Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Glia. Author manuscript; available in PMC 2011 January 1.
Published in final edited form as:
PMCID: PMC2814838

Maintaining retinal astrocytes normalizes revascularization and prevents vascular pathology associated with oxygen induced retinopathy


Astrocytes are well known modulators of normal developmental retinal vascularization. However, relatively little is known about the role of glial cells during pathological retinal neovascularization (NV), a leading contributor to vision loss in industrialized nations. We demonstrate that the loss of astrocytes and microglia directly correlates with the development of pathological NV in a mouse model of oxygen induced retinopathy (OIR). These two distinct glial cell populations were found to have cooperative survival effects in vitro and in vivo. The intravitreal injection of myeloid progenitor cells, astrocytes, or astrocyte-conditioned media rescued endogenous astrocytes from degeneration that normally occurs within the hypoxic, vaso-obliterated retina following return to normoxia. Protection of the retinal astrocytes and microglia was directly correlated with accelerated revascularization of the normal retinal plexuses and reduction of pathological intravitreal NV normally associated with OIR. Using astrocyte-conditioned media, several factors were identified which may contribute to the observed astrocytic protection and subsequent normalization of the retinal vasculature, including VEGF and bFGF. Injection of VEGF or bFGF at specific doses rescued the retinas from developing OIR-associated pathology, an effect which was also preceded by protection of endogenous glia from hypoxia-induced degeneration. Together, these data suggest that vascular-associated glia are also required for normalized revascularization of the hypoxic retina. Methods developed to target and protect glial cells may provide a novel strategy by which normalized revascularization can be promoted and the consequences of abnormal NV in retinal vascular diseases can be prevented.


The abnormal growth of blood vessels and associated vascular leakage in diabetic retinopathy, exudative age-related macular degeneration, retinopathy of prematurity, and vascular occlusions, are major causes of vision loss (Adamis et al. 1999; Das and McGuire 2003; Friedlander 2007). Numerous studies have demonstrated a role for VEGF in the development of pathological neovascularization (NV) in the eye (Aiello et al. 1995; Caldwell et al. 2003; Ferrara and Kerbel 2005; Ng et al. 2006; Witmer et al. 2003), and anti-VEGF drugs are currently in clinical use to slow the progression of retinal vascular disease (Avery et al. 2006; Ng et al. 2006). However, potential problems with anti-VEGF therapy could limit its utility including the need for repeated injections as well as potential secondary effects on other, non-endothelial, cell types (Zachary 2005). In addition to its role in angiogenesis and vascular permeability, VEGF has also been shown to have important neuroprotective activities in the retina (Robinson et al. 2001) and is a critical modulator of vascular associated cells such as astrocytes and microglia.

Cell-based therapy represents a newly emerging method by which vascular and neuronal degenerative diseases may be treated (Friedlander 2005; Friedlander et al. 2007; Otani and Friedlander 2005). Intravitreal injection of vascular-related progenitor cells prevents vascular regression and protects neurons in mouse models of retinal degeneration (Otani et al. 2004; Otani et al. 2002). Injection of myeloid progenitors facilitates vascular repair in models of ischemic retinopathy by accelerating normal revascularization of the superficial and deep retinal vascular plexuses, and decreasing intravitreal vascular pathogenesis (Ritter et al. 2006). These cell based therapies have the potential to correct the underlying vascular abnormalities associated with ischemic retinopathies rather than simply treat the complications resulting from abnormal NV. While we have gained some insight regarding the vasculo- and neurotrophic properties of these cells, the precise mechanism by which they facilitate vascular rescue models of ischemic retinopathy remains unclear.

The neovascular response to hypoxic stress in vascular disease models varies significantly from one mouse strain to the next (Rohan et al. 2000). This is exemplified by the response of albino BALB/cByJ and pigmented C57BL/6J mice to models of oxygen induced retinopathy (OIR) (Ritter et al. 2006; Smith et al. 1994) (Figure 1). In the mouse OIR model, a transient period of hyperoxia during retinal vascular development causes the attenuation of normal retinal vessel development and degeneration of the newly formed, immature vasculature. Upon return to normoxia, the central retinal area where the vasculature has regressed becomes hypoxic leading to development of pathological NV at the interface between the perfused and non-perfused retina (Figure 1A-E) (Smith et al. 1994). This model closely resembles retinopathy of prematurity (ROP), a condition which can be observed in premature infants who are placed in hyperoxic chambers thereby introducing the risk for developing retinopathy after removal from the hyperoxic chambers (Smith 2002). In the OIR model, the retinal vasculature of both BALB/cByJ and C57BL/6J mice becomes obliterated to a similar extent during exposure to hyperoxia (Figure 1F). However, in retinas from BALB/cByJ mice, the vascular obliterated regions revascularize more quickly and pathological, pre-laminar neovascular tufts do not form. In contrast, the retinas of pigmented C57BL/6J mice are characterized by slower revascularization of the normal plexuses and the formation of pathological, pre-laminar neovascular tufts (Figure 1F).

Figure 1
BALB/cByJ and C57Bl/6J mouse strains react differently to OIR. A) Mouse model of OIR: 7 day old pups (post-natal day 7; P7) and their mothers are placed in 75% oxygen chambers for 5 days followed by return to ambient room air at P12. The retinas are analyzed ...

