In mammalian eyes, the pupillary membrane and hyaloid vessels, including the hyaloid artery, tunica vasculosa lentis, and vasa hyaloidea propria, nourish the immature lens, retina, and vitreous (Fig. 1). These vessels are known to regress during the later stages of ocular development, pre-natally in humans and within the first few weeks of post-natal development in rodents, to provide an optically clear path between the lens and retina (Ito and Yoshioka, 1999). The development of the retinal vasculature coincides with the start of the regression of the hyaloid vasculature (Zhu et al., 2000).
There are many critical molecules involved in the physiological regulation of blood vessel formation as well as regression (Saint-Geniez and D'Amore, 2004). Vascular endothelial growth factor (VEGF), for example, has been implicated in both the normal development and maintenance of the vasculature as well as in its dysfunction (Thurston and Gale, 2004; Ferrara and Davis-Smyth, 1997). Mice with only one copy of the VEGF gene die before birth because of blood vessel abnormalities, indicating that normal amounts of VEGF are absolutely required for normal vascular development (Carmeliet et al., 1996).
Vasculogenesis and angiogenesis are the two processes by which vascular development occurs. The retina is vascularized by a two-step process, initially by vasculogenesis, followed by angiogenesis (Flower et al., 1985; Chan-Ling et al., 1990; McLeod et al., 1987). Whether vessels form by vasculogenesis or angiogenesis, the primitive vessels are subsequently remodeled. Vascular remodeling, including the growth of new vessels and regression of others, is a complex process that involves several critical factors; it plays a major role in the early development of the vascular system (Yancopoulos et al., 2000).
Crystallins are the most abundant soluble proteins of the lens of the eye (Piatigorsky, 2002). Although highly specialized for lens, the crystallins are also expressed in other tissues (Srinivasan et al., 1992; Deretic et al., 1994; Head et al., 1995; Sinha et al., 1998; Magabo et al., 2000; Jones et al., 1999; Crabb et al., 2002; Xi et al., 2003). Three major families of crystallins, α, β, and γ, are ubiquitously represented in all vertebrates (Hejtmancik and Piatigorsky, 1994). There are also taxon-specific crystallins that serve as metabolic enzymes in other tissues. The two α-crystallins (αA and αB) belong to the small heat shock protein family of molecular chaperones (Sax and Piatigorsky, 1994; Horwitz, 1992). The β- and γ-crystallins are evolutionarily and structurally related members of a βγ superfamily, which also includes micro-organism stress proteins as well as vertebrate proteins that appear to be associated with processes of cell differentiation and morphological change (Clout et al., 1997; Ray et al., 1997). The non-lens functions of β- and γ-crystallins are yet to be fully determined.
Astrocytes have multiple functions that include regulation of blood vessel structure and function as well as involvement in pathologic processes (Nedergaard et al., 2003). While oligodendrocytes of the glial cell family have been shown to express crystallins, there has not been a report indicating expression of β and γ-crystallins by astrocytes (Dabir et al., 2004). The glial cells of the optic nerve contain the same cell types found in the white matter throughout the Central Nervous System (CNS) (Mi et al., 2001). Type-1 astrocytes develop in the embryonic optic nerve from astrocyte precursor cells (APC) (Raff et al., 1984; Miller et al., 1985, 1989; Mi and Barres, 1999). The majority of the APCs differentiate into GFAP-immunopositive astrocytes.
Astrocytes modulate the differentiation of vascular endothelial cells (Lattera et al., 1990) and contribute to the glial limitans that lines blood vessels (Stewart and Tuor, 1994; Rungger-Brandle et al., 1993). Astrocytes are also involved in the formation and preservation of the blood-brain and blood-retinal barriers (Janzer and Raff, 1987). It has been postulated that astrocytes guide and modulate vascular growth (Jiang et al., 1995). They migrate ahead of the vessels (vascular “front”), and are thus in a position to respond to local environmental signals (Provis, 2001). Although all of the vascular cells may participate in the remodeling process, the astrocytes may play a prominent part, because they are capable of sensing changes within their immediate milieu.
