Molecular Vision 2007: 13:2310-2319
Received 11 July 2006 | Accepted 12 December 2007 | Published 21 December 2007
1Department of Ophthalmology, Harkness Eye Institute, Columbia University, New York, NY;2Kentucky Lions Eye Center, University of Louisville, Louisville, KY;3Department of Chemistry and Biochemistry, Northern Illinois University, DeKalb, IL.
Correspondence: Lucian V. Del Priore, MD, PhD, Department of Ophthalmology, Harkness Eye Institute of Columbia University, 635 West 165th Street, New York, NY, 10032; Phone: 212-305-2923; Fax: 212-342-1724; email: email@example.com.
Purpose: We have shown previously that aging of human Bruch's membrane affects the attachment, survival and gene expression profile of the overlying retinal pigment epithelium (RPE). Herein we determine the effects of Bruch's membrane aging on RPE phagocytosis of rod outer segments.
Methods: Explants of human Bruch's membrane were prepared from cadaver donor eyes (aged 9–81years) within 48 h of death, and 6 mm punches were embedded with the basal lamina in a 96-well plate. Approximately 50,000 ARPE-19 cells per well were seeded onto the explant surface and cultured for two weeks until they reached confluence. In addition, ARPE-19 were also seeded onto RPE-derived extracellular matrix (RPE-ECM) that was unmodified or modified by nonenzymatic nitration. Bovine rod outer segments were purified by sucrose gradient centrifugation, labeled with 10 ug/ml fluorescein isothiocyanate, and added to ARPE-19 cultured on Bruch's membrane or RPE-ECM for 24 h. Phagocytic activity was quantified by flow cytometry of harvested cells.
Results: The ability of RPE to phagocytose rod outer segments decreased as a function of aging of Bruch's membrane; mean phagocytotic activity of ARPE-19 on younger Bruch's membrane was significantly higher than on older Bruch's membrane (129.7 ± 34.8 versus 67.4 ± 4.2 arbitrary units, respectively; p<0.01). Nitrite treatment of RPE-ECM decreased rod outer segment phagocytosis compared to untreated RPE-ECM and mimicked the effects of aging of human Bruch's membrane.
Conclusions: Aging of human Bruch's membrane decreases rod outer segment phagocytosis by ARPE-19. This effect can be mimicked by nonenzymatic nitration of extracellular matrix in vitro. Our observations may have implications for understanding the role of aging changes within Bruch's membrane on pathogenesis of age-related macular degeneration and other disorders.
In the normal human eye, the retina pigment epithelium (RPE) forms a hexagonal cell monolayer that lines Bruch's membrane internally and separates the neural retina from the choriocapillaris . The RPE is responsible for maintaining the integrity of the neural retina, choriocapillaris, and Bruch's membrane . The apical RPE processes surround the distal tip of the outer segments . The RPE performs several crucial functions important for maintaining the outer retina, including phagocytosis of the distal tips of outer segments, recycling of visual pigment, and transferring nutrients from the choriocapillaris to the neural retina [2-4]. In addition, the RPE is responsible for maintaining the integrity of the choriocapillaris, as surgical RPE removal or pharmacological RPE damage by intravitreal injection of ornithine or iodate leads to secondary choriocapillaris atrophy [5,6]. The integrity of the RPE monolayer is maintained if there is proper attachment of the RPE to one another and to Bruch's membrane . The attachment between the RPE and inner aspects of Bruch's membrane are mediated by integrin receptors on the RPE and ligands within the basal lamina that include laminin, fibronectin, and collagen type IV ; cell-cell attachment is mediated by cadherins present on the lateral RPE cell border .
Age-related macular degeneration (AMD) is characterized by cellular changes in the RPE, choriocapillaris, and outer retina and by structural changes within Bruch's membrane that include diffuse thickening, accumulation of drusen, basal laminar, and basal linear deposits, collagen cross-linking in the inner and outer collagen layer, calcification, and fragmentation of the elastin layer and lipidization [9-13]. Cellular changes in advanced AMD include atrophy of the RPE, choriocapillaris, and outer retina in nonexudative AMD and the development of choroidal neovascularization in exudative AMD [14-18]. The connection between the ultrastructural changes observed in Bruch's membrane with age and cellular changes that develop in AMD is not known. However, it is known that the structural changes within Bruch's membrane precede cellular changes in the RPE by one or two decades [9,10,12,19] and that age-related changes within Bruch's membrane can induce changes in the attachment, survival, proliferation, and gene expression profile of the overlying RPE [20-24]. The chronology of these changes suggests that changes in Bruch's membrane may induce changes in the behavior of the overlying RPE.