Astrocytes play a critical role during normal inner retinal vascularization (Dorrell et al. 2002; Fruttiger et al. 1996; Provis et al. 2000; Stone et al. 1995) and degeneration of retinal astrocytes in ischemic tissues is associated with failure of the blood retinal barrier in oxygen induced retinopathies (Chan-Ling and Stone 1992). In this study, we examine the potential role of astrocytes in OIR and explore the potential utility of cell therapy for the treatment of retinal ischemia. We demonstrate a strong correlation between astrocyte survival, rapid retinal revascularization, and lack of pathological NV in BALB/cByJ retinas when compared with retinas from C57BL/6J mice. We also describe trophic cross-talk between astrocytes and retinal microglia and provide evidence that the rescue potential of myeloid progenitor cells may, at least in part, be facilitated by protecting retinal astrocytes from degeneration in the C57BL/6J mouse retina. We further support a role of astrocytes in promoting normalized revascularization of the OIR retina by data demonstrating that other intervention strategies that protect the endogenous astrocytes from hypoxia-related degeneration, including the injection of astrocytes, astrocyte-conditioned media, and even low levels of VEGF or bFGF, also rescue the OIR phenotype. We therefore conclude that there is a strong correlation between astroglial survival and normalized retinal revascularization following oxygen-induced obliteration and suggest a potential new therapeutic strategy for treating ischemic retinopathies that targets and rescues vasculature-associated glial cells rather than targeting and blocking the growth of vascular endothelial cells.

Materials and Methods

Intravitreal injections and the OIR model

All animal protocols were approved by the IACUC committee at the Scripps Research Institute, La Jolla, California. The OIR model was used as previously described (Banin et al. 2006; Smith et al. 1994). Briefly, post-natal day 7 (P7) pups and their mothers were exposed to 75% oxygen for 5 days and returned to normoxia at P12 (Figure 1A). Exposure of the mouse pups to hyperoxia results in obliteration of the newly developed central retinal vessels (Figure 1B-C). When the mice are returned to normoxia, the retina becomes hypoxic due to the absence of normal retinal vasculature, resulting in pathological neovascular growth by P17 (Figure 1D-E). Appropriate cells (250,000 cells / eye), astrocyte-conditioned media, or various doses of bFGF or VEGF were injected intravitreally on post-natal day 7 (P7) or on post-natal day 12 if noted in the results and figure legends. Mice were euthanized and retinas isolated at various timepoints following return to normoxia. Following fixation using either 10 minutes of methanol at 4 °C (GFAP) or 1 hour of 4% paraformaldehyde (CD11b, isolectin GS), retinas were blocked for 2 hours and stained for astrocytes using GFAP-specific antibodies (Sigma, St. Louis, MO; methanol fixation, 10 minutes), microglia using CD11b-specific antibodies (BD Biosciences, San Jose, CA), and / or vessels using the fluorescently conjugated isolectin Griffonia Simplicifolia IB-4 (isolectin GS; Invitrogen). Isolectin GS can be used to effectively label murine endothelial cells and activated macrophages (Maddox et al. 1982), thus allowing us to differentiate resting microglia (CDllb positive) from activated macrophages (CDllb and lectin positive). Quantification of the remaining obliterated areas, used as a measure of the rate of normal retinal revascularization, and quantification of the areas of pathological neovascular growth were performed at P17 as previously described (Banin et al. 2006).

Isolation of primary astrocytes

Astrocytes were isolated using a published protocol (Gebicke-Haerter et al. 1989). Briefly, cerebral hemispheres were dissected from brains of neonatal mice. The dorsal cortex was isolated, avoiding the underlying white matter, and the tissue was minced in Hank's Balanced Salt Solution using forceps and by repeat pipeting with a P1000 pipette. After centrifugation at 1000 × G for 5 minutes, 2.5 mL trypsin-EDTA (Invitrogen, Carlsbad, CA) per dissected brain was added and the tissue was dissociated at 37 °C for 30 minutes followed by deactivation of the trypsin with an equal volume of growth media (DMEM + 10% FBS + Glutamax + Pen-Strep; Invitrogen). The resulting cells were plated at low density in 75 cm2 flasks and grown to confluency. These cells consisted mainly of astrocytes and microglia. To separate these two populations, the flasks were shaken in a culture chamber (humidified with 5% O2) overnight at 225 rpm. The microglia-containing media was removed leaving behind the attached astrocytes (approximately 95% pure), which were grown further and used within 3 passages. Alternatively, the astrocytic cell line, C8-D1A (ATCC) was used. These cells were grown in normal DMEM culture media containing 10% FBS. Primary or C8-D1A astrocytes were intravitreally injected at a concentration of 250,000 cells per eye using a 33-gauge needle and injecting a total volume of 0.5 μL per eye.

Isolation of bone marrow-derived myeloid progenitor cells (BM-MPCs)

Whole bone marrow was isolated from the femurs and tibias of 6-8 week old C57BL/6J mice. The isolated bone marrow cells were incubated with a fluorescently labeled antibody directed against CD44 (BD Biosciences — Pharmingen, clone IM7). Fluorescence-activated cell sorting (FACS) was performed on a FACSAria (BD) and was used to isolate CD44hi subpopulations that correspond to the myeloid progenitor fraction of the whole bone marrow (Ritter et al. 2006). 250,000 cells per eye were injected intravitreally using a 33 gauge needle in a total volume of 0.5 μL per injection.

Injection of astrocyte-conditioned media

Primary murine astrocytes or astrocytes from the murine C8-D1A cell line (ATCC) were grown in normal DMEM culture media containing 10% FBS until approximately 80% confluent. The cells were then washed thoroughly with PBS to remove any remaining serum followed by culture in serum-free DMEM. After 7 days, the astrocyte-conditioned, serum-free media was collected and cellular debris was removed by centrifugation at 5000*G for 5 minutes. The remaining supernatant was collected and concentrated 100-fold using 10 kDa molecular weight cutoff centrifuge filters (Pierce). The 100x concentrated astrocyte-conditioned media was injected intravitreally in 0.5 μL volumes. Control, non-conditioned serum-free DMEM media was also concentrated 100-fold and injected intravitreally in 0.5 μL volumes as a negative control.

