Biochemical Alterations in the Retinas of Very Low-Density Lipoprotein Receptor Knockout Mice: An Animal Model of Retinal Angiomatous Proliferation

Objective  To identify and characterize biochemical alterations in the retinas of very low-density lipoprotein receptor (VLDLr) knockout mice in an animal model of retinal angiomatous proliferation.

Methods  Immunohistochemical analysis, Western blot analysis, reverse transcriptase–polymerase chain reaction, and electrophoretic mobility shift assay were used to identify and characterize the altered gene and protein expression as well as signal cascades involved in the pathogenesis of neovascularization in the retinas of VLDLr mice.

Results  Expression of the angiogenic factors vascular endothelial growth factor and basic fibroblast growth factor was significantly greater in the lesion area, and Müller cells around the lesion area were activated, as indicated by increased expression of glial fibrillary acidic protein. Expression of the proinflammatory cytokine IL-18 (interleukin 18) and the inflammation mediator intercellular adhesion molecule-1 was increased before significant intraretinal neovascularization. Furthermore, phosphorylation of Akt and mitogen-activated protein kinase and translocalization of nuclear factor kappa B were greater in VLDLr knockout mouse retinas.

Conclusion  An inflammatory process is involved in the development of neovascularization in the VLDLr knockout mouse retina.

Clinical Relevance  Understanding the molecular mechanisms underlying these biochemical alterations in the retinas of VLDLr knockout mice will provide a foundation for developing novel therapeutic approaches to retinal angiomatous proliferation.

In the past decade, a new form of exudative age-related macular degeneration (AMD), retinal angiomatous proliferation (RAP), has been described. Retinal angiomatous proliferation is a neovascularization that begins in the retina and extends through the subretinal space, eventually communicating with choroidal neovascularization.^1-3 The pathologic vascular changes in RAP were first described as deep retinal vascular anomalous complex⁴ and then as occult retinal choroidal anastomosis,⁵ and now RAP is the most common term for this lesion. It comprises 10% to 15% of neovascular AMD cases.⁶ The neovasculature may become associated with a pigment epithelial detachment and eventually with choroidal neovascularization to form a retinal-choroidal anastomosis.¹ The etiology and the biochemical mechanisms responsible for RAP are still unknown, and the efficacy of any treatment has not yet been established.^1-3,7

Recently a mouse model of subretinal neovascularization with retinal-choroidal anastomosis was described in a gene-targeted mutant mouse homozygous for a disruption of the gene encoding the very low-density lipoprotein receptor (VLDLr). Normal and heterozygous mice do not have this phenotype.⁸ Very low-density lipoprotein receptor is a receptor for several apolipoprotein E–containing lipoprotein ligands⁹ and may facilitate the binding of triglyceride-rich lipoproteins in the capillary bed, leading to the subsequent delivery of triglyceride-derived free fatty acids to the underlying tissues active in free fatty acid metabolism.¹⁰ Deficiency of VLDLr leads to a hypertriglyceridemic phenotype in mice under conditions of dietary stress.¹¹ The morphologic alterations in the retina of the VLDLr knockout mouse are similar to those of RAP in humans in terms of the neovascularization starting in the outer plexiform layer, progressing to the subretinal space, and then anastomosing with choroidal vessels.⁶ To study the molecular and cellular mechanisms underlying this pathogenic process, and the role of the VLDLr in the regulation of intraretinal neovascularization, multidisciplinary biochemical approaches, including Western blot analysis, reverse transcriptase–polymerase chain reaction (RT-PCR), and immunohistochemical analysis, were used to examine the altered molecules and signal cascades involved in inflammatory and angiogenic responses. The balance¹² between vascular endothelial growth factor (VEGF) (angiogenesis) and pigment epithelium–derived factor (PEDF) (antiangiogenesis) in the retina has been reported to play a crucial role in the pathogenesis of retinal neovascularization in many eye diseases.^13,14 It is known that the expression and secretion of VEGF and PEDF are increased when Müller cells¹⁵ are activated. Glial fibrillary acidic protein (GFAP) is the marker of activated Müller cells. The purpose of this study was to investigate the effect of VLDLr knockout on the expression of VEGF, PEDF, intercellular adhesion molecule-1 (ICAM-1), basic fibroblast growth factor (bFGF), and GFAP and the activation of the Akt/mitogen-activated protein kinase (MAPK)/nuclear factor kappa B (NF-κB) signal pathway.

