Functional analyses of rare genetic variants in complement component C9 identified in patients with age-related macular degeneration

Functional analyses of rare genetic variants in complement component C9 identified in patients... Abstract Age-related macular degeneration (AMD) is a progressive disease of the central retina and the leading cause of irreversible vision loss in the western world. The involvement of abnormal complement activation in AMD has been suggested by association of variants in genes encoding complement proteins with disease development. A low-frequency variant (p.P167S) in the complement component C9 (C9) gene was recently shown to be highly associated with AMD; however, its functional outcome remains largely unexplored. In this study, we reveal five novel rare genetic variants (p.M45L, p.F62S, p.G126R, p.T170I and p.A529T) in C9 in AMD patients, and evaluate their functional effects in vitro together with the previously identified (p.R118W and p.P167S) C9 variants. Our results demonstrate that the concentration of C9 is significantly elevated in patients’ sera carrying the p.M45L, p.F62S, p.P167S and p.A529T variants compared with non-carrier controls. However, no difference can be observed in soluble terminal complement complex levels between the carrier and non-carrier groups. Comparing the polymerization of the C9 variants we reveal that the p.P167S mutant spontaneously aggregates, while the other mutant proteins (except for C9 p.A529T) fail to polymerize in the presence of zinc. Altered polymerization of the p.F62S and p.P167S proteins associated with decreased lysis of sheep erythrocytes and adult retinal pigment epithelial-19 cells by carriers’ sera. Our data suggest that the analyzed C9 variants affect only the secretion and polymerization of C9, without influencing its classical lytic activity. Future studies need to be performed to understand the implications of the altered polymerization of C9 in AMD pathology. Introduction Age-related macular degeneration (AMD; MIM# 603075), a progressive eye disorder, is the major cause of irreversible vision loss in the western world (1). The disease is multifactorial, involving both environmental and genetic factors in its pathogenesis (2). Genetic alterations are estimated to account for 46–71% of variability in disease risk (3). A large component of the heritability of AMD can be explained by genetic variants in the alternative pathway of the complement system. A recently published genome-wide association study detected 52 (45 common and 7 rare) variants at 34 genomic regions that are independently associated with AMD. More than one third of these variants reside in or near genes encoding for components of the complement system (4). Complement is a crucial part of innate immunity, providing clearance of foreign and altered-self structures. Activation of its cascade results in enzymatic-cleavage of the central component, C3 into C3a and C3b. C3b is a crucial component of C3 and C5 convertases, allowing further propagation of the cascade into the terminal pathway where C5b-8 complexes incorporate into the membrane. After binding several copies of complement component C9 (C9) to C5b-8, the pore-forming membrane attack complex (MAC)/terminal complement complex (TCC) is assembled (5,6). MAC kills the target cell by inducing cell lysis, or at a reduced, the so-called sublytic concentration, it can provoke a wide array of physiologic responses ranging from apoptosis to pro-inflammatory cytokine secretion (7–10). The potential involvement of C9 in AMD pathology has been suggested by increased MAC deposition in the retina of AMD patients and the correlation between the amount of MAC and the loss of RPE cells (11,12). In addition, sublytic MAC deposition on RPE cells has been reported to induce secretion of pro-inflammatory cytokines and vascular endothelial growth factor, contributing to the development of advanced AMD (13,14). Recently, three rare genetic variants in C9 were reported in association with AMD (15–17) namely p.R95X, p.R118W and p.P167S. The p.P167S variant was reported to be highly associated with AMD risk in multiple studies (4,15,17–19). More recently, a genetic burden of C9 variants was described in two separate AMD cohorts. The first study identified 13 rare variants (P-value 2.4×10−08) (18) and the second study (Corominas et al., manuscript submitted) revealed 17 rare variants (P-value 5.01×10−03) in C9; however in both studies the burden did not remain significant after correction for multiple comparisons. In a recent study, we demonstrated that the p.P167S variant leads to increased serum concentration of the protein (17). However, the functional consequences of the p.P167S variant and of other genetic alterations in C9 remain unclear. In this study, we aimed to further elucidate the functional effects of C9 variants in vitro in order to understand the role of C9 in AMD pathogenesis. Results Genetic alterations identified in C9 Through whole-exome sequencing in 793 unrelated individuals (662 AMD cases and 131 controls), we identified five novel rare variants in C9: p.M45L, p.F62S, p.G126R, p.T170I and p.A529T, in addition to previously reported variants p.R118W and p.P167S (Table 1). Genotyping of these 7 variants in 1896 unrelated AMD cases and 1499 unrelated control individuals (Supplementary Material, Table S1) identified 127 rare variant carriers. The identified variants are present in different domains of the protein: the thrombospondin type 1 (TSP1), the low-density lipoprotein receptor type A (LDLRA), the MAC/perforin (MACPF) and the epidermal growth factor (EGF)-like domains (Fig. 1A). All the identified variants are non-synonymous point mutations resulting in amino acid changes in the mature protein (Table 1). Table 1. Overview of rare genetic variants in C9 investigated in the study Variant ExAc freq (%) PhyloPa Granthama SIFTb PolyPhenb CADDcphred Literature (1st report) refSNP Amino acidd cDNAd Novel rs41271047 p.M45L c.133A>T 0.21 0.734 15 T B 9.639 Novel rs140251849 p.F62S c.185T>C 0.01 2.187 155 T D 25 Geerlings et al. (17) rs147701327 p.R118W c.352C>T 0.02 1.393 101 D P 28 Novel rs199939436 p.G126R c.376G>A 0.03 3.758 125 D D 34 Novel rs34882957 p.P167S c.499C>T 0.470 3.279 74 D D 25.3 Seddon et al. (15) NA p.T170I c.509C>T NA 2.087 89 D D 24.7 Novel rs137891079 p.A529T c.1585G>A 0.050 −1.191 58 D D 0.028 Novel Variant ExAc freq (%) PhyloPa Granthama SIFTb PolyPhenb CADDcphred Literature (1st report) refSNP Amino acidd cDNAd Novel rs41271047 p.M45L c.133A>T 0.21 0.734 15 T B 9.639 Novel rs140251849 p.F62S c.185T>C 0.01 2.187 155 T D 25 Geerlings et al. (17) rs147701327 p.R118W c.352C>T 0.02 1.393 101 D P 28 Novel rs199939436 p.G126R c.376G>A 0.03 3.758 125 D D 34 Novel rs34882957 p.P167S c.499C>T 0.470 3.279 74 D D 25.3 Seddon et al. (15) NA p.T170I c.509C>T NA 2.087 89 D D 24.7 Novel rs137891079 p.A529T c.1585G>A 0.050 −1.191 58 D D 0.028 Novel a Higher PhyloP [range −14; 6.4] and Grantham [range 0–215] scores correlate with a higher conservation. b Sorting Intolerant from Tolerant (SIFT) and PolyPhen2 classification: T, tolerated; B, benign; D, damaging; P, pathogenic. c Combined Annotation Dependent Depletion (CADD) scores the deleteriousness of genetic variants in the human genome (log10 scale, i.e. 10 predicts variant in top 10%, 20 in top 1%, 30 in top 0.1%, etc. of reference most deleterious variants). d Signal peptide included in numbering. NA, non-available. Table 1. Overview of rare genetic variants in C9 investigated in the study Variant ExAc freq (%) PhyloPa Granthama SIFTb PolyPhenb CADDcphred Literature (1st report) refSNP Amino acidd cDNAd Novel rs41271047 p.M45L c.133A>T 0.21 0.734 15 T B 9.639 Novel rs140251849 p.F62S c.185T>C 0.01 2.187 155 T D 25 Geerlings et al. (17) rs147701327 p.R118W c.352C>T 0.02 1.393 101 D P 28 Novel rs199939436 p.G126R c.376G>A 0.03 3.758 125 D D 34 Novel rs34882957 p.P167S c.499C>T 0.470 3.279 74 D D 25.3 Seddon et al. (15) NA p.T170I c.509C>T NA 2.087 89 D D 24.7 Novel rs137891079 p.A529T c.1585G>A 0.050 −1.191 58 D D 0.028 Novel Variant ExAc freq (%) PhyloPa Granthama SIFTb PolyPhenb CADDcphred Literature (1st report) refSNP Amino acidd cDNAd Novel rs41271047 p.M45L c.133A>T 0.21 0.734 15 T B 9.639 Novel rs140251849 p.F62S c.185T>C 0.01 2.187 155 T D 25 Geerlings et al. (17) rs147701327 p.R118W c.352C>T 0.02 1.393 101 D P 28 Novel rs199939436 p.G126R c.376G>A 0.03 3.758 125 D D 34 Novel rs34882957 p.P167S c.499C>T 0.470 3.279 74 D D 25.3 Seddon et al. (15) NA p.T170I c.509C>T NA 2.087 89 D D 24.7 Novel rs137891079 p.A529T c.1585G>A 0.050 −1.191 58 D D 0.028 Novel a Higher PhyloP [range −14; 6.4] and Grantham [range 0–215] scores correlate with a higher conservation. b Sorting Intolerant from Tolerant (SIFT) and PolyPhen2 classification: T, tolerated; B, benign; D, damaging; P, pathogenic. c Combined Annotation Dependent Depletion (CADD) scores the deleteriousness of genetic variants in the human genome (log10 scale, i.e. 10 predicts variant in top 10%, 20 in top 1%, 30 in top 0.1%, etc. of reference most deleterious variants). d Signal peptide included in numbering. NA, non-available. Figure 1. View largeDownload slide Schematic illustration of C9 domains with the identified C9 variants. SP, signal peptide; TSP1, thrombospondin type 1; LDLRA, low-density lipoprotein receptor type A; MACPF, the MAC/perforin domain; EGF, the epidermal growth factor-like domains. Figure 1. View largeDownload slide Schematic illustration of C9 domains with the identified C9 variants. SP, signal peptide; TSP1, thrombospondin type 1; LDLRA, low-density lipoprotein receptor type A; MACPF, the MAC/perforin domain; EGF, the epidermal growth factor-like domains. In our case–control cohort, none of the rare C9 variants were individually associated with AMD pathogenesis (Supplementary Material, Table S2). Nevertheless, regarding the previously shown high association of the p.P167S variant with AMD in large case–control studies (4,15,18,20), and the recently reported burden of rare variants in C9 [Corominas et al., manuscript submitted and (18)], we set out to identify the functional effect of these variants on the C9 protein, in order to place them into the context of AMD pathogenesis. To this end, 128 serum and 95 plasma samples of the identified carriers were tested in functional assays and compared with 156 sera (78 with AMD and 78 without the disease) and 155 plasma samples (77 with AMD and 78 without AMD) of age-matched non-carrier individuals (Supplementary Material, Table S2). Serum C9 and plasma TCC level of C9 carriers To analyze whether the identified genetic variants affect protein synthesis and secretion, the concentration of C9 was measured in sera of 127 AMD patients carrying rare genetic variants in C9 and compared with 156 non-carriers with (n = 78) or without (n = 78) AMD. We found that the p.M45L, p.F62S, p.P167S and p.A529T variants lead to a significantly increased C9 level compared with non-carrier controls (Fig. 2A). Sera carrying the p.R118W and p.G126R variants did not significantly differ in C9 levels between carriers and non-carriers (Fig. 2A). Figure 2. View largeDownload slide Measurement of serum C9 and plasma sTCC levels in C9 carriers. Concentrations of C9 in sera (A) and sTCC level in plasma (B) of patients carrying rare genetic variants in C9 were measured by ELISA and compared with non-carriers with or without AMD. Differences with P < 0.05 were considered statistically significant and signed with black (carriers versus non-carriers with AMD) or with light grey (carriers versus non-carriers without AMD) asterisks (Kruskal–Wallis test with Dunn’s multiple comparison, nsP > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). Data are shown as median μg/ml secreted C9 or AU/ml sTCC level (the latter on log10 scale) with interquartile range of four (C9 ELISA) or three (sTCC ELISA) independent experiments. Figure 2. View largeDownload slide Measurement of serum C9 and plasma sTCC levels in C9 carriers. Concentrations of C9 in sera (A) and sTCC level in plasma (B) of patients carrying rare genetic variants in C9 were measured by ELISA and compared with non-carriers with or without AMD. Differences with P < 0.05 were considered statistically significant and signed with black (carriers versus non-carriers with AMD) or with light grey (carriers versus non-carriers without AMD) asterisks (Kruskal–Wallis test with Dunn’s multiple comparison, nsP > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). Data are shown as median μg/ml secreted C9 or AU/ml sTCC level (the latter on log10 scale) with interquartile range of four (C9 ELISA) or three (sTCC ELISA) independent experiments. Binding of C9 to soluble C5b-8 complexes results in formation of soluble TCC (sTCC), which is a sign of ongoing complement activation. Therefore, we measured sTCC levels in patient and control plasma samples of carriers and non-carriers of rare C9 variants. Despite the differences in serum C9 concentration, however, we did not detect any alteration in the systemic level of sTCC in plasma of C9 carriers compared with non-carrier controls (Fig. 2B). Expression and secretion of wild-type and mutant C9 proteins by human embryonic kidney 293 cells To study the effect of the identified variants on protein secretion, human embryonic kidney 293 (HEK293) cells were transfected with either the wild-type (WT) or mutant C9 constructs and concentrations of expressed C9 in cell lysates and secreted C9 in supernatants were analyzed using western blot and enzyme-linked immunosorbent assay (ELISA), respectively. Five of the mutant C9 proteins (p.M54L, p.F62S, p.R118W, p.G126R, p.P167S) were expressed at similar levels as the WT protein in monomer forms, while the C9 p.T170I and p.A529T mutants showed decreased concentration in the cell lysates (Fig. 3A and B). Furthermore, polymerized form of C9 could be observed in case of C9 p.G126R and p.P167S (Fig. 3B). In agreement with results of C9 ELISA using patients’ sera, C9 p.F62S was secreted at higher concentration into the cell supernatant than the WT protein (Fig. 3C). In addition to its reduced expression in the cell lysate, concentration of the p.T170I mutant protein in the supernatant was also reduced. Level of the C9 p.R118W protein in the supernatant was decreased, while its expression in the cell lysate was normal. Figure 3. View largeDownload slide Expression and secretion of recombinant WT and mutant C9 proteins. HEK293 cells were transiently transfected with WT or mutant C9-pCEP4 constructs. As negative control, empty pCEP4 vector was used. Expression of C9 in the cell lysates was measured by western blot. Results shown are either densitometric analysis of three transfections and illustrated as % of WT C9 expression with mean ± SD (A) or one representative western blot (B). (C) Concentration of C9 in the secreted supernatants was measured by ELISA. Graph shows mean ± SD % of WT C9 concentration and is result of three independent transfections measured in duplicates. Differences with P < 0.05 were considered statistically significant and compared with WT C9 (one-way ANOVA with Dunnett’s multiple comparison, nsP > 0.05, *P < 0.05, **P < 0.01, ****P < 0.0001). Figure 3. View largeDownload slide Expression and secretion of recombinant WT and mutant