Author contributions: D.E.B.: Conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript; S.T.H.: Conception and design, collection and/or assembly of data, data analysis and interpretation, final approval of manuscript; T.J.R.: Conception and design, collection and/or assembly of data, data analysis and interpretation, final approval of manuscript; A.M.F.: Collection and/or assembly of data, final approval of manuscript; C.R.H.: Collection and/or assembly of data, final approval of manuscript; L.V.J.: Conception and design, financial support, final approval of manuscript; D.O.C.: Conception and design, financial support, final approval of manuscript.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLS EXPRESS August 5, 2009.
Human induced pluripotent stem cells (iPSCs) have great promise for cellular therapy, but it is unclear if they have the same potential as human embryonic stem cells (hESCs) to differentiate into specialized cell types. Ocular cells such as the retinal pigmented epithelium (RPE) are of particular interest because they could be used to treat degenerative eye diseases, including age-related macular degeneration and retinitis pigmentosa. We show here that iPSCs generated using Oct4, Sox2, Nanog, and Lin28 can spontaneously differentiate into RPE cells, which can then be isolated and cultured to form highly differentiated RPE monolayers. RPE derived from iPSCs (iPS-RPE) were analyzed with respect to gene expression, protein expression, and rod outer segment phagocytosis, and compared with cultured fetal human RPE (fRPE) and RPE derived from hESCs (hESC-RPE). iPS-RPE expression of marker mRNAs was quantitatively similar to that of fRPE and hESC-RPE, and marker proteins were appropriately expressed and localized in polarized monolayers. Levels of rod outer segment phagocytosis by iPS-RPE, fRPE, and hESC-RPE were likewise similar and dependent on integrin αvβ5. This work shows that iPSCs can differentiate into functional RPE that are quantitatively similar to fRPE and hESC-RPE and further supports the finding that iPSCs are similar to hESCs in their differentiation potential. STEM CELLS 2009;27:2427–2434
The derivation of human induced pluripotent stem cells (iPSCs) has created the potential for patient-specific, immune-matched cells for regenerative medicine [1, 2]. Human iPSCs have been derived via expression of Oct4, Sox2, Nanog, and Lin28 by Yu et al.  and others and via expression of Oct4, Sox2, cMyc, and Klf4 by Takahashi et al.  as well as other investigators. Although iPSCs generated by both methods have been differentiated into derivatives of the three embryonic germ layers, and iPSC-derived cells have shown efficacy in animal models of disease [3–7], their differentiation abilities are relatively unexplored . Functional retinal pigmented epithelial (RPE) cells have been derived from human embryonic stem cells (hESC-RPE) and have been shown to rescue visual function in the dystrophic rat [9–11]. More recently, iPSCs reprogrammed using the Yamanaka factors have been shown to give rise to ocular cells, including RPE cells . These studies support the idea that stem cell–derived RPE are good candidates for the treatment of age-related macular degeneration (AMD) and other degenerative eye diseases [9, 10, 12–15].
AMD is the leading cause of blindness among people over 60 years of age . Presenting first with the formation of drusen (deposits forming between the RPE and Bruch's membrane), AMD progresses from the dysfunction and death of RPE cells to photoreceptor loss and deficits in high-acuity vision. The RPE plays many roles in visual function: absorption of stray light with pigment granules, formation of the blood–retina barrier with tight junctions, transport of nutrients and ions, secretion of growth factors and transport molecules, isomerization of retinol in the visual cycle, and phagocytosis of rod outer segments (ROS) . Antiangiogenic pharmaceuticals can slow the wet or exudative form of the disease, but this applies to only 5% of patients diagnosed with AMD. For the more common dry or nonexudative form of AMD, there is no effective treatment option to date . Autologous transplantation of RPE/choroid from the periphery of the eye to the macula has demonstrated the potential for RPE cell replacement as a therapy for AMD [18–24]. Use of iPSC-derived RPE (iPS-RPE) as a therapy would reduce surgical complexity and maintain the benefit of immune-matched autologous cells.
In this report, we show that iPSCs reprogrammed using the Thomson factors  spontaneously differentiate into RPE, and we compare these iPSC-RPE to hESC-RPE and fetal RPE (fRPE); we find a high degree of similarity in gene expression patterns and in protein distribution. Furthermore, we show that iPS-RPE are functionally similar to fRPE and hESC-RPE in the phagocytosis of ROS, a process critical for proper visual function. Our results demonstrate that iPSCs generated using the Thomson factors are similar to hESCs in their ability to differentiate into functional RPE, suggesting that iPS-RPE may also be beneficial in cellular therapy. This work was presented in abstract form at the 2008 annual Optical Society of America Vision Meeting .
