The classical and lectin pathways of the complement system are important for the elimination of pathogens and apoptotic cells and stimulation of the adaptive immune system. Upon activation of these pathways, complement component C4 is proteolytically cleaved, and the major product C4b is deposited on the activator, enabling assembly of a C3 convertase and downstream alternative pathway amplification. Although excessive activation of the lectin and classical pathways contributes to multiple autoimmune and inflammatory diseases and overexpression of a C4 isoform has recently been linked to schizophrenia, a C4 inhibitor and structural characterization of the convertase formed by C4b is lacking. In this study, we present the nanobody hC4Nb8 that binds with picomolar affinity to human C4b and potently inhibits in vitro complement C3 deposition through the classical and lectin pathways in human serum and in mouse serum. The crystal structure of the C4b:hC4Nb8 complex and a three-dimensional reconstruction of the C4bC2 proconvertase obtained by electron microscopy together rationalize how hC4Nb8 prevents proconvertase assembly through recognition of a neoepitope exposed in C4b and reveals a unique C2 conformation compared with the alternative pathway proconvertase. On human induced pluripotent stem cell–derived neurons, the nanobody prevents C3 deposition through the classical pathway. Furthermore, hC4Nb8 inhibits the classical pathway-mediated immune complex delivery to follicular dendritic cells in vivo. The hC4Nb8 represents a novel ultrahigh-affinity inhibitor of the classical and lectin pathways of the complement cascade under both in vitro and in vivo conditions.

This article is featured in In This Issue, p.1477

The complement system is part of the innate immune response and a critical component for our first line of defense against invading pathogens. Its origin dates back to the emergence of multicellular organisms (1). Complement is activated through three pathways: the classical pathway (CP), the lectin pathway (LP), and the alternative pathway (AP). The CP and the LP are initiated by a pattern recognition molecule (PRM) that binds to either pathogen-associated molecular patterns or to danger-associated molecular patterns (2). Pattern recognition triggers a proteolytic cascade in which a central event is the cleavage of the protein complement factor C3, leading to deposition of the fragment C3b on the activator surface, phagocytosis of the activator, and ultimately to lysis if the activator is a susceptible cell (3, 4).

The AP may be activated by spontaneous hydrolysis of an internal thioester in C3 (5), but the AP also strongly amplifies the initial deposition of C3b deposited on activators through the CP and LP (6). The LP is initiated by binding of one of five different PRMs to specific carbohydrate patterns on the activator (7). The proteases MASP-1 and MASP-2, which are associated with the LP PRMs, are activated upon clustering on the activator (8), and MASP-2 cleaves C4 into the small fragment C4a and the large fragment 190-kDa C4b (Fig. 1A). This cleavage initiates a conformational change, leading to exposure of a thioester group in the nascent C4b that may react with a nucleophile on the activator surface (9). The CP evolved from the LP, and through evolution, the PRM C1q acquired the capability of recognizing Ag-bound IgG and IgM, but a number of other C1q-binding and CP-activating structures have been reported (1, 10). Upon C1q activator binding, the associated C1r and C1s proteases are activated, and C1s, in turn, cleaves C4 to C4b, presumably in a manner very similar to that of MASP-2 in the LP (11).

Activator-bound C4b binds to the zymogen C2 in a Mg2+-dependent manner (12, 13). The C4b2 complex is the CP proconvertase, the precursor of the C3 convertase. The CP C3 convertase is generated through a second cleavage conducted by C1s or MASP-1/MASP-2 (Fig. 1A), releasing C2b and leaving the C3 convertase C4b2a attached to the surface (14). Cleavage of C3 by C4b2a initiates a conformational change in the nascent C3 similar to that in C4b, exposing a thioester that links C3b to the surface (15). On host surfaces, C3b is degraded by the protease factor I (FI) with the help of cofactors factor H (FH), CD46/MCP, and complement receptor (CR) 1 (CD35), resulting in the late opsonins iC3b and C3dg, both ligands for CR3 (16) (Fig. 1A). In lymph nodes, iC3b/C3dg enables uptake of immune complexes (ICs) through interaction with CR3 on the surface of subcapsular sinus macrophages, from which the ICs are passed on to naive B cells and subsequently to follicular dendritic cells (FDCs) through CR2 recognition of iC3b/C3dg (17). The resulting transfer of complement opsonized Ags allows for long-term Ag presentation on FDCs, leading to formation of B cell germinal centers, in which B cell clonal expansion, class switch recombination, and somatic hypermutation for the production of Abs takes place (18). C4b is degraded in a similar manner by FI together with cofactors C4b binding protein (C4BP) and CR1 to C4c and C4d. However, no effector function of these C4b degradation products have been identified, but C4b, together with C3b, functions as ligand for CR1 on RBCs and contributes to clearance of ICs (19). Whereas the function of the CP for pathogen clearance and tissue homeostasis has been investigated for decades (20), more recent studies have also identified an important role for this pathway during development in the brain (21, 22). Complement contributes to synaptic pruning in the CNS as a developmental mechanism for refinement of synaptic circuits, enabling full cognitive ability in adulthood (23, 24). The pruning process is supported by the interaction between iC3b deposited on synapses with CR3 on microglia, the CNS resident macrophages. Deficiencies of C1q, C3, and C4 have been demonstrated to affect synaptic pruning in mice, linking the CP directly to the process of neuronal connectivity refinement (25, 26).

A rapidly growing research topic concerns the molecular mechanisms underlying disease development as a consequence of complement dysregulation. Diseases in which CP dysregulation is known to play a significant role include systemic lupus erythematosus, hereditary angioedema, ischemia–reperfusion injury, sepsis, autoimmune hemolytic anemia, glomerulonephritis, Ab-mediated graft rejection, Alzheimer disease, schizophrenia (SCZ), and cold agglutinin disease (2731). Numerous complement inhibitors targeting the alternative and terminal pathways have been described. However, with respect to inhibitors specific for the CP and LP, many fewer molecules have been described, and there is currently no well-established C4-specific complement inhibitor (32).

Nanobodies are the Ag-binding domain of H chain–only Abs present in camelids (33). They offer a versatile platform to target virtually any Ag, as they can be selected in vitro by phage display after in vivo llama immunization, isolation of lymphocytes, mRNA extraction, and cDNA library generation. After selection, the clones can be tested by ELISA on Ag-coated plates and positive hits sequenced and cloned into bacterial expression vectors. This workflow frequently allows production of multiple extremely high-affinity binders. Nanobodies tend to have a relatively longer CDR3 ranging between 15 and 20 residues, evolved to compensate for the presence of only three CDRs instead of six. This extralong loop allows penetration of cryptic epitopes often important for protein function, a property that is also facilitated by the small size of nanobodies (33).

In this work, we present a potent C4b-specific–inhibiting nanobody (hC4Nb8) and describe the crystal structure of the Ag:nanobody complex, which, together with a three-dimensional (3D) reconstruction of the C4b2 proconvertase, rationalizes its mode of action with respect to inhibition of the complement CP. It is also demonstrated that the hC4Nb8 exerts efficient CP complement inhibition in vivo in a human C4 (hC4) knock-in transgenic mouse model and in vitro in the context of the CNS.

Human C4, C4b, C2, and CR1 CCP1-3 were purified as previously described (3436). For production of the nanobodies, after three immunization boosts with a total of 400 μg of C4 and 400 μg of C4b in 3 wk intervals, blood was withdrawn from the llama and used to purify peripheral blood leukocytes and isolate mRNA. The phage display library was generated as described previously (37). Nanobody hC4Nb8 was selected by two rounds of phage display. In the first round, 1 μg of C4 in 100 μl of PBS was coated in one well of a Nunc MaxiSorp plate. The well was then blocked by addition of 2% (w/w) BSA in PBS containing 0.1% Tween 20 (PBS-T). The well was washed with 3 × 300 μl of PBS-T and incubated with the phage library for 1 h. Subsequently, the well was washed 15 times with 300 μl of PBS-T and 15 times with 300 μl of PBS. Binders were eluted by addition of 100 μl of 0.2 M glycine (pH 2.2) for 10 min. The eluted phages were neutralized by addition of 15 μl of 1 M Tris (pH 9.1) and used for infection on ER2738 cells. The second round of selection was done essentially as the first round but now using only 0.1 μg of C4. After the last round of selection, 96 colonies were incubated in a 96-well tray, induced by addition of IPTG to a final concentration of 1 mM, and incubated overnight at 30°C. On the following day, the plate was centrifuged at 2000 rpm for 10 min, and the supernatant was used in an ELISA to test for binding to C4. For the ELISA, 10 μg of C4 in 10 ml of PBS was added to a 96-well Nunc MaxiSorp plate (100 μl/well) and incubated overnight at 4°C. The plate was blocked by addition of 200 μl of 2% BSA in PBS-T (blocking solution) and incubated at room temperature (RT) for 2 h. The plate was then washed three times in PBS-T and 50 μl of supernatant plus 50 μl of blocking solution was added to each well. After 1 h of incubation at RT, the plate was washed three times in PBS-T, and 100 μl of HRP-conjugated E-Tag Ab (Bethyl Laboratories) diluted 1:10,000 in blocking solution was added to each plate. The plate was incubated for 1 h at RT and washed three times in PBS-T, followed by addition of 100 μl 3,3′,5,5′-tetramethylbenzidine substrate per well. The reaction was stopped by addition of 100 μl of 1 M HCl, and absorbance was measured at 450 nm. Phagemids of positive clones from the ELISA assays were purified and sequenced. The sequence coding for hC4Nb8 was amplified by PCR using a forward primer with restriction site NdeI 5′-GGGAATTCCATATGCAGGTGCAGCTCGTGGAGACG-3′ and reverse primer with restriction site XhoI 5′-CCCAAACTCGAGTGAGGAGACGGTGACCTGG-3′. Amplified inserts and vector pET22b+ were cut with the corresponding restriction enzymes. Inserts were ligated into the vector using DNA ligase (Thermo Fisher Scientific) according to supplier’s instructions. The resulting plasmids were sequenced by dideoxyribonucleotide sequencing with sequencing primer T7. Plasmids were transformed in LOBSTR (38) chemically competent cells and grown on an Luria broth (LB) agar plate containing ampicillin (100 μg/ml) and chloramphenicol (35 μg/ml). A single colony was picked from the plate and used to inoculate an LB medium preculture containing antibiotics and 0.4% glucose and grown at 37°C overnight with shaking at 150 rpm. The preculture was diluted 50 times into 2 L LB Broth medium (Sigma-Aldrich) containing antibiotics and 0.4% glucose. Cells were grown at 37°C and 150 rpm until OD600 = 0.6–0.8; protein expression was induced by addition of 0.5 mM IPTG. Protein expression was carried out for 18 h at 20°C and 150 rpm. The cells were pelleted by centrifugation at 6000 rpm for 20 min and resuspended in PBS containing 400 mM NaCl and 20 mM imidazole (resuspension buffer). For nanobody purification, the resuspended pellet was sonicated, and cell debris was removed by centrifugation. The supernatant was filtered through 0.45-μm cellulose filters and loaded on HisTrap FF Crude 5-ml column (GE Healthcare) equilibrated in resuspension buffer. The column was washed with 50 ml of resuspension buffer and eluted with three fractions of 5-, 11-, and 11-ml volume of resuspension buffer supplemented with 400 mM imidazole. The eluate was dialyzed against 2 l of 20 mM NaOAc and 50 mM NaCl (pH 5.5), purified further by cation exchange chromatography on a Source15S (GE Healthcare) 1-ml column, and eluted with a 15-ml linear gradient from 50 to 500 mM NaCl. The fractions containing the nanobody were pooled after SDS-PAGE analysis and concentrate; the sample was further purified by gel filtration on a 24-ml Superdex 75 Increase column (GE Healthcare) equilibrated in 20 mM HEPES and 150 mM NaCl (pH 7.5). For the in vivo experiments, endotoxin was removed from the samples as in (39), and the endotoxin level that was quantified by the Thermo Fisher Scientific Endotoxin Quantitation Kit was lower than 2 endotoxin units/ml for all samples. For immunostaining the nanobody insert fused to the IgY H chain was cloned into pcDNA3.1, and the plasmid DNA was purified using the QIAGEN Gigaprep Purification Kit. HEK293f cells were transfected with 1:3 DNA/PEI ratio, and protein expression was carried out for 5 d. All the purification steps were carried out using single-use plastic to avoid endotoxin contamination. The cells were harvested by centrifugation at 4000 rpm for 15 min, and the supernatant was filtered through 0.2-μm filters, the pH was adjusted to 8.5, and the sample was loaded on a HisTrap excel 5-ml column (GE Healthcare) equilibrated in 50 mM Tris-HCl (pH 8.5), 500 mM NaCl, and 10 mM imidazole. The protein was eluted with 30 ml of equilibration buffer supplemented with 500 mM imidazole. The sample was diluted 10-fold in 50 mM Tris-HCl (pH 8.5) and loaded on a HisTrap FF 5-ml column (GE Healthcare) equilibrated in 50 mM Tris-HCl (pH 8.5), 500 mM NaCl, and 10 mM imidazole. The protein was eluted with 10 ml of equilibration buffer supplemented with 500 mM imidazole; the elution was concentrated using 50-kDa MWCO Filters (Sartorius Stedim Biotech) and desalted in PBS using PD-10 Desalting Columns (GE Healthcare) according to manufacturer’s instructions. The endotoxin level calculated using the Thermo Fisher Scientific Endotoxin Quantitation Kit was below 2 endotoxin units/ml. For site-specific biotinylation, the AVI-tag peptide (GLNDIFEAQKIEWHE) was introduced at the C-terminal end of hC4Nb8 with the QuikChange Lightning Site-Directed Mutagenesis Kit (Agilent Technologies) with forward primer 5′-GCACCACGGCCTGAACGATATTTTTGAAGCGCAGAAAATTGAATGGCATGAATGAGATCCGGCTGC-3′ and reverse primer 5′-GATCTCATTCATGCCATTCAATTTTCTGCGCTTCAAAAATATCGTTCAGGCCGTGGTGGTGGTGGTG-3′. The protein was purified as the wild-type (WT) hC4Nb8 and dialyzed against 20 mM HEPES-NaOH and 150 mM NaCl (pH 7.5). The buffer was supplemented with 0.15 mM biotin, 2 mM ATP, and 5 mM MgCl2, BirA ligase was added at 1:5 w/w ratio, and the biotinylation reaction was carried out for 16 h at 25°C. BirA ligase was removed by ion exchange chromatography, and after testing the biotinylation success by a pull-down assay of hC4Nb8-AVI-biotin with streptavidin-agarose beads (Thermo Fisher Scientific), the biotinylated nanobody was stored at −80°C until use. Single point mutations were introduced in the hC4Nb8 WT sequence using the QuickChange Lightning Site-Directed Mutagenesis Kit (Agilent Technologies).