The VLDLr⁻/⁻ mice were purchased from Jackson Laboratory (Bar Harbor, Me). All the animals were born and raised in a 12-hour-on and 12-hour-off bright cyclic light environment at an average illumination of 50 lux. For genotype analysis, we used the Jackson Laboratory protocol (http://www.jax.org). The VLDLr⁻/⁻ descendents were used as the experimental group and the VLDLr^+/+ descendents were used as controls.

Histologic analyses were performed as described previously.^16-18 For the leakage study, mice were deeply anesthetized using a ketamine-xylazine mixture and were perfused through the left ventricle using a fluorescein-dextran stock solution, 50 mg/mL (Sigma-Aldrich Corp, St Louis, Mo), at a 0.03-mL/g body weight concentration. Animals were killed 1 minute after perfusion. The eyes were enucleated and fixed in 4% paraformaldehyde for at least 3 hours. The corneas and lenses were then removed, and the peripheral retinas were dissected and flat-mounted on microscope slides for examination under a fluorescence microscope.

Immunohistochemical staining was performed as described previously.¹⁹ Briefly, the eye was enucleated and then fixed with 4% paraformaldehyde in phosphate-buffered saline for 4 hours, then dissected to remove the corner and vitreous to get eyecup. Sections were incubated with rabbit anti-VEGF (1:250), bFGF (1:50), PEDF (1:50), and protein kinase C (1:100) polyclonal antibodies or mouse anti-GFAP (1:600) and rhodopsin (1:1000) monoclonal antibodies. The anti-VEGF, bFGF, and PEDF were purchased from Santa Cruz Biotechnology (Santa Cruz, Calif). The anti-GFAP was from Chemicon (Temecula, Calif). Anti–protein kinase C was obtained from Sigma-Aldrich Corp. Anti-rhodopsin (R1D4) was provided by Robert Molday, PhD (University of British Columbia, Vancouver). The secondary anti–rabbit or anti–mouse antibodies were labeled with correspondent fluorescein to view the detected antigens. Control sections were treated in the same way with omission of primary antibody or with rabbit IgG. The sections were finally mounted using fluorescent mounting media with 4′-6-diamidino-2-phenylindole and were viewed using fluorescence microscopy.

The cornea and the lens were removed from the eyes. The inner layers of the retinas (neural retina) were separated from the eyecups. Neural retinas were lysed, then run on sodium dodecyl sulfate–polyacrylamide gel electrophoresis. After transferring, blots were incubated with polyclonal anti-VEGF, bFGF, PEDF, MAPK, Akt, and pAkt (Ser473) or monoclonal anti–p44/42MAPK antibodies. The anti–MAPK, Akt, and pAkt (Ser473) and p44/42MAPK antibodies were from Cell Signaling Technology Inc (Danvers, Mass). Western blot analyses were performed as described previously.²⁰ All the primary and secondary antibodies were diluted in 0.1% Tween-20 in Tris-buffered saline with 2.5% nonfat dry milk.

Electrophoretic mobility shift assays were performed as described in a previous publication.²⁰ Semiquantitative RT-PCR analyses were performed according to previous publications.^21,22

Results are expressed as mean ± SD. Differences were assessed using 1-way analysis of variance and the t test. A P<.05 was considered significant.

Histologic evaluation by means of light microscopy was performed on paraffin-embedded retinal sections stained with hematoxylin-eosin. The retinas of adult wild-type mice (Figure 1A) and 12- to 14-day-old VLDLr knockout mice (Figure 1B) were normal. The accumulation of cell debris in the subretinal space and the thickening of retinal pigment epithelial cells were observed before the development of intraretinal neovascularization in the retinas of VLDLr knockout mice. New vessels appearing in the outer plexiform layer were observed in 3-week-old VLDLr knockout mice (Figure 1C). These vessels migrated into the subretinal space with a disorganized inner nuclear layer by 4 weeks of age (Figure 1D). Between 6 and 8 weeks of age, the outer and the inner segments were disrupted in the lesion area, and migration of retinal pigment epithelial cells into the lesion area was also observed. In addition, retinal-choroidal anastomoses were clearly observed by 6 to 8 weeks of age (Figure 1E). The regression of neovascularization and the morphologic alterations mentioned previously herein were seen at 6 months, with significant loss of photoreceptors as indicated by a reduction in the thickness of the outer nuclear layer (ONL) (Figure 1F).