Semi-quantification of astrocyte rescue in the retinal vascular obliterated areas

The extent of astrocyte persistence in the vascular obliterated areas was quantified by analyzing the OIR retinas at P14, 2 days following removal from hyperoxia. Retinas were dissected, flat-mounted, and stained for GFAP to visualize the astrocytes. Masked observers scored the extent of astrocyte persistence in the obliterated zones on a 0 – 6 scale. Values of 0 or 1 indicated no or very few astrocytes were observed in the vascular obliterated zone. Values of 2 - 4 indicated retinas with substantial numbers of astrocytes remaining in the vascular obliterated zones (although still fewer than normal), but indicated that these astrocytes appeared abnormal, lacked the normal cellular processes, and the standard network observed in a normal astrocytic template was not observed. Values of 5 - 6 indicated retinas with large numbers of astrocytes which were observed to form a mostly to completely normal astrocytic template within the vascular obliterated zones.

Western blot and proteomic analysis

Serum-free astrocyte-conditioned media was collected and concentrated 100-fold as described for the intravitreal injections. 20 uL samples of the concentrated conditioned media or concentrated non-conditioned control media were then separated by polyacrylamide gel electrophoresis. After Coomassie staining, prominent bands were cut using sterile technique and analyzed by LC-MS/MS mass spectrometry to identify proteins in the astrocyte-conditioned media. In separate gels, the samples were transferred to nitrocellulose paper. After blocking with 5% milk buffer in PBS + 0.1% Tween, the blots were incubated with rabbit anti-mouse bFGF (NOVUS, 008-100) or rabbit anti-mouse pan VEGF (Chemicon, AB1876). The blots were then washed, incubated with HRP conjugated secondary antibodies, washed again, and visualized by chemiluminescence (Pierce, 34075).

LC-MS/MS analysis

Proteins were solubilized in 0.1% Rapigest (Waters Corportation) and reduced in 10 mM D,L-dithiothrietol (Sigma) for one hour followed by alkylation with 55 mM iodoacetamide (Sigma) for 30 minutes in the dark. Proteins were trypsin digested at 37°C using a 1:30 (w/w) enzyme to substrate ratio. The peptides were analyzed by reverse phase chromatography prior to mass spectrometry analysis using Zorbax SB-C18 stationary phase (Agilent; 5 μm particles, 15 cm length, 75 μm inner diameter). The reverse phase gradient separation was performed using water and acetonitrile (0.1% formic acid) mobile phase gradients.

The data-dependent MS/MS data was obtained on a LTQ linear ion trap mass spectrometer using a home built nanoelectrospray source. One MS spectrum was followed by 4 MS/MS scans on the most abundant ions after the application of the dynamic exclusion list. Protein identification was performed using Mascot (Matrix Science, London, UK, version 2.1.04) at the 95% confidence level with a calculated false positive rate of < 1% as determined using a reversed concatenated protein database.

Tandem mass spectra were extracted by the Xcalibur software. All MS/MS samples were analyzed using Mascot (Matrix Science, London, UK; version 2.1.04). Scaffold (version Scaffold-01_06_03, Proteome Software Inc., Portland, OR) was used to additionally validate MS/MS based peptide and protein identifications. Peptide identifications were accepted if they could be established at greater than 95.0% probability as specified by the Peptide Prophet algorithm (Keller et al. 2002). Protein identifications were accepted if they could be established at greater than 99.0% probability and contained at least 2 identified peptides.


Astrocyte and microglia loss in the hypoxic retina correlates with OIR pathology

Over the past two decades, models of oxygen-induced retinopathy (OIR) have become common for studying hypoxia-induced angiogenesis (Chan-Ling and Stone 1992; Madan and Penn 2003; McLeod et al. 1998; Penn et al. 1993; Smith et al. 1994). Vascular changes observed in the mouse model of OIR are consistent, reproducible, and quantifiable (Banin et al. 2006), and its use has been extended to the study of angiogenesis in general, as well as testing various angiostatic molecules (Eichler et al. 2006; Ferrara and Kerbel 2005). We analyzed two strains of mice, BALB/cByJ and C57BL/6J, for differences in the response of vascular-related cells in OIR. As previously noted (Ritter et al. 2006), retinas from both strains demonstrated equal levels of hyperoxia-induced vascular obliteration. However, the retinal vascular response to the ensuing hypoxia after removal from hyperoxia was dramatically different in the two strains (Figure 1F). The retinal vasculature of BALB/cByJ mice revascularized rapidly reproducing the characteristic superficial and deep vascular plexuses without pathological NV by 5 days post-hyperoxia (P17). In contrast, retinas of C57BL/6J mice exhibited slower revascularization of the normal retinal plexuses and extensive pathological NV (Figure 1F).

We next analyzed the retinal astrocytes to determine if there were also strain differences in the survival and persistence of retinal astroglia. In both BALB/cByJ and C57BL/6J mice, retinal astrocytes survived through hyperoxia and were observed in the vaso-obliterated areas upon return to normoxia at P12 (Figure 2A). In retinas from C57BL/6J mice, astrocytes in the vaso-obliterated zone quickly began to degenerate and were mostly absent by 48 hours following return to normoxia (P14) (Figure 2 B-C). The loss of astrocytes is accompanied by an increase in Müller cell GFAP reactivity which is observed as spotted staining from the Müller cell end-feet within the superficial vascular plexus. These activated Müller glia are observed as GFAP positive cells with processes spanning the entire retina (Figure 2D). Activation of Müller glia and subsequent GFAP reactivity is commonly associated with retinal stress and vascular disease (Dorrell et al. 2009; Rungger-Brandle et al. 2000). In contrast, astrocytic density remained high within the vaso-obliterated zone of retinas from BALB/cByJ throughout revascularization and Müller cell activation was not observed (Figure 2B-C). Thus, astroglia survival correlates with accelerated physiological revascularization and decreased pathological changes in the BALB/cByJ mouse retinas. Differences in myeloid cell persistence have also been observed in the BALB/cByJ and C57BL/6J mouse retinas following OIR (Ritter et al. 2006). These results suggest that the persistence of retinal astrocytes and microglia within the vaso-obliterated areas may help facilitate normal revascularization, as observed in the BALB/cByJ strain, whereas their degeneration may contribute to the pathological neovascularization normally observed in the C57BL/6J mouse retina.

