A novel fusion protein reduces kidney complement in experimental C3 glomerulopathy
Talat H Malik, Karolina Kwiatkowska, Hannah J Lomax-Browne, Charlotte M Bottomley, Matthew Bright, Zhang Sung Tean, Ahmet K Akturk, Ian E Alexander, Grant J Logan, Matthew C Pickering
TL;DR
A new fusion protein targeting kidney complement reduces damage in a mouse model of C3 glomerulopathy without affecting the rest of the body.
Contribution
A novel fusion protein targeting glomerular complement C3 with tissue-specific inhibition in kidney disease.
Findings
AAV-mediated fusion protein reduced glomerular C3b/iC3b/C3c and properdin in FH-deficient mice.
FHR51-9FH1-5 ameliorated C3 glomerulopathy in a CFHR5 nephropathy mouse model.
In vitro, the fusion protein bound C3 and reduced C3a generation in alternative pathway assays.
Abstract
Complement activation contributes to kidney damage in many types of glomerulonephritis and complement inhibition therapy is approved for IgA nephropathy and C3 glomerulopathy. However, inhibition is not specific to the kidney resulting in unnecessary systemic complement inhibition and increased infection risk. To develop an effective inhibitor of glomerular complement we combined complement factor H-related protein 5 (FHR51-9), which binds to glomerular complement C3, with the complement regulatory domains of the key negative regulator of C3 activation complement factor H (FH1-5). One week after adeno-associated virus (AAV) mediated expression of the FHR51-9FH1-5 fusion protein in factor H (FH)-deficient mice, glomerular C3b/iC3b/C3c was significantly reduced and properdin resolved completely compared to controls. There was no change to circulating C3 levels and FHR51-9FH1-5 was…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5- —Wellcome Trust Senior Fellowship in Clinical Science
- —Rebecca L. Cooper Foundation
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsComplement system in diseases · Renal Diseases and Glomerulopathies · Monoclonal and Polyclonal Antibodies Research
Introduction
C3 glomerulopathy (C3G) encompasses a group of kidney diseases characterized by abnormal glomerular C3 deposition due to dysregulation of the complement alternative pathway (AP) [1, 2]. This pathway is constitutively active, so effective AP regulation is essential to prevent uncontrolled activation and damage to host tissues. Factor H (FH) is the main AP inhibitor and is a 155 kDa plasma glycoprotein consisting of 20 short consensus repeat (SCR) domains. SCR domains 1–4 mediate complement regulation whilst domains 19 and 20 are required for efficient interaction with surface complement and polyanions. The causes of AP dysregulation in C3G are multiple [3]. Some potentiate AP activation such as C3 nephritic factors and anti-factor B autoantibodies. Others impair FH function and include FH deficiency [4–6], FH autoantibodies [7, 8], paraproteins [9], and abnormal factor H-related (FHR) proteins [10].
FHR proteins, like FH, are composed of SCR domains. Unlike FH, they lack complement regulatory domains but can interact with surface complement. Whilst their function is incompletely understood, in some circumstances these proteins interfere with surface complement regulation resulting in complement-mediated disease [10]. A key example is CFHR5 nephropathy, a familial C3G, characterized by a heterozygous internal duplication within the CFHR5 gene resulting in the production of an abnormal FHR5 protein (FHR5mut) [11, 12]. FHR5 is a FH-related protein that has long been recognized to interact with kidney complement: it is found in association with glomerular C3 in multiple types of glomerular diseases [13]. FHR5 binds to C3b, iC3b, and C3d [14] and in vitro can promote C3 activation by facilitating C3 convertase formation [15]. We modelled CFHR5 nephropathy in vivo by generating mouse strains expressing human FH with either FHR5 (hFH-FHR5) or FHR5mut (hFH-FHR5mut) in the absence of mouse FH-FHR proteins [16]. The hFH-FHR5mut phenotype spontaneously recapitulated key features of CFHR5 nephropathy: abnormal glomerular C3 deposition and mesangial and glomerular basement membrane electron-dense deposits. C3G has been associated with other changes in the FHR protein family [10]. An abnormal FHR1 protein causes C3G through its increased capacity to interact with C3-opsonized surfaces and promote AP activation independently of ligand competition with FH [17].
Definitive treatment strategies for C3G have focused on complement inhibition since the accumulation of glomerular C3 is central to the pathogenesis and preventing this would be predicted to protect the kidney from further damage. Iptacopan, a factor B inhibitor, and pegcetacoplan, a C3 inhibitor, have shown efficacy in clinical trials [18–20]. Experimental approaches include FH-derived engineered proteins with enhanced surface complement regulation [21–23] and inhibitors that target sites of complement activation [24, 25]. Homodimeric mini-FH (HDM-FH) combines FH SCR1-5 and SCR18-20 with FHR1 dimerization domains and was shown to reduce glomerular complement in the FH-deficient and hFH-FHR5mut C3G mouse models [16, 23]. Targeted molecules include a fusion protein (CR2-FH) [25] comprised of the complement regulatory domains of FH (SCR1-5) linked to the C3 fragment-binding domains of complement receptor 2 (CR2 SCR1-4) and antibody fusion proteins that combine FH SCR1-5 with an anti-C3d monoclonal antibody (C3d-mAb-2fH) [24]. Both have shown efficacy in reducing glomerular complement in mouse [24, 25] and rat [24] models.
