Young bone marrow Sca‐1 cells protect aged retina from ischaemia‐reperfusion injury through activation of FGF2

Abstract Retinal ganglion cell apoptosis and optic nerve degeneration are prevalent in aged patients, which may be related to the decrease in bone marrow (BM) stem cell number/function because of the possible cross‐talk between the two organs. This pathological process is accelerated by retinal ischaemia‐reperfusion (I/R) injury. This study investigated whether young BM stem cells can regenerate and repair the aged retina after acute I/R injury. Young BM stem cell antigen 1 positive (Sca‐1+) or Sca‐1− cells were transplanted into lethally irradiated aged recipient mice to generate Sca‐1+ and Sca‐1− chimaeras, respectively. The animals were housed for 3 months to allow the young Sca‐1 cells to repopulate in the BM of aged mice. Retinal I/R was then induced by elevation of intraocular pressure. Better preservation of visual function was found in Sca‐1+ than Sca‐1− chimaeras 7 days after injury. More Sca‐1+ cells homed to the retina than Sca‐1− cells and more cells differentiated into glial and microglial cells in the Sca‐1+ chimaeras. After injury, Sca‐1+ cells in the retina reduced host cellular apoptosis, which was associated with higher expression of fibroblast growth factor 2 (FGF2) in the Sca‐1+ chimaeras. Young Sca‐1+ cells repopulated the stem cells in the aged retina and diminished cellular apoptosis after acute I/R injury through FGF2 and Akt signalling pathways.


| INTRODUCTION
Aging diminishes the functional capacity of stem cells and reduces the ability of the aged organism to repair tissue following injury. 1 Glaucoma, a sight-threatening disorder associated with death of retinal ganglion cells (RGCs) and degeneration of optic nerve fibres, affects millions of people worldwide and advanced age is widely recognised as one of the major risk factors for many of the leading causes of vision loss, including glaucoma. 2,3 The incidence of glaucoma increases exponentially with age. 4,5 Retinal ischaemia-reperfusion (I/R) is a pathophysiological process contributing to the cellular damage in glaucoma. Damage to any type of retinal neuron initiates the loss of vision and therapeutic modalities that can reverse these degenerative processes are required to prevent or reverse vision loss. Moreover, aged and diseased tissues provide an unfavourable microenvironment for regeneration. [6][7][8] Therefore, an intervention to rejuvenate the aged retina may be an important and innovative treatment strategy. 1,9 Recently, stem cell therapy for degenerative retinal diseases has provided hope to restore vision for otherwise incurable disease processes. Bone marrow (BM) is a potential accessible source of autologous stem cells for retinal cell regeneration without the disadvantage of the immune barrier inherent to allogeneic cell transplants. 10 The in vivo mobilisation of BM-derived stem cells (BMSC) can contribute to retinal repair and this regenerative therapy can be implemented without invasive procedures. [11][12][13] BMSCs have been reported to express a variety of cytokines and growth factors, which have powerful trophic and protective functions for neural tissue-derived cells.
However, the number of these cells is relatively low and complete recovery of the damaged tissues remains elusive, especially in older patients who have limited function of endogenous stem and progenitor cells. 14 Recently, we evaluated the effect of cardiac-resident BMSCs on the recovery of cardiac function after injury in aged animals. 15 We found that BM reconstitution in old recipient mice with young donor cells improved cardiac functional recovery following injury. Furthermore, we demonstrated that the young BM cells from the reconstituted BM stably integrated in the aged myocardium prior to injury and were required for functional recovery after injury. 15 Subsequently, we identified the stem cell antigen 1 (Sca-1 + ) cell as the young BM cell type which had the greatest ability to home to the myocardium and protect the aged recipient mouse heart. 16 Sca-1 has been widely used as a marker to isolate haematopoietic stem cells.