In the human eye in vivo, the RPE and Bruch's membrane must by necessity age together, thus making it impossible to separate the effects of Bruch's membrane aging from the effects of cellular aging. Studying the effects of aged Bruch's membrane in the human eye in vivo creates a chicken-versus-the-egg (which comes first) dilemma since cellular aging of the RPE may initiate aging of Bruch's membrane which in turn may accelerate RPE aging . In the current study we used a system previously developed in our laboratory to study the effects of age-related changes within Bruch's membrane on the behavior of human RPE [20-22]. In essence, RPE from tissue culture were seeded onto acellular human Bruch's membrane explants harvested from donors of different ages. After repopulation of the surface, RPE monolayers were then fed purified labeled outer segments to determine the effects of Bruch's membrane aging on RPE phagocytic ability. This experimental design allowed us to separate the effects of substrate aging from the effects of RPE aging. Our results suggest that age-related changes in Bruch's membrane can decrease the phagocytosis ability of the RPE, and have important implications for understanding age-related changes in RPE function observed in AMD.
Immortalized human RPE (ARPE-19), obtained from American Type Culture Collection (Manassas, VA), were cultured and propagated in Dulbecco's Modified Eagle Medium (DMEM; Gibco, Grand Island, NY) containing 10% fetal bovine serum (FBS), 100 IU/ml penicillin G, 100 ug/ml streptomycin, 100 ug/ml gentamicin, and 2.5 ug/ml Amphotericin B (Sigma, St. Louis, MO). The cells were incubated in a humidified atmosphere of 5% CO2 and 95% air at 37 °C, and the culture medium was changed three times per week. Approximately 50,000 cells were plated into 96-well plates (Nalge Nunc International, Rochester, NY) and allowed to reach confluence (typically two weeks). Eight wells of the culture were immunostained with mouse antihuman ZO-1 antibody using fluorescence conjugated secondary goat antimouse antibody (Invitrogen Corp, Carlsbad, CA).
Bruch's membrane explants were obtained from young (aged 9–46 years) and older (aged 74–81 years) donors (Table 1) obtained through the National Disease Research Interchange (Philadelphia, PA). All donor eyes were enucleated within 10 h and shipped in a sterile container at 4 °C and received in our laboratory within 48 h of death. Two Bruch's membrane explants were harvested from each donor with the basal lamina layer of Bruch's membrane on the apical surface [20-22]. The posterior pole of each eyecup was inspected visually with direct and retroillumination under a dissecting microscope, and globes were discarded if there were any evidence of subretinal bleeding, extensive drusen, or irregular pigmentation of the macular RPE. The eyes were placed in carbon dioxide-free media (Gibco), and an incision was made through the sclera 3 mm posterior to the limbus and extended circumferentially. Four radial incisions were then made through the sclera, and the sclera was peeled away. The anterior segment and vitreous were removed and discarded. The choroid-Bruch's membrane-RPE complex was removed after trimming its attachment to the optic nerve. Native RPE were removed by bathing the explant with 0.02 M ammonium hydroxide in a 60-mm polystyrene Petri dish (Falcon; Becton Dickinson, Lincoln Park, NJ) for 20 min at room temperature followed by washing three times in phosphate buffered saline (PBS). The explant was then floated in carbon dioxide-free media over an unlaminated, hydrophobic 125–175 micron-thick polytetrafluoroethylene membrane (Millipore, Bedford, MA) with 0.5 micron pores with the basal lamina of the RPE facing the membrane. The curled edges were flattened from the choroidal side with fine forceps without touching Bruch's membrane. Next, 4% agarose (Sigma) was poured onto the Bruch's membrane-choroid complex from the choroidal side, and the tissue was kept at 4 °C for 2–3 min to allow the agarose to solidify. The polytetrafluoroethylene membrane was peeled off. Next, 6 mm circular buttons were trephined on a Teflon sheet from peripheral Bruch's membrane and placed on 4% agarose at 37 °C in untreated polystyrene wells of a 96-well plate (Corning Costar Corp., Cambridge, MA). Allowing the agarose to solidify within 2–3 min at room temperature enabled the Bruch's membrane explant to stabilize. The acellular explants were rinsed gently with PBS three times for 5 min before they were sterilized with 20,000 rad of gamma radiation sterilized and stored at 4 °C.
