Embryonic stem cells

ESC are derived from a transient population of cells in the early embryo and are characterized by pluripotency (i.e., the ability to differentiate into derivatives of the three primary germ cells layers: ectoderm, endoderm, and mesoderm) and the ability to self-renew indefinitely (Fig. 35.1). Murine ESC were first derived and cultured in 1981,^11,12 and have become important research tools for the creation of transgenic and knockout mice and for the study of early development in mammals (Table 35.1).¹³ In 1995, James Thomson’s team derived primate ESC from rhesus monkey blastocysts,¹⁴ but it was not until 1998 that human ESC (HESC) were derived and cultured from the inner cell mass of human blastocysts¹⁵ or from human primordial germ cells.¹⁶ Human blastocysts develop approximately 5 days postfertilization and the inner cell mass consists of 20–50 cells at days 5 and 6.¹⁷ In 2006, a method for culture of HESC from single blastomeres without destruction of the embryo was reported.¹⁸ Methods for the culture of murine and HESC typically utilized fibroblast feeder cells to facilitate the growth and survival of the stem cells but recently protocols have been developed for the feeder-free and serum-free culture.¹³

Fig. 35.1 Culture of human embryonic stem cells and their differentiation into endoderm, ectoderm, and mesoderm.

Table 35.1 Embryonic stem cell (ESC) time lines

iPSC, induced pluripotent stem cells; FDA, Food and Drug Administration.

The National Institutes for Health (NIH) have developed guidelines to establish policy and procedures under which NIH will fund research using HESC, and to help ensure that NIH-funded research in this area is ethically responsible, scientifically worthy, and conducted in accordance with applicable law. These guidelines were developed in response to Executive Order 13505, were issued on March 9, 2009, and became effective on July 7, 2009. As of November 2011, there were 136 HESC lines eligible for use in NIH-supported research, as listed in the NIH Human Embryonic Stem Cell Registry (http://stemcells.nih.gov/research/registry/).

Induced pluripotent stem cells

One of the most significant advances in the field of stem cell biology came with the observation by Yamanaka and colleagues that the addition of four critical transcription factors, under appropriate conditions, to adult somatic cells (e.g. skin keratinocytes or fibroblasts) would generate an induced pluripotent state from which a number of differentiated cells types could then be derived^18a. By generating such induced pluripotent stem cells (iPSC), it may now be possible to produce autologous grafts for use in treating a variety of diseases, including those of the retina. This is discussed in greater detail below.

Adult tissue stem cells

The concept that adult tissues contain pluripotent stem cells that can serve as a source of regenerative tissue useful in the repair of aging adult organ systems is an important scientific advance.¹⁹ For example, tissues traditionally thought of as not having regenerative capacity (e.g., brain and myocardium) are the ones most often touted as being the most likely to benefit from treatments with neural or myocardial “stem cells.” In fact, clinical trials using adult bone marrow-derived stem cells to regenerate infarcted myocardium have reported success in improving cardiac function,²⁰ presumably on the basis of bone marrow stem cells differentiating into myocardium. Unfortunately, the idea that adult bone marrow-derived hematopoietic stem cells (HSC) can transdifferentiate into myocardial tissue is not supported by experimental data.^21,22 While this does not invalidate the clinical observations, it does raise a cautionary note on how such results can be interpreted and emphasizes the need to establish rigorous standards by which such clinical studies are evaluated.²³

While there is an extensive literature on stem cells giving rise to nervous,²⁴ muscle,²⁵ vascular,^26,27 and hematopoietic tissue, work on retinal stem cells is more limited. Nonetheless, a literature has emerged over the past decade that strongly supports the potential for exploiting progenitor cells to maintain, and perhaps regenerate, abnormal retinal tissue. These studies describe four basic populations of cells that may contain dormant progenitor cells which, under appropriate circumstances, may have therapeutic application in the treatment of retinal disease: (1) retinal stem cells that can give rise to photoreceptors and other retinal neurons; (2) Müller/glial stem cells that can differentiate into retinal neurons; (3) retinal pigment epithelial (RPE) stem cells that can not only serve to replace diseased RPE but perhaps also be stimulated to differentiate into photoreceptors; and (4) endothelial progenitor cells (EPC) that can contribute to the retinal vasculature and exert a neurotrophic effect. Since there are a large number of reviews on retinal stem cells, this topic will not be discussed in great detail here. Glial and RPE stem cell biology are relatively new areas of investigation and, thus, there is not yet a lot of data on this potentially exciting, but understudied, area. Adult bone marrow-derived HSC containing EPC is an area of recent interest with broad-ranging potential application and will be discussed in slightly more detail given the therapeutic potential of these cells.

Differentiation of HESC into photoreceptors

Studies evaluating the in vivo efficacy of HESC-derived photoreceptors are at an early stage of development. Differentiation of HESC into retinal progenitor cells has been achieved, as has further in vitro differentiation into photoreceptor-like cells.⁵² As compared to RPE, photoreceptor transplants would require integration into the existing outer nuclear layer and formation of functional synapses to act as a replacement therapy. Conversely, there may be some potential therapeutic effect by providing neurotrophic support to adjacent cells without true synaptic integration. When HESC-derived retinal progenitor cells were cocultured with explants from retinal degeneration mice (Aipl1^–/–), they showed incorporation into the retinas, had morphologic characteristics of photoreceptor cells, and were immunoreactive for recoverin – a finding that was only rarely found when cells were cocultured with wild-type retinas.⁵² Subsequent experiments showed that when these cells were injected into the subretinal space of adult Crx^–/– mice (a model of Leber congenital amaurosis), the HESC-derived retinal cells differentiate into photoreceptor-like cells that were immunoreactive for recoverin and rhodopsin and restored light responses with an electroretinogram-like signal in the transplanted eye.⁵³ While the integration of HESC-derived photoreceptors may be a relatively rare event, these results demonstrate in principle that HESC can be used as a source of cells for photoreceptor replacement therapy.

