BASIC RESEARCH www.jasn.org

Complement Factor H-Related 5-Hybrid Anchor Properdin and Activate Complement at Self-Surfaces

† † Qian Chen,* Melanie Manzke,* Andrea Hartmann,* Maike Büttner, Kerstin Amann, ‡ † Diana Pauly, Michael Wiesener, Christine Skerka,* and Peter F. Zipfel*§

*Department of Infection Biology, Leibniz Institute for Natural Product Research and Infection Biology, Jena, Germany; †Department of Nephrology and Hypertension, Friedrich-Alexander University of Erlangen-Nuremberg, Germany; ‡University Hospital Regensburg, Regensburg, Germany; and §Friedrich Schiller University, Jena, Germany

ABSTRACT C3 glomerulopathy (C3G) is a severe kidney disease for which no specific therapy exists. The causes of C3G are heterogeneous, and defective complement regulation is often linked to C3G pathogenesis. Copy number variations in the complement factor H-related (CFHR) cluster on 1q32 and CFHR5 mutant proteins associate with this disease. Here, we identified CFHR5 as a pattern recognition that binds to damaged human endothelial cell surfaces and to properdin, the human complement activator. We found the two N-terminal short consensus repeat domains of CFHR5 contact properdin and mediate dimer formation. These properdin-binding segments are duplicated in two mutant CFHR5 proteins,

CFHR2-CFHR5Hyb from German patients with C3G and CFHR5Dup from Cypriot patients with C3G. Each of these mutated proteins assembled into large multimeric complexes and, compared to CFHR5, bound damaged human cell surfaces and properdin with greater intensity and exacerbated local complement activation. This enhanced surface binding and properdin recruitment was further evidenced in the mesangia of a transplanted and explanted kidney from a German patient with a CFHR2-CFHR5Hyb protein. Enhanced properdin staining correlated with local complement activation with C3b and C5b-9 deposition on the mesangial cell surface in vitro. This gain of function in complement activation for two disease-associated CFHR5 mutants describes a new disease mechanism of C3G, which is relevant for defining appropriate treatment options for this disorder.

J Am Soc Nephrol 27: 1413–1425, 2016. doi: 10.1681/ASN.2015020212

Factor H-related proteins (CFHRs) are important uremic syndrome,10–15 deficiency of CFHR plasma regulators of the human complement system.1–3 proteins and factor H autoantibody positive hemo- CFHR5 was initially identified within glomerular lyticuremicsyndrome,16–20 age-related macular immune deposits in kidney biopsies derived from degeneration,21–23 IgA nephropathy,24 as well as sys- patients with membranoproliferative glomerulo- temic lupus erythematosus.25 CNVs in the CFHR5 nephritis as well as other glomerular diseases.4 In gene are identified in patients with C3G,5–7 a diseased glomeruli where complement activation oc- curred on the glomerular surface, CFHR5 colocalized with both C3b and terminal complement complex Received February 25, 2015. Accepted August 4, 2015. 4 C5b-9, thus suggesting that CFHR5-directed com- Published online ahead of print. Publication date available at plement activation in the glomeruli is relevant for www.jasn.org. disease pathogenesis. Correspondence: Prof. Peter F. Zipfel, Department of Infection In addition, copy number variations (CNVs) in Biology, Leibniz Institute for Natural Products Research and In- the human gene cluster that includes the five CFHR fection Biology, Beutenbergstr. 11a, 07745 Jena, Germany. are defined in several human diseases including Email: [email protected] – C3 glomerulopathy (C3G),5 9 atypical hemolytic Copyright © 2016 by the American Society of Nephrology