MATERIALS AND METHODS
Pluripotent Stem Cell Culture
Induced pluripotent stem cell lines iPS(IMR90)-3, iPS(IMR90)-4, iPS(foreskin)-1, and iPS(foreskin)-2 (gift of J. Thomson, University of Wisconsin and UC Santa Barbara) were maintained in Dulbecco's modified Eagles medium (DMEM)/F12 containing 2 mM GlutaMAX-I, 10% knockout serum replacement, 0.1 mM MEM NEAA, 0.1 mM β-mercaptoethanol (Invitrogen, Carlsbad, CA, http://www.invitrogen.com), and 100 ng/ml recombinant zebrafish basic fibroblast growth factor (bFGF; gift of J. Thomson) on a mouse embryonic fibroblast feeder layer treated with mitomycin C (Sigma-Genosys, Cambridge, U.K., http://www. sigmaaldrich.com/Brands/Sigma_Genosys.html) or a matrigel-coated tissue culture plate (BD Biosciences, San Diego, http://www.bdbiosciences.com). The hESC line H9 (WiCell Research Institute, Madison, WI, http://www.wicell.org) was maintained in DMEM/F12 containing 2 mM GlutaMAX-I, 10% knockout serum replacement, 0.1 mM MEM NEAA, 0.1 mM β-mercaptoethanol (Invitrogen), and 4 ng/ml bFGF (Peprotech, Rocky Hill, NJ, http://www.peprotech.com) on a human foreskin fibroblast (Hs27, ATCC) feeder layer treated with mitomycin C (Sigma-Genosys).
Differentiation, Enrichment, and Culture of Pigmented Cells
Spontaneous differentiation of pluripotent stem cells was induced by removal of bFGF from the medium. The first signs of pigment were seen 20--35 days after removal of bFGF. These small spots of pigment grew into larger pigmented foci. After sufficiently large foci of pigmented cells had developed, they were mechanically dissected and dissociated into single cells using 0.05% Trypsin/EDTA (Invitrogen) (foci were large enough for dissection 60--90 days after removal of bFGF; see supplementary online information for exact lengths of time before enrichment). Cell suspensions were strained through a sterile 30-μm strainer cap (BD Falcon, Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com) to remove clumps and seeded at a minimum of 6.3 × 104 cell/cm2 onto tissue culture plates coated with 0.1% gelatin for 1 hour at room temperature. Cells were maintained in DMEM (high glucose) containing 7% knockout serum replacement, 0.1 mM MEM NEAA, 2 mM GlutaMAX I, 0.1 mM β-mercaptoethanol (Invitrogen), and 5% standard fetal bovine serum (HyClone, Logan, UT, http://www.hyclone.com). bFGF (10 ng/ml; Peprotech, Rocky Hill, NJ, http://www.peprotech.com) was included in the medium until cells reached confluence . Cells were subcultured monthly by dissociation with 0.05% trypsin/EDTA (Invitrogen) and replated at a minimum of 6.3 × 104 cells/cm2; 10 ng/ml bFGF (Peprotech) was included in the medium until the cultures reached confluence. Studies were performed on cells at passage 0 (enrichment), passage 1, or passage 2. Three separate enrichments of iPS-RPE were examined, one derived from iPS(IMR90)-3 and two from iPS(IMR90)-4. Seven separate enrichments of hESC-RPE were examined, all from H9. One additional enrichment of iPS-RPE from iPS(IMR90)-4 was performed after 8 months in bFGF-free culture, specifically to examine the protein expression of RPE65 after this length of time.
Fetal Human RPE Cell Culture
Fetal human RPE cells from a 21-week-old donor were a kind gift of D. Bok (University of California Los Angeles). Cells were cultured in fetal human RPE media  on 0.1% gelatin-coated tissue culture plates for 30 days prior to use; all fRPE analyses used fRPE at passage 0.