Fresh bovine eyes were transported on ice from a local slaughterhouse to the laboratory. The eyes were washed once with Hanks’ basal salt solution (HBSS; Gibco), which also contained 100 U/ml penicillin and 0.1 mg/ml streptomycin. The anterior half of the eye was removed and discarded. The vitreous was removed, and the retina was allowed to gently float off the RPE by pipetting HBSS into the subretinal space. The retina was removed after cutting at the optic nerve. It was then put into a glass tube on ice that contained isolation medium composed of 20% sucrose, 20 mM Tris-Cl, 2 mM MgCl2, and 130 mM NaCl (ph 7.2), and it was homogenized. One milliliter of the retinal homogenate was layered on top of a 10%–60% linear sucrose gradient and centrifuged at 141,000 g for 2 h at 4 °C in a Beckman SW-40 rotor (Beckman Coulter, Inc., Fullerton, CA). Rod outer segments (ROS) sedimented at between 27-50% sucrose gradient. Bands containing rod outer segments were collected and diluted with HBSS. The rod outer segments were pelleted by centrifugation at 7700 g for 10 min at 4 °C, suspended in the buffered salt solution, and again pelleted at 800 g for 10 min [26,27]. Purified ROS were stored at 4 °C in Dulbecco’s modified Eagles medium (DMEM) containing 17% sucrose, 100 ug /ml streptomycin, and 100 U/ml penicillin. The rod outer segments were used within 4 weeks. Prior to use, they were centrifuged at 9,000 g for 10 min to remove storage medium .
Isolated rod outer segments were pelleted in HBSS, suspended in serum-free culture medium (DMEM) and transferred to a 1.5 ml microcentrifuge tube, where it was stained with fluorescein isothiocyanate (FITC; Sigma) using a modification of the method of Boyle and McLaughlin . A 2 mg/ml stock solution of FITC in 0.1 mol/L sodium bicarbonate at pH 9.0–9.5 was prepared under dim red light, filter-sterilized, and stored in aliquots at –20 °C. The FITC stock was added to the suspended rod outer segments (final concentration 10 ug/ml) and incubated for 1 h at room temperature in the dark with gentle stirring. The FITC-stained rod outer segments were pelleted for 4 min at 7,000 rpm in a microcentrifuge, suspended in growth medium, counted, and diluted to a final concentration of 1.0 × 107/ml using a Coulter Z1 cell counter (Beckman Coulter). Microscopic examination of the labeled rod outer segments showed intact outer segments after labeling (data not shown).
RPE cells were fed FITC-labeled outer segments two weeks after the RPE were seeded onto Bruch's membrane explants. For this purpose, confluent RPE cultures were overlaid with FITC-rod outer segments (107/ml in DMEM with 20% fetal bovine serum) and incubated at 37 °C for 24 h. After 24 h the cells were rinsed with DMEM with 20% FBS to remove FITC-labeled ROS from the medium. Eight cultures were fixed at this point using 2% paraformaldehyde in 0.1 M phosphate buffer at pH 7.4 for subsequent imaging.
Visualization of the RPE border was facilitated by labeling fixed cells with Vybrant Cell-labeling Solution Di-I (Molecular Probes, Eugene, OR) before imaging. Briefly, two weeks after cells were cultured on Bruch's membrane, they were labeled with Vybrant CM-Di-I in DMEM with 10% FBS for 20 min at 37 °C and washed three times with cell culture medium. The size of RPE cells were measured using Image J software (NIH, Bethesda, MD) at 20x objective.