Studies have shown that Noggin (inhibitor of BMP pathway) or Dickkopf (Dkk)-1 (an antagonist of wnt signaling pathway) promote anterior neural identity and then insulin-like growth factor (IGF)-1 promotes the formation of retinal progenitor cells.⁵⁴ Tom Reh’s lab reported derivation of retinal and photoreceptor-like cells from HESC-derived embryoid bodies grown in suspension culture, through addition of Noggin, Dkk-1, and IGF-1 into neuronal differentiation medium in adherent culture.^52,55 After 21 days of induction in the differentiation medium, they found the expression of retinal progenitor cell-related transcription factors such as Rx, Otx2, Pax6, Chx10, and Crx was significantly increased.⁵⁵ When further cultured, these cells formed neural rosettes that were picked and, allowed to self-aggregate, they formed rosettes with retinal progenitors in the center, differentiated cells showing photoreceptor markers such as recoverin, and RPE differentiation at the periphery.⁵⁵ Microarray analysis of HESC-derived retinal cells showed a very high correlation between genes expressed in human fetal retina and HESC-derived retinal cells.⁵⁶

Investigators from the RIKEN Research Institute reported the successful derivation of retinal progenitor cell and photoreceptors from either mouse or HESC using an even more highly defined stepwise method also using suspension culture followed by adherence culture.^57,58 The final steps included use of retinoic acid and taurine to induce photoreceptor differentiation. With this induction method, they found that HESC can be differentiated into photoreceptors showing both rhodopsin and recoverin immunoreactivity in 150 days.⁵⁸

Differentiation of HESC into three-dimensional embryonic and retinal tissues

Murine ESC can be differentiated into eye-like structures with cells that have properties of the crystalline lens, neural retina, and RPE.⁵⁹ Dissociated cells from these eye-like structures were shown to integrate into the retina, especially after retinal injury.⁶⁰ In April 2011, investigators reported the remarkable discovery that murine ESC aggregates grown in presence of added basement membrane components formed hollow spheres that underwent evagination into vesicles and then invagination into optic cup structures. The self-organizing optic cups had RPE and neural retinal domains and generated stratified neural retinal tissues⁶¹ (Fig. 35.2). Later in 2011, it was reported that three-dimensional populations of human retinal progenitor cells can be isolated from early forebrain cultures derived from HESC, and that a transient population of optic vesicle structures arose during the time appropriate for human retinogenesis. These optic vesicle structures could be differentiated into multiple retinal cell types, including RPE and photoreceptors.⁶²

Fig. 35.2 Murine embryonic stem cells can self-assemble into optic cup-like structures. These complex primitive retinal structures are engineered to express green fluorescence protein. (Reproduced with permission from The Scientist – Magazine of the Life Sciences. classic.the-scientist.com/news/display/58105/ © The Scientist.)

Clearly, there would be great therapeutic value in isolating retinal stem cells from the adult mammalian eye and then using them to regenerate diseased retinas. Recent reports suggest that, if done early enough, some level of visual function may be obtained in animal eyes with retinal degeneration treated with multipotent progenitor cells isolated from retinas of neonatal mice.^62a

While at the cellular level it may some day be possible to prepare biological retinal prostheses, there would remain extensive technical challenges in establishing functional neuronal connections between these implanted progenitor cells or sheets of photoreceptors and the host’s nerve fiber layer prior to re-establishing visual pathways that would lead to functional vision. If stem cell-derived retinas turn out to be more resistant to genetic and environmental damage than endogenous diseased tissue, it may be possible to do “prophylactic” transplants with these cells early enough in the disease process to insure that normal, functional visual pathways are established and maintained.

RPE cell-based delivery of trophic (and other) factors

Immortalized human RPE cell lines have been created to facilitate RPE cell transplantation as well as their use in cell-based therapeutic approaches. These cells enjoy an extended lifespan after being stably transfected with a plasmid encoding the simian virus 40 large T antigen and many of the factors expressed by functional RPE cells in vivo are observed to be expressed by these transformed cell lines.⁶⁷ When these cells are transplanted subretinally into a rat model of retinal degeneration (the Royal College of Surgeons (RCS) rat), loss of visual function is attenuated⁶⁸ and cortically dependent visual function is preserved long-term.⁶⁹ These RPE cell lines can be transfected with plasmids encoding a variety of trophic factors shown to have protective effects on photoreceptors^70,71 and then encapsulated into polymer devices that permit diffusion of cell products into the tissue into which they are transplanted. When transformed RPE cell lines are transfected with a plasmid encoding one such factor, ciliary neurotrophic factor (CNTF), and transplanted directly into the vitreous of dogs with retinal degeneration, photoreceptor degeneration is reduced.⁷² Furthermore, production of this factor and implantation of the encapsulation device into the vitreous of normal rabbits did not lead to toxic effects on either the electroretinogram or retinal histology in these animals at doses that protect photoreceptors in dogs with retinal degeneration.⁷³ Such cell-based delivery devices may be used to provide trophic factors for the treatment of a variety of retinal degenerative diseases and, in fact, have recently been used in a human clinical trial evaluating the efficacy of CNTF in the treatment of retinal degeneration.^73a In this respect, the implanted encapsulated cell devices function as a stem cell might, providing factors critical to the prevention of, or recovery from, retinal degenerative disease.