In a microarray analysis comparing gene expression in the retina of Nuc1 mutant rats (Sinha et al., 2005) with wildtype littermates, we found increased expression of a number of β- and γ-crystallin genes in Nuc1. These crystallins were subsequently found to be localized near vessels both in Nuc1, in which the hyaloid vasculature is abnormally retained, as well as in the wildtype retina. Specifically, the astrocytes at the vascular front express not only VEGF, as expected, but also β- and γ-crystallins. We further demonstrate that in human PFV (Persistent Fetal Vasculature) disease (Goldberg, 1997), in which the hyaloid vascular system also fails to regress normally, astrocytes similarly express β- and γ-crystallins, as well as VEGF. In an in-vitro model of chemical hypoxia, we show that astrocytes expressing VEGF also express β- and γ-crystallin. Our data suggest, for the first time, that β- and γ-crystallins may be involved in mediating vascular stabilization, remodeling, or survival in the developing mammalian eye.
Persistent Hyaloid Vascular System in Nuc1 Homozygous Rats
In the wild-type rats, the hyaloid artery, a constituent of the transient vasculature that nourishes the immature lens, retina, and vitreous, regressed by day 35 (Fig. 2a). In contrast, in Nuc1 homozygous rats, the hyaloid artery persisted at the surface of the optic nerve head projecting into the vitreous until adulthood (Fig. 2b,c). The pupillary membrane (PM), a temporary capillary network on the anterior surface of the lens, also known as the anterior tunica vasculosa lentis, normally regresses during the second week after birth. As shown in Figure 2d, in the normal eye, the PM was absent by post-natal day 25. However, it persisted in Nuc1 at the same post-natal stage (Fig. 2e, arrows). Nuc1 rats also displayed iris hyperplasia (Fig. 2e, arrowhead) and disrupted lens structure (Fig. 2e). By four months of age, the iris and ciliary body were dragged towards the center of the posterior chamber (Fig. 2g, arrow), inducing dragging and folding of the peripheral retina (Fig. 2g, arrowhead).
β- and γ-Crystallin Expression in the Retina and Hyaloid Vessels
To confirm microarray data (not shown) that indicated a more than twofold increase in the expression of β- and γ-crystallins in Nuc1 homozygous rat retina, we performed Real-time RT-PCR on 25-day-old retinas of Nuc1 homozygous (inclusive of the hyaloid artery) and wild-type Sprague Dawley rats. As shown in Figure 3a and b, mRNA levels of βA1/A3, βA4, βB1, βB3, γA, γB, γC, γD, and γEF crystallins were significantly increased in Nuc1 retina compared to the wild-type retina. Incorporation of 35S-amino acids by organ cultured 20-day retina also indicated a marked increase in crystallin synthesis in Nuc1 homozygotes compared to wild-type (Fig. 3c). Crystallin protein expression was confirmed by Western blotting (Fig. 3d).
The retained hyaloid artery and vasa hyaloidea propria in Nuc1 assumed a configuration typical of the struts of an umbrella or the guide ropes of a parachute (Fig. 4a). Distinct staining patterns with the crystallin antibodies were obtained for the retained hyaloid artery (Fig. 4a,c, arrow) and pupillary membrane (Fig. 4b) in Nuc1. When the γ-crystallin staining pattern (similar data were obtained with β-crystallin) in the retained hyaloid artery (Fig. 4c, arrow) was compared to the section stained with isolectin-B4 (Fig. 4d), it clearly indicated that crystallin was localized around the blood vessels [see merged picture with γ-crystallins and Isolectin-B4 (Fig. 4e). Isolectin B4 has been shown to label blood vessels in the rat (Ashwell et al., 1989).