C9 proteins. HEK293 cells were transiently transfected with WT or mutant C9-pCEP4 constructs. As negative control, empty pCEP4 vector was used. Expression of C9 in the cell lysates was measured by western blot. Results shown are either densitometric analysis of three transfections and illustrated as % of WT C9 expression with mean ± SD (A) or one representative western blot (B). (C) Concentration of C9 in the secreted supernatants was measured by ELISA. Graph shows mean ± SD % of WT C9 concentration and is result of three independent transfections measured in duplicates. Differences with P < 0.05 were considered statistically significant and compared with WT C9 (one-way ANOVA with Dunnett’s multiple comparison, nsP > 0.05, *P < 0.05, **P < 0.01, ****P < 0.0001). Lytic activity of the recombinant, mutant C9 proteins In order to explore the impact of the identified variants on the protein function, independent of other serum components, the mutant C9 proteins were recombinantly produced by HEK293F cells and purified by Ni2+-affinity chromatography and gel filtration from the secreted supernatant. The isolated proteins were visualized by both Silver staining and western blotting (Supplementary Material, Fig. S1), confirming the presence of pure, monomeric C9. Firstly, the recombinant, purified C9 proteins were compared in their ability to lyse erythrocytes. To this end, sensitized sheep erythrocytes were treated with C9-depleted serum, which was reconstituted with the WT or mutant C9 proteins. Six of the mutant C9 proteins (p.M45L, p.F62S, p.R118W, p.G126R, p.T170I and p.A529T) had normal lytic activity, while the C9 p.P167S protein showed a slight, but significant reduction in erythrocyte lysis (Fig. 4A). Figure 4. View largeDownload slide Lytic activity of recombinant C9 proteins. Sheep erythrocytes (A) or ARPE-19 cells (B) were treated with C9-depleted serum supplemented with the recombinant WT or mutant C9 variants. (A) Lysis of erythrocytes was analyzed via measurement of released hemoglobin at 405 nm. Data are expressed as median with interquartile range of water-induced maximum lysis and are results of three independent experiments carried out in duplicate. (B) MAC-induced cytotoxicity was measured via lactate-dehydrogenase (LDH) release from ARPE-19 cells. Data expressed are mean ± SD of lysis buffer-induced maximum lysis and are results of three independent experiments measured in duplicate. Differences with P < 0.05 were considered statistically significant (one-way ANOVA with Dunnett’s multiple comparison, nsP > 0.05, **P < 0.01). Figure 4. View largeDownload slide Lytic activity of recombinant C9 proteins. Sheep erythrocytes (A) or ARPE-19 cells (B) were treated with C9-depleted serum supplemented with the recombinant WT or mutant C9 variants. (A) Lysis of erythrocytes was analyzed via measurement of released hemoglobin at 405 nm. Data are expressed as median with interquartile range of water-induced maximum lysis and are results of three independent experiments carried out in duplicate. (B) MAC-induced cytotoxicity was measured via lactate-dehydrogenase (LDH) release from ARPE-19 cells. Data expressed are mean ± SD of lysis buffer-induced maximum lysis and are results of three independent experiments measured in duplicate. Differences with P < 0.05 were considered statistically significant (one-way ANOVA with Dunnett’s multiple comparison, nsP > 0.05, **P < 0.01). To further investigate the effects in a model more relevant for AMD, we compared the lytic activity of the WT and mutant C9 proteins on the retinal pigment epithelial cell line ARPE-19. To induce MAC deposition, ARPE-19 cells were pre-incubated with a function-blocking anti-CD59 antibody before treatment with C9-depleted serum, supplemented with either the WT or mutant recombinant C9 proteins. In agreement with the results obtained using sheep erythrocytes (Fig. 4A), we did not observe any differences between the WT and mutant C9 proteins in their cytotoxic activity on ARPE-19 cells (Fig. 4B), suggesting that the variants do not affect the classical, lytic function of the protein. Lytic activity of sera of C9 carriers Next, we tested whether the physiological difference in C9 concentration in sera of carriers of C9 variants may result in elevated lysis of the target cells, a phenomenon that may lead to pathological changes and destruction of the retina. To measure the C9-dependent lytic activity of sera independent of other complement components, a modified hemolytic assay was designed. Sensitized erythrocytes or ARPE-19 cells were incubated with C9-depleted serum in dextrose gelatin veronal buffer (DGVB++), which allows complement activation and deposition of C5b-8 complexes on the cell membrane. Thereafter, the cells were incubated with sera of carriers or non-carrier controls diluted in ethylenediaminetetraacetic acid (EDTA) gelatin veronal buffer (EDTA-GVB), which blocks complement activation and novel C5b-8 complex formation, but allows integration of C9 in the pre-formed C5b-8 complexes and induces lysis dependent on both C9 concentration and its functional activity in the serum. In spite of the significantly increased C9 concentration in sera of carriers of the p.M45L, p.F62S, p.P167S and p.A529T variants, the elevated C9 level did not cause increased lysis of erythrocytes (Fig. 5A) or ARPE-19 cells (Fig. 5B and C). On the contrary, we observed a slight but significant decrease in lytic activity of C9 p.F62S on both erythrocytes and ARPE-19 cells compared with non-carrier controls without AMD. Furthermore, sera of carriers of the C9 p.P167S variant showed decrease in lytic activity on erythrocytes compared with non-carriers with or without AMD. Figure 5. View largeDownload slide C9-dependent lytic activity of sera carrying rare genetic variants in C9. Lytic activity of sera was measured via incubation of sheep erythrocytes (A) or ARPE-19 cells (B, C) with C9-depleted serum supplemented with EDTA-GVB diluted sera. (A) Lysis of erythrocytes was analyzed via measurement of released hemoglobin at 405 nm. Data are expressed as median with interquartile range of water-induced maximum lysis and are results of four independent experiments. MAC-induced cytotoxicity of ARPE-19 cells was measured via LDH release (B) and by analysis of AnnexinV+ ZombieAqua+ (late apoptotic) cells by flow cytometry (C). Data expressed are median with interquartile range of lysis buffer-induced maximum LDH release (B) and % of late apoptotic cells (C) of two independent experiments measured in duplicate. Differences with P < 0.05 were considered statistically significant and signed with black (carriers versus non-carriers with AMD) or with light grey (carriers versus non-carriers without AMD) asterisks (Kruskal–Wallis test with Dunn’s multiple comparison, nsP > 0.05, *P < 0.05, **P < 0.01, ****P <0.0001). Figure 5. View largeDownload slide C9-dependent lytic activity of sera carrying rare genetic variants in C9. Lytic activity of sera was measured via incubation of sheep erythrocytes (A) or ARPE-19 cells (B, C) with C9-depleted serum supplemented with EDTA-GVB diluted sera. (A) Lysis of erythrocytes was analyzed via measurement of released hemoglobin at 405 nm. Data are expressed as median with interquartile range of water-induced maximum lysis and are results of four independent experiments. MAC-induced cytotoxicity of ARPE-19 cells was measured via LDH release (B) and by analysis of AnnexinV+ ZombieAqua+ (late apoptotic) cells by flow cytometry (C). Data expressed are median with interquartile range of lysis buffer-induced maximum LDH release (B) and % of late apoptotic cells (C) of two independent experiments measured in duplicate. Differences with P < 0.05 were considered statistically significant and signed with black (carriers versus non-carriers with AMD) or with light grey (carriers versus non-carriers without AMD) asterisks (Kruskal–Wallis test with Dunn’s multiple comparison, nsP > 0.05, *P < 0.05, **P < 0.01, ****P <0.0001). Polymerization of the recombinant, mutant C9 proteins During purification of recombinant C9 proteins we observed that the p.P167S mutant tends to aggregate and form high molecular weight multimers. To evaluate and compare the polymerization tendency of the mutant C9 proteins, recombinant, purified C9 was incubated in the presence or absence of zinc ions, which accelerate C9 polymerization and are used as dietary supplement to reduce AMD progression (21). We observed that the p.P167S mutant spontaneously aggregates (Fig. 6), while the other mutant proteins polymerize normally in the absence of zinc. However, the p.M45L, p.F62S, p.R118W, p.G126R and p.T170I mutant proteins showed impaired polymerization in the presence of zinc in contrast to C9 p.P167S, which showed higher polymerization than the WT protein. Polymerization ability of the p.A529T mutant protein was unaltered. Figure 6. View largeDownload slide Polymerization of WT and mutant C9 proteins in the presence or absence of Zn2+ ions. Recombinant WT or mutant C9 (5 μM) were allowed to polymerize in the presence or absence of 15 μM Zn2+ ions at 37°C for 4 h. The proteins were separated by electrophoresis on a 5–10% polyacrylamide gradient gel and polymers were visualized by Silver staining. (A) One representative experiment of three is shown. (B) Densitometric ratio of polymer and monomer WT and mutant C9 variants were calculated by ImageLab software. Results are illustrated as mean ± SD polymer/monomer ratios of C9 mutant proteins of three independent experiments. Differences between WT and mutant C9 proteins with P < 0.05 were considered statistically significant and signed with light grey (spontaneous polymerization) or with black (Zn2+-induced polymerization) asterisks (two-way ANOVA with Dunnett’s multiple comparison, nsP > 0.05, **P < 0.01, ****P < 0.0001). Figure 6. View largeDownload slide Polymerization of WT and mutant C9 proteins in the presence or absence of Zn2+ ions. Recombinant WT or mutant C9 (5 μM) were allowed to polymerize in the presence or absence of 15 μM Zn2+ ions at 37°C for 4 h. The proteins were separated by electrophoresis on a 5–10% polyacrylamide gradient gel and polymers were visualized by Silver staining. (A) One representative experiment of three is shown. (B) Densitometric ratio of polymer and monomer WT and mutant C9 variants were calculated by ImageLab software. Results are illustrated as mean ± SD polymer/monomer ratios of C9 mutant proteins of three independent experiments. Differences between WT and mutant C9 proteins with P < 0.05 were considered statistically significant and signed with light grey (spontaneous polymerization) or with black (Zn2+-induced polymerization) asterisks (two-way ANOVA with Dunnett’s multiple comparison, nsP > 0.05, **P < 0.01, ****P < 0.0001). Discussion In this study, we identified five novel, non-synonymous rare genetic variants in C9 (p.M45L, p.F62S, p.G126R, p.T170I and p.A529T), in addition to the previously described C9 p.R118W (17) and p.P167S (15) variants. Although the identified C9 variants (except for C9 p.M45L) were more prevalent in individuals affected by AMD compared with control individuals, the limited number of subjects analyzed in combination with the low allele frequency of the C9 variants left their individual association to AMD statistically inconclusive (Supplementary Material, Table S2). Nevertheless, the C9 variants revealed in our study further expand the repertoire of genetic variants in genes encoding components of the complement cascade and confirm the earlier reported correlations of C9 p.R118W and C9 p.P167S to enhanced diseased risk. In previous studies, three genetic variants in C9 have been associated with AMD. The nonsense mutation p.R95X is most common among Japanese and leads to complete or partial C9 deficiency (22). This variant is protective for AMD, conferring a 4.7-fold reduction in disease risk, and is correlated to decreased vascular endothelial growth factor levels (16). The p.R118W C9 variant was identified with a relatively low allelic odds-ratio (1.12) and normal C9 level in the carriers compared with non-carriers (17). C9 p.P167S increases risk for AMD (4,15,17–19) and results in elevated serum C9 level in carriers compared with non-carriers (17). In our cohort, we did not detect any significant alteration in systemic complement activation levels, measured in the form of plasma sTCC between the carriers and non-carriers (Fig. 2B). This is in accordance with previous studies, reporting no significant difference in systemic sTCC level between AMD patients and non-carrier control groups (23–25). Nevertheless, increased MAC was previously observed in Bruch’s membrane and the choriocapillaris of human eyes, gradually increasing with age. The elevated MAC was especially evident in eyes affected by AMD, even more so in carriers of the CFH risk allele (11,12). These findings suggest that changes in MAC may only be detected locally in the eye and further strengthen the role of local rather than systemic complement activation in AMD pathology, which support recent data showing increased complement activation in aqueous humor rather than in plasma of patients with AMD (26,27). We found that carriers of the p.M45L, p.F62S, p.P167S and p.A529T variants have increased serum C9 level compared with non-carrier controls (Fig. 2A) while C9 concentrations of p.R118W and p.G126R was unaltered compared with sera of non-carrier controls. These data confirm our previous results reporting normal C9 level in sera carrying the C9 p.R118W variant and elevated levels in carriers of the C9 p.P167S variant (17). Interestingly, secretion of the recombinant C9 proteins by HEK293 cells was higher only for the C9 p.F62S mutant (Fig. 3C). The differences between concentrations of various mutants of C9 in sera and HEK293 cell supernatants can be partially explained by the multicellular source and consumption of C9 in vivo in comparison to the single cell origin of the protein in HEK293 cell cultures. Despite this discrepancy, data obtained from transfected HEK293 cells contribute to the better understanding of the in vivo observed alterations in C9 concentration between the distinct carriers. For example, C9 p.F62S was expressed at the same level as WT C9 (Fig. 3A and B) but increased in the secreted supernatant (Fig. 3C). These results indicate that the mutation causes increased protein secretion and explain the elevated concentration of C9 p.F62S in the carrier’s sera (Fig. 2A). Furthermore, although secretion of recombinant C9 p.R118W and p.T170I was lower than of the WT protein, expression of these mutant proteins was also lower (in case of p.T170I almost completely absent), indicating that the in vitro observed decreased secretion is mainly owing to their diminished transcription. Indeed, serum concentration of C9 p.R118W did not differ from non-carrier controls. Unfortunately, no serum samples were available for the p.T170I carrier (n = 1) and hence, its mutational effect on C9 secretion in vivo needs to be explored in future studies. Despite the significantly elevated C9 level in carriers of the p.M45L, p.F62S, p.P167S and p.A529T variants (Fig. 2A), we could not detect an increase in C9-dependent lytic activity of patients’ sera carrying rare-genetic variants in C9 (Fig. 5). More unexpectedly, sera