Here our aim was to efficiently target a complement regulator to sites of glomerular C3 activation to achieve glomerular-specific complement inhibition. We chose to harness the ability of FHR5 to interact with glomerular C3. Consequently, we combined FHR5 with the FH SCR1-5. This fusion protein, termed FHR5_1-9_FH_1-5_, ameliorated glomerular complement activation in FH-deficient mice and in the hFH-FHR5mut C3G mouse model. FHR5_1-9_FH_1-5_ is an effective means of targeting glomerular C3 and could achieve complement inhibition in the presence of either FH deficiency or the FHR5mut protein.
Materials and methods
Mice
delFH-FHR [16], delFH [26], hFH-FHR5, [16] and hFH-FHR5mut [16] mice were generated as previously described. All mice were housed in specific pathogen-free conditions and procedures performed according to institutional guidelines and approved by the United Kingdom Home Office. All experiments were conducted in accordance with the ARRIVE guidelines [27]. Male and female mice were used except for experiments using the hFH-FHR5 and hFH-FHR5mut strains where we used only male mice since, as reported [16], transgene expression is insufficient in the female strain. All in vivo experiments were performed twice except for the 20-week experimental timepoint in hFH-FHR5mut mice presented in supplemental data.
Adeno-associated virus (AAV) constructs and recombinant FHR51-9FH1-5
The cDNA for FHR5mut, green fluorescent protein (GFP) and FHR5_1-9_FH_1-5_ containing a Kozak sequence in each case, were inserted into an AAV vector construct with a modified promoter sequentially containing one copy of the ApoE enhancer, a human 1-α-antitrypsin promoter and SV40 intron [28]. We utilized the hepatotropic serotype 8 capsid [29] for AAV8-GFP and AAV8-FHR5_1-9_FH_1-5_ and hepatotropic serotype KP1 capsid [30] for AAV-KP1-FHR5mut to enable robust transgene expression in the adult mouse liver following a single intra-peritoneal injection of the AAV vector. AAV vector constructs were prepared as previously described [16]. Recombinant FHR5_1-9_FH_1-5_ consists of full length FHR5 (SCR1-9) linked to SCR1-5 of human FH (Supplementary Figure S1). It was generated by GeneArt (Invitrogen), subcloned into a CMV expression vector, confirmed by sequencing and transfected into HEK-293T cells using lipofectamine 3000 (Thermo Fisher Scientific). The FHR5_1-9_FH_1-5_ protein was purified from supernatant using a heparin column, quantified, and shown to react with anti-FH and FHR5 antibodies and glomerular C3 ex vivo (Supplementary Figure S1).
Complement phenotyping
Plasma was collected from mice in the presence of EDTA by cardiac puncture. For western blotting, samples were separated using SDS-PAGE under reducing conditions for C3 and non-reducing conditions for human FHR5 and mouse C5. Detection antibodies: goat anti-mouse C3 (MP Biomedicals), rabbit anti-human CFHR5 (Abnova), and goat anti-human C5 (Quidel). Secondary antibodies: mouse anti-goat/sheep IgG-HRP (Sigma) and swine anti-rabbit IgG-HRP (Dako). Blots were visualized using enhanced chemiluminescence (ECL) substrate (Pierce, Thermo).
Mouse C3, human FH, and FHR5 levels in mouse plasma were measured by ELISA as previously described [16]. To measure FHR5_1-9_FH_1-5_ levels, ELISA plates were coated with 2 µg/mL of anti-FH monoclonal OX24 antibody (Thermo) in 0.2 M carbonate-bicarbonate buffer then blocked with PBS-1% BSA. Mouse plasma (1:1000 or 1:2000) was added in PBS–0.1% Tween-20–1% BSA. After incubation and washing, bound FHR5_1-9_FH_1-5_ was detected using goat polyclonal anti-CFHR5 antibody (R&D systems), followed by mouse anti-goat/sheep IgG-HRP (Sigma) and finally TMB substrate (BD). Results were quantified using a standard curve generated using recombinant FHR5_1-9_FH_1-5_ protein.
Renal phenotyping
Immunofluorescence staining was performed on 5 µm cryosections mounted using Vectashield medium with 4′,6-diamidino-2-phenylindole (Vector Laboratories). Antibodies: C3b/iC3b/C3c—fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse C3 (MP Biomedicals); C3d-biotinylated-goat anti-mouse C3d (R&D systems) and Alexa-488 conjugated streptavidin; human FH—polyclonal goat anti-human FH (Quidel) and FITC-conjugated mouse anti-goat/sheep IgG (Sigma); FHR5/FHR5mut—rabbit monoclonal anti-human CFHR5 antibody (Abnova) and FITC-conjugated goat anti-rabbit IgG (Invitrogen); mouse properdin—FITC-conjugated goat anti-human properdin (Caltag Medsystems, GAHu/PPD/FITC); mouse C5—goat anti-human C5 (Quidel) and FITC conjugated mouse anti-goat/sheep IgG (Sigma). Immunofluorescence quantification was performed on images (Leica DM4B optical microscope, 10 glomeruli per section) using ImageJ and values expressed as arbitrary fluorescence units (AFU).