Figure 2
Astrocytes degenerate in the avascular, hypoxic areas of the central retina following return of the OIR retinas to normoxia. A) Staining with GFAP demonstrates the presence of astrocytes in the vaso-obliterated areas of retinas from both C57Bl/6J and ...

Injection of BM-MPCs mediates astrocyte survival in C57BL/6J mice

We have previously demonstrated that intravitreal injection of bone marrow-derived myeloid progenitor cells (BM-MPCs) leads to enhanced physiological revascularization and reduced pathological NV in C57BL/6J mouse retinas following hyperoxia, resulting in a phenotype that more closely resembles that of BALB/cByJ mice (Ritter et al. 2006) (Figure 3A). Based on our findings that correlate astrocyte persistence in the retina with accelerated physiological revascularization and reduced pathology, we tested whether BM-MPCs also protected astrocytes from degeneration within the hypoxic, vascular-obliterated zones of C57BL/6J following OIR. Injection of BM-MPCs at P7 resulted in astrocyte survival and maintenance of a relatively normal astrocytic template within the vaso-obliterated central retina (Figure 3B-C). This was correlated with accelerated revascularization of the normal retinal plexuses and reduced pathological NV. These results also support a correlation between astrocyte survival and normal revascularization of the ischemic retina; astrocyte survival at P14 occurred in retinas with BM-MPC treatment which subsequently resulted in a normalized vascular phenotype by P17. Vehicle (PBS) control injection also rescued the astrocytes from degeneration compared to non-injected OIR controls (Figure 3C), but at significantly lower levels than observed with BM-MPC injection. This correlates with a slight OIR rescue consistently observed by vehicle or sham injection compared to non-injected retinas (Ritter et al. 2006).

Figure 3
Injected bone marrow-derived myeloid progenitor cells (BM-MPCs) protect retinal astrocytes and rescue the retinal vasculature following OIR. A) Injected BM-MPCs rescue the OIR phenotype by reducing the area of obliteration in C57BL/6J retinas at P17 (demonstrating ...

Astrocytes mediate survival of myeloid cells and endogenous astrocytes

Bone-marrow derived progenitor cells target specifically to sites of gliosis, reactive astrocyte activity and growth (Dorrell et al. 2004; Otani et al. 2002). Having demonstrated that BM-MPC injections promote astrocyte and vascular rescue, presumably by replacing the degenerated microglia population and protecting the endogenous astrocytes from degeneration, we next tested whether intravitreal injection of astrocytes can also promote cell survival and facilitate physiological revascularization in the OIR model. Both primary astrocytes, isolated from neonatal mouse brains, and the astrocytic cell line C8-D1A promoted growth of myeloid cells in vitro. BM-MPCs survived and proliferated in co-culture with astrocytes whereas the same cells did not survive using comparable culture conditions without astrocytes (Figure 4A). Serum-free media pre-conditioned by astrocytes also facilitated myeloid progenitor cell growth in vitro, suggesting that myeloid cell survival is facilitated by astrocyte secreted factors. Thus, it appears that the trophic effects of astrocytes and the BM-MCPs occurs in a reciprocal manner, at least in vitro.

Figure 4
Injection of astrocytes or astrocyte-conditioned media results in normalized revascularization in the OIR model. A) Astrocytes facilitate growth and survival of BM-MPCs in culture. B) Injection of astrocyte-conditioned media results in persistence of ...

Having demonstrated that the persistence of astrocytes and microglia within the vascular obliterated zones of the OIR retinas was highly correlated with normalized revascularization, we wanted to determine if astrocytes or astrocyte conditioned media could also protect microglia and endogenous astrocytes in vivo. Intravitreal injection of astrocytes into C57BL/6J OIR mice at P7 led to increased density of the endogenous microglia (CD11b) and the presence of activated macrophages within the vaso-obliterated zones at P14 (2 days following removal from OIR). As previously observed, microglia were present upon removal from hyperoxia but were substantially reduced by P14 within the hypoxic, vascular-obliterated zones of control injected C57BL/6J OIR mice (Figure 4B). Relatively normal astrocytic templates were also observed following astroglia cell injections indicating that these treatments also promote survival of endogenous astrocytes. Similar to the results observed following injection of astrocytes, injection of astrocyte-conditioned media also promoted survival of the endogenous microglia and maintenance of the endogenous astrocytic template within the vascular-obliterated zones of C57BL/6J retinas (Figure 4 B-C). Based on semi-quantitative scores from each of four masked observer's, astrocyte survival within the vascular obliterated zones of OIR retinas was significantly enhanced in retinas injected with astrocyte-conditioned media (Figure 4C). These data demonstrate that astrocyte-secreted factors promote the survival of myeloid cells in vitro, and endogenous retinal microglia and astroglia in vivo.

Intravitreal injection of astrocytes mediates rescue of the OIR phenotype

We next determined whether intravitreal injection of astrocytes also resulted in modified retinal vasculature following OIR. Injection of either C8-D1A astrocytes or primary astrocytes at P7 (prior to exposure to hyperoxia) protected the retinas from the pathological changes observed in OIR to an extent comparable to that observed when BM-MPCs were injected (Figure 4D-E). Revascularization of the normal retinal plexuses was accelerated, as demonstrated by a decrease in the remaining vaso-obliterated areas at P17, and pathological NV was also inhibited, as shown by a decrease in areas of neovascular tuft formation. It is important to note that the extent of vascular obliteration between all retinas was identical at P12, indicating that these cells did not prevent hyperoxia-induced vaso-obliteration, but rather accelerated subsequent revascularization of the normal retinal vascular plexuses. Injection of astrocytes did not lead to any adverse immune reactions or formation of intraocular tumors. In contrast, other cell types such as primary fibroblasts demonstrated no evidence of vascular rescue and injection of non-differentiated mouse embryonic stem cells resulted in massive inflammation and retinal toxicity due to extensive, uncontrolled growth of these cells (data not shown).