This subpopulation of BM cells has been demonstrated to promote neuronal fibre growth in vitro, and cells expressing early ocular markers were highly enriched within the murine Sca-1 + cells. 17 Here, we investigated the importance of young BM-derived Sca-1 + cells for retinal homoeostasis and for retinal repair and regeneration after injury in aged mice. We reconstituted old mice with BM-derived Sca-1 + or Sca-1 − cells from young green fluorescence protein (GFP + ) mice to generate Sca-1 + [Y(Sca1 + )-O] and Sca-1 − [Y(Sca1 − )-O] chimaeras, respectively. Chimaeric mice received retinal I/R to induce injury to the optic nerve. We found that BM chimaerism established with young Sca-1 + cells was associated with better restoration of retina progenitors and improved healing of the aged retina after injury.
The improved regeneration of aged retina involved activation of the fibroblast growth factor 2 (FGF2)-Akt pathway. Suppressing the activation of this pathway using an FGF2 neutralising antibody blocked the beneficial effects of Sca-1 + cells on retinal repair.

| MATERIALS AND METHODS
A detailed methodology can be found in the Supplemental Material.

| Animal and BM reconstitution procedures
The Animal Care Committee of the University Health Network approved all experimental procedures, which were carried out according to the Guide for the Care and Use of Laboratory Animals (NIH, 8th Edition, 2011). Young nucleated BM cells were harvested from femurs and tibias of C57BL/6-Tg-GFP or C57BL/6 wild-type mice (2-3 months, the Jackson Laboratory, Bar Harbor, ME, USA), and separated into Sca-1 positive-and negative-labelled fractions by immunomagnetic activated cell sorting following the manufacturer's instructions (Stem cell Technology, Vancouver, BC, Canada). The purity of positive cells was confirmed by flow cytometry (90.7 ± 1.7%).

| Model of retinal ischaemia-reperfusion injury
Retinal I/R injury was induced as previously described. 18 Briefly, a normal saline reservoir was constructed with a 500 mL IV bottle of balanced salt solution (HBSS containing 0.1% heparin sodium) which was connected to a primary set of pre-pierced Y-site tubing. The Y-site tubing connected to five 30-gauge ½ inch needles with a 10-inch segment of 30-gauge tubing through a five-valve manifold.
The reservoir was hung from an IV pole extension and was elevated to maintain a height of 1.5 m which would subject the eye to hydrostatic pressure of 110 mmHg. At 3 months after BM reconstitution, mice were intubated and ventilated with 1% isoflurane mixed with oxygen. The pupils were dilated with 1% tropicamide. To induce I/R injury, the anterior chamber of each eye was cannulated with the 30-gauge infusion needle approximately halfway between the zonule fibres and the apex of the cornea under a surgical microscope. The infusion was continued for 1 hour to induce retinal ischaemia. The SHAO ET AL. | 6177 microscope was used to verify that no leakage had occurred. Ocular distention was visually confirmed by observing that the eye subjected to I/R was larger than the contralateral eye. Together, the absence of leakage and the presence of ocular distention indicated a successful elevation of intraocular pressure which was also confirmed by whitening of the iris and loss of the red reflex. The needle was withdrawn and the intraocular pressure normalised, resulting in reperfusion, which was confirmed by the reappearance of the red reflex. The control mice underwent the same procedure with the needle inserted into the anterior chamber, but saline was not infused under pressure.
A schematic timetable illustrating the procedures conducted in the in vitro and in vivo studies can be found in Figure S1.

| Statistical analysis
Analyses were performed with Prism version 6.0 software (GraphPad software Inc., San Diego, CA, USA). All values are expressed as mean ± SEM. Student's t test was used for two-group comparisons.
Comparisons of parameters amongst three or more groups were analysed using one-way ANOVA for single-factor variables followed by Tukey or two-way ANOVA for two-factor variables with repeated measures over time, followed by Bonferroni post-hoc tests. Differences were considered statistically significant at P < 0.05. A sample size analysis was conducted to determine the appropriate sample size needed to reliably detect a significant difference between experimental groups.

