The results of recent studies have implicated local inflammation and complement activation as the processes involved in the pathogenesis of age-related macular degeneration (AMD). We have demonstrated that amyloid β (Aβ), which is deposited in drusen, causes an imbalance in the angiogenesis-related factors in retinal pigment epithelial cells. We have also shown that neprilysin gene-disrupted mice accumulate Aβ, and develop several features of AMD. The purpose of this study was to investigate the mechanisms involved in the development of AMD that are triggered by Aβ. Our results showed that Aβ binds to complement factor I which inhibits the ability of factor I to cleave C3b to inactivated iC3b. Factor H and factor I are soluble complement-activation inhibitors, and preincubation of factor I with Aβ in the presence of factor H abolished the ability of Aβ to cleave C3b, and also abolished the ability of factor I to cleave FGR-AMC. In contrast, Aβ did not affect the function of factor H even after binding. The production of iC3b was significantly decreased when C3b and factor H were incubated with the eyes from neprilysin gene-disrupted mice as compared with when C3b and factor H were incubated with eyes from age-matched wild-type mice. These results suggest that Aβ activates the complement system within drusen by blocking the function of factor I leading to a low-grade, chronic inflammation in subretinal tissues. These findings link four factors that have been suggested to be associated with AMD: inflammation, complement activation, Aβ deposition, and drusen.
The complement system is a component of the humoral immune system involved in host defense, and it can be activated through three distinct pathways: the classical, the alternative, and the lectin. Activation of these three pathways leads to the cleavage of C3 into C3a and C3b (1), which finally leads to the formation of the C5b-9 membrane attack complex (MAC)3 (1, 2, 3). The MAC is a lytic complex that is lethal not only to foreign pathogens but also to local host cells and tissues. Thus, the regulation of complement activation is important for the maintenance of tissue homeostasis and is mediated by a family of complement proteins.
There has recently been an increased interest in the involvement of complement in the development and progression of many diseases including age-related macular degeneration (AMD). AMD is the leading cause of irreversible vision loss, affecting 30∼50 million elderly individuals worldwide (4, 5, 6, 7). Two major clinical phenotypes of AMD are recognized: a nonexudative (dry) type and an exudative (wet) type. Recent studies have demonstrated that local and chronic inflammation induced by the complement pathway plays a central role in the formation of drusen and the development of AMD (8, 9, 10, 11).
Polymorphisms in the genes coding for complement factor H and factor B have been shown to be predictors of the risk of developing AMD (12, 13, 14, 15, 16, 17). Factor H and factor B are key components of the alternative complement pathway. Factor H, the most important soluble complement activation inhibitor, blocks the binding of factor B to C3b, supports the dissociation of the C3bBb complex (decay-accelerating activity), and acts as a cofactor in the cleavage of C3b by factor I (18, 19). Another susceptible locus for AMD is the hypothetical gene called LOC387715/HTRA1 (17, 20, 21, 22). The population risk associated with variants of factor H, factor B, and LOC387715/HTRA1 is at least 50% (17). In addition, a very recent study confirmed that functional polymorphisms in the C3 gene were strongly associated with AMD (23). These genetic studies in patients with AMD, together with the finding that proteins associated with inflammation and immune-mediated processes are deposited in drusen and Bruch’s membrane in eyes with AMD (24), support the hypothesis that inflammation and complement activation are involved in the pathogenesis of AMD. Also, animal experiments, using a laser-induced choroidal neovascularization model, showed that the activation of the complement alternative pathway was critical for the development of choroidal neovascularization (25), a major feature of the neovascular form of AMD.
One of the earliest clinical hallmarks of AMD is subretinal extracellular deposits, known as drusen, which accumulate beneath the retinal pigmented epithelium (RPE) (26). Epidemiological studies have shown that numerous and/or confluent drusen significantly increase the risk for the development of AMD (27, 28). However, the mechanism of the development of AMD from drusen has not been precisely determined. Recent proteomic analyses have shown that drusen contain many proteins, such as cholesterol, apolipoproteins B and E, C-reactive protein, clusterin, vitronectin, and amyloid β (Aβ) (24, 29, 30). Among these proteins, we focused on Aβ as the primary stimulus that causes the development of AMD (31).