Quantitative Real-Time Polymerase Chain Reaction
Total RNA was isolated from iPS-RPE cell cultures derived from iPS(IMR90-3) (one replicate) and iPS(IMR90-4) (two replicates) at passage 0 (30 days after enrichment), three replicates of hESC-RPE from H9's at passage 1, and three replicates of fRPE using the Qiagen RNeasy Plus Mini Kit (Qiagen, Hilden, Germany, http://www1.qiagen.com). cDNA was synthesized from 1 μg of RNA using the iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA, http://www.bio-rad.com). Primer pairs were designed to create a 75–200 base pair product (Beacon Design 4.0; Premier Biosoft International, Palo Alto, CA, http://www.premierbiosoft. com). Quantitative real-time polymerase chain reaction (PCR) was carried out on a Bio-Rad MyIQ Single Color Real-Time PCR Detection System using the SYBR Green method . Triplicate 20-μl reactions were run in a 96-well plate with half of the cDNA synthesis reaction used per plate. Primer specificity was confirmed with melting temperature analysis, gel electrophoresis, and direct sequencing (Iowa State DNA Facility, Ames, IA). Data was normalized to the geometric mean of the “housekeeping” genes: glyceraldehyde phosphate dehydrogenase (GAPDH), peptidylprolyl isomerase A (PPIA), hydroxymethylbilane synthase (HMBS) and glucose phosphate isomerase (GPI) . Primer sequences are listed in supplemental online information.
iPS-RPE at passage 1 or passage 2 were grown on gelatin-coated chambered slides for 1 month. For fixation, slides were washed with phosphate-buffered saline (PBS), fixed with 4% paraformaldehyde in 0.1 M sodium cacodylate buffer (pH 7.4) for 15 minutes at 4°C, and stored in PBS at 4°C until labeling. Slides were washed with PBS, blocked with PBS containing 1% BSA, 0.1% NP40, and 1% normal goat or donkey serum in PBS for 1 hour at 4°C, and probed with anti--ZO-1 (Zymed Labs, Invitrogen), anti-Otx2 (10 μg/ml; R&D Systems Inc., Minneapolis, MN, http://www.rndsystems.com), anti-pigment epithelium-derived factor (PEDF) (10 μg/ml; Lifespan Biosciences, Inc., Seattle, WA, http://www.lsbio.com), anti-EMMPRIN (1:50; Zymed Labs, Invitrogen), or anti-αv integrin (1:100; Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com) for 1 hour (ZO-1) or overnight (all others) at 4°C. Slides were incubated with an appropriate Alexa Fluor (Invitrogen)–conjugated secondary antibody (1:200) for 30 minutes at 4°C, stained with Hoechst (2 μg/ml) (Invitrogen) for 5 minutes at room temperature, washed with PBS, and then imaged at room temperature using an Olympus BX51 fluorescence microscope or an Olympus Fluoview 500 confocal microscope (Olympus, Tokyo, http://www. olympus-global.com).
iPS-RPE protein cell lysates derived from iPS(IMR90)-4 at passage 0 (30 days in culture) (lysis buffer: 50 mM Tris-Cl (pH 7.6), 150 mM NaCl, 5 mM EDTA, 1% IGEPAL, 0.2% SDS, and EDTA-free protease inhibitor cocktail set V (Calbiochem, San Diego, http://www.emdbiosciences.com)) were quantified using a Micro BCA Protein Assay Kit (Pierce, Rockford, IL, http://www.piercenet.com). Total protein was separated via 8% SDS-PAGE, transferred to nitrocellulose membranes (Bio-Rad), blocked for 1 hour at room temperature in tris-buffered saline tween-20 containing 5% milk, then probed with anti-Mitf (1:100; Lab Vision, Fremont, CA, http://www.labvision.com), anti-bestrophin (1:1,000; Novus Biologicals, LLC, Littleton, CO, http://www.novusbio.com), anti-tyrosinase (1:200; Santa Cruz Biotechnologies), or anti-RPE65 (1:5,000; Chemicon, Temecula, CA, http://www.chemicon.com) for 1 hour at room temperature in block solution. To visualize labeling, blots were incubated with secondary anti-mouse Ig conjugated to HRP (1:10,000; GE Healthcare, Piscataway, NJ, http://www.gehealthcare.com), followed by antigen detection via chemiluminescence (Pierce) and exposure of the membranes to radiographic film. Lysates for RPE65 labeling were obtained from iPSCs allowed to differentiate for 8 months in iPSC media in the absence of bFGF. Pigmented and nonpigmented cells were present in these cultures, and pigmented cells were enriched just prior to protein extraction. Melanocyte protein lysate was collected from MeWo cells (HTB-65; American Type Culture Collection, Manassas, VA, http://www.atcc.org). Adult human RPE/choroid protein lysate was a kind gift from the Center for the Study of Macular Degeneration (University of California Santa Barbara).
