An Evaluation of the Complement-Regulating Activities of Human Complement Factor H (FH) Variants Associated With Age-Related Macular Degeneration

Purpose Factor H (FH, encoded by CFH) prevents activation of the complement system's alternative pathway (AP) on host tissues. FH impedes C3 convertase (C3bBb) formation, accelerates C3bBb decay, and is a cofactor for factor I (FI)–catalyzed C3b cleavage. Numerous CFH variants are associated with age-related macular degeneration (AMD), but their functional consequences frequently remain undetermined. Here, we conduct functional comparisons between a control version of FH (not AMD linked) and 21 AMD-linked FH variants. Methods Recombinantly produced, untagged, full-length FH versions were assayed for binding to C3b and decay acceleration of C3bBb using surface-plasmon resonance, FI-cofactor activity using a fluorescent probe of C3b integrity, suppression of C5b-9 assembly on an AP-activating surface, and inhibition of human AP-mediated lysis of sheep erythrocytes. Results All versions were successfully purified despite below-average yields for Arg2Thr, Arg53Cys, Arg175Pro, Arg175Gln, Ile221Val, Tyr402His, Pro503Ala, Arg567Gly, Gly1194Asp, and Arg1210Cys. Compared to control FH, Arg2Thr, Leu3Val, Ser58Ala, Asp90Gly, Asp130Asn, Gln400Lys, Tyr402His, Gly650Val, Ser890Ile, and Thr965Met showed minimal functional differences. Arg1210C, Arg53His, Arg175Gln, Gly1194Asp, Pro503Ala, Arg53Cys, Arg576Gly, and Arg175Pro (in order of decreasing efficacy) underperformed, while Ile221Val, Arg303Gln, and Arg303Trp were “marginal.” We newly identified variants toward the center of the molecule, Pro503Ala and Arg567Gly, as potentially pathogenic. Conclusions Our approach could be extended to other variants of uncertain significance and to assays for noncanonical FH activities, aiming to facilitate selection of cohorts most likely to benefit from therapeutic FH. This is timely as recombinant therapeutic FH is in development for intravitreal treatment of AMD in patients with reduced FH functionality.


Supplementary Methods I: Expression and purification of recombinant FH variants
Samples of the following prioritized 21 FH variant proteins associated with enhanced AMD risk, and a control FH protein (ID P08603 in UniProtKB), were prepared: Tyr402His Arg53His Pro503Ala Ser58Ala Arg567Gly Asp90Gly Gly650Val Asp130Asn Ser890Ile Arg175Gln Thr956Met Arg175Pro Gly1194Asp Ile221Val Arg1210Cys Arg303Trp * Amino-acid substitution in the N-terminal signal sequence, which is cleaved from mature, secreted protein All proteins were produced (without affinity tags) by transient expression using a standard method in human cell culture (HEK293 cell lines), 1 secreted (with cleavage of N-terminal signal sequence, residues 1-18), and purified to homogeneity ( Supplementary Fig. 1A) from conditioned media using a proprietary affinitypurification method. Control FH and all FH variants were produced and purified in an identical manner; yields of purified proteins were determined by measuring absorbance of light at 280 nm in a Nanodrop (Thermo Fisher Scientific, Waltham, MA, USA) and applying calculated extinction coefficients.
Supplementary Methods II: C3b binding by surface plasmon resonance (SPR) For each experiment, 0.25 µM variant FH (or control FH) in 10 mM 4-(2hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer, 150 nM NaCl, 0.05% (v/v) surfactant P20, was injected in duplicate over amine-coupled C3b (300 RUs on a C1 chip) for 150 seconds at 50 μL/minute followed by a dissociation time of 600 seconds. The sensor chip surface was regenerated between individual sample injections by three 30-second 50 µL/minute injections of 1 M NaCl, with a surface stabilization period of 150 seconds following the third NaCl injection. Injections of control FHfor pairwise comparisons with each variant FH -were replicated at frequent, regular intervals to help take account of any drift in chip or instrument performance over the duration of the experiment. The baseline drift was corrected by subtracting the signal obtained from an injection of 0 µM FH. Data were analyzed using the Biacore Evaluation software and a 1:1 steady-state binding model.