Aβ peptides vary in length from 39- to 43-aa residues and are produced by the sequential proteolytic processing of amyloid precursor proteins. Studies have shown that Aβ is the main component of senile plaques in the brain of patients with Alzheimer’s disease (AD), and the transition of Aβ from the monomeric form to the oligomeric or aggregated form in the brain is a key event in the pathogenesis of AD (32, 33, 34). The results of our earlier study demonstrated that exposure of RPE cells to Aβ induced a marked increase in vascular endothelial growth factor and a marked decrease of pigment epithelium-derived factor, an antiangiogenic factor, in the RPE cells (31). Importantly, there was an increase in the deposition of Aβ in the subretinal space of senescent neprilysin gene-disrupted mice, and these mice developed several features of human eyes with AMD (31). These results suggest that Aβ deposited in drusen may be a key contributor to the development of AMD. However, the underlying molecular mechanism on how Aβ leads to the features of AMD has not been determined.
The pathogenesis of AD has also been considered to be one of the inflammatory processes caused by complement activation (35). Aβ is colocalized with activated complement components, the C5b-9 MAC, in the brain of AD patients (36). Aβ activates the alternative pathway by triggering the formation of covalent, ester-linked complexes of Aβ with C3b/iC3b, an activation product of the third complement component (C3) (36). Also in AMD, Johnson et al. (29) demonstrated that Aβ is colocalized with the activation-specific fragments of complement C3 in unique substructural domains, the “amyloid vesicles,” within the drusen thereby identifying them as potential sites of complement activation.
These findings led us to hypothesize that AMD results from an abnormal inflammatory process which includes an unregulated activation of the complement pathway triggered by Aβ, as suggested for AD. Although there has been a publication demonstrating the colocalization of complement proteins and Aβ in eyes with AMD (29), the precise mechanism of how complement is activated by Aβ in AMD has not been determined. Based on the human pathological studies and our previous animal experiments, we have explored the influence of Aβ on the activation of the alternative pathway, and we will present new evidence that Aβ binds and inactivates the function of complement factor I. This then results in an unregulated activation of complement in subretinal tissues.
Materials and Methods
1–40 (HCl form) and Aβ40–1 were obtained from Peptide Institute. The mAb against Aβ (4G8) was purchased from SIGNET; the iC3b ELISA kit, purified human factor H, and purified human factor I were obtained from Quidel; purified human C3b was obtained from Biogenesis; mAb against human factor I (MCA507, OX21), mAb against human factor H (MCA509, OX24), and polyclonal antiserum against human C3 (AHC007) were obtained from Serotec; recombinant substrate FGR-AMC was obtained from American Diagnostica; Microfluor white plates were obtained from Packard Bioscience; the0.4-μm pore-size transwell was obtained from Corning; the XTT cell proliferation assay kit was obtained from Cayman Chemical.
Human RPE cell cultures
Primary cultures of human RPE cells were a gift from Dr. P. A. Campochiaro (Wilmer Eye Institute, Johns Hopkins University, Baltimore, MD). The cultures used for the experiments were at the third or fourth passage at the time of separation. Cultures were shown to be pure populations of RPE cells by immunocytochemical staining for cytokeratins (data not shown). Differentiated RPE cells were established as described (37, 38). RPE cell cultures were maintained in DMEM supplemented with 10% FBS and 10 ng/ml basic FGF. The cells exhibited epithelial morphology and expressed CRALBP, a marker for differentiated RPE cells (38).
RPE cells were subcultured in 12-well tissue-culture plates at a density of 6 × 104 cells/well. Seven days after reaching confluence, the medium was changed, and cells were incubated in serum-free DMEM (500 μl) in the presence (50 μM) or absence of Aβ1–40 peptide (HCl form). After 24 h, the medium was collected and stored at −20°C until use for Western blot analysis and ELISA. The cells were homogenized, and their RNA and protein were extracted. Each condition was examined in triplicate, and results were repeated in at least three independent experiments.
Cell viability assay
Seven days after reaching cellular confluence, the medium was changed, and cells were incubated in serum-free DMEM in the presence (25 or 50 μM) or absence of Aβ1–40 peptide (HCl form) for 24 h. The XTT assay was used to measure viability using XTT reagent (Cayman Chemical), which was added for 1 h at 37°C and was quantified using an ELISA plate reader. The average absorbance measured for medium plus treatment was subtracted from each test sample. Each experiment was performed at least three times on different days.