| BM-derived Sca-1 + cells had greater homing and differentiation capabilities after acute intraocular hypertensive injury
To determine the homing capacity of the young BM Sca-1 + cells to the retina of the aged recipient mice, Sca-1 + [Y(Sca-1 + )-O] and Sca-1 − [Y(Sca-1 − )-O] chimaeras were generated using young BM GFP cells. At 3 months after BM reconstitution, the reconstitution rate for Sca-1 + and Sca-1 − chimaeras was 48.47 ± 1.85% and 31.58 ± 3.11% in BM and 76.97 ± 1.81% and 47.76 ± 3.87% in blood, respectively. GFP expression allowed the tracking of BMderived cell migration into the host retina at 3 months after BM reconstitution. At baseline without injury, only a few GFP cells were found in the retina in either Sca-1 + or Sca-1 − chimaeras ( Figure 1A).
After the induction of I/R injury, more donor-derived GFP + cells were found in the host retina, especially in the inner layers of the retina in both the Sca-1 + and the Sca-1 − chimaeras ( Figure 1A). Further quantification of the GFP + cells in the injured retina 3 and 7 days after injury revealed a significantly greater number of GFP + cells in the Sca-1 + group than the Sca-1 − group, indicating better homing capability of the Sca-1 + cells ( Figure 1B).
Next, to evaluate the differentiation potential of the BM Sca-1 + cells, immunostaining was performed to examine if the GFP + cells were also positive for the neuron marker, NeuN, the microglia marker, F4/80, or the glia marker, GFAP. As shown in Figure 1C, there were GFP + cells which were also positive for NeuN, F4/80 and GFAP, indicating that the homed young cells had the ability to differentiate into all three cell lineages. Quantification of the number of double-positive cells showed that nearly 49.9 ± 4.54% of GFP + cells also expressed the microglial marker, and 15.25 ± 1.45% expressed the glial marker in the Sca-1 + chimaeric retina. Conversely, in the Sca-1 − chimaeras, the corresponding percentages were 32.66 ± 6.45% and 7.34 ± 0.82%, respectively, indicating that more homed cells differentiated into microglia and glia in the Sca-1 + than Sca-1 − group ( Figure 1D). On the other hand, only 1% of GFP + cells expressed the neuron marker NeuN in both the Sca-1 − and Sca-1 + chimaeric retina with no difference between the two groups ( Figure 1D). Under all stimulus conditions, the average P2 latency response was longer compared to that at baseline in all chimaeric groups ( Figure 2H).
The amplitude of P2 decreased significantly at 7 days after I/R injury compared with that at baseline in all three chimaeric groups ( Figure 2I). However, the amplitude of P2 in Sca-1 − chimaeras (15.2 ± 4.5 μv) was even lower than in unsorted-BM chimaeras (21.3 ± 1.7 μv) at 7 days after I/R injury. The amplitude of P2 in Sca-1 + chimaeras (27.1 ± 2.5 μv) was the greatest compared to the other two groups. The shorter P2 latency and the increase in P2 amplitude suggest that Sca-1 + chimaeras had better preservation of visual function.

| Homed BM Sca-1 + cells reduced cellular apoptosis in the host retina
To further elucidate the possible mechanisms responsible for the repair of the aged retina after I/R injury in the Sca-1 + and Sca-1 − chimaeric mice, TUNEL staining was performed at 3 days after injury and more apoptotic cells were observed mostly in the RGC layer of the retina of Sca-1 − than Sca-1 + chimaeric mice ( Figure 3A and B). The protein expression of Bcl2, the downstream protective mediator of cellular apoptosis, increased in both groups 3 days post-I/R injury, but was significantly greater in the Sca-1 + than Sca-1 − chimaeric retina ( Figure 3C and D). Conversely, the protein expression of Bax, the downstream mediator of cell apoptosis, was significantly lower in Sca-1 + than Sca-1 − chimaeras at 3 days after I/R injury ( Figure 3E and F).