Embryonic stem cells as a source of RPE

Over the past decade, a number of methods have been reported to achieve the differentiation of HESC into retinal cells.^{52,58,74–76} Some of these methods were developed for differentiation of HESC into one type of retinal cell such as the RPE^58,74–76 or photoreceptors,^58,75,76 while other methods have focused on the efficient generation of retinal progenitor cells.⁵²

Differentiation of HESC into RPE

Spontaneous differentiation of HESC into RPE is the simplest and most commonly used method to produce RPE from HESC.⁷⁵ HESC colonies are first allowed to overgrow in growth factor-supplemented HESC medium until the borders of the colonies contact each other. The medium is then changed to basic HESC medium without basic fibroblast growth factor supplementation and is changed every other day for several months until the RPE cells are seen as small pigmented colonies in the culture dish (Fig. 35.2). RPE can also be derived from HESC using a two-stage induction method: HESC are first differentiated towards a neuroectodermal fate in suspension culture using neural differentiation medium, and then differentiated into RPE. The neural precursors are then plated in cell culture dishes for further differentiation into RPE. During these stages of induction and differentiation, the RPE cells appear as early as 4 weeks and reach a large enough number of cells for subculture at approximately 8–10 weeks.⁷⁵ Nicotinamide and Activin A (a member of the transforming growth factor-beta superfamily) can also be used to direct the induction of RPE from HESC.⁸³

RPE cells can be easily identified and isolated from other differentiated cells in HESC cultures because of their unique pigmentation, hexagonal shape, and pattern of growth. The RPE patches can be either mechanically incised out of the cultures or enzymatically dissociated from the cultures and can be grown to confluence and passaged retaining typical pigmentation and morphology (Fig. 35.3).

Fig. 35.3 Spontaneous differentiation of pigmented retinal pigment epithelium from human embryonic stem cells after withdrawal of basic fibroblast growth factor. In dish on left (A), multiple foci of pigmented cells are seen. These colonies can be picked, enriched, and expanded to pure populations of retinal pigment epithelium, as shown on right (B).

(A, courtesy of Dennis Clegg, PhD, University of California Santa Barbara.)

Characterization of HESC-derived RPE in vitro

The characteristics that define human RPE cells have been outlined in Chapter 16 (Cell biology of the retinal pigment epithelium) and reviewed by others.⁷⁴ HESC-derived RPE develop a typical hexagonal shape and become highly pigmented when they attain confluence (Fig. 35.3B). The cells can further differentiate and become highly polarized when grown for extended periods of time on Transwell inserts and other substrates (Fig. 35.4).^77,78 Polarized HESC-derived RPE show apical microvilli, are joined in the apical regions by tight junctions, show apically distributed sodium/potassium ATPase (Na⁺/K⁺ATPase), and have a high transepithelial resistance.⁷⁸ HESC-derived RPE express a host of RPE-specific and RPE highly expressed genes including visual cycle genes (RPE65, RDH 11, CRALBP); RPE membrane channel and transporter genes (BEST1, SLC); pigment biosynthesis and melanin biosynthesis genes (GPR143, TYRP1, dopachrome tautomerase (DCT), SILV); and phagocytosis-associated genes (LAMP2, VDP, Mertk, GULP1).⁷⁵

Fig. 35.4 Human embryonic stem cell-derived retinal pigment epithelium can be differentiated into a polarized monolayer. These monolayers show tight junction proteins (ZO-1), as shown by confocal microscopy (A), apical microvilli, as shown by scanning electron microscopy (B), and apical microvilli, apical melanosomes, and tight junctions, as shown by transmission electron microscopy (C, D).

Efficacy of HESC-derived RPE cells in vivo

The most commonly used model to evaluate therapeutic efficacy of HESC RPE is the RCS rat. The primary defect in the RCS rat is in the RPE; thus this model provides the ability to evaluate the effectiveness of RPE cell replacement therapy. The RCS rat has a recessively inherited mutation in the receptor tyrosine kinase gene Mertk, leading to impaired phagocytosis of shed photoreceptor outer segments with buildup of outer-segment material in the subretinal space, and subsequent secondary degeneration of photoreceptors between postnatal days 20 and 60.⁷⁹ Using subretinal injections of cell suspensions of HESC RPE, several groups have shown reproducible survival of the transplanted HESC RPE in the subretinal space (>220 days in one study), and positive labeling of these cells with human-specific markers and RPE-specific genes (e.g., RPE65).^80–83 The cells appear to disperse in the subretinal space and, while clumps of cells or multilayered grafts are often formed, in some cases they were arranged in an apparent monolayer. Cells showed focal rhodopsin staining, suggesting that they are phagocytosing photoreceptor outer segments.^81–83 The transplanted HESC RPE were associated with histologic and functional rescue of the photoreceptors, as measured by delay in the loss of nuclei in the outer nuclear layer, and retention of electroretinogram and optomotor responses in the treated eye compared to untreated eye.^80–83 Each of these studies utilized immune suppression (typically systemic cyclosporine, with or without systemic corticosteroid) to prevent immune rejection of the xenograft. Although the subretinal space is thought to be an immune-privileged site, such privilege may be compromised at the time of surgery or by the disease process.^84,85

Pivotal safety studies are necessary prior to receiving approval from the Food and Drug Administration (FDA) for a clinical trial. Studies using HESC-derived RPE cell suspensions under good manufacturing practices and good laboratory practice have been performed to establish karyotypic stability in the cells, lack of infectious and adventitious agents in the product, and lack of teratoma and/or tumor formation by the cells in immune-deficient mice.⁸¹ It should be noted that G-banding karyotypic analysis may not show any significant abnormalities while high-resolution DNA analysis may show culture-induced copy number changes and loss of heterozygosity; the significance of these changes for the purpose of cell therapy is unknown.¹⁸ Importantly, none of the studies that utilized highly differentiated HESC-derived RPE in immune-deficient animals found evidence of teratoma formation.^80–83 However, one study in which murine ESC differentiated toward RPE fate for only 7 days and injected in the subretinal space of rd12 (a spontaneous mutant model of Rpe65 Leber congenital amaurosis and having the same genetic background as the murine ESC) developed teratoma tumors in some animals.⁸⁶

