Rat Bone Marrow Stromal Cells Differentiate Into Retinal Neuron Phenotypes In Vitro And Induce Rescue Effects In Vivo

Degeneration of the retina and optic nerve is a major cause of visual impairment, and is characterized by a progressive cell loss and an associated dysfunction of retinal neurons. To date, no effective treatment has been discovered to prevent the progressive retinal degeneration or to restore visual function. In recent years, stem cell-replacement therapy has become one of the most promising therapeutic applications for degenerative neural diseases. Bone marrow stromal cells (BMSCs) are multipotent adult stem cells, and offer a distinct advantage by having autologous immunological characteristics as well as a relatively easy isolation and expansion. Previous studies have shown that multipotent BMSCs can differentiate into osteoblasts, adipocytes, cardiac muscle cells and trans-differentiate into neural lineages, both in vivo and in vitro. Although no definitive evidence has been given, BMSCs trans-differentiation into neurons offers interesting new possibilities for generating large numbers of BMSCs-derived retinal neurons.Two main induction protocols have recently been described to induce BMSCs to differentiate into retinal neurons. One method is co-culture induction, which mimics the local microenvironment, and the other is chemical induction of BMSCs. The development of rat retinal neurons can be divided into embryonic and neonatal stages. The development of the neonatal retina forms a second peak in retinal cell development and differentiation, and is the major stage for rod differentiation. It is believed that neonatal retina may provide a high probability of mimicking the developing retinal microenvironment and, therefore, is suitable for inducing BMSCs to differentiate into retinal neurons. Co-culturing with neonatal retinal cells can induce retinal stem cells to differentiate into rods and BMSCs to retinal ganglion cells. However, whether neonatal retinal co-cultures can induce BMSCs differentiation into other retinal neuronal types is still unclear. Several chemical agents have been used in induction protocols. A classic method for neuronal induction has been the use of basic fibroblast growth factor (bFGF). Compared to other protocols, bFGF induction has the advantage of inducing a greater number of neurons to differentiate. A combination of bFGF with brain-derived neurotrophic factor and neuronal growth factor (NGF) can induce BMSCs to differentiate into various retinal neuronal cell types, in particular rods. It has been suggested that bFGF can induce BMSCs into both neuronal and retinal neuronal lineages. However, whether BMSCs can differentiate into retinal neurons using a bFGF induction protocol has not been determined. Furthermore, a comparison between co-culture and chemical induction methods with respect to BMSCs differentiation into retinal neurons has not been carried out. Several studies have focused on transplantation of un-inducted BMSCs into the subretinal space of degenerating retinae or laser induced retinal damage in rats. The transplanted cells survived, migrated and integrated into host retina and there was an improvement in retinal function. These rescue effects, although very promising, need further study.Based on the considerations above, our study initially focused on: 1) the isolation, cultivation, identification and biological characteristics of postnatal (30 days old) BMSCs taken from the femurs of Long-Evans rats. 2) Using cultured BMSCs, we observed and compared the effects on differentiation of bFGF cytokine induction with those following induction by co-culturing with neonatal retina. 3) After completion of the in vitro experiments, a third series of experiments compared the effects of transplanting un-induced cells or induced cells into degenerating retinae. Survival, migration, differentiation and rescue effects were compared between these two methods. Our results are as follows.1. BMSC cultures: BMSCs from femurs of adult Long-Evans rats were successfully cultivated and the number of cells amplified. The cells reached confluence after 7-10 days in primary culture. The passaged BMSCs acquired stable properties, a high purity and the highly differentiated potency of the cells.2. Induced cell cultures: Rat BMSCs co-cultured in a Transwell system with neonatal retinal neurons acquired an increasing number of neuronal-like cell characteristics over time. The induced cells were positive for MAP-2 (neuronal marker), Thy1.1 (retinal ganglion cell marker) and opsin (photoreceptor markers), but negative for PKCα(bipolar cell marker) and GFAP (glial marker). The percentage of MAP-2 and Thy1.1 cells was significantly higher at day 5 than at other time points (P<0.01; 10.8% and 3.8%, respectively). Opsin-positive cells were first observed at day 5 and only formed 1.9% of the population, which was significantly higher than that seen on day 7 (P<0.01).The bFGF cytokine