RT-PCR for C3, C5, factor H, factor H-like protein (FHL), factor B, and factor I
Total RNA was extracted from cultured RPE cells using TRIzol reagent. cDNA was synthesized from 2 μg of total RNA using You-Prime First-Strand Beads according to the manufacturer’s protocol, and the reaction products were subjected to PCR amplification using the GeneAmp PCR system (Cetus; PerkinElmer). The mRNAs of human C3, C5, factor H, FHL (a truncated form of factor H (39)), factor B, and factor I were amplified with the following primers: for C3 (40), 5′-TCACCGTCAACCACAAGCTGCTACC-3′ (forward) and 5′-TTTCATAGTAGGCTCGGATCTTCCA-3′ (reverse); for C5 (40), 5′-GTGGCATTAGCAGCAGTGGACAGTG-3′ (forward) and 5′-GCAGGCTCCATCGTAACAACATTTC-3′ (reverse); for factor H (41), 5′-TCTGCATGTTGGCCTTCCTGTC-3′ (forward) and 5′-CTTCCTTGTAAATCTCCACCTG-3′ (reverse); for FHL (40), 5′-CAGAAGTTCAGAGGGTAAAGCT-3′ (forward) and 5′-TACTGGCTGGATACCTGCTCCG-3′ (reverse); for factor B (42), 5′-CAACAGAAGCGGAAGATCGTC-3′ (forward) and 5′-TATCTCCAGGTCCCGCTTCTC-3′ (reverse); and for factor I (43), 5′-GGCAGGTGGCAATTAAGGATG-3′ (forward) and 5′-GGTGTATCCAGTCTACTACTGT-3′ (reverse).
The PCR for factor H and FHL was repeated for 30 cycles, and each cycle included denaturation at 94°C for 1 min, annealing at 58°C for 1 min, and primer extension at 72°C for 1 min. The PCR for C3 and C5 was repeated for 30 cycles, and each cycle included denaturation at 94°C for 1 min, annealing at 62°C for 1 min, and primer extension at 72°C for 1 min. The PCR for factor I was repeated for 30 cycles, and each cycle included denaturation at 94°C for 1 min, annealing at 60°C for 1 min, and primer extension at 72°C for 1 min. The PCR for factor B was repeated for 35 cycles, and each cycle included denaturation at 94°C for 1 min, annealing at 59°C for 1 min, and primer extension at 72°C for 1 min. The expected size of the reaction products was 186 bp for C3, 315 bp for C5, 424 bp for FHL, 354 bp for factor H, 884 bp for factor B, and 146 bp for factor I.
LightCycler real-time PCR
The cDNA was subjected to quantitative PCRs on a LightCycler Instrument (Roche Diagnostics) using the QuantiTect SYBR Green PCR kit (Qiagen) for C3, factor H, and factor I. PCR amplifications were performed with specific primers in a total volume of 20 μl containing 2 μl of sense and antisense primer mixture (0.5 μM of each primer), 10 μl of 2× SYBR Green QPCR Master Mix (Qiagen), 1 μl of 1/10 diluted cDNA, and 7 μl of nuclease-free PCR-grade water. The mixture was used as a template for the amplification after initial denaturation at 95°C and 40∼50 cycles (95°C for 10 s, 58∼62°C for 15 s, and 72°C for 30 s). The primer sequences of human GAPDH were 5′-ACCACAGTCCATGCCATCAC-3′ (forward) and 5′-TCCACCACCCTGTTGCTGTA-3′ (reverse). The primer sequences of C3, factor H, factor B, and factor I were the same as used for semiquantitative RT-PCR. SYBR Green fluorescence was measured, and quantification of each PCR product was expressed relative to GAPDH.
ELISA measurements of C3 fragments
The amount of iC3b secreted from RPE cells into the conditioned medium was determined with a commercial ELISA kit according to the manufacturer’s instructions. The absorbance was measured at 405 nm in a Bio-Rad Model 450 microplate reader. Serial dilutions of recombinant C3 and iC3b were measured and a standardized curve was constructed.