An expression plasmid encoding mouse C4B with the endogenous signaling peptide was designed with a C-terminal His-tag. The construct was obtained by PCR amplification of the full C4 insert from pC427A-K1324N, kindly provided by D. Isenman, and subcloning into pcDNA3.1(+) using NotI and XbaI restriction sites. The recombinant mC4 was expressed in HEK293f cells maintained at 37°C, 8% CO2, and 125 rpm in serum‐free FreeStyle 293 Expression Medium (Invitrogen). Cells were transiently transfected using final concentrations of 2 mg/l polyethyleneimine (PEI 25K; Polysciences) and 1 mg/l plasmid DNA. The conditioned medium was harvested 4 d posttransfection and adjusted to pH 7.8 with 25 mM HEPES-NaOH. The secreted mC4-His was applied to a 5 ml of HisTrap (GE Healthcare) in 300 mM NaCl and 20 mM HEPES (pH 7.8), washed with 25 mM and eluted with 300 mM imidazole, followed by a final polishing step on a 24-ml Superdex 200 Increase (GE Healthcare) equilibrated in 20 mM HEPES (pH 7.5) and 150 mM NaCl. To generate mC4b, the same protocol was used as for hC4b generation (35).

To test binding to C4 and C4b, the nanobody was preincubated with the Ag in a 5-fold molar excess for 5 min at 4°C, and the sample was analyzed by analytical gel filtration on a 24-ml Superdex 200 Increase column (GE Healthcare) equilibrated in 20 mM HEPES-NaOH and 150 mM NaCl (pH 7.5). Complex formation was monitored by the shift to a lower retention volume of the peak compared with a control run in which the same amount of C4 or C4b was injected. To monitor proconvertase formation in the presence of hC4Nb8, the nanobody was incubated with the proconvertase for 5 min at 4°C at a C2/C4b/Nb ration of 1.3:1:6 M, and the sample was injected onto a Superdex 200 Increase column (GE Healthcare) equilibrated in 20 mM HEPES-NaOH, 150 mM NaCl, and 2 mM MgCl2 (pH 7.5). Disruption of proconvertase formation was monitored by an increased intensity of the C4b (11.5 ml) and C2 (13 ml) peaks on the chromatograms and a decrease in the C4b2 peak (10.5 ml).

For the FI cleavage assay, C4b was mixed with CR1 CCP1-3, FI (Complement Technology), and hC4Nb8 at the molar ratio of 1:10:0.5:1 in 20 mM HEPES-NaOH and 50 mM NaCl (pH 7.5) and incubated for 2, 4, 8, 16, and 24 h at 37°C. A control reaction with C4b/CR1/FI = 1:10:0.5 was carried out in parallel and analyzed by reducing SDS-PAGE analysis at the 24 h time point (shown in Fig. 2). The cleavage reactions were stopped by addition of 3 mM PMSF and reducing SDS-PAGE loading buffer. The experiment was repeated twice.

For a test of the influence of nanobodies on the human CP, each well in a 96-well MaxiSorp plate (catalog no. 446612; Thermo Fisher Scientific) was coated with 100 μl of 15 μg/ml heat-aggregated human IgG diluted in 50 mM sodium carbonate (pH 9.6) (Ampliqon) and incubated overnight. Residual binding sites in the wells were blocked by TBS (10 mM Tris [pH 7.4] and 140 mM NaCl) supplemented 1 mg/ml human serum albumin for 1 h at RT, then washed trice with TBS-T. Nanobodies were diluted in veronal buffer (4 mM barbital, 145 mM NaCl, 2 mM CaCl2 and 1 mM MgCl2 [pH 7.4]) supplemented with 0.2% normal human serum (NHS), and 100 μl was added to wells in duplicates. The wells were incubated for 1.5 h at 37°C in a humidity box, then washed trice with TBS-T containing 5 mM CaCl2 (TBS-T/Ca). Deposited C3 was detected using 100 μl of biotinylated rabbit anti-C3d Ab (cat no. A0063; Dako [anti-C3d]) diluted to 0.5 μg/ml in TBS-T and incubated for 2 h at RT followed by three washes in TBS-T. Then, 100 μl of 1 μg/ml europium-labeled streptavidin (catalog no. 1244-360; PerkinElmer) diluted in TBS-T supplemented with 25 μM EDTA were added to the wells and incubated for 1 h at RT. The wells were washed trice in TBS-T, then 200 μl of enhancement buffer (catalog no. Q99800; Ampliqon) was added to each well. The fluorescence signal, read as time-resolved fluorometry, was measured using a VICTOR5 Multilabel Plate Reader (PerkinElmer) with excitation wavelength of 350 nm and emission 610 nm; the output was in counts per second. The test for the influence of the nanobodies on the LP assay was performed in a similar manner, except that a serum concentration of 1% or 50% was used, as described before (37). In this assay, the surface is coated with mannan instead of IgG. The effect of nanobodies on the murine LP was similarly tested on mannan-coated surfaces. In this case, murine serum from male C57Bl6 mice diluted to 0.33% was used, and the deposited C3 was quantified using a rat anti-mouse C3 Ab (catalog no. CL7503NA; CEDARLANE) at 0.25 μg/ml Tris-buffered saline containing 0.1% Tween 20 and 5 mM Ca2+, followed by wash and incubation with biotinylated rabbit anti-mouse IgG (E0468; Dako) at 0.25 μg/ml Tris-buffered saline containing 0.1% Tween 20 and 5 mM Ca2+. As for the assays described above, this is followed by europium-labeled streptavidin and reading of the signal. All experiments were carried out in duplicates.

Five micrograms of biotinylated hC4Nb8-AVI (0.87 mg/ml) were diluted in 150 μl of PBS and incubated with 20 μl of Streptavidin Plus UltraLink resin beads (Thermo Fisher Scientific) equilibrated in PBS for 30 min at 4°C with inversion. The supernatant was removed after centrifugation, and the beads were washed three times for 10 min at 4°C with 150 μl of PBS. For each pull-down experiment, 15 μl of marmoset plasma (kindly provided by Q. Zhang and G. Feng) was thawed on ice and incubated with 20 μl of streptavidin-agarose beads, either conjugated to hC4Nb8 or unconjugated, and equilibrated in PBS for 30 min at 4°C with inversion. After incubation, the serum flow through (FT) was removed, and the beads were washed six times for 10 min at 4°C (W1–W6) with 150 μl of PBS. After the washes, the samples were analyzed by reducing SDS-PAGE. For the SDS-PAGE analysis, the input and the FT were diluted 1:100 in PBS, and 10 μl were loaded onto the gel. Eight microliters of W2 and ten microliters of W4 to W6 were loaded on the gel, respectively. Residual proteins bound to the beads were analyzed by loading 10 μl of the beads resuspended in 20 μl of PBS. A 4× concentration of Laemmli buffer containing 2-ME was added, and the samples were boiled at 95°C for 5 min. Ten microliters of Precision Plus Protein Kaleidoscope Prestained Protein Standards was used as molecular mass marker (kDa).

Deglycosylated C4b, prepared according to (35), was mixed in a 1:5 M ratio with hC4Nb8. The complex was purified by size-exclusion chromatography (SEC) on a 24-ml Superdex 200 Increase column equilibrated in 20 mM HEPES-NaOH and 150 mM NaCl (pH 7.5). The fractions containing the complex were concentrated to 8 mg/ml and used for crystallization screening with ProPlex commercial screens (Molecular Dimensions), Structure Screens (Molecular Dimensions), or PEGRx screens (Hampton Research). The screens were dispensed using an Oryx robot (Douglas Instruments), in which the drops contained a 0.15-μl sample and 0.15-μl reservoir. After 5 d, plate-shaped crystals appeared in 100 mM HEPES-NaOH (pH 7), 10% w/w PEG4000, and 10% v/v 2-propanol. Crystals were cryoprotected in the reservoir buffer containing 20% ethylene glycol or 25% glycerol, and x-ray diffraction data from cryocooled crystals were collected at PETRA III beamlines P13 and P14 (European Molecular Biology Laboratory, Hamburg, Germany). The diffraction data were processed using XDS (40). C4b coordinates (Protein Data Bank identification number [PDB ID] 5JPN) without the C345c domain were used for initial structure determination by molecular replacement in phenix.phaser (41). The nanobody in PDB ID 5IMM with 80% sequence identity without CDRs was used as starting model for the nanobody and placed manually with Coot (42) after density modification with phenix.density_modification (43), and the CDRs were built manually. Refinement was carried out in phenix.refine (44) with options rigid-body, translation-liberation-screw-rotation, individual sites, individual B factors, and use of torsion angle noncrystallographic symmetry restraints. Data collection and refinement statistics are presented in Table I. Structure factors and coordinates are deposited under PDB ID 6YSQ (www.rcsb.org).

The samples were applied to glow-discharged carbon-coated copper GC400 grids for 5 s at 20 μg/ml. After application, the sample was blotted and stained with 3 μl of 2% w/v uranyl formate by two sequential rounds of staining immediately followed with blotting, followed then by a 1 min stain and blot before the grid was air-dried. The micrographs were recorded automatically using Leginon on a Tecnai T12 G2 transmission electron microscope operating at 120 kV and equipped with a TemCam-F416 detector (Tietz Video Image Processing Systems). The defocus range was −0.7 to −1.7 μm, exposure time was 750 ms, and original magnification was 67,000, yielding a pixel size of 3.15 Å. The data were processed without performing contrast transfer function correction, and particles were automatically picked by CisTEM (45) and extracted with a box size of 96 × 96 pixels. The particles used for 3D classification were selected after reference-free two-dimensional (2D) classification in RELION 3.0.7 (46), using a mask of 250 Å. The remaining particles were used for 3D classification based on automated generation of the initial model in RELION for the C4b2 complexes or with C4b low pass filtered to 25 Å for the hC4Nb8:C4b complex. Fitting of the of the crystal structure of C4b (PDB ID 5JPN) (35) into the 3D electron microscopy (EM) envelope of the hC4Nb8:C4b complex or of the hC4Nb8:C4b crystal structure in the 3D EM envelope of the C4b2 complex was performed manually in PyMOL (version 2.3.0). A total of 29,274 and 6,631 particles were used for the hC4Nb8:C4b or C4b2 3D reconstruction, respectively. For fitting of C4b, C2a, and C2b in the NSEM envelope, C3bB (PDB ID 2XWJ) was first fitted in UCSF Chimera version 1.14 using the built in “Fit in Map” function (47). C3bB models without the C3b TE domain, the FB serine protease (SP) domain, or missing both of these domains were also used for fitting the EM envelope; however no difference was obtained in the output compared with the fitting of the full complex. A coordinate file was written for the fitted C3bB, and all the following steps were carried out in PyMOL (version 2.3.0). C4b in the C4b:hC4Nb8 complex was aligned on the MG6 domain of C3b, the FB CCP domains were used to model the C2 CCP domains, whereas the structure of C2a (PDB ID 2I6S) was superimposed on the VWA domain in the fitted C3bB. The structure of the cobra venom factor-FB (CVFB) complex (PDB ID 3HRZ) was superimposed on the fitted C3bB. Although an improved fit to the 3D reconstruction could be obtained by further manual adjustment of the C4b C345c and the C2 VWA-SP domains, this was avoided in the model presented in this study for simplicity and because of the limited resolution.

The surface plasmon resonance (SPR) measurements were conducted on a Biacore T200 instrument (GE Healthcare). Streptavidin was diluted to 10 μg/ml in 10 mM sodium acetate (pH 4.5) and immobilized to 100 response units (RU) on the carboxymethylated dextran surface of a Sensor Chip CM5 (XanTec bioanalytics GmbH) using an amine coupling kit. The biotinylated nanobody was injected on the immobilized streptavidin at 30 μg/ml, giving 24 RU of captured nanobody. The binding measurements were performed in 20 mM HEPES, 150 mM NaCl, 3 mM MgCl2, and 0.05% Tween 20 (pH 7.5) at 30 μl/min flow rate. At the end of each concentration, measurement of the surface was regenerated by injection of 100 mM glycine (pH 2.7) for three cycles of 10 s contact time. Sensorgrams were recorded at concentrations of 56.25, 112.5, 225, 450, and 900 nM for C4, and at 1.5625, 3.125, 6.25, 12.5, and 25 nM for C4b. Fitting of the data were performed for all the measured concentrations simultaneously, using BIAevaluation software. The apparent KD were calculated from the ratio between the association and dissociation rate constants. The presented data are the mean ± SE from three replicate experiments. For the competition experiments, the coupling procedure was the same, but the RU of immobilized nanobody were 50. The complexes were preincubated for 1 h on ice prior to injection on the sensor chip at the following molar ratios: C4b/C2 = 1:1.1; C4b/CR1 = 1:100; C4b/hC4Nb8 WT and mutants = 1:1; the running buffer was 20 mM HEPES, 150 mM NaCl, 3 mM MgCl2, and 0.05% Tween 20 (pH 7.5); and all experiments were done in triplicates. The calorimetric titration experiments were carried out using a MicroCal VP-ITC instrument in 20 mM HEPES-NaOH and 150 mM NaCl (pH 7.5). The proteins were either dialyzed or repurified by gel filtration in running buffer right before the experiments. All experiments were carried out at 30°C, and the ligands were titrated with 30 injections of a 10-μl volume. C4 was used at a 2.5-μM concentration, C4b was used at 1.25 and 2.5 μM, whereas hC4Nb8 and the W53A plus R107A mutant were used at 12.5 and 25 μM. All experiments were conducted in duplicates. For subtraction of heats of dilution, 12.5 and 25 μM solutions of the ligands were titrated in running buffer. Before each run, the samples were centrifuged at 24,000 rcf for 5 min at 4°C, degassed for 5 min, and centrifuged again prior to concentration measurement on a NanoDrop ND 1000 Spectrophotometer (Saveen Werner).