The fluorescence (excitation=488 nm, emission=530 nm) of 10,000 of unfixed cells/well was then assayed immediately on a FACSan (Becton Dickinson Immunocytometry System, San Jose, CA) using a live gate to exclude cell fragments, rod outer segments particles, and other unwanted debris. A logarithmic scale of relative fluorescent intensity was used, and rod outer segments phagocytosis was calculated by subtracting the geometric mean autofluorescence of control cells from the geometric mean fluorescence of cells challenged with FITC-labeled rod outer segments. Cellular autofluorescence was determined for each population by analyzing cells from wells with no added rod outer segments. Following rod outer segments challenge, unbound and cell membrane surface-bound ROS were removed by washing the wells three times with PBS, treated for 10 min with 0.25% trypsin containing 1 mM edetic acid , pelleted, washed again with PBS, and suspended in 500 ul of PBS with 10 mM glucose at pH 8.0.
Affymetrix DNA microarray was performed as follows . Briefly, ARPE19 were seeded onto 10 independent Bruch's membrane explants (five young donors age 31–47 years and five older donors aged 71–81 years). ARPE-19 cells were seeded onto the explants surface as described in the previous section. Total RNA isolation and microarray target (U133plus2, Affymetrix, Santa Clara, CA) hybridization, washing, staining, and scanning probe arrays were done following a standard Affymetrix GeneChip Expression Analysis Manual. Gene expression analyses, including global normalization and scaling, were performed using the Affymetrix GCOS Manager software .
To prepare tissue culture wells coated with RPE- extracellular matrix (ECM), we plated ARPE-19 into 96-well plates and allowed them to grow to confluence. RPE cells were removed by adding 20 mM ammonium hydroxide buffer for 20 min, followed by washing with PBS three times, and drying to obtain 96 well plates coated with RPE-derived ECM. Sodium nitrite-treated ECM was prepared by adding 100 mM sodium nitrite in pH 4 acetate buffer for one week. Wells were then washed at least four times with PBS, further incubated with PBS for 4 h, and then washed at least two additional times to completely remove nitrite .
Two weeks after seeding onto Bruch's membrane explants ARPE-19 formed a hexagonal monolayer with prominent expression of ZO-1 at the cell border. As Figure 1 shows, ARPE-19 appeared similar after culturing onto a younger (20 years old) donor Bruch's membrane versus an older (80 years old) donor Bruch's membrane. As evident in Figure 2, ingestion of the outer segment material was more prominent in ARPE-19 seeded onto younger versus older Bruch's membrane. Phase contrast microscopy did not show apparent morphological differences between ARPE-19 seeded onto younger (Figure 2A) versus older (Figure 2B) Bruch's membrane. Fluorescence microscopy of ARPE-19 seeded onto a 16-year-old Bruch's membrane showed intense labeling of ingested outer segments (2C, solid arrow), whereas cells seeded onto an 81-year-old Bruch's membrane had less intense labeling after feeding fluorescent-labeled outer segments (2D, solid arrow). Flow cytometry of harvested ARPE-19 revealed that the population of cells containing labeled outer segments was higher in RPE cultured on young Bruch's membrane (Figure 3, top panel) compared to ARPE-19 cultured on older Bruch's membrane (Figure 3, bottom panel).
The average fluorescence intensity per cell, which is a measure of the phagocytosis ability of ARPE-19, increases as a function of incubation time (Figure 4) up to 24 hours after feeding of outer segments. This observation is consistent with the study reported elsewhere . The average fluorescent intensity per cell decreased as the age of human Bruch's membrane increased (Figure 5), indicating that aging of human Bruch's membrane decreases the ability of the RPE to ingest outer segment material.
We also determined the effects of Bruch's membrane aging on the expression levels of MerTK, which is a receptor tyrosine kinase involved in phagocytosis of outer segments by RPE. Our results demonstrate that the expression level of MerTK was affected by the age of Bruch's membrane; the expression level of MerTK mRNA was lower in RPE cultured on older Bruch's membrane than on younger Bruch's membrane (61.53 ± 9.2 versus 100 ± 11.2 arbitrary units, respectively; p=0.026, Figure 6).
We then compared the effects of nonenzymatic nitration of the substrate on the phagocytosis ability of ARPE-19. The ingestion of labeled rod outer segments increased with rod outer segment concentration on both unmodified and nitrite-modified RPE derived ECM (Figure 7) with the plateau value higher on unmodified ECM. At all rod outer segment concentrations tested, the phagocytosis of outer segments was higher on unmodified ECM versus nitrite-modified ECM (Figure 7).