An alternative approach to the use of HESC RPE cell suspensions would be to transplant monolayer sheets of highly differentiated and polarized RPE resting on a biodegradable or nonbiodegradable membrane. The reasoning for such an approach would be that these cells would have a better chance to integrate with the host photoreceptor outer segments as an intact monolayer and thus improve the functionality of the graft.⁷⁷ It has been shown that HESC RPE can be polarized in vitro to develop tight junctions with high transepithelial resistance and elaborate extensive apical microvilli (Fig. 35.4).⁷⁸ It has also been observed that highly polarized HESC RPE show increased secretion of pigment epithelial-derived growth factor, a factor with neurotrophic and antiangiogenic activity.⁷⁸ HESC RPE are capable of specifically phagocytosing photoreceptor outer segments.^86a Polarized HESC RPE show increased phagocytosis of bovine rod outer segments in vitro compared to nonpolarized cultures.⁷⁸ Studies implanting polarized HESC RPE grown on a nonbiodegradable substrate (e.g., parylene) show retention of an intact monolayer in vivo and prominent integration with host photoreceptors (Fig. 35.5).⁷⁷ The relative efficacy of HESC-derived RPE cell suspensions versus polarized sheets is under active investigation by several groups.

Fig. 35.5 Human embryonic stem cell (HESC)-derived retinal pigment epithelium (RPE) can be grown to confluence and polarized on a nonbiodegradable substrate. In (A), the pigment cells form an intact monolayer. The cells with substrate can be implanted in the subretinal space of Royal College of Surgeons rats where they survive and protect host photoreceptors from degeneration. In (B) (hematoxylin and eosin stain), the HESC-derived RPE sitting on the nonbiodegradable membrane integrate into the host retina. Note the interface between the transplanted RPE and the host photoreceptor outer segments (arrows). In (C), the transplanted HESC-derived RPE can be identified using the human-specific marker TRA-1–85 (immunofluorescence microscopy with human marker TRA-1–85 in green, and nuclear counterstain 6’-diamidino-2-pheninlindole hydrochloride (DAPI) in blue).

The effect of the local tissue microenvironment on transplant survival is an important issue when considering how best to deliver stem cells and cells derived from stem cells. Cross-talk between cells and extracellular matrix is critical to maintaining the differentiated cell type as well as insuring appropriate function. Many studies, in the retina as well as other tissues, have identified a number of factors that may be critical to the successful engraftment of stem cell-derived, and other, cell types, including cardiomyocytes,⁸⁷ spinal cord,⁸⁸ and brain⁸⁹ neurons and photoreceptors.⁹⁰

The use of induced pluripotent stem cells as a source of autologous RPE (and other cell type) grafts

One of the most significant recent breakthroughs in stem cell research has been the observation that adult human somatic tissues may be induced to express pluripotency after reactivation of four transcription factors.⁹¹ These adult tissue-derived stem cells, called induced pluripotent stem cells (iPSC), can be differentiated into a variety of tissues that may be used as autologous grafts for therapeutic reimplantation in various diseases.⁹² The use of iPSC circumvents most of the ethical and practical problems associated with large-scale clinical use of ESC, and patients with iPSC-derived autologous transplants may not require lifelong immunosuppressive treatment to prevent graft rejection. However, safety concerns regarding iPSC include malignant transformation resulting from reactivation of the four reprogramming transcription factors (including c-MYC) randomly integrated into the genome at multiple loci following retroviral transduction. In fact, iPSC generation and transformation share many molecular mechanisms, and a high incidence of tumorigenesis is observed in iPSC-derived chimeric mice.^93–96 The risk of tumorigenesis cannot be attributed solely to c-MYC reactivation: transgenic mice derived from iPS without c-MYC still form tumors.⁹⁷

Exhaustive comparative analyses of multiple human (h) iPSC and HESC lines reveal that, while many hiPSC and HESC lines share very similar transcriptomic and epigenetic profiles, others are heterogeneous. The differences observed are randomly distributed and limit the differentiation capacity of the cells.⁹⁸ Furthermore, recent evidence shows that, in hiPSC, reprogramming and selection pressure to obtain rapidly proliferating cell-lines may induce chromosomal aneuploidy in nonrandomly distributed loci that may further limit the differentiation capacity and promote tumorigenicity of iPSCs.^99–101 Genetic instability in iPSC is correlated with higher passage numbers, so reprogramming methods that are inefficient and require multiple passages may increase the risk of tumorigenesis.^99,100 Furthermore, high cytotoxicity is observed when reprogramming methods that utilize repeated transfections (reprogramming methods involving mRNA strands or recombinant proteins). Consequently, these methods could induce even higher selection pressures for highly proliferative cells with mutations; however this idea has not yet been directly tested.