The transient hyaloid artery at post-natal day 9 in normal (Fig. 5a) and Nuc1 homozygous rats (Fig. 5b) showed distinct crystallin expression (Fig. 5,a,b, red, arrows) surrounding the blood vessels (Fig. 5a,b, green, asterisk). In rats, the hyaloid artery starts involuting around post-natal day 7 and regresses completely by 3 weeks of post-natal development (Cairns, 1959). While the hyaloid vessels undergo involution, the retinal vessels continue growing and reach the adult form by the time the hyaloid vasculature has completely regressed. The vascularization of the retina is restricted to the inner part of the retina. At post-natal day 20, crystallin expression is also localized surrounding the retinal vessels in both normal (Fig. 5c) and Nuc1 homozygous rats (Fig. 5d).
VEGF and β- and γ-Crystallin Expression in GFAP-Immunopositive Astrocytes of the Retained Vasculature in Nuc1 Homozygote Rats
We further analyzed the cellular identity of crystallin-immunopositive cells in the retained hyaloid vasculature of the 5-week-old Nuc1 homozygote. Previous studies have shown that astrocytes migrate ahead of the vessels and are present on the external surface of the blood vessels (Stone et al., 1995). Therefore, we used antibodies to GFAP to localize the astrocytes in the retained hyaloid artery and retinal vessels of Nuc1 homozygotes. VEGF has also been shown previously to be expressed by astrocytes (Wechsler-Reya and Barres, 1997) and has an important function in vascular remodeling. Our data demonstrate VEGF immunopositive reactivity in the retained hyaloid tissue of Nuc1 (Fig. 6a), co-localizing with GFAP (Fig. 6b,c). Co-localization of GFAP and γ-crystallin reactivity (Fig. 6d,f) clearly demonstrated that astrocytes in the retained hyaloid of Nuc1 also expressed crystallins.
In the Nuc1 homozygous retina, confocal microscopy with crystallin specific antibodies also showed expression at the internal limiting membrane (Fig. 7b) and in ganglion cells (Fig. 7b, arrowheads). GFAP immunopositive staining was also present in the internal limiting membrane of the Nuc1 retina but not in the ganglion cells (Fig. 7a). Co-localization of crystallins and GFAP (Fig. 7c) showed that GFAP+ cells associated with vessels at the inner limiting membrane also expressed crystallins. VEGF expression was localized primarily at the internal limiting membrane (Fig. 7e). A similar pattern of staining was observed with antibodies to GFAP (Fig. 7d) and co-localization indicated that GFAP+ astrocytes do express VEGF (Fig. 7f, arrows).
VEGF and β- and γ-Crystallin Expression in Human PFV
Failure of regression of the hyaloid vasculature in the human eye leads to persistent fetal vasculature (PFV) disease. The ocular phenotypes of Nuc1 and human PFV are similar. To determine if β- and γ-crystallins are also expressed in the GFAP+ astrocytes of the retained hyaloid tissue of PFV patients, we used a similar approach as indicated above for Nuc1. We immunolocalized VEGF, GFAP, and crystallins from human PFV patients, on serial paraffin sections, all of which showed a positive staining pattern comparable to our Nuc1 rat data. Figure 8 shows a representative PFV tissue section (one of five patient eyes examined) from a female who died six days after birth with a diagnosis of trisomy 13 (Patau's syndrome) with microphthalmia and multiple congenital malformations, including PFV. There was a persistent pupillary membrane, and the lens showed cataractous changes with posterior distortion (not shown). Firmly attached to the capsule was a mesenchymal fibrovascular tissue, including the well-preserved hyaloid system, as shown in part in Figure 8a and at higher magnification in Figure 8b. Within that tissue, many well-differentiated rosettes were present (Fig. 8a, short arrows).
Immunohistochemical study showed that cells surrounding the hyaloid tissue were positive with VEGF (Fig. 8c and d), GFAP (Fig. 8e), and with crystallin antibodies (Fig. 8f), clearly indicating an expression pattern of VEGF and β, γ-crystallins by GFAP+ astrocytes similar to that seen in our Nuc1 model.
