Neuropilin-1 Acts as a Receptor for Complement Split Products

Abstract

Complement split products (CSPs), such as the fragments C4d and C3d, which are generated as a consequence of complement regulatory processes, are established markers for disease activity in autoimmunity or antibody-mediated graft rejection. Since immunoglobulin-like transcript 4 (ILT4) was previously shown to interact with soluble CSPs, but not with CSPs covalently-bound to target surfaces following classical complement activation, the present study aimed to identify novel cellular receptors interacting with covalently-deposited CSPs. By applying an unbiased screening approach using a cDNA mammalian expression library generated from human monocyte-derived dendritic cells and probed with recombinant human C4d, we identified neuropilin-1 (NRP1) as a novel receptor for C4d, C3d, and iC3b. NRP1, a highly conserved type 1 transmembrane protein, plays important roles in the development of the nervous and cardiovascular system as well as in tumorigenesis through interaction with its established binding partners, such as vascular endothelial growth factor (VEGF) and semaphorin 3A (Sema3A). NRP1 is also expressed on immune cells and serves as a marker for murine Tregs. Although NRP1 contains domains homologous to ones found in some complement proteins, it has not been linked to the complement system. We demonstrate that binding of C4d to NRP1 expressing cells was dose-dependent and saturable, and had a K_D value of 0.71 μM. Importantly, and in contrast to ILT4, NRP1 interacted with CSPs that were covalently bound to target surfaces in the course of complement activation, therefore representing a classical complement receptor. The binding site of CSPs was mapped to the b1 domain of the coagulation factor V/VIII homology domain of NRP1. Taken together, our results demonstrate a novel role for NRP1 as a receptor for CSPs deposited on surfaces during complement activation. Further work is required to elucidate the functional consequences of the NRP1-CSP interactions in immunity.

Keywords: C3d; C4d; complement receptors; complement split products; iC3b; neuropilin-1.

Figure 1

Identification of NRP1 as a receptor for C4d. (A) C4d-reactive cells enriched from a BW cell pool expressing a moDCs-cDNA library by multiple rounds of cell sorting. Sorting gates are shown. (B) A single cell clone derived from the C4d-reactive BW cell pool was probed with rh-C4Ad and rh-C4Bd and analyzed via flow cytometry. (C) PCR-amplification of retroviral inserts of a C4d-binding clone. (D) BW cells expressing a 5 kb retroviral insert encoding NRP1 were probed with a NRP1 mAb (monoclonal) or biotinylated rh-C4Ad, rh-C4Bd or ih-C4d (20 μg/ml each; open histograms: reactivity of NRP1 mAb or C4d to BW control cells; gray histograms: reactivity of NRP1 mAb or C4d to BW NRP1 cells). Biotinylation of rh-C4Ad and rh-C4Bd employed the NHS-biotin procedure, except for ih-C4d, which was specifically biotinylated on the thioester carbonyl moiety employing amine-PEG2-biotin reagent. (E) Monocytes and moDCs analyzed for NRP1 expression (open histograms: isotype control; gray histograms: NRP1 mAb). MFI, mean fluorescence intensity.

Figure 2

Interaction of NRP1 and mNRP1 with complement split products C4d, C3d, and iC3b. (A) Flow cytometric analysis of BW cells transduced to express high levels of human NRP1. Interaction of indicated CSPs (20 μg/ml each) with BW control cells (open histograms) and BW cells expressing NRP1 (gray histograms). Expression of NRP1 was verified with a monoclonal NRP1 antibody. (B) Interaction of plate-bound CSPs (465 nM each) with soluble recombinant human NRP1-immunoglobulin fusion protein (rh-NRP1-Ig) and complement receptor Ig fusion protein (rh-CRIg-Ig) analyzed in an ELISA-based assay. (C) Binding of murine NRP1 (mNRP1) mAb and recombinant human CSPs (rh-C4d, ih-iC3b, and ih-C3d) to BW control cells (open histograms) and BW cells expressing mNRP1 (gray histograms). Data shown is representative of two independently performed experiments.