Angiographic analysis with fluorescein-dextran perfusion showed obvious leakage in the retinas of 6-week-old VLDLr knockout mice, and this neovasculature was usually localized in the ONL and in the subretinal space (Figure 2). Leakage spots were mainly distributed in the central area of the retina near the optic nerve (Figure 2C). Red blood cells were often found surrounding the leakage spot. No leakage spots were found in the retinas of wild-type mice.

To investigate the change in VEGF in the process of pathologic neovascularization in VLDLr knockout mice, immunoperoxidase staining was conducted to detect the expression of VEGF. A strong VEGF-positive signal in the lesion area was shown in knockout mouse retinas, whereas no clear VEGF staining was observed in wild-type mouse retinas (Figure 3A-D). Semiquantitative RT-PCR analysis showed that VEGF messenger RNA expression was greater in the retinas of 4-week-old VLDLr knockout mice (Figure 3E). The data from Western blotting showed that VEGF protein expression was significantly greater in the retinas of 4-week-old VLDLr knockout mice compared with aged-matched wild-type mice (Figure 3F and G).

Glial fibrillary acidic protein, a marker for activation of Müller cells in the retina, was barely detectable in the retinas of wild-type mice; however, clear vertical thread-like GFAP-positive staining was observed in the retinal lesion area as early as 3 weeks of age in VLDLr knockout mice (Figure 4A). By 4 weeks of age, GFAP-positive staining was also observed on the ONL, which may be the end-feet of Müller cells (Figure 4B). By 8 weeks of age, GFAP staining was even observed in the lesion area in the subretinal space (Figure 4C). The positive staining observed around blood vessels may be activated astrocytes.

Immunohistochemical analysis and Western blot analysis were used to compare the levels of bFGF in wild-type and VLDLr knockout mice. Strong bFGF staining was observed in the lesion area of the VLDLr mouse retina at age 4 weeks. The expression was mostly concentrated in the ONL, with some bFGF-positive cells in the inner nuclear layer (Figure 5A). Immunoblotting showed significantly greater expression of bFGF in the retina of VLDLr knockout mice compared with wild-type mice (Figure 5D and E).

Expression of PEDF was compared in wild-type and VLDLr mutant retinas. Strong PEDF-positive staining was observed in the inner segment and ganglion cells in the retinas of wild-type mice (Figure 6A). The PEDF-positive staining was also observed in retinal Müller cells of wild-type mice. In VLDLr knockout mice, the PEDF distribution pattern was similar to that in wild-type mice, except in the lesion area, where the inner segment was disrupted (Figure 6C). We found no significant difference in the expression of PEDF in retinas between VLDLr knockout and wild-type mice (Figure 6G and H). Double immunofluorescent staining for PEDF and GFAP demonstrated that the PEDF-positive signal overlapped the GFAP-positive signal in some cells in the lesion area (Figure 6D and F).

To investigate the signaling pathways involved in the pathogenesis of angiogenesis in the retinas of VLDLr knockout mice, the activation of Akt and the MAPK signaling pathway was studied using Western blot analysis. Phosphorylated Akt levels were significantly greater in the retinas of 3-week-old VLDLr knockout mice, whereas total Akt protein levels were not changed compared with those of wild-type mice (Figure 7A and B). Similarly, phosphorylated MAPK 44/42 levels were significantly greater in the retinas of 3-week-old VLDLr knockout mice without alteration in the level of total MAPK (Figure 7C and D). Furthermore, the retinas of 3-week-old VLDLr knockout mice demonstrated significant increases in the translocation of NF-κB from cytoplasm to the nucleus as measured by means of electrophoretic mobility shift assay (Figure 7E and F).

To examine ICAM-1 gene and protein expression in the retinas of VLDLr knockout mice, RT-PCR and Western blot analyses were performed. Expression of the ICAM-1 gene and protein was significantly greater in the retinas of 3-week-old VLDLr knockout mice (Figure 8A-C). However, the expression of vascular cell adhesion molecule-1 in the retinas of VLDLr knockout mice was not changed (data not shown).

To test the possibility of whether the expression of IL-18 (interleukin 18) is altered in the retinas of VLDLr knockout mice, expression of the IL-18 gene and protein was examined in the retinas of 3-week-old wild-type and VLDLr knockout mice by means of RT-PCR and Western blot analysis. The IL-18 gene expression was greater in the retinas of VLDLr knockout mice (Figure 8D-F).