Induction of VEGF and β- and γ-Crystallins in Cultured Human Astrocytes Exposed to 3-Nitropropionic Acid
It has been shown that astrocytes are sensitive to hypoxia, which induces them to release VEGF (Chow et al., 2001; Sandercoe et al., 2003). To determine if astrocytes also express β- and γ-crystallins in response to hypoxia, we used 3-nitropropionic acid, an irreversible inhibitor of mitochondrial succinate dehydrogenase activity (Cavaliere et al., 2001). It can be considered a model substance to study hypoxic neuronal damage (Riepe et al., 1996). As shown in Figure 9, when fetal human astrocytes in culture were exposed to 3-NP, there was induction not only of VEGF within 6 hr, as expected, but also β- and γ-crystallins. Strong cytoplasmic staining was demonstrated with both VEGF and β-crystallin antibodies (γ-crystallin also showed a similar staining pattern); nuclei were counterstained with DAPI. Control cells showed little or no VEGF or β, γ-crystallin staining.
Programmed hyaloid vascular regression is an essential element of the development of the eye; apoptosis plays an important role in this process. Vascular regression normally occurs within 3 weeks after birth in rats (Latker and Kuwabara, 1981). In Nuc1, the programmed regression of the vascular network is inhibited. We have shown that the Nuc1 mutation also prevents the normal programmed loss of nuclei from lens fiber cells and affects the regulation of cell numbers and maturation of retinal neurons (Sinha et al., 2005). At present, the mechanistic relationship among these various phenotypes is unclear. However, it appears that Nuc1 may be an eye-specific regulator of apoptosis, since the mutation has shown no obvious effects outside of the eye. Even homozygous mutants appear grossly normal, except for their eyes, and both males and females are able to reproduce. Nuc1 may govern the maturation of several ocular tissues. The only other animal model that shows similar features is the Apaf-1 (CED-4 homolog) null mouse (Cecconi et al., 1998). That model similarly shows alteration in lens development and abnormal cell number regulation in the retina; it inhibits hyaloid regression, but also exhibits delayed inter-digital mesenchymal cell death, which is not seen in Nuc1.
Several mouse models with persistent hyaloid vessels have been reported (Saint-Geniez and D'Amore, 2004). In these models, the hyaloid vessels might persist to compensate for the absence or lack of retinal vessels. In contrast, the Nuc1 mutant rats exhibited the initial network of retinal vessels but with a possible abnormality in the remodeling process (Gehlbach et al., unpublished data). Further, Nuc1 exhibits inhibition of the normal regression of the entire fetal intraocular vasculature and not just part of it, as reported in those mouse models.
Another factor believed to be important in the regression of the hyaloid vascular system is the macrophage population. Persistent vasculature was reported in transgenic mice where macrophages were disrupted by directed diphtheria toxin expression, using macrophage-specific promoter elements (Lang and Bishop, 1993). That study suggested that macrophages are required for normal hyaloid vessel regression. Our studies do indicate that macrophages may play an essential role in the pathophysiology of Nuc1 (Hose et al., 2005), although we have no evidence to suggest direct involvement in the persistence of the fetal vasculature.
Crystallins, initially regarded as lens-specific structural proteins, now are thought to be multifunctional proteins with physiological roles in non-lens tissues as well. A recent study has shown that αA and αB-crystallins function as distinct anti-apoptotic regulators (Mao et al., 2004). Several other laboratories have also reported the possible involvement of α-crystallins in cellular apoptosis (Mehlen et al., 1996; Golenhofen et al., 1999; Hoover et al., 2000; Andley et al., 2000; Mo et al., 2001; Kamradt et al., 2001; Bruey et al., 2000). While our microarray data (not shown) showed upregulation of αB-crystallin, a ubiquitously expressed small heat shock protein known to be induced by stress, it also showed a marked upregulation of β, γ-crystallins. Real-time RT PCR and Western blots confirmed this result. Interestingly, our studies showed that while many members of the β- and γ-crystallin family are upregulated in Nuc1, they are also differentially regulated. Moreover, metabolic labeling studies with 35S amino acids indicated that newly synthesized crystallin proteins were increased in organ-cultured Nuc1 retina relative to wildtype retina. Double-labeling experiments with isolectin B4 and β- and γ-crystallin antibodies of the retained hyaloid artery in Nuc1 clearly showed that β-and γ-crystallin expression is concentrated surrounding the vasculature. During normal development, β and γ-crystallin expression was also associated with the hyaloid artery as shown in the 9-day wildtype eye. Moreover, the retinal vasculature in both normal and Nuc1 rats showed crystallin expression surrounding the blood vessels. It is possible that β- and γ-crystallin have a role in vascular development and/or survival.
