Clinical background
Advances in the treatment of diseases involving the ocular anterior segment, particularly the lens and cornea, have greatly decreased the prevalence of visual impairment caused by dysfunction of these structures. Unfortunately, treatment of diseases impacting structures of the posterior segment, particularly the retina and optic nerve, have not advanced to the same extent and as a consequence these conditions are now the major source of incurable blindness in the developed world ( Box 77.1 ). The reasons for this are not hard to discern in that both the retina and optic nerve are components of the central nervous system (CNS) and it has long been appreciated that the mammalian CNS exhibits a very restricted capacity for endogenous regeneration. Furthermore, what little capacity exists diminishes further with postnatal maturation. Additional barriers to the development of effective restorative treatments for the retina and its central projections are numerous and include the complex phenotypes of retinal cells, particularly photoreceptors and ganglion cells, as well as the need for an unusually precise cytoarchitecture. This is true both in terms of outer segment packing and the topographic organization of output fibers projecting to the visual centers of the brain.
- •
Degenerative diseases impacting structures of the posterior segment of the eye, particularly the retina and optic nerve, are currently the major source of incurable blindness in the developed world
- •
There is an urgent need to reconstruct, via cellular replacement, the damaged or lost layers of the retina
As a consequence of the many challenges involved, there are at present no restorative treatments for retinal cell loss. This situation may change; however, a growing body of experimental data suggests that many of the barriers to retinal repair in mammals are not insurmountable. In particular, stem cell transplantation and tissue engineering have recently emerged as promising strategies for repopulating the cellular constituents of the retina. Alternatively, it is conceivable that an electronic prosthesis could bypass the damaged components of the retina and convey visual information directly to downstream visual neurons, although this fascinating strategy will need to overcome difficulties faced when attempting to use artificial constructs to stimulate high-resolution visual acuity. For now, it appears that the most straightforward method of restoring retinal function resulting from cell loss is to replace those cells through transplantation.
Etiology
Work in animal models and limited studies in human subjects have shown that a number of different tissues and cell types can survive as allografts in the ocular posterior segment, in either the vitreous cavity or subretinal space. The general facility of graft survival seen likely results in part from the degree of immune privilege afforded in these locations. Vitreal delivery may be adequate for some applications and cell types; however, subretinal placement is preferable for restoration of photoreceptors and the retinal pigment epithelium (RPE), as is needed in retinitis pigmentosa (RP) and to varying extents in retinal detachment and age-related macular degeneration. In terms of relevant cell types for outer retinal repopulation, both photoreceptors and RPE cells survive transplantation beneath the retina; however, the reluctance of donor photoreceptors to make functional connections with the surviving host circuitry and failure of donor RPE cells to reform a polarized monolayer on Bruch’s membrane have frustrated attempts to achieve functional repair of the outer retina using grafts of freshly isolated cells alone. Here we will focus on the challenges facing photoreceptor replacement and consider the advantages of using cultured allogeneic retinal progenitor cells (RPCs) as donor material.
To repopulate the outer nuclear layer with functional photoreceptors, there is a fundamental problem to overcome. This is the physical barrier to neurite outgrowth presented by hypertrophy of the outer limiting membrane (OLM) following photoreceptor loss. The OLM is not in fact a membrane per se but rather is an emergent structural element formed by the joined outer ends of retinal Müller cells. In the setting of photoreceptor degeneration, the OLM undergoes thickening in association with upregulation of the markers neurocan and CD44. Regenerating neurites originating from either above or below have great difficulty crossing this barrier. In fact, the phenomena of glial hypertrophy and scar formation are common in the setting of CNS disease and injury and have frequently been implicated in the failure of endogenous regenerative mechanisms to bridge a lesion.
We have shown that grafted CNS progenitor cells are not impeded by a hypertrophied OLM and can migrate across this barrier in large numbers. The ability to migrate into the mature, diseased retina is one remarkable characteristic of CNS progenitor cells that recommends them as a potential tool for use in retinal repopulation. Moreover, these cells not only migrate into the retina, but also exhibit widespread integration into the local cytoarchitecture, with tropism for regions of injury or disease. In the case of RPCs, there is also the potential to differentiate into cells with morphological features and marker expression characteristic of photoreceptors.