To generate RPE from hiPSC for treating diseases such as AMD in which the onset occurs in 55+-year-old humans, the source material (likely fibroblasts or keratinocytes) will be older and more difficult to reprogram. The RPE derivation process is long and all clones will need to be carefully screened before transplantation. Consequently, this treatment strategy could be very economically taxing to patients and healthcare systems. Therefore, to generate any tissue of interest, including RPE from hiPSC, one or two specific reprogramming protocols may need to be adopted and optimized to insure reliable and safe derivations of that cell type. Furthermore, experimental animals with implants should be monitored over extended periods of time to ensure that no tumors form. We have recently reported the generation of iPSCs reprogrammed using two or one factors (OCT4/KLF4 or OCT4 only) and small molecules.^102,103 Recent advances in stem cell biology have demonstrated that iPSC can be derived from somatic cells such as skin fibroblasts by using retroviral vectors to introduce reprogramming factors (Oct4, Sox2, Nanog, Lin28) into somatic cells. However, random viral integration and spontaneous reactivation of reprogramming factors during differentiation pose a cancer risk and limit the potential use of iPSC derived in this manner in human therapy. In addition to this risk, therapeutic application of current technology is also limited by its low efficiency and slow kinetics. Finally, there is the requirement for numerous growth factors and animal cell feeder layers to obtain even limited numbers of iPSC. To overcome these hurdles, increase the efficiency of generating iPSC-derived RPE, and avoid using lentiviral vectors, it may be possible to use an episomal, or nonintegrating, vector to deliver the required transcription factors or eliminate the need to deliver such factors by using small molecules to induce pluripotency.

Potential problems associated with the use of iPSC to generate RPE grafts

Current human iPSC technology requires retroviral vectors for delivering reprogramming factors into somatic cells.^{90,91,104–107} Random viral integration and spontaneous reactivation of reprogramming factors pose cancer risk and impair differentiation, limiting the potential use of iPSC in human therapy.¹⁰⁸ While recent reports have shown that adenoviral¹⁰⁹ and plasmid¹¹⁰ vectors can reprogram mouse embryonic fibroblasts (MEFs) at very low efficiency, there is no evidence that these methods can be successfully applied to reprogram human somatic cells. To overcome this bottleneck, episomal expression vectors that encode the four human reprogramming factors (Oct4, Sox2, Nanog, Lin28) have been developed.^111–113 Transfection of MEFs and human fibroblasts with this vector produced iPSC in which the exogenous factors are eliminated by LoxP/Cre-mediated deletion.¹¹³

The recent development of the iPSC reprogramming technology allows, for the first time, the generation of autologous cellular central nervous system grafts from readily accessible cell sources such as skin biopsies.^90,91 AMD may represent an ideal target for autologous iPSC RPE transplantation therapy as RPE cell dysfunction in this disease is caused by cellular aging processes rather than a monogenic defect. While several genetic risk alleles for AMD have been identified,¹¹⁴ manifestation age of AMD even in the presence of these risk alleles is above 55 years. Hence, iPSC RPE cell grafts from AMD patients would be expected to represent younger, and presumably healthier, RPE that have not yet themselves become damaged by the aging processes. Furthermore, RPE cell grafts, unlike neuronal cell grafts, would not require synaptic integration into the retinal neuronal network to resume rescue activity and can be delivered locally into the subretinal space, thus avoiding the need for graft cell migration to their final localization.

Recent studies using transcriptomic analyses demonstrated significant differences between gene transcription in HESC/hiPSC-derived RPE and hRPE.¹¹⁵ While useful for monitoring the differentiation status of cells, variations in gene expression at the transcriptional level do not necessarily provide insight into cellular activities. Functional consequences of transcriptional activity can be analyzed with novel metabolomic-based analyses that quantitatively measure the activities of endogenous biochemical pathways.^116–118 Application of sophisticated in vivo imaging techniques such as scanning laser ophthalmoscopy, optical coherence tomography, and adaptive optics, coupled with focal electroretinography, will permit the detailed evaluation of therapeutic stem cell treatments.

Synthetic small molecules to enhance iPSC and RPE production

Identification and optimization of chemical tools for HESC iPSC derivation and long-term self-renewal, as well as their differentiation to a lineage-specific functional cell type (e.g., RPE) in chemically defined conditions will provide higher consistency/robustness in HESC iPSC culture, facilitate the derivation and manufacturing of clinical-grade human stem cells for therapy, speed up understanding of the self-renewal and differentiation mechanisms, and help identify and control signaling inputs that direct self-renewal or differentiation of HESCs/iPSCs. The discovery and mechanistic characterization of synthetic small molecules that support self-renewal of ESCs in the absence of feeder cells, serum products, and conventional growth factors^119,120 have been described. From reprogramming screens, small molecules that can substitute for reprogramming transcription factors and enhance reprogramming efficiency and kinetics in the generation of iPSCs have also been described.¹²¹ These data demonstrate proof of concept that rationally designed, unbiased screens may provide new tools and reveal novel mechanisms and that a single small molecule can be identified from a functional screen to achieve a desired complex phenotype by modulating one or multiple targets. While it has been demonstrated that a single small molecule can have significant positive effects on HESC self-renewal/survival, it is more likely that a combination of several small molecules (each with different mechanisms of action) will generate a more robust/efficient condition.

Recently, four, two, and one factor-derived iPSC (4F-, 2F-, and 1F-iPSCs, respectively) were differentiated into fully functional cuboidal-shaped pigmented cells in polarized monolayers that express RPE-specific markers.¹²² 1F-iPS RPE strongly resemble primary human fetal RPE (hfRPE) based on proteomic and untargeted metabolomic analyses, and, utilizing novel in vivo imaging technology coupled with electroretinography, it has been demonstrated that 1F-iPS RPE mediate anatomical and functional rescue of photoreceptors after transplantation in an animal model of RPE-mediated retinal degeneration (Fig. 35.6). 1F-iPS RPE cells were injected subretinally as a suspension and formed a monolayer dispersed between host RPE cells. Furthermore, 1F-iPS-RPE do not simply provide trophic support to rescue photoreceptors as previously speculated, but actually phagocytose photoreceptor outer segments in vivo and restore visual cycling (based on high-resolution mass spectrometry-based detection of recycled photoreceptor protein and lipid endproducts and electron microscopic analysis). Thus, 1F-iPS-RPE grafts may be superior to conventional iPS-RPE for clinical use since 1F-iPS-RPE closely resemble hfRPE, mediate anatomical and functional photoreceptor rescue in vivo, and are generated using a reduced number of potentially oncogenic reprogramming factors.

