Contribution Of Bone Marrow-derived Cells To Choroidal Neovascularization And Their Potential Application In Therapeutic Strategy

Background choroidal neovascularization (CNV), which refers to the formation of new blood vessels that arise from the choriocapillaries through Bruch's membrane into the subretinal space causing damage of topic choriod and retina, exudation of fluid and hemorrhage, is now known to occur as a final common pathway in nearly 40 ophthalmic diseases leading to irreversible visual loss. The pathogenesis of CNV, which includes multiple cell-types, cytokines and signal transduction pathways, is quite complicated and still poorly understood.Postnatal Vessels are formed through two distinct processes. Angiogenesis involve the remodeling of established capillary networks and arterioles, while vasculogenesis involves the differentiation of stem/progenitor cells into mature vascular cells. Though being'quiescence'generally, these stem/progenitor cells can be mobilized into circulation by physiological or pathologic stimulus to contribute to tissue repair and neoplasia formation at perienchyma. Recent studies have indicated that CNV occur through not only angiogenesis, but also vasculogenesis, that is, both bone marrow-derived cells (BMCs) and cells in situ participate in CNV development. However, little is known about the dynamic conduct of BMCs in the CNV microenvironment. BMCs, a heterogeneous cell population, is comprise of multiple stem/progenitor cell-types. Which types of them contribute to CNV and what role they play in the process? Whether they could be the new targets for CNV treatment? Lots of issues need to be studied.Objectives This study aims⑴to investigate the contribution of BMCs to CNV, and the dynamic conduct of BMCs in CNV microenvironment;⑵to investigate the effects of nicotine on BMCs'contribution to CNV and the underlying mechanism;⑶to investigate the specific recruitment of MSCs to CNV and their contribution to CNV formation, to confirm the potential of MSCs being cell vectors;⑷to explore a noval therapeutic strategy of applicating MSCs as delivery vehicles for CNV treatment.Methods⑴Green fluorescent protein (GFP) chimeric mice were developed by transplanting unpurified bone marrow cells from GFP +/+ transgenic mice to wild-type adult C57 mice. The qualified chimeric mice underwent laser rupture of Bruch membrane to induce CNV. Choroidal flatmount was performed to detect BMCs recruitment to and participation in CNV lesions, and immunofluorescent staining was performed to detect the phenotype of GFP+ cells (including vascular endothelial cells (VEC), vascular smooth muscle cells (VSMC) and macrophages) and expression of vascular endothelia growth factor (VEGF) and basic fibroblast cell growth factor (bFGF) in CNV;⑵GFP chimeric mice were developed. CNV was induced by lasering, and nicotine was administered orally in drinking water of mice in nicotine group on that very day. The chimeric mice in control group were normal raised. Four weeks later, histopathologic study and choroidal flatmount were performed to measure the CNV severity and BMCs recruitment. The differentiation of BMCs in CNV and local expressions of VEGF, bFGF and vascular cell adhesion molecule-1 (VCAM-1) were detected by immunofluorescent staining;⑶MSCs derived from GFP transgenic C57 mice were enriched and cultured. 4.0×106 GFP-MSCs or unpurified bone marrow cells were injected into wild type C57 mice through tail veins 0.5-1 hour after laser photocoagulation induction of CNV. To compare the migration of these two cell-types, GFP+ cells were detected in peripheral blood, bone marrow, heart, liver, spleen and lung. To study the time window of MSCs recruitment, mice were randomly divided into three groups. Mice in one group received single injection of 4.0×106 GFP-MSCs 0.5-1 hours after laser photocoagulation. The second group received twice injection at 0.5-1 hours and 3 days after lasering, and mice in another group received three times injections at 0.5-1 hours, 3 days and 6 days after lasering, respectively. Migration of GFP-MSCs in vivo was observed and compared. Expression of SDF-1, an important chemotactic factor for stem cells, in eyes after lasering was detected by using ELISA and immunofluorescent staining. Seven days after laser, choroidal flatmount was performed to investigate the participate of MSCs in CNV, and the differentiation of MSCs in CNV were analyzed by cell-markers immunofluorescent staining;⑷Adenoviral vectors (Ad) expressing human pigment epithelial-derived factor (PEDF) were generated with a reporter gene (GFP). Control vectors that do not express PEDF were constructed and produced in parallel (AdNull). MSCs were transduced. Transduction efficiency was detected by inverted microscopy and flow cytometry. Concentrations of human PEDF released by the transduced MSCs were measured by ELISA. AdPEDF/ MSCs, AdNull/MSCs, GFP-MSCs (4.0×106 cells/mouse) or 0.4ml PBS were injected into C57 mice tail veins 0.5-1 hour after laser photocoagulation. Human PEDF expression in mice eyes was examined by ELISA