Supplementary Methods III: Decay accelerating activity (DAA) assay
The SPR experiments were performed using 10 mM HEPES, 150 mM NaCl, 0.05% (v/v) surfactant P20, 1 mM MgCl2 (pH 7.4) (GE Healthcare). C3bBb was assembled on the CM5 chip (GE Healthcare) bearing 2845 RUs of immobilized (amine-coupled) C3b molecules by performing a 220-second 10 μL/minute injection of a solution containing 500 nM FB (Complement Technology, Inc.) and 50 nM FD (Complement Technology, Inc.). The decay of C3bBb was subsequently monitored over an initial 230-second dissociation phase in the absence of any FH, allowing observation of intrinsic convertase decay (i.e., a decline in RUs attributable to loss of Bb). Subsequently, a solution of 20 nM FH (or the control) was injected at 10 µL/minute for 550 seconds, allowing observation of accelerated dissociation. The surface was regenerated between measurements by a 30-second 10 μL/minute injection of 0.1 μM plasma-purified FH followed by three, 30-second, 10 μL/minute injections of 1 M NaCl, and then a surface-stabilization period of 120 seconds. The SPR response arising from the binding of variant or control FH to immobilized C3b, in the absence of the convertase, was subtracted from the corresponding convertase decay response. Data were processed using the Biacore Evaluation software and then normalized to compensate for the small drift in signal (assumed to arise from the gradual leaching of C3b from the surface over multiple measurements). Normalization of data was achieved by comparing responses at 220 seconds (i.e., at the time when the injection of FB and FD ceased) and adjusting these to be 1.00. The normalized responses were then plotted against time, and the resultant plots overlaid at the time point of variant/control FH injection.

Supplementary Methods IV: Fluorescent cofactor assay
Stock solutions of variant FH and control FH were serially diluted in Tris-buffered saline (1X TBS; 50 mM Tris, 150 mM NaCl, pH 7.4) to achieve concentrations ranging from 0.025 to 0.5 mg/mL. Subsequently, assay components were added to opaque half-area black polystyrene plates in the following order in a 50 µL final reaction volume: 0.02 mg purified human C3b (Complement Technology, Inc.), 5 µM ANS (final concentration) (Acros Organics B.V.B.A., Geel, Belgium), control/variant FH to give final concentrations between 5 and 100 µg/mL, and 0.1 µg of FI (Complement Technology, Inc.). Reactions were mixed briefly by shaking at 4000 revolutions per minute and monitored over 30 minutes at 30-second intervals at 30 °C. Fluorescence readings were recorded in kinetic mode with excitation set to 386 nm and emission set to 472 nm.
Supplementary Methods V: Hemolysis assay Human FH-depleted serum (20 µL) (Complement technology, Inc.) was preincubated at room temperature for ten minutes with either variant FH or control FH at concentrations ranging from 25 to 600 µg/mL using gelatin-containing veronal buffer (GVB 0 Complement Technology, Inc.) with 0.1 mM EDTA [GVB 0.1 mM E ], as the diluent. Sheep erythrocytes (Complement Technology, Inc.) were washed once in GVB 0 and re-suspended to 2.1x10 8 cells/mL in GVB 0.1 mM E . A total of 2.1x10 7 sheep erythrocytes were added to each well of a 96-well plate followed by Mg·EGTA at a final concentration of 10 mM in a total reaction volume of 200 µL. Wells containing reaction mixes in heat-inactivated FH-depleted serum and 1% v/v triton-X (BioVision, Milpitas, CA, United States) in water were included as negative and positive controls for lysis, respectively. Samples were incubated at 37 °C for one hour with shaking, followed by the addition of 150 µL of GVB 10 mM E to stop further cell lysis. The samples were centrifuged at 1000 x g for five minutes at 7 °C. A volume of 150 µL of pre-cleared supernatant was transferred to a clear-bottom 96-well plate. The extent of hemolysis (hemoglobin released) was then measured, in triplicate samples for each FH concentration, as absorbance at 412 nm, corrected for background absorbance measured at 690 nm.

Supplementary Tables
Supplementary Table 1. FH variants data summary.  Figure 1 (main text), are represented by ovals numbered 1-20. The modular location of the amino acid-residue substitution in each variant is indicated, along with a pie-chart style representation of its functional impact as reflected in its CES (the larger the red segment, the greater the impact).
Abbreviations: CCPs, complement control protein modules; CES, combined efficacy score; FH, complement factor H Supplementary Figure 9. Stacked plots showing the relative contributions to CES, plotted in order of ascending CES. The upper panel includes values from each of the five functional measurements and the protein yield. For the lower panel, protein yields were omitted from the CES calculations. Variants that have moved significantly within the ranking as a result are indicated by red dotted lines. Thus, the "defective" variant R53H moves still further down the ranking when considering only the functional assay data. R1210C and R303Q swap positions at the boundary between defective and marginally impacted categories. Also of some note is R2T, which Abbreviations: AP, alternative pathway; C3b, complement component 3b; CES, combined efficacy score; DAA, decay accelerating activity