Fig. 35.6 One-factor-derived induced pluripotent stem cell retinal pigment epithelium (1F-iPS-RPE) strongly resembles human fetal RPE (hfRPE) and mediates functional and anatomical rescue of photoreceptors after transplantation in Royal College of Surgeons (RCS) rats. (A) 10 000+ unique metabolic features were detected using mass spectrometry for several different cell types. The percentage of dysregulation (fold change ≥2, P ≤ 0.001) between these cell types was plotted. (B) Polarized 1F-iPS-RPE cells labeled with anti-pigment epithelial-derived growth factor antibodies (red) and 6’-diamidino-2-pheninlindole hydrochloride (DAPI: blue). Scale bar = 50 µm. (C) Polarized 1F-iPS-RPE cells are present and correctly oriented in the subretinal space 7 months after transplantation. (D) Focal electroretinogram measurements were taken from the same eye over a region where cells were not implanted (uninjected region) and over a region where 1F-iPS-RPE was implanted 5 weeks after transplantation. The data were plotted as percent amplitude (µv) above sham-injected levels. (E and F) Photoreceptors are rescued in RCS eyes injected with 1F-iPS-RPE. (E) Few photoreceptors can be detected in control RCS eyes based on Recoverin (green) and (DAPI: blue) staining. (F) In contralateral eyes injected with 1F-iPS-RPE cells, however, 11 ± 1.1 rows of photoreceptors are preserved. (E and F) Scale bars = 100 µm.

(Data and images courtesy of Peter Westenskow, PhD and Martin Friedlander, MD, PhD, The Scripps Research Institute, adapted from reference 122).

Bone marrow-derived EPC can contribute to retinal and choroidal neovascularization

The observation that HSCs contain a pool of EPC capable of incorporating into retinal vasculature has recently been demonstrated by several groups. The study by Grant and colleagues was the first direct demonstration that systemically administered HSC can function as hemangioblasts during hypoxia-stimulated retinal neovascularization.¹³⁴ In these studies, single HSCs isolated from bone marrow of adult transgenic mice expressing green fluorescent protein (GFP) were injected intravenously into mice that had been sublethally irradiated to destroy host bone marrow. Serial long-term transplants were performed to be certain that the reconstituted bone marrow arose from the injected stem cells. After recovering from the radiation treatment and demonstrating that engraftment was established, the retina in these mice was treated with thermal laser to stimulate retinal neovascularization and it was demonstrated that the engrafted GFP bone marrow-derived stem cells contributed to the neovascularization. This model demonstrated that circulating, undifferentiated precursor cells can be recruited to sites of retinal neovascularization and, along with proliferation of local endothelial cells, can contribute to new blood vessel growth and development. The relative contribution of circulating precursor cells and endogenous retinal vascular endothelial cells to newly forming vasculature in human disease remains unknown; the experiments of Grant and colleagues¹³¹ demonstrate that circulating cells can incorporate into laser-stimulated retinal neovascularization, but the role of these cells in nonirradiated hosts where the proliferation of local inflammatory, precursor, and endothelial cells is not impaired by lethal irradiation remains unclear. Studies from several groups have demonstrated, using the same irradiation/bone marrow reconstitution model, that circulating stem cells can also contribute to laser-stimulated choroidal neovascularization.^135–137

If, indeed, circulating EPC contribute to pathological neovascularization in ischemic and inflammatory retinopathies such as diabetic retinopathy and AMD, would inhibition of their targeting to these sites reduce abnormal angiogenesis? In order to evaluate this potential therapeutic intervention better, it is necessary to know what factors are responsible for appropriate targeting of EPC. One study¹³⁸ has demonstrated a role for the adhesion molecule, R-cadherin, in the targeting of HSC to the retinal vasculature. When small-molecule antagonists or function-blocking antibodies to R-cadherin are used to pretreat Lin^– HSC prior to intravitreal injection, the cells no longer target sites of angiogenesis and participate in the formation of new retinal blood vessels. Other studies suggest that certain adhesion molecules, such as the integrin alpha 4 beta 1, may play a role in targeting of circulating EPC to sites of abnormal angiogenesis during vascularization of tumors and these same integrins may be potential therapeutic targets if, indeed, circulating EPC contribute to pathological ocular angiogenesis. Interfering with the function of such targeting molecules used by EPC to target sites of pathological neovascularization in combination with cell-based therapies to produce angiostatic molecules locally could significantly reduce abnormal angiogenesis. Unfortunately, inhibition of neovascularization under such ischemic conditions may serve to promote ongoing ischemia: would it be better to coax the newly forming vessels into functional ones that could alleviate hypoxia or make the endogenous vasculature and neurons more resistant to hypoxic damage?