Five days after cellular confluence, the medium was changed to serum-free DMEM. Then, the cells were incubated with or without Aβ1–40 (HCl form; 50 μM) or Aβ40–1 for 24 h. In some experiments, exogenous C3b (5 μg/ml) was added to the medium, and in other experiments, cells were preincubated with or without Aβ1–40 (HCl form; 50 μM) for 12 h. After the preincubation, C3b (5 μg/ml) and various concentrations of purified human factor H or factor I were added to the medium. The conditioned media were collected after another 12 h, and the amount of iC3b was measured.
For transwell membrane experiments, cultured RPE cells were seeded onto cell culture inserts (0.4-μm pore size; Costar; Corning). Five days after cellular confluence, the medium was changed to serum-free DMEM. In this experiment, the exogenous C3b and Aβ1–40 were added into the lower chamber. The conditioned media were also collected from the lower chamber.
Because it has already been reported that mouse complement factor I can cleave human C3b in the presence of human complement factor H (44), the ability of cleaving human C3b by the eye samples from both 12-mo-old neprilysin gene-disrupted mice (which leads to an increased deposition of Aβ as we demonstrated previously (31)) and age-matched wild-type mice was examined. Briefly, eyes were enucleated and the anterior segment as well as the neural retina was removed from the eyecup before sclera and the vascular choroid were carefully removed as much as possible using forceps. Electron microscopic examination confirmed that the intact RPE-Bruch’s membrane-choriocapillaris complex and only a small amount of vascular choroid were removed in preliminary experiments (data not shown). The RPE-Bruch’s membrane-choriocapillaris complex was pooled from four eyes, respectively, from wild-type or neprilysin gene-disrupted mice, and then were transferred into 1.5-ml assay tubes containing 40 μl of 10 μM phosphate buffer (pH 7.4; containing 145 mM NaCl) including 3.2 μg of human C3b and 3.2 μg of human factor H. The mixtures were centrifuged and incubated at 37°C for 12 h before being submitted to human iC3b EIA.
Western blot analyses
For the Western blot analysis of C3, factor H, and factor I, RPE cells were maintained in serum-free medium for 24 h; the conditioned media were collected and cells were lysed in 1 ml of sample buffer (125 mM Tris-HCl (pH 6.8), 2% SDS, 5% glycerol, 0.003% bromphenol blue, and 1% 2-ME). The final protein concentration was determined using a BCA assay (Pierce) according to the manufacturer’s instructions. Equal amounts of secreted protein (8 μg) or a sample of the whole cell lysate (80 μg) were separated by 8% SDS-PAGE and electrophoretically transferred onto nylon membranes. The nylon membranes containing the transferred proteins were pretreated with 5.0% nonfat dried milk in 50 mM Tris (pH 8.0) and then incubated overnight with a monoclonal Ab against human C3 (1/1000 dilution), mAb against human factor H (1/200 dilution), or mAb against human factor I (1/200 dilution). SDS-PAGE analysis for C3, factor H, and factor I was performed under nonreduced conditions.
Measurement of factor I cofactor activity
The cleavage of C3b by factor I in the presence of factor H was performed in 10 mM phosphate buffer (pH 7.4) containing 145 mM NaCl, according to the method of Sahu et al. (45) In a typical assay, C3b (80 ng/μl), factor H (80 ng/μl), and factor I (3.5 ng/μl) were mixed in 10 μM phosphate buffer (pH 7.4; containing 145 mM NaCl). The total volume in each reaction tube was adjusted to 20 μl, and the tube was incubated at 37°C. In some experiments, factor I and/or factor H were preincubated with Aβ1–40 (200 μM) for 1 h, and then incubated with C3b. After another 1 h, samples (5 μl) were removed and mixed with sample buffer containing DTT, boiled for 5 min, and subjected to electrophoresis on an 8% SDS-PAGE gel. The cleavage products were made visible by silver staining of the gel.
Amidolytic assay using synthetic substrate
The cleavage of the synthetic FGR-AMC by factor I in the presence or absence of Aβ was examined. This substrate is cleaved by factor I in the absence of any cofactors (46). FGR-AMC (50 μM; final concentration) in 20 mM HEPES at pH 8.5 was added to factor I (0.5 μM; final concentration) in the same buffer in a white Microfluor plate. The final reaction volume was 200 μl. The reaction was performed in the presence or absence of Aβ1–40 (200 μM, final concentration). The amidolytic activity of factor I was measured using a microtiter plate reader (Fluoroskan; Thermo Life Sciences) by excitation at 355 nm and continuous monitoring of emission at 460 nm for 1 h or more at 37°C. Activity of factor I was expressed as fluorescence of total free AMC released from FGR-AMC by the cleavage of factor I.
