Promises of stem cell therapy for retinal degenerative diseases

Fetal stem cells are fetal retinal cells, at the exact developmental time when these cells are born and about to form intrinsic connections. Previous studies have shown that, before the formation of synaptic connections, retinal ganglion cells can regenerate after axotomy and navigate through the optic chiasm [8]. It has been proposed that immature photoreceptors might also have the capacity to reconnect themselves to the central neural system (CNS) after transplantation. Fetal retinal progenitor cells (RPCs) derived from a range of mammalian species, including rats [9, 10], pigs [11], and humans, [12] have been tried. It has been shown that rodent fetal RPCs are able to propagate extensively, expressing photoreceptor markers. Transplantation of fetal RPCs resulting in the survival and differentiation of the grafted tissue has been proven to be associated with behavioral benefits in retinal dystrophic recipients [13, 14]. Fetal neurons appear to show higher survival capacities than adult neurons [15]. For human fetal-derived retinal cells, Young [12] isolated proliferating human retinal progenitor cells (hRPCs) from 10th to 13rd week of gestation, and demonstrated that they could be expanded in tissue culture. However, their proliferating capacity was weak, and population declined quickly. Recently, Aftab et al. [15] have shown that donor tissues taken from 16th to 18th week of gestation give the longest in-vitro survival time, and the highest number of cells. After transplantation, these cells were integrated into the recipient retina, and differentiated into rhodopsin positive cells. This result supported the potential of hRPC transplantation for degenerative diseases. Nevertheless, ethical issues still exist, and the supply of such cells is still limited.

An alternative is to use embryonic stem cells and induced pluripotent stem cells (ESC/iPSC). ESCs/iPSCs have a great potential to differentiate into any of the 200 or more adult cell types. Hence ESC/iPSC provides potentially unlimited cell sources for the generation of retinal cells. In-vitro differentiation of ESC/iPSC into functional retinal cell types is achievable by defined step-wise protocols [16‐19]. ESCs could be induced to differentiate into eye-like structures that contained cells with properties of crystalline lens, neural retina, and RPE [20]. Further studies have indicated that cells from these eye-like structures could be further differentiated into RGCs when transplanted into the vitreous of an injured adult mouse retina [21]. Recently, the success of defined differentiation of human ESC-derived RPE cells (hESC-RPE) has been reported [22]. Following transplantation in animal models, restoration of vision had been reported and no tumor formation was seen [19, 23]. In 2010, the FDA approved the first clinical trial using hESC-RPE for the treatment of dry AMD and Stardgart’s disease (STGD) in humans. Hopefully results will be available in the near future. The main advantage over adult-derived RPE cell lines is the ability to produce differentiated RPE cells in vitro, which is less immunogenic.

Transplantation of hESC-derived RPE cells has proved to be a milestone in clinical therapeutics. Nevertheless, its use is still limited by ethical controversies and the risk of rejection. Induced pluripotent stem cells (iPSC) offer an alternative cellular source for patient-specific treatment without the risk of rejection and ethical problems [24, 25]. Nonetheless, clinical application of iPSCs is limited by the risks of proviral integrations and potential insertional mutagenesis during delivery of reprogrammed factors using virus. To overcome these issues, efforts toward the generation of “clinical grade” iPSC have been proposed. Recently, the reprogramming technologies in iPSC generation have been rapidly improved by the use of chemicals, plasmids, synthesized mRNAs, and direct protein delivery [26‐29]. In the future, transplantation of photoreceptors with or without RPE cells derived from these sources provides enormous potential for treating retinal degenerations. Personalized treatment strategy is potentially possible with the use of iPSCs, assuming that the risks associated are minimized.

It is known that lower vertebrates, such as teleosts or amphibians, have the ability to regenerate new retinal neurons throughout life, from a region called the ciliary marginal zone (CMZ) [30, 31]. It was also thought that the adult mammalian ciliary body (CB) might harbor retinal stem cell. In 2000, two independent groups discovered that the ciliary epithelium (CE) of the murine eye contains multipotent retinal stem cells [32, 33]. It was shown that single pigmented cells from the CE of mouse retina could clonally proliferate in vitro and form sphere colonies. These cells have the ability to be induced into retinal-specific cell types, including rod photoreceptors, bipolar neurons, and Müller glia.

Similar multipotent retinal stem cells were later identified in other mammalian species, including pigs and humans [34, 35]. These cells were proliferative, but to a lesser extent than fetal or ESC-derived retinal stem cells. When these cells were transplanted into adult mice, new photoreceptors were induced [35].