All animal experiments were carried out in agreement with the institutional guidelines at Harvard Medical School, following approval of ethical protocols by the local Institutional Animal Care and Use Committee (protocol numbers IS00000748, IS00000111, and IS00002660) and per applicable laws and regulations. Five- to-ten- week-old mixed sex C57BL6/C4 knockout (KO) mice backcrossed with hC4A and hC4B as previously described (M. Yilmaz, E. Yalcin, J. Presumey, E. Aw, C. W. Whelan, B. Stevens, S. A. McCarroll, and M. C. Carroll, manuscript in preparation) were used in all experiments (hC4 AB/−). For immunostaining experiments, 14-μm spleen sections were cryosectioned and fresh frozen at −80°C and re-equilibrated at RT for 20 min prior to fixation with acetone. Afterward, acetone evaporation blocking buffer containing PBS, 2% w/v BSA, 5% v/v FCS, and 0.1% Tween 20 was added to each slide and incubated for 1 h at RT, after which the primary Ab consisting of rabbit anti-hC4c (Dako) or IgY-hC4Nb8 was added. The Abs were diluted 1:300 in PBS-T, and the slides were incubated for 1 h at RT before washing three times for 3 min in PBS, 0.1% Tween 20, and addition of secondary anti-rabbit–Alexa 568 or anti-chicken–Alexa 488 at 1:300 dilution in PBS and 0.1% Tween 20. After 1 h of incubation at RT, the slides were washed three times for 5 min in PBS and 0.1% Tween 20 and mounting medium (electron microscopy) without DAPI was added, and the slides were sealed and dried at 4°C overnight. The slides were imaged at magnification ×20 on a confocal microscope (Olympus FluoView FV1000 confocal system).

For the hemolytic assay, the experiments were repeated in triplicates (n = 3 mice). Approximately 100 μl of mouse blood was collected into a tube containing 5 μl of 0.1 M EDTA and kept on ice for the whole experimental procedure. The serum (supernatant) was collected after centrifugation at 10,000 rpm for 8 min, and samples were diluted 1:60, 1:240, 1:1000, or 1:50,000 for treated and nontreated conditions. Each experimental condition contained a 25-μl sample (mouse serum ± 10 μg/ml hC4Nb8), 25 μl C4-deficient guinea pig sera and 15 μl of EA cells (1 × 108 cells/ml; 65 μl working volume). The absorbance value obtained from mixing 50 μl of water and 15 μl EA cells represented complete hemolysis (100%), whereas 50 μl gelatin veronal buffer and 15 μl of EA cells represented the background (negative control). The plate was incubated at 37°C for 30 min, centrifuged at 1600 rpm for 5 min, the supernatant was collected, and absorbance was measured at 415 nm.

For passive immunization experiments, mice were injected with 1 mg of rabbit anti-PE Ab (Rockland Immunochemicals) 24 h before subcutaneous injection of 5 μg of hC4Nb8 or Lag16 and 1 μg of PE in 10 μl of HBSS in each leg. The mice were sacrificed 24 h after administration of the treatment and popliteal lymph nodes were collected in 4% paraformaldehyde; after 2 h they were transferred to 30% w/v sucrose and incubated at 4°C overnight. No distinction was made between lymph nodes coming from the left or from the right leg. The lymph nodes were cryosectioned in 14 μm sections and the slides were fresh frozen at −80°C, prior to staining with the different markers with the same protocol used for spleen staining. In the first round of experiments, the primary Abs used were 8C12-biotin for FDCs and B220-FITC for the B cell follicle, whereas in the second round of experiments the FDCs were labeled with 7E9-pacific blue. The Abs were diluted 1:300 in PBS-T and the slides were incubated for 1 h at RT. The slides were washed three times for 3 min in PBS, 0.1% Tween 20 prior to addition of streptavidin conjugated to Alexa488 at 1–300 dilution in PBS, 0.1% Tween 20. After 1 h incubation at RT, the slides were washed three times for 5 min in PBS, 0.1% Tween 20 and mounting medium (Electron microscopy) without DAPI was added and the slides were sealed and dried at 4°C overnight. The slides were imaged on a confocal microscope (Olympus FluoView FV1000 confocal system) at 40× magnification and one B cell follicle was imaged in each field of view. Four images were collected for each mouse without side segregation of the lymph nodes. In the first round of experiments one slide was imaged for each mouse, whereas in the second round three slides per mouse were imaged, for a total of 12 images per mouse. For data analysis, the images were analyzed in ImageJ. The images were thresholded and the colocalized area of the FDC channel with the PE channel was calculated. The area around the FDC network was isolated, transformed to binary area and the ratio with the binary PE colocalized area was calculated and plotted in GraphPad Prism 6.0. The statistical analysis was carried out in GraphPad Prism 6.0 with one-way ANOVA and multiple comparisons, in which the mean of each column was compared with the mean of every other column (Supplemental Table II).

Neurons were differentiated from induced pluripotent stem cells by expression of Neurogenin2 and grown on a Matrigel matrix for 14 d after differentiation. All the assay steps were carried out using an INTEGRA VIAFLO 96/384 with 96-well tips and a 300-μl head. All aspirations and dispenses over cells were done at the lowest speed. The cells were sensitized to complement deposition by addition of anti-NCAM Ab (AB5032; Millipore) at 2.5 μg/ml in 10 mM HEPES-NaOH, 140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 2 mM CaCl2 (pH 7.2) (binding buffer), and incubation at RT for 20 min to generate ICs. The cells were then washed with 100 μl of binding buffer in each well prior to addition of 10% NHS or C4-depleted serum (C4-Dpl NHS; Complement Technology) diluted in gelatin veronal buffer including Ca2+ and Mg2+ (Complement Technology) with the desired nanobody dilution (hC4Nb8 or Lag16). Complement deposition was carried out for 30 min at 37°C. The cells were washed twice with prewarmed neurobasal medium and live stained with FITC-conjugated anti-C3c Ab (F0201; Dako) diluted 1:200 for 20 min at 37°C. The cells were washed once in PBS and fixed for 7 min in 4% paraformaldehyde, washed again in PBS, and stained for 1 h at 4°C with eFluor 660–Tuj1 (anti–β-tubulin) Ab (50-4510-82; Invitrogen) in 50 μl per well. The cells were washed once in PBS, and the plate was kept at 4°C until imaging. The plates were imaged on a PerkinElmer Opera Phenix instrument; 11 fields of view per well were collected at 63× magnification in z-stack mode. Data were analyzed using a pipeline created in the PerkinElmer Harmony software. Briefly, images of C3 and Tuj1 were thresholded and masked: Tuj1+ neurites were identified by the FindImageRegion module using a common threshold, spots of C3 were segmented with the FindSpots module and masked with the Tuj1+ neurite region (C3/Tuj1+ objects) and area was quantified per C3/Tuj1+ object and then summed across all C3/Tuj1+ objects. C3 deposition was then calculated per well as the sum of C3/Tuj1+ object area as a percentage of total Tuj1+ neurite area. Three plates were analyzed for a total of (n = 24) wells per dose of hC4Nb8 or Lag16 in 10% NHS and (n = 6) wells per hC4Nb8 or Lag16 in C4-Dpl NHS. Statistical analysis was carried out in R version 3.6.1 with one-way or two-way ANOVA and Tukey honestly significantly different post hoc tests, in which the mean of each condition was compared with the mean of every other condition. ANOVA was performed both across serum conditions (10% NHS and C4u-Dpl) at matching 20 μg/ml doses of hC4Nb8 and Lag16 (two-way) and within the deposition-permissive serum (10% NHS) condition to examine dose effects of hC4Nb8 versus Lag16 control (one-way) (Supplemental Table III).

Bio-layer interferometry (BLI) measurements were carried out using an Octet RED96 System (ForteBio) with the plate at 30°C and shaking at 1000 rpm. The running buffer for all experiments was 20 mM HEPES-NaOH, 150 mM NaCl, 3 mM MgCl2, and 0.05% Tween 20 (pH 7.5). The AVI-tagged site-specifically biotinylated hC4Nb8 was immobilized on a streptavidin biosensor (ForteBio) at 1.25 μg/ml, and mC4b was present in the 2-fold dilution series from 12.5 to 0.78 nM in the association phase. Durations of 600 s for the association and 1200 s for the dissociation phase were used. The data were processed by subtraction of a 0 nM measurement and fitted in the Octet System Data Analysis software with a 1:1 Langmuir binding model.

A llama was immunized with a total of 400 μg of hC4 and 400 μg of hC4b in three consecutive immunization boosts separated by intervals of 3 wk. The llama blood was collected, and lymphocytes were isolated for mRNA extraction and cDNA synthesis. Nanobodies were cloned as fusions to protein pIII expressed on the surface of M13 phages. Potential complement inhibitors were isolated by two rounds of phage display selection on plates coated with hC4. The positive clones were identified by ELISA, sequenced, and cloned into a bacterial expression vector. The inhibitory activity of multiple nanobodies was investigated in CP and LP deposition assays on aggregated IgG and mannan-coated surfaces, respectively. After these preliminary experiments, hC4Nb8 appeared to be the strongest inhibitor of CP- and LP-driven C3 fragment deposition but not C4 fragment deposition in human serum (Fig. 1B–E). The nanobody also inhibited LP-driven C3 deposition in murine serum (Fig. 1F). To test whether the nanobody would inhibit complement in conditions where C4 is present in high concentrations and thereby competes with C4b for hC4Nb8 binding, we carried out C3 deposition assays in 50% NHS, triggering LP activation on a mannan-coated surface. In this study, we included the C3-specific nanobody hC3Nb1 (37), which efficiently suppresses the activity of the AP C3 convertase while allowing the CP C3 convertase to turn over C3. In this manner, we were able to selectively measure the LP contribution without an overwhelming AP-driven C3 deposition (Fig. 1G). Also, in these near-physiological conditions, we observe that the hC4Nb8 nanobody inhibits C3 deposition in a dose-dependent manner. The level of C3 deposition in the presence of both hC4Nb8 and hC3Nb1 is comparable to the addition of 10 mM EDTA in 50% NHS. To evaluate whether hC4Nb8 may act as a C4-specific inhibitor in nonhuman primate disease models, we also tested the cross-reactivity of hC4Nb8 with marmoset (Callithrix jacchus) C4 (CjC4) by pull-down experiments. We consistently observed presence of the CjC4 bands (CjC4 α-, β-, and γ-chains) on the hC4Nb8 streptavidin beads in eight experiments with plasma from different animals compared with the bands observed with streptavidin beads without hC4Nb8 (Fig. 1H). This result demonstrates that hC4Nb8 cross-reacts with CjC4.

FIGURE 1.

The hC4Nb8 nanobody is a potent inhibitor of the CP and LP C3 convertase. (A) The proteolytic cascade initiated upon activation of the CP and LP. The proteolytic degradation of C3b to iC3b and C3dg is enabled on host cells because of the presence of cofactors. (B and C) The hC4Nb8 has no effect on C4b deposition but potently inhibits C3b deposition upon CP activation of NHS by the IgG surface. (D and E) Same as in (B) and (C) but with LP-driven C4b and C3b deposition from NHS onto a mannan surface. (F) In murine serum, hC4Nb8 also inhibits C3 convertase activity (i.e., suppresses C3 fragment deposition via the LP onto a mannan surface). The percentages of C4b and C3b deposition are relative to the signal at the same serum dilution without nanobody addition. The average C4 concentration in the NHS dilutions used for the assays in (B)–(E) is depicted by a dashed line. The nanobody hC3Nb1 W102A is an inactive mutant of the AP inhibitor hC3Nb1 (37); (G) hC4Nb8 inhibits LP-driven C3b deposition onto a mannan surface in 50% NHS to levels comparable to addition of 10 mM EDTA when the AP is simultaneously inhibited by the hC3Nb1 nanobody. (H) SDS-PAGE analysis of pull-down experiment from Cj plasma using biotinylated hC4Nb8 bound to streptavidin beads (+hC4Nb8, beads lane), whereas no C4 binding is observed when using streptavidin beads but no biotinylated hC4Nb8 (−hC4Nb8, beads lane). The α-, β-, and γ-chains of Cj C4 are assigned based on comparison with the molecular mass of the these chains in hC4, but a minor content of α′-chain from C4b is possible. Bds, beads; Inp., input; W, wash fractions.

FIGURE 1.

The hC4Nb8 nanobody is a potent inhibitor of the CP and LP C3 convertase. (A) The proteolytic cascade initiated upon activation of the CP and LP. The proteolytic degradation of C3b to iC3b and C3dg is enabled on host cells because of the presence of cofactors. (B and C) The hC4Nb8 has no effect on C4b deposition but potently inhibits C3b deposition upon CP activation of NHS by the IgG surface. (D and E) Same as in (B) and (C) but with LP-driven C4b and C3b deposition from NHS onto a mannan surface. (F) In murine serum, hC4Nb8 also inhibits C3 convertase activity (i.e., suppresses C3 fragment deposition via the LP onto a mannan surface). The percentages of C4b and C3b deposition are relative to the signal at the same serum dilution without nanobody addition. The average C4 concentration in the NHS dilutions used for the assays in (B)–(E) is depicted by a dashed line. The nanobody hC3Nb1 W102A is an inactive mutant of the AP inhibitor hC3Nb1 (37); (G) hC4Nb8 inhibits LP-driven C3b deposition onto a mannan surface in 50% NHS to levels comparable to addition of 10 mM EDTA when the AP is simultaneously inhibited by the hC3Nb1 nanobody. (H) SDS-PAGE analysis of pull-down experiment from Cj plasma using biotinylated hC4Nb8 bound to streptavidin beads (+hC4Nb8, beads lane), whereas no C4 binding is observed when using streptavidin beads but no biotinylated hC4Nb8 (−hC4Nb8, beads lane). The α-, β-, and γ-chains of Cj C4 are assigned based on comparison with the molecular mass of the these chains in hC4, but a minor content of α′-chain from C4b is possible. Bds, beads; Inp., input; W, wash fractions.