To examine the binding of factor I or factor H and Aβ, human purified factor I (210 ng) or human purified factor H (4.8 μg) and Aβ1–40 (200 μM; final concentration) were added to 10 μM phosphate buffer (pH 7.4; containing 145 mM NaCl) to a final volume of 60 μl. After 1-h incubation at 37°C, the mixtures were reacted with 200 μl of factor I Ab or 200 μl of factor H Ab and further incubated at 4°C with gentle end-over-end mixing for 90 min. After incubation, 100 μl of 50% agarose slurry (protein G type) was added to both mixtures and further incubated at 4°C for 30 min with gentle end-over-end mixing. After 30 min of incubation, each mixture was centrifuged, the supernatant was removed, and the agarose containing factor I-Aβ complex or factor H-Aβ complex was washed with PBS three times before being loaded in diluted sample buffer (by PBS, 1/1) which did not contain DTT. The samples was heated for 5 min at 100°C and centrifuged. Then, 50 μl of the supernatant from each sample was respectively subjected to Western blot analysis using factor I Ab (MCA507, 1/200) or factor H Ab (MCA509, 1/200), and Aβ (4G8, 1/1000) Ab. Complex immunoreactivity was made visible by exposure of x-ray film to blots incubated with ECL reagent.
Data are expressed as means ± SEM. The significance level was set at p < 0.05. Statistical analysis was performed with the Mann-Whitney U test.
Expression of C3, C5, factor H, FHL, factor B, and factor I in RPE cells
The liver is the primary site of synthesis of the majority of the complement proteins in the human plasma. Extrahepatic biosynthesis of complement proteins may be an important factor in triggering and perpetuating local inflammatory reactions, especially in tissues that are shielded from the plasma components by a blood-tissue barrier. RT-PCR demonstrated the constitutive expression of the mRNAs for C3, C5, factor H, FHL, factor B, and factor I (Fig. 1⇓) in the RPE cells. These findings suggest that RPE cells have a capacity to modulate the activation of the alternative pathway of complement.
C3b degradation is inhibited in Aβ-treated RPE cells
In all three pathways of complement activation, the pivotal step is the conversion of complement C3 to C3b, which is then covertly coupled to pathogens and Ag-Ab complexes. C3b is cleaved by factor I to iC3b in the presence of factor H and other cofactors (47). To determine whether C3b was cleaved by the RPE cells, we measured the level of iC3b in the supernatant by ELISA. Without exogenous C3b and Aβ1–40, the average level of iC3b was 0.95 ± 0.08 μg/ml in the supernatants of RPE cells (Fig. 2⇓). When the RPE cells were incubated with 50 μM Aβ1–40 for 24 h, the amount of iC3b was significantly reduced to 0.13 ± 0.01 μg/ml (p = 0.007). When C3b (5 μg/ml) was added to the medium, the average amount of iC3b in the RPE cell supernatant was significantly increased to 2.00 ± 0.14 μg/ml. However, when 50 μM Aβ1–40 was added to the medium, the amount of iC3b was significantly decreased to 0.78 ± 0.07 μg/ml (p = 0.013). In contrast, Aβ40–1 alone had no effect on the cleavage of C3b.
When we added the Aβ1–40 into the lower chamber using the transwell culture system, the amount of iC3b within the medium from the lower chamber was significantly decreased to 0.11 ± 0.01 μg/ml as compared with 0.175 ± 0.005 μg/ml within the medium from the nontreated RPE cells (p = 0.034) (Fig. 3⇓). When C3b (5 μg/ml) was added into the lower medium, the average amount of iC3b was significantly increased to 1.5 ± 0.3 μg/ml. However, when 50 μM Aβ1–40 was added to the lower medium together with C3b, the amount of iC3b was significantly decreased to 0.2 ± 0.02 μg/ml (p = 0.0016) (Fig. 3⇓). In addition, the 50 μM Aβ1–40 did not affect the cell viability as determined by the XTT assay (data not shown).