New vessel formation is a concert of various angiogenetic stimuli and inhibitors that interact with endothelial cells. Among the angiogenetic factors, VEGF is the essential growth factor for the formation of new blood vessels if overexpressed.²³ The link between VEGF levels and neovascularization has been observed in experimental models of neovascularization for the cornea, iris, retina, and choroids.²⁴ Moreover, when VEGF is selectively inhibited in these models, blood vessel growth in these tissues is suppressed.⁶ Vascular endothelial growth factor is a major regulator of blood vessel formation and function. It controls several processes in endothelial cells, such as proliferation, survival, and migration.²⁵ To initiate angiogenesis, the balance between the positive and negative regulators is likely to be shifted such that mitogenic factors are enhanced or inhibitory factors are decreased. Enhanced VEGF levels were consistently correlated with the stimulation of neovascularization in the retina. Up-regulation of VEGF was detected by Lara et al²⁶ using RT-PCR. In the present study, the gene and protein expression of VEGF was increased in the retinas of VLDLr knockout mice at age 4 weeks as detected by RT-PCR and Western blot analysis. This indicates that VEGF may play a crucial role in retina neovascularization in VLDLr knockout mice. We do not have a clear explanation at present as to why there were no differences in VEGF or VEGF-A expression reported by Wang et al²⁷ and Hu et al²⁸ in retinas of wild-type and VLDLr knockout mice using RT-PCR and in situ hybridization studies.

Basic fibroblast growth factor is a potent angiogenic molecule. It stimulates smooth muscle cell growth, wound healing, and tissue repair.²⁹ Basic fibroblast growth factor is localized in the retina, lens, photoreceptors, aqueous and vitreous humors, and corneal epithelium.^30,31 It acts as a neurotrophic factor to support photoreceptor survival and may participate in photoreceptor signal transduction.^32,33 In the retinas of VLDLr knockout mice, strong bFGF staining was located mainly in the ONL. Some bFGF-positive cells were also observed in the inner nuclear layer at the lesion area. This suggests that bFGF may participate in angiogenic and neuroprotective processes.

Glial fibrillary acidic protein is an established indicator of retinal stress. Increased GFAP expression in macroglia has been described in human retinas with AMD.³⁴ In the normal mammalian retina, GFAP is marginally detectable in Müller cells.^35,36 When under stress, activated Müller cells express GFAP. Neovascularization during hypoxic conditions is mediated by Müller cells via their release of VEGF and transforming growth factor β or via their direct contact to endothelial cells.³⁷ Activation of Müller cells in the retinas of VLDLr knockout mice may contribute to the strong expression of VEGF in the lesion area.

It is believed that breaking the balance between angiogenetic stimulators and inhibitors, especially between VEGF and PEDF, leads to neovascularization. Pigment epithelium–derived factor is a potent and broadly acting neurotrophic factor and angiogenic inhibitor, which inhibits proliferation of endothelial cells.³⁸ It was reported recently that extracellular phosphorylation converts PEDF from a neurotrophic to an antiangiogenic factor.³⁹ Decreased expression of PEDF during active neovascular episodes in VLDLr mutants was reported,²⁸ whereas an increase in the expression of PEDF was observed in the initial stage of angiogenesis in the retina of this animal.²⁷ The present data show that a decrease in PEDF expression can only be localized in neovascular areas, whereas PEDF expression in the normal retina is intact, with no significant change in whole retinal PEDF expression detected by Western blot and RT-PCR. Significantly increased PEDF expression at the initial neovascularization stage might be the result of a protective mechanism that tries to inhibit the pathologic angiogenesis, whereas in the late stage of neovascularization, a strong decrease in the expression of PEDF can be explained by the protective mechanism being disrupted in VLDLr mutant mice. At 4 to 6 weeks of age, VLDLr knockout mice might be at a transition stage in which the level of PEDF transits from the initial high level to a later low level. Further experiments are needed to carefully characterize the regulation and expression of different types of PEDF in the process of pathologic neovascularization in VLDLr knockout mice.

The MAPK/Akt/NF-κB signaling cascade is important for mediating diverse cellular responses to growth factors, physical and chemical stresses, and inflammatory cytokines.^40,41 In the present study, activated MAPK levels in the retinas of VLDLr knockout mice were significantly elevated, suggesting an angiogenic role for MAPK in retinal neovascularization in this mouse model. It is known that MAPK is involved in transcriptional regulation of proinflammation cytokine expression.⁴² On activation of MAPK, transcription factors present in the cytoplasm or the nucleus are phosphorylated and activated, leading to the expression of target genes, and resulting in a biological response. Indeed, the translocation of NF-κB from the cytoplasm to the nucleus was observed in the retinas of VLDLr knockout mice in the present study, indicating activation of NF-κB, and leading to the expression of target genes. It has also been demonstrated that Akt is responsible for the activation of NF-κB during inflammatory processes.⁴³ Furthermore, the involvement of Akt in inflammatory processes has been suggested in patients with nonbacterial inflammation of the nervous system.⁴⁴ The increased activation of Akt in the retinas of the VLDLr knockout mice may also play a role in the pathogenesis of neovascularization.