C3 binding assays
C3 binding assays were carried out in Maxisorp ELISA plates. These assays were performed with human proteins. Wells were coated overnight at 4°C with 5 µg/mL C3, C3b, iC3b, C3d (Complement Technologies) or 10% BSA in 0.1 M carbonate-bicarbonate buffer, pH 9.6. All washing steps were carried out with PBS-Tween (PBST, 0.1% Tween-20) and wells were blocked with 2% BSA (w/v) in PBST for 1 h at room temperature. Next, serial dilutions (from 1 to 64 nM) of FHR5_1-9_FH_1-5_ in blocking buffer were incubated for 1 h at room temperature. FHR5_1-9_FH_1-5_ binding was detected with both mouse anti-CFHR5 (Abnova) and sheep anti-human FH (Abcam), followed by rabbit anti-mouse-HRP (Dako) and donkey anti-sheep-HRP (Jackson), respectively.
C3 complement activation and cofactor assays
C3 complement activation assays modified from [17] were carried out in Maxisorp ELISA plates. These assays were performed with human proteins. Wells were coated overnight at 4°C with 5 µg/mL C3b (Complement Technologies) in 0.1 M carbonate–bicarbonate buffer, pH 9.6. The assays were performed in complement buffer (Tris buffered saline; TBS pH 7.4, 2 mM calcium chloride, 5 mM magnesium chloride) and washing steps between each incubation were carried out with TBS-Tween (TBST, 0.1% Tween-20). Following blocking with 3% BSA (w/v) in TBS for 2 h at room temperature, factor B (10 µg/mL, Complement Technologies), factor D (0.1 µg/mL, Complement Technologies), and properdin (2 µg/mL, Complement Technologies) were incubated for 1 h at 37°C, followed by incubation with 10 µg/mL C3 (Complement Technologies) for 1 h at 37°C. The supernatant was collected and diluted 1/10 and generated C3a was measured using a MicroVue C3a plus EIA kit (Quidel). When testing effects on complement activation, FH (0.04, 0.08, 0.16, 0.32, 0.64, and 1.28 µM) or FHR5_1-9_FH_1-5_ (16, 32, 64, 128, 256, and 512 nM) were included in the assay for 30 min at 37°C, in an independent step before the addition of C3. When testing cofactor activity, FH (0.04, 0.08, 0.16, 0.32, 0.64, and 1.28 µM) or FHR5_1-9_FH_1-5_ (16, 32, 64, 128, 256, and 512 nm) were included in the assay with FI (1560 nM, Complement Technologies) for 30 min at 37°C, in an independent step immediately after blocking.
To investigate FHR5_1-9_FH_1-5_ acting as a fluid phase cofactor for FI, it was tested alongside FH in a C3b cofactor assay [31]. 512 nM of FHR5_1-9_FH_1-5_ or 1.28 µM of FH was added to 1μg of C3b, and 780 nM of FI, in a total volume of 15μl of PBS. The assays were incubated for 30 min at 37°C. The assay products were separated on a 10% SDS-PAGE gel under reducing conditions, followed by western blotting to PVDF membrane and detection with goat anti-human C3-HRP (MP Biomedicals).
Statistical analysis
Statistical analyses were performed using GraphPad Prism (version 10.0; GraphPad). Student’s t-test was used for two group comparisons and for multiple groups we used ANOVA with Dunnett’s multiple comparisons test.
Results
FHR51-9FH1-5 ameliorated glomerular C3b/iC3b/C3c deposition in both delFH-FHR and delFH mice
Mice lacking FH and the mouse FHR proteins (delFH-FHR) [16], like FH-deficient mice (delFH) [26] have uncontrolled AP activation with consequent depletion of plasma C3 and C5 and accumulation of C3 within glomeruli. To test the ability of FHR5_1-9_FH_1-5_ to ameliorate complement activation in vivo we examined the effects of AAV8-delivered FHR5_1-9_FH_1-5_ in male and female delFH-FHR mice (Fig. 1, Supplementary Figure S2). One-week after AAV8-FHR5_1-9_FH_1-5_ injection, the FHR5_1-9_FH_1-5_ protein was detectable in the circulation (median level 3.98 and 1.74 µg/mL in male and female mice, respectively). No change in plasma C3 levels was evident by ELISA (Fig. 1c, Supplementary Figure S2d) but a faint band consistent with the intact C3 α-chain was evident and plasma C5 became detectable by western blotting (Supplementary Figure S3). Glomerular properdin staining was absent and C3b/iC3b/C3c significantly reduced compared with controls in both male and female animals treated with AAV8-FHR5_1-9_FH_1-5_ (Fig. 1d and e, Supplementary Figure S2e). C5 staining was unchanged in both groups whilst glomerular C3d reduced significantly in female but not male animals (Fig. 1d and e, Supplementary Figure S2e). To detect glomerular FHR5_1-9_FH_1-5_, we stained kidney sections with anti-FH and anti-FHR5 antibodies. Capillary wall staining was evident with both antibodies at 1-week post-injection only in AAV8-FHR5_1-9_FH_1-5_ treated mice and staining colocalized with glomerular C3d indicating the presence of the fusion protein in glomeruli (Fig. 1f, Supplementary Figure S2f). We saw similar effects of AAV8-FHR5_1-9_FH_1-5_ in male and female delFH mice (Fig. 2, Supplementary Figure S4). One-week post-injection the FHR5_1-9_FH_1-5_ protein was detected in the circulation (median level 5.4 and 2.5 µg/mL in male and female mice, respectively), plasma C3 levels remained unchanged but C5 became detectable. The fusion protein was detected in glomeruli in association with C3d; glomerular properdin was absent and C3b/iC3b/C3c significantly reduced compared to controls with no change in C3d and borderline significant reduction in C5 staining. One female mouse in the AAV8-FHR5_1-9_FH_1-5_ treated group was found dead 24 h after intraperitoneal injection: cause of death undetermined. Taken together, these data showed that the FHR5_1-9_FH_1-5_ protein was effective at reducing glomerular C3 activation in FH deficiency. We next examined the effects of AAV-FHR5_1-9_FH_1-5_ on glomerular C3 in hFH-FHR5mut mice.