Bone marrow-derived EPC can exert a vasculotrophic rescue

In order to address the potential utility of these cells in relieving hypoxia and help rebuild a damaged retinal vasculature, Lin^– HSCs have been injected directly into the eyes of newborn mice while they were forming their retinal vasculature (Fig. 35.7); in this environment, these cells can target activated astrocytes, a hallmark of many ocular vascular and degenerative diseases. Once targeted to this template of activated astrocytes, the Lin^– HSCs participate in normal developmental angiogenesis in both neonatal mice and injury-induced neovascularization in the adult.¹³⁹

Fig. 35.7 Bone marrow-derived endothelial progenitor cell target sites of retinal gliosis and incorporate into developing blood vessels to form mosaic vessels. When these cells are taken from the bone marrow of adult mice transgenic for enhanced green fluorescent protein (eGFP, yellow) and injected into mice transgenic for GFP-glial fibrillary acidic protein (a marker of astrocytes, green), the stem cells selectively target to the underlying astrocytes (A). This also happens when these cells are injected into adult mice (B), in which a needle or laser is used to scar the retina and stimulate a focal gliosis (red), suggesting that these cells (green) may also be useful to treat injured adult retinas. Two weeks after injection of adult bone marrow-derived stem cells into the neonatal mouse retina (C), mosaic blood vessels (orange-yellow) consisting of stem cells and endogenous retinal vascular endothelial cells are observed along with vessels consisting only of endogenous vessels (red).

(Adapted with permission from Otani A, Kinder K, Ewalt K, et al. Bone marrow-derived stem cells target retinal astrocytes and can promote or inhibit retinal angiogenesis. Nat Med 2002;8:1004–10.)

Is there a population of adult bone marrow-derived progenitor cells that can be used to target, incorporate into, and “stabilize” a degenerative, abnormal retinal vasculature? Studies¹³⁹ suggest that this, in fact, may be possible. The HSC fraction used in these studies not only inhibited angiogenesis when engineered to express an antiangiogenic, but also rescued and stabilized (e.g., “matures”) degenerating vessels. More surprisingly, it was also observed that by preventing vascular degeneration there is a trophic rescue effect on the photoreceptors themselves,¹²⁴ suggesting that autologous bone marrow grafts of HSC fractions containing EPC may provide trophic effects on associated neural tissue that goes beyond simple nutrition. Such observations could provide a rationale for the use of HSC in the treatment of a variety of inherited retinal degenerations such as retinitis pigmentosa.

The potential use of HSC for cell-based therapies in the eye presents a technical challenge; stem cells are notoriously difficult to transfect and, thus, it will be necessary to improve transfection efficiency. Another enigma in the circulating stem cell field is the issue of HSC “homing”; R-cadherin is clearly involved, but all of the molecular signals have not yet been identified. Identification of these signals would be of immense benefit in terms of exploiting the potential use of HSC in therapeutic angiogenesis as well as directed cell therapy. Finally, which cell type in adult bone marrow actually adheres to astrocytes and incorporates into the developing vasculature? Many of the surface markers routinely used to identify cell populations vary from laboratory to laboratory and, certainly, expression of these markers depends on the environment into which the cells are placed.¹⁴⁰ Thus, sorting cells based on cell surface antigen expression too early in their lineage may lead to cases of mistaken identity or loss of potentially useful cells. Functional definition of stem cell populations, as done in the Otani¹³⁹ study, surely provides an assessable endpoint but warrants further investigation. Finally, the use of stem cells as drug delivery vehicles has the potential to deliver drugs selectively and potently to the back of the eye in physiologically meaningful doses. Since this is where most vision-threatening pathologies occur, the use of these cells, as trophic cellular devices or drug delivery vehicles, holds great promise for millions of patients suffering from currently untreatable diseases.

Recent reports in the literature demonstrate that Lin^– HSC contain a population of EPC that can promote angiogenesis by targeting reactive astrocytes and incorporate into an established template without disrupting retinal structure. Thus, genetically modified, autologous EPC transplanted into ischemic or abnormally vascularized eyes may stably incorporate into new vessels and continuously deliver therapeutic molecules locally for prolonged periods of time. Such local delivery of pharmacological agents in physiologically meaningful doses may represent a new paradigm for treating currently untreatable ocular diseases.

Bone marrow-derived EPC can exert a neurotrophic rescue in retinal degeneration

Inherited degenerations of the retina affect as many as 1 in 3500 individuals and are characterized by progressive night blindness, visual field loss, optic nerve atrophy, arteriolar attenuation, altered vascular permeability, and central loss of vision, often progressing to complete blindness. Molecular genetic analysis of these diseases has identified mutations in over 110 different genes, accounting for only a relatively small percentage of the known affected individuals^141,142; many of these mutations are associated with enzymatic and structural components of the phototransduction machinery, including rhodopsin,¹⁴³ cGMP phosphodiesterase,¹⁴⁴ rds peripherin,¹⁴⁵ and RPE65.¹⁴⁶ Despite these observations, there are still no effective treatments to slow or reverse the progression of these diseases. Recent advances in gene therapy have led to successful reversal of the rds¹⁴⁷ and rd¹⁴⁸ phenotypes in mice and the RPE65 phenotype in dogs¹⁴⁹ when the wild-type transgene is delivered to photoreceptors or the RPE in animals with a specific mutation. The potential use of calcium channel blockers,¹⁵⁰ trophic factors,¹⁵¹ and dietary supplements¹⁵² has also been explored in recent studies. Most inherited human retinal degenerations specifically affect rod photoreceptors but there is concomitant loss of cones, the principal cellular component of the macula. Cone-specific survival factors have been described¹⁵³ and may facilitate cone survival in mouse models of retinal degeneration.