Retinal transplantation
To achieve functional repair in patients afflicted with retinal degenerative disorders, there is an urgent need to reconstruct, via cellular replacement, the damaged or lost layers of the retina. While restorative repair of the retina is a daunting challenge, a range of data suggests that such a goal is now feasible. Two approaches for targeted repopulation of the retina are considered below and include stem/progenitor cell transplantation and retinal sheet transplantation, both of which pose a variety of advantages and disadvantages ( Box 77.2 ).
- •
Two promising approaches for targeted repopulation of the retina following injury/disease include transplantation of stem/progenitor cells and transplantation of intact retinal sheets
Stem/progenitor cell transplantation
Over the past decade, stem/progenitor cell transplantation as a means of inducing tissue reconstruction and functional regeneration has garnered extensive interest in the field of regenerative medicine. Within the retina in particular, many exciting advances have been made. One significant achievement came in 2004 when we were able to show that a subset of transplanted RPCs developed into a variety of mature retinal neurons, including retinal ganglion and photoreceptor cells. Since then, numerous studies reporting varying degrees of success have utilized an assortment of different cell types ranging from the fate-restricted photoreceptor precursor to the pluripotent embryonic stem (ES) cell. ES cells are of particular interest due to their ability to undergo unlimited expansion and subsequent tissue-specific differentiation. These inherent properties may allow one to generate a sufficiently large number of cells in order to perform clinical transplantation from single isolations rather than requiring multiple new donations, as is potentially the problem when using more terminally differentiated cell types. However, as cited above, these cells are pluripotent, meaning they can be induced to generate cell types for each of the three germ layers and as such, are not retina-specific. Thus, protocols for retina/cell type-specific differentiation are required. In light of this, many labs have been aggressively searching for the proper method of retinal cell induction. One of the first published reports of RPC, and subsequent retinal cell generation, from human ES cells came from Lamba and collegues in 2006. In this publication, the group was able to show that they could reliably produce healthy functioning photoreceptors, albeit at low levels, using a relatively simple induction protocol. Since then, similar studies using variations of Reh’s methods have reported an increased generation of retina-specific cell types, photoreceptor cells in particular, in a range of organisms including primate and human.
Retinal sheet transplantation
Like stem cells, retinal sheets, including full-thickness and photoreceptor only, have also been used in an attempt to achieve retinal reconstruction and visual restoration following injury and disease. The goal of these techniques is to deliver retinal tissue with proper laminar structure and cellular organization directly to the site of injury in an attempt to form new functional connections between remaining host tissue and healthy donor material. Although extensive connections and subsequent visual restoration have not yet been achieved by using this technique, significant progress, particularly when using developing tissue, has been made. For instance, when embryonic rat retina was transplanted into the subretinal space of degenerating animals, the tissue was shown to survive without immune rejection, develop normally, continue to respond to light for up to 3 months after transplantation, and form limited functional connections with the host. Apart from functional integration, similar results have also been reported in a pig model of RP. For example, Ghosh et al have shown that transplanted retinal tissue isolated from healthy fetal donors can survive and maintain proper laminar structure and cellular organization within the subretinal space of the host for up 6 months post-transplantation. However, an issue that remains when using full-thickness retinal grafts is the introduction of redundancy into the system. For instance, in diseases such as RP where there is a progressive loss of photoreceptor cells with sparing of the inner retinal circuitry (albeit temporarily), full-thickness transplants would result in replication of a majority of the retinal cell types. Thus, in a situation such as this, transplantation and subsequent integration of photoreceptor sheets, void of all other retinal structures, would be beneficial. Although extensive research has yet to be carried out using this approach, studies with promising findings have been reported. For instance, as with full-thickness retinal grafts, transplantation of photoreceptor sheets has enjoyed prolonged survival and structural preservation.
As promising as the abovementioned studies may be, issues such as inadequate cellular/axonal integration following stem/progenitor cell and retinal sheet transplantation remain. Modest functional integration may be accounted for by a variety of factors, including the lack of cellular support and survival following bolus stem cell injection, and inadequate growth responses of retinal tissue transplantation. Most important, however, is the presence of a postinjury inhibitory extracellular CNS environment.