We then compared the effects of Bruch's membrane aging and nonenzymatic nitration of RPE-ECM on ARPE-19 phagocytosis (Figure 8). Using the Student t-test, we found the average fluorescent intensity per cell on younger Bruch's membrane (129.7 ± 34.8 arbitrary units) was higher than on older Bruch's membrane (67.4 ± 4.2 arbitrary units, respectively; p<0.01). This may imply that aging decreases ARPE-phagocytosis by about 50% compared to the baseline value on young Bruch's membrane. Interestingly, the mean fluorescence intensity of ARPE19 cells (reflecting amount of ingested FITC-labeled ROS) is reduced by about 50% when cultured on nonenzymatic nitration of RPE derived ECM (126.7 ± 12.8 versus 60.8 ± 8.2 arbitrary units for unmodified versus nitrite-treated RPE-ECM, respectively; student t-test p < 0.01, Figure 8).
We then determined the average cell area on aged versus young Bruch's membrane, and unmodified and nitrite–treated RPE-ECM, to ensure that the effects we observed were not due to a simple difference in surface area between the cells. The mean area of RPE cells on younger Bruch's membrane was similar to the cell area on older Bruch's membrane (Figure 9). Culturing the cells on harvested ECM led to a larger surface area compared to RPE cultured on human Bruch's membrane; however there was no difference between RPE cultured on untreated versus nitrite-treated extracellular matrix (Figure 9).
AMD is a complex disorder characterized by cellular changes in the RPE, choriocapillaris, and outer retina and by structural changes within Bruch's membrane that include diffuse thickening, accumulation of basal laminar and basal linear deposits, collagen cross-linking within the inner and outer collagen layers, and elastin fragmentation [12,13]. Cellular changes that occur in AMD include RPE and choriocapillaris atrophy, photoreceptor atrophy, and the development of choroidal neovascularization [9-13]. In the aging eye, the connection between aging changes in the basement membrane and aging changes within cells is not known. However, it is known that the aging changes within Bruch's membrane occur several decades before cellular changes, thus suggesting that Bruch's membrane dysfunction can induce changes in the adjacent cells.
We have previously developed a system for isolating the effects of basement membrane aging from the effects of cellular aging from the effects of cellular aging. We have this system to show that RPE attachment and proliferation decrease and apoptosis rate increases when cultured on aged Bruch's membrane [20-22]; the gene expression profile of human RPE is also altered by aging of Bruch's membrane . Herein we demonstrate that aging of Bruch's membrane has a direct effect on an important physiologic function of human RPE, namely, phagocytosis of outer segments. To our knowledge, this is the first direct evidence that aging of human Bruch's membrane leads to a diminution in RPE phagocytic ability.
Phagocytosis is an important physiologic process performed by the RPE, which is one of the most active phagocytic epitheliums in the human body. In vivo, approximately 3%–5% of the distal tips of photoreceptor outer segments are shed daily, and proper RPE phagocytosis is necessary to maintain the health and integrity of the neural retina and choriocapillaris [2,4,33]. In vivo, impairment of RPE phagocytosis would lead to improper handling of photoreceptor outer segments with accumulation of material within the RPE and deposition of abnormal material within Bruch's membrane. Lipofuscin accumulation has been associated with age-related macular degeneration [34,35]. Improper handling of ingested photoreceptor material may create a positive feedback loop in which aging changes in Bruch's membrane may reduce RPE phagocytosis, leading to further accumulation of material within Bruch's membrane, further impairing phagocytosis. Interestingly, a decrease in phagocytic ability may be associated with an increase in RPE melanogenesis, and clinically, RPE hyperpigmentation is a risk factor for the development of AMD. As human age, alterations in melanogenesis occur in the epidermis throughout the human body .