In addition to the vasculotrophic properties described above, these cells have also recently been reported to prevent completely retinal vascular degeneration, ordinarily observed in mouse models of retinal degeneration, and that the vascular rescue correlates with neuronal rescue (Fig. 35.8). The inner nuclear layer remains nearly normal and the outer nuclear layer containing photoreceptors is significantly preserved, with the rescued cells being predominantly cones. Detectable, albeit severely abnormal, electroretinogram recordings are observed in rescued mice at times when they are never observed in control-treated, or untreated, rd/rd eyes. This rescue effect is also observed when human bone marrow-derived Lin^– HSCs are used to treat severe combined immunodeficient mice with retinal degeneration. Large-scale genomic analysis of rescued and nonrescued eyes revealed significant upregulation of antiapoptotic genes. It is important to note that the injected bone marrow-derived progenitor cells are never observed anywhere but in or near blood vessels; since these cells are derived from GFP transgenic mice in which all the cells are green, it is easy to determine which structures in the retina are derived from the injected progenitor cells. Green donor cells are never observed in the nuclear layers or the rescued retinas. These findings demonstrate that this newly described neurotrophic effect correlates with preservation of the vasculature and suggest that autologous bone marrow-derived EPCs may be useful in the treatment of currently untreatable retinal degenerative diseases as well as ones in which abnormal angiogenesis cause the vision loss.

Fig. 35.8 Adult bone marrow-derived stem cells rescue degenerating blood vessels and retinal function, and exert a profound neurotrophic rescue effect in a mouse model of retinal degeneration. (A and B) Representative cases of rescued and nonrescued retinas 2 months after injection. Retinal section of stem cell-injected right eye (A) and control cell-injected left eye (B) of the same animal are shown (green: CD31-stained vasculature, red: 6’-diamidino-2-pheninlindole hydrochloride (DAPI)-stained nuclei). (C and D) Electroretinographic recordings were used to measure the function of stem cell- or control cell-injected retinas from the same eyes shown in A and B. (E–H) Rescued outer nuclear layer (ONL) in a mouse model of retinal degeneration (rd1/rd1) following intravitreal injection of adult bone marrow-derived stem cells consists predominantly of cones. Control (CD31 hematopoietic stem cell-injected) eyes are identical to noninjected rd1/rd1 retinas, without any staining for cone (E) or rod (G) opsin. Stem cell-treated contralateral eyes in the same animals have a markedly reduced, but clearly present, ONL that is predominantly comprised of cones, as evidenced by positive immunoreactivity for cone red/green opsin (F). A small number of rods are also observed (H).

(Adapted with permission from Otani A, Kinder K, Ewalt K, et al. Bone marrow-derived stem cells target retinal astrocytes and can promote or inhibit retinal angiogenesis. Nat Med 2002;8:1004–10 and Otani A, Dorrell MI, Kinder K, et al. Rescue of retinal degeneration by intravitreally injected adult bone marrow-derived lineage negative hematopoietic stem cells. J Clin Invest 2004;114:765–74.)

The concept that tissue-specific vasculature has trophic effects that go beyond that expected from simply providing vascular “nourishment” is supported in numerous recent reports in the literature. For example, liver endothelial cells can be induced to produce, after vascular endothelial cell growth factor receptor-1 activation, growth factors critical to hepatocyte regeneration and maintenance in the face of hepatic injury.¹²³ Similar interactions between vascular endothelial cells and adjacent hepatic parenchymal cells are necessary for liver organogenesis, well before the formation of functional blood vessels.¹²⁵ In individuals with retinal degeneration, the presence of endothelial progenitors derived from bone marrow HSC populations may make the vasculature more resistant to degeneration and at the same time facilitate retinal neuronal survival. In humans with this disease, slowing the rate of degeneration even to a limited extent may provide years of additional sight. While the animal studies discussed above had significant, but not dramatic, preservation of electroretinogram signals, even such minimal signs of electrical function in the retina may be sufficient to support vision.

Clinically, it is widely appreciated that there may be substantial loss of photoreceptors and other neurons while still preserving functional vision. At some point, the critical threshold is crossed and vision is lost. Since nearly all of the human inherited retinal degenerations are of early, but slow, onset, it may be possible to identify individuals with retinal degeneration, treat them intravitreally with an autologous bone marrow stem cell graft, and delay retinal degeneration with concomitant loss of vision. This would be particularly effective if selective rescue of cones occurs. Using such Lin^– HSC induction of antiapoptotic factor, synthesis occurs in a paracrine fashion; it may be possible to apply this approach to the treatment of other visual neuronal degenerative disorders such as glaucoma in which there is retinal ganglion cell degeneration.

Adult bone marrow-derived stem cells may have wide utility in the treatment of retinal vascular diseases and even inherited retinal degenerations. While the use of these cells to target neovasculature and contribute to the stabilization of otherwise friable vessels in ischemic retinopathies may seem intuitive, the neurotrophic effect observed is surprising but reasonable given the “protective umbrella” effect afforded by these heat shock protein-laden cells. Thus, potential applications of these cells include not only as cell-based therapeutic delivery vehicles, but also as possible stabilizing elements in an otherwise unstable neovasculature of the type observed in ischemic retinopathies. In fact, these cells will selectively target sites of hypoxia-driven neovascularization, as demonstrated in mouse models of these diseases (Fig. 35.9). While current retinal vascular disease therapy is based largely on ablating the new vessels with angiostatics or thermal and nonthermal lasers, a novel paradigm would include the use of these stem cells to target, incorporate into, and stabilize neovasculature.

Fig. 35.9 Adult bone marrow-derived stem cells target sites of ischemia in a mouse model of retinopathy of prematurity. Stem cells taken from the bone marrow of adult enhanced green fluorescent protein transgenic mice are injected into mice placed into high oxygen between day 7 and day 12 after birth. Shortly after injection of the stem cells on postnatal day 6, the green stem cells are observed to align with the red remodeling vasculature. If these cells can stabilize the vasculature and make it resistant to hypoxic damage, this may have utility as a novel paradigm for “rescuing” retinal vessels in ischemic retinopathies such as retinopathy of prematurity and diabetes.

(Courtesy of Eyal Banin, David Friedlander, and Martin Friedlander, Scripps Research Institute.)