Another source of retinal stem cells was later discovered within the iris epithelium by Haruta and colleagues in 2001. The iris epithelium might harbor discrete heterogeneous populations of cells endowed with innate neural stem cell properties, including the ability to differentiate into retinal specific neurons [36, 37].

After the discovery of adult retina-specific stem cells, a number of laboratories have sought to expand numbers of such adult retina-specific stem cells and optimize sub-retinal differentiation. However, there are several obstacles to the use of such cells. Firstly, the percentage of actively proliferating cells in the CE is very few (<2%) [38]. Secondly, self-renewal and proliferation rates would decrease gradually with subsequent passages [35, 38]. Thirdly, there may be a risk of tumour formation, as reported by Djojosubroto [39] in a recent study. Furthermore, Gualdoni [40] found that the expansion of CE-derived cells quickly led to the loss of retinal progenitor cell markers and hence reduced the potential of photoreceptor differentiation. Further investigations are needed to delineate the intrinsic mechanisms governing adult stem cell self-renewal and differentiation, as well as genetic stability.

Other adult stem cells have also been reported to be capable of inducing retinal regeneration. These include neural progenitor cells (NPC) [41, 42], hematopoietic stem cells (hSC) [43], and mescenchymal stem cells (MSC) [44]. NPCs have been shown to promote the recovery from retinal injury and to express retinal phenotypic neurochemical markers [45]. However, reports have shown that NPC lacks the ability to differentiate into mature retinal neurons [46]. Furthermore, the shortage of adult NPC sources has further limited its application.

Autologous transplantation using hSC or MSCs has the advantage of reducing the risk of rejection and avoiding ethical controversies. Using retinal ischemia-reperfusion models, anatomical integration has been reported by intravitreal injection of hSCs [43, 47] and MSCs [48]. Animal studies have demonstrated that MSCs is capable of integrating into the ganglion cells and nerve fiber layers. Due to the fact that MSC derived from an elderly donor has limited functions, MSC derived from human embryonic stem cells or iPSCs serves as an alternative source [49, 50].

Functional retinal differentiation from hSCs or MSCs is still highly debatable. More evidence has suggested that improvements with the use of adult hSCs or MSCs may actually be attributed to the secreted neurotrophic factors and anti-inflammatory cytokines in situ, instead of direct functional retinal differentiation [44, 51].

Retina-specific cell types can be derived from various cell sources. Different cell sources and important growth factors and chemical modulators used to promote retinal cell differentiation are summarized in Table 1.

Cells of bone marrow origin have also been used for retinal regeneration. Bone marrow contains subsets of non-haematopoietic lineages, which are capable of multi-lineage differentiation into cells of non-haematopoietic capabilities. These include mesenchymal, endothelial, and very small embryonic/epiblast-like stem cells (VSEL). These cells proliferate and act in response to tissue injury or damage. Although most of these cells are organ-restricted, some appear to retain multipotential capacities [52]. In a mouse model, Li [53] showed that adult bone marrow-derived stem cells (BMSC) could be induced into RPE lineage in vitro. When infused back in vivo, these BMSC-derived RPE cells can home onto the focal areas of RPE damage, and form a monolayer on Bruch’s membrane.

Retinal pigment epithelium is essential for the maintenance of neural retinal function. There is evidence suggesting that retinal degeneration can be treated with subretinal injections of RPE cells. Transplantation of RPE was first reported in the late 1980s [54, 55], on dystrophic Royal College of Surgeons (RCS) rats. These rats have defective RPE cells, which eventually lead to photoreceptor cell death. Since then, significant progress has been made regarding autologous RPE transplantation. Recent advances in stem cell culture and differentiation techniques have made possible the generation of RPE cells from pluripotent stem cells. It was shown that using RPE cells derived from hESC lines effectively improves photoreceptor survival in RPE-defective RCS rats [56]. These RPE cells also prevented the onset of secondary degenerative events [57].

Since the first case report of human homologous and autologous RPE transplantation for the treatment of exudative AMD in 1991 [58], more than 30 homologous and 230 autologous RPE grafts have been performed [59]. The technique of harvesting RPE–choroid patch graft from the periphery, followed by insertion under the macula through retinotomy, was first described by van Meurs in 2003 [60]. It is now the most popular method of choice for RPE transplantation. Recent reports have shown that autologous RPE–choroid graft is able to produce sustainable long-term improvement in terms of vision and microperimetry performance, for neovascular age-related macular degeneration [61]. Other than autologous grafts, allogeneic and cultured HLA-typed cadaveric RPE cells have also been proposed for clinical application [62].