Several investigators have analyzed VEGF expression during normal and pathological vascular development. Our studies demonstrate that the retained hyaloid artery in Nuc1 express VEGF as well as β-and γ-crystallin. We localized both crystallin and VEGF expression in the retained hyaloid artery in Nuc1 to the GFAP-positive astrocytes. Astrocytes are now regarded as multifunctional housekeeping cells that interact with the vasculature to form a gliovascular network. While, expression of both VEGF and crystallins was seen in the GFAP-positive astrocytes of the inner limiting membrane of the Nuc1 retina, crystallin expression was evident in other parts of the retina including the ganglion cells. Crystallins have been shown by others to be expressed in the retina (Deretic et al., 1994; Sinha et al., 1998; Jones et al., 1999; Xi et al., 2003); however, the function of crystallins in the neural retina remains to be determined. We have recently shown that in Nuc1 the retina is thicker than normal and shows reduced programmed cell death during development (Sinha et al., 2005). Others have shown that the formation of the retinal vasculature is associated with the thickening of the retina (Dreher et al., 1992). Crystallin production may be upregulated in the Nuc1 retina as a result of increased vascularization, to meet the higher metabolic demand of the abnormally thickened tissue.
Apoptosis plays a major role in the remodeling of organs. We have recently shown that in Nuc1, the normal apoptotic-like process in the lens and retina appeared to be inhibited (Sinha et al., 2005). It is possible that β- and γ-crystallins synthesized by the astrocytes foster vessel survival or stabilization and thus inhibit regression of the hyaloid system. The vasculature is capable of sensing changes within its milieu and remodels itself, when necessary, through local production of mediators that influence structure as well as function. Vascular remodeling is an active process of structural alterations that may subsequently contribute to the pathophysiology of vascular diseases (Gibbons and Dzau, 1994).
Retention of fetal vasculature in humans is characteristic of PFV disease (Goldberg, 1997). We provide evidence that GFAP-immunopositive astrocytes in the retained vessels of PFV patients express VEGF and also β-, γ-crystallins. Thus, increased expression of VEGF by GFAP+ astrocytes in the retained hyaloid tissue, a finding not previously reported, suggests VEGF-mediated remodeling of the vessels in PFV. Moreover, expression of the crystallins by the same astrocytes suggests that crystallins may also participate in the vascular remodeling process.
It has been postulated that “physiological hypoxia” is required for formation and survival of the transient embryonic vasculature of the eye (Chow et al., 2001). Physiological levels of hypoxia are also the stimulus for normal development of the retinal vasculature (Chan-Ling and Stone, 1993). It is possible that a defect in hyaloid vascular regression in both Nuc1 and human PFV may lead to increased VEGF and β, γ-crystallin production due to a hypoxic environment. To determine whether hypoxia could induce crystallin synthesis in astrocytes, we cultured human astrocytes and exposed the cells to 3-nitropropionic acid (3-NP), a substance shown to induce neuronal hypoxia (Chauhan et al., 2003). Interestingly, the astrocytes expressed high levels of VEGF and β-, γ-crystallins within 6 hr of exposure to 3-NP. Indeed, it has also been shown that astrocytes respond to the changing physiological levels of hypoxia during development and secrete VEGF. Although VEGF is known to be regulated by hypoxia, there is as yet no other evidence that hypoxia regulates β- and γ-crystallins. However, the presence of a hypoxia-response element (HRE) in the promoter of γs (unpublished observation), a γ-crystallin family member that is stress inducible in the retina (Sinha et al., 1998), raises the possibility that β- and γ-crystallins could also be regulated by hypoxia.