J Am Soc Nephrol 27: 1413–1425, 2016 ISSN : 1046-6673/2705-1413 1413 BASIC RESEARCH www.jasn.org devastating kidney disease characterized by C3 deposition in Monocytic THP1 cells express properdin,28 and native, cell- glomeruli. We previously reported two related German C3G derived properdin (green fluorescence) was evenly distributed on patients, both with the same 25 kbp chromosomal segment the cell surface (Figure 1F, panels on the left). CFHR5 (red fluo- deleted, which spans from the CFHR2- to the CFHR5 gene. rescence) bound to the THP1 surface and showed a rather spotty Both patients express a CFHR2-CFHR5 hybrid protein, which distribution (Figure 1F, upper row). CFHR5 colocalized with has the first two domains (i.e., short consensus repeats 1–2 properdin at the surface, as revealed by the yellow signal upon [SCR1–2]) of CFHR2 linked to all nine SCR domains of merging the images (Figure 1F, upper row, right panel). CFHR2 CFHR5. This CFHR2-CFHR5 hybrid protein deregulates the did not bind to the THP1 surface (Figure 1F, middle row). Thus C3 convertase in the fluid phase.5 Patients of a large Cypriot CFHR5 binds the complement activator properdin and at the cell cohort express a CFHR5 mutant protein, which has the first surface CFHR5 colocalized with native, cell-derived properdin. two domains, i.e., SCR 1 and 2, duplicated.6,7 Thus both the CFHR2-CFHR5 hybrid and the CFHR5 mutant have the C3G-CFHR5 Mutant Proteins Efficiently Bind Properdin N-terminal interaction segment duplicated, indicating a func- Mutant CFHR5 proteins with a duplicated N-terminal dimer- tional role of these duplicated interaction segments in disease ization region, derived either from CFHR2 or CFHR5, are pathology. Therefore, it is of interest to address the function of associated with C3G, and this suggests an important role of CFHR5 on target surfaces and to define the pathogenic mech- CFHR5 in C3G pathology. Todefine how these mutant proteins anism of CFHR5 mutant-associated C3G. cause pathology, both C3G-associated CFHR5 mutants were generated and recombinantly expressed. CFHR212-CFHR5 RESULTS (termed CFHR2-CFHR5Hyb)hasCFHR212 fused to intact CFHR5 and was identified in two German C3G patients; sim- CFHR5 Binds to Modified Surfaces and Enhances C3b ilarly CFHR512-CFHR5 (CFHR5Dup), which has the interac- Deposition tion segment of CFHR5 (CFHR512) duplicated, was initially Given the sequence similarities of the C-terminal domains of described in Cypriot patients. In addition, CFHR212- CFHR proteins with the surface recognition region of factor H, CFHR5Δ12 (termed CFHR2-CFHR5Var) was generated, which fi we analyzed surface binding of both CFHR5 and CFHR2. as a variant of CFHR2-CFHR5Hyb harbors the rst interaction i.e., CFHR5 derived from normal human serum (NHS), and segment, CFHR212, but lacks the second interaction seg- i.e., CFHR2 to a lesser degree, bound to the surface of nonhuman ment, CFHR512 (Figure 2A). Chinese hamster ovary (CHO) cells and to necrotic cells, but To determine how these CFHR5 mutant proteins affect fi veryweakly to intact human umbilical veinendothelial (HUVE) complement-mediated C3b surface deposition, rst properdin cells (Figure 1A, lanes 2–4). Surface binding was confirmed by binding was tested. When assayed either by ELISA or by SPR, flow cytometry using recombinant proteins. CFHR5 bound both the German and the Cypriot CFHR5 mutants bound dose-dependently to CHO and to necrotic HUVECells (Figure properdin about twice as strongly, as compared with CFHR2- 1B, Supplemental Figure 1A). CFHR5 bound to the surface of CFHR5Var or CFHR5 (Figure 2, B and C). This stronger fi necrotic human cells with 2.7-fold higher intensity than to interaction was con rmed when binding was assessed to fl CHO cells. CFHR5 bound with higher intensity than CFHR2 properdin-expressing THP1 cells, either by ow cytometry or (Supplemental Tables 1 and 2). by confocal microscopy (Supplemental Figures 3 and 4, CFHR5 attached to necrotic HUVE cells enhanced C3 Supplemental Table 6). Taken together, both C3G mutants deposition when the cells were challenged with NHS. This with the duplicated interaction segment bound with high effect was observed on the surface of HUVE cells but not for and with similar intensity to properdin. Apparently two copies CFHR5 attached to CHO cells (Supplemental Figure 1B). Thus of the N-terminal interaction region enhance properdin binding. surface-bound CFHR5 allows C3 convertase formation and To localize the region relevant for properdin binding, i.e., complement activation. In this set-up CFHR2 did not enhance additional mutants were generated, CFHR212-CFHR512, C3b deposition (Figure 1, B and C, Supplemental Tables 3 and 4). which represents exclusively the duplicated interaction seg- ments of CFHR2-CFHR5Hyb;CFHR512, as well as CFHR212, CFHR5 Binds Properdin representing exclusively the interaction segments of either Given this activating effect, we hypothesized that CFHR5 interacts CFHR5 or CFHR2. Recombinant CFHR2-CFHR5Hyb, when with properdin, the only known human complement activa- separated on a calibrated size exclusion column, eluted in tor.26,27 Indeed, when analyzed by ELISA, properdin bound to fractions with masses ranging from .700 to ca. 440 kDa immobilized CFHR5, and binding was dose-dependent (Figure (Figure 3, A and B). As monomer CFHR2-CFHR5Hyb has a 1D). CFHR2 bound with much lower intensity. In a reverse mass of 70 kDa, the large complexes are composed of more setting, CFHR5 bound to immobilized properdin (Supplemen- than ten, and the ca. 440 kDa complexes of six protein units. tal Figure 2). CFHR5::properdin interaction was also confirmed The recombinant CFHR212-CFHR512 fragment, which as a mono- by surface plasmon resonance (SPR) and CFHR5 formed strong mer has a mass of 30 kDa, eluted in fractions corresponding and stable complexes with properdin (Figure 1E). CFHR2 in the to molecular masses ranging from .700 to ca. 120 kDa, again same settings did not interact with properdin. demonstrating formation of higher-ordered complexes that