Close modal

The strong inhibition of C3 deposition in human serum suggested that hC4Nb8 prevents either proconvertase assembly or interferes with C4b2a recognition of the C3 substrate. To gain functional insight into the mechanism of inhibition, we evaluated the binding of the nanobody to native C4 and C4b by SEC. For C4b, the presence of the nanobody resulted in a clear shift in the elution volume, indicating formation of a stable C4b-hC4Nb8 complex. In contrast, C4 eluted identically in the presence and absence of hC4Nb8, suggesting that the epitope may only be accessible in C4b or that the affinity for native C4 is too low to be detected in this assay (Fig. 2A, 2B). We next investigated the formation of the CP and LP proconvertase C4b2 in the presence of the nanobody (Fig. 2C, 2D). In the presence of molar excess of hC4Nb8, C2 was unable to interact with C4b, suggesting that one mechanism whereby hC4Nb8 inhibits C3 deposition is by interfering with proconvertase assembly.

FIGURE 2.

hC4Nb8 binds selectively to C4b and disrupts C4b interactions with C2 and CR1. (A) SEC binding assay with hC4b, (B) hC4, and (C) hC4b2. The chromatograms for elution of C4, C4b, or C4b2 in the presence of a 5-fold molar excess of hC4Nb8 are depicted with a full line, whereas the chromatograms of C4b, C2, and hC4Nb8 are depicted with a dashed line. The C4b2 chromatogram is depicted with a dashed and dotted line. (D) Reducing SDS-PAGE analysis of the chromatogram in (C) (i.e., hC4Nb8 mixed with hC4b2) with retention volume of each sample given on top of the lane. (E) CR1 cofactor activity on FI-mediated cleavage assay with a 1:10:0.2:1 = C4b/CR1/FI/hC4Nb8 M ratio. The C4b/FI ratio corresponds to that present with physiological C4 and FI concentrations. Incubation time is given above the gel. The positions of the C4b chains (α′-, β-, and γ-) and C4d and CR1 are indicated.

FIGURE 2.

hC4Nb8 binds selectively to C4b and disrupts C4b interactions with C2 and CR1. (A) SEC binding assay with hC4b, (B) hC4, and (C) hC4b2. The chromatograms for elution of C4, C4b, or C4b2 in the presence of a 5-fold molar excess of hC4Nb8 are depicted with a full line, whereas the chromatograms of C4b, C2, and hC4Nb8 are depicted with a dashed line. The C4b2 chromatogram is depicted with a dashed and dotted line. (D) Reducing SDS-PAGE analysis of the chromatogram in (C) (i.e., hC4Nb8 mixed with hC4b2) with retention volume of each sample given on top of the lane. (E) CR1 cofactor activity on FI-mediated cleavage assay with a 1:10:0.2:1 = C4b/CR1/FI/hC4Nb8 M ratio. The C4b/FI ratio corresponds to that present with physiological C4 and FI concentrations. Incubation time is given above the gel. The positions of the C4b chains (α′-, β-, and γ-) and C4d and CR1 are indicated.

Close modal

Based on the knowledge that the nanobody competes with C2 binding and, by comparison, with the known structures of C3b in complex with FH and FI (48), we hypothesized that hC4Nb8 inhibited CR1-mediated FI cleavage of C4b in vitro. In the experimental conditions used, the cleavage of C4 to C4c and C4d was complete after 2 h of incubation at 37°C (result not shown). When present in a 1:1 M ratio with respect to C4b, the nanobody strongly inhibits cleavage by FI, and substrate cleavage is only starting to become apparent after 24 h of incubation at 37°C (Fig. 2E).

To define the structural basis for the inhibition of CP proconvertase assembly exerted by hC4Nb8, we crystallized the hC4Nb8:C4b complex and collected x-ray diffraction data extending to a maximum resolution of 3.3 Å (Table I). The structure was determined by molecular replacement, using the structure of C4b (35, 49) as search model, whereas the nanobody was modeled from PDB ID 5IMM. The final atomic model (Fig. 3A) was obtained after a few iterations of rebuilding–refinement cycles, and refined to an Rfree value of 27%. An example of the electron density for the C4b:hC4Nb8 interface is displayed in Fig. 3B. The asymmetric unit of the unit cell contains two copies of the C4b:hC4Nb8 complex that do not differ significantly with respect to the C4b:hC4Nb8 interaction.

Table I.
Refinement statistics for structure determination of the hC4Nb8:C4b complex
Data Collection
Resolution range 48.26−3.3 (3.418−3.3) 
Space group P 1 21 1 
Unit cell 131.2 89.51 231.2 90 97.552 90 
Total reflections 538,319 (49,811) 
Unique reflections 79,980 (7,914) 
Multiplicity 6.7 (6.7) 
Completeness (%) 99.87 (99.94) 
Mean I/σ(I) 11.73 (0.65) 
Wilson B-factor 147.94 
Rmerge 0.09005 (2.284) 
Rmeas 0.09786 (2.473) 
CC1/2 0.998 (0.341) 
Refinement
 
Reflections used in refinement 80,456 (7,913) 
Reflections used for Rfree 1,566 (155) 
Rwork/Rfree 0.2212 (0.3613)/0.2698 (0.3852) 
Number of nonhydrogen atoms 26,297 
Macromolecules 26,141 
Ligands 156 
Protein residues 3,384 
Root mean square (bonds) 0.004 
Root mean square (angles) 0.91 
Ramachandran favored, allowed, outliers (%) 92.08, 7.20, 0.72 
Clash score 3.36 
Average B-factor 195.53 
Data Collection
Resolution range 48.26−3.3 (3.418−3.3) 
Space group P 1 21 1 
Unit cell 131.2 89.51 231.2 90 97.552 90 
Total reflections 538,319 (49,811) 
Unique reflections 79,980 (7,914) 
Multiplicity 6.7 (6.7) 
Completeness (%) 99.87 (99.94) 
Mean I/σ(I) 11.73 (0.65) 
Wilson B-factor 147.94 
Rmerge 0.09005 (2.284) 
Rmeas 0.09786 (2.473) 
CC1/2 0.998 (0.341) 
Refinement
 
Reflections used in refinement 80,456 (7,913) 
Reflections used for Rfree 1,566 (155) 
Rwork/Rfree 0.2212 (0.3613)/0.2698 (0.3852) 
Number of nonhydrogen atoms 26,297 
Macromolecules 26,141 
Ligands 156 
Protein residues 3,384 
Root mean square (bonds) 0.004 
Root mean square (angles) 0.91 
Ramachandran favored, allowed, outliers (%) 92.08, 7.20, 0.72 
Clash score 3.36 
Average B-factor 195.53 

Rmeas = Σhkl(N/(N−1))1/2Σi|Ii(hkl)−Ι̅(hkl)|/ΣhklΣiIi(hkl). CC1/2 is the correlation between random half-datasets. Statistics for the highest-resolution shell are shown in parentheses.

FIGURE 3.

The crystal structure of the hC4Nb8:hC4b complex at 3.3 Å resolution. (A) The nanobody (sand) binds to the Nt-α′ region (spheres) of C4b (blue) and two β-strands in the MG6 domains. (B) Omit 2mFo-DFc map for the epitope region and the hC4Nb8 CDR3 contoured at 1 σ. (CE) Detailed presentations of interactions formed between C4b and the hC4Nb8 CDR1, CDR2, and CDR3 regions, respectively. The majority of the electrostatic interactions are formed by CDR3. Putative hydrogen bonds and electrostatic interactions are indicated by dashed lines.

FIGURE 3.

The crystal structure of the hC4Nb8:hC4b complex at 3.3 Å resolution. (A) The nanobody (sand) binds to the Nt-α′ region (spheres) of C4b (blue) and two β-strands in the MG6 domains. (B) Omit 2mFo-DFc map for the epitope region and the hC4Nb8 CDR3 contoured at 1 σ. (CE) Detailed presentations of interactions formed between C4b and the hC4Nb8 CDR1, CDR2, and CDR3 regions, respectively. The majority of the electrostatic interactions are formed by CDR3. Putative hydrogen bonds and electrostatic interactions are indicated by dashed lines.

Close modal

The C4b epitope for hC4Nb8 is primarily formed by residues Pro-773-Asn781 in the last half of the α′ N-terminal region (Nt-α′) of C4b and the following two β-strands in the MG6 domain. Analysis with PISA indicates a buried surface area upon complex formation of 1736 Å2 and a shape complementarity of 0.63 between the two proteins. Both hydrophobic and polar residues in hC4Nb8 are engaged in interactions with C4b. In hC4Nb8, Trp53 from CDR2 and Ile105 from CDR3 form the hydrophobic core of the complex, together with Phe777, Phe778, and Trp798 presented by the C4b MG6 domain (Fig. 3C–E). A hydrogen bond appears to bridge Arg31 in the hC4Nb8 CDR1 with C4b Ser776 (Fig. 3C), whereas a second hydrogen bond possibly bridges Ser54 in hC4Nb8 CDR2 and the main chain of Val774 in C4b (Fig. 3D). Electrostatic interactions are formed between hC4Nb8 CDR3 residues Glu102, Glu110, and Arg107 and residues Arg785 and Glu787 residues in the C4b Nt-α′ region MG6 domain (Fig. 3E). Of notice, at the resolution of 3.3 Å, hydrogen bonds can only be considered as putative. In addition, water molecules present at the intermolecular interface cannot be modeled but most likely form additional bridging contacts between hC4Nb8 and C4b.

As the Nt-α′ region undergoes an extensive conformational change upon C4 cleavage, the binding site for hC4Nb8 is a C4b-specific neoepitope. In native C4, the future Nt-α′ region is positioned in a channel formed by the MG2 and MG3 domains and the linker region, whereas it becomes exposed in C4b (11, 35). Moreover, if hC4Nb8 were to bind the C4 MG6 domain in the same manner as observed in the C4b complex, the C4 MG7 domain would apparently sterically clash with the nanobody (Fig. 4A, 4B), suggesting that only weak binding to C4 may occur, in line with our SEC experiment (Fig. 2A). The observed hC4Nb8 epitope (Fig. 4C, left) also rationalizes the complete lack of inhibition with respect to C4 deposition (Fig. 1B, 1D). A projection of the hC4Nb8 binding site in C4b onto C4 reveals that the nanobody epitope is located far from the three regions in C4 contacted by MASP-2 and by homology C1s: 1) the scissile bond region, 2) the sulfotyrosine region in the C-terminal part of the C4 α-chain, and 3) a basic patch centered around residue 1720 in the C345c domain of C4 (11). The hC4b:hC4Nb8 atomic model is also consistent with the observed inhibition of CR1-mediated FI degradation of C4b. A structural alignment of C4b with C3b in the structure of the C3b:FH:FI complex (48) suggests that hC4Nb8 prevents binding of the CCP1-2 domains of the C4b regulators CR1, C4BP, and MCP and the subsequent interaction with FI required for degradation of C4b to C4c and C4d.

FIGURE 4.

The hC4Nb8 epitope is a neoepitope in C4b and overlaps with the C2 binding site. (A) structure of hC4Nb8 (sand) in complex with C4b (gray) with MG6 and Nt-α′ in blue; (B) model of hC4Nb8 in complex with C4 (gray) with MG6 and Nt-α′ in blue. hC4Nb8 is shown as transparent surface in sand. Notice the predicted clash with the MG7 domain. (C) Left, Epitope of hC4Nb8 on C4b with residues interacting with hC4Nb8 in sand and remaining residues in blue. Right, Sequence alignment of the epitope regions (top, MG6; bottom, the Nt-α′ region and downstream residues in the MG6 domain). Interface residues are labeled with a star. His582, Lys793 in mouse C4 and Lys793, Arg798 in Slp are colored red. (D) Comparison of the C4b:hC4Nb8 structure with the NSEM envelope (gray transparent surface) of the C4b2 complex reveals that hC4Nb8 overlaps directly with the predicted C2b (red) binding site.

FIGURE 4.

The hC4Nb8 epitope is a neoepitope in C4b and overlaps with the C2 binding site. (A) structure of hC4Nb8 (sand) in complex with C4b (gray) with MG6 and Nt-α′ in blue; (B) model of hC4Nb8 in complex with C4 (gray) with MG6 and Nt-α′ in blue. hC4Nb8 is shown as transparent surface in sand. Notice the predicted clash with the MG7 domain. (C) Left, Epitope of hC4Nb8 on C4b with residues interacting with hC4Nb8 in sand and remaining residues in blue. Right, Sequence alignment of the epitope regions (top, MG6; bottom, the Nt-α′ region and downstream residues in the MG6 domain). Interface residues are labeled with a star. His582, Lys793 in mouse C4 and Lys793, Arg798 in Slp are colored red. (D) Comparison of the C4b:hC4Nb8 structure with the NSEM envelope (gray transparent surface) of the C4b2 complex reveals that hC4Nb8 overlaps directly with the predicted C2b (red) binding site.

Close modal

The interaction of C2 with C4b is mediated through two contact points (50), with one of the binding sites in the C2b CCP domains and the second provided by the MIDAS site in the C2a VWA domain interacting with the C4b C terminus presented by the C345c domain. Located next to the hC4Nb8 epitope are multiple acidic residues (E763, E764, D768, E769, E770, E771) in the C4b Nt-α′ previously identified as important for the interaction with C2 (51) (Fig. 4C, left). Our SEC assays showed that the nanobody interferes with the C2 binding to C4b but cannot distinguish mechanistically between direct competition with C2 or an allosteric effect of hC4Nb8. To obtain the structural basis for hC4Nb8 inhibition of C2 binding, we reconstituted the CP C3 proconvertase and determined a 3D reconstruction at a resolution of 21 Å of the C4b2 complex with negative stain EM (Fig. 5). Alignment of the hC4b:hC4Nb8 complex with this 3D reconstruction reveals a major spatial overlap between hC4Nb8 and a volume of the 3D reconstruction likely to contain the C2 CCP domains (Fig. 4D). Fitting of the CCP domains of the C2 homolog FB in the EM envelope suggests that steric clashes arise because of the dramatic overlap of hC4Nb8 with the C2 CCP2 domain (Fig. 4D, inset), suggesting that the nanobody prevents C2 from binding through direct competition rather than by allosteric effects.

FIGURE 5.

