Keywords

1 Introduction

The retina consists of diverse cell types stratified into organized functional tiers. Rod and cone photoreceptors and retinal pigment epithelium (RPE) cells work in conjunction to perform the energy-demanding and complicated biochemical process of converting light energy into electrical signals that ultimately are processed into “vision.” The neurosensory retina rests on a monolayer of RPE cells that associates with the photoreceptors on one side while partitioning the retina from the choriocapillaris on the other. The RPE performs multiple critical diverse functions essential for maintaining photoreceptor homeostasis (for review, see [1]). In fact, death or dysfunction of RPE cells can induce devastating secondary effects on photoreceptors characteristic of age-related macular degeneration (AMD) . AMD is the leading cause of vision loss in the elderly in industrialized countries [2, 3], and demographic analyses predict that it will become more widespread [4].

AMD is a multifactorial polygenic disease characterized by a broad spectrum of signs and symptoms (for review see [5]). However, RPE dysfunction and photoreceptor degeneration are shared characteristics that may be amenable to cell replacement strategies for a majority of AMD patients if delivery could be optimized and graft rejection prevented [6–9]. Compared with photoreceptor transplantation, RPE transplantation strategies are more simple since RPE cells do not need to integrate into neuronal networks that begin to degenerate and/or remodel during retinal degeneration (for review see [10]). Pluripotent stem cells may provide an excellent source of RPE, and banks of histocompatibility antigen-typed RPE derived from human embryonic stem cells (hESC-RPE) cells can be generated and used to intervene therapeutically after a patient is diagnosed with the disease.

Several NIH approved pluripotent embryonic stem cells (ESCs) are currently available, however their widespread use is still controversial due to ethical concerns. These concerns may be obviated due to a remarkable observation that the transgenic manipulation of only four transcription factors, OSKM: Oct4, Sox2, Klf4, or c-Myc [11], or OSNL: Oct4, Sox2, Nanog, and Lin28 [12], in somatic cells can “reprogram” them into induced pluripotent stem cells (iPSCs) from which autologous RPE grafts can be generated. These iPSC-RPE could be derived from individual patients and used for therapeutic transplantation since the progression of AMD is relatively slow. However, there are some prevailing safety and immunogenic concerns regarding the use of iPSCs that must first be resolved [13–20].

2 Stem Cell Derived RPE Cells Have Been Well Characterized

Pigmented RPE can be readily derived from hESCs [21–25] and hiPSCs [24, 26–30]. These cells spontaneously differentiate with appropriate morphologies and functionality, but may be more efficiently generated if exogenous factors are added to the differentiation media [21, 24, 29, 30]. hES– and hiPS–RPE cells have been well characterized. They express RPE-specific terminal differentiation markers [21–30] in polarized planes [25, 26, 29]. Using high resolution mass spectrometry-based metabolomic analyses as a measure of functional genomics, we have shown that iPS-RPE strongly resemble primary human RPE cells [29].

hES–RPE and hiPS–RPE also function as well as their human primary RPE counterparts in vitro. As observed in primary RPE [31], tightly coupled stem cell-derived RPE form fluid-filled domes, demonstrating that the ion pumps are vectorially functional [32]. Stem cell-derived RPE can also phagocytose photoreceptor outer segments in vitro [21, 25–27, 29, 33]. We have developed a flow cytometry-based method to measure the dynamics of RPE phagocytosis. Using this strategy we have shown that iPS-RPE synthesize phagocytosis receptors that phagocytose outer segments as effectively as primary human RPE do [34].

An important measure of RPE function is to determine if the cells operate in a diseased context in vivo. Studies in animal models have provided very encouraging evidence that stem cell-derived RPE transplantation can effectively promote photoreceptor cell anatomical and functional rescue in dystrophic retinas [21, 23, 25, 27, 29]. Additionally, the extremely preliminary results of a Phase 1/2 clinical trial using hES–RPE managed by Advanced Cell Technology Inc. (ACT) have shown that after four months the transplanted cells do not induce any obvious adverse side effects and may have integrated into the subretinal space of the treated patients [35].

3 Current and Future Prospects for RPE Cell Transplantation

In the majority of studies published to date (and in the ACT managed clinical trial) a bolus of RPE cells were injected into the subretinal space. With a few exceptions (including in our study in which we see a monolayer of RPE cells integrated in the subretinal space up to 17 months post injection [29]), transplanted RPE cells generally survive only for a few months after implantation and not integrate into monolayers [21, 27, 36]. For these reasons the use of alternative approaches are being advocated, including the use of intact RPE sheets or RPE cells grown on porous engineered Bruch’s membrane mimics [37, 38]. The surgical techniques required to deliver sheets of cells are inherently more complicated. Furthermore, it must be demonstrated that the scaffolds can stably support RPE cells over extended periods of time and allow adequate RPE and choriocapillaris crosstalk.

The use of iPS–RPE, and personalized medicine in general, could be very expensive. One alternative approach may be to bank reduced complexity human leukocyte antigen (HLA)-homozygous iPSC (or perhaps even hESC cell-lines). It has been estimated that approximately 75 and 140 unique donors would be needed to cover ~ 80 % and ~ 90 % of the Japanese population, and roughly 64,000–160,000 individuals would need to be typed to find the donors [39]. However, it is also important to consider that long-term immunosuppressive therapies are expensive, associated with complications, and not well tolerated by elderly patients [40, 41]. Therefore, the expenses for both techniques, and potential consequences of life-long immunosuppression, should be directly compared before any approach is deemed to be cost prohibitive.

While still extremely premature, an alternative therapeutic approach may involve the replacement of whole degenerated regions of human retinas with large patches of intact ocular tissues grown from stem cells in 3D cultures. The idea of transplanting stratified layers of functional neural networks may be more realistic based on two ground-breaking recent reports demonstrating that aggregates of mouse and, more recently, human ES cells grown in 3-D cultures self-assembled into structures strongly resembling optic cups (rudimentary sensory retinas) with neural and RPE domains [42, 43]. If this approach could be optimized, theoretically entire autologous maculas may be generated as therapeutic interventions for AMD .

4 Conclusions

RPE that function in vitro and in vivo to maintain photoreceptor homeostasis can be readily generated from stem cells. While few conclusions can be drawn until long-term studies in human patients have been completed, and RPE derivation and delivery techniques are optimized, the evidence collected in animal models that RPE grafts can prevent continued retinal degeneration and maintain visual function is very encouraging. Therefore, stem cell-based RPE transplantation therapies for untreatable retinal degenerative diseases such as AMD may ultimately prove to be not only realistic, but also therapeutically effective.