of the p.F62S and p.P167S C9 carriers caused decreased lysis of sheep erythrocytes and ARPE-19 cells. The lytic activity of the other C9 proteins was normal despite increased serum levels (Fig. 5A and C). Since one would assume that increased C9 concentration results in enhanced MAC formation and thus, killing of the target cells, these results are surprising. However, it should be considered that the stoichiometry of C5b-8 and C9 within MAC is strictly regulated. One C5b-8 complex can bind only a limited number of C9 molecules (28), indicating that even when more C9 is available, it does not necessarily lead to more C5b-9 complex formation if the level of earlier complement components remains unaltered. Thus, owing to limited availability of C5b-8 complexes, the slightly enhanced levels of C9 p.M45L and p.A529T do not lead to enhanced pore formation and lysis (Figs 4 and 5). Nevertheless, the reason for decreased lytic activity of C9 p.F62S and p.P167S appears to be more complex and probably connected to altered polymerization propensity of these variants. Hemolytic assays performed with the recombinant C9 mutant proteins purified only on nickel column—containing both monomer and polymer proteins—displayed a significantly reduced lytic activity of the C9 p.F62S and p.P167S mutant proteins (data not shown), confirming our observations using serum samples (Fig. 5). We also observed during expression in vitro that these variants tend to aggregate. Hence, to exclude the functionally inactive polymers from our analysis, the recombinant C9 proteins were further purified by gel filtration (Supplementary Material, Fig. S1). Notably, applying the pure, monomer forms of C9, we could not detect any difference in lytic activity of the WT and mutant proteins (Fig. 4). These results suggest that the genetic alterations do not alter the classical, lytic function of C9 and that decreased lytic activity of C9 p.F62S and p.P167S is presumably caused by altered polymerization of these variants (Fig. 6). C9 has a tendency to polymerize, leading to formation of poly(C9) and a rapid loss of hemolytic activity (29). Although we did not observe the presence of poly(C9) in sera of the p.P167S carriers, probably owing to low serum concentration of C9, our in vitro data suggest that the C9 p.P167S mutant protein has an increased propensity to aggregate. Formation of C9 p.P167S polymerization was observed already during expression and purification of the recombinant protein. In contrast, the C9 p.F62S had a decreased capacity to polymerize in the presence of zinc ions (Fig. 6). This altered polymerization of C9 p.F62S and p.P167S in contrast to the WT and other mutant C9 proteins may be explained by localization of the variants in the mature protein (Fig. 1): the TSP1 and MACPF domains have been reported as main drivers of C9 polymerization during MAC formation (30). Thus the p.F62S variant, and in particular the p.P167S mutation owing to the substitution of a proline residue, may alter the structure of C9 in a way that results in altered polymerization propensity and hence, impaired pore formation and lysis. Interestingly, despite the almost complete lack of zinc-induced polymerization, the hemolytic activity of C9 p.F62S carriers’ sera was only slightly reduced (Fig. 5) and unaltered in assays in which the recombinant protein was used as a source of C9 (Fig. 4). However, it is important to emphasize that experiments carried out with the recombinant proteins measure only the functional effect of the mutations without the influence of changes in the protein concentration in vitro (Fig. 2). This may also explain the different results of assays using either sera (Fig. 5) or the purified proteins (Fig. 4): the increased concentration of C9 p.F62S and p.P167S in the serum may facilitate the formation of polymers, rendering the protein inactive and causing decreased lysis (Fig. 5). The latter effect is not present when recombinant C9 variants are applied at the same concentration. Although decreased lytic activity of the p.F62S variant could be observed only in assays measuring lysis of sheep erythrocytes but not in lactate dehydrogenase (LDH) release of ARPE-19 cells, the differences between sensitivity of erythrocytes’ and ARPE-19 cells’ membranes to lysis together with the biological variability of serum samples can partially explain the different results of the two assays. Nevertheless, both the p.F62S and p.P167S variants showed tendency to decreased lytic activity and probably the lack of statistically significant difference between control samples and the C9 p.F62S carriers in analysis of ARPE-19 cells is owing to the fact that in the latter case, experiments were carried out with a lower number of patients sample. Taken together, we have shown that several identified C9 variants, including the p.P167S variant highly associated with AMD, affect the serum level and polymerization of C9 without influencing its classical lytic activity. These results, alongside with normal sTCC level of the carriers’ plasma and the lack of correlation between C9 concentration and lytic activity of patients’ sera suggest that the variants influence disease pathology probably locally in the eye and not by increased lytic activity. This may happen mainly via enhanced formation of drusen owing to increased aggregation of certain mutant C9 proteins, but differences in sub-lytic MAC deposition and hence, C9-dependent changes in the inflammatory milieu of the retina cannot be ruled out. Further, one should consider that although ARPE-19 cells are regarded as reliable and widely used alternatives to native RPE, they are not optimal model of AMD (31). Hence, future experiments—using primary RPE cells, induced pluripotent stem cells from the carriers and complex retinal tissues—need to be performed to understand the precise implication of C9 in AMD pathology. Materials and Methods Genetic analysis and patient selection We performed whole-exome sequencing for 793 unrelated individuals (662 cases and 131 controls), as described in detail previously (17). Filtering steps were implemented to uncover the coding non-synonymous variants of C9 (NM_001737.3). Frequency filters, using public database ExAC (32) were implemented to ensure selection of rare variants (minor allele frequency <1%). We identified seven unique rare variants in C9, which were screened in an additional cohort consisting of 1234 unrelated case and 1368 unrelated control individuals using custom-made competitive allele-specific PCR assays according to the manufacturers’ recommendations (KASP SNP Genotyping System, LGC). Furthermore, family members of eight different families (n = 24) of which the proband carried a C9 variant were screened using Sanger sequencing (Supplementary Material, Fig. S3). All individuals included in the study underwent clinical evaluation and were graded for ‘AMD’ or ‘no AMD’ according to the Cologne Image Reading Center protocol. Control individuals without AMD were 60 years or older (33). Collection of serum and plasma samples To analyze the functional effect of rare genetic variants in C9, 128 serum and 95 plasma samples of 132 carriers were collected. Furthermore, we obtained serum (n = 156) and plasma (n = 155) of 157 individuals that did not carry a rare genetic variant in C9 (Supplementary Material, Table S2). Samples were obtained by a standard coagulation/centrifugation protocol and frozen at −80°C within 1 h after collection. Genomic DNA was isolated from peripheral blood samples according to standard procedures. This study was approved by local ethics committees on Research Involving Human Subjects and met the criteria of the Declaration of Helsinki. Cell lines Freestyle HEK293F cells (Invitrogen) were cultured in Freestyle Expression medium (Invitrogen) according to the manufacturer’s instructions. Cells were passaged every third day and transfected at Passage 10. HEK293 cells (ATCC) were grown in DMEM medium supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Gibco). The retinal pigment epithelial cell line ARPE-19 (ATCC) was cultured in DME/F12 medium (Hyclone) supplemented with 10% heat-inactivated FBS. Cells used for functional assays were between Passages 10 and 20. The cells were mycoplasma free and tested regularly with the VenorGEM Classic kit (Minerva Biolabs). ELISA for measurement of C9 concentration C9 concentrations in serum samples or supernatants of HEK293 cells secreting recombinant C9 proteins were determined by ELISA. Maxisorp microtiter plates (96-well, ThermoScientific) were coated with 1 μg/ml mouse anti-human C9 (Hycult Biotech, #HM2111) in 50 mM sodium carbonate (pH 9.6) overnight, at 4°C. Between each of the following steps, the plates were washed four times with Immunowash [50 mM Tris–HCl (pH 8.0), 150 mM NaCl, 0.1% Tween 20]. After coating, plates were blocked in Quench solution [Immunowash supplemented with 3% fish gelatin (Norland Products) for 1 h at room temperature (RT) and incubated with serum samples (diluted 1:200 in Quench) or with supernatants of HEK293 cells for 1 h at RT. As standard, serum-purified C9 (prepared in our lab) was applied. After incubation, bound proteins were detected using goat anti-human C9 (Complement Technologies, #A226, 1:4000 in serum or 1:2000 in HEK293 supernatant C9 ELISA in Quench,) followed by HRP-conjugated rabbit anti-goat Ig (1:2000 in Quench, DAKO). As substrate, 1,2-phenylenediamine dihydrochloride (OPD, Kem-En-Tec) was used and absorbance at 490 nm was measured using a Cary50 MPR microplate reader (Varian). ELISA for measurement of sTCC concentration Maxisorp immunoplates (96-well) were coated with 0.5 μg/ml mouse anti-human TCC (#aE11, Hycult) in PBS (pH7) (GE Healthcare) at RT, overnight. Between each of the following steps, plates were washed four times with PBST (PBS + 0.2% Tween 20). After coating, the plates were incubated with plasma samples diluted 5× in AG buffer (PBS + 0.02% NaN3 + 0.2% Tween 20 + 0.02 M Na2EDTA) for 1 h at RT. As standard, zymosan activated serum (ZAS-93, 1000 AU/ml, Hycult) was used. Plasma sTCC was detected using biotinylated mouse anti-human C6 (Quidel, #A219, 1:2000 in PBST) biotinylated using EZ-Link Sulfo-NHS-Biotin from Pierce, followed by HRP-conjugated streptavidin (1:1000 in PBST, R&D Sytems). As chromogen, OPD was used. C9 cDNA clones for recombinant proteins To determine if the genetic variants affect the function of the protein, both the WT and mutant C9 proteins were expressed in vitro. To this end, full-length cDNA encoding human C9 with an N-terminal His-tag was purchased from Invitrogen in pMA-T vector. The identified variants in C9, i.e. p.M45L, p.F62S, p.R118W, p.G126R, p.P167S, p.T170I and p.A529T were introduced using the QuikChange site-directed mutagenesis kit (Agilent Technologies), according to the manufacturer’s instructions. The primers used are listed in Supplementary Material, Table S3. The variants were confirmed by automated Sanger DNA sequencing (GATC Biotec). WT and mutant C9 cDNA sequences were then subcloned into the eukaryotic pCEP4 expression vector (Invitrogen), suitable for transfection of HEK293F cells. Expression and purification of recombinant C9 proteins HEK293F cells were transiently transfected with the WT or mutant C9-pCEP4 constructs using FreeStyle Max Reagent according to the manufacturer’s instructions (Invitrogen). Secreted supernatants were collected after 2, 4, 6, 8 days, pooled and stored at −20°C. Recombinant WT and mutant C9 proteins were isolated from the collected media using a column of Ni-NTA Superflow resin (Qiagen) equilibrated with 50 mM NaH2PO4, 300 mM NaCl, 10 mM imidazole, pH 8 (binding buffer). After extensive washing with binding buffer, bound proteins were eluted with 50 mM NaH2PO4, 300 mM NaCl, 500 mM imidazole, pH 8 (elution buffer). The presence of C9 in the eluted fractions was verified by 7.5% SDS-PAGE and silver staining. To separate monomer and polymer C9 proteins by gel filtration, C9-containing fractions of Ni2+-affinity chromatography were pooled, concentrated using an Amicon Filter unit with 50 kDa cut-off (Millipore) and loaded onto Sephacryl-100 column (GE healthcare) to separate monomer and polymer C9 proteins. All preparative works were done at 4°C. The purity of monomer WT and mutant C9 proteins was analyzed by SDS-PAGE and western blot. SDS-PAGE and western blot The purified recombinant C9 proteins were separated by gel electrophoresis under reducing (25 mM DTT) conditions and transferred to a PVDF membrane using semi-dry blotting apparatus (BioRad). The membranes were blocked with Quench solution and C9 was visualized using a polyclonal, goat anti-human C9 antibody from Complement Technologies (#A226, 1:20 000 dilution in Quench) followed by incubation with HRP-conjugated rabbit anti-goat Ig (diluted 1:10 000 in Quench). Bound antibodies were visualized by the enhanced chemiluminescence method (Millipore) and analyzed with the ImageLab software (BioRad). Transient transfection of HEK293 cells Transient transfection of HEK293 cells with empty vector, WT or mutant C9-pCEP4 constructs was accomplished using Lipofectamine 2000 reagent (Invitrogen), according to the manufacturer’s instructions. Briefly, cells were seeded on 12-well plates before transfection and grew until 90% confluency. On the day of transfection, DMEM + 10% FBS was changed to OptiMem (Gibco) and cells were transfected with 6 µg plasmid complexed with 6 µl Lipofectamine 2000 reagents. Eight hours after transfection, cells were washed into DMEM + 10% FBS and further cultured for 2 days in DMEM + 10% FBS and 3 days in OptiMem medium. After collecting the supernatant for C9 ELISA, cells were lysed in RIPA buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% NP40, 0.5% deoxycholate) containing protease inhibitors (Roche) for 60 min on ice. After centrifugation at 15 000g for 15 min at 4°C, 60 µg of each cell lysate was run on non-reducing 10% SDS-PAGE. C9 was visualized by western blot as described for purified C9 mutant proteins, except that goat anti-human C9 antibody was used at 1:5000 dilution. Lysis of erythrocytes by recombinant WT and mutant C9 proteins Sheep erythrocytes (Håtunalab) were washed with DGVB++ (2.5 mM veronal buffer, pH 7.3, 72 mM NaCl, 140 mM glucose, 0.1% gelatin, 1 mM MgCl2 and 0.15 mM CaCl2), pelleted and incubated with 1 ml of DGVB++ containing amboceptor (Behring) diluted 1:1000 for 20 min at 37°C. Sensitized erythrocytes were washed three times, pelleted and resuspended in DGVB++ to obtain a cell suspension of which 10 µl lysed by 90 µl of water gives 1.2–1.4 absorbance at 405 nm. The suspension (10 μl) was incubated with 50 µl of 2% C9-depleted serum (Complement Technologies) reconstituted with 50 ng/ml WT or mutant C9 proteins (preliminary titrated concentration of recombinant WT C9 causing lysis of 50% of erythrocytes). After incubation for 30 min at 37°C, cells were overlayed with another 50 µl of DGVB++ and centrifuged. Released hemoglobin levels of the collected supernatants (80 µl) were measured at 405 nm using Cary50 MPR microplate reader. C9-induced lysis was calculated by dividing (Abs 405 nm of C9 reconstituted samples—Abs 405 nm of C9 depl. serum treated samples alone) with Abs 405 nm of water-treated samples and illustrated as % of maximal (i.e. water induced) lysis. Lysis