and immunofluorescent staining. One week after laser, CNV lesion severity was measured by quantitative analysis of both histopathology and choroidal flatmount. To test the impact of PEDF secreted by transduced MSCs on the RPE proliferation and migration, coculture system of AdPEDF/MSCs and RPE cells was established.Results⑴Large number of GFP-BMCs appeared in the CNV lesions and integrated into CNV, constituting 16.22% of the total vascular area. Most GFP-BMCs appeared in CNV lesions (including choroid beneath CNV lesion), and a few GFP-BMCs were in neurosensory retina over CNV, corneoscleral limbus, ciliary body, optic disc and sclera, retina and choroid distant from CNV. GFP+ cells, which were immunoreactive forαSMA or CD31, appeared in CNV lesions only. However, F4/80+/GFP+ cells can be also detected in neurosensory retina over CNV, corneoscleral limbus and ciliary body. The constituent ratio of those three cell-types in total GFP-BMCs in CNV altered as CNV developed. The maximal ratios of CD31-labeled cells and F4/80-labeled cells presented at 2 week, while the ratio ofαSMA-labeled cells upgraded continuously. Immunofluorescent staining showed some BMCs in CNV were VEGF or bFGF positive.⑵Nicotine administration resulted in larger diameter and surface area of CNV. Nicotine-exposed mice demonstrated increased area and density of GFP+BMCs, increased GFP+ vascular cells area, and decreased ratio of BMCs expressing F4/80 in CNV. Furthermore, the expression of VEGF and bFGF within CNV and VCAM-1 in choroid beneath CNV was up-regulated in nicotine-exposed mice.⑶GFP-labeled BMCs , rather than MSCs, were found in lungs, livers and spleens at all time point. The quantity of MSCs recruited to CNV after singer, twice, or thrice MSCs-injections was compared, and there was no statistic difference between each two groups, while lots of GFP-MSCs were found in spleens of mice in the latter two groups. SDF-1 expression in eyes of mice models was examined. Initially, SDF-1 was expressed in RPE layer. As CNV formed, positive staining presented in CNV. The level of SDF-1 expression increased rapidly in the first 24h, reached peak at the second day, and then decreased sharply. Flow cytometry analysis showed MSCs underwent a transient retention in bone marrow. Choroidal flatmount indicated that 1 day after laser, GFP-MSCs were found dispersed around the laser sites. Subsequently, a directional movement of GFP-MSCs coming closer toward the CNV lesions presented. On day 7, GFP-MSCs were found in CNV and appeared to participate in vascular structure formation. Most GFP-MSCs were located in CNV between choroid and photoreceptors. In differentiation analysis, positive expression was found for cell-markers of VEC, VSMC, macrophage, fibroblast and epithelial cell in MSCs examined.⑷Reporter GFP expression in MSCs was found 24 hours after transduction by using inverted fluorescent microscope. Flow cytometry analysis further showed (73.6±5.3)% of the total cell population was GFP positive. Human PEDF production by AdPEDF/MSCs persisted for at least 8 days in vitro. After transplantation of AdPEDF/MSCs, human PEDF positive staining was found in MSCs and the nearby extracellular matrix in CNV. ELISA analysis showed that PEDF expression persisted for at least one week, the average production of human PEDF in the eyes was steady and the level was 4 folds higher than the quantity required to elicit anti-angiogenic effects for CNV. Histopathology and flatmount analysis indicated that there was significant reduction in CNV thickness, diameter and surface area in the AdPEDF group. The proliferation and migration of RPE cells which cocultured with AdPEDF/MSCs were enhanced.Conclusion⑴BMCs play an important role in CNV by differentiating into multiple cell-types to compose CNV and secreting angiogenic factors. CNV microenvironment recruits BMCs, supports and regulates the differentiation of BMCs. There may be interaction between BMCs and the CNV microenvironment.⑵Nicotine promotes recruitment and incorporation of BMCs into CNV, and affects differentiation of BMCs in CNV. These effects may be partly due to regulation of related factors (e.g. VEGF or VCAM-1) expression by nicotin.⑶Comparing with unpurified BMCs, MSCs could be specifically recruited into CNV lesions, without stagnation in other organs, indicating their predominant potential for specific recruitment to CNV. Single lasering induction results in a finite time window for MSCs recruitment, it may due to the expression pattern of chemotatic factor, such as SDF-1, in CNV. After laser photocoagulation, MSCs were specifically recruited into CNV lesions gradually, and differentiated into multiple cell types required for CNV formation.⑷Genetic engineered MSCs localized and produced anti-angiogenic factor in CNV lesion. Thus, the growth of CNV was inhibited in vivo. The effect may be mediated, partly at least, by RPE cells, which function as an important regulator for CNV development. MSCs could serve as delivery vehicles of anti-angiogenic agents for the treatment of a range of CNV-associated diseases.