Grafted c-kit + /SSEA1 ? eye-wall progenitor cells delay retinal degeneration in mice by regulating neural plasticity and forming new graft-to-host synapses

In the present study, we used FACS to isolate the c-kit⁺/SSEA1^? subpopulation of cells from the eye walls of newborn mice. Our results showed that these c-kit⁺/SSEA1^– cells expressed RPC markers, retained the capacity for cell division, and continued to express high levels of TERT after 20 passages. After transplantation into the subretinal space of rd1 mice, c-kit⁺/SSEA1^? cells differentiated into photoreceptors and increased the overall levels of rhodopsin and recoverin. Our data indicated that synapses had formed between engrafted c-kit⁺/SSEA1^? cell-derived photoreceptors and host bipolar cells and that neural plasticity had been markedly improved, ameliorating the morphological abnormalities of the INL neurons and the degeneration of visual function in rd1 mice significantly.

Retinal degeneration is caused by the progressive loss of the sensory cells of the retina, the photoreceptors, and it accounts for approximately 50% of all cases of blindness in the developed world [1–3]. Treatment efforts include attempts to replace damaged cells by transplantation and strategies for reactivating endogenous stem cell populations to generate new photoreceptors. To date, the efficiency of reactivation and the potential of the newly generated cells are low and are insufficient for the widespread repair of the mature mammalian eye after injury or disease [44–46]. As an alternative, in vitro cell culture can serve as a source of donor cells for cell replacement therapies.

During the early stages after cell transplantation, transplanted stem/progenitor cells release growth factors such as BDNF and NGF to increase the survival and function of the remaining structure [47, 48]. Meanwhile, transplanted cells might promote immunomodulation to suppress microglial activation [36]. They can also release immunoregulatory cytokines and chemokines and express immune-relevant receptors to alleviate the inflammatory response [49]. This bystander neuroprotection by neurotrophic support and/or immunomodulation plays a fundamental role in the therapeutic efficacy of stem cells at early stages. Regarding mechanisms of long-term stem cell-mediated therapy, transplanted cells might migrate from the transplantation site to the ONL, differentiate into photoreceptors, form new synaptic connections to host retinal neurons, and integrate into retinal circuits. The donor cells and their interactions with the recipient retinal microenvironment determine the outcome of the integration.

One major challenge for the recent studies has been to identify the appropriate donor cell population. Several groups have therefore examined the therapeutic effects of cells isolated from embryonic retinas followed by various expansion and differentiation protocols [50, 51]. However, these immature cells failed to integrate into the retina. Nrl⁺ postmitotic photoreceptor precursor cells labeled using green fluorescent protein (GFP) were then demonstrated to be able to correctly integrate into the retina in both the immature, developmental environment and the adult environment [19, 20, 52]. However, the stage equivalent to these precursors occurs early in the second trimester of human gestation. Obvious ethical concerns and extremely limited supply make these cells difficult to use clinically on a large scale. Relying on a three-dimensional differentiation protocol, photoreceptor precursors harvested from three-dimensional retinal cups may be safer than the photoreceptor precursors differentiated directly from embryonic stem cells (ESC) because of the organ-specific character [21]. However, this isolation procedure still depends on the genetic modification model to obtain a rhodopsin/GFP⁺ population, which will not be easy to use in the clinic. In summary, these findings demonstrate that integration is feasible if the correct stage cell is provided. However, this is not a necessary condition for integration. Another study showed that cells isolated from adult mouse retina (4–8 weeks old), forming a heterogeneous pool, could also morphologically integrate into the recipient retina [34]. Whether using retina-isolated cells or ESC-derived cells, the isolation protocol is more effective when a suitable surface marker is used to purify RPCs. In our present study, the eye-wall c-kit⁺/SSEA1^? cells were isolated after birth, and no tumor formation was observed in any of the experiments. Furthermore, c-kit⁺/SSEA1^? cells were capable of integrating into the recipient retina and permitted long-term visual restoration. As a homogeneous population isolated from postnatal retina, c-kit⁺/SSEA1^? cells are safe, effective, and mass-producible, meet the needs of research and clinical contexts, and might be promising donor cells in the future.

In addition to appropriate donor cells, correct integration requires successful migration of transplanted cells from the transplantation site (usually the subretinal space) through the outer limiting membrane (OLM) into the ONL [16]. Two determining factors of this process have been identified: the integrity of the OLM and the extent of recipient retinal gliosis [53–56]. The OLM, via adherens junctions between the terminal processes of the Müller glial cells and the inner segments of the photoreceptors, forms a barrier and restricts the integration of transplanted photoreceptors. It has been reported that disrupting the integrity of the OLM, via pharmacological disruption or transcriptional gene silencing of the OLM-related protein ZO-1, can improve the integration efficacy [53, 55, 56]. However, these two strategies would not be ideal for clinical applications. Toxic effects of pharmacological intervention are detrimental to Müller glial cells [57]. In addition, the OLM and the RPE layer share the adherens junction protein ZO-1, making it also a less than ideal target. On the other hand, reactive gliosis of Müller glial cells leading to glial scarring can decrease retinal integration as degeneration progresses. GFAP^?/?Vim^?/? mice lacking GFAP and vimentin expression show markedly reduced levels of scarring. Subsequently, migration of transplanted cells is clearly increased [54, 58]. Confusingly, gliosis may also have beneficial effects and promote the survival of transplanted cells and remaining cone photoreceptors [59]. In an rd1 RP model, rd1 mice demonstrate severe glial scarring but also show an increase in disturbances of the OLM [53]. Although rods die rapidly, dendrites of rod bipolar cells disappear significantly more slowly [60]. In our present study, the overall levels of synaptophysin and PSD-95 increased significantly, suggesting that eye-wall c-kit⁺/SSEA1^? cells promote structural plasticity after degeneration.