Addition of factor I restores iC3b production in Aβ-treated RPE cells
To confirm the direct contribution of either factor I or factor H in the decrease of C3b degradation by Aβ, we then examined the effect of purified factor I or factor H on Aβ-induced degradation of C3b. After changing the medium to serum-free medium, RPE cells were preincubated with (50 μM) or without Aβ1–40 for 12 h. Then, 5 μg of rC3b was added to the medium, and the supernatant was collected after another 12 h. The amount of iC3b was determined by ELISA. The concentration of iC3b in the supernatants of nontreated RPE cells was 2.30 ± 0.10 μg/ml (Fig. 4⇓), but the iC3b was significantly decreased to 0.87 ± 0.05 μg/ml in Aβ-treated RPE cells (p = 0.013). The addition of human purified factor I (1, 2, and 5 μg) significantly restored iC3b generation, but the addition of purified factor H (1, 2, and 5 μg) did not increase the amount of iC3b, although the high concentration of factor H (5 μg) was able to restore iC3b generation in part.
Effects of Aβ on expression of gene and protein of C3, factor H, and factor I in human RPE cells
To determine whether the decrease of iC3b in the Aβ-treated RPE cells was due to a modified production of different complement-related proteins, we determined the expression of the mRNA and protein of C3, factor H, and factor I. The results showed that the expressions of the mRNA of C3, factor H, and factor I were not significantly different between RPE cells with and without Aβ treatment by real-time PCR (p > 0.05; Fig. 5⇓, upper lane).
A representative photograph of a Western blot of C3, factor H, and factor I protein is shown in Fig. 5⇑. Compared with nontreated RPE cells, the amount of C3 protein in the supernatants of Aβ-treated cells was similar, although it was slightly higher than that in the cell lysates. The secretion of factor H protein was decreased more in the supernatant of RPE cells treated with Aβ1–40 than that in nontreated cells, and factor H protein was also increased in the Aβ-treated RPE cell lysates compared with nontreated cell lysates. The secretion of the protein of factor I was decreased more in the supernatant of Aβ-treated RPE cells than in nontreated cells; however, factor I protein was slightly increased in the cell lysates of Aβ-treated cells than in nontreated cell lysates.
Binding of Aβ with factor H and factor I
Because it has already been reported that Aβ binds to factor H (48), we investigated whether Aβ also forms a complex with factor I. Coimmunoprecipitation analysis detected a band in the 88-kDa area, which corresponded to the molecular mass of factor I, both by Abs against factor I as well as against Aβ (Fig. 6⇓A). Also, a band in the 155-kDa area, which corresponded to the molecular mass of factor H, was detected by the Abs against factor H and Aβ (Fig. 6⇓B). These findings indicated that Aβ binds to factor I as well as factor H to form a complex.
Aβ affects activity of factor I but not factor H
We next analyzed whether Aβ affected the ability of factor I and factor H to cleave C3b. When C3b was incubated in only phosphate buffer, we detected two bands corresponding to the α- (110-kDa) and β- (70-kDa) chains of C3b (Fig. 7⇓, first lane). The addition of factor H or factor I alone did not cause the degradation of C3b (Fig. 7⇓, second and third lanes). When we incubated C3b with both factor H and factor I, two degradative fragments of the α-chain, viz., 68- and 43-kDa, appeared (Fig. 7⇓, fourth lane). The absence of the 68- and 43-kDa fragments in reactions with C3b and factor I or factor H alone showed that both factor I and factor H are necessary for the cleavage of C3b under these conditions (Fig. 7⇓, fourth lane). The presence of the 68- and 43-kDa fragments indicated that factor I mediated the cleavage of C3b between Arg1298 and Ser1299 in the generation of iC3b2. Finally, when we simultaneously incubated C3b with factor H, factor I, and Aβ1–40, the α-chain of C3b was also cleaved into two fragments (Fig. 7⇓, fifth lane).
Next, we investigated whether preincubation of factor H or factor I with Aβ affected its ability to cleave of C3b. When factor H was preincubated with Aβ1–40 for 1 h, and then incubated with C3b and factor I for an additional 1 h, the degradation of the α-chain occurred (Fig. 7⇑, seventh lane). However, when factor I was preincubated with Aβ1–40 for 1 h before with C3b and factor H for another hour, the degradation of α-chain did not occur (Fig. 7⇑, ninth lane). These results indicated that the preincubation with Aβ affected the ability of factor I, but not factor H, to cleave C3b into iC3b.