Intravitreally injected BM-MPCs target the retina in large numbers (Ritter et al. 2006). In contrast, when we injected primary astrocytes obtained from eGFP mice, very few of the injected astrocytes became localized to the retina following intravitreal injection. The majority of injected astrocytes remained in the vitreous (data not shown). Despite this, astrocytes effectively normalized the vascular phenotype in the C57Bl/6J OIR retinas at levels similar tothe retina-targeting BM-MPCs (Figure 4E), indicating that retinal targeting and direct cell-cell contact is not required for astrocyte-mediated rescue. This suggests that the injected astrocytes may secrete factors that mediate the observed rescue effects. To test this further, we injected serum-free media preconditioned by C8-D1A astrocytes and analyzed the effects on the retinal vasculature of C57BL/6J mice after OIR. Injection of astrocyte-conditioned serum-free media also resulted in vascular rescue as demonstrated by accelerated revascularization of the normal retinal plexuses and reduced intravitreal pathological NV. This effect was observed whether the astrocyte-conditioned media was injected at P7 (just prior to exposure to hyperoxia) or at P12 (upon return to normoxia) (Figure 4D). This result is consistent with our earlier findings demonstrating that astrocyte-conditioned media facilitated survival of both microglia and astrocytes in the vaso-obliterated zones of OIR retinas. These data further strengthen the correlation between astrocyte and microglial survival and normalized revascularization in the OIR model.

Injection of bFGF or VEGF rescues the OIR phenotype

To determine what secreted factor(s) might mediate the OIR rescue effect, we analyzed the serum-free conditioned media using LC-MS/MS proteomics. Several molecules were identified in the astrocyte-conditioned, serum-free media (Table 1). Multiple extracellular matrix molecules were found, along with cell adhesion molecules and various molecules involved in protein transport and folding. Several secreted molecules were also identified, including vascular-related molecules such as basic fibroblast growth factor (bFGF), vascular endothelial growth factor (VEGF), several other cytokine and signaling-related molecules, and molecules involved in coagulation or inflammation (Table 1). The presence of bFGF and VEGF in the conditioned media was confirmed by western blot analysis (Figure 5A).

Figure 5
Injection of small quantities of bFGF or VEGF can protect retinal astrocytes and rescue the retinal vasculature in the OIR model. A) bFGF and VEGF is secreted by astrocytes and is present in the astrocyte-conditioned, serum-free media. B-C) Injection ...
Table 1
Molecules Identified in the Astrocyte-Conditioned, Serum-Free Media

Both astrocytes (Araujo and Cotman 1992) and microglia (Liu et al. 1998) are known to express bFGF and astrocytic expression of bFGF can protect endothelial cells from oxygen-induced degeneration in the mouse OIR model (Yamada et al. 2000). However, we have shown rescue effects resulting from injection of astrocyte-conditioned media at P12, after the mice return from hyperoxia and after vascular obliteration is complete. To test the possibility that bFGF might participate in the astrocyte-mediated vascular rescue effect, varying levels of bFGF were intravitreally injected into OIR mice at P12. Low doses (10 or 30 ng per eye) of bFGF mediated accelerated revascularization of the obliterated areas in a dose dependent manner and inhibited neovascular tuft formation (Figure 5B). Injection of 30 ng bFGF per eye was optimal while injection of lower or higher doses of bFGF was less effective (Figure 5B).

VEGF was also injected at varying doses for comparison with bFGF injection. Astrocyte secretion of VEGF is known to mediate retinal vascular development (Gerhardt et al. 2003; Provis et al. 1997; Stone et al. 1995). However, overexpression of VEGF during the ischemic phase of OIR is also known to be a critical driving factor in the formation of pathology (McLeod et al. 2002; Smith et al. 1999; Werdich et al. 2004) and blocking VEGF activity has proven to be effective in preventing OIR-induced pathological neovascularization (Aiello et al. 1995; Ishida et al. 2003; Ng et al. 2006; Ozaki et al. 2000). Surprisingly, we found that injection of VEGF at relatively low levels actually rescued the OIR phenotype by enhancing revascularization and reducing areas of neovascular tufts (Figure 5C). Similar to injection of bFGF, injection of higher doses of VEGF no longer rescued the C57BL/6J OIR phenotype (Figure 5C). Injection of the optimal concentrations of either bFGF (30 ng/eye) or VEGF (50 ng/eye), were found to significantly increase survival of the retinal astrocytes within the obliterated vascular zone (Figure 5D-E), again demonstrating a correlation between astrocyte survival with accelerated revascularization and reduced OIR pathology.


Strong associations exist between retinal blood vessels and vascular associated cells such as astrocytes, microglia, and pericytes. Retinal astrocytes play a critical role during normal retinal vascular development by expressing VEGF and initiating endothelial cell proliferation (Provis et al. 1997; Stalmans et al. 2002; Stone et al. 1995). Endothelial cell guidance along the underlying astrocytic template during retinal vascular development is mediated by specific VEGF isoforms expressed by retinal astrocytes (Gerhardt et al. 2003; Stalmans et al. 2002) and specific R-cadherin mediated cell-cell adhesion (Dorrell et al. 2002; Dorrell and Friedlander 2006). It is likely that astrocytes also play an important role in promoting normal revascularization during ischemic injury (Chan-Ling and Stone 1992).

