration. Exp. Eye Res. 91, 449-455.

Luo, J., Baranov, P., Patel, S., Ouyang, H., Quach, J., Wu, F., Qiu, A., Luo, H.,Hicks, C., Zeng, J. et al. (2014). Human retinal progenitor cell transplantationpreserves vision. J. Biol. Chem. 289, 6362-6371.

Ma, Y., Yao, N., Liu, G., Dong, L., Liu, Y., Zhang, M., Wang, F., Wang, B., Wei, X.,Dong, H. et al. (2014). Functional screen reveals essential roles of miR-27a/24 indifferentiation of embryonic stem cells. EMBO J. 34, 361-378.

Merz, E., Oberstein, A. and Wellek, S. (2000). Age-related reference ranges forfetal foot length. Ultraschall Med. 21, 79-85.

Nakano, T., Ando, S., Takata, N., Kawada, M., Muguruma, K., Sekiguchi, K., Saito,K., Yonemura, S., Eiraku, M. and Sasai, Y. (2012). Self-formation of optic cups andstorable stratified neural retina from human ESCs. Cell Stem Cell 10, 771-785.

O’Brien, K. M., Schulte, D. and Hendrickson, A. E. (2003). Expression ofphotoreceptor-associated molecules during human fetal eye development. Mol.Vis. 9, 401-409.

Pearson, R. A., Barber, A. C., Rizzi, M., Hippert, C., Xue, T., West, E. L., Duran,Y., Smith, A. J., Chuang, J. Z., Azam, S. A. et al. (2012). Restoration of visionafter transplantation of photoreceptors. Nature 485, 99-103.

Pinilla, I., Lund, R. D. and Sauve, Y. (2004). Contribution of rod and conepathways to the dark-adapted electroretinogram (ERG) b-wave following retinaldegeneration in RCS rats. Vision Res. 44, 2467-2474.

Rangel, E. B., Gomes, S. A., Dulce, R. A., Premer, C., Rodrigues, C. O.,Kanashiro-Takeuchi, R. M., Oskouei, B., Carvalho, D. A., Ruiz, P., Reiser, J.et al. (2013). C-kit(+) cells isolated from developing kidneys are a novel populationof stem cells with regenerative potential. Stem Cells 31, 1644-1656.

2177

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2169-2178 doi:10.1242/jcs.169086

Journal

ofCe

llScience

Schmitt, S., Aftab, U., Jiang, C., Redenti, S., Klassen, H., Miljan, E., Sinden, J.and Young, M. (2009). Molecular characterization of human retinal progenitorcells. Invest. Ophthalmol. Vis. Sci. 50, 5901-5908.

Schwartz, S. D., Hubschman, J. P., Heilwell, G., Franco-Cardenas, V., Pan,C. K., Ostrick, R. M., Mickunas, E., Gay, R., Klimanskaya, I. and Lanza, R.(2012). Embryonic stem cell trials for macular degeneration: a preliminary report.Lancet 379, 713-720.

Schwartz, S. D., Regillo, C. D., Lam, B. L., Eliott, D., Rosenfeld, P. J., Gregori,N. Z., Hubschman, J. P., Davis, J. L., Heilwell, G., Spirn, M. et al. (2015).Human embryonic stem cell-derived retinal pigment epithelium in patients withage-relatedmacular degeneration and Stargardt’smacular dystrophy: follow-up oftwo open-label phase 1/2 studies. Lancet 385, 509-516.

Sowden, J. C. (2014). ESC-derived retinal pigmented epithelial cell transplants inpatients: so far, so good. Cell Stem Cell 15, 537-538.

Stanzel, B. V., Liu, Z., Somboonthanakij, S., Wongsawad, W., Brinken, R., Eter,N., Corneo, B., Holz, F. G., Temple, S., Stern, J. H. et al. (2014). HumanRPE stem cells grown into polarized RPE monolayers on a polyester matrixare maintained after grafting into rabbit subretinal space. Stem Cell Reports. 2,64-77.

Tian, C., Zhao, T., Zeng, Y. and Yin, Z. Q. (2011). Increased Muller cell de-differentiation after grafting of retinal stem cell in the sub-retinal space of RoyalCollege of Surgeons rats. Tissue Eng. Part A 17, 2523-2532.

Tzameret, A., Sher, I., Belkin, M., Treves, A. J., Meir, A., Nagler, A., Levkovitch-Verbin,H.,Barshack, I.,Rosner,M.andRotenstreich,Y. (2014). Transplantationofhuman bone marrow mesenchymal stem cells as a thin subretinal layer amelioratesretinal degeneration in a rat model of retinal dystrophy. Exp. Eye Res. 118, 135-144.

Wan, J., Zheng, H., Xiao, H. L., She, Z. J. and Zhou, G. M. (2007). Sonic hedgehogpromotes stem-cell potential of Muller glia in the mammalian retina. Biochem.Biophys. Res. Commun. 363, 347-354.

Wright, A. J. and Andrews, P. W. (2009). Surface marker antigens in thecharacterization of human embryonic stem cells. Stem Cell Res. 3, 3-11.

Xia, J., Liu, H., Fan, X., Hu, Y., Zhang, Y., Wang, Z., Zhou, X., Luo, M. and Gu, P.(2012). An in vitro comparison of two different subpopulations of retinal progenitorcells for self-renewal and multipotentiality. Brain Res. 1433, 38-46.

Xu, Y., Chen, C., Jin, N., Zhu, J., Kang, L., Zhou, T., Wang, J., Sheng, A., Shi, J.,Gu, Z. et al. (2013). Muller glia cells activation in rat retina after optic nerve injury:spatiotemporal correlation with transcription initiation factor IIb. J. Mol. Neurosci.51, 37-46.

2178

RESEARCH ARTICLE Journal of Cell Science (2015) 128, 2169-2178 doi:10.1242/jcs.169086

Journal

ofCe

llScience