ROS phagocytosis assays were performed as previously described . Bovine eyes were obtained fresh from a local slaughterhouse; ROS were purified from retinal extracts and fluorescently labeled using the FluoReporter FITC Protein Labeling Kit (Invitrogen). Cells were seeded in quadruplicate on gelatin-coated wells in a 96-well plate at a concentration of 25–50,000 cells per well and allowed to grow to confluence for 3–4 weeks. Cells were then challenged with 1 × 106 FITC-labeled ROS per well with or without 50 μg/ml anti-αvβ5 (ab24694; Abcam, Cambridge, U.K., http://www.abcam.com) or 50 μg/ml IgG1 control (ab9404; Abcam) for 5 hours at 37°C in 5% CO2. Wells were then vigorously washed five times with warm PBS to remove unbound ROS. Photomicrographs of total ROS uptake (ROS bound and internalized) were obtained using an Olympus IX70 inverted microscope equipped with a FITC excitation filter. To determine the level of ROS internalization, an equal volume of 0.4% trypan blue was added to the PBS for 10 minutes to quench extracellular fluorescence, followed by four gentle washes with PBS. The internalized ROS was then documented in fluorescence photomicrographs. Fluorescence intensity was quantified with pixel densitometry using ImageJ software (NIH, Bethesda, MD) for photomicrograph analysis. Photomicrographs from three wells for each condition were averaged within each assay. Assays were performed in triplicate on iPS-RPE, hESC-RPE (four replicates), fRPE, and human foreskin fibroblast (Hs27) cells. Separate experiments were normalized to the positive control ARPE-19 cell line, which was assayed in each experiment.
Differentiation and Enrichment of Putative RPE from iPSCs
To test the ability of iPSCs to differentiate into RPE, we removed bFGF from the medium and allowed the cells to overgrow and spontaneously differentiate. Using this technique, RPE cells are reproducibly generated from hESC cultures after 6–8 weeks in bFGF-free conditions . Depending on the iPSC line used, pigment onset typically occurred 20–35 days after removal of bFGF. The iPS(IMR90) lines, derived from fetal lung fibroblasts, usually pigmented 1–3 weeks earlier than the iPS(foreskin) lines, which were derived from postnatal foreskin fibroblasts. Over time, pigmented cell foci expanded and exhibited an epithelial phenotype (Fig. 1). Pigmented foci were large enough for mechanical dissection after 2–3 months of bFGF-free culture (see supplemental online information for exact timelines). Pigmented areas in iPS(foreskin) lines rarely expanded to a size sufficient for enrichment, so only iPS(IMR90) lines were used in these studies.
Upon single-cell dissociation, enriched cells seeded on gelatin-coated tissue culture plates lost their epithelial morphology and pigment until confluence was reestablished, consistent with observations of both hESC-derived RPE and primary RPE culture [9, 30]. In addition to reestablishment of morphology and pigment, confluent cultures began to form domes, indicative of fluid transport through an epithelial sheet of cells with tight junctions and seen with primary RPE culture . Cells could be serially passaged using trypsin for several passages (4–5), regaining epithelial morphology and pigment after each passage. After 4–5 passages, the ability of iPS-RPE to regain a pigmented epithelial phenotype declined and they became more fibroblastic in morphology. As discussed in Methods and Materials, characterization of iPS-RPE was performed on cells derived from iPS(IMR90), no later than passage 2. hESC-RPE for comparison were derived in the same manner, as previously reported .
iPS-RPE Express RPE Genes and Proteins and Are Polarized
Quantitative PCR was used to examine expression of genes involved in crucial RPE functions, as well as those important in both the parental iPS(IMR90-4) cell line and fRPE (Fig. 2).