We have demonstrated that normalized revascularization of the vaso-obliterated retina in the mouse OIR model strongly correlates with the survival of astroglia and microglia during hypoxic insult. In each case, astrocyte and microglia survival within the vaso-obliterated, hypoxic retina correlates with accelerated revascularization of the normal retinal plexuses and reduced pathological intravitreal NV. Vascular degeneration in the central retinal vasculature is similar in both BALB/cByJ and C57Bl/6J mouse retinas after exposure to hyperoxia. Therefore, the neovascular drive is likely to be comparable in the two strains upon return to normoxia. However, in BALB/cByJ mice, the retinal astrocytes and retinal microglia survive within the vaso-obliterated areas of the retina following return to normoxia. Subsequently, normal retinal revascularization occurs much more quickly and without formation of pathological pre-retinal NV. In C57Bl/6J mice, the retinal astrocytes and retinal microglia disappear from the avascular, hypoxic areas of the retina and this is associated with slower recovery and formation of large pathological, pre-retinal vascular tufts.

The correlation between the survival of astrocytes within the hypoxic, vaso-obliterated areas after return to normoxia and subsequent revascularization in the mouse OIR model is also demonstrated by analyzing treatments that preserve the endogenous astrocytic template and subsequently normalize retinal revascularization. We hypothesize that in the absence of astrocytes, no template remains to guide blood vessel growth during the revascularization resulting in disorganized vessels and the formation of pathological, intravitreal NV. However, when astrocytes survive within the central vaso-obliterated retina, the remaining astrocytic template facilitates normalized revascularization of the characteristic retinal vascular plexuses. Endothelial filopodia can be observed extended along the astrocytic template (Figure 5F) during revascularization suggesting that the vessels are indeed identifying underlying guidance cues provided by the astrocytes. This facilitates similar retinal vascular guidance mechanisms to those that initially produce the characteristic vascular plexuses during development, a process known to be mediated by endothelial filopodia and astrocytic guidance cues (Dorrell et al. 2002; Fruttiger et al. 1996; Gerhardt et al. 2003). BM-MPCs, astrocytes and astrocyte-conditioned media all protect the endogenous astrocytes and subsequently promote normalized revascularization in C57BL/6J mouse retinas.

Proteomic analysis of the astrocyte-conditioned media was used to identify proteins that may mediate the rescue effect (Table 1). Both bFGF and VEGF were identified in the astrocyte-conditioned media (Figure 4) and are known to be secreted by retinal astrocytes and microglia (Scherer and Schnitzer 1994; Yamada et al. 2000). We have demonstrated that intravitreal injection of either bFGF or VEGF at low doses mimics the astrocyte and astrocyte-conditioned media-mediated rescue effects. It is unlikely that such therapy would be used to treat vascular-associated retinopathies since both bFGF and VEGF are known to contribute to neovascularization in these diseases. In fact, most therapies currently used to treat ischemic retinopathies are designed to inhibit their activity. However, our data demonstrates that at specific doses, these growth factors can actually have beneficial effects, presumably by protecting critical vascular-related glial cells from hypoxia-induced cell death. Our data suggests that it is important to consider possible detrimental effects that treatments blocking similar growth factors might have on vascular-related cells such as microglia and astrocytes. Inhibition of VEGF or bFGF certainly can block endothelial cell proliferation and reduce vascular permeability (Aiello et al. 1995; Caldwell et al. 2003; Witmer et al. 2003), but may also lead to loss of the important microglia and astrocytes as well and long-term repercussions. Thus, it may be important to identify other potential therapies to treat ischemic diseases of the eye.

Cell-based therapy may be one avenue by which ocular vascular diseases can be treated in the future. Recently, a population of Müller glia were found that can function as neuronal stem cells (Bernardos et al. 2007). Perhaps, in the future, these cells or other glial progenitors may be found that can replace lost astrocytes. In the meantime, cell therapy with astrocytes or myeloid progenitors may be a useful approach since these cells can react to the local environment, potentially secrete appropriate levels of growth factors important for normalized vascular regrowth, and lead to retinal vascular stabilization. Astrocyte injection rescues the vascular phenotype whereas other cell types such as fibroblasts or embryonic stem cells have no effect. However, we do not observe incorporation of the injected astrocytes into the retina. Thus, the use of astrocytes as a long-term cell therapy in cases of continuing ischemia may not be practical. Previous studies from our laboratory have demonstrated that intravitreal injection of bone marrow-derived myeloid progenitor cells (BM-MPCs) leads to accelerated and normalized revascularization, as well as inhibition of pathological NV following hyperoxia-induced vascular obliteration in the mouse OIR model (Ritter et al. 2006). We now demonstrate that this vascular rescue is preceded by BM-MPC mediated rescue of retinal astrocytes within the vascular obliterated zones of C57Bl/6J mouse retinas, suggesting a potential mechanism of action for the observed rescue activity. By preventing astrocyte cell death in the hypoxic zones, and re-populating the central retina with microglia, the BM-MPCs may help facilitate normalized revascularization of the central retina in C57/Bl-6 mice which would otherwise develop pathological NV. Myeloid progenitors incorporate into the retina and become functionally active microglia and thus may have continued, long-term beneficial effects on the survival of critical vascular-associated glial cells. This may provide a paradigm for long-term therapy which can normalize revascularization of the ischemic retina and alleviate problems associated with the underlying disease.


This work was supported by grants to MF from the National Eye Institute of the National Institutes of Health (R01EY11254 and R24EY017540), the MacTel Foundation and the V. Kann Rassmussen Foundation. MID was supported by a fellowship from the California Institute for Regenerative Medicine (CIRM).