In addition, FluoroGold-labelling of viable RGCs demonstrated that the number of RGCs decreased at 3 and 7 days after reperfusion. However, the number of viable RGCs was significantly greater in the retina of the Sca-1 + than Sca-1 − group at 3 and 7 days after I/R injury ( Figure 3G and H). Collectively, these results indicate that homed BM Sca-1 + cells protected the host retina from cellular apoptosis.

| Homed BM-derived Sca-1 cells improved repair of the retina through the FGF2 pathway
Recent studies demonstrated that growth factors have beneficial effects on cell survival in the retina under pathological F I G U R E 1 BM-derived Sca-1 + cells had greater homing and differentiation capabilities after acute ischaemia-reperfusion injury. Bone marrow (BM) Sca-1 + or Sca-1 − cells from young GFP (green fluorescent protein, green, 2 × 10 6 ) transgenic mice were used to reconstitute old wild-type mice, generating Sca-1 + and Sca-1 − chimaeras, respectively. Acute ischaemia-reperfusion (I/R) injury was induced 12 weeks later. Progenitor cells in the retina of recipients were evaluated 3 and 7 days post-I/R injury. Characterisation and quantification by immunolabelling of retinal sections for GFP (A and B) and GFP/NeuN, GFP/F4/80, GFP/GFAP (Glial Fibrillary Acidic Protein) double-positive cells (C and D). BM Sca-1 + cells had greater capability to home to the retina than Sca-1 − cells (n = 4/group; A and B). There was more cell differentiation into microglia (F4/80) and glia (GFAP) in Sca-1 + than Sca-1 − chimaeras after retinal injury (C and D; n = 4/group). INL: inner nuclear layer. Data analysis used two-way ANOVA followed by Bonferroni post-hoc tests for multiple comparisons (B and D). Data shown are mean ± SEM. **P < 0.01, *P < 0.05 F I G U R E 3 Homed BM Sca-1 + cells rescued the host retina from cellular apoptosis. Cellular apoptosis was evaluated 3 days post-ischaemiareperfusion (I/R) injury using Tunel staining. (A) Representative images of apoptotic cells (Tunel + , red) and (B) quantification of Tunel + cells in Sca-1 + and Sca-1 − chimaeras with or without I/R injury (n = 4/group). Bcl-2 and Bax protein expression was evaluated by Western blot. (C) Representative Western blot image and (D) quantification of Bcl-2 protein expression in Sca-1 + and Sca-1 − chimaeras with or without I/R injury. GAPDH was used as a loading control (n = 4/group). (E) Representative Western blot image and (F) quantification of Bax protein expression in chimaeric mice with or without I/R injury. GAPDH used as a loading control (n = 3/group). (G) Representative images and (H) quantification of viable RGCs using FluoroGold (FG)-labelling showed that the number of RGCs decreased at 3 and 7 days after reperfusion. However, the number of RGCs that survived was significantly greater in the retinas of Sca-1 + than Sca-1 − chimaeric mice at 3 and 7 days after I/R injury. RGC: retinal ganglion cell; RGCL: retinal ganglion cell layer; INL: inner nuclear layer; ONL: outer nuclear layer. Data analysis used two-way ANOVA followed by Bonferroni post-hoc tests for multiple comparisons (B, D, F and H). Data shown are mean ± SEM. **P < 0.01, *P < 0.05 F I G U R E 2 Homed BM Sca-1 + cells improved healing of injured retina. The bone marrow (BM) of irradiated old wild-type mice was reconstituted with 2 × 10 6 young BM Sca-1 + , Sca-1 − or unsorted cells, generating Sca-1 + , Sca-1 − , and unsorted chimaeras, respectively. Acute ischaemia-reperfusion (I/R) injury was induced 12 weeks later. The