In the proconvertase, C2 adopts an extended conformation distinct from the closed and open states of FB. (A) Examples of 2D classes used in the 3D classification. The 2D classes are 400 Å in each direction. (B) Resolution estimate based on Fourier shell correlation calculation in RELION. (CE) Fitting of the C4b structure in the C4b:hC4Nb8 complex and of the crystal structures of C2b and C2a in the NSEM density obtained from C4b2 reveals significant differences in the location of the SP domain (C) compared with FB open conformation in C3bB (PDB ID 2XWJ) (D) and to FB closed conformation in CVFB (PDB ID 3HRZ) (E). The noncatalytic subunit is in blue, the zymogen CCP domains are in red, whereas the VWA-SP domains are in green in all panels. The location of the SP domain is denoted by a star.

FIGURE 5.

In the proconvertase, C2 adopts an extended conformation distinct from the closed and open states of FB. (A) Examples of 2D classes used in the 3D classification. The 2D classes are 400 Å in each direction. (B) Resolution estimate based on Fourier shell correlation calculation in RELION. (CE) Fitting of the C4b structure in the C4b:hC4Nb8 complex and of the crystal structures of C2b and C2a in the NSEM density obtained from C4b2 reveals significant differences in the location of the SP domain (C) compared with FB open conformation in C3bB (PDB ID 2XWJ) (D) and to FB closed conformation in CVFB (PDB ID 3HRZ) (E). The noncatalytic subunit is in blue, the zymogen CCP domains are in red, whereas the VWA-SP domains are in green in all panels. The location of the SP domain is denoted by a star.

Close modal

The AP functional homolog of C2, FB, adopts two different conformations named closed and open. These were captured in the crystal structures of the CVFB complex (52) and the C3bB complex (53), respectively. The open conformation allows cleavage by the fluid phase protease FD, and it is adopted by FB only after binding to C3b (52, 53). The two FB conformations mainly differ by the location of the catalytic SP domain. In the CVFB complex, the SP domain is close to the first two CCP domains in FB, whereas in the open conformation the SP domain has rotated 80° and contacts the C3b CUB domain. Importantly, our 3D reconstruction reveals that the C2 SP domain appears to be in a rather different position compared with these two FB conformations. The C2 conformation is extended with the SP domain far from both the C2 CCP domains and the C4b CUB domain (Fig. 5C–E). Our model of the C4b2 complex derived by rigid-body modeling against small-angle x-ray scattering data (34) also suggested a C2 conformation different from the known FB conformations for both unbound and C4b bound C2, which is also in agreement with the cleavability of C2 in absence of C4b binding (54). In summary, our structural studies show that hC4Nb8 binds to a neoepitope in C4b formed by residues in the Nt-α′ and the MG6 domain presented upon C4 cleavage by MASP-2 or C1s. This epitope overlaps strongly with the binding site for C2 CCP domains, rationalizing the observed hC4Nb8 inhibition of CP C3 convertase formation and activity.

The C3 deposition assays, our SEC analysis, and the crystal structure of the hC4b:hC4Nb8 complex collectively suggested that the nanobody bound with high affinity to C4b and with much weaker affinity to C4, but accurate values for binding constants and rate constants are needed to predict the potential of hC4Nb8 for inhibition of the CP C3 convertase in vivo. We therefore conducted SPR experiments in which the nanobody was C-terminally AVI-tagged, site-specifically biotinylated, and immobilized on a sensor chip coated with amine-coupled streptavidin. Binding kinetics were studied by flowing 2-fold dilution series of the analytes C4 and C4b over the immobilized ligand hC4Nb8 surface. For C4b, the sensorgrams could be fitted well with a 1:1 binding model, giving a fitted value of KD = 15.5 ± 1.3 pM (Fig. 6A). In contrast, sensorgrams obtained with C4 could not be fitted with this model, and 40-times-higher concentrations of C4 were required to obtain a response in the same range as acquired for C4b, resulting in very high analyte concentrations, possibly giving rise to unspecific binding effects (Fig. 6B).

FIGURE 6.

hC4Nb8 binds with a picomolar dissociation constant to C4b. (A) C4b injected as analyte at concentrations 25, 12.5, 6.25, 3.13, and 1.56 nM over an SPR sensor with immobilized hC4Nb8. (B) Same as in (A) but with C4 as analyte injected at concentrations of 900, 450, 225, 112.5, and 56.25 nM, respectively. Signal is shown as dashed lines, fitting is shown in full lines. (C) SPR competition experiments, in which C4b preincubated with CR1 (dashed line) or C2 (dashed and dotted line) were injected onto the hC4Nb8-coated sensor. All the experiments were carried out in triplicates at 25°C. (D) BLI triplicate measurement of KD, in which the biotinylated hC4Nb8 was immobilized on a streptavidin sensor and mC4b in the binding phase was present in the 2-fold dilution series from 12.5 to 0.78 nM. Signal is shown as dashed lines; fitting is shown as full lines. (E) Pull-down assay of C4b by mutants of hC4Nb8 to rank the single mutants for a weaker interaction with C4b; reducing SDS-PAGE analysis of the pull-down eluates are shown.

FIGURE 6.

hC4Nb8 binds with a picomolar dissociation constant to C4b. (A) C4b injected as analyte at concentrations 25, 12.5, 6.25, 3.13, and 1.56 nM over an SPR sensor with immobilized hC4Nb8. (B) Same as in (A) but with C4 as analyte injected at concentrations of 900, 450, 225, 112.5, and 56.25 nM, respectively. Signal is shown as dashed lines, fitting is shown in full lines. (C) SPR competition experiments, in which C4b preincubated with CR1 (dashed line) or C2 (dashed and dotted line) were injected onto the hC4Nb8-coated sensor. All the experiments were carried out in triplicates at 25°C. (D) BLI triplicate measurement of KD, in which the biotinylated hC4Nb8 was immobilized on a streptavidin sensor and mC4b in the binding phase was present in the 2-fold dilution series from 12.5 to 0.78 nM. Signal is shown as dashed lines; fitting is shown as full lines. (E) Pull-down assay of C4b by mutants of hC4Nb8 to rank the single mutants for a weaker interaction with C4b; reducing SDS-PAGE analysis of the pull-down eluates are shown.

Close modal

We also used the SPR approach to confirm the competition of the nanobody with C2 suggested by both our SEC analysis and our structural studies (Figs. 2C, 4D) and competition with CR1 in the assisted FI cleavage assay (Fig. 2C, 2D). The extent of competition was quantified as a percentage of signal for C4b binding to immobilized hC4Nb8 in the presence of C2 or CR1 compared with the signal for an experiment in which only C4b was injected. The signal was almost fully recovered for the C4b:CR1 complex, whereas for C4b2 only 17% of the signal was recovered (Fig. 6C, dashed and dotted line), suggesting that C2 is a much stronger competitor than the CR1 CCP1-3 fragment used in this study. These results are in qualitative agreement with previously determined dissociation constants for complexes of C4b with C2 and the CR1 ectodomain, 72 nM (50) and 0.9 μM, respectively (55).

Sequence alignment of human, marmoset, and mouse C4 reveals that residues 791–794 in the epitope are the least conserved in mouse C4 (Fig. 4C, right). Possibly, the presence of a lysine in mouse C4 at the position corresponding to Gln793 that is near to Arg107 in the hC4Nb8 CDR3 (red in Fig. 4C) introduces electrostatic repulsion weakening the binding to mouse C4b. This could contribute to the less efficient inhibition of LP-driven C3b deposition observed in murine serum (Fig. 1F) compared with human serum. We therefore measured the binding affinity of the hC4Nb8 nanobody to murine C4b by BLI (Fig. 6D). Our data show that the nanobody binds to mouse C4b with a dissociation constant KD = 0.176 nM, which is comparable to the dissociation constant of 16 pM observed for hC4b by SPR. Overall, our SPR experiments documented that hC4Nb8 binds with picomolar affinity to hC4b and provided further evidence of hC4Nb8 competition with the physiological C4b binding partners C2 and CR1. The difference in the inhibitory efficacy of hC4Nb8 in human and murine serum appears not to stem from a significantly weaker interaction with murine C4b.

In light of the limitations of the SPR approach described above, we decided to take advantage of isothermal titration (ITC) experiments to determine the effect of mutations in hC4Nb8 and to clarify the strength of the interaction between hC4Nb8 and native C4. As ITC consumes substantial amounts of material, we first conducted pull-down experiments with very long incubation times to establish which hC4Nb8 single mutations had the strongest effect on the affinity for C4b. The WT hC4Nb8 nanobody was immobilized on streptavidin beads, whereas C4b, together with different hC4Nb8 mutants, was added in the supernatant. After prolonged incubation at 4°C, SDS-PAGE analysis of material eluted from the beads by boiling was conducted. After normalization by hC4Nb8 band intensity, the effect of the mutations on the hC4Nb8:C4b interaction could be ranked in the order W53A > R107A > E102A > E104A > A75R = R31A = WT (Fig. 6E). Based on this evidence, the double-mutant W53A, R107A was selected for further analysis with ITC. The structure of the C4b:hC4Nb8 complex suggests that the introduction of the W53A, R107A mutation in the nanobody will primarily affect the interaction of hC4Nb8 with the C4/C4b MG6 domain.

Before proceeding to the titration experiments, the efficacy of the double-mutant W53A, R107A was examined in a CP C3 deposition assay. This demonstrated that hC4Nb8 W53A, R107A needed to be present in an ∼4-fold higher concentration to inhibit the CP C3 convertase to the same degree as the WT nanobody (Fig. 7A). In the ITC experiments, the dissociation constant for the C4b:hC4Nb8 complex could not be determined, as the heat-evolved measure jumped from the nonbound to the bound state without intermediate points needed for the fitting and binding constant determination (Fig. 7B). This observation is in agreement with the 16 pM KD measured by SPR, which is outside of the limits that can be measured directly by ITC (56). Assuming that thermodynamic parameters determined by SPR and ITC can be compared, the binding constant for the C4b complex decreased 3000-fold for the double-mutant W53A, R107A, going from 15.5 pM to 48 nM (Fig. 7C). For both WT hC4bNb8 and the W53, R107A double-mutant, binding was highly exothermic, with enthalpies of −17 and −25 kcal∙mol−1, respectively. The titration of hC4Nb8 into native C4 resulted in a titration curve to which a dissociation constant for the hC4Nb8:C4 complex of 2 μM could be fitted (Fig. 7D), confirming the very low response for C4 binding in SPR (Fig. 6B). The double-mutation W53A, R107A completely abolished binding of hC4Nb8 to native C4 (Fig. 7E). The five order of magnitude difference in affinity for C4 compared with C4b provides further evidence for the relevance of our crystal structure, as the epitope of the nanobody encompasses the Nt-α′ region of C4b, which only becomes exposed after C4 activation, whereas it is tucked inside a cavity formed by the MG2 and MG3 domains in native C4 (11) (Fig. 4B). The 2-μM measured KD for binding to C4 is likewise in agreement with the absence of complex formation in our SEC experiment.

FIGURE 7.

Validation of the crystal structure and determination of the dissociation constant for the C4:hC4Nb8 complex. (A) Assay for deposition of C3 fragments in 0.2% NHS using hC4Nb8 WT and W53A, R107A mutant as inhibitor. The C4 concentration at the tested serum dilution is marked with a vertical dashed line. (B and C) ITC experiments to quantify the interaction of the ligands hC4Nb8 WT (B and D) and the W53A, R107A mutant (C and E) with the Ags C4b and (D and E) C4. Titrations were performed at 30°C in a 1.8-ml sample cell and 300-μl total titrant volume. All the experiments were carried out in duplicate.

FIGURE 7.

Validation of the crystal structure and determination of the dissociation constant for the C4:hC4Nb8 complex. (A) Assay for deposition of C3 fragments in 0.2% NHS using hC4Nb8 WT and W53A, R107A mutant as inhibitor. The C4 concentration at the tested serum dilution is marked with a vertical dashed line. (B and C) ITC experiments to quantify the interaction of the ligands hC4Nb8 WT (B and D) and the W53A, R107A mutant (C and E) with the Ags C4b and (D and E) C4. Titrations were performed at 30°C in a 1.8-ml sample cell and 300-μl total titrant volume. All the experiments were carried out in duplicate.

Close modal

We next addressed whether hC4Nb8 may be useful for the analysis of complement-driven immune response and pathogenesis. We genetically fused the nanobody to the Fc moiety of chicken Ig IgY, creating a bivalent C4-specific reagent for immunostaining of tissue sections from transgenic mice expressing hC4, whereas deleted for murine C4. The hC4Nb8–IgY was used for staining of a spleen section in parallel with a validated commercial anti-C4 polyclonal Ab. The staining revealed colocalization between the two imaging protocols, confirming the binding of hC4Nb8 to hC4 or hC4b in transgenic mouse tissue (Fig. 8A). The hC4Nb8 nanobody was also tested in hemolytic assays to detect its inhibitory effect in the transgenic hC4 mouse serum. Incubation of the serum with 10 μg/ml hC4Nb8 inhibited CP-driven hemolysis of EA cells (n = 3) compared with nontreated serum at all the dilutions tested (1:60, 1:240, 1:960) (Fig. 8B).

FIGURE 8.

hC4Nb8 binds to transgenic mice hC4 ex vivo and blocks the CP in vivo. (A) Immunostaining of spleen sections from a hC4 transgenic mouse. Red: anti-hC4 Ab; green: IgY-hC4Nb8. Scale bar, 40 μm. White arrows indicate examples of colocalization. (B) Hemolytic assay on Ab-coated erythrocytes in hC4 AB/minus mouse serum treated with 10 μg/ml hC4Nb8 (red) or untreated (black) (n = 3). (C) Analysis of all the images in which the FDC binary area colocalized with PE divided by the total FDC binary area was plotted, summary of two independent experiments (****p < 0.0001, ***p < 0.001).

FIGURE 8.

hC4Nb8 binds to transgenic mice hC4 ex vivo and blocks the CP in vivo. (A) Immunostaining of spleen sections from a hC4 transgenic mouse. Red: anti-hC4 Ab; green: IgY-hC4Nb8. Scale bar, 40 μm. White arrows indicate examples of colocalization. (B) Hemolytic assay on Ab-coated erythrocytes in hC4 AB/minus mouse serum treated with 10 μg/ml hC4Nb8 (red) or untreated (black) (n = 3). (C) Analysis of all the images in which the FDC binary area colocalized with PE divided by the total FDC binary area was plotted, summary of two independent experiments (****p < 0.0001, ***p < 0.001).