  • Adamis AP, Aiello LP, D'Amato RA. Angiogenesis and ophthalmic disease. Angiogenesis. 1999;3(1):9–14. [PubMed]
  • Aiello LP, Pierce EA, Foley ED, Takagi H, Chen H, Riddle L, Ferrara N, King GL, Smith LE. Suppression of retinal neovascularization in vivo by inhibition of vascular endothelial growth factor (VEGF) using soluble VEGF-receptor chimeric proteins. Proc Natl Acad Sci U S A. 1995;92(23):10457–61. [PubMed]
  • Araujo DM, Cotman CW. Basic FGF in astroglial, microglial, and neuronal cultures: characterization of binding sites and modulation of release by lymphokines and trophic factors. J Neurosci. 1992;12(5):1668–78. [PubMed]
  • Avery RL, Pieramici DJ, Rabena MD, Castellarin AA, Nasir MA, Giust MJ. Intravitreal bevacizumab (Avastin) for neovascular age-related macular degeneration. Ophthalmology. 2006;113(3):363–372. e5. [PubMed]
  • Banin E, Dorrell MI, Aguilar E, Ritter MR, Aderman CM, Smith AC, Friedlander J, Friedlander M. T2-TrpRS inhibits preretinal neovascularization and enhances physiological vascular regrowth in OIR as assessed by a new method of quantification. Invest Ophthalmol Vis Sci. 2006;47(5):2125–34. [PubMed]
  • Bernardos RL, Barthel LK, Meyers JR, Raymond PA. Late-stage neuronal progenitors in the retina are radial Muller glia that function as retinal stem cells. J Neurosci. 2007;27(26):7028–40. [PubMed]
  • Caldwell RB, Bartoli M, Behzadian MA, El-Remessy AE, Al-Shabrawey M, Platt DH, Caldwell RW. Vascular endothelial growth factor and diabetic retinopathy: pathophysiological mechanisms and treatment perspectives. Diabetes Metab Res Rev. 2003;19(6):442–55. [PubMed]
  • Chan-Ling T, Stone J. Degeneration of astrocytes in feline retinopathy of prematurity causes failure of the blood-retinal barrier. Invest Ophthalmol Vis Sci. 1992;33(7):2148–59. [PubMed]
  • Das A, McGuire PG. Retinal and choroidal angiogenesis: pathophysiology and strategies for inhibition. Prog Retin Eye Res. 2003;22(6):721–48. [PubMed]
  • Dorrell MI, Aguilar E, Friedlander M. Retinal vascular development is mediated by endothelial filopodia, a preexisting astrocytic template and specific R-cadherin adhesion. Invest Ophthalmol Vis Sci. 2002;43(11):3500–10. [PubMed]
  • Dorrell MI, Aguilar E, Jacobson R, Yanes O, Gariano R, Heckenlively J, Banin E, Ramirez GA, Gasmi M, Bird A. Antioxidant or neurotrophic factor treatment preserves function in a mouse model of neovascularization-associated oxidative stress. J Clin Invest. 2009 others. [PMC free article] [PubMed]
  • Dorrell MI, Friedlander M. Mechanisms of endothelial cell guidance and vascular patterning in the developing mouse retina. Prog Retin Eye Res. 2006;25(3):277–95. [PubMed]
  • Dorrell MI, Otani A, Aguilar E, Moreno SK, Friedlander M. Adult bone marrow-derived stem cells utilize R-cadherin to target sites of neovascularization in the developing retina. Blood. 2004 [PubMed]
  • Eichler W, Yafai Y, Wiedemann P, Fengler D. Antineovascular agents in the treatment of eye diseases. Curr Pharm Des. 2006;12(21):2645–60. [PubMed]
  • Ferrara N, Kerbel RS. Angiogenesis as a therapeutic target. Nature. 2005;438(7070):967–74. [PubMed]
  • Friedlander M. Stem Cells and Retinal Diseases. In: Ryan SJH DR, Schachat AP, Wilkinson P, editors. Retina. 4th Edition Elsevier Publishers; 2005.
  • Friedlander M. Fibrosis and diseases of the eye. J Clin Invest. 2007;117(3):576–86. [PMC free article] [PubMed]
  • Friedlander M, Dorrell MI, Ritter MR, Marchetti V, Moreno SK, El-Kalay M, Bird AC, Banin E, Aguilar E. Progenitor cells and retinal angiogenesis. Angiogenesis. 2007;10(2):89–101. [PubMed]
  • Fruttiger M, Calver AR, Kruger WH, Mudhar HS, Michalovich D, Takakura N, Nishikawa S, Richardson WD. PDGF mediates a neuron-astrocyte interaction in the developing retina. Neuron. 1996;17(6):1117–31. [PubMed]
  • Gebicke-Haerter PJ, Bauer J, Schobert A, Northoff H. Lipopolysaccharide-free conditions in primary astrocyte cultures allow growth and isolation of microglial cells. J Neurosci. 1989;9(1):183–94. [PubMed]
  • Gerhardt H, Golding M, Fruttiger M, Ruhrberg C, Lundkvist A, Abramsson A, Jeltsch M, Mitchell C, Alitalo K, Shima D. VEGF guides angiogenic sprouting utilizing endothelial tip cell filopodia. J Cell Biol. 2003;161(6):1163–77. others. [PMC free article] [PubMed]
  • Ishida S, Usui T, Yamashiro K, Kaji Y, Amano S, Ogura Y, Hida T, Oguchi Y, Ambati J, Miller JW. VEGF164-mediated inflammation is required for pathological, but not physiological, ischemia-induced retinal neovascularization. J Exp Med. 2003;198(3):483–9. others. [PMC free article] [PubMed]
  • Keller A, Nesvizhskii AI, Kolker E, Aebersold R. Empirical statistical model to estimate the accuracy of peptide identifications made by MS/MS and database search. Anal Chem. 2002;74(20):5383–92. [PubMed]
  • Liu X, Mashour GA, Webster HF, Kurtz A. Basic FGF and FGF receptor 1 are expressed in microglia during experimental autoimmune encephalomyelitis: temporally distinct expression of midkine and pleiotrophin. Glia. 1998;24(4):390–7. [PubMed]