of erythrocytes by sera of C9 carriers and non-carrier controls To analyze C9-dependent lysis of erythrocytes by sera of carriers and non-carriers, erythrocytes were prepared and sensitized as described above. To induce deposition of C5b-8 complexes, 10 μl of the erythrocyte suspension was incubated with 50 µl of 2% C9-depleted serum (Complement Technologies) diluted in DGVB++ at 37°C for 30 min, shaking at 650 rpm. After incubation, cells were washed three times with 40 mM EDTA-GVB buffer (2.5 mM veronal buffer, pH 7.3, 72 mM NaCl, 140 mM glucose and 40 mM EDTA) and MAC assembly was induced by serum samples, diluted 1:2000 in 40 mM EDTA-GVB (the buffer blocks complement activation and thus prevents formation of novel C5b-8/9 complexes from the applied sera). After incubation for 30 min at 37°C, cells were overlayed with another 50 µl of 40 mM EDTA-GVB and centrifuged. Released hemoglobin level of the collected supernatants (80 µl) was measured at 405 nm using a Cary50 MPR microplate reader. Treatment of ARPE-19 cells using recombinant WT and mutant C9 proteins ARPE-19 cells were plated at 106 cells/ml concentration on 96-well plates (Nunc) in DME/F12 medium supplemented with 10% FBS. After 1 day, cells were washed and medium was changed to DME/F12 without FBS. The next day, cells were centrifuged and treated with 3 µg/well function blocking anti-CD59 antibody (IBGRL) for 45 min at 4°C to enhance MAC formation. After incubation, cells were treated with 10% C9-depleted serum reconstituted with 10 µg/ml of purified WT or mutant C9 proteins in DGVB++ for 2 h at 37°C, shaking at 150 rpm. After 2 h, cells were centrifuged at 1500g for 3 min, and supernatants were collected to measure LDH release using the Pierce LDH cytotoxicity detection kit (ThermoScientific). Treatment of ARPE-19 cells by sera of C9 carriers and non-carrier controls ARPE-19 cells were plated and primed for MAC deposition as described above. To induce C5b-8 deposition, anti-CD59 sensitized cell were treated with 20% C9-depleted serum in DGVB++ for 45 min at 37°C, shaking at 150 rpm. After incubation, cells were washed three times in EDTA-GVB and treated with serum samples, diluted 2.5 times in 40 mM EDTA-GVB to supply C9 in the pre-formed C5b-8 complex and induce lysis. After 2 h, supernatants were collected to measure LDH release and cells were analyzed for MAC deposition by flow cytometry. LDH assay LDH release was measured in 80 μl volume using the Pierce LDH cytotoxicity detection kit (ThermoScientific) according to the manufacturer’s instruction. The following formula was used to determine percent of maximum LDH release: LDH activity of C9 depleted serum alone treated samples was substracted from LDH activity of C9 reconstituted samples and divided by the total LDH activity (maximum LDH release induced by 1× lysis buffer of Pierce LDH cytotoxicity kit) and multiplied by 100. Flow cytometry To detect MAC deposition and viability of ARPE-19 cells after induction of MAC assembly, cells were washed two times in PBS and incubated with a neoantigen-specific rabbit anti-human C5b-9 antibody (Complement Technologies, 1:400) for 30 min at 4°C. The cells were washed twice in binding buffer (10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl2) and incubated with AlexaFluor647-conjugated goat anti-rabbit Ab (ThermoScientific, 1:2000), Annexin V-FITC (Immunotools, 2 µl/sample) and ZombieAqua fixable viability dye (BioLegend, 1:2000) for 30 min at RT in the dark. The stained cells were washed two times and analyzed by CytoFlex flow cytometer (Beckman Coulter). Data were analyzed with FlowJo software (Tree Star) and expressed as geometric mean fluorescence intensity (gMFI) of C5b-9 signal or percent of ZombieAqua and Annexin V double positive, late apoptotic cells. C9 polymerization Purified C9 proteins (5 μM) were allowed to polymerize in TBS in the presence or absence of 15 μM Zn2+ ions at 37°C for 4 h, shaking at 350 rpm. After incubation, samples were separated by 5–10% gradient gel electrophoresis and visualized by silver staining. Statistical analysis Results investigating the relationship between serum C9 concentration, plasma sTCC level and lytic activity of serum samples and the carrier status were analyzed using the Kruskal–Wallis test with Dunn’s multiple comparison. Data comparing expression, secretion and functional activity of recombinant WT and mutant C9 proteins were analyzed by one-way ANOVA with Dunnett’s multiple comparison. C9 polymerization was compared using two-way ANOVA with Dunnett’s multiple comparison test. Data were analyzed, and graphs prepared using Prism Software version 7. Supplementary Material Supplementary Material is available at HMG online. Conflict of Interest statement. None declared. Funding This work was supported by the Wenner-Gren Foundation, The Royal Physiographic Society of Lund (37352) and the Lars Hierta's Memorial Foundation (FO2016-0079), awarded to M.K. The research leading to these results has received funding from the European Research Council under the European Union's Seventh Framework Programme (FP/2007-2013)/ERC Grant Agreement n. 310644 (MACULA), the Foundation Fighting Blindness USA (grant C-GE-0811-0548-RAD04), the Swedish Research Council (2016-01142) and a grant for clinical research (ALF). References 1 Chakravarthy U. , Evans J. , Rosenfeld P.J. ( 2010 ) Age related macular degeneration . BMJ , 340 , c981. Google Scholar CrossRef Search ADS PubMed 2 Sobrin L. , Seddon J.M. ( 2014 ) Nature and nurture—genes and environment—predict onset and progression of macular degeneration . Prog. Retin. Eye Res ., 40 , 1 – 15 . Google Scholar CrossRef Search ADS PubMed 3 Seddon J.M. , Silver R.E. , Kwong M. , Rosner B. ( 2015 ) Risk prediction for progression of macular degeneration: 10 common and rare genetic variants, demographic, environmental, and macular covariates . Invest. Ophthalmol. Vis. Sci ., 56 , 2192 – 2202 . Google Scholar CrossRef Search ADS PubMed 4 Fritsche L.G. , Igl W. , Bailey J.N. , Grassmann F. , Sengupta S. , Bragg-Gresham J.L. , Burdon K.P. , Hebbring S.J. , Wen C. , Gorski M. et al. ( 2016 ) A large genome-wide association study of age-related macular degeneration highlights contributions of rare and common variants . Nat. Genet ., 48 , 134 – 143 . Google Scholar CrossRef Search ADS PubMed 5 Ricklin D. , Hajishengallis G. , Yang K. , Lambris J.D. ( 2010 ) Complement: a key system for immune surveillance and homeostasis . Nat. Immunol ., 11 , 785 – 797 . Google Scholar CrossRef Search ADS PubMed 6 Sonnen A.F. , Henneke P. ( 2014 ) Structural biology of the membrane attack complex . Subcell. Biochem ., 80 , 83 – 116 . Google Scholar CrossRef Search ADS PubMed 7 Kilgore K.S. , Schmid E. , Shanley T.P. , Flory C.M. , Maheswari V. , Tramontini N.L. , Cohen H. , Ward P.A. , Friedl H.P. , Warren J.S. ( 1997 ) Sublytic concentrations of the membrane attack complex of complement induce endothelial interleukin-8 and monocyte chemoattractant protein-1 through nuclear factor-kappa B activation . Am. J. Pathol ., 150 , 2019 – 2031 . Google Scholar PubMed 8 Niculescu F. , Badea T. , Rus H. ( 1999 ) Sublytic C5b-9 induces proliferation of human aortic smooth muscle cells: role of mitogen activated protein kinase and phosphatidylinositol 3-kinase . Atherosclerosis , 142 , 47 – 56 . Google Scholar CrossRef Search ADS PubMed 9 Triantafilou K. , Hughes T.R. , Triantafilou M. , Morgan B.P. ( 2013 ) The complement membrane attack complex triggers intracellular Ca2+ fluxes leading to NLRP3 inflammasome activation . J. Cell Sci ., 126 , 2903 – 2913 . Google Scholar CrossRef Search ADS PubMed 10 Vlaicu S.I. , Tegla C.A. , Cudrici C.D. , Danoff J. , Madani H. , Sugarman A. , Niculescu F. , Mircea P.A. , Rus V. , Rus H. ( 2013 ) Role of C5b-9 complement complex and response gene to complement-32 (RGC-32) in cancer . Immunol. Res ., 56 , 109 – 121 . Google Scholar CrossRef Search ADS PubMed 11 Mullins R.F. , Dewald A.D. , Streb L.M. , Wang K. , Kuehn M.H. , Stone E.M. ( 2011 ) Elevated membrane attack complex in human choroid with high risk complement factor H genotypes . Exp. Eye. Res ., 93 , 565 – 567 . Google Scholar CrossRef Search ADS PubMed 12 Mullins R.F. , Schoo D.P. , Sohn E.H. , Flamme-Wiese M.J. , Workamelahu G. , Johnston R.M. , Wang K. , Tucker B.A. , Stone E.M. ( 2014 ) The membrane attack complex in aging human choriocapillaris: relationship to macular degeneration and choroidal thinning . Am. J. Pathol ., 184 , 3142 – 3153 . Google Scholar CrossRef Search ADS PubMed 13 Lueck K. , Wasmuth S. , Williams J. , Hughes T.R. , Morgan B.P. , Lommatzsch A. , Greenwood J. , Moss S.E. , Pauleikhoff D. ( 2011 ) Sub-lytic C5b-9 induces functional changes in retinal pigment epithelial cells consistent with age-related macular degeneration . Eye (London) , 25 , 1074 – 1082 . Google Scholar CrossRef Search ADS 14 Kunchithapautham K. , Rohrer B. ( 2011 ) Sublytic membrane-attack-complex (MAC) activation alters regulated rather than constitutive vascular endothelial growth factor (VEGF) secretion in retinal pigment epithelium monolayers . J. Biol. Chem ., 286 , 23717 – 23724 . Google Scholar CrossRef Search ADS PubMed 15 Seddon J.M. , Yu Y. , Miller E.C. , Reynolds R. , Tan P.L. , Gowrisankar S. , Goldstein J.I. , Triebwasser M. , Anderson H.E. , Zerbib J. et al. ( 2013 ) Rare variants in CFI, C3 and C9 are associated with high risk of advanced age-related macular degeneration . Nat. Genet ., 45 , 1366 – 1370 . Google Scholar CrossRef Search ADS PubMed 16 Nishiguchi K.M. , Yasuma T.R. , Tomida D. , Nakamura M. , Ishikawa K. , Kikuchi M. , Ohmi Y. , Niwa T. , Hamajima N. , Furukawa K. et al. ( 2012 ) C9-R95X polymorphism in patients with neovascular age-related macular degeneration . Invest. Ophthalmol. Vis. Sci ., 53 , 508 – 512 . Google Scholar CrossRef Search ADS PubMed 17 Geerlings M.J. , Kremlitzka M. , Bakker B. , Nilsson S.C. , Saksens N.T. , Lechanteur Y.T. , Pauper M. , Corominas J. , Fauser S. , Hoyng C.B. et al. ( 2017 ) The functional effect of rare variants in complement genes on C3b degradation in patients with age-related macular degeneration . JAMA Ophthalmol ., 135 , 39 – 46 . Google Scholar CrossRef Search ADS PubMed 18 Yu Y. , Wagner E.K. , Souied E.H. , Seitsonen S. , Immonen I.J. , Happola P. , Raychaudhuri S. , Daly M.J. , Seddon J.M. ( 2016 ) Protective coding variants in CFH and PELI3 and a variant near CTRB1 are associated with age-related macular degenerationdagger . Hum. Mol. Genet ., 25 , 5276 – 5285 . Google Scholar PubMed 19 Saksens N.T. , Geerlings M.J. , Bakker B. , Schick T. , Daha M.R. , Fauser S. , Boon C.J. , de Jong E.K. , Hoyng C.B. , den Hollander A.I. ( 2016 ) Rare genetic variants associated with development of age-related macular degeneration . JAMA Ophthalmol ., 134 , 287 – 293 . Google Scholar CrossRef Search ADS PubMed 20 Saksens N.T. , Lechanteur Y.T. , Verbakel S.K. , Groenewoud J.M. , Daha M.R. , Schick T. , Fauser S. , Boon C.J. , Hoyng C.B. , den Hollander A.I. ( 2016 ) Analysis of risk alleles and complement activation levels in familial and non-familial age-related macular degeneration . PLoS One , 11 , e0144367. Google Scholar CrossRef Search ADS PubMed 21 Smailhodzic D. , van Asten F. , Blom A.M. , Mohlin F.C. , den Hollander A.I. , van de Ven J.P. , van Huet R.A. , Groenewoud J.M. , Tian Y. , Berendschot T.T. et al. ( 2014 ) Zinc supplementation inhibits complement activation in age-related macular degeneration . PLoS One , 9 , e112682. Google Scholar CrossRef Search ADS PubMed 22 Horiuchi T. , Nishizaka H. , Kojima T. , Sawabe T. , Niho Y. , Schneider P.M. , Inaba S. , Sakai K. , Hayashi K. , Hashimura C. et al. ( 1998 ) A non-sense mutation at Arg95 is predominant in complement 9 deficiency in Japanese . J. Immunol ., 160 , 1509 – 1513 . Google Scholar PubMed 23 Reynolds R. , Hartnett M.E. , Atkinson J.P. , Giclas P.C. , Rosner B. , Seddon J.M. ( 2009 ) Plasma complement components and activation fragments: associations with age-related macular degeneration genotypes and phenotypes . Invest. Ophthalmol. Vis. Sci ., 50 , 5818 – 5827 . Google Scholar CrossRef Search ADS PubMed 24 Smailhodzic D. , Muether P.S. , Chen J. , Kwestro A. , Zhang A.Y. , Omar A. , Van de Ven J.P. , Keunen J.E. , Kirchhof B. , Hoyng C.B. et al. ( 2012 ) Cumulative effect of risk alleles in CFH, ARMS2, and VEGFA on the response to ranibizumab treatment in age-related macular degeneration . Ophthalmology , 119 , 2304 – 2311 . Google Scholar CrossRef Search ADS PubMed 25 Scholl H.P. , Charbel Issa P. , Walier M. , Janzer S. , Pollok-Kopp B. , Borncke F. , Fritsche L.G. , Chong N.V. , Fimmers R. , Wienker T. et al. ( 2008 ) Systemic complement activation in age-related macular degeneration . PLoS One , 3 , e2593. Google Scholar CrossRef Search ADS PubMed 26 Ristau T. , Paun C. , Ersoy L. , Hahn M. , Lechanteur Y. , Hoyng C. , de Jong E.K. , Daha M.R. , Kirchhof B. , den Hollander A.I. et al. ( 2014 ) Impact of the common genetic associations of age-related macular degeneration upon systemic complement component C3d levels . PLoS One , 9 , e93459. Google Scholar CrossRef Search ADS PubMed 27 Schick T. , Steinhauer M. , Aslanidis A. , Altay L. , Karlstetter M. , Langmann T. , Kirschfink M. , Fauser S. ( 2017 ) Local complement activation in aqueous humor in patients with age-related macular degeneration . Eye (London) , 31 , 810 – 813 . Google Scholar CrossRef Search ADS 28 Tegla C.A. , Cudrici C. , Patel S. , Trippe R. III , Rus V. , Niculescu F. , Rus H. ( 2011 ) Membrane attack by complement: the assembly and biology of terminal complement complexes . Immunol. Res ., 51 , 45 – 60 . Google Scholar CrossRef Search ADS PubMed 29 Podack E.R. , Tschopp J. ( 1982 ) Polymerization of the ninth component of complement (C9): formation of poly(C9) with a tubular ultrastructure resembling the membrane attack complex of complement . Proc. Natl. Acad. Sci. USA ., 79 , 574 – 578 . Google Scholar CrossRef Search ADS 30 Dudkina N.V. , Spicer B.A. , Reboul C.F. , Conroy P.J. , Lukoyanova N. , Elmlund H. , Law R.H. , Ekkel S.M. , Kondos S.C. , Goode R.J. et al. ( 2016 ) Structure of the poly-C9 component of the complement membrane attack complex . Nat. Commun ., 7 , 10588. Google Scholar CrossRef Search ADS PubMed 31 Samuel W. , Jaworski C. , Postnikova O.A. , Kutty R.K. , Duncan T. , Tan L.X. , Poliakov E. , Lakkaraju A. , Redmond T.M. ( 2017 ) Appropriately differentiated ARPE-19 cells regain phenotype and gene expression profiles similar to those of native RPE cells . Mol. Vis ., 23 , 60 – 89 . Google Scholar PubMed 32 Lek M. , Karczewski K.J. , Minikel E.V. , Samocha K.E. , Banks E. , Fennell T. , O’Donnell-Luria A.H. , Ware J.S. , Hill A.J. , Cummings B.B. et al. ( 2016 ) Analysis of protein-coding genetic variation in 60,706 humans . Nature , 536 , 285 – 291 . Google Scholar CrossRef Search ADS PubMed 33 Ristau T. , Ersoy L. , Hahn M. , den Hollander A.I. , Kirchhof B. , Liakopoulos S. , Fauser S. ( 2014 ) Nongenetic risk factors for neovascular age-related macular degeneration . Invest. Ophthalmol. Vis. Sci ., 55 , 5228 – 5232 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Human Molecular Genetics Oxford University Press

Functional analyses of rare genetic variants in complement component C9 identified in patients with age-related macular degeneration

Loading next page...