We first examined transcription factors that reflect the states of pluripotency (Oct4, Nanog), early eye field development (Pax6, Rax, Six3), or differentiated RPE (Mitf, Otx2). iPS-RPE did not express transcripts for the pluripotency transcription factors Oct4 or Nanog, but low levels of Rax and Six3, transcription factors involved in early eye field development, were seen in iPS-RPE, hESC-RPE, and fRPE. Interestingly, Pax6 transcripts were seen in iPS-RPE and hESC-RPE but not in fRPE. Pax6 is expressed during RPE development but is turned off as the RPE matures, suggesting that iPS-RPE and hESC-RPE may be similar to immature RPE . The transcription factors Mitf and Otx2 are necessary for RPE differentiation  and are expressed at similar levels in iPS-RPE, hESC-RPE, and fRPE. Otx2 is also expressed in undifferentiated iPSCs, as has been found in hESCs .
The expression of genes involved in the many functions of differentiated RPE was also examined. Tyrosinase, Tyrp1, Tyrp2, and SILVER, genes involved in pigment synthesis , were expressed in iPS-RPE at levels equal to or higher than levels in hESC-RPE and fRPE. CRALBP and RPE65 are genes involved in the visual cycle and are expressed by terminally differentiated RPE . Transcripts for these genes were present at lower levels in iPS-RPE and hESC-RPE than fRPE, again suggesting that these cells may represent immature RPE. The genes encoding the tight junction proteins claudin and ZO-1 were expressed at similar levels in iPSCs, iPS-RPE, hESC-RPE, and fRPE. Transthyretin and PEDF are both secreted apically towards the neural retina by RPE cells, the former for binding of retinoids, the latter as an antiangiogenic neurotrophin [36, 37]. Transcripts for both these proteins showed high levels in iPS-RPE, hESC-RPE, and fRPE. mRNAs encoding bestrophin and EMMPRIN, proteins localized to the membrane of the RPE , were also expressed at similar levels in iPS-RPE, hESC-RPE, and fRPE. Transcripts between the three enriched iPS-RPE cell lines were similar but not identical (supplemental online information).
To confirm that iPS-RPE expressed RPE proteins along with their transcripts, immunoblotting and immunocytochemistry were used to examine expression and localization (Fig. 3). The tight junction protein ZO-1 was detected on the lateral margins of most cells, as expected (Fig. 3A). Expression was not uniform throughout the entire population, which was attributed in part to blocking of fluorescence by highly pigmented cells (Fig. 3A, arrowheads). EMMPRIN was diffusely expressed, consistent with membrane localization. PEDF staining was both diffuse and punctuate, as has been previously reported . Expected apical localization of the integrin αv subunit, relative to nuclear Otx2, is shown in Figure 3B. The transcription factor Mitf was detected in immunoblots of iPS-RPE as a doublet at 65–70 kDA (Fig. 3C), consistent with isoforms found in RPE . In contrast, MeWo cells (a melanocyte cell line) express shorter isoforms, with a doublet detected at 52–56 kDA (Fig. 3C). Bestrophin and tyrosinase proteins were both expressed by iPS-RPE. Although low levels of mRNA were present for the visual cycle gene RPE65 in 1-month-old cultures, protein expression could not be detected at this time point. However, RPE65 protein expression was observed in iPS-RPE when cells were cultured for 8 months in the absence of bFGF and without enrichment (Fig. 3C).
iPS-RPE Internalize ROS
Daily phagocytosis of shed photoreceptor outer segments is an essential function of the RPE . An ROS phagocytosis assay  was used to see if iPS-RPE could perform this function (Fig. 4). Isolated bovine ROS were fluorescently labeled with FITC and incubated with iPS-RPE, hESC-RPE, fRPE, or the human fibroblast cell line Hs27 for 5 hours. Free outer segments were washed away, and external fluorescence was quenched with trypan blue to visualize internalized outer segments. iPS-RPE phagocytosis of ROS was as efficient as hESC-RPE and fRPE and approximately three times more efficient than the negative control Hs27 cells.
RPE are known to utilize integrin αvβ5 in the initial binding of outer segments, prior to internalization [29, 41]. To study the role of integrin αvβ5 in the phagocytosis of ROS by iPS-RPE and hESC-RPE, we coincubated the cells and ROS with either a control IgG antibody or an inhibitory antibody (P1F6) against integrin αvβ5. Addition of the P1F6 antibody blocked ROS internalization by all cell types, while addition of the control antibody did not affect ROS internalization. This indicates that, like primary fRPE, the mechanism of iPS-RPE and hESC-RPE phagocytosis is integrin αvβ5–dependent. The efficiency of ROS internalization between the three enrichments of iPS-RPE cell lines was similar but not identical (supplemental online information).