AAV-delivered FHR51-9FH1-5 to delFH-FHR mice. (a) Schematic showing the FHR51-9FH1-5 protein consists of FHR5 (nine SCR domains) linked to the first 5 SCR domains of FH. The complement regulatory domains of FH (red) are contained within the first 4 domains of the FH protein. 0.25 × 1012 vector genomes of either AAV8-FHR51-9FH1-5 (n = 4) or AAV8-GFP (n = 4) were administered to 21-week-old male delFH-FHR mice which were then culled 1 week later. AAV—adeno-associated virus; GFP—green fluorescent protein; ip—intraperitoneal; M—male; SCR—short consensus repeat; vg—vector genomes. (b) Plasma FHR51-9FH1-5 protein levels quantified by ELISA and plasma western blot (arrow) probed with a rabbit anti-human FHR5 antibody. Horizontal bars denote median values. (c) Plasma C3 levels in AAV8-GFP (black dots) and AAV8-FHR51-9FH1-5 (red dots) groups. Horizontal bars denote median values. (d) Glomerular scoring (arbitrary fluorescence units; AFU) for C3b/iC3b/C3c, C3d, C5, and properdin staining and (e) representative glomerular images. P values derived from unpaired t-test. Horizontal bars denote median values and range of values is indicated. (f) Representative glomerular co-staining of FH and FHR5 in AAV8-GFP and AAV8-FHR51-9FH1-5 groups (left panel). Representative co-staining of C3d with either FH or FHR5 in AAV8-FHR51-9FH1-5 treated animals (right panel). Scale bars = 20 µm.
AAV-delivered FHR51-9FH1-5 to delFH mice. (a) 0.25 × 1012 vector genomes of either AAV8-FHR51-9FH1-5 (n = 3M, 4F) or AAV8-GFP (n = 3M, 4F) were administered to 8-week-old delFH mice which were then culled 1 week later. AAV—adeno-associated virus; F—female; GFP—green fluorescent protein; ip—intraperitoneal; M—male; SCR—short consensus repeat; vg—vector genomes. One female mouse in the AAV8-FHR51-9FH1-5 treated group was found dead 24 h after intra-peritoneal injection: cause of death undetermined so this group reduced to n = 3 animals. (b) Plasma FHR51-9FH1-5 protein levels quantified by ELISA and plasma western blot (arrow) probed with a rabbit anti-human FHR5 antibody. Horizontal bars denote median values. (c) Plasma C3 levels in AAV8-GFP (black symbols) and AAV8-FHR51-9FH1-5 (red symbols) groups. Horizontal bars denote median values. (d) Glomerular scoring for C3b/iC3b/C3c, C3d, C5 and properdin staining and (e) representative glomerular images. AFU = arbitrary fluorescence units, P values derived from unpaired t-test. Horizontal bars denote median values and range of values is indicated. (f) Representative glomerular co-staining of FH and FHR5 in AAV8-GFP and AAV8-FHR51-9FH1-5 groups (left panel). Representative co-staining of C3d with either FH or FHR5 in AAV8-FHR51-9FH1-5 treated animals (right panel). Scale bars = 20 µm.
