视网膜色素上皮细胞的干细胞高效的推导

The various cell types that occupy the retina are organized in a functional architecture. The photoreceptors in the back of the retina are responsible for converting light into electrical impulses through phototransduction. However, phototransduction cannot occur without the coordinated efforts of other neighboring cell types including Mueller glia and retinal pigment epithelium (RPE) cells. A monolayer of RPE cells partitions the sensory retina from the choriocapillaris, the primary blood supply for photoreceptors, and are ideally situated to control multiple functions important for photoreceptor homeostasis. In fact, the RPE and photoreceptors are so co-dependent they are widely considered to be one single functional unit. (For a review of all the diverse functions of the RPE see Strauss, 2005¹.) Death or dysfunction of retinal pigment epithelium cells can induce age-related macular degeneration (AMD), the leading cause of permanent vision loss in industrialized countries^2-4.

AMD is a multifactorial disease of RPE, photoreceptors, and the choroidal vasculature; risk factors are diverse and include combinations of environmental and genetic influences^5,6. Treatments for AMD are very limited, but one promising potential therapy is RPE cell replacement^7,8. While the outcomes have been mixed, the transplantation of RPE cells in AMD patients (and in other patients with retinal degeneration) and also in rodent models of retinal degeneration, has the potential to transiently prevent significant photoreceptor atrophy^9-23. (The animal model commonly used for these studies are Royal College of Surgeons (RCS) rats, which harbor a mutation in the MerTK gene. This mutation renders RPE cells incapable of phagocytosing photoreceptor outer segments and promotes retinal degeneration²⁴.) While the reported survival rates of implanted RPE in the subretinal space of RCS rats and mice vary greatly, there is potential for them to survive for several months or years^9,10,12,20.

RPE cells can be obtained in sufficient numbers for transplantation by deriving them from pluripotent stem cells^9-14,25-28. Several independent groups have demonstrated that these cells function in similar ways to their somatic counterparts, and long term studies suggest that they are safe upon implantation in rat and mouse disease models^{9,10,12,14,19,20,25,29-32}. The use of induced pluripotent stem cells instead of embryonic stem cells may be advantageous since ethical issues and immunological challenges associated with allogeneic RPE may be obviated^33,34. Another exciting application for iPS technology is disease modeling³⁵. The ability to interrogate large numbers of RPE cells derived from patients with RPE diseases could be invaluable for understanding their molecular bases. This type of study has been performed recently with Best disease patient RPE and could pave the way for similar studies of inherited maculopathies³⁶.

The derivation of RPE from stem cells is a relatively simple process and can be done entirely in xeno-free conditions. The simplest strategy is to derive monolayers of RPE cells spontaneously, however, the yield can be significantly improved using directed differentiation techniques. But these techniques involve the use of exogenous protein factors that can be expensive and often generated in bacteria or other non-human sources^10,12,37. In our studies we followed an established protocol¹⁰ that utilizes nicotinamide and Activin A, a signaling factor that has been shown to be sufficient for RPE specification³⁸. Here we will demonstrate that the small molecule IDE-1 can adequately replace Activin A, thus reducing costs and alleviating concerns associated with the use of recombinant proteins³⁹. Additionally, we utilize xeno-free serum replacement, and we culture the differentiating RPE cells on a synthetic xeno-free substrate. RPE cells have been shown previously to differentiate very effectively using this approach⁴⁰.

When differentiated as a monolayer, we visualize pigmented colonies containing RPE cells after as early as five weeks¹². Once they reach sufficient size, they can be manually excised and transferred to another dish for expansion. RPE cells are notorious for dedifferentiating with each passage, and the use of anything older than five passages should be avoided (we find that sufficient numbers of cells for characterization and transplantation in animal models can easily be generated after two or three passages). Once fully differentiated, we employ multiple techniques to characterize the cells anatomically and functionally to ensure that they will serve as adequate replacements for diseased RPE. The description of these techniques, and protocol for implanting the iPS-RPE in the subretinal space of rodents, are beyond the scope of this methods paper and have been previously published^12,32,41.

While developing standardized protocols for effective derivation of iPS-RPE is clearly important for the clinics, there is also significant preclinical work to still be done in animal models. There are concerns regarding immunogenicity of iPS-derived cells, and multiple different implantation techniques, including implanting cells on artificial substrates, are being explored^42,43. For these reasons, we feel that the publication of standardized protocols is beneficial to facilitate both clinical and preclinical studies. Especially if direct comparisons will be done of iPS-RPE cells derived in different labs by different research groups.