Figure 3

Classical receptor-ligand interaction between C4d and NRP1. (A) BW NRP1 cells or BW control cells were incubated with increasing concentrations of rh-C4Ad. Cell-bound rh-C4Ad was measured by flow cytometry using a monoclonal C4d antibody for detection. Mean and standard deviation of triplicate measurements are shown. (B) Dissociation constant (K_D) of the C4d-NRP1 interaction was assessed via Langmuir binding isotherm equation for single class 1:1 binding equilibrium based on data shown in A. Error estimates of the fitted parameters to the mean of triplicate data points, and the correlation coefficient of the fit to the model are indicated. (C,D) Binding of biotinylated rh-C4Ad or rh-C4Bd (20 μg/ml each) to BW control cells or BW NRP1 in absence or presence of unlabeled rh-C4d (C) or a polyclonal NRP1 antibody (10 μg/ml) (D). (E) Binding of biotinylated rh-C4Bd (20 μg/ml each) to BW control cells or BW mNRP1 in absence or presence of a polyclonal NRP1 antibody (10 μg/ml). In (A–E) probe biotinylation was performed with iodoacetyl-LC-biotin specifically targeting the thioester cysteine. MFI, mean fluorescence intensity.

Figure 4

Interaction between NRP1 and C4d deposited to surfaces via thioester binding following classical complement activation. (A) C4d biotinylated at its intrinsic thioester carbonyl via amine-PEG2-biotin reagent (ih-C4d-bio) was immobilized to a solid phase either via streptavidin (+SA) or in random orientation without SA (–SA) and detected by ELISA using a monoclonal C4d antibody. (B) Interaction between C4d, immobilized to a solid phase as shown in (A), and rh-NRP1-Ig and rh-ILT4-Ig. Rh-CTLA-4-Ig fusion proteins and goat anti-human-IgG (ctrl) served as negative controls. (C) C4d biotinylated via its intrinsic thioester was tetramerized using streptavidin-APC and probed with BW cells expressing ILT4 or NRP1 and analyzed via flow cytometry. (D) Antibody-coated sheep erythrocytes (EA) bearing C4d were prepared using purified C1, C4, C4 binding protein (C4BP), and factor I (FI) as described in Material and Methods. Deposition of C4d was detected with a monoclonal C4d antibody (upper panel). Interaction between C4d-loaded EA and control EA with immunoglobulin fusion proteins (rh-ILT4-Ig; rh-NRP1-Ig) is shown in the lower panel. (E) Binding of goat anti-human-IgG (ctrl) or rh-NRP1-Ig (20 or 60 μg/ml) to C4d-EA in the presence or absence of an anti-C4d and anti-C4c antibodies. Data shown are representative for two independently performed experiments. gMFI, geometric mean of fluorescence intensity.

Figure 5

The b1 region of neuropilin-1 is essential for the interaction with CSPs. (A) Schematic representation of wildtype (wt) NRP1 and seven deletion variants. (B) Flow cytometric analysis of BW cells expressing wildtype NRP1 and NRP1 deletion variants depicted in (A) using a polyclonal NRP1 antibody. (C) BW control cells (ctrl) and BW cells expressing wt-NRP1 or NRP1 deletion variants were probed with the indicated biotinylated rh-C4Bd, ih-C3d, and ih-iC3b molecules (20 μg/ml each) followed by SA-PE and analyzed via flow cytometry. The data are from three independent experiments performed in triplicates. (D) Binding of an immunoglobulin fusion protein representing the CUB domain (a1a2) of NRP1 (rh-NRP1-CUB-Ig; 10 μg/ml) and two differently generated fusion proteins representing the extracellular part (a1a2 and b1b2) of NRP1 (rh-NRP1-Ig (generated in house) and rh-NRP1-Ig (c.) (commercial); 10 μg/ml each) to immobilized rh-C4Ad, rh-C4Bd and ih-C3d (465 nM each) was assessed by ELISA. gMFI, geometric mean of fluorescence intensity; SEA, the SerGluAla C-terminal tripeptide represents a PDZ domain-binding motif.

Figure 6

NRP1 interacts with MHC class I molecules. (A) Binding of tetramers representing MHC class I molecule HLA-A*02:01 (CMV CLGGLLTMV) with BW control cells and BW cells expressing NRP1 or ILT4 in the presence or absence of a polyclonal NRP1 antibody via flow cytometry. (B) Data of two independently performed binding experiments is shown. gMFI, geometric mean of fluorescence intensity.