FHR51-9FH1-5 ameliorated glomerular C3b/iC3b/C3c deposition in hFH-FHR5mut male mice
Following AAV8-FHR5_1-9_FH_1-5_ administration to hFH-FHR5 and hFH-FHR5mut mice, plasma FHR5_1-9_FH_1-5_ was detectable at 1 week and increased further at the 6-week cull point (Fig. 3a and b, Supplementary Figure S5a). Plasma C3 levels did not change (Supplementary Figure S5b). No abnormal glomerular staining for C3b/iC3b/C3c, C3d, C5, or properdin was evident in either AAV8-GFP or AAV8-FHR5_1-9_FH_1-5_ treated hFH-FHR5 mice. However, the C3b/iC3b/C3c staining along Bowman’s capsule and within the tubulointerstitial areas that is seen in wild-type mice and in the unmanipulated hFH-FHR5 strain [16], was markedly reduced following administration of AAV8-FHR5_1-9_FH_1-5_ (Fig. 3c, Supplementary Figure S6). Consistent with the spontaneous phenotype [16], AAV8-GFP treated hFH-FHR5mut mice had abnormal glomerular C3b/iC3b/C3c, C3d, C5, properdin staining (Fig. 3d). In contrast, glomerular C3b/iC3b/C3c, C3d, and C5 were significantly reduced and properdin was absent in AAV8-FHR5_1-9_FH_1-5_ treated hFH-FHR5mut mice (Fig. 3d and e). The presence of the FHR5mut protein in the glomeruli of hFH-FHR5mut mice results in strong staining with anti-FHR5 antibodies [16] and this was evident in the AAV8-GFP treated hFH-FHR5mut mice where FHR5 reactivity colocalized with C3b/iC3b/C3c, C3d, and FH (Fig. 3f and g). In the AAV8-FHR5_1-9_FH_1-5_ treated hFH-FHR5mut mice, glomerular FHR5 staining was present and of similar magnitude to that of the AAV8-GFP treated animals and colocalized with C3d and FH (Fig. 3f and g). Glomerular FH staining was significantly increased compared to AAV8-GFP treated animals (Fig. 3f). Taken together with the reduction in glomerular C3b/iC3b/C3c at this time-point, the FHR5 and FH staining likely reflects the presence of the fusion protein in the kidney. To determine if longer exposure to the FHR5_1-9_FH_1-5_ protein could completely resolve glomerular C3d we culled hFH-FHR5mut mice 20 weeks after administration of AAV8-FHR5_1-9_FH_1-5_ (Supplementary Figure S7). Like the 6-week time point, glomerular C3b/iC3b/C3c was significantly reduced and properdin staining was absent. Whilst reduced, glomerular C3d was still detectable but C5 staining was now negative. These data indicated that the FHR5_1-9_FH_1-5_ protein ameliorated pre-existing FHR5mut-mediated glomerular complement activation. We next examined if the FHR5_1-9_FH_1-5_ protein could prevent FHR5mut-triggered glomerular complement activation.
AAV-delivered FHR51-9FH1-5 to hFH-FHR5 and hFH-FHR5mut mice. (a) 1 × 1012 vector genomes of either AAV8-FHR51-9FH1-5 or AAV8-GFP were administered to 10-week-old hFH-FHR5 or hFHR5mut mice which were then culled 6 weeks later. AAV—adeno-associated virus; M—male; GFP—green fluorescent protein; ip—intraperitoneal; vg—vector genomes. (b) Plasma FHR51-9FH1-5 protein levels quantified by ELISA. Horizontal bars denote median values. Representative glomerular images for C3b/iC3b/C3c, C3d, C5, and properdin staining in (c) hFH-FHR5 and (d) hFH-FHR5mut groups. (e) Glomerular scoring for C3b/iC3b/C3c, C3d, C5, and properdin staining in the AAV8-FHR51-9FH1-5 treated hFH-FHR5mut animals. AFU = arbitrary fluorescence units, P values derived from unpaired t-test. Horizontal bars denote median values and range of values is indicated. (f) Representative glomerular staining for FH and FHR5 in AAV8-GFP and AAV8-FHR51-9FH1-5 treated hFH-FHR5 or hFHR5mut mice (left panel). Glomerular scoring for FH and FHR5 staining in the AAV8-FHR51-9FH1-5 treated hFH-FHR5mut animals (right panel). P values derived from unpaired t-test. Horizontal bars denote median values and range of values is indicated. (g) Representative co-staining of FHR5 with C3b/iC3b/C3c, C3d, and FH in hFH-FHR5mut animals treated with AAV-GFP (left panel) or AAV8-FHR51-9FH1-5 (right panel). Scale bars = 20 µm.
FHR51-9FH1-5 prevented the development of FHR5mut protein triggering glomerular C3 deposition in hFH-FHR5 male mice
We have previously shown that AAV-delivered FHR5mut results in the development of abnormal glomerular C3 in hFH-FHR5 mice [16]. To test if the FHR5_1-9_FH_1-5_ could prevent the glomerular C3 deposition in this model, we administered AAV8-FHR5_1-9_FH_1-5_ to hFH-FHR5 male mice 1 week before injection of AAV-KP1-FHR5mut (Fig. 4). Plasma FHR5_1-9_FH_1-5_ was detectable 1 week after the AAV8-FHR5_1-9_FH_1-5_ injection in all six hFH-FHR5 animals (Fig. 4b). Following the AAV-KP1-FHR5mut injection, plasma FHR5mut levels increased 1 week later and at the 7-week cull point (Fig. 4c). We were able to detect the sequential appearance of the FHR5_1-9_FH_1-5_ and FHR5mut proteins by plasma western blotting using an anti-FHR5 antibody (Fig. 4d). Plasma C3 levels reduced following the AAV-KP1-FHR5mut injection in both the AAV8-GFP and AAV8-FHR5_1-9_FH_1-5_ groups (Fig. 4e), a phenomenon that we have previously seen [16]. As expected in the AAV8-GFP/AAV-KP1-FHR5mut group, glomerular staining showed abnormal C3b/iC3b/C3c, C3d, C5, and properdin staining but in the AAV8-FHR5_1-9_FH_1-5_/AAV-KP1-FHR5mut animals glomerular C3b/iC3b/C3c, C3d, and C5 staining was significantly reduced and properdin staining was absent (Fig. 4f and g). Glomerular FHR5 and FH staining which co-localized with C3b/iC3b/C3c were present in both groups, although FHR5 staining was reduced in the AAV8-FHR5_1-9_FH_1-5_/AAV-KP1-FHR5mut group (Fig. 4h and i). These data showed that the FHR5_1-9_FH_1-5_ protein ameliorated FHR5mut-triggered glomerular complement activation.