At the moment, we do not know which molecular change or changes in Bruch's membrane are responsible for the age-dependent decline in RPE phagocytosis ability we have observed. However it is tempting to speculate that nitrite modification of the substrate may be responsible for this effect since nitrite modification of ECM decreases the phagocytosis ability of the RPE to an extent comparable to Bruch's membrane aging. We have shown previously that nitrite modification decreases the ability of RPE to attach to extracellular matrix, inhibits cell proliferation, and increases cell apoptosis and necrosis . Nitrite-mediated collagen damage leads to collagen cross-linking in vitro and may underlie the changes that we have seen. There is an age-related linear decline in collagen solubility, which reflects an increase in cross-linking, with increasing patient age . In vivo, pathological collagen cross-linking is nonenzymatic , and cross-linking agents include nonreducing sugars, reactive oxygen species, and reactive nitrogen species. Pentosidine, which results from nonenzymatic glycation, increases with increasing age in human Bruch's membrane [40,41]. Nitrite-mediated nonenzymatic collagen cross-linking, shown in vitro  and likely to occur in vivo, and has been associated with chronic inflammation and cigarette smoking [42,43], which are associated with the development of AMD [44,45]. We cannot exclude the possibility that other changes are responsible for the effect of Bruch's membrane aging on RPE phagocytosis, since the proteomics of aging Bruch's membrane are not understood completely. In addition to nitrite modification, numerous other changes occur within Bruch's membrane with age including posttranslational modification of existing proteins, deposition of abnormal proteins and lipids, topographic reorganization of protein structure, and changes in the presentation of ligand binding sites necessary for cell attachment [12,13].
Our results suggest that expression of the gene encoding for MerTK, a tyrosine kinase known to be involved in RPE phagocytosis, decreases with aging of human Bruch's membrane. Several studies have demonstrated that MerTK plays an important role in phagocytosis of outer segments by the RPE. The Royal College of Surgeons rat (RCS rat) is a well characterized animal model of retinal degeneration that arises from a defect in phagocytosis, with accumulation of debris in the subretinal space leading to progressive loss of photoreceptors . The defect results from a mutation in the Mertk gene, which is normally expressed in the RPE . In the RCS rat, a small deletion of DNA disrupts the gene encoding the receptor tyrosine kinase Mertk . Mice homozygous for a targeted disruption of a homologous gene manifest a retinal dystrophy phenotype similar to RCS rats . Tissue culture studies suggest that the RPE is the site of action of Mertk, and Mertk protein plays a key role in RPE phagocytosis. Mertk protein is absent from RCS but not wild-type cultured RPE cells. Delivery of rat Mertk to cultured RCS RPE cells by means of a recombinant adenovirus restored the cells to complete phagocytic competency . Mutation of the human ortholog, MERTK, were detected in 3 of 328 DNA samples from patients with various retinal dystrophies diagnosed with retinitis pigmentosa, thus demonstrating that a defect in RPE phagocytosis can lead to human retinal disease . A MERTK missense mutation in humans can lead to a severe cone-rod dystrophy with childhood onset when in compound heterozygous form with a R722X allele . The similarity in phenotypes between the two rodent models, and the detected presence of a similar phenotype in humans with mutations in the homologous gene, suggests that alterations in phagocytosis can lead to retinal degeneration in both humans and rodents . Correction of this phagocytic defect with gene therapy can prolong photoreceptor cell survival in the RCS rat [47,52,53].
In summary, we have shown that aging of human Bruch's membrane leads to a decline in the phagocytic ability of RPE cells, and that the magnitude of this effect is comparable to the effects of nonenzymatic nitration of RPE-derived extracellular matrix. Currently, AMD is thought of as a cellular disease in which pathological alterations in the RPE, outer retina, or possibly macrophages lead to secondary change in Bruch's membrane; our own studies offer a different paradigm to interpret aging changes in this disease. In our model, age-related changes in human Bruch's membrane, that may include but are not limited to nonenzymatic nitration, could lead to extensive disruption of RPE function, including RPE phagocytosis with accumulation of material beneath the RPE. Interestingly, a clinical trial is currently under way to determine if systemic dialysis can be used to alter Bruch's membrane in age-related macular degeneration . Preliminary studies suggest that soft drusen may resolve and cell function may improve after this procedure, which may improve the function of the RPE, photoreceptors, and outer retina . Additional studies are required to determine the ultimate effects of dialysis and the use of other techniques to refurbish human Bruch's membrane on the function of the RPE and other cells .
This study was supported by the Robert L. Burch III Fund, the Macula Society, the Foundation Fighting Blindness, the Hickey’s Family Fund, and unrestricted funds from Research to Prevent Blindness. Dr. Tezel is the recipient of an RPB Career Development Award.