In conclusion, it is tempting to speculate that β, γ-crystallins may function in mediating blood vessel remodeling and/or survival in the developing eye. While cellular functions of α-crystallins in non-lens tissues have been established, the function of β/γ crystallins remains unclear. The possibility that they may have a role in vascular remodeling is important, because such remodeling is fundamental to normal ocular development and to the pathogenesis of numerous diseases. Any influence that crystallins may have on such processes could have potential clinical importance.
Experiments were performed using postnatal Nuc1 and wild-type Sprague Dawley rats, as described earlier (Sinha et al., 2005) in accordance with the Guide for the Care and Use of Laboratory Animals (National Academy Press). Experiments involving normal and PFV human tissue conformed to the guidelines set forth in the Declaration of Helsinki for the use of human tissue in research. Post-mortem samples were fixed in 10% formalin, and 5-μm sections were cut and either stained with PAS or processed for immunofluorescence.
Real Time RT- PCR
Real time RT- PCR was used to determine the expression of β- and γ-crystallins in wild type and Nuc1 homozygote retinas. Total RNA from samples was reverse transcribed using SuperScript II Reverse Transcriptase (Invitrogen, La Jolla, CA). For Real-time PCR analysis, LightCycler FastStart DNA Master SYBR Green kit (Roche Diagnostics) and the Light Cycler from Roche Diagnostics were used. Primer sets for β- and γ-crystallin family members were taken from published sequences and are available upon request. Since β- and γ-crystallin family members share close homology, we selected non-homologous regions using the ComAlign software. Primer pairs based on the respective rat sequences of each crystallin gene from the Ensembl database were then designed using the primer 3 software. Hypoxanthine PhosphoRibosyl Transferase (HPRT) was used as an internal control. SYBR green was incorporated into the reaction mixture to facilitate measurement of product. The integrity of PCR product was verified by melting curve analysis. Real-time PCR values were determined by reference to a standard curve that was generated by Real-time PCR amplification of serially diluted cDNAs using β- and γ-crystallin and HPRT primers. Values obtained for levels of β- and γ-crystallins were normalized to the levels of HPRT mRNA.
To determine if new synthesis of crystallins in Nuc1 homozygotes was different from that in the wild type, 20-day-old retinas in organ culture were incubated with 200 μci of 35S-labeled amino-acids (Easy Tag Protein labeling mix, Perkin Elmer Life Sciences, Oak Brook, IL) for 3 hr. Retinal samples were then rinsed and homogenized in 20 mM Tris (pH 7.1) containing protein inhibitors (Roche) in a loose-fitting plastic tissue grinder to disrupt cells, but not to disrupt the nuclei. The samples were centrifuged at maximum speed in an Eppendorf microfuge for 15 min to remove debris and cell nuclei. To each supernatant was added a few microliters of DNase (Roche) to digest any remaining DNA. After 30 min of incubation at room temperature, SDS sample buffer with reducing agent (Novex) was added. The samples were placed in a boiling water bath for 2 min and then loaded on a Nu-PAGE (4–12%) Bis-Tris gradient gel (Invitrogen). Gels were stained with Coomassie brilliant blue, dried, and autoradiographed using Kodak Biomax film.