light/dark exploration (A and B) and optomoter (C-E) tests revealed better preservation of visual behaviour in the unsorted than Sca-1 − chimaeras, and the Sca-1 + chimaeras had the best preserved visual behaviour at 7 days after I/R injury (n = 10/group). (F) Flash-visual evoked potential (VEP) was used to evaluate nerve conduction and axon potential propagation through the visual system. Flash VEP waveform latency showed significant delay in the initiation of positive peak (time to P2; G and H) and decreased amplitude of P2 (I) following visual stimulation in Sca-1 − chimaeras compared to unsorted-BM chimaeras 7 days after I/R injury (n = 6/group). However, the Sca-1 + chimaeras had better preservation of visual function than both the unsorted-BM and Sca-1 − chimaeras. Data analysis used two-way ANOVA followed by Bonferroni post-hoc tests for multiple comparisons (A-E, H and I). Data shown are mean ± SEM. **P < 0.01, *P < 0.05 conditions. 19 To elucidate the possible factors responsible for the BM Sca-1 + cell-mediated protection from retinal injury, BM Sca-1 + , and Sca-1 − cells were successfully isolated from young donor mice ( Figure 4A) and differential expression profiles of growth factors for the BM-derived Sca-1 + and Sca-1 − subset of cells were compared using RT-qPCR. As shown in Figure 4B, the mRNA levels of FGF2, insulin-like growth factor 1 (IGF-1), ciliary neurotrophic factor (CNTF), nerve growth factor (NGF), FGF1, stem cell factor (SCF), and neuron-derived neurotrophic factor (NDNF) were significantly higher in Sca-1 + than Sca-1 − cells and the mRNA level of  The expression profiles of growth factors were determined by RT-qPCR. The mRNA levels of FGF2 (fibroblast growth factor 2), IGF-1 (insulin-like growth factor 1), CNTF (ciliary neurotrophic factor), NGF (nerve growth factor), FGF1 (fibroblast growth factor 1), SCF (stem cell factor) and NDNF (neuron-derived neurotrophic factor) were significantly higher in Sca-1 + than Sca-1 − cells. Amongst these growth factors, the mRNA level of FGF2 increased most dramatically in the BM Sca-1 + cells relative to the Sca-1 − cells. (C-P) After 48 hours of culture, FGF2, IGF-1, CNTF, NGF, FGF1, SCF, and NDNF protein levels in the culture medium and cell lysate were determined by ELISA. FGF2 protein level was significantly greater in the medium and cell lysate of BM Sca-1 + than Sca-1 − cells (n = 4/group). Data analysis used un-paired t test. Data shown are mean ± SEM. *P < 0.05, **P < 0.01 retina of Sca-1 + compared to Sca-1 − chimaeric mice ( Figure 7A and B). However, the number of viable RGCs decreased significantly in Sca-1 + chimaeras when an FGF2 neutralising antibody was added compared to Sca-1 + chimaeras without the antibody, indicating that the FGF2 neutralising antibody partially reversed the protective effect of BM-derived Sca-1 cells. Bcl-2 ( Figure 7C and D) and Bax ( Figure 7E and F) protein expression was evaluated in the Sca-1 + and Sca-1 − chimaeras at baseline, after I/R injury, and after I/R injury in the presence of an FGF2 antibody. The protein level of Bcl-2 decreased significantly in the retinas of Sca-1 + chimaeras whereas the protein level of Bax increased significantly after the application of an FGF2 neutralising antibody compared to the Sca-1 + chimaeras without the antibody. The FGF2 protein expression was significantly higher in the retinas of Sca-1 + than Sca-1 − chimaeras 3 days after I/ R injury. However, FGF2 protein expression was significantly decreased with an FGF2 antibody ( Figure 7G and H). Finally, to further confirm activation of the FGF2 pathway, total Akt and p-Akt (phospho Ser473-Akt) expression were measured ( Figure 7I-K). The ratio of p-Akt/total Akt expression was significantly higher in the retinas of Sca-1 + than Sca-1 − chimaeras 3 days after I/R injury. However, the ratio of p-Akt/total Akt expression was significantly decreased after application of an FGF2 antibody. Together, these results were similar to those observed with the in vitro co-culture