We have derived RPE from induced pluripotent stem cells that are highly similar to primary fetal human RPE and RPE from human embryonic stem cells. These cells spontaneously arise from differentiating iPSCs generated with the Thomson factors  after 20–35 days, a timeline slightly faster than normal human RPE development in utero (approximately 40 days to pigmentation) and similar to spontaneous differentiation from human embryonic stem cells [9, 42]. Gross morphology of iPS-RPE is highly similar to that of primary RPE cultures as well as adult human RPE [43, 44].
iPS-RPE express RPE gene transcripts at levels similar to those of cultured fetal human RPE and hESC-RPE. Immunocytochemistry and immunoblotting show that RPE proteins are also expressed by iPS-RPE, in a polarized manner, as observed in normal RPE [43, 44]. Immunocytochemistry showed nonuniform labeling, likely due in part to fluorescence blocking by pigment granules, but also potentially due to heterogeneity in maturity states. Like hESC-RPE , iPS-RPE dedifferentiate and lose pigmentation upon passage and then regain maturity upon confluence.
It is interesting to note that protein expression of RPE65 is only found after long-term culture (8 months), although mRNA is consistently present under shorter culture conditions (30 days). RPE65 is an essential gene in visual function, mutations of which cause Leber's congenital amaurosis [45, 46]. Similar reports of low/absent RPE65 protein expression, despite mRNA expression in primary RPE cell culture, were made upon discovery of the gene, suggesting post-transcriptional regulation for the mRNA by surrounding ocular tissues [47, 48]. We observed long-term culture-dependent (8 month) RPE65 protein expression when iPS-RPE cells were not isolated from other differentiated cells. It is possible that neural retinal cells were in these cultures, regulating RPE65 protein expression. Length of time in culture also appears to have an important role in RPE65 expression as enriched cultures of hESC-RPE grown for greater that 7 months without passage had marked increases in RPE65 protein expression [Hikita et al., manuscript in preparation]. It is unclear whether the increase in RPE65 protein expression in our 8-month unenriched iPS-RPE culture was caused by signaling from other cells or by length of time in culture. It is also unclear whether RPE cultured for long time periods are functionally different from RPE from shorter cultures, aside from RPE65 protein expression. We expect that long-term RPE cultures in the presence of appropriate periocular tissue will be the most terminally differentiated. It will be interesting to see whether transplantation of iPS-RPE cells into animal models increases RPE65 protein expression. We believe this is likely the case, as we have seen that transplantation of iPS-RPE into the Royal College of Surgeon's (RCS) rat, a model of retinal dystrophy [46, 47], rescues visual function [Carr et al., manuscript in preparation]. Because RPE65 is required for visual function, it is likely that these cells produce the functional RPE65 protein in this model, although further testing is required to show this.
Daily phagocytosis of shed photoreceptor outer segments by the RPE is essential for the removal of proteins damaged by ultraviolet radiation and recycling of important nutrients . Loss of this RPE function by mutation of the gene for the internalization receptor MerTK is the cause of blindness in the RCS rat [49, 50]. iPS-RPE can perform this critical function in vitro as effectively as fetal human RPE, acting through the same integrin αvβ5--dependent mechanism. Importantly, phagocytosis activity in vivo has been demonstrated for hESC-RPE in the RCS rat model , and recently we have shown that iPS-RPE can also rescue vision in this animal [Carr et al., manuscript in preparation], demonstrating that phagocytosis of photoreceptor outer segments by iPS-RPE occurs in vivo.
The propensity of different hESC lines to differentiate into specific cell types has been shown to be dependent on the cell line [51, 52]. We have also seen a difference in the propensity to spontaneously give rise to RPE between different iPSC lines. Specifically, the iPS(foreskin) lines had a lower propensity to spontaneously differentiate into RPE than the iPS(IMR90) lines. Additionally, among the different enrichments of RPE from iPS(IMR90), different levels of transcript expression and levels of phagocytosis were seen. The cause for these differences between enrichments is unclear. They may be due to the different iPS(IMR90) clones used, the length of time prior to enrichment of RPE, or different non-RPE cell types that spontaneously arise within differentiating cultures.