SDS-PAGE and Western Blot Analysis
The eyes were enucleated from 20-day-old wild type and Nuc1 homozygous rats after euthanization. The retina from each eye was dissected and rinsed in PBS and homogenized in SDS sample preparation buffer. After the supernatant fractions were heated in a boiling waterbath for 2 min, approximately 100 μg protein from each preparation was loaded on 4–12% Bis-Tris Nu-PAGE gels (Invitrogen). The gels were stained with Coomassie brilliant blue. For Western blotting, proteins were transferred to Nitrocellulose membranes (Bio-Rad Laboratories, Richmond, CA), blocked with 10% milk diluent, and incubated with the primary antibody overnight at 4°C. HRP-conjugated secondary antibodies and 4-CN substrate (Kirkegaard and Perry Laboratories) were used for visualization. A cocktail containing β- and γ-crystallin antibodies, each at 1:800 dilution was used. The antisera were raised in rabbits using calf β- or γ-crystallin protein as antigen.
The primary antibodies used in this study included rabbit polyclonal antibodies to β and γ-crystallin (1:500), VEGF (Santa Cruz, sc-152; 1:100), GFAP (glial fibrillary acidic protein) (Dako; 1:1,000) for single labeling and the mouse monoclonal GFAP (Santa Cruz; 1:200) for double labeling. Frozen sections were incubated with primary antibodies overnight at 4°C, washed with PBS, and incubated with donkey anti-rabbit secondary antibodies conjugated to either Cy-2 or Cy-3 (Jackson ImmunoRes, West Grove, PA, 1:200) for 1 hr at room temperature. Crystallin antibodies were also immunoabsorbed with respective crystallin proteins and used as an additional control in the present study. For double labeling with biotinylated isolectin B4 (Sigma, St. Louis, MO) and crystallin, streptavidin-Cy2 and Cy3 conjugated secondary antibodies (Jackson ImmunoRes, 1:200) were used. For double labeling with other primary antibodies (two primary antibodies from two different species), Cy2 or Cy3 conjugated secondary antibodies were used. The sections were finally counterstained with Hoechst and mounted with DAKO fluorescent mounting medium. For visualization of blood vessels, rats were anesthetized and perfused with PBS containing 50 mg/ml of fluorescein-labeled dextran (average molecular weight 500,000; Sigma, St. Louis, MO) as previously described (Tobe et al., 1998). The eyes were removed, immersed in OCT compound without fixation, and sectioned. The 7-μm sections were then immunolabeled with either β- or γ-crystallin antibodies (1:1,000 dilution). Fluorescent digital images were taken with a Zeiss microscope (Axioskop II). Confocal microscopy was done on Zeiss LSM 510.
Culture of Human Astrocytes and 3-Nitropropionic Acid Treatment
The SVG cell line used in this study was derived from human fetal astrocytes transformed with SV40 large T antigen (Major et al., 1985; Tornatore et al., 1996). Cultures of the human fetal astrocyte cell line (SVG) were maintained in Dulbecco's modified Eagle's medium with 2 mM L-glutamine, 10% fetal bovine serum, and streptomycin-penicillin-fungizone solutions. Cells grown in 60-mm dishes were treated with 10 μM 3-nitropropionic acid (3NP) for 6, 12, and 24 hr at 37°C and then fixed with paraformaldehyde (4%) for 15 min and blocked with 3% BSA at 4°C overnight. The fixed cells were processed following standard techniques for immunofluorescence using VEGF and crystallin antibodies as indicated earlier.
This work was supported in part by Helena Rubeinstein Foundation (to D.S.), a pediatric ophthalmology research grant from Knights Templar Eye Foundation, Inc. (to D.S.), Juvenile Diabetes Research Foundation International (to P.G.), Alexander and Margaret Stewart Trust (to P.G.), NIH KO8EY13420 (to P.G.), Guerrieri Retinal Research Fund (to M.F.G.), and Research to Prevent Blindness (an unrestricted grant to Wilmer Eye Institute). We thank Drs. Peggy Zelenka, Paul Russell, Panagiotis Tsonis, and Gerard Lutty for critical reading and discussion regarding this manuscript. We are grateful to Dr. Ashok Chauhan for his help and support with the cell culture studies and Ms. Rhonda Grebe of the Wilmer Confocal facility.