 
/lp/ou_press/functional-analyses-of-rare-genetic-variants-in-complement-component-fPN5UYD0Sm
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oup.com
ISSN
0964-6906
eISSN
1460-2083
D.O.I.
10.1093/hmg/ddy178
Publisher site
See Article on Publisher Site

Abstract

Abstract Age-related macular degeneration (AMD) is a progressive disease of the central retina and the leading cause of irreversible vision loss in the western world. The involvement of abnormal complement activation in AMD has been suggested by association of variants in genes encoding complement proteins with disease development. A low-frequency variant (p.P167S) in the complement component C9 (C9) gene was recently shown to be highly associated with AMD; however, its functional outcome remains largely unexplored. In this study, we reveal five novel rare genetic variants (p.M45L, p.F62S, p.G126R, p.T170I and p.A529T) in C9 in AMD patients, and evaluate their functional effects in vitro together with the previously identified (p.R118W and p.P167S) C9 variants. Our results demonstrate that the concentration of C9 is significantly elevated in patients’ sera carrying the p.M45L, p.F62S, p.P167S and p.A529T variants compared with non-carrier controls. However, no difference can be observed in soluble terminal complement complex levels between the carrier and non-carrier groups. Comparing the polymerization of the C9 variants we reveal that the p.P167S mutant spontaneously aggregates, while the other mutant proteins (except for C9 p.A529T) fail to polymerize in the presence of zinc. Altered polymerization of the p.F62S and p.P167S proteins associated with decreased lysis of sheep erythrocytes and adult retinal pigment epithelial-19 cells by carriers’ sera. Our data suggest that the analyzed C9 variants affect only the secretion and polymerization of C9, without influencing its classical lytic activity. Future studies need to be performed to understand the implications of the altered polymerization of C9 in AMD pathology. Introduction Age-related macular degeneration (AMD; MIM# 603075), a progressive eye disorder, is the major cause of irreversible vision loss in the western world (1). The disease is multifactorial, involving both environmental and genetic factors in its pathogenesis (2). Genetic alterations are estimated to account for 46–71% of variability in disease risk (3). A large component of the heritability of AMD can be explained by genetic variants in the alternative pathway of the complement system. A recently published genome-wide association study detected 52 (45 common and 7 rare) variants at 34 genomic regions that are independently associated with AMD. More than one third of these variants reside in or near genes encoding for components of the complement system (4). Complement is a crucial part of innate immunity, providing clearance of foreign and altered-self structures. Activation of its cascade results in enzymatic-cleavage of the central component, C3 into C3a and C3b. C3b is a crucial component of C3 and C5 convertases, allowing further propagation of the cascade into the terminal pathway where C5b-8 complexes incorporate into the membrane. After binding several copies of complement component C9 (C9) to C5b-8, the pore-forming membrane attack complex (MAC)/terminal complement complex (TCC) is assembled (5,6). MAC kills the target cell by inducing cell lysis, or at a reduced, the so-called sublytic concentration, it can provoke a wide array of physiologic responses ranging from apoptosis to pro-inflammatory cytokine secretion (7–10). The potential involvement of C9 in AMD pathology has been suggested by increased MAC deposition in the retina of AMD patients and the correlation between the amount of MAC and the loss of RPE cells (11,12). In addition, sublytic MAC deposition on RPE cells has been reported to induce secretion of pro-inflammatory cytokines and vascular endothelial growth factor, contributing to the development of advanced AMD (13,14). Recently, three rare genetic variants in C9 were reported in association with AMD (15–17) namely p.R95X, p.R118W and p.P167S. The p.P167S variant was reported to be highly associated with AMD risk in multiple studies (4,15,17–19). More recently, a genetic burden of C9 variants was described in two separate AMD cohorts. The first study identified 13 rare variants (P-value 2.4×10−08) (18) and the second study (Corominas et al., manuscript submitted) revealed 17 rare variants (P-value 5.01×10−03) in C9; however in both studies the burden did not remain significant after correction for multiple comparisons. In a recent study, we demonstrated that the p.P167S variant leads to increased serum concentration of the protein (17). However, the functional consequences of the p.P167S variant and of other genetic alterations in C9 remain unclear. In this study, we aimed to further elucidate the functional effects of C9 variants in vitro in order to understand the role of C9 in AMD pathogenesis. Results Genetic alterations identified in C9 Through whole-exome sequencing in 793 unrelated individuals (662 AMD cases and 131 controls), we identified five novel rare variants in C9: p.M45L, p.F62S, p.G126R, p.T170I and p.A529T, in addition to previously reported variants p.R118W and p.P167S (Table 1). Genotyping of these 7 variants in 1896 unrelated AMD cases and 1499 unrelated control individuals (Supplementary Material, Table S1) identified 127 rare variant carriers. The identified variants are present in different domains of the protein: the thrombospondin type 1 (TSP1), the low-density lipoprotein receptor type A (LDLRA), the MAC/perforin (MACPF) and the epidermal growth factor (EGF)-like domains (Fig. 1A). All the identified variants are non-synonymous point mutations resulting in amino acid changes in the mature protein (Table 1). Table 1. Overview of rare genetic variants in C9 investigated in the study Variant ExAc freq (%) PhyloPa Granthama SIFTb PolyPhenb CADDcphred Literature (1st report) refSNP Amino acidd cDNAd Novel rs41271047 p.M45L c.133A>T 0.21 0.734 15 T B 9.639 Novel rs140251849 p.F62S c.185T>C 0.01 2.187 155 T D 25 Geerlings et al. (17) rs147701327 p.R118W c.352C>T 0.02 1.393 101 D P 28 Novel rs199939436 p.G126R c.376G>A 0.03 3.758 125 D D 34 Novel rs34882957 p.P167S c.499C>T 0.470 3.279 74 D D 25.3 Seddon et al. (15) NA p.T170I c.509C>T NA 2.087 89 D D 24.7 Novel rs137891079 p.A529T c.1585G>A 0.050 −1.191 58 D D 0.028 Novel Variant ExAc freq (%) PhyloPa Granthama SIFTb PolyPhenb CADDcphred Literature (1st report) refSNP Amino acidd cDNAd Novel rs41271047 p.M45L c.133A>T 0.21 0.734 15 T B 9.639 Novel rs140251849 p.F62S c.185T>C 0.01 2.187 155 T D 25 Geerlings et al. (17) rs147701327 p.R118W c.352C>T 0.02 1.393 101 D P 28 Novel rs199939436 p.G126R c.376G>A 0.03 3.758 125 D D 34 Novel rs34882957 p.P167S c.499C>T 0.470 3.279 74 D D 25.3 Seddon et al. (15) NA p.T170I c.509C>T NA 2.087 89 D D 24.7 Novel rs137891079 p.A529T c.1585G>A 0.050 −1.191 58 D D 0.028 Novel a Higher PhyloP [range −14; 6.4] and Grantham [range 0–215] scores correlate with a higher conservation. b Sorting Intolerant from Tolerant (SIFT) and PolyPhen2 classification: T, tolerated; B, benign; D, damaging; P, pathogenic. c Combined Annotation Dependent Depletion (CADD) scores the deleteriousness of genetic variants in the human genome (log10 scale, i.e. 10 predicts variant in top 10%, 20 in top 1%, 30 in top 0.1%, etc. of reference most deleterious variants). d Signal peptide included in numbering. NA, non-available. Table 1. Overview of rare genetic variants in C9 investigated in the study Variant ExAc freq (%) PhyloPa Granthama SIFTb PolyPhenb CADDcphred Literature (1st report) refSNP Amino acidd cDNAd Novel rs41271047 p.M45L c.133A>T 0.21 0.734 15 T B 9.639 Novel rs140251849 p.F62S c.185T>C 0.01 2.187 155 T D 25 Geerlings et al. (17) rs147701327 p.R118W c.352C>T 0.02 1.393 101 D P 28 Novel rs199939436 p.G126R c.376G>A 0.03 3.758 125 D D 34 Novel rs34882957 p.P167S c.499C>T 0.470 3.279 74 D D 25.3 Seddon et al. (15) NA p.T170I c.509C>T NA 2.087 89 D D 24.7 Novel rs137891079 p.A529T c.1585G>A 0.050 −1.191 58 D D 0.028 Novel Variant ExAc freq (%) PhyloPa Granthama SIFTb PolyPhenb CADDcphred Literature (1st report) refSNP Amino acidd cDNAd Novel rs41271047 p.M45L c.133A>T 0.21 0.734 15 T B 9.639 Novel rs140251849 p.F62S c.185T>C 0.01 2.187 155 T D 25 Geerlings et al. (17) rs147701327 p.R118W c.352C>T 0.02 1.393 101 D P 28 Novel rs199939436 p.G126R c.376G>A 0.03 3.758 125 D D 34 Novel rs34882957 p.P167S c.499C>T 0.470 3.279 74 D D 25.3 Seddon et al. (15) NA p.T170I c.509C>T NA 2.087 89 D D 24.7 Novel rs137891079 p.A529T c.1585G>A 0.050 −1.191 58 D D 0.028 Novel a Higher PhyloP [range −14; 6.4] and Grantham [range 0–215] scores correlate with a higher conservation. b Sorting Intolerant from Tolerant (SIFT) and PolyPhen2 classification: T, tolerated; B, benign; D, damaging; P, pathogenic. c Combined Annotation Dependent Depletion (CADD) scores the deleteriousness of genetic variants in the human genome (log10 scale, i.e. 10 predicts variant in top 10%, 20 in top 1%, 30 in top 0.1%, etc. of reference most deleterious variants). d Signal peptide included in numbering. NA, non-available. Figure 1. View largeDownload slide Schematic illustration of C9 domains with the identified C9 variants. SP, signal peptide; TSP1, thrombospondin type 1; LDLRA, low-density lipoprotein receptor type A; MACPF, the MAC/perforin domain; EGF, the epidermal growth factor-like domains. Figure 1. View largeDownload slide Schematic illustration of C9 domains with the identified C9 variants. SP, signal peptide; TSP1, thrombospondin type 1; LDLRA, low-density lipoprotein receptor type A; MACPF, the MAC/perforin domain; EGF, the epidermal growth factor-like domains. In our case–control cohort, none of the rare C9 variants were individually associated with AMD pathogenesis (Supplementary Material, Table S2). Nevertheless, regarding the previously shown high association of the p.P167S variant with AMD in large case–control studies (4,15,18,20), and the recently reported burden of rare variants in C9 [Corominas et al., manuscript submitted and (18)], we set out to identify the functional effect of these variants on the C9 protein, in order to place them into the context of AMD pathogenesis. To this end, 128 serum and 95 plasma samples of the identified carriers were tested in functional assays and compared with 156 sera (78 with AMD and 78 without the disease) and 155 plasma samples (77 with AMD and 78 without AMD) of age-matched non-carrier individuals (Supplementary Material, Table S2). Serum C9 and plasma TCC level of C9 carriers To analyze whether the identified genetic variants affect protein synthesis and secretion, the concentration of C9 was measured in sera of 127 AMD patients carrying rare genetic variants in C9 and compared with 156 non-carriers with (n = 78) or without (n = 78) AMD. We found that the p.M45L, p.F62S, p.P167S and p.A529T variants lead to a significantly increased C9 level compared with non-carrier controls (Fig. 2A). Sera carrying the p.R118W and p.G126R variants did not significantly differ in C9 levels between carriers and non-carriers (Fig. 2A). Figure 2. View largeDownload slide Measurement of serum C9 and plasma sTCC levels in C9 carriers. Concentrations of C9 in sera (A) and sTCC level in plasma (B) of patients carrying rare genetic variants in C9 were measured by ELISA and compared with non-carriers with or without AMD. Differences with P < 0.05 were considered statistically significant and signed with black (carriers versus non-carriers with AMD) or with light grey (carriers versus non-carriers without AMD) asterisks (Kruskal–Wallis test with Dunn’s multiple comparison, nsP > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). Data are shown as median μg/ml secreted C9 or AU/ml sTCC level (the latter on log10 scale) with interquartile range of four (C9 ELISA) or three (sTCC ELISA) independent experiments. Figure 2. View largeDownload slide Measurement of serum C9 and plasma sTCC levels in C9 carriers. Concentrations of C9 in sera (A) and sTCC level in plasma (B) of patients carrying rare genetic variants in C9 were measured by ELISA and compared with non-carriers with or without AMD. Differences with P < 0.05 were considered statistically significant and signed with black (carriers versus non-carriers with AMD) or with light grey (carriers versus non-carriers without AMD) asterisks (Kruskal–Wallis test with Dunn’s multiple comparison, nsP > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). Data are shown as median μg/ml secreted C9 or AU/ml sTCC level (the latter on log10 scale) with interquartile range of four (C9 ELISA) or three (sTCC ELISA) independent experiments. Binding of C9 to soluble C5b-8 complexes results in formation of soluble TCC (sTCC), which is a sign of ongoing complement activation. Therefore, we measured sTCC levels in patient and control plasma samples of carriers and non-carriers of rare C9 variants. Despite the differences in serum C9 concentration, however, we did not detect any alteration in the systemic level of sTCC in plasma of C9 carriers compared with non-carrier controls (Fig. 2B). Expression and secretion of wild-type and mutant C9 proteins by human embryonic kidney 293 cells To study the effect of the identified variants on protein secretion, human embryonic kidney 293 (HEK293) cells were transfected with either the wild-type (WT) or mutant C9 constructs and concentrations of expressed C9 in cell lysates and secreted C9 in supernatants were analyzed using western blot and enzyme-linked immunosorbent assay (ELISA), respectively. Five of the mutant C9 proteins (p.M54L, p.F62S, p.R118W, p.G126R, p.P167S) were expressed at similar levels as the WT protein in monomer forms, while the C9 p.T170I and p.A529T mutants showed decreased concentration in the cell lysates (Fig. 3A and B). Furthermore, polymerized form of C9 could be observed in case of C9 p.G126R and p.P167S (Fig. 3B). In agreement with results of C9 ELISA using patients’ sera, C9 p.F62S was secreted at higher concentration into the cell supernatant than the WT protein (Fig. 3C). In addition to its reduced expression in the cell lysate, concentration of the p.T170I mutant protein in the supernatant was also reduced. Level of the C9 p.R118W protein in the supernatant was decreased, while its expression in the cell lysate was normal. Figure 3. View largeDownload slide Expression and secretion of recombinant WT and mutant C9 proteins. HEK293 cells were transiently transfected with WT or mutant C9-pCEP4 constructs. As negative control, empty pCEP4 vector was used. Expression of C9 in the cell lysates was measured by western blot. Results shown are either densitometric analysis of three transfections and illustrated as % of WT C9 expression with mean ± SD (A) or one representative western blot (B). (C) Concentration of C9 in the secreted supernatants was measured by ELISA. Graph shows mean ± SD % of WT C9 concentration and is result of three independent transfections measured in duplicates. Differences with P < 0.05 were considered statistically significant and compared with WT C9 (one-way ANOVA with Dunnett’s multiple comparison, nsP > 0.05, *P < 0.05, **P < 0.01, ****P < 0.0001). Figure 3. View largeDownload slide Expression