In this study, we successfully differentiated RPE-like cells from several human pluripotent stem cell lines without the use of animal cells or serum during the differentiation. In addition, we reported RPE differentiation in xeno-free and defined culture conditions. Putative RPE cells arise spontaneously from undifferentiated cells by the removal of bFGF from the culture conditions. The hESC lines used in this study have a similar background; all the cell lines were cultured on hFF feeder cells and in KO-SR containing culture medium without serum. The hiPSC line, FiPS 5–7, was cultured on top of hFF feeder cells from passage 26 onwards. Other published RPE differentiation methods have harnessed mainly hESC and hiPSC lines derived and cultured on MEFs or with FBS (). Only very recently, Idelson and coworkers published a protocol differentiating hESCs toward RPE cells from hESCs grown on hFF feeder cells and with KO-SR containing culture medium [16
Our results demonstrate that the appearance of the first pigmented cells was relatively fast after the removal of bFGF, both in hESC and hiPSC lines, varying from 10 to 21 days. Most of the published data describe the appearance of the first pigmented cells usually around 2 to 8 weeks [1
]. Thus, according to their pigmentation, hESCs and hiPSCs cultured on hFF seem to have equally good differentiation rate to those of other published cell lines cultured on MEFs. We analyzed the percentage of the pigment containing cell clusters to indicate the differentiation rate of each cell line on D21–D28 after initiation of the differentiation. Although calculations only give a rough estimate of the pigmentation rate, the results demonstrated differences between analyzed cell lines. Idelson and coworkers published comparable results with an equal amount of pigmented cell clusters (<15%) after four weeks of differentiation without NIC supplementation [16
]. For comparison, Klimanskaya and coworkers reported the presence of pigmented islands in less than 1% of hESC-derived cell aggregates after 4–8 weeks of differentiation [1
During the time course of differentiation, we already detected the expression of early eye lineage markers PAX6 and RAX, as well as early RPE marker MITF at D7 after removal of bFGF from all analyzed hESC lines. Although the expression of PAX6 was also detected during a later stage of the differentiation, our data demonstrated that the mature RPE cell specific markers, RPE65 and bestrophin, were detected from all but one of the hESC lines on D28 at the latest. In addition, other RPE cell markers—OTX2v1, PMEL, PEDF, and tyrosinase—were also expressed early in the differentiation process of hESC lines, with tyrosinase having the latest appearance of the studied genes. Surprisingly, the expression of RAX was not detected from hiPSC-RPE (FiPS 5–7) until D56. On the other hand, RPE cell specific markers RPE65, bestrophin, OTX2v1 PMEL, PEDF, and tyrosinase were expressed in hiPSC-RPE derived cells similarly as to hESC-RPE cells.
According to the pigmentation rate, we chose the most promising hESC line, Regea 08/023, and hiPSC line, FiPS 5–7, for more detailed evaluation. Manually selected and long-term cultured cells formed a monolayer of pigmented cobblestone-like cells that had a very similar morphology to previously published hESC-RPE and hiPSC-RPE cells [1
]. After dissociation into single cells, the differentiated RPE-like cells started to proliferate and were observed to undergo morphological changes that are typical in EMT. EMT is related to normal development and tissue repair, but also pathological processes such as cancer and proliferative vitreoretinopathy. The EMT process has been recently described in isolated RPE cells [35
]; thus it is a natural feature when RPE cells lose cell-cell contacts. Our results indicate that putative hESC-RPE and hiPSC-RPE cells also seem to undergo a similar process.
The gene expression analyses of hESC-RPE and hiPSC-RPE cells revealed that the cells expressed eye precursor genes RAX
, and RPE cell markers. The expression of RAX
was not constantly detected from hiPSC-RPE (FiPS 5–7) until D56 with RT–PCR. However, more quantitative qRT–PCR analysis showed that RAX
expression was similar to that of hESC-RPE cells. Both cell lines also expressed the marker of neural crest-derived melanocytes (SOX10
), indicating the presence of other ectoderm derivatives. SOX10
expression by selected hESC and hiPSC derived RPE-like cells is most probably the expression of other types of neural cells, which easily contaminate pluripotent human cell cultures [36
]. Selected and long-term cultured hiPSC-RPE cells expressed mesodermal marker and pluripotent marker OCT3/4
and nanog, indicating that the cell population contained undifferentiated cells. We showed that neither hESC-RPE nor hiPSC-RPE were homogenous, even after 196 days in culture, according to gene expression, although morphologically the cells seemed to be of uniform quality. Thus, the manual selection, which we used in this study, is not sufficient to gain a pure population of putative RPE cells. Consequently, it is essential to develop more specific differentiation methods and more efficient purification and selection methods for these cells.
Microscopy of cells differentiated for 83 days showed that the cells were highly organized and pigmented. The immunocytochemical localization of proteins essential for mature and functional RPE cells was identical to previously described results [16
]. Of note, the cells expressed RPE65, which is essential for the regeneration of the visual pigment required for both rod- and cone-mediated vision [38
]. The expression of the RPE65 protein is typically lost in cultured RPE cells [39
], and has been reported in only a few papers describing the differentiation of RPE cells from human ESCs [16
]. The expression of RPE65 protein in a culture environment by human iPSC-RPE has been previously reported only by western blotting and was detected as late as after eight months of differentiation [32
]. In the present study, both hESC-RPE and hiPSC-RPE cells showed RPE65 immunostaining at D83, suggesting that the differentiated cells are functional and closely resemble the native RPE cells.