| DISCUSSION
In the current study, we investigated the effects of young BM- F I G U R E 5 Homed BM-derived Sca-1 + cells prevented retinal injury through activation of FGF2 pathway. BM Sca-1 + and Sca-1 − cells from young GFP (green fluorescent protein, green, 2 × 10 6 ) mice were used to reconstitute irradiated old wild-type mice, generating Sca-1 + and Sca-1 − chimaeras. Acute ischaemia-reperfusion (I/R) injury was induced 12 weeks later. (A-G) The expression profiles of growth factors in the retinas of Sca-1 + and Sca-1 − chimaeras with or without ischaemiareperfusion (I/R) injury were determined by RT-qPCR. The mRNA levels of FGF2 (fibroblast growth factor 2), IGF-1 (insulinlike growth factor 1) and CNTF (ciliary neurotrophic factor) were significantly higher in Sca-1 + than Sca-1 − chimaeras. (H-I) FGF2 protein expression, as assessed by Western blot, was significantly higher in the retinas of Sca-1 + than Sca-1 − chimaeras 3 days after I/R injury. NGF: nerve growth factor, SCF: stem cell factor, NDNF: neuron-derived neurotrophic factor. n = 4/group; Data analysis used two-way ANOVA followed by Bonferroni post-hoc tests for multiple comparisons (A-G and I). Data shown are mean ± SEM. *P < 0.05, **P < 0.01 Rejuvenation of aged BM by reconstitution with young Sca-1 + cells reduced the functional deterioration following retinal injury. Pressure injury stimulates the release of neurotransmitters which contribute to apoptosis of neuronal cells and active BMSCs protect them from retinal apoptosis. 20 These studies provide preliminary evidence for a new treatment strategy to mitigate the effects of increased intraocular pressure injury.
As reported in previous studies, normally sighted animals showed an aversion to light in the light/dark exploration behavioural test. 21,22 Here we showed that the Sca-1 − chimaeras failed to demonstrate light aversion and reduced their movements in the majority of light/ dark testing paradigms compared to the unsorted-BM chimaeras after retinal I/R injury. However, the Sca-1 + chimaeras showed a robust aversion to light and more movements compared to both the Sca-1 − and unsorted-BM chimaeras. Additionally, these results were consistent with those from the optomotor response test, which was used to evaluate visual defects in this mouse model. 23,24 Most animals compensate for a globally moving environment by turning their head to stabilise the images on the retina. An involuntary head movement was recorded as an oculomotor response, which offered a simple and rapid method to assess visual acuity. In our experiments, the Sca-1 − , unsorted-BM, and Sca-1 + chimaeric animals produced an optomotor response from 0.03 to 0.26 cpd before retina impairment. After retinal damage (by increased intraocular pressure), the Sca-1 − and unsorted-BM chimaeras displayed almost no optomotor response at those same spatial frequencies of the grating test.
However, the Sca-1 + chimaeras still demonstrated head responses at the lowest grating.