and secretion of recombinant WT and mutant C9 proteins. HEK293 cells were transiently transfected with WT or mutant C9-pCEP4 constructs. As negative control, empty pCEP4 vector was used. Expression of C9 in the cell lysates was measured by western blot. Results shown are either densitometric analysis of three transfections and illustrated as % of WT C9 expression with mean ± SD (A) or one representative western blot (B). (C) Concentration of C9 in the secreted supernatants was measured by ELISA. Graph shows mean ± SD % of WT C9 concentration and is result of three independent transfections measured in duplicates. Differences with P < 0.05 were considered statistically significant and compared with WT C9 (one-way ANOVA with Dunnett’s multiple comparison, nsP > 0.05, *P < 0.05, **P < 0.01, ****P < 0.0001). Lytic activity of the recombinant, mutant C9 proteins In order to explore the impact of the identified variants on the protein function, independent of other serum components, the mutant C9 proteins were recombinantly produced by HEK293F cells and purified by Ni2+-affinity chromatography and gel filtration from the secreted supernatant. The isolated proteins were visualized by both Silver staining and western blotting (Supplementary Material, Fig. S1), confirming the presence of pure, monomeric C9. Firstly, the recombinant, purified C9 proteins were compared in their ability to lyse erythrocytes. To this end, sensitized sheep erythrocytes were treated with C9-depleted serum, which was reconstituted with the WT or mutant C9 proteins. Six of the mutant C9 proteins (p.M45L, p.F62S, p.R118W, p.G126R, p.T170I and p.A529T) had normal lytic activity, while the C9 p.P167S protein showed a slight, but significant reduction in erythrocyte lysis (Fig. 4A). Figure 4. View largeDownload slide Lytic activity of recombinant C9 proteins. Sheep erythrocytes (A) or ARPE-19 cells (B) were treated with C9-depleted serum supplemented with the recombinant WT or mutant C9 variants. (A) Lysis of erythrocytes was analyzed via measurement of released hemoglobin at 405 nm. Data are expressed as median with interquartile range of water-induced maximum lysis and are results of three independent experiments carried out in duplicate. (B) MAC-induced cytotoxicity was measured via lactate-dehydrogenase (LDH) release from ARPE-19 cells. Data expressed are mean ± SD of lysis buffer-induced maximum lysis and are results of three independent experiments measured in duplicate. Differences with P < 0.05 were considered statistically significant (one-way ANOVA with Dunnett’s multiple comparison, nsP > 0.05, **P < 0.01). Figure 4. View largeDownload slide Lytic activity of recombinant C9 proteins. Sheep erythrocytes (A) or ARPE-19 cells (B) were treated with C9-depleted serum supplemented with the recombinant WT or mutant C9 variants. (A) Lysis of erythrocytes was analyzed via measurement of released hemoglobin at 405 nm. Data are expressed as median with interquartile range of water-induced maximum lysis and are results of three independent experiments carried out in duplicate. (B) MAC-induced cytotoxicity was measured via lactate-dehydrogenase (LDH) release from ARPE-19 cells. Data expressed are mean ± SD of lysis buffer-induced maximum lysis and are results of three independent experiments measured in duplicate. Differences with P < 0.05 were considered statistically significant (one-way ANOVA with Dunnett’s multiple comparison, nsP > 0.05, **P < 0.01). To further investigate the effects in a model more relevant for AMD, we compared the lytic activity of the WT and mutant C9 proteins on the retinal pigment epithelial cell line ARPE-19. To induce MAC deposition, ARPE-19 cells were pre-incubated with a function-blocking anti-CD59 antibody before treatment with C9-depleted serum, supplemented with either the WT or mutant recombinant C9 proteins. In agreement with the results obtained using sheep erythrocytes (Fig. 4A), we did not observe any differences between the WT and mutant C9 proteins in their cytotoxic activity on ARPE-19 cells (Fig. 4B), suggesting that the variants do not affect the classical, lytic function of the protein. Lytic activity of sera of C9 carriers Next, we tested whether the physiological difference in C9 concentration in sera of carriers of C9 variants may result in elevated lysis of the target cells, a phenomenon that may lead to pathological changes and destruction of the retina. To measure the C9-dependent lytic activity of sera independent of other complement components, a modified hemolytic assay was designed. Sensitized erythrocytes or ARPE-19 cells were incubated with C9-depleted serum in dextrose gelatin veronal buffer (DGVB++), which allows complement activation and deposition of C5b-8 complexes on the cell membrane. Thereafter, the cells were incubated with sera of carriers or non-carrier controls diluted in ethylenediaminetetraacetic acid (EDTA) gelatin veronal buffer (EDTA-GVB), which blocks complement activation and novel C5b-8 complex formation, but allows integration of C9 in the pre-formed C5b-8 complexes and induces lysis dependent on both C9 concentration and its functional activity in the serum. In spite of the significantly increased C9 concentration in sera of carriers of the p.M45L, p.F62S, p.P167S and p.A529T variants, the elevated C9 level did not cause increased lysis of erythrocytes (Fig. 5A) or ARPE-19 cells (Fig. 5B and C). On the contrary, we observed a slight but significant decrease in lytic activity of C9 p.F62S on both erythrocytes and ARPE-19 cells compared with non-carrier controls without AMD. Furthermore, sera of carriers of the C9 p.P167S variant showed decrease in lytic activity on erythrocytes compared with non-carriers with or without AMD. Figure 5. View largeDownload slide C9-dependent lytic activity of sera carrying rare genetic variants in C9. Lytic activity of sera was measured via incubation of sheep erythrocytes (A) or ARPE-19 cells (B, C) with C9-depleted serum supplemented with EDTA-GVB diluted sera. (A) Lysis of erythrocytes was analyzed via measurement of released hemoglobin at 405 nm. Data are expressed as median with interquartile range of water-induced maximum lysis and are results of four independent experiments. MAC-induced cytotoxicity of ARPE-19 cells was measured via LDH release (B) and by analysis of AnnexinV+ ZombieAqua+ (late apoptotic) cells by flow cytometry (C). Data expressed are median with interquartile range of lysis buffer-induced maximum LDH release (B) and % of late apoptotic cells (C) of two independent experiments measured in duplicate. Differences with P < 0.05 were considered statistically significant and signed with black (carriers versus non-carriers with AMD) or with light grey (carriers versus non-carriers without AMD) asterisks (Kruskal–Wallis test with Dunn’s multiple comparison, nsP > 0.05, *P < 0.05, **P < 0.01, ****P <0.0001). Figure 5. View largeDownload slide C9-dependent lytic activity of sera carrying rare genetic variants in C9. Lytic activity of sera was measured via incubation of sheep erythrocytes (A) or ARPE-19 cells (B, C) with C9-depleted serum supplemented with EDTA-GVB diluted sera. (A) Lysis of erythrocytes was analyzed via measurement of released hemoglobin at 405 nm. Data are expressed as median with interquartile range of water-induced maximum lysis and are results of four independent experiments. MAC-induced cytotoxicity of ARPE-19 cells was measured via LDH release (B) and by analysis of AnnexinV+ ZombieAqua+ (late apoptotic) cells by flow cytometry (C). Data expressed are median with interquartile range of lysis buffer-induced maximum LDH release (B) and % of late apoptotic cells (C) of two independent experiments measured in duplicate. Differences with P < 0.05 were considered statistically significant and signed with black (carriers versus non-carriers with AMD) or with light grey (carriers versus non-carriers without AMD) asterisks (Kruskal–Wallis test with Dunn’s multiple comparison, nsP > 0.05, *P < 0.05, **P < 0.01, ****P <0.0001). Polymerization of the recombinant, mutant C9 proteins During purification of recombinant C9 proteins we observed that the p.P167S mutant tends to aggregate and form high molecular weight multimers. To evaluate and compare the polymerization tendency of the mutant C9 proteins, recombinant, purified C9 was incubated in the presence or absence of zinc ions, which accelerate C9 polymerization and are used as dietary supplement to reduce AMD progression (21). We observed that the p.P167S mutant spontaneously aggregates (Fig. 6), while the other mutant proteins polymerize normally in the absence of zinc. However, the p.M45L, p.F62S, p.R118W, p.G126R and p.T170I mutant proteins showed impaired polymerization in the presence of zinc in contrast to C9 p.P167S, which showed higher polymerization than the WT protein. Polymerization ability of the p.A529T mutant protein was unaltered. Figure 6. View largeDownload slide Polymerization of WT and mutant C9 proteins in the presence or absence of Zn2+ ions. Recombinant WT or mutant C9 (5 μM) were allowed to polymerize in the presence or absence of 15 μM Zn2+ ions at 37°C for 4 h. The proteins were separated by electrophoresis on a 5–10% polyacrylamide gradient gel and polymers were visualized by Silver staining. (A) One representative experiment of three is shown. (B) Densitometric ratio of polymer and monomer WT and mutant C9 variants were calculated by ImageLab software. Results are illustrated as mean ± SD polymer/monomer ratios of C9 mutant proteins of three independent experiments. Differences between WT and mutant C9 proteins with P < 0.05 were considered statistically significant and signed with light grey (spontaneous polymerization) or with black (Zn2+-induced polymerization) asterisks (two-way ANOVA with Dunnett’s multiple comparison, nsP > 0.05, **P < 0.01, ****P < 0.0001). Figure 6. View largeDownload slide Polymerization of WT and mutant C9 proteins in the presence or absence of Zn2+ ions. Recombinant WT or mutant C9 (5 μM) were allowed to polymerize in the presence or absence of 15 μM Zn2+ ions at 37°C for 4 h. The proteins were separated by electrophoresis on a 5–10% polyacrylamide gradient gel and polymers were visualized by Silver staining. (A) One representative experiment of three is shown. (B) Densitometric ratio of polymer and monomer WT and mutant C9 variants were calculated by ImageLab software. Results are illustrated as mean ± SD polymer/monomer ratios of C9 mutant proteins of three independent experiments. Differences between WT and mutant C9 proteins with P < 0.05 were considered statistically significant and signed with light grey (spontaneous polymerization) or with black (Zn2+-induced polymerization) asterisks (two-way ANOVA with Dunnett’s multiple comparison, nsP > 0.05, **P < 0.01, ****P < 0.0001). Discussion In this study, we identified five novel, non-synonymous rare genetic variants in C9 (p.M45L, p.F62S, p.G126R, p.T170I and p.A529T), in addition to the previously described C9 p.R118W (17) and p.P167S (15) variants. Although the identified C9 variants (except for C9 p.M45L) were more prevalent in individuals affected by AMD compared with control individuals, the limited number of subjects analyzed in combination with the low allele frequency of the C9 variants left their individual association to AMD statistically inconclusive (Supplementary Material, Table S2). Nevertheless, the C9 variants revealed in our study further expand the repertoire of genetic variants in genes encoding components of the complement cascade and confirm the earlier reported correlations of C9 p.R118W and C9 p.P167S to enhanced diseased risk. In previous studies, three genetic variants in C9 have been associated with AMD. The nonsense mutation p.R95X is most common among Japanese and leads to complete or partial C9 deficiency (22). This variant is protective for AMD, conferring a 4.7-fold reduction in disease risk, and is correlated to decreased vascular endothelial growth factor levels (16). The p.R118W C9 variant was identified with a relatively low allelic odds-ratio (1.12) and normal C9 level in the carriers compared with non-carriers (17). C9 p.P167S increases risk for AMD (4,15,17–19) and results in elevated serum C9 level in carriers compared with non-carriers (17). In our cohort, we did not detect any significant alteration in systemic complement activation levels, measured in the form of plasma sTCC between the carriers and non-carriers (Fig. 2B). This is in accordance with previous studies, reporting no significant difference in systemic sTCC level between AMD patients and non-carrier control groups (23–25). Nevertheless, increased MAC was previously observed in Bruch’s membrane and the choriocapillaris of human eyes, gradually increasing with age. The elevated MAC was especially evident in eyes affected by AMD, even more so in carriers of the CFH risk allele (11,12). These findings suggest that changes in MAC may only be detected locally in the eye and further strengthen the role of local rather than systemic complement activation in AMD pathology, which support recent data showing increased complement activation in aqueous humor rather than in plasma of patients with AMD (26,27). We found that carriers of the p.M45L, p.F62S, p.P167S and p.A529T variants have increased serum C9 level compared with non-carrier controls (Fig. 2A) while C9 concentrations of p.R118W and p.G126R was unaltered compared with sera of non-carrier controls. These data confirm our previous results reporting normal C9 level in sera carrying the C9 p.R118W variant and elevated levels in carriers of the C9 p.P167S variant (17). Interestingly, secretion of the recombinant C9 proteins by HEK293 cells was higher only for the C9 p.F62S mutant (Fig. 3C). The differences between concentrations of various mutants of C9 in sera and HEK293 cell supernatants can be partially explained by the multicellular source and consumption of C9 in vivo in comparison to the single cell origin of the protein in HEK293 cell cultures. Despite this discrepancy, data obtained from transfected HEK293 cells contribute to the better understanding of the in vivo observed alterations in C9 concentration between the distinct carriers. For example, C9 p.F62S was expressed at the same level as WT C9 (Fig. 3A and B) but increased in the secreted supernatant (Fig. 3C). These results indicate that the mutation causes increased protein secretion and explain the elevated concentration of C9 p.F62S in the carrier’s sera (Fig. 2A). Furthermore, although secretion of recombinant C9 p.R118W and p.T170I was lower than of the WT protein, expression of these mutant proteins was also lower (in case of p.T170I almost completely absent), indicating that the in vitro observed decreased secretion is mainly owing to their diminished transcription. Indeed, serum concentration of C9 p.R118W did not differ from non-carrier controls. Unfortunately, no serum samples were available for the p.T170I carrier (n = 1) and hence, its mutational effect on C9 secretion in vivo needs to be explored in future studies. Despite the significantly elevated C9 level in carriers of the p.M45L, p.F62S, p.P167S and p.A529T variants (Fig. 2A), we could not detect an increase in C9-dependent lytic activity of patients’ sera