AAV-delivered FHR51-9FH1-5 and hFHR5mut to hFH-FHR5 mice. (a) 0.25 × 1012 vector genomes of either AAV8-FHR51-9FH1-5 or AAV8-GFP were administered to 7-week-old hFH-FHR5 mice followed 1 week later by AAV-KP1-FHR5mut and all animals culled 7 weeks after the first AAV injection. AAV—adeno-associated virus; M—male; GFP—green fluorescent protein; ip—intraperitoneal; vg—vector genomes. (b) Plasma FHR51-9FH1-5 protein levels quantified by ELISA. (c) Plasma FHR5 levels. Horizontal bars denote median values. (d) Representative western blot of plasma probed with a rabbit anti-human FHR5 antibody. Arrows indicate FHR5, FHR5mut and the FHR51-9FH1-5 proteins. (e) Plasma C3 levels. Horizontal bars denote median values. Representative glomerular images (f) and scoring (g) for C3b/iC3b/C3c, C3d, C5, and properdin staining. AFU = arbitrary fluorescence units, P values derived from unpaired t-test. Horizontal bars denote median values and range of values is indicated. (h) Representative glomerular staining for FH and FHR5 and glomerular scoring for FH and FHR5 staining. P values derived from unpaired t-test. Horizontal bars denote median values and range of values is indicated. (i) Representative co-staining of FHR5 with C3b/iC3b/C3c and FH in hFH-FHR5 animals treated with AAV-GFP-AAV-KP1-FHR5mut (left panel) or AAV8-FHR51-9FH1-5-AAV-KP1-FHR5mut (right panel). Scale bars = 20 µm.
Recombinant FHR51-9FH1-5 interacts with immobilized C3 ligands, inhibits the alternative pathway convertase and acts as a cofactor for factor I
To determine the mechanism of action of the FHR5_1-9_FH_1-5_ protein, we first examined the interaction of recombinant FHR5_1-9_FH_1-5_ with surface-immobilized C3, C3b, iC3b, and C3d in vitro. FHR5_1-9_FH_1-5_ bound to these ligands in a dose-dependent manner with the highest binding to C3d (Fig. 5a and b). To test the ability of FHR5_1-9_FH_1-5_ to regulate the AP C3 convertase we generated a convertase by sequential addition of factors B, D, and properdin to surface-immobilized C3b. We then added C3 and measured C3a production. Addition of either FH or the FHR5_1-9_FH_1-5_ resulted in dose-dependent reduction in C3a (Fig. 5c and d). To investigate cofactor activity, we added FI with either FH or FHR5_1-9_FH_1-5_ to surface-immobilized C3b and then tested the ability of the surface to form an AP convertase by sequential addition of factors B, D, and properdin followed by C3 and measurement of C3a production. FH and FHR5_1-9_FH_1-5_ resulted in dose-dependent reduction in C3a (Fig. 5e and f) indirectly indicating that the surface C3b had been converted to iC3b through the actions of FI. Using western blotting of C3 under reducing conditions to enable us to differentiate C3 activation fragments, FHR5_1-9_FH_1-5_ in combination with FI, resulted in cleavage of C3b indicating that FHR5_1-9_FH_1-5_ can act as a cofactor for FI in the fluid phase (Fig. 5g).
*FHR51-9FH1-5 binds to C3, destabilizes C3 convertase and acts as a cofactor for factor I. Binding profiles of increasing amounts of FHR51-9FH1-5 to surface bound C3, C3b, iC3b, C3d and BSA (negative control), detected with both anti-CFHR5 (a) and anti-FH (b) antibodies. Data shown are mean values of triplicate measurements; error bars denote the standard deviation. C3a generated by convertase formation on surface-bound C3b: ELISA plates were coated with purified C3b. Surface C3 convertase was formed by addition of purified factor B (FB), factor D (FD), and properdin (P). Then increasing amounts of either FH (c) or the FHR51-9FH1-5 (d) were added followed by addition of C3 to enable C3a generation. To investigate factor I (FI) co-factor activity, FI along with increasing amounts of either FH (e) or the FHR51-9FH1-5 (f) were added prior to the convertase formation step. NC—negative control, no FB, FD, or P added. PC—positive control, no FH, or FHR51-9FH1-5 added. P values were determined by one-way ANOVA with Dunnett’s multiple comparisons test; ***P < 0.0001. Fluid phase co-factor assay products were visualized by western blotting (g). Cofactor activity was indicated by the cleavage of C3b as determined by the presence of C3 α-chain fragments (arrow). Negative controls: C3, C3b, C3b with FH, C3b with FI, and FHR51-9FH1-5 alone. Positive control: C3b incubated with FH and FI.