The VEP has been employed to identify and monitor glaucomatous nerve damage 25,26 and the components of the flash-VEP are F I G U R E 6 In vitro co-culture of BM-derived Sca-1 + cells with retina explants decreased retinal cell apoptosis. Retina explants were cocultured with BM-derived Sca-1 + or Sca-1 − cells under normoxia and hypoxia conditions. (A) Representative images of apoptotic cells (Tunel + , red) and NeuN + (green) neurons. Nuclei were stained with DAPI. (B) Co-culture with BM Sca-1 + cells protected the retina cells from hypoxiainduced apoptosis. However, this protective effect was lost after application of an FGF2 (Fibroblast Growth Factor 2) neutralising antibody (Ab, n = 6/group). (C and D) Bcl-2 protein expression was significantly greater in retina explants co-cultured with BM Sca-1 + than Sca-1 − cells under both normoxia and hypoxia conditions as assessed by Western blot. However, this effect was blocked by an FGF2 antibody (Ab). GAPDH was used as a loading control (n = 3/group). (E-G) Total Akt and p-Akt (phospho Ser473-Akt) expression was assessed by Western blots. The ratio of p-Akt/total Akt expression was significantly greater in retina explants co-cultured with BM Sca-1 + than Sca-1 − cells under both normoxia and hypoxia conditions. The ratio of p-Akt/total Akt expression was significantly decreased after the application of an FGF2 Ab (n = 3/group). Data analysis used two-way ANOVA followed by Bonferroni post-hoc tests for multiple comparisons (B, D, F and G). Data shown are mean ± SEM. **P < 0.01, *P < 0.05 SHAO ET AL. which may contribute to a variety of pathologic processes. 15,28,29 We found that Sca-1 + cells from young donors showed superior capacity and greater efficiency than Sca-1 − cells to home to and integrate into the aged retina to repopulate the stem cell population in the aged eye after injury. These Sca-1 + cells migrated to the retina, expressed microglial and glial markers and improved retinal response to injury. Other studies have also identified the ability BMSCs, including haematopoietic stem cells, to cross lineage boundaries and to express tissue-specific proteins in the retina and contribute to retinal repair after damage. 11,13,[30][31][32][33] Amongst the circulating BMSCs that migrated to the retina, BMderived microglia in the retina are of particular interest since they have been reported to be recruited by apoptotic neuronal cells in the retina. 33,34 Microglia, monocyte-derived macrophages, and other innate immune cell types play an important role in central nervous system (CNS) homoeostasis during development, adulthood, and aging. 35 A proper balance between the inflammatory responses of these innate immune cells is essential for efficient tissue repair, and immune modulation may be an effective way to promote repair and enhance regenerative therapies. 7,36 Microglia express many versatile receptors 37 39 Indeed, we also found that BM Sca-1 + cells differentiated into microglia after injury. These differentiated Sca-1 + cells may also act as a unique endogenous neuroprotective agent or they may interact with Müller glial cells to control neurotrophic factor production to improve host retinal cell survival.
Delivery of young BM cells to the injured eye may be difficult.
In this study, we describe a new and very effective means to deliver these retinal-protective cells. Rather than direct injection, we reconstituted the BM with these young cells and then employed the normal method of engraftment into the retina. This preliminary study provided proof of principle evidence that this approach is effective. Future studies will need to determine how long after injury BM reconstitution remains effective at regenerating the injured retina.
In this study, we also identified some of the potential mechanisms responsible for the regenerative effects of the young BM Sca-1 + cells. This study demonstrated that BM Sca-1 + cells expressed more growth factors than BM Sca-1 − cells. Consistent with this finding, other studies have shown a significantly higher expression of BDNF and FGF2 mRNA in the Sca-1 + Lin − CD45 − -derived BM stem cell population which could enhance survival and neurite outgrowth of dopaminergic neurons. 