Close modal

After these preliminary validation experiments, we adopted a local immunization strategy to evaluate the function of hC4Nb8 in vivo. The transport of ICs to FDCs depends on activation of the CP (18). In our experiments, hC4-expressing or C4KO mice were passively immunized with anti-PE Ab injected i.p. After 24 h, PE was injected s.c. in the mice hock along with hC4Nb8 or the Lag16 control nanobody. Sections of the popliteal lymph nodes were stained with anti-B220 (CD45R) Ab in Far Red (647 nm), 7E9 Ab binding to CD21/CD35 on FDCs was stained in Pacific Blue (405 nm), and PE was stained in the red channel (568 nm). The colocalized deposition of red fluorescent PE with blue FDCs on popliteal lymph nodes was quantified by ImageJ after thresholding the images (57). The overall analysis of two independent experiments is presented in Fig. 8C. The hC4Nb8 nanobody inhibits PE colocalization on the FDCs compared with the Lag16 control nanobody (p < 0.0001), and colocalization is comparable to that observed for C4KO mice with a 0.1 ratio (Fig. 8C). In contrast, the calculated average ratio of colocalized area was 0.2 when the Lag16 control nanobody was administered. The total FDC area was not dependent on the treatment received by the mouse (Supplemental Table I).

To assess the ability of the nanobody to inhibit complement deposition in the context of the CNS, the nanobody was assayed in neuronal complement deposition assays [modified from (58)]. Human induced pluripotent stem cell–derived neurons sensitized to complement deposition were treated with NHS and live stained for C3c, and the proportion of neuronal processes (red in Fig. 9A) colocalized with C3 deposition (green in Fig. 9A) was quantified. In this experimental setup, there is a strong dose-dependent inhibitory effect of the hC4Nb8 nanobody on deposition of C3 onto neurites compared with the Lag16 control nanobody (p < 0.0001, Fig. 9B).

FIGURE 9.

Incubation of hC4Nb8 with 10% NHS inhibits C3 deposition on neurites in a dose-dependent manner. (A) Example image of C3c deposition (green) colocalizing with Tuj1+ neurites (red) in 10% NHS. (B) Analysis of all wells in the presence of hC4Nb8 or control (Lag16, Ctrl Nb) nanobody in 10% NHS (n = 24 wells per dose) or in C4-Dpl NHS (n = 6 wells per condition) shown with 95% confidence intervals. hC4Nb8 at 5, 10, and 20 μg/ml doses led to significantly less C3c deposition compared with control (Lag16) nanobody (p < 0.0005, p < 0.0001, p < 0.001). Under C4-Dpl conditions; both hC4Nb8 and control nanobody had significantly decreased deposition relative to the control condition in 10% NHS (p < 0.0001), further demonstrating that nanobody inhibition is specific to the CP and LP.

FIGURE 9.

Incubation of hC4Nb8 with 10% NHS inhibits C3 deposition on neurites in a dose-dependent manner. (A) Example image of C3c deposition (green) colocalizing with Tuj1+ neurites (red) in 10% NHS. (B) Analysis of all wells in the presence of hC4Nb8 or control (Lag16, Ctrl Nb) nanobody in 10% NHS (n = 24 wells per dose) or in C4-Dpl NHS (n = 6 wells per condition) shown with 95% confidence intervals. hC4Nb8 at 5, 10, and 20 μg/ml doses led to significantly less C3c deposition compared with control (Lag16) nanobody (p < 0.0005, p < 0.0001, p < 0.001). Under C4-Dpl conditions; both hC4Nb8 and control nanobody had significantly decreased deposition relative to the control condition in 10% NHS (p < 0.0001), further demonstrating that nanobody inhibition is specific to the CP and LP.

Close modal

Overall, our functional experiments with hC4Nb8 in different contexts demonstrate that the nanobody can be applied broadly in experiments in which complement inhibition is needed. In vivo, we have shown that hC4Nb8 inhibits the CP-dependent transport of ICs to FDCs in lymph nodes, suggesting that a functional effect could be observed for other effector functions of complement as well. Our demonstration of in vitro inhibition of complement on neuronal cell cultures also opens the possibility of using the nanobody in sophisticated cell-based assays modeling complement-mediated synaptic pruning.

In contrast to the steadily growing number of inhibitors targeting the AP and the terminal pathway developed with the aim of regulating complement activation as a therapeutic strategy (27, 59), the availability of CP- and LP-specific inhibitors is more limited. Except for a 56-kDa chimeric protein based on the endogenous LP regulator MAP-1 and C4BP that is an efficient regulator of C4b in vitro, there is no well-characterized man-made C4-specific inhibitor (60). In this study, we present, to our knowledge, the first structurally characterized inhibitor of the CP and LP C3/C5 convertases acting at the level of C4b. Our 14-kDa hC4Nb8 nanobody is highly specific for C4b and potently inhibits CP- and LP-driven C3 deposition in human serum as well as LP-driven C3 deposition in mouse serum. Our structural and functional studies show that the dominating mechanism by which hC4Nb8 inhibits the CP C3 convertase is by preventing assembly of the proconvertase. If the active C4b2a was formed, comparison with our prior structural model of the CP C3 convertase (34) suggests that the nanobody would most likely also compete with the catalytic C2a subunit and hence act as a decay acceleration factor. A possible disadvantage of hC4Nb8 is that our experiments with FI degradation of C4b suggest that the nanobody will prevent the endogenous regulators C4BP, CR1, and MCP from assisting FI with C4b conversion to C4d on host cells. This may be less of an issue, as FI regulation of C3b remains intact, and the FI cleavage-resistant C4b present is effectively prevented by the hC4Nb8 from forming CP C3 and C5 convertases.

As a powerful demonstration of the efficacy of the nanobody framework for the development of complement inhibitors, hC4Nb8 binds a neoepitope in C4b with a dissociation constant of 16 pM, using its three CDR regions. For comparison, the therapeutic mAb eculizumab, with six CDR regions and used for over a decade in the clinic for treatment of paroxysmal nocturnal haemoglobinuria and atypical haemolytic uremic syndrome, binds its Ag complement C5 with a dissociation constant of 18 pM (61). Nanobodies are 14–15-kDa proteins and therefore normally have short circulation times in vivo because of renal clearance (62). Nevertheless, we could show that hC4Nb8 is efficient in vivo in the passive immunization model, in which the nanobody was administrated together with the PE Ag (57). The main reason for this in vivo efficacy is likely to be the picomolar KD for the hC4Nb8 interaction with C4b deposited at the CP and LP activation site. In addition, although 1 × 105 times weaker than the C4b:hC4Nb8 interaction, complex formation with the abundant native C4 present at 2–3 μM in plasma potentially delays clearance compared with an unbound nanobody. The uptake of complement iC3b opsonized Ags examined in the passive immunization model is hijacked by HIV in the chronic stage of AIDS (63, 64), suggesting a therapeutic potential for hC4Nb8 administration in the context of retroviral infection.

The hC4Nb8 offers a small versatile protein module that may help to deduce the molecular mechanism behind complement-driven pathogenesis, in which evidence suggests activation through the CP or LP. For in vivo applications, hC4Nb8 could simulate C4 KO conditions for a short period of time, avoiding the development of auto-reactive B cells observed in C4KO mice (65). To improve its circulation time, PEGylation or fusion of the nanobody to an IgG Fc or an albumin binding molecule are all established means to increase circulation times to days (reviewed in Ref. 66). In addition, fusion of nanobodies with fragments of endogenous complement inhibitors or targeting molecules is easily achieved by genetic engineering and could further increase tissue specificity and complement inhibition, as demonstrated for the MAP-1:C4BP fusion protein (60). For future, more general animal studies not involving the hC4 knock-in mouse used in this study, the effect of hC4Nb8 administration should be carefully investigated. In vivo, the protease MASP-2 can cleave C3 in a C4-bypass pathway (67). Because the affinity of hC4Nb8 for mC4b is still in the subnanomolar range, the weaker inhibition observed in C57BL/6 serum could arise from this unconventional activation pathway or from stronger competition by mouse C2 compared with human C2. The C57BL/6 serum does not contain the C4 homolog sex-limited protein (Slp) which has hemolytic activity (6870); however, in other mouse strains in which Slp is expressed, the nanobody may show even less efficiency because in Slp an arginine substitutes for a tryptophan within the hC4Nb8 epitope (Fig. 4C).

The CP is known to contribute to pathogenesis in ischemia–reperfusion injury, sepsis, autoimmune hemolytic anemia, glomerulonephritis, Ab-mediated rejection, Alzheimer disease, SCZ (2730), multiple cancer models (71), glaucoma (72), and cold agglutinin disease (31). The LP is known to contribute to pathogenesis in rheumatic heart disease (73), inflammatory arthritis (74), viral infection (75), infection by protozoan microbes (76), pneumococcal infection (77), and renal disease (78), and very recent findings implicate uncontrolled activation of the LP and propagation into the alternative and terminal pathways in patients infected with COVID-19 (reviewed in Ref. 79).

There are currently three mAbs under preclinical development or in clinical trials that, like hC4Nb8, prevent CP and LP C3 convertase activity: the C2-specific mAb PRO-02 (Prothix) inhibiting the C3 convertase, whereas the MASP-2–specific OMS721 (narsoplimab) and the C1s-specific BIVV009 (sutimlimab) (59) interfere with both C4 and C2 cleavage. Several C1q inhibitors (ANX005, ANX007, ANX009) are also in the pipeline for the treatment of neurologic disease (Annexon Biosciences). Systemic and permanent therapeutic regulation of the CP C3 convertase could be problematic because C1q and C4 deficiencies predispose for systemic lupus erythematosus, and both the CP and LP are important for clearance of pathogens. The drawbacks of C4b inhibition compared with inhibition of the upstream C1 complex may be smaller, as C4-independent functions of C1q (80) are preserved. Direct C3 cleavage by MASP-2 activated through the LP can also still occur (81); the AP will still be able to initiate through C3 tickover (82), and the activity of the AP convertases will not be affected.

Of special interest for future studies resolving the role of complement in neurodegenerative disease, we demonstrated in this study in a cell culture model, that hC4Nb8 can prevent C3b deposition from human serum in a dose-dependent manner on neurites to a level close to that observed with C4-deficient serum, in which residual C3b deposition is likely to occur through low level AP activity. Recent genome-wide association studies linked SCZ to a higher copy number and brain mRNA levels of the C4A isotype (25), as well as to a single nucleotide polymorphism in the CNS-specific complement regulator CSMD1 (83), which possibly performs a function analogous to that of CR1 in the periphery (58, 84). In the proposed disease mechanism, either C4A overexpression and/or reduced levels of CSMD1 would lead to increased synaptic pruning during development, thereby producing incorrectly refined synaptic circuits and disease onset in early adulthood. The molecular understanding of SCZ is only starting, and therapeutic intervention, at present, focuses on alleviating symptoms (30). Complement-mediated pruning defects are also implicated in epilepsy (85), and several studies have underlined the role of C4 in the pathogenesis of other CNS diseases. Mice deficient in C4 and not C1q or C3 have shown improved recovery after traumatic brain injury (86), and elevated C4d levels were measured in amyotrophic lateral sclerosis patients’ cerebrospinal fluid (87). It is long known that increased levels of the components C1q, C3, and C4 associate with Alzheimer disease (88, 89). Recent evidence implies the CP in spinal motor circuit refinement during development and in the pathogenesis of spinal muscular atrophy (90).

If the hC4Nb8 can be administrated within the CNS without breaking the blood–brain barrier, therapeutic blockade of the CP C3 convertase may be a promising strategy for treating SCZ and other neurodegenerative diseases. In addition to the perspective of developing hC4Nb8 into a therapeutic agent, it also bears substantial potential for in vitro research involving complement. Because of its specificity and very tight binding, it could be applied for diagnostic purposes to quantitate C4b deposition. In vitro, hC4Nb8 would be a high-affinity C4b and C4c binder in ELISAs and flow cytometry experiments, and introduction of, for example, a hemagglutinin tag could allow direct detection with secondary Abs in immunostaining (91). For in vivo localization, a radioactive tag could be added and followed by positron emission tomography imaging (92). The presence of divalent cations in cell culture assays is necessary for long-term cell viability, replacing EGTA and EDTA with the hC4Nb8 nanobody to specifically block the CP and LP C3 convertase would therefore eliminate cytotoxic effects. In addition, in synaptic pruning models and other cellular models involving Mg2+ and Ca2+-dependent binding of iC3b to CR3, phagocytosis can be evaluated whereas inhibiting the CP C3 convertase with hC4Nb8.

In conclusion, the hC4Nb8 nanobody is a C4-specific complement-blocking reagent with potential for very broad applications. The linkage of C4 with several neurologic diseases leads to a rise in interest toward CP therapeutic inhibition targeting the CNS. Our study provides evidence that hC4Nb8 holds promise as a research tool for the study of the molecular mechanisms of disease pathogenesis and as a potential therapeutic inhibitor. Further investigations are needed to understand whether administration of hC4Nb8 could ameliorate CP- and LP-mediated disease pathogenesis in animal models. The cross-reactivity of hC4Nb8 with marmoset C4 opens the possibility of applying the reagent in the study of CNS disease with nonhuman primate models.

We acknowledge Christine Schar for excellent assistance with SPR and ITC experiments, the beamline staff at the European Molecular Biology Laboratory (Hamburg, Germany) and Karen Margrethe Nielsen for technical support, and Neal Lojek for the differentiation of neuronal cell cultures. The authors are grateful for the pC427A-K1324N mC4 expression plasmid, a kind gift from David Isenman, Toronto University.

This work was supported by Lundbeck Foundation BRAINSTRUC Grant R155-2015-2666 and the Graduate School of Science and Technology, Aarhus University.

The online version of this article contains supplemental material.