  • Madan A, Penn JS. Animal models of oxygen-induced retinopathy. Front Biosci. 2003;8:d1030–43. [PubMed]
  • Maddox DE, Shibata S, Goldstein IJ. Stimulated macrophages express a new glycoprotein receptor reactive with Griffonia simplicifolia I-B4 isolectin. Proc Natl Acad Sci U S A. 1982;79(1):166–70. [PubMed]
  • McLeod DS, D'Anna SA, Lutty GA. Clinical and histopathologic features of canine oxygen-induced proliferative retinopathy. Invest Ophthalmol Vis Sci. 1998;39(10):1918–32. [PubMed]
  • McLeod DS, Taomoto M, Cao J, Zhu Z, Witte L, Lutty GA. Localization of VEGF receptor-2 (KDR/Flk-1) and effects of blocking it in oxygen-induced retinopathy. Invest Ophthalmol Vis Sci. 2002;43(2):474–82. [PubMed]
  • Ng EW, Shima DT, Calias P, Cunningham ET, Jr., Guyer DR, Adamis AP. Pegaptanib, a targeted anti-VEGF aptamer for ocular vascular disease. Nat Rev Drug Discov. 2006;5(2):123–32. [PubMed]
  • Otani A, Dorrell MI, Kinder K, Moreno SK, Nusinowitz S, Banin E, Heckenlively J, Friedlander M. Rescue of retinal degeneration by intravitreally injected adult bone marrow-derived lineage-negative hematopoietic stem cells. J Clin Invest. 2004;114(6):765–74. [PMC free article] [PubMed]
  • Otani A, Friedlander M. Retinal vascular regeneration. Semin Ophthalmol. 2005;20(1):43–50. [PubMed]
  • Otani A, Kinder K, Ewalt K, Otero FJ, Schimmel P, Friedlander M. Bone marrow-derived stem cells target retinal astrocytes and can promote or inhibit retinal angiogenesis. Nat Med. 2002;8(9):1004–10. [PubMed]
  • Ozaki H, Seo MS, Ozaki K, Yamada H, Yamada E, Okamoto N, Hofmann F, Wood JM, Campochiaro PA. Blockade of vascular endothelial cell growth factor receptor signaling is sufficient to completely prevent retinal neovascularization. Am J Pathol. 2000;156(2):697–707. [PubMed]
  • Penn JS, Tolman BL, Lowery LA. Variable oxygen exposure causes preretinal neovascularization in the newborn rat. Invest Ophthalmol Vis Sci. 1993;34(3):576–85. [PubMed]
  • Provis JM, Leech J, Diaz CM, Penfold PL, Stone J, Keshet E. Development of the human retinal vasculature: cellular relations and VEGF expression. Exp Eye Res. 1997;65(4):555–68. [PubMed]
  • Provis JM, Sandercoe T, Hendrickson AE. Astrocytes and blood vessels define the foveal rim during primate retinal development. Invest Ophthalmol Vis Sci. 2000;41(10):2827–36. [PubMed]
  • Ritter MR, Banin E, Moreno SK, Aguilar E, Dorrell MI, Friedlander M. Myeloid progenitors differentiate into microglia and promote vascular repair in a model of ischemic retinopathy. J Clin Invest. 2006;116(12):3266–76. [PMC free article] [PubMed]
  • Robinson GS, Ju M, Shih SC, Xu X, McMahon G, Caldwell RB, Smith LE. Nonvascular role for VEGF: VEGFR-1, 2 activity is critical for neural retinal development. Faseb J. 2001;15(7):1215–7. [PubMed]
  • Rohan RM, Fernandez A, Udagawa T, Yuan J, D'Amato RJ. Genetic heterogeneity of angiogenesis in mice. Faseb J. 2000;14(7):871–6. [PubMed]
  • Rungger-Brandle E, Dosso AA, Leuenberger PM. Glial reactivity, an early feature of diabetic retinopathy. Invest Ophthalmol Vis Sci. 2000;41(7):1971–80. [PubMed]
  • Smith LE. Pathogenesis of retinopathy of prematurity. Acta Paediatr Suppl. 2002;91(437):26–8. [PubMed]
  • Smith LE, Shen W, Perruzzi C, Soker S, Kinose F, Xu X, Robinson G, Driver S, Bischoff J, Zhang B. Regulation of vascular endothelial growth factor-dependent retinal neovascularization by insulin-like growth factor-1 receptor. Nat Med. 1999;5(12):1390–5. others. [PubMed]
  • Smith LE, Wesolowski E, McLellan A, Kostyk SK, D'Amato R, Sullivan R, D'Amore PA. Oxygen-induced retinopathy in the mouse. Invest Ophthalmol Vis Sci. 1994;35(1):101–11. [PubMed]
  • Stalmans I, Ng YS, Rohan R, Fruttiger M, Bouche A, Yuce A, Fujisawa H, Hermans B, Shani M, Jansen S. Arteriolar and venular patterning in retinas of mice selectively expressing VEGF isoforms. J Clin Invest. 2002;109(3):327–36. others. [PMC free article] [PubMed]
  • Stone J, Itin A, Alon T, Pe'er J, Gnessin H, Chan-Ling T, Keshet E. Development of retinal vasculature is mediated by hypoxia-induced vascular endothelial growth factor (VEGF) expression by neuroglia. J Neurosci. 1995;15(7 Pt 1):4738–47. [PubMed]
  • Werdich XQ, McCollum GW, Rajaratnam VS, Penn JS. Variable oxygen and retinal VEGF levels: correlation with incidence and severity of pathology in a rat model of oxygen-induced retinopathy. Exp Eye Res. 2004;79(5):623–30. [PubMed]
  • Witmer AN, Vrensen GF, Van Noorden CJ, Schlingemann RO. Vascular endothelial growth factors and angiogenesis in eye disease. Prog Retin Eye Res. 2003;22(1):1–29. [PubMed]
  • Yamada H, Yamada E, Ando A, Seo MS, Esumi N, Okamoto N, Vinores M, LaRochelle W, Zack DJ, Campochiaro PA. Platelet-derived growth factor-A-induced retinal gliosis protects against ischemic retinopathy. Am J Pathol. 2000;156(2):477–87. [PubMed]
  • Zachary I. Neuroprotective role of vascular endothelial growth factor: signalling mechanisms, biological function, and therapeutic potential. Neurosignals. 2005;14(5):207–21. [PubMed]