For therapeutic use it will be important to establish specific protocols for derivation of RPE from stem cells that consistently give rise to cells of similar quality. In that light, it will be necessary to have a panel of assays to determine the quality of the cells derived. In addition to the assays used in the current study (gene/protein expression, quantitative ROS phagocytosis), other assays could include measurements of transepithelial resistance , enzyme-linked immunosorbent assays to detect polarized secretion of trophic factors (PEDF, vascular endothelial growth factor) , and a retinoid metabolism assay . Relating these assays to functional levels in animal models will give a robust screening method for each batch of stem cell RPE.
Multiple groups have derived RPE from hESC [9, 10, 12–15]. Data presented here shows that iPS-RPE are similar to previous reports of hESC-RPE in gene expression and function. We have found that hESC-RPE cell lines that we have derived are quantitatively similar to iPS-RPE with respect to mRNA levels of RPE markers and to phagocytosis ability. More recently, we have used microarray analysis to compare genome-wide transcript levels between iPS-RPE, hESC-RPE, and fRPE, and we find that stem cell--derived RPE share a common mRNA expression pattern (Radeke et al., manuscript in preparation). Both iPS-RPE and hESC-RPE, similar to primary cultures of RPE , lose their phenotype and senesce after repeated passaging. However, we estimate that a single six-well plate of differentiating stem cells can give rise to approximately 107 RPE cells after two passages. Because the macula is small, it would require RPE cells on the order of 105 to cover the entire area, so it is feasible to generate a quantity of cells needed for therapies. We are investigating several possible ways to increase the amount of RPE derived from a given enrichment: improved efficiency through directed differentiation, improved culture techniques to maintain RPE stability over multiple passages, or the use of a larger number of undifferentiated stem cells.
Although the iPSCs used in this study are not suitable for human trials because of transgene integration, nonintegrative iPS cells have recently been derived [56, 57]. Although generation of patient-specific iPS lines would be a costly and challenging paradigm for future health care, iPS-RPE may be superior to hESC-RPE because there would be less chance of immune rejection. One challenge associated with iPSCs derived from the patient is the possibility of reintroducing any genetic defects that contributed to the disease. Combining iPS technology with gene therapy is a possible solution . For AMD this could mean replacing a faulty Factor H haplotype, which has been associated with increased propensity to develop the disease . However, any alteration of the genome carries with it a bevy of risks that will have to be weighed against the need. In the case of AMD and other age-related degenerative diseases, gene therapy may not be necessary. For a disease that takes 50 years or more to progress, reintroduction of cells harboring a slow-acting genetic defect may be of no concern.
Degenerative diseases do carry their own challenges for cell replacement therapy, however, the central challenge being the window of intervention. Transplantation of cells has to be early enough in disease progression to maintain function in a dying tissue. Additionally, degenerative diseases may lead to a dysfunctional microenvironment such as drusen deposits and a dystrophic Bruch's membrane in AMD . We are investigating synthetic substrates that could be cotransplanted with stem cell-RPE to act as a surrogate Bruch's membrane.
While this manuscript was in preparation, a report of RPE derived from iPSCs was published . The iPSCs used in that report were produced using the Yamanaka transcription factors , whereas our iPSCs were derived using the Thomson transcription factors . Although the iPS-RPE from Yamanaka iPSCs have not yet been fully characterized, this indicates that iPS-RPE derivation from iPSCs is not dependent on the reprogramming mechanism.
Proper gene expression and cellular function of iPS-RPE suggests that the cells are a viable candidate for cellular therapy to treat degenerative eye diseases such as AMD and retinitis pigmentosa. It is still unclear which pluripotent cell type, embryonic or induced, will prove most effective for therapies. We show here that iPSCs generated by expression of Oct4, Sox2, Nanog, and Lin28 are similar to hESCs in their ability to produce functional RPE, and that iPS-RPE are remarkably similar to primary fetal human RPE. A quantitative in vitro phagocytosis assay shows that both iPS-RPE and hESC-RPE are highly similar to fRPE in the levels and mechanism of photoreceptor outer segment phagocytosis.
We thank James Thomson and Jessica Antosiewicz-Bourget for cells, reagents and advice; Dean Bok for cells; and Peter Coffey, Don Anderson, and Monte Radeke for helpful suggestions and advice. This work was supported by the California Institute for Regenerative Medicine grants T300009, C4-00521-1; Army Research Office; Millipore Corporation; Advanced Cell Technology; NIH 5R24EY014799-05; NIH NCRR Shared Instrumentation Grant 1S10RR017753-07.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.