carrying rare-genetic variants in C9 (Fig. 5). More unexpectedly, sera of the p.F62S and p.P167S C9 carriers caused decreased lysis of sheep erythrocytes and ARPE-19 cells. The lytic activity of the other C9 proteins was normal despite increased serum levels (Fig. 5A and C). Since one would assume that increased C9 concentration results in enhanced MAC formation and thus, killing of the target cells, these results are surprising. However, it should be considered that the stoichiometry of C5b-8 and C9 within MAC is strictly regulated. One C5b-8 complex can bind only a limited number of C9 molecules (28), indicating that even when more C9 is available, it does not necessarily lead to more C5b-9 complex formation if the level of earlier complement components remains unaltered. Thus, owing to limited availability of C5b-8 complexes, the slightly enhanced levels of C9 p.M45L and p.A529T do not lead to enhanced pore formation and lysis (Figs 4 and 5). Nevertheless, the reason for decreased lytic activity of C9 p.F62S and p.P167S appears to be more complex and probably connected to altered polymerization propensity of these variants. Hemolytic assays performed with the recombinant C9 mutant proteins purified only on nickel column—containing both monomer and polymer proteins—displayed a significantly reduced lytic activity of the C9 p.F62S and p.P167S mutant proteins (data not shown), confirming our observations using serum samples (Fig. 5). We also observed during expression in vitro that these variants tend to aggregate. Hence, to exclude the functionally inactive polymers from our analysis, the recombinant C9 proteins were further purified by gel filtration (Supplementary Material, Fig. S1). Notably, applying the pure, monomer forms of C9, we could not detect any difference in lytic activity of the WT and mutant proteins (Fig. 4). These results suggest that the genetic alterations do not alter the classical, lytic function of C9 and that decreased lytic activity of C9 p.F62S and p.P167S is presumably caused by altered polymerization of these variants (Fig. 6). C9 has a tendency to polymerize, leading to formation of poly(C9) and a rapid loss of hemolytic activity (29). Although we did not observe the presence of poly(C9) in sera of the p.P167S carriers, probably owing to low serum concentration of C9, our in vitro data suggest that the C9 p.P167S mutant protein has an increased propensity to aggregate. Formation of C9 p.P167S polymerization was observed already during expression and purification of the recombinant protein. In contrast, the C9 p.F62S had a decreased capacity to polymerize in the presence of zinc ions (Fig. 6). This altered polymerization of C9 p.F62S and p.P167S in contrast to the WT and other mutant C9 proteins may be explained by localization of the variants in the mature protein (Fig. 1): the TSP1 and MACPF domains have been reported as main drivers of C9 polymerization during MAC formation (30). Thus the p.F62S variant, and in particular the p.P167S mutation owing to the substitution of a proline residue, may alter the structure of C9 in a way that results in altered polymerization propensity and hence, impaired pore formation and lysis. Interestingly, despite the almost complete lack of zinc-induced polymerization, the hemolytic activity of C9 p.F62S carriers’ sera was only slightly reduced (Fig. 5) and unaltered in assays in which the recombinant protein was used as a source of C9 (Fig. 4). However, it is important to emphasize that experiments carried out with the recombinant proteins measure only the functional effect of the mutations without the influence of changes in the protein concentration in vitro (Fig. 2). This may also explain the different results of assays using either sera (Fig. 5) or the purified proteins (Fig. 4): the increased concentration of C9 p.F62S and p.P167S in the serum may facilitate the formation of polymers, rendering the protein inactive and causing decreased lysis (Fig. 5). The latter effect is not present when recombinant C9 variants are applied at the same concentration. Although decreased lytic activity of the p.F62S variant could be observed only in assays measuring lysis of sheep erythrocytes but not in lactate dehydrogenase (LDH) release of ARPE-19 cells, the differences between sensitivity of erythrocytes’ and ARPE-19 cells’ membranes to lysis together with the biological variability of serum samples can partially explain the different results of the two assays. Nevertheless, both the p.F62S and p.P167S variants showed tendency to decreased lytic activity and probably the lack of statistically significant difference between control samples and the C9 p.F62S carriers in analysis of ARPE-19 cells is owing to the fact that in the latter case, experiments were carried out with a lower number of patients sample. Taken together, we have shown that several identified C9 variants, including the p.P167S variant highly associated with AMD, affect the serum level and polymerization of C9 without influencing its classical lytic activity. These results, alongside with normal sTCC level of the carriers’ plasma and the lack of correlation between C9 concentration and lytic activity of patients’ sera suggest that the variants influence disease pathology probably locally in the eye and not by increased lytic activity. This may happen mainly via enhanced formation of drusen owing to increased aggregation of certain mutant C9 proteins, but differences in sub-lytic MAC deposition and hence, C9-dependent changes in the inflammatory milieu of the retina cannot be ruled out. Further, one should consider that although ARPE-19 cells are regarded as reliable and widely used alternatives to native RPE, they are not optimal model of AMD (31). Hence, future experiments—using primary RPE cells, induced pluripotent stem cells from the carriers and complex retinal tissues—need to be performed to understand the precise implication of C9 in AMD pathology. Materials and Methods Genetic analysis and patient selection We performed whole-exome sequencing for 793 unrelated individuals (662 cases and 131 controls), as described in detail previously (17). Filtering steps were implemented to uncover the coding non-synonymous variants of C9 (NM_001737.3). Frequency filters, using public database ExAC (32) were implemented to ensure selection of rare variants (minor allele frequency <1%). We identified seven unique rare variants in C9, which were screened in an additional cohort consisting of 1234 unrelated case and 1368 unrelated control individuals using custom-made competitive allele-specific PCR assays according to the manufacturers’ recommendations (KASP SNP Genotyping System, LGC). Furthermore, family members of eight different families (n = 24) of which the proband carried a C9 variant were screened using Sanger sequencing (Supplementary Material, Fig. S3). All individuals included in the study underwent clinical evaluation and were graded for ‘AMD’ or ‘no AMD’ according to the Cologne Image Reading Center protocol. Control individuals without AMD were 60 years or older (33). Collection of serum and plasma samples To analyze the functional effect of rare genetic variants in C9, 128 serum and 95 plasma samples of 132 carriers were collected. Furthermore, we obtained serum (n = 156) and plasma (n = 155) of 157 individuals that did not carry a rare genetic variant in C9 (Supplementary Material, Table S2). Samples were obtained by a standard coagulation/centrifugation protocol and frozen at −80°C within 1 h after collection. Genomic DNA was isolated from peripheral blood samples according to standard procedures. This study was approved by local ethics committees on Research Involving Human Subjects and met the criteria of the Declaration of Helsinki. Cell lines Freestyle HEK293F cells (Invitrogen) were cultured in Freestyle Expression medium (Invitrogen) according to the manufacturer’s instructions. Cells were passaged every third day and transfected at Passage 10. HEK293 cells (ATCC) were grown in DMEM medium supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Gibco). The retinal pigment epithelial cell line ARPE-19 (ATCC) was cultured in DME/F12 medium (Hyclone) supplemented with 10% heat-inactivated FBS. Cells used for functional assays were between Passages 10 and 20. The cells were mycoplasma free and tested regularly with the VenorGEM Classic kit (Minerva Biolabs). ELISA for measurement of C9 concentration C9 concentrations in serum samples or supernatants of HEK293 cells secreting recombinant C9 proteins were determined by ELISA. Maxisorp microtiter plates (96-well, ThermoScientific) were coated with 1 μg/ml mouse anti-human C9 (Hycult Biotech, #HM2111) in 50 mM sodium carbonate (pH 9.6) overnight, at 4°C. Between each of the following steps, the plates were washed four times with Immunowash [50 mM Tris–HCl (pH 8.0), 150 mM NaCl, 0.1% Tween 20]. After coating, plates were blocked in Quench solution [Immunowash supplemented with 3% fish gelatin (Norland Products) for 1 h at room temperature (RT) and incubated with serum samples (diluted 1:200 in Quench) or with supernatants of HEK293 cells for 1 h at RT. As standard, serum-purified C9 (prepared in our lab) was applied. After incubation, bound proteins were detected using goat anti-human C9 (Complement Technologies, #A226, 1:4000 in serum or 1:2000 in HEK293 supernatant C9 ELISA in Quench,) followed by HRP-conjugated rabbit anti-goat Ig (1:2000 in Quench, DAKO). As substrate, 1,2-phenylenediamine dihydrochloride (OPD, Kem-En-Tec) was used and absorbance at 490 nm was measured using a Cary50 MPR microplate reader (Varian). ELISA for measurement of sTCC concentration Maxisorp immunoplates (96-well) were coated with 0.5 μg/ml mouse anti-human TCC (#aE11, Hycult) in PBS (pH7) (GE Healthcare) at RT, overnight. Between each of the following steps, plates were washed four times with PBST (PBS + 0.2% Tween 20). After coating, the plates were incubated with plasma samples diluted 5× in AG buffer (PBS + 0.02% NaN3 + 0.2% Tween 20 + 0.02 M Na2EDTA) for 1 h at RT. As standard, zymosan activated serum (ZAS-93, 1000 AU/ml, Hycult) was used. Plasma sTCC was detected using biotinylated mouse anti-human C6 (Quidel, #A219, 1:2000 in PBST) biotinylated using EZ-Link Sulfo-NHS-Biotin from Pierce, followed by HRP-conjugated streptavidin (1:1000 in PBST, R&D Sytems). As chromogen, OPD was used. C9 cDNA clones for recombinant proteins To determine if the genetic variants affect the function of the protein, both the WT and mutant C9 proteins were expressed in vitro. To this end, full-length cDNA encoding human C9 with an N-terminal His-tag was purchased from Invitrogen in pMA-T vector. The identified variants in C9, i.e. p.M45L, p.F62S, p.R118W, p.G126R, p.P167S, p.T170I and p.A529T were introduced using the QuikChange site-directed mutagenesis kit (Agilent Technologies), according to the manufacturer’s instructions. The primers used are listed in Supplementary Material, Table S3. The variants were confirmed by automated Sanger DNA sequencing (GATC Biotec). WT and mutant C9 cDNA sequences were then subcloned into the eukaryotic pCEP4 expression vector (Invitrogen), suitable for transfection of HEK293F cells. Expression and purification of recombinant C9 proteins HEK293F cells were transiently transfected with the WT or mutant C9-pCEP4 constructs using FreeStyle Max Reagent according to the manufacturer’s instructions (Invitrogen). Secreted supernatants were collected after 2, 4, 6, 8 days, pooled and stored at −20°C. Recombinant WT and mutant C9 proteins were isolated from the collected media using a column of Ni-NTA Superflow resin (Qiagen) equilibrated with 50 mM NaH2PO4, 300 mM NaCl, 10 mM imidazole, pH 8 (binding buffer). After extensive washing with binding buffer, bound proteins were eluted with 50 mM NaH2PO4, 300 mM NaCl, 500 mM imidazole, pH 8 (elution buffer). The presence of C9 in the eluted fractions was verified by 7.5% SDS-PAGE and silver staining. To separate monomer and polymer C9 proteins by gel filtration, C9-containing fractions of Ni2+-affinity chromatography were pooled, concentrated using an Amicon Filter unit with 50 kDa cut-off (Millipore) and loaded onto Sephacryl-100 column (GE healthcare) to separate monomer and polymer C9 proteins. All preparative works were done at 4°C. The purity of monomer WT and mutant C9 proteins was analyzed by SDS-PAGE and western blot. SDS-PAGE and western blot The purified recombinant C9 proteins were separated by gel electrophoresis under reducing (25 mM DTT) conditions and transferred to a PVDF membrane using semi-dry blotting apparatus (BioRad). The membranes were blocked with Quench solution and C9 was visualized using a polyclonal, goat anti-human C9 antibody from Complement Technologies (#A226, 1:20 000 dilution in Quench) followed by incubation with HRP-conjugated rabbit anti-goat Ig (diluted 1:10 000 in Quench). Bound antibodies were visualized by the enhanced chemiluminescence method (Millipore) and analyzed with the ImageLab software (BioRad). Transient transfection of HEK293 cells Transient transfection of HEK293 cells with empty vector, WT or mutant C9-pCEP4 constructs was accomplished using Lipofectamine 2000 reagent (Invitrogen), according to the manufacturer’s instructions. Briefly, cells were seeded on 12-well plates before transfection and grew until 90% confluency. On the day of transfection, DMEM + 10% FBS was changed to OptiMem (Gibco) and cells were transfected with 6 µg plasmid complexed with 6 µl Lipofectamine 2000 reagents. Eight hours after transfection, cells were washed into DMEM + 10% FBS and further cultured for 2 days in DMEM + 10% FBS and 3 days in OptiMem medium. After collecting the supernatant for C9 ELISA, cells were lysed in RIPA buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% NP40, 0.5% deoxycholate) containing protease inhibitors (Roche) for 60 min on ice. After centrifugation at 15 000g for 15 min at 4°C, 60 µg of each cell lysate was run on non-reducing 10% SDS-PAGE. C9 was visualized by western blot as described for purified C9 mutant proteins, except that goat anti-human C9 antibody was used at 1:5000 dilution. Lysis of erythrocytes by recombinant WT and mutant C9 proteins Sheep erythrocytes (Håtunalab) were washed with DGVB++ (2.5 mM veronal buffer, pH 7.3, 72 mM NaCl, 140 mM glucose, 0.1% gelatin, 1 mM MgCl2 and 0.15 mM CaCl2), pelleted and incubated with 1 ml of DGVB++ containing amboceptor (Behring) diluted 1:1000 for 20 min at 37°C. Sensitized erythrocytes were washed three times, pelleted and resuspended in DGVB++ to obtain a cell suspension of which 10 µl lysed by 90 µl of water gives 1.2–1.4 absorbance at 405 nm. The suspension (10 μl) was incubated with 50 µl of 2% C9-depleted serum (Complement Technologies) reconstituted with 50 ng/ml WT or mutant C9 proteins (preliminary titrated concentration of recombinant WT C9 causing lysis of 50% of erythrocytes). After incubation for 30 min at 37°C, cells were overlayed with another 50 µl of DGVB++ and centrifuged. Released hemoglobin levels of the collected supernatants (80 µl) were measured at 405 nm using Cary50 MPR microplate reader. C9-induced lysis was calculated by dividing (Abs 405 nm of C9 reconstituted samples—Abs 405 nm of C9 depl. serum treated samples alone) with Abs 405 nm of water-treated