Pathophysiology
Glial scar formation
Unlike the peripheral nervous system (PNS), the regenerative capacity of the CNS following injury is extremely limited. Amongst other reasons, the paucity of regeneration can be attributed to the presence of an inhospitable extracellular environment ( Box 77.3 ). Unlike the PNS, the CNS is plagued by an abundance of myelin-associated extracellular matrix (ECM) proteins such as myelin-associated glycoprotein (MAG), Nogo, and Omgp that are well known for their ability to inhibit axonal extension and cellular migration. For instance, in 1988, Caroni and Schwab identified the first of these myelin-associated molecules, later termed Nogo, as being a potent inhibitor of fibroblast cell migration and neurite extension.
- •
The inability of the injured central nervous system to regenerate can in part be attributed to the presence of an inhospitable extracellular environment, predominated by the presence of inhibitory glial scar/myelin-related extracellular matrix proteins that prevent axon extension and transplant integration
The abovementioned ECM molecules typically exert their action by binding a common complex of cell surface receptors, in the case of Nogo consisting of the Nogo receptor (NgR) in conjunction with p75 (low-affinity neurotrophin receptor), Lingo, and TROY. Collectively, the binding of these molecules stimulates an inhibitory intracellular signaling cascade which utilizes the small GTPase-dependent enzyme RhoA. Activation of RhoA and its downstream effectors ultimately stimulates growth cone collapse, axon retraction, and cellular repulsion by negative regulation of the actin cytoskeleton. Thus, chemical and/or enzymatic inhibition/neutralization of these ECM molecules or their downstream targets has been shown to alleviate inhibitory myelin-associated growth inhibition in a variety of CNS compartments, including the optic nerve. For instance, in the absence of Nogo-induced RhoA activation, by using either AAV-induced dominant negative NgR expression or NgR-null mice, significantly enhanced retinal ganglion cell axon extension was observed. Similarly, animals vaccinated with spinal cord homogenates rich in myelin-associated proteins could extend axons significantly further than control animals following optic nerve injury. Likewise, when serum from vaccinated animals was used to treat purified cultures of retinal ganglion cells in vitro, it was found that MAG-induced growth cone collapse and axon retraction were alleviated. In light of these findings, regeneration of the retinal ganglion cell layer following injury could potentially benefit from negative regulation of the aforementioned molecules.
As initially suggested, retinal degenerative diseases such as age-related macular degeneration and RP predominantly affect the photoreceptor layer of the retina which, unlike most other CNS tissues, is void of myelin and the associated oligodendrocytes. Thus, the environmental inhibitors mentioned above which impede ganglion cell and optic nerve regeneration are not a factor. Why then are attempts at stimulating retinal regeneration via photoreceptor or stem cell transplantation still largely unsuccessful? The lack of functional integration following transplantation is in large part due to the presence of injury-induced glial scar formation. The retina, like other nervous system compartments, undergoes a process known as reactive gliosis. This is an injury-initiated event that involves the infiltration of a variety of cell types, with the most prominent being activated astroglia. Activation, a process that results in the upregulation of the intermediate filament proteins glial fibrillary acidic protein and vimentin, is crucial for glial scar formation.
In the retina, the major glial cell type responsible for reactive scarring following disease-induced retinal degeneration is the Müller glial cell. During retinal reactive gliosis, hypertrophic Müller glia undergo the activation and upregulation of the above-mentioned intermediate filament proteins ( Figure 77.1A ), after which they respond by extending projections from their original location (forming the OLM) into the subretinal space. Here, these projections proceed to form a dense fibrotic barrier that contains a variety of growth-inhibitory extracellular matrix/adhesion molecules, including the chondroitin sulfate proteoglycan (CSPG) Neurocan and the hyaluronan-binding glycoprotein CD44 ( Figure 77.1B and C ). Both Neurocan and CD44 have previously been shown to function as chemical inhibitors to axon growth and cellular migration, thus preventing regeneration and functional synapse formation. For instance, we have previously shown that an abundance of these molecules, deposited by reactive glial cells at the outer limits of the degenerative mouse retina, prevent neurite extension and subsequent integration following retinal transplantation ( Box 77.4 ).