Abbreviations used in this article:

AP

alternative pathway

BLI

bio-layer interferometry

C4BP

C4b binding protein

C4-Dpl

C4-depleted serum

CjC4

Callithrix jacchus C4

CP

classical pathway

CR

complement receptor

CVFB

cobra venom factor-FB

2D

two-dimensional

3D

three-dimensional

EM

electron microscopy

FDC

follicular dendritic cell

FH

factor H

FI

factor I

FT

flow through

hC4

human C4

IC

immune complex

ITC

isothermal titration

KO

knockout

LB

Luria broth

LP

lectin pathway

NHS

normal human serum

PBS-T

PBS containing 0.1% Tween 20

PDB ID

Protein Data Bank identification number

PRM

pattern recognition molecule

RT

room temperature

RU

response unit

SCZ

schizophrenia

SEC

size-exclusion chromatography

Slp

sex-limited protein

SP

serine protease

SPR

surface plasmon resonance

WT

wild-type.

1
Dodds
,
A. W.
,
M.
Matsushita
.
2007
.
The phylogeny of the complement system and the origins of the classical pathway.
Immunobiology
212
:
233
243
.
2
Bajic
,
G.
,
S. E.
Degn
,
S.
Thiel
,
G. R.
Andersen
.
2015
.
Complement activation, regulation, and molecular basis for complement-related diseases.
EMBO J.
34
:
2735
2757
.
3
van Lookeren Campagne
,
M.
,
C.
Wiesmann
,
E. J.
Brown
.
2007
.
Macrophage complement receptors and pathogen clearance.
Cell. Microbiol.
9
:
2095
2102
.
4
Doorduijn
,
D. J.
,
S. H. M.
Rooijakkers
,
D. A. C.
Heesterbeek
.
2019
.
How the membrane attack complex damages the bacterial cell envelope and kills gram-negative bacteria.
BioEssays
41
: e1900074.
5
Pangburn
,
M. K.
,
R. D.
Schreiber
,
H. J.
Müller-Eberhard
.
1981
.
Formation of the initial C3 convertase of the alternative complement pathway. Acquisition of C3b-like activities by spontaneous hydrolysis of the putative thioester in native C3.
J. Exp. Med.
154
:
856
867
.
6
Lachmann
,
P. J.
2009
.
The amplification loop of the complement pathways.
In
Advances in Immunology.
F. W.
Alt
,
K. F.
Austen
,
T.
Honjo
,
F.
Melchers
,
J. W.
Uhr
,
E. R.
Unanue
, eds.
Academic Press
,
Waltham, MA
, p.
115
149
.
7
Kjaer
,
T. R.
,
S.
Thiel
,
G. R.
Andersen
.
2013
.
Toward a structure-based comprehension of the lectin pathway of complement.
Mol. Immunol.
56
:
413
422
.
8
Degn
,
S. E.
,
T. R.
Kjaer
,
R. T.
Kidmose
,
L.
Jensen
,
A. G.
Hansen
,
M.
Tekin
,
J. C.
Jensenius
,
G. R.
Andersen
,
S.
Thiel
.
2014
.
Complement activation by ligand-driven juxtaposition of discrete pattern recognition complexes.
Proc. Natl. Acad. Sci. USA
111
:
13445
13450
.
9
Kjaer
,
T. R.
,
L. T. M.
Le
,
J. S.
Pedersen
,
B.
Sander
,
M. M.
Golas
,
J. C.
Jensenius
,
G. R.
Andersen
,
S.
Thiel
.
2015
.
Structural insights into the initiating complex of the lectin pathway of complement activation.
Structure
23
:
342
351
.
10
Kojouharova
,
M.
,
K.
Reid
,
M.
Gadjeva
.
2010
.
New insights into the molecular mechanisms of classical complement activation.
Mol. Immunol.
47
:
2154
2160
.
11
Kidmose
,
R. T.
,
N. S.
Laursen
,
J.
Dobó
,
T. R.
Kjaer
,
S.
Sirotkina
,
L.
Yatime
,
L.
Sottrup-Jensen
,
S.
Thiel
,
P.
Gál
,
G. R.
Andersen
.
2012
.
Structural basis for activation of the complement system by component C4 cleavage.
Proc. Natl. Acad. Sci. USA
109
:
15425
15430
.
12
Sitomer
,
G.
,
R. M.
Stroud
,
M. M.
Mayer
.
1966
.
Reversible adsorption of C’2 by EAC’4: role of Mg2+, enumeration of competent SAC’4, two-step nature of C’2a fixation and estimation of its efficiency.
Immunochemistry
3
:
57
69
.
13
Milder
,
F. J.
,
H. C. A.
Raaijmakers
,
M. D. A. A.
Vandeputte
,
A.
Schouten
,
E. G.
Huizinga
,
R. A.
Romijn
,
W.
Hemrika
,
A.
Roos
,
M. R.
Daha
,
P.
Gros
.
2006
.
Structure of complement component C2A: implications for convertase formation and substrate binding.
Structure
14
:
1587
1597
.
14
Kerr
,
M. A.
1980
.
The human complement system: assembly of the classical pathway C3 convertase.
Biochem. J.
189
:
173
181
.
15
Law
,
S. K.
,
A. W.
Dodds
.
1997
.
The internal thioester and the covalent binding properties of the complement proteins C3 and C4.
Protein Sci.
6
:
263
274
.
16
Bajic
,
G.
,
L.
Yatime
,
R. B.
Sim
,
T.
Vorup-Jensen
,
G. R.
Andersen
.
2013
.
Structural insight on the recognition of surface-bound opsonins by the integrin I domain of complement receptor 3.
Proc. Natl. Acad. Sci. USA
110
:
16426
16431
.
17
Carroll
,
M. C.
,
D. E.
Isenman
.
2012
.
Regulation of humoral immunity by complement.
Immunity
37
:
199
207
.
18
Gonzalez
,
S. F.
,
S. E.
Degn
,
L. A.
Pitcher
,
M.
Woodruff
,
B. A.
Heesters
,
M. C.
Carroll
.
2011
.
Trafficking of B cell antigen in lymph nodes.
Annu. Rev. Immunol.
29
:
215
233
.
19
Krych-Goldberg
,
M.
,
J. P.
Atkinson
.
2001
.
Structure-function relationships of complement receptor type 1.
Immunol. Rev.
180
:
112
122
.
20
Galvan
,
M. D.
,
M. C.
Greenlee-Wacker
,
S. S.
Bohlson
.
2012
.
C1q and phagocytosis: the perfect complement to a good meal.
J. Leukoc. Biol.
92
:
489
497
.
21
Presumey
,
J.
,
A. R.
Bialas
,
M. C.
Carroll
.
2017
.
Complement system in neural synapse elimination in development and disease.
In
Advances in Immunology.
F. W.
Alt
, ed.
Academic Press
,
Cambridge, MA
, p.
53
79
.
22
Veerhuis
,
R.
,
H. M.
Nielsen
,
A. J.
Tenner
.
2011
.
Complement in the brain.
Mol. Immunol.
48
:
1592
1603
.
23
Johnson
,
M. B.
,
B.
Stevens
.
2018
.
Pruning hypothesis comes of age.
Nature
554
:
438
439
.
24
Feinberg
,
I.
1982–1983
.
Schizophrenia: caused by a fault in programmed synaptic elimination during adolescence?
J. Psychiatr. Res.
17
:
319
334
.
25
Schizophrenia Working Group of the Psychiatric Genomics Consortium
.
2016
.
Schizophrenia risk from complex variation of complement component 4.
Nature
530
:
177
183
.
26
Stevens
,
B.
,
N. J.
Allen
,
L. E.
Vazquez
,
G. R.
Howell
,
K. S.
Christopherson
,
N.
Nouri
,
K. D.
Micheva
,
A. K.
Mehalow
,
A. D.
Huberman
,
B.
Stafford
, et al
.
2007
.
The classical complement cascade mediates CNS synapse elimination.
Cell
131
:
1164
1178
.
27
Ricklin
,
D.
,
J. D.
Lambris
.
2013
.
Complement in immune and inflammatory disorders: therapeutic interventions.
J. Immunol.
190
:
3839
3847
.
28
Wouters
,
D.
,
S.
Zeerleder
.
2015
.
Complement inhibitors to treat IgM-mediated autoimmune hemolysis.
Haematologica
100
:
1388
1395
.
29
Lintner
,
K. E.
,
Y. L.
Wu
,
Y.
Yang
,
C. H.
Spencer
,
G.
Hauptmann
,
L. A.
Hebert
,
J. P.
Atkinson
,
C. Y.
Yu
.
2016
.
Early components of the complement classical activation pathway in human systemic autoimmune diseases.
Front. Immunol.
7
:
36
.
30
Woo
,
J. J.
,
J. G.
Pouget
,
C. C.
Zai
,
J. L.
Kennedy
.
2020
.
The complement system in schizophrenia: where are we now and what’s next?
Mol. Psychiatry
25
:
114
130
.
31
Jäger
,
U.
,
S.
D’Sa
,
C.
Schörgenhofer
,
J.
Bartko
,
U.
Derhaschnig
,
C.
Sillaber
,
P.
Jilma-Stohlawetz
,
M.
Fillitz
,
T.
Schenk
,
G.
Patou
, et al
.
2019
.
Inhibition of complement C1s improves severe hemolytic anemia in cold agglutinin disease: a first-in-human trial.
Blood
133
:
893
901
.
32
Zelek
,
W. M.
,
L.
Xie
,
B. P.
Morgan
,
C. L.
Harris
.
2019
.
Compendium of current complement therapeutics.
Mol. Immunol.
114
:
341
352
.
33
Muyldermans
,
S.
2013
.
Nanobodies: natural single-domain antibodies.
Annu. Rev. Biochem.
82
:
775
797
.
34
Mortensen
,
S.
,
J. K.
Jensen
,
G. R.
Andersen
.
2016
.
Solution structures of complement C2 and its C4 complexes propose pathway-specific mechanisms for control and activation of the complement proconvertases.
J. Biol. Chem.
291
:
16494
16507
.
35
Mortensen
,
S.
,
R. T.
Kidmose
,
S. V.
Petersen
,
Á.
Szilágyi
,
Z.
Prohászka
,
G. R.
Andersen
.
2015
.
Structural basis for the function of complement component C4 within the classical and lectin pathways of complement.
J. Immunol.
194
:
5488
5496
.
36
Dodd
,
I.
,
D. E.
Mossakowska
,
P.
Camilleri
,
M.
Haran
,
P.
Hensley
,
E. J.
Lawlor
,
D. L.
McBay
,
W.
Pindar
,
R. A. G.
Smith
.
1995
.
Overexpression in Escherichia coli, folding, purification, and characterization of the first three short consensus repeat modules of human complement receptor type 1.
Protein Expr. Purif.
6
:
727
736
.
37
Jensen
,
R. K.
,
R.
Pihl
,
T. A. F.
Gadeberg
,
J. K.
Jensen
,
K. R.
Andersen
,
S.
Thiel
,
N. S.
Laursen
,
G. R.
Andersen
.
2018
.
A potent complement factor C3-specific nanobody inhibiting multiple functions in the alternative pathway of human and murine complement.
J. Biol. Chem.
293
:
6269
6281
.
38
Andersen
,
K. R.
,
N. C.
Leksa
,
T. U.
Schwartz
.
2013
.
Optimized E. coli expression strain LOBSTR eliminates common contaminants from His-tag purification.
Proteins
81
:
1857
1861
.
39
Dougan
,
M.
,
J. R.
Ingram
,
H.-J.
Jeong
,
M. M.
Mosaheb
,
P. T.
Bruck
,
L.
Ali
,
N.
Pishesha
,
O.
Blomberg
,
P. M.
Tyler
,
M. M.
Servos
, et al
.
2018
.
Targeting cytokine therapy to the pancreatic tumor microenvironment using PD-L1-specific VHHs.
Cancer Immunol. Res.
6
:
389
401
.
40
Kabsch
,
W.
2010
.
XDS.
Acta Crystallogr. D Biol. Crystallogr.
66
:
125
132
.
41
McCoy
,
A. J.
,
R. W.
Grosse-Kunstleve
,
P. D.
Adams
,
M. D.
Winn
,
L. C.
Storoni
,
R. J.
Read
.
2007
.
Phaser crystallographic software.
J. Appl. Cryst.
40
:
658
674
.
42
Emsley
,
P.
,
K.
Cowtan
.
2004
.
Coot: model-building tools for molecular graphics.
Acta Crystallogr. D Biol. Crystallogr.
60
:
2126
2132
.
43
Terwilliger
,
T. C.
2000
.
Maximum-likelihood density modification.
Acta Crystallogr. D Biol. Crystallogr.
56
:
965
972
.
44
Afonine
,
P. V.
,
R. W.
Grosse-Kunstleve
,
N.
Echols
,
J. J.
Headd
,
N. W.
Moriarty
,
M.
Mustyakimov
,
T. C.
Terwilliger
,
A.
Urzhumtsev
,
P. H.
Zwart
,
P. D.
Adams
.
2012
.
Towards automated crystallographic structure refinement with phenix.refine.
Acta Crystallogr. D Biol. Crystallogr.
68
:
352
367
.
45
Grant
,
T.
,
A.
Rohou
,
N.
Grigorieff
.
2018
.
cis TEM, user-friendly software for single-particle image processing.
eLife
7
: e35383.
46
Scheres
,
S. H. W.
2012
.
RELION: implementation of a Bayesian approach to cryo-EM structure determination.
J. Struct. Biol.
180
:
519
530
.
47
Pettersen
,
E. F.
,
T. D.
Goddard
,
C. C.
Huang
,
G. S.
Couch
,
D. M.
Greenblatt
,
E. C.
Meng
,
T. E.
Ferrin
.
2004
.
UCSF chimera--a visualization system for exploratory research and analysis.
J. Comput. Chem.
25
:
1605
1612
.
48
Xue
,
X.
,
J.
Wu
,
D.
Ricklin
,
F.
Forneris
,
P.
Di Crescenzio
,
C. Q.
Schmidt
,
J.
Granneman
,
T. H.
Sharp
,
J. D.
Lambris
,
P.
Gros
.
2017
.
Regulator-dependent mechanisms of C3b processing by factor I allow differentiation of immune responses.
Nat. Struct. Mol. Biol.
24
:
643
651
.
49
Croll
,
T. I.
,
G. R.
Andersen
.
2016
.