samples and illustrated as % of maximal (i.e. water induced) lysis. Lysis of erythrocytes by sera of C9 carriers and non-carrier controls To analyze C9-dependent lysis of erythrocytes by sera of carriers and non-carriers, erythrocytes were prepared and sensitized as described above. To induce deposition of C5b-8 complexes, 10 μl of the erythrocyte suspension was incubated with 50 µl of 2% C9-depleted serum (Complement Technologies) diluted in DGVB++ at 37°C for 30 min, shaking at 650 rpm. After incubation, cells were washed three times with 40 mM EDTA-GVB buffer (2.5 mM veronal buffer, pH 7.3, 72 mM NaCl, 140 mM glucose and 40 mM EDTA) and MAC assembly was induced by serum samples, diluted 1:2000 in 40 mM EDTA-GVB (the buffer blocks complement activation and thus prevents formation of novel C5b-8/9 complexes from the applied sera). After incubation for 30 min at 37°C, cells were overlayed with another 50 µl of 40 mM EDTA-GVB and centrifuged. Released hemoglobin level of the collected supernatants (80 µl) was measured at 405 nm using a Cary50 MPR microplate reader. Treatment of ARPE-19 cells using recombinant WT and mutant C9 proteins ARPE-19 cells were plated at 106 cells/ml concentration on 96-well plates (Nunc) in DME/F12 medium supplemented with 10% FBS. After 1 day, cells were washed and medium was changed to DME/F12 without FBS. The next day, cells were centrifuged and treated with 3 µg/well function blocking anti-CD59 antibody (IBGRL) for 45 min at 4°C to enhance MAC formation. After incubation, cells were treated with 10% C9-depleted serum reconstituted with 10 µg/ml of purified WT or mutant C9 proteins in DGVB++ for 2 h at 37°C, shaking at 150 rpm. After 2 h, cells were centrifuged at 1500g for 3 min, and supernatants were collected to measure LDH release using the Pierce LDH cytotoxicity detection kit (ThermoScientific). Treatment of ARPE-19 cells by sera of C9 carriers and non-carrier controls ARPE-19 cells were plated and primed for MAC deposition as described above. To induce C5b-8 deposition, anti-CD59 sensitized cell were treated with 20% C9-depleted serum in DGVB++ for 45 min at 37°C, shaking at 150 rpm. After incubation, cells were washed three times in EDTA-GVB and treated with serum samples, diluted 2.5 times in 40 mM EDTA-GVB to supply C9 in the pre-formed C5b-8 complex and induce lysis. After 2 h, supernatants were collected to measure LDH release and cells were analyzed for MAC deposition by flow cytometry. LDH assay LDH release was measured in 80 μl volume using the Pierce LDH cytotoxicity detection kit (ThermoScientific) according to the manufacturer’s instruction. The following formula was used to determine percent of maximum LDH release: LDH activity of C9 depleted serum alone treated samples was substracted from LDH activity of C9 reconstituted samples and divided by the total LDH activity (maximum LDH release induced by 1× lysis buffer of Pierce LDH cytotoxicity kit) and multiplied by 100. Flow cytometry To detect MAC deposition and viability of ARPE-19 cells after induction of MAC assembly, cells were washed two times in PBS and incubated with a neoantigen-specific rabbit anti-human C5b-9 antibody (Complement Technologies, 1:400) for 30 min at 4°C. The cells were washed twice in binding buffer (10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl2) and incubated with AlexaFluor647-conjugated goat anti-rabbit Ab (ThermoScientific, 1:2000), Annexin V-FITC (Immunotools, 2 µl/sample) and ZombieAqua fixable viability dye (BioLegend, 1:2000) for 30 min at RT in the dark. The stained cells were washed two times and analyzed by CytoFlex flow cytometer (Beckman Coulter). Data were analyzed with FlowJo software (Tree Star) and expressed as geometric mean fluorescence intensity (gMFI) of C5b-9 signal or percent of ZombieAqua and Annexin V double positive, late apoptotic cells. C9 polymerization Purified C9 proteins (5 μM) were allowed to polymerize in TBS in the presence or absence of 15 μM Zn2+ ions at 37°C for 4 h, shaking at 350 rpm. After incubation, samples were separated by 5–10% gradient gel electrophoresis and visualized by silver staining. Statistical analysis Results investigating the relationship between serum C9 concentration, plasma sTCC level and lytic activity of serum samples and the carrier status were analyzed using the Kruskal–Wallis test with Dunn’s multiple comparison. Data comparing expression, secretion and functional activity of recombinant WT and mutant C9 proteins were analyzed by one-way ANOVA with Dunnett’s multiple comparison. C9 polymerization was compared using two-way ANOVA with Dunnett’s multiple comparison test. Data were analyzed, and graphs prepared using Prism Software version 7. Supplementary Material Supplementary Material is available at HMG online. Conflict of Interest statement. None declared. Funding This work was supported by the Wenner-Gren Foundation, The Royal Physiographic Society of Lund (37352) and the Lars Hierta's Memorial Foundation (FO2016-0079), awarded to M.K. The research leading to these results has received funding from the European Research Council under the European Union's Seventh Framework Programme (FP/2007-2013)/ERC Grant Agreement n. 310644 (MACULA), the Foundation Fighting Blindness USA (grant C-GE-0811-0548-RAD04), the Swedish Research Council (2016-01142) and a grant for clinical research (ALF). References 1 Chakravarthy U. , Evans J. , Rosenfeld P.J. ( 2010 ) Age related macular degeneration . BMJ , 340 , c981. Google Scholar CrossRef Search ADS PubMed 2 Sobrin L. , Seddon J.M. ( 2014 ) Nature and nurture—genes and environment—predict onset and progression of macular degeneration . Prog. Retin. Eye Res ., 40 , 1 – 15 . Google Scholar CrossRef Search ADS PubMed 3 Seddon J.M. , Silver R.E. , Kwong M. , Rosner B. ( 2015 ) Risk prediction for progression of macular degeneration: 10 common and rare genetic variants, demographic, environmental, and macular covariates . Invest. Ophthalmol. Vis. Sci ., 56 , 2192 – 2202 . Google Scholar CrossRef Search ADS PubMed 4 Fritsche L.G. , Igl W. , Bailey J.N. , Grassmann F. , Sengupta S. , Bragg-Gresham J.L. , Burdon K.P. , Hebbring S.J. , Wen C. , Gorski M. et al. ( 2016 ) A large genome-wide association study of age-related macular degeneration highlights contributions of rare and common variants . Nat. Genet ., 48 , 134 – 143 . Google Scholar CrossRef Search ADS PubMed 5 Ricklin D. , Hajishengallis G. , Yang K. , Lambris J.D. ( 2010 ) Complement: a key system for immune surveillance and homeostasis . Nat. Immunol ., 11 , 785 – 797 . Google Scholar CrossRef Search ADS PubMed 6 Sonnen A.F. , Henneke P. ( 2014 ) Structural biology of the membrane attack complex . Subcell. Biochem ., 80 , 83 – 116 . Google Scholar CrossRef Search ADS PubMed 7 Kilgore K.S. , Schmid E. , Shanley T.P. , Flory C.M. , Maheswari V. , Tramontini N.L. , Cohen H. , Ward P.A. , Friedl H.P. , Warren J.S. ( 1997 ) Sublytic concentrations of the membrane attack complex of complement induce endothelial interleukin-8 and monocyte chemoattractant protein-1 through nuclear factor-kappa B activation . Am. J. Pathol ., 150 , 2019 – 2031 . Google Scholar PubMed 8 Niculescu F. , Badea T. , Rus H. ( 1999 ) Sublytic C5b-9 induces proliferation of human aortic smooth muscle cells: role of mitogen activated protein kinase and phosphatidylinositol 3-kinase . Atherosclerosis , 142 , 47 – 56 . Google Scholar CrossRef Search ADS PubMed 9 Triantafilou K. , Hughes T.R. , Triantafilou M. , Morgan B.P. ( 2013 ) The complement membrane attack complex triggers intracellular Ca2+ fluxes leading to NLRP3 inflammasome activation . J. Cell Sci ., 126 , 2903 – 2913 . Google Scholar CrossRef Search ADS PubMed 10 Vlaicu S.I. , Tegla C.A. , Cudrici C.D. , Danoff J. , Madani H. , Sugarman A. , Niculescu F. , Mircea P.A. , Rus V. , Rus H. ( 2013 ) Role of C5b-9 complement complex and response gene to complement-32 (RGC-32) in cancer . Immunol. Res ., 56 , 109 – 121 . Google Scholar CrossRef Search ADS PubMed 11 Mullins R.F. , Dewald A.D. , Streb L.M. , Wang K. , Kuehn M.H. , Stone E.M. ( 2011 ) Elevated membrane attack complex in human choroid with high risk complement factor H genotypes . Exp. Eye. Res ., 93 , 565 – 567 . Google Scholar CrossRef Search ADS PubMed 12 Mullins R.F. , Schoo D.P. , Sohn E.H. , Flamme-Wiese M.J. , Workamelahu G. , Johnston R.M. , Wang K. , Tucker B.A. , Stone E.M. ( 2014 ) The membrane attack complex in aging human choriocapillaris: relationship to macular degeneration and choroidal thinning . Am. J. Pathol ., 184 , 3142 – 3153 . Google Scholar CrossRef Search ADS PubMed 13 Lueck K. , Wasmuth S. , Williams J. , Hughes T.R. , Morgan B.P. , Lommatzsch A. , Greenwood J. , Moss S.E. , Pauleikhoff D. ( 2011 ) Sub-lytic C5b-9 induces functional changes in retinal pigment epithelial cells consistent with age-related macular degeneration . Eye (London) , 25 , 1074 – 1082 . Google Scholar CrossRef Search ADS 14 Kunchithapautham K. , Rohrer B. ( 2011 ) Sublytic membrane-attack-complex (MAC) activation alters regulated rather than constitutive vascular endothelial growth factor (VEGF) secretion in retinal pigment epithelium monolayers . J. Biol. Chem ., 286 , 23717 – 23724 . Google Scholar CrossRef Search ADS PubMed 15 Seddon J.M. , Yu Y. , Miller E.C. , Reynolds R. , Tan P.L. , Gowrisankar S. , Goldstein J.I. , Triebwasser M. , Anderson H.E. , Zerbib J. et al. ( 2013 ) Rare variants in CFI, C3 and C9 are associated with high risk of advanced age-related macular degeneration . Nat. Genet ., 45 , 1366 – 1370 . Google Scholar CrossRef Search ADS PubMed 16 Nishiguchi K.M. , Yasuma T.R. , Tomida D. , Nakamura M. , Ishikawa K. , Kikuchi M. , Ohmi Y. , Niwa T. , Hamajima N. , Furukawa K. et al. ( 2012 ) C9-R95X polymorphism in patients with neovascular age-related macular degeneration . Invest. Ophthalmol. Vis. Sci ., 53 , 508 – 512 . Google Scholar CrossRef Search ADS PubMed 17 Geerlings M.J. , Kremlitzka M. , Bakker B. , Nilsson S.C. , Saksens N.T. , Lechanteur Y.T. , Pauper M. , Corominas J. , Fauser S. , Hoyng C.B. et al. ( 2017 ) The functional effect of rare variants in complement genes on C3b degradation in patients with age-related macular degeneration . JAMA Ophthalmol ., 135 , 39 – 46 . Google Scholar CrossRef Search ADS PubMed 18 Yu Y. , Wagner E.K. , Souied E.H. , Seitsonen S. , Immonen I.J. , Happola P. , Raychaudhuri S. , Daly M.J. , Seddon J.M. ( 2016 ) Protective coding variants in CFH and PELI3 and a variant near CTRB1 are associated with age-related macular degenerationdagger . Hum. Mol. Genet ., 25 , 5276 – 5285 . Google Scholar PubMed 19 Saksens N.T. , Geerlings M.J. , Bakker B. , Schick T. , Daha M.R. , Fauser S. , Boon C.J. , de Jong E.K. , Hoyng C.B. , den Hollander A.I. ( 2016 ) Rare genetic variants associated with development of age-related macular degeneration . JAMA Ophthalmol ., 134 , 287 – 293 . Google Scholar CrossRef Search ADS PubMed 20 Saksens N.T. , Lechanteur Y.T. , Verbakel S.K. , Groenewoud J.M. , Daha M.R. , Schick T. , Fauser S. , Boon C.J. , Hoyng C.B. , den Hollander A.I. ( 2016 ) Analysis of risk alleles and complement activation levels in familial and non-familial age-related macular degeneration . PLoS One , 11 , e0144367. Google Scholar CrossRef Search ADS PubMed 21 Smailhodzic D. , van Asten F. , Blom A.M. , Mohlin F.C. , den Hollander A.I. , van de Ven J.P. , van Huet R.A. , Groenewoud J.M. , Tian Y. , Berendschot T.T. et al. ( 2014 ) Zinc supplementation inhibits complement activation in age-related macular degeneration . PLoS One , 9 , e112682. Google Scholar CrossRef Search ADS PubMed 22 Horiuchi T. , Nishizaka H. , Kojima T. , Sawabe T. , Niho Y. , Schneider P.M. , Inaba S. , Sakai K. , Hayashi K. , Hashimura C. et al. ( 1998 ) A non-sense mutation at Arg95 is predominant in complement 9 deficiency in Japanese . J. Immunol ., 160 , 1509 – 1513 . Google Scholar PubMed 23 Reynolds R. , Hartnett M.E. , Atkinson J.P. , Giclas P.C. , Rosner B. , Seddon J.M. ( 2009 ) Plasma complement components and activation fragments: associations with age-related macular degeneration genotypes and phenotypes . Invest. Ophthalmol. Vis. Sci ., 50 , 5818 – 5827 . Google Scholar CrossRef Search ADS PubMed 24 Smailhodzic D. , Muether P.S. , Chen J. , Kwestro A. , Zhang A.Y. , Omar A. , Van de Ven J.P. , Keunen J.E. , Kirchhof B. , Hoyng C.B. et al. ( 2012 ) Cumulative effect of risk alleles in CFH, ARMS2, and VEGFA on the response to ranibizumab treatment in age-related macular degeneration . Ophthalmology , 119 , 2304 – 2311 . Google Scholar CrossRef Search ADS PubMed 25 Scholl H.P. , Charbel Issa P. , Walier M. , Janzer S. , Pollok-Kopp B. , Borncke F. , Fritsche L.G. , Chong N.V. , Fimmers R. , Wienker T. et al. ( 2008 ) Systemic complement activation in age-related macular degeneration . PLoS One , 3 , e2593. Google Scholar CrossRef Search ADS PubMed 26 Ristau T. , Paun C. , Ersoy L. , Hahn M. , Lechanteur Y. , Hoyng C. , de Jong E.K. , Daha M.R. , Kirchhof B. , den Hollander A.I. et al. ( 2014 ) Impact of the common genetic associations of age-related macular degeneration upon systemic complement component C3d levels . PLoS One , 9 , e93459. Google Scholar CrossRef Search ADS PubMed 27 Schick T. , Steinhauer M. , Aslanidis A. , Altay L. , Karlstetter M. , Langmann T. , Kirschfink M. , Fauser S. ( 2017 ) Local complement activation in aqueous humor in patients with age-related macular degeneration . Eye (London) , 31 , 810 – 813 . Google Scholar CrossRef Search ADS 28 Tegla C.A. , Cudrici C. , Patel S. , Trippe R. III , Rus V. , Niculescu F. , Rus H. ( 2011 ) Membrane attack by complement: the assembly and biology of terminal complement complexes . Immunol. Res ., 51 , 45 – 60 . Google Scholar CrossRef Search ADS PubMed 29 Podack E.R. , Tschopp J. ( 1982 ) Polymerization of the ninth component of complement (C9): formation of poly(C9) with a tubular ultrastructure resembling the membrane attack complex of complement . Proc. Natl. Acad. Sci. USA ., 79 , 574 – 578 . Google Scholar CrossRef Search ADS 30 Dudkina N.V. , Spicer B.A. , Reboul C.F. , Conroy P.J. , Lukoyanova N. , Elmlund H. , Law R.H. , Ekkel S.M. , Kondos S.C. , Goode R.J. et al. ( 2016 ) Structure of the poly-C9 component of the complement membrane attack complex . Nat. Commun ., 7 , 10588. Google Scholar CrossRef Search ADS PubMed 31 Samuel W. , Jaworski C. , Postnikova O.A. , Kutty R.K. , Duncan T. , Tan L.X. , Poliakov E. , Lakkaraju A. , Redmond T.M. ( 2017 ) Appropriately differentiated ARPE-19 cells regain phenotype and gene expression profiles similar to those of native RPE cells . Mol. Vis ., 23 , 60 – 89 . Google Scholar PubMed 32 Lek M. , Karczewski K.J. , Minikel E.V. , Samocha K.E. , Banks E. , Fennell T. , O’Donnell-Luria A.H. , Ware J.S. , Hill A.J. , Cummings B.B. et al. ( 2016 ) Analysis of protein-coding genetic variation in 60,706 humans . Nature , 536 , 285 – 291 . Google Scholar CrossRef Search ADS PubMed 33 Ristau T. , Ersoy L. , Hahn M. , den Hollander A.I. , Kirchhof B. , Liakopoulos S. , Fauser S. ( 2014 ) Nongenetic risk factors for neovascular age-related macular degeneration . Invest. Ophthalmol. Vis. Sci ., 55 , 5228 – 5232 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Human Molecular GeneticsOxford University Press

Published: May 14, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off