40 When we compared a variety of neurotrophic factors expressed by the BM-derived Sca-1 + and Sca-1 − stem cells, the mRNA of FGF2, IGF-1, CNTF, NGF, FGF1, SCF, and NDNF were significantly higher in the Sca-1 + subset. Amongst all these growth factors, FGF2 was identified as a particularly potent neuroprotectant molecule with the highest expression in BM Sca-1 + cells. FGF2 has previously been described as a neurotrophic factor, 41 and is expressed throughout the retina and CNS. FGF2 also exerts a direct neuroprotective effect and can stimulate progenitor cell formation in the retina. [42][43][44] Moreover, it was reported that a reduction F I G U R E 7 Blocking the FGF2 pathway reversed the effect of BM Sca-1 + cells on retinal injury in vivo. The bone marrow (BM) of irradiated old wild-type mice was reconstituted with 2 × 10 6 BM Sca-1 + or Sca-1 − cells from young GFP (green fluorescent protein) transgenic donors, generating Sca-1 + and Sca-1 − chimaeras, respectively. Acute ischaemia-reperfusion (I/R) was induced 12 weeks later. The number of RGCs that survived was determined using FluoroGold-(FG) labelling. (A) Representative images and (B) quantification of viable RGCs in Sca-1 + and Sca-1 − chimaeras at baseline, 7 days after I/R injury, and in the presence and absence of an FGF2 (Fibroblast Growth Factor 2) neutralising antibody (Ab, n = 6/group). Bcl-2 (C and D, n = 3/group) and Bax (E and F, n = 4/group) protein expression was evaluated by Western blot and GAPDH was used as a loading control. (G and H) FGF2 protein expression, as assessed by Western blot, was significantly greater in the retinas of Sca-1 + than Sca-1 − chimaeras 3 days after I/R injury. However, FGF2 protein expression significantly decreased with the addition of an FGF2 Ab (n = 3/group). (I-K) Total Akt and p-Akt (phospho Ser473-Akt) expression was assessed by Western blot. The ratio of p-Akt/total Akt expression was significantly greater in the retinas of Sca-1 + than Sca-1 − chimaeras 3 days after I/R injury. The ratio of p-Akt/total Akt expression was significantly decreased after application of an FGF2 Ab (n = 4/group). Data analysis used two-way ANOVA followed by Bonferroni post-hoc tests for multiple comparisons (B, D, F, H, J and K). Data shown are mean ± SEM. **P < 0.01, *P < 0.05 in FGF2 expression could contribute to the age-related impaired function of human mesenchymal-derived progenitor cells. 45 Our data showed robust expression of the FGF2 protein in the retina of Sca-1 + chimaeras in response to acute I/R injury that was greater than that seen in Sca-1 − chimaeras and was associated with decreased cellular apoptosis and greater neuronal preservation in the Sca-1 + chimaeras. Although the evidence supports that FGF2 is the major factor involved in the BM Sca-1 + cell-mediated protection of the retina from I/R injury, possible effects from other cytokines secreted by BM Sca-1 cells cannot be completely excluded. In addition, effective replacement of BM stem/progenitor cells provided a continuous source of homing of the cells to the site of the injury in a temporal and spatial manner when the repair was needed. Timing of the application and dose of FGF2 needs to be established. Also, it would be technically challenging to continuously supply FGF2 to the retina at the right time and the right place. Reconstituting the BM with these young stem/progenitor cells and then employing the normal method of engraftment into the retina still provided an effective approach to protect against retinal injury.
Associated with the increase in FGF2 protein in Sca-1 + chimaeras, the anti-apoptotic protein Bcl2 was up-regulated and the apoptotic protein Bax was down-regulated. Our data suggest that the improved FGF2 expression along with the homing of young BM-derived Sca-1 + cells may potentiate the regenerative capacity of the aged retina after acute I/R damage. Indeed, Akt, the central mediator for cell survival and tissue regeneration was activated in response to the elevation of the FGF2 protein in the Sca-1 + chimaeric retina. Akt is involved in cellular survival pathways by inhibiting apoptotic processes. Akt also positively regulates some transcription factors to allow expression of prosurvival genes. Akt phosphorylates and activates the expression of caspase inhibitors, 46 and also stimulates the recruitment of Bcl2 to promote cell survival. 47 Here we showed that BM Sca-1 + cells activated Akt possibly through the FGF2 pathway, thereby protecting retinal cells against acute I/R injury. By utilising a neutralising antibody, we confirmed that FGF2 signalling is necessary for the neuroprotective effects of BM-derived Sca-1 + cells.

| CONCLUSION
We identified a subset of BM-derived Sca-1 cells that contributes to

ACKNOWLEDG EMENTS
We thank Dr. Leigh Botly for assistance with manuscript writing, preparation, and editing. This work was supported by grants from the Canadian Institutes of Health Research (143323 to R-KL) and the National Natural Science Foundation of China (81500713).

CONFLI CT OF INTEREST
The authors confirm that there are no conflicts of interest.