Effects of Aβ on the amidolytic activity of factor I
To further explore whether Aβ affected the function of factor I, we examined the amidolytic activity of factor I. This assay determines the levels of amidolytic activity of factor I against the FGR-AMC substrate without any cofactors. The addition of Aβ1–40 abolished the amidolytic activity of factor I for 1–12 h (Fig. 8⇓).
C3b degradative activity is decreased in the eyes of neprilysin gene-disrupted mice
Finally, we incubated the RPE-Bruch’s membrane-choriocapillaris complex from the eyes of 12-mo-old neprilysin gene-disrupted mice or age-matched wild-type mice with exogenously added human C3b and factor H. The results demonstrated that the production of iC3b was significantly decreased to 14.2 ± 1.25 μg/ml when C3b and factor H were incubated with the eyes from neprilysin gene-disrupted mice compared with when C3b and factor H were incubated with the eyes from wild-type mice (29.0 ± 2.5 μg/ml; p = 0.049; Fig. 9⇓).
In our previous study regarding the developmental mechanism of AMD from drusen, (31) we focused on Aβ accumulated within drusen, and demonstrated that senescent neprilysin gene-disrupted mice which accumulate an excess amount of Aβ manifested similar features of human AMD. Although the mechanism for why Aβ induced the development of AMD was unclear, we hypothesized that Aβ-induced complement activation in subretinal tissue plays an important role for the development of AMD, and explored the mechanism of Aβ-induced complement activation in the present study.
Our results showed that the generation of inactivated C3b, and thus iC3b, was inhibited in RPE cells exposed to Aβ (Fig. 2⇑). This finding was confirmed in experiments using the RPE-Bruch’s membrane-choriocapillaris complex from neprilysin gene-disrupted mice (Fig. 9⇑), which accumulated Aβ deposition in RPE cells as well as within the sub-RPE deposits, as we reported previously (31). The mechanism for this inhibition is the binding of Aβ to factor I which abolishes its ability to cleave its substrates as shown in two functional assays: the cofactor assay and the amidolytic assay.
Factor I is an 88-kDa serum glycoprotein which is a serine protease and cleaves the α-chains of C3b and C4b in the presence of its cofactor proteins (46). By inactivating C3b and C4b through limited proteolytic cleavage and thereby preventing the formation of the C3 and C5 convertases, factor I inhibits the alternative and the classical component pathways. Thus, the dysfunction of factor I accelerates the formation of the C3 convertase in the alternative pathway (45) resulting in unregulated complement activation.
Interestingly, the simultaneous incubation of Aβ with C3b, factor H, and factor I did not inhibit the C3b degradation in the cofactor assay. However, preincubation of factor I with Aβ abolished its ability to cleave the α-chain of C3b in the cofactor assay and to cleave FGR-AMC in the amidolytic assay. The HCl form of Aβ is known to be aggregated within 30 min when incubated in medium by The Peptide Institute (Osaka, Japan). This might suggest that the aggregation of Aβ might be necessary for the binding to factor I to Aβ as shown for the binding of factor H to Aβ (48). However, the mechanism of how the function of factor I is disturbed after binding to Aβ requires further investigation. As best we know, the binding of factor I to Aβ has not been reported, while the binding of factor H to the aggregated fibrillar form of Aβ has been reported (48). Aggregated fibrillar Aβ is generally considered to be a rather sticky complex capable of binding proteins nonspecifically, although a charge-based interaction might be one possibility as suggested by Strohmeyer and associates (48).
Western blot analyses showed that exposure of RPE cells to Aβ decreased the amount of factor H and factor I secreted into the supernatant of cultured RPE cells (Fig. 5⇑), while Western blot analyses using cell lysates showed that the amount of C3, factor H, and factor I was increased in Aβ-treated RPE cells (Fig. 5⇑). This difference might be because Aβ binds to these molecules and traps them possibly on the cell membrane as reported (43). Thus, not only the dysfunction of factor I by binding to Aβ, but also a decrease of the secreted amount of free factor H and factor I into the subretinal space might be involved in the process of the unregulated activation of complement.