The mammalian retina contains highly specialized photoreceptors that are capable of capturing photons and transducing them into electrical signals. Replacing photoreceptors is much more challenging than RPE cells because the connection between photoreceptors and neurons are lost, and also because of the potential scarring response elicited. In 1999, Kwan [63] successfully transplanted photoreceptors in a mouse model of retinal degeneration. Outer segment reconstitution was observed in the recipient mice, and the mice were able to perform a simple light–dark discrimination test. Improvements in visual function have also been reported following transplantation of multipotent retinal progenitor cells derived from green fluorescent protein (GFP)-expressing mice [13].

Other transplantation strategies include grafting of retinal sheets, with or without attached RPE, into eyes of various rodent models of retinal degeneration [64]. Most recently, a few reports have demonstrated the ability of human neonatal photoreceptor precursors to differentiate and integrate into the outer nuclear layer (ONL) in both the intact and the degenerating retina of mature mice [14, 65]. Evidences has shown that photoreceptors and embryonic retinal tissue, when transplanted into the subretinal space, can form new synapses with existing host neurons. However, at the moment, photoreceptor transplants remain in the stage of laboratory science. The development of a combined tissue-engineered scaffold targeting both RPE and photoreceptors may be a promising direction for future research. Recent studies on stem cell therapy in retinal diseases are summarized in Table 2.

Integration of graft cells into host tissue is another major obstacle at the moment. Wang et al. [55] reported their results following injection of both RPE cells of cell line ARPE-19 and human Schwann cells (hSC). By 15 weeks after injection, only hSC was able to form a monolayer of cells at the level of RPE. Despite displaying diploid properties and expressing RPE-specific markers, ARPE-19 cells did not show properties of RPE cells when injected and failed to form a monolayer of cells [71]. Klimanskaya [72] also observed that ESC-derived RPE cells remained aggregated when grafted, and failed to form a monolayer over the defective RPE in the subretinal space.

The reason for the behavior of grafted cells in failing subretinal expansion may be due to the high levels of retinoids within the host RPE cells. Retinoid is known to be able to inhibit cellular division, which may have also inhibited the transplanted cells from integrating. To overcome this limited integration, transplanting RPE cell sheets instead of RPE cells has been proposed [73]. Concomitant pharmacological modulations of the extracellular matrix may also improve host integration [74]. As transplantation techniques advance, future studies may incorporate more adjuncts to improve the integration of grafted cells into the host.

Inflammation is activated in retinal degeneration primarily by microglia and macrophages [75, 76]. When a lesion is identified, these cells form a barrier around the lesion, separating damaged tissue from the undamaged [77]. Activated microglia and macrophages produce inflammatory cytokines including IL-1b, IL-6, tumour necrosis factor-alpha (TNF-a), nitric oxide and reactive oxygen species [75]. Inflammatory cytokines have been implicated in signalling migratory pathways for progenitor cells, in the hope of replacing the defective or dead cells [78]. However, excessive cytokines maybe toxic to grafted cells. They have to overcome the pro-inflammatory reaction as well as gliotic barriers, which may hinder its proliferation. Furthermore, microglial cells can serve as antigen-presenting cells, and the immune reaction produced may destroy allografts or xenografts. This poses a major obstacle to the long-term survival of allogeneic RPE or photoreceptor grafts. The eye and brain have been considered to be immune-privileged sites, partly because of the existence of the blood–retinal, blood–brain barrier. However, these barriers are often compromised in injured or diseased retina, and blood vessels may become ‘leaky’ [79]. At present, many transplanted photoreceptor progenitors die following allografts [14]. Graft rejection without immunosuppression may lead to graft rejection [80]. Therefore, immunosuppression may be required at least until the blood–retinal barrier has regained its function after surgery.

At the moment, no evidences of tumorigenesis have been reported when human ESC-derived retinal progenitor cells were used [23, 56, 68, 81, 82]. However, there has been one report of tumor formation when neurally selected mouse ESC was transplanted into rodent retina [83]. Therefore, it is vital that a stringent selection process to eliminate any undifferentiated ESC before applying into human trials.

Another potential risk for tumorigenicety is if the cell cultivation period has been prolonged. Djojosubroto [39] reported that chromosomal aberration accumulated rapidly upon prolonged cultivation in ciliary body-derived cells. This maybe potentially tumorigenic; hence, great care must be taken not to prolong the cultivation period.