Discussion
We have shown that AAV8-FHR5_1-9_FH_1-5_ treatment reduced glomerular complement in C3G mouse models driven by either FH deficiency (the delFH-FHR and delFH strains) or FHR5mut, the abnormal FHR5 protein associated with CFHR5 nephropathy [12]. In the FH deficiency models, we examined the phenotype after 1 week due to concerns about potential immunogenicity of the FHR5_1-9_FH_1-5_ protein. However, even at this early timepoint, glomerular properdin was absent and glomerular C3b/iC3b/C3c significantly reduced. Plasma C3 levels did not change but in the FH-deficient strains we detected increased plasma C5. In previous studies, the administration of fusion proteins utilizing the regulatory domains of FH (e.g. complement receptor 2_1-4_-FH_1-5_ [25], IgG-FH_1-5_, and anti-P-FH_1-5_ [32]) also resulted in reduced glomerular C3 together with increases in circulating C5 that persisted for longer than increases in either plasma C3 or factor B. It appears that regulation of the C5 convertase is more readily achieved by FH-containing fusion proteins than regulation of the fluid-phase C3 convertase. Notably, the C5 convertase is partially properdin-dependent in delFH mice [33, 34] and in C3G circulating properdin levels correlate with surface C5 convertase activity [35, 36]. In this study, we interpret the resolution of glomerular properdin to indicate reduction in both C3 and C5 convertase activity, the latter resulting in the appearance of C5 in circulation. Using an anti-C5 antibody, we detected a significant reduction in glomerular C5 reactivity in delFH but not delFH-FHR mice. C5 within the membrane attack complex is known to persist in tissues for long periods even after cessation of complement activation [37]. Notably, the AAV-delivery consistently resulted in higher levels of circulating FHR5_1-9_FH_1-5_ in male compared to female mice on both the delFH-FHR and delFH background. AAV-delivered protein resulting in higher levels in male mice has been reported previously [38] and thought to be due to an androgen-dependent pathway [39]. Although the level of FHR5_1-9_FH_1-5_ in circulation was higher in the male compared to the female FH-deficient strains the phenotypic response was identical. In the delFH-FHR strain it was straightforward to determine if the FHR5_1-9_FH_1-5_ protein was present in glomeruli since this strain does not have any mouse FH or FHR proteins that could interact with detection antibodies. Our staining data showed the protein in a glomerular location comparable to that seen for mouse C3d in both FH-deficient strains suggesting that the fusion protein interacts via its FHR5 domains with glomerular C3 fragments and negatively regulates glomerular complement activation using the FH_1-5_ domains.
We have previously shown that male hFH-FHR5mut mice develop spontaneous C3G whereas those expressing the normal FHR5 protein (hFH-FHR5) do not [16]. The advantage of these humanized models was not only that they represent a model of C3G distinct from FH deficiency, but also that the FHR5_1-9_FH_1-5_ protein would be unlikely to trigger an immune response enabling us to examine the longer-term effects of AAV8-FHR5_1-9_FH_1-5_. Over 6 weeks, AAV8-FHR5_1-9_FH_1-5_ significantly reduced glomerular C3b/iC3b/C3c, C3d and C5, and properdin completely resolved. Over 20 weeks, AAV8-FHR5_1-9_FH_1-5_ resolved C5 staining completely but, although reduced, there was still glomerular C3d staining. In experimental nephritis models, glomerular C3d is known to persist for long periods following cessation of complement activation [40]. It is likely that, provided complement activation has stopped, clearance of C3d from non-cellular surfaces is dependent on surface turnover or remodelling. Detecting the presence of the FHR5_1-9_FH_1-5_ protein in glomeruli of hFH-FHR5mut mice using anti-FHR5 antibodies was complicated by the fact that unmanipulated hFH-FHR5mut animals have strong glomerular staining for FHR5 due to the presence of glomerular FHR5mut protein. In the AAV8-FHR5_1-9_FH_1-5_ treated hFH-FHR5mut mice, we could not detect any difference in glomerular FHR5 staining compared with AAV8-GFP treated hFH-FHR5mut mice. In both cases, FHR5 staining was marked. However, we did detect a significant increase in glomerular FH staining in the AAV8-FHR5_1-9_FH_1-5_ treated hFH-FHR5mut mice and the glomerular FH and FHR5 staining colocalized. We concluded that these observations represented glomerular FHR5_1-9_FH_1-5_ protein. We did not detect any difference in glomerular IgG staining between the hFH-FHR5mut mice treated with either AAV-GFP or AAV-FHR5_1-9_FH_1-5_ at cull at either 6 or 20 weeks (data not shown). However, we do not know if long-term administration of FHR5_1-9_FH_1-5_ would be immunogenic in human serum where additional FHR proteins would be present. In these experiments, we included male hFH-FHR5 mice which do not develop any glomerular abnormalities. However, we did observe that the normal C3b/iC3b/C3c staining along Bowman’s capsule and within the tubulointerstitial areas that is seen in this strain, and in wild-type mice, was markedly reduced following administration of AAV8-FHR5_1-9_FH_1-5_. This staining is known to be AP activation dependent [41] and we speculate that the fusion protein ameliorated this physiological activation in this strain. The FHR5_1-9_FH_1-5_ protein was also able to ameliorate glomerular complement activation in hFH-FHR5 mice treated with AAV-hFHR5mut. In this model, where AAV-hFHR5mut triggers glomerular C3 activation in hFH-FHR5 mice, our data showed that prior expression of AAV-FHR5_1-9_FH_1-5_ ameliorated the subsequent glomerular complement activation. When we first reported this model, we noted that the expression of FHR5mut results in a fall in circulating C3 levels in the hFH-FHR5 mice [16]. This was seen again and notably occurred irrespective of whether the mice had received AAV8-FHR5_1-9_FH_1-5_. So, whilst AAV8-FHR5_1-9_FH_1-5_ administration reduced glomerular complement in this model, it had no impact on the reduction in circulating C3 driven by FHR5mut. This reinforced the notion that FHR5_1-9_FH_1-5_ protein is effective at reducing complement activation in the kidney but not in the circulation.