induction protocol induced greater numbers of BMSCs to assume a neuronal-like appearance at day 1 compared to retinal co-cultures; however, bFGF induction produced a marked cell death in the cultures, and the surviving adherent cells were fewer in number after day 1 compared to the co-culture method (P<0.01). The induced cells were positive for MAP-2, Thy1.1 and opsin, but were negative for PKCαand GFAP. In contrast to the co-culture method, the highest percentage of MAP-2 and Thy1.1 cells was observed on day 1. The percentage of MAP-2 and Thy1.1 positive cells in the cultures was significantly higher following bFGF induction compared to co-culture induction at different time points (P<0.01). Opsin-positive cells constituted 2.3% of the sample but were only present on day 1 after bFGF induction, in contrast to retinal co-culturing, where they were first observed on day 5.Control cells were grown in growth medium (DMEM/F12 +10%FBS) and showed no morphological changes and were negatively stained for all cell markers.3. Transplantation: The labeled un-induced BMSCs or bFGF induced cell mixtures composed of BMSCs, phenotypic differentiated neurons and retinal neurons were transplanted into subretianl space in RCS and control normal rats. Survival, migration, differentiation and rescue effects of transplanted cells in the host retinal were evaluated at different time points.Cell survival. Following transplantation into the subretinal space in RCS and normal rats, both un-induced BMSCs and cell mixtures (composed of phenotypically differentiated neurons and retinal neurons derived from BMSCs cultures) survived for 3 months post-transplantation. The number of surviving cells in the degenerating RCS rat retina was significantly higher compared to transplants in the normal rat retina (P<0.05). Thus, transplanted cells had a greater capacity to survive in degenerating retinae compared to normal retinae. In addition, the survival rate of cell mixtures in RCS or normal rat retinae was significantly higher than that of un-induced BMSCs (P<0.05, P<0.01).Cell migration from the transplant: After 3 months post-transplantation in RCS rats, un-induced BMSCs had migrated as far as the outer nuclear layer (ONL) and retinal ganglion cell layer (GCL); in contrast migration into the INL in normal rats was rare. The transplanted cells migrated significantly longer distances and at a faster rate in RCS rats compared to that in normal rats (P<0.05, P<0.01). The cell mixtures had reached the GCL after 2 months post-transplantation in RCS rats, and were rarely seen in the GCL of normal rats after 3 months post transplantation. At all time points examined, the spread of cell mixtures was significantly larger (P<0.01) than that of un-inducted cells in both degenerating and normal retinae.Cell markers: Un-induced BMSCs expressed Thy1.1,opsin,PKCαand GFAP in RCS rats, but only GFAP in normal rats. In contrast, the induced cell mixtures only expressed Thy1.1 and GFAP in RCS rats, and GFAP in normal rats.Rescue effects: The number of host cells surviving in the ONL after cell transplantation was significantly higher than that seen after a sham operation (P<0.05). This suggests that a number of cells were rescued from cell death and degeneration. At one postoperative month the latency and amplitude of the Rod-ERG b wave showed significantly more recovery in transplanted animals compared to sham operated animals (P<0.05). A significant increase in the amplitude of the Max-ERG b wave was also observed 1 and 2 months post-transplantation (P<0.05). However, by 3 months the improvement in the Rod-ERG and Max-ERG b wave latency and amplitude had disappeared (P>0.05). UN-induced and bFGF induced cell mixtures made no significant difference (P>0.05) to photoreceptor rescue and flash-ERG function.Conclusions:1. BMSCs obtained from adult Long-Evans rat femurs are easily cultured to a high purity and have favorable biological characteristics and multipotency. Thus these stem cells are suitable candidates for further experimental manipulation.2. The neonatal retina and bFGF induction protocol can induce rat BMSCs to differentiate into neurons, retinal ganglion cells and rods. The microenvironment of neonatal retinal is suitable for the long-term survival of induced cells. However, the bFGF induction protocol can induce rat BMSCs to differentiate into neurons and retinal neurons when a short induction time is used. Thus provides a new method for further clinical transplantation and effective treatment for patients in a short time.3. The bFGF induced cell mixtures containing phenotypic differentiated cells have stronger survival and migration abilities, and can rescue degeneration of photoreceptors and partly improve visual function in RCS rats after transplantation for 1 and 2 months. It provides a new direction for clinical transplantation in the future to further optimize the BMSCs, obtain an ideal method for transplantation and improve visual function.