One of the most important functions of the RPE cells is the phagocytosis of POS. In vivo, photoreceptor cells undergo a daily renewal process and RPE cells take care of the waste disposal by phagocytosing nonfunctional POS. Several groups have published the phagocytotic activities of hESC-RPE and hiPSC-RPE in vitro using latex beads [1
] or POS isolated from animals [32
]. It has been further suggested that the only proof for the specific phagocytosis activity of RPE cells is the capability for POS phagocytosis [42
]. The hESC-RPE and hiPSC-RPE cells generated in our study possess the relevant molecular functions required for the phagocytosis of isolated POS, thus demonstrating their functionality in vitro.
In addition, the functionality of hESC-RPE and hiPSC-RPE cells was shown by PEDF secretion, with both differentiation methods described here. PEDF secreted by RPE cells is antiangiogenic and neuroprotective, protecting retinal neurons from light damage, oxidative stress, and glutamate excitotoxicity [43
]. According to our knowledge, the secretion of PEDF from putative hESC-RPE or hiPSC-RPE cells has not been previously published anyone other than Klimanskaya and coworkers [1
]. Our data demonstrated that the differentiated pigmented cells were able to secrete high levels of PEDF into the culture medium, while nonpigmented cells did not secrete any detectable amounts of PEDF. There is no published data about the amount of PEDF secreted by putative hESC/hiPSC-RPE [1
]. Tong and coworkers have reported that the ARPE19 line secretes PEDF at around the 5–35 ng/ml level to the culture medium [44
], and Maminishkis and coworkers have stated that isolated human fetal RPE cells secrete PEDF at the 600 ng/ml level [45
]. These levels have not been standardized to the cell number included in the experiment; thus, direct comparison of our results is not possible. However, the hESC-RPE and hiPSC-RPE cells in our study secrete reasonable levels (40 ng/ml- 430 ng/ml) of PEDF, demonstrating their functionality in relation to growth factor production in vitro.
Versatile determination of the maturation and polarization capacity of the derived cells in a specific culture protocol is important [32
]. TEER is one of the assessments that can be used for this [46
]. TEER values have been regularly assessed from human retinal explants [47
] and primary and immortalized RPE cell lines reaching 206 Ωcm2
and 100 Ωcm2
, respectively [48
]. To our knowledge, the TEER of putative hESC-RPE or hiPSC-RPE cells has not yet been evaluated. In our culture, the putative hESC-RPE cells reached TEER values of 310 Ωcm2
after 268 days of differentiation and the putative hiPSC-RPE cells reached TEER values of 74 Ωcm2
after 287 days of differentiation. When the TEER values of hESC/hiPSC-RPE are compared to values derived from other RPE cell cultures, it appears that putative hESC-RPE cells are at least as good or even superior to immortalized RPE cell lines in forming a highly polarized monolayer. Further indications of a high degree of cellular polarization of differentiated cells are the separation of Na+
ATPase to the apical and bestrophin to the basal side of the cellular monolayer, and junctional localization of apical tight junction protein ZO-1.
We further demonstrated that xeno-free and defined differentiation conditions can be used for the induction and maturation of pluripotent stem cell derived RPE cells. With this method, we were able to differentiate RPE-like cells from hESCs and hiPSCs. The differentiated cells showed similar gene and protein expression profiles, as well as cellular polarization, to the RPE cells differentiated using RPEbasic conditions containing xenogeneic substances. Our results demonstrated that the appearance of the pigmented cells was slightly faster in RPEregES conditions as compared to the cells in RPEbasic conditions. In addition, the percentage of the pigmented clusters after 23 days of differentiation of both the hESC and hiPSC lines were higher in RPEregES conditions, with the hiPSC line having over 10% difference between the two conditions. These results suggested that our defined and xeno-free culture medium described previously [27
] may even enhance the differentiation of RPE cells. This may be partly explained by the fact that the medium contains Activin A (5 ng/ml), a known inducer of RPE cell fate [26
]. Our defined and xeno-free differentiation method is one step forward when optimizing proper culture conditions for clinically eligible RPE cells for transplantation. It remains to be studied whether supplements such as NIC [16
] and Wnt and Nodal antagonists [21
] increase RPE differentiation from hESCs and hiPSCs in our culture systems.
In conclusion, we have demonstrated a progressive differentiation protocol for the production of functional RPE-like cells from hESCs and hiPSCs. Our results demonstrate that putative hESC-RPE and hiPSC-RPE express genes and proteins characteristic for RPE cells, can phagocytose POS, are able to secrete PEDF and form highly polarized and tight epithelial structures in vitro. Our results show that highly mature RPE-like cells can be differentiated in xeno-free and defined culture systems that are easier to translate under the Good Manufacturing Practice production systems needed for clinical use. Furthermore, defined conditions will greatly elucidate the further development of more efficient differentiation protocols and the use of cells in drug screening and toxicology studies.