Our experiments in vivo provided compelling evidence that FHR5_1-9_FH_1-5_ could regulate glomerular complement in models of C3G and regulate the physiological activation of complement within tubulointerstitial areas. To examine the binding of the fusion protein to C3 and C3 proteolytic fragments together with testing FH_1-5_ function such as FI cofactor activity we generated recombinant protein. In vitro recombinant FHR5_1-9_FH_1-5_ bound to C3 and its fragments in a dose-dependent manner with the highest binding to C3d. Like FH, recombinant FHR5_1-9_FH_1-5_ negatively regulated the C3 convertase in vitro and had FI cofactor activity.
There are limitations in our study. Our in vivo data is derived from a humanized mouse model which may not recapitulate the actions of the fusion protein in humans. FHR5_1-9_FH_1-5_ could interact with complement deposits on any endogenous or exogenous surface so might not mitigate the infection risk associated with systemic complement inhibition. The binding of FHR5_1-9_FH_1-5_ to kidney surfaces is complement-dependent. This means a degree of complement activation will necessarily occur before the protein can prevent further activation. Whether this initial activation could result in kidney damage is unclear. Our approach contrasts with designing complement inhibitors which target tissue independently of local complement activation (reviewed in [42, 43]). Our in vivo experiments utilize AAV-delivery of the inhibitor and, although we show proof of principle data in an animal model, we have not determined the optimal pharmacodynamics or feasibility of this approach in C3G. This will require labelling of the protein so we can accurately track its tissue distribution in vivo without confounding from endogenous FHR5 reactivity. Importantly, our models of C3G are driven by either FH deficiency or the FHR5mut protein so we do not know if FHR5_1-9_FH_1-5_ would be effective in other causes of C3G, most obviously nephritic factors.
We set out to develop a protein that would target glomerular C3 and inhibit further complement activation at that site. The advantages of targeting complement inhibition to tissue sites rather than systemically inhibiting complement include reduced infection risk and preserving the physiological activities of complement at non-targeted sites [42, 43]. For kidney disease in which complement plays a primary role in kidney damage, such as C3G, or is an important contributor to kidney injury (e.g. IgA nephropathy), targeting complement inhibition to glomeruli would seem ideal. Our data provide strong preclinical evidence that FHR5_1-9_FH_1-5_ is a rational approach to achieving this.
Supplementary Material
uxag015_Supplementary_Data
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Fakhouri F, Frémeaux-Bacchi V, Noël LH, Cook HT, Pickering MC. C 3 glomerulopathy: a new classification. Nat Rev Nephrol 2010, 6, 494–9.20606628 10.1038/nrneph.2010.85 · doi ↗ · pubmed ↗
- 2Pickering MC, D'Agati VD, Nester CM, Smith RJ, Haas M, Appel GB, et al C 3 glomerulopathy: consensus report. Kidney Int 2013, 84, 1079–89.24172683 10.1038/ki.2013.377PMC 3842953 · doi ↗ · pubmed ↗
- 3Roquigny J, Meuleman MS, El Sissy C, Martins PV, Meri S, Duval A, et al Acquired and genetic drivers of C 3 and C 5 convertase dysregulation in C 3 glomerulopathy and immunoglobulin-associated MPGN. Nephrol Dial Transplant 2024, 40, 1081–90.
- 4Levy M, Halbwachs-Mecarelli L, Gubler MC, Kohout G, Bensenouci A, Niaudet P, et al H deficiency in two brothers with atypical dense intramembranous deposit disease. Kidney Int 1986, 30, 949–56.2950269 10.1038/ki.1986.278 · doi ↗ · pubmed ↗
- 5Lopez-Larrea C, Dieguez MA, Enguix A, Dominguez O, Marin B, Gomez E. A familial deficiency of complement factor H. Biochem Soc Trans 1987, 15, 648–9.
- 6Rusai K, Zaller V, Szilagyi A, Kain R, Prohaszka Z, Cook HT, et al A rare case: childhood-onset C 3 glomerulonephritis due to homozygous factor H deficiency. CEN Case Rep 2013, 2, 234–8.28509298 10.1007/s 13730-013-0070-5PMC 5411554 · doi ↗ · pubmed ↗
- 7Blanc C, Togarsimalemath SK, Chauvet S, Le Quintrec M, Moulin B, Buchler M, et al Anti-factor H autoantibodies in C 3 glomerulopathies and in atypical hemolytic uremic syndrome: one target, two diseases. J Immunol 2015, 194, 5129–38.25917093 10.4049/jimmunol.1402770 · doi ↗ · pubmed ↗
- 8Jokiranta TS, Solomon A, Pangburn MK, Zipfel PF, Meri S. Nephritogenic lambda light chain dimer: a unique human miniautoantibody against complement factor H. J Immunol 1999, 163, 4590–6.10510403 · pubmed ↗