Intercellular adhesion molecule-1 plays an important role in the inflammatory response.⁴⁵ Correlation between IL-18 and ICAM-1 has also been reported in autoimmune diseases.⁴⁶ It has been suggested that the higher level of ICAM-1 protein in the macular area may impart greater susceptibility of the macula to immune cell–mediated damage in AMD in humans.⁴⁷ We demonstrate that the expression of ICAM-1 was significantly greater in the retinas of VLDLr knockout mice before significant development of intraretinal neovascularization, suggesting that ICAM-1 may be involved in early angiogenesis in the retinas of VLDLr knockout mice.

Interleukin 18, previously known as interferon-inducing factor and a recently described member of the IL-1 cytokine superfamily, is now recognized as an important regulator of innate and acquired immune responses.⁴⁸ The role of IL-18 in the eye is not clear. Increased bioactive corneal IL-18 production can be induced by a variety of proinflammatory agents and may play an important role in initiating interferon-mediated inflammatory responses.⁴⁹ It is known that the MAPK/Akt/NF-κB signaling cascades are involved in mediating diverse cellular responses to growth factors, physical and chemical stresses, and inflammatory cytokines, including IL-18.⁵⁰ The present finding that the expression of IL-18 in the retinas of VLDLr knockout mice is significantly increased before significant development of neovascularization in the retinas of VLDLr knockout mice suggests a possible role of IL-18 in the pathogenesis of AMD.

The VLDLr, a lipid clearance receptor, binds to apolipoprotein E–containing lipoproteins, mediates fatty acid entry into peripheral tissues, and facilitates the hydrolysis of triglycerides.⁵¹ The VLDLr also plays an important role in signal transduction. Except for apolipoprotein E, several ligands and their signal pathways related to angiogenesis have been reported to be regulated by the VLDLr.⁵² In agreement with Heckenlively et al,⁸ the present histologic studies show that VLDLr knockout mouse retinal neovascularization appeared in the outer plexiform layer in postnatal 3 weeks, and these vessels migrated into the subretinal space in postnatal week 4 and anastomosed with choroidal vessels in postnatal weeks 6 to 8.

The accumulation of cell debris in the subretinal space could up-regulate inflammatory response factors (ICAM-1 and IL-18) and activate MAPK/Akt/NF-κB signaling cascades in VLDLr knockout mice, which will cause a mild subclinical form of inflammation, subsequently activate Müller cells and up-regulate VEGF and bFGF expression, and, finally, lead to neovascularization in the retinas of VLDLr knockout mice.

Elucidating the molecular mechanisms underlying these biochemical alterations in the retinas of VLDLr knockout mice will improve our understanding of the pathogenesis of retinal neovascularization observed in AMD, including RAP. It will provide a foundation for developing novel therapeutic approaches to not only RAP but also other forms of choroidal and retinal neovascularization.

Correspondence: Wei Cao, MD, PhD, Department of Ophthalmology, University of Oklahoma Health Sciences Center, Dean A. McGee Eye Institute, 608 Stanton L. Young Blvd, Oklahoma City, OK 73104 (wei-cao@ouhsc.edu).

Submitted for Publication: June 30, 2006; final revision received September 25, 2006; accepted October 31, 2006.

Author Contributions: Dr Cao had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Financial Disclosure: None reported.

Funding/Support: This study was supported by grant P20 RR017703 from the COBRE program of the National Center for Research Resources, National Institutes of Health; core grant EY12190 from the National Institutes of Health; the Foundation Fighting Blindness; an unrestricted grant from Research to Prevent Blindness to the Department of Ophthalmology, University of Oklahoma Health Science Center; and grant HR06-012 from the Oklahoma Center for the Advancement of Science and Technology.

Acknowledgment: We thank Li Kong, MD, PhD, Mark Dittmar, BSc, and Linda Boone, BSc, for technical support and Dr Molday for providing the antirhodopsin.