The extrahepatic biosynthesis of complement is considered to be an important source of complement which triggers and perpetuates local inflammatory reactions, especially in tissues that are shielded from the plasma components by a blood-tissue barrier such as the retina. By using specific primers and RT-PCR, we demonstrated that human RPE cells constitutively express the major components of the complement alternative pathway, including C3, factor B, factor H, FHL, C5, and factor I (Figs. 1⇑ and 2⇑). A Medline search identified only one publication that reported the synthesis of factor H by human and mouse RPE cells (49). However, we did not find any publications for the expression of the other complement components by RPE cells.
The expression of various complement components in human RPE cells suggests that most of the complement proteins and their regulators that are present within the drusen and in the subretinal space can be synthesized by RPE cells, although some of the proteins may be derived from the choroidal circulation. Therefore, it is highly likely that the RPE cells regulate the activation of the complement pathway and maintain this aspect of retinal homeostasis under physiological conditions. However, once the RPE cells become senescent or pathological, the regulation of the complement pathway might be disturbed.
Although recent genetic studies suggested an important role of the complement system for the pathogenesis of AMD (12, 13, 14, 15, 16, 17, 20, 21, 22, 23), we suspect that this is probably due to the tissue specificity of the macular area in the eye. The retina is shielded from the plasma components by a blood-tissue barrier like the brain, and furthermore, the retina of the macular area is completely avascular. In such kind of shielded tissue, once unregulated complement activation induced by the accumulated Aβ occurs, it might be difficult to stop the amplification loop. Subsequently, the chronic exposure of RPE cells to bioactive fragments of complement components might induce pathological damage on RPE cells. In addition, chronic activation of complement might cause a migration of monocytes that might be involved in the degradation of Bruch’s membrane.
Chronic inflammation induced by complement activation is also considered to be a significant contributor to the neurodegenerative processes in the brains of patients with AD (35, 36, 43, 50, 51). In any case, the pathogenesis of AD and AMD may have many similarities, and the observations made in our study might also apply to the brain of AD patients.
For AMD to develop, an interaction between the environment and genetic susceptibility has been considered to be critical. Very recently, Campochiaro and colleagues demonstrated that oxidative stress reduces the ability of IFN-γ, an inflammatory cytokine, to increase the expression of factor H in RPE cells (52). Also, in an investigation of a potential trigger for complement activation in AMD, Zhou et al. (53) explored the idea that the complex mixture of products resulting from photo-oxidation of A2E might include a range of fragments that could be recognized by the complement system as “foreign,” and could activate the complement system, leading to low-grade inflammation. A2E is a bis-retinoid pigment which accumulates as lipofuscin in RPE. They found that iC3b and C3a were elevated in the medium overlying ARPE-19 cells that had accumulated A2E and were irradiated to induce A2E photo-oxidation. In our study, we showed that Aβ directly inactivated factor I. Taken together (52, 53), these data suggest that there are important interactions between environmental exposures and genetic susceptibilities in the pathogenesis of AMD.
In conclusion, we have demonstrated that Aβ binds to factor I and blocks its ability to inhibit the inactivation of C3b, resulting in unregulated complement activation. Our findings provide new insights into the pathogenesis and the treatment strategies of AMD, especially to prevent the progression to AMD from its earliest clinical hallmark of drusen deposits.
We thank Prof. Duco Hamasaki for critical discussion and revision of the final manuscript. We also thank Prof. Peter A. Campochiaro and Sean F. Hackett for human RPE cells. We thank M. Sekiguchi, K. Watanabe, and R. Fujioka (RIKEN Brain Institute) for technical assistance.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported in part by Research Grants 19390441, 18023037, and 19659445 from the Japan Society for the Promotion of Science, Tokyo, Japan. This work was supported in part by Grants-in-Aid for Scientific Research on Priority Areas, Research on Pathomechanisms of Brain Disorders 18023037 (to N.I.) and 20023031 (to N.I.) from the Ministry of Education, Culture, Sports, Science, and Technology, Japan.
↵2 Address correspondence and reprint requests to Dr. Kyoko Ohno-Matsui, Department of Ophthalmology and Visual Science, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113, Japan. E-mail address:
↵3 Abbreviations used in this paper: MAC, membrane attack complex; AMD, age-related macular degeneration; RPE, retinal pigmented epithelium; Aβ, amyloid β; AD, Alzheimer’s disease; FGF, fibroblast growth factor; FHL, factor H-like protein.
- Received November 13, 2007.
- Accepted April 29, 2008.
- Copyright © 2008 by The American Association of Immunologists