Re-evaluation of low-resolution crystal structures via interactive molecular-dynamics flexible fitting (iMDFF): a case study in complement C4.
Acta Crystallogr. D Struct. Biol.
72
:
1006
1016
.
50
Laich
,
A.
,
R. B.
Sim
.
2001
.
Complement C4bC2 complex formation: an investigation by surface plasmon resonance.
Biochim. Biophys. Acta
1544
:
96
112
.
51
Pan
,
Q.
,
R. O.
Ebanks
,
D. E.
Isenman
.
2000
.
Two clusters of acidic amino acids near the NH2 terminus of complement component C4 alpha′-chain are important for C2 binding.
J. Immunol.
165
:
2518
2527
.
52
Janssen
,
B. J. C.
,
L.
Gomes
,
R. I.
Koning
,
D. I.
Svergun
,
A. J.
Koster
,
D. C.
Fritzinger
,
C.-W.
Vogel
,
P.
Gros
.
2009
.
Insights into complement convertase formation based on the structure of the factor B-cobra venom factor complex.
EMBO J.
28
:
2469
2478
.
53
Forneris
,
F.
,
D.
Ricklin
,
J.
Wu
,
A.
Tzekou
,
R. S.
Wallace
,
J. D.
Lambris
,
P.
Gros
.
2010
.
Structures of C3b in complex with factors B and D give insight into complement convertase formation.
Science
330
:
1816
1820
.
54
Wallis
,
R.
,
A. W.
Dodds
,
D. A.
Mitchell
,
R. B.
Sim
,
K. B. M.
Reid
,
W. J.
Schwaeble
.
2007
.
Molecular interactions between MASP-2, C4, and C2 and their activation fragments leading to complement activation via the lectin pathway.
J. Biol. Chem.
282
:
7844
7851
.
55
Clemenza
,
L.
,
D. E.
Isenman
.
2004
.
The C4A and C4B isotypic forms of human complement fragment C4b have the same intrinsic affinity for complement receptor 1 (CR1/CD35).
J. Immunol.
172
:
1670
1680
.
56
Freyer
,
M. W.
,
E. A.
Lewis
.
2008
.
Isothermal titration calorimetry: experimental design, data analysis, and probing macromolecule/ligand binding and kinetic interactions.
Methods Cell Biol.
84
:
79
113
.
57
Phan
,
T. G.
,
I.
Grigorova
,
T.
Okada
,
J. G.
Cyster
.
2007
.
Subcapsular encounter and complement-dependent transport of immune complexes by lymph node B cells.
Nat. Immunol.
8
:
992
1000
.
58
Escudero-Esparza
,
A.
,
N.
Kalchishkova
,
E.
Kurbasic
,
W. G.
Jiang
,
A. M.
Blom
.
2013
.
The novel complement inhibitor human CUB and Sushi multiple domains 1 (CSMD1) protein promotes factor I-mediated degradation of C4b and C3b and inhibits the membrane attack complex assembly.
FASEB J.
27
:
5083
5093
.
59
Ricklin
,
D.
,
D. C.
Mastellos
,
E. S.
Reis
,
J. D.
Lambris
.
2018
.
The renaissance of complement therapeutics.
Nat. Rev. Nephrol.
14
:
26
47
.
60
Hertz
,
C. E.
,
R.
Bayarri-Olmos
,
N.
Kirketerp-Møller
,
S.
van Putten
,
K.
Pilely
,
M. O.
Skjoedt
,
P.
Garred
.
2018
.
Chimeric proteins containing MAP-1 and functional domains of C4b-binding protein reveal strong complement inhibitory capacities.
Front. Immunol.
9
:
1945
.
61
Schatz-Jakobsen
,
J. A.
,
Y.
Zhang
,
K.
Johnson
,
A.
Neill
,
D.
Sheridan
,
G. R.
Andersen
.
2016
.
Structural basis for eculizumab-mediated inhibition of the complement terminal pathway.
J. Immunol.
197
:
337
344
.
62
Debie
,
P.
,
N.
Devoogdt
,
S.
Hernot
.
2019
.
Targeted nanobody-based molecular tracers for nuclear imaging and image-guided surgery.
Antibodies (Basel)
8
:
12
.
63
Kacani
,
L.
,
W. M.
Prodinger
,
G. M.
Sprinzl
,
M. G.
Schwendinger
,
M.
Spruth
,
H.
Stoiber
,
S.
Döpper
,
S.
Steinhuber
,
F.
Steindl
,
M. P.
Dierich
.
2000
.
Detachment of human immunodeficiency virus type 1 from germinal centers by blocking complement receptor type 2.
J. Virol.
74
:
7997
8002
.
64
Huber
,
G.
,
Z.
Bánki
,
S.
Lengauer
,
H.
Stoiber
.
2011
.
Emerging role for complement in HIV infection.
Curr. Opin. HIV AIDS
6
:
419
426
.
65
Chatterjee
,
P.
,
A. F.
Agyemang
,
M. B.
Alimzhanov
,
S.
Degn
,
S. A.
Tsiftsoglou
,
E.
Alicot
,
S. A.
Jones
,
M.
Ma
,
M. C.
Carroll
.
2013
.
Complement C4 maintains peripheral B-cell tolerance in a myeloid cell dependent manner.
Eur. J. Immunol.
43
:
2441
2450
.
66
Steeland
,
S.
,
R. E.
Vandenbroucke
,
C.
Libert
.
2016
.
Nanobodies as therapeutics: big opportunities for small antibodies.
Drug Discov. Today
21
:
1076
1113
.
67
Asgari
,
E.
,
C. A.
Farrar
,
N.
Lynch
,
Y. M.
Ali
,
S.
Roscher
,
C.
Stover
,
W.
Zhou
,
W. J.
Schwaeble
,
S. H.
Sacks
.
2014
.
Mannan-binding lectin-associated serine protease 2 is critical for the development of renal ischemia reperfusion injury and mediates tissue injury in the absence of complement C4.
FASEB J.
28
:
3996
4003
.
68
Beurskens
,
F. J.
,
M.
Tosi
,
T.
Shiroisho
,
D. M.
Robins
,
M. C.
Carroll
,
H.
van Dijk
.
2000
.
Mouse complement components C4 and Slp act synergistically in a homologous hemolytic C4 assay.
Eur. J. Immunol.
30
:
1507
1511
.
69
Beurskens
,
F. J.
,
J. D.
Kuenen
,
F.
Hofhuis
,
A. C.
Fluit
,
D. M.
Robins
,
H.
Van Dijk
.
1999
.
Sex-limited protein: in vitro and in vivo functions.
Clin. Exp. Immunol.
116
:
395
400
.
70
Stavenhagen
,
J.
,
F.
Loreni
,
C.
Hemenway
,
M.
Kalff
,
D. M.
Robins
.
1987
.
Molecular genetics of androgen-dependent and -independent expression of mouse sex-limited protein.
Mol. Cell. Biol.
7
:
1716
1724
.
71
Roumenina
,
L. T.
,
M. V.
Daugan
,
F.
Petitprez
,
C.
Sautès-Fridman
,
W. H.
Fridman
.
2019
.
Context-dependent roles of complement in cancer.
Nat. Rev. Cancer
19
:
698
715
.
72
Williams
,
P. A.
,
J. R.
Tribble
,
K. W.
Pepper
,
S. D.
Cross
,
B. P.
Morgan
,
J. E.
Morgan
,
S. W. M.
John
,
G. R.
Howell
.
2016
.
Inhibition of the classical pathway of the complement cascade prevents early dendritic and synaptic degeneration in glaucoma.
Mol. Neurodegener.
11
:
26
.
73
Beltrame
,
M. H.
,
S. J.
Catarino
,
I.
Goeldner
,
A. B. W.
Boldt
,
I. J.
de Messias-Reason
.
2015
.
The lectin pathway of complement and rheumatic heart disease.
Front. Pediatr.
2
:
148
.
74
Banda
,
N. K.
,
S.
Acharya
,
R. I.
Scheinman
,
G.
Mehta
,
M.
Takahashi
,
Y.
Endo
,
W.
Zhou
,
C. A.
Farrar
,
S. H.
Sacks
,
T.
Fujita
, et al
.
2017
.
Deconstructing the lectin pathway in the pathogenesis of experimental inflammatory arthritis: essential role of the lectin ficolin B and mannose-binding protein-associated serine protease 2.
J. Immunol.
199
:
1835
1845
.
75
Mason
,
C. P.
,
A. W.
Tarr
.
2015
.
Human lectins and their roles in viral infections.
Molecules
20
:
2229
2271
.
76
Evans-Osses
,
I.
,
I.
de Messias-Reason
,
M. I.
Ramirez
.
2013
.
The emerging role of complement lectin pathway in trypanosomatids: molecular bases in activation, genetic deficiencies, susceptibility to infection, and complement system-based therapeutics.
ScientificWorldJournal
2013
: 675898.
77
Ali
,
Y. M.
,
N. J.
Lynch
,
K. S.
Haleem
,
T.
Fujita
,
Y.
Endo
,
S.
Hansen
,
U.
Holmskov
,
K.
Takahashi
,
G. L.
Stahl
,
T.
Dudler
, et al
.
2012
.
The lectin pathway of complement activation is a critical component of the innate immune response to pneumococcal infection.
PLoS Pathog.
8
: e1002793.
78
Gaya da Costa
,
M.
,
F.
Poppelaars
,
S. P.
Berger
,
M. R.
Daha
,
M. A.
Seelen
.
2018
.
The lectin pathway in renal disease: old concept and new insights.
Nephrol. Dial. Transplant.
33
:
2073
2079
.
79
Risitano
,
A. M.
,
D. C.
Mastellos
,
M.
Huber-Lang
,
D.
Yancopoulou
,
C.
Garlanda
,
F.
Ciceri
,
J. D.
Lambris
.
2020
.
Complement as a target in COVID-19? [Published erratum appears in 2020 Nat. Rev. Immunol. 20: 448.]
Nat. Rev. Immunol.
20
:
343
344
.
80
Lu
,
J.
,
U.
Kishore
.
2017
.
C1 complex: an adaptable proteolytic module for complement and non-complement functions.
Front. Immunol.
8
:
592
.
81
Yaseen
,
S.
,
G.
Demopulos
,
T.
Dudler
,
M.
Yabuki
,
C. L.
Wood
,
W. J.
Cummings
,
L. W.
Tjoelker
,
T.
Fujita
,
S.
Sacks
,
P.
Garred
, et al
.
2017
.
Lectin pathway effector enzyme mannan-binding lectin-associated serine protease-2 can activate native complement C3 in absence of C4 and/or C2.
FASEB J.
31
:
2210
2219
.
82
Lachmann
,
P. J.
,
E.
Lay
,
D. J.
Seilly
.
2018
.
Experimental confirmation of the C3 tickover hypothesis by studies with an Ab (S77) that inhibits tickover in whole serum.
FASEB J.
32
:
123
129
.
83
Schizophrenia Working Group of the Psychiatric Genomics Consortium
.
2014
.
Biological insights from 108 schizophrenia-associated genetic loci.
Nature
511
:
421
427
.
84
Kraus
,
D. M.
,
G. S.
Elliott
,
H.
Chute
,
T.
Horan
,
K. H.
Pfenninger
,
S. D.
Sanford
,
S.
Foster
,
S.
Scully
,
A. A.
Welcher
,
V. M.
Holers
.
2006
.
CSMD1 is a novel multiple domain complement-regulatory protein highly expressed in the central nervous system and epithelial tissues.
J. Immunol.
176
:
4419
4430
.
85
Chu
,
Y.
,
X.
Jin
,
I.
Parada
,
A.
Pesic
,
B.
Stevens
,
B.
Barres
,
D. A.
Prince
.
2010
.
Enhanced synaptic connectivity and epilepsy in C1q knockout mice.
Proc. Natl. Acad. Sci. USA
107
:
7975
7980
.
86
You
,
Z.
,
J.
Yang
,
K.
Takahashi
,
P. H.
Yager
,
H.-H.
Kim
,
T.
Qin
,
G. L.
Stahl
,
R. A. B.
Ezekowitz
,
M. C.
Carroll
,
M. J.
Whalen
.
2007
.
Reduced tissue damage and improved recovery of motor function after traumatic brain injury in mice deficient in complement component C4.
J. Cereb. Blood Flow Metab.
27
:
1954
1964
.
87
Tsuboi
,
Y.
,
T.
Yamada
.
1994
.
Increased concentration of C4d complement protein in CSF in amyotrophic lateral sclerosis.
J. Neurol. Neurosurg. Psychiatry
57
:
859
861
.
88
Walker
,
D. G.
,
P. L.
McGeer
.
1992
.
Complement gene expression in human brain: comparison between normal and Alzheimer disease cases.
Brain Res. Mol. Brain Res.
14
:
109
116
.
89
Morgan
,
B. P.
2018
.
Complement in the pathogenesis of Alzheimer’s disease. [Published erratum appears in 2018 Semin. Immunopathol. 40: 517.]
Semin. Immunopathol.
40
:
113
124
.
90
Vukojicic
,
A.
,
N.
Delestrée
,
E. V.
Fletcher
,
J. G.
Pagiazitis
,
S.
Sankaranarayanan
,
T. A.
Yednock
,
B. A.
Barres
,
G. Z.
Mentis
.
2019
.
The classical complement pathway mediates microglia-dependent remodeling of spinal motor circuits during development and in SMA.
Cell Rep.
29
:
3087
3100.e7
.
91
Dong
,
J.-X.
,
Y.
Lee
,
M.
Kirmiz
,
S.
Palacio
,
C.
Dumitras
,
C. M.
Moreno
,
R.
Sando
,
L. F.
Santana
,
T. C.
Südhof
,
B.
Gong
, et al
.
2019
.
A toolbox of nanobodies developed and validated for use as intrabodies and nanoscale immunolabels in mammalian brain neurons.
eLife
8
: e48750.
92
Rashidian
,
M.
,
M. W.
LaFleur
,
V. L.
Verschoor
,
A.
Dongre
,
Y.
Zhang
,
T. H.
Nguyen
,
S.
Kolifrath
,
A. R.
Aref
,
C. J.
Lau
,
C. P.
Paweletz
, et al
.
2019
.
Immuno-PET identifies the myeloid compartment as a key contributor to the outcome of the antitumor response under PD-1 blockade.
Proc. Natl. Acad. Sci. USA
116
:
16971
16980
.

A.Z., S.T., N.S.L., and G.R.A. are listed as inventors on a patent describing the hC4Nb8 nanobody (patent number WO2019238674A1). The other authors have no financial conflicts of interest.

Supplementary data