1414 Journal of the American Society of Nephrology J Am Soc Nephrol 27: 1413–1425, 2016 www.jasn.org BASIC RESEARCH

Figure 1. CFHR5 binds to modified cell surfaces and interacts with properdin. (A) Intact CHO cells (lane 2), necrotic (lane 3) and living HUVE cells (lane 4) were incubated with NHS. Bound CFHR5, CFHR2, CFHR1 and factor H (arrows) were visualized by Western blotting. (B) Cells were incubated with recombinant CFHR5 or CFHR2 (50 and 200 nM each). Bound CFHR5 (left panel) or CFHR2 (right panel) was detected with CFHR5 or CFHR2 mAb by flow cytometry. Antibody binding in the absence of proteins was used as the negative control (dashed lines, buffer). (C) CFHR5- or CFHR2-loaded necrotic HUVE cells were incubated with NHS, C3b deposition was measured. C3b deposition to unloaded necrotic HUVE cells is shown in a dotted histogram (indicated as ‘0’). C3b deposition increased on CFHR5- attached necrotic HUVEC surface (left panel). CFHR2 did not affect complement activation (right panel). (D) Properdin bound dose- dependently to immobilized CFHR5 (circles), to CFHR2 (triangles) and to BSA (asterisk) as analyzed by ELISA. Data represented mean6SEM, n=3. (E) CFHR5, but not CFHR2, bound to immobilized properdin as evaluated with SPR. (F) When analyzed by confocal microscope, properdin (green fluorescence, left panels) was identified evenly distributed at the surface of THP1 cells. Bound CFHR5 (red, middle panel) and properdin colocalize on the surface of THP1 cells (right panel, upper row). Binding of CFHR2 was not detectable (middle row). Antibody binding in the absence of CFHR5 or CFHR2 was used as the negative control (red, Co). consist of .23 to ca. 4 units (Figure 3C). Thus two tandem identified in fractions that corresponded to masses of 350–150 arranged interaction segments lead to multimer formation. In kDa (Figure 3, D and E), demonstrating that CFHR5 in plasma NHS and in patient serum #635 (HS#635), CFHR5 was also forms oligomeric complexes and that these complexes are

J Am Soc Nephrol 27: 1413–1425, 2016 CFHR5 Anchors Properdin 1415 BASIC RESEARCH www.jasn.org

Figure 2. C3G-associated CFHR5 mutants bind more strongly to properdin. (A) Domain composition of C3G-associated CFHR5 mutants, i.e., CFHR2-CFHR5Hyb and CFHR5Dup, a genetically engineered, truncated variant (CFHR2-CFHR5Var) that has CFHR512 of the CFHR2-CFHR5Hyb deleted and CFHR5. (B) Binding of properdin to immobilized, recombinant CFHR5 disease proteins (CFHR2- CFHR5Hyb and CFHR5Dup), CFHR2-CFHR5Var and to CFHR5 was analyzed by ELISA. Data represented mean6SEM, n=3. (C) Binding of CFHR2-CFHR5Hyb and CFHR5Dup, CFHR2-CFHR5Var and CFHR5 to immobilized properdin was evaluated by SPR. The two disease- associated CFHR5 mutants bind strongly to immobilized properdin, whereas CFHR2-CFHR5Var and CFHR5 bind with lower intensity. composed of between six to three units. In patient serum, properdin matrices (Figure 4B). In this set-up both C3G mutants CFHR2-CFHR5Hyb was identified with a profile that ranged enhanced C3b surface deposition by about 40%. Thus the from .800 to ca. 150 kDa, confirming formation of higher- CFHR5 C3G mutants anchor properdin efficiently, form potent ordered, multimeric complexes (Figure 3E). C3 convertases and aggravate C3b opsonization. To map the domains of CFHR5 and CFHR5 mutants CFHR2-CFHR5Hyb derived from HS#635 bound strongly relevant for properdin interaction, binding of the deletion to CHO cells and to necrotic HUVE cells (Supplemental mutants to immobilized properdin was analyzed by SPR Figure 5A), therefore the aggravating effect of the recombinant (Figure 3F). CFHR512, but not CFHR212, bound to immobi- CFHR5 mutants for surface-mediated C3b opsonization was lized properdin. Both CFHR512 and CFHR212 are reported to evaluated on the CHO surface. Compared with CFHR5, both form dimers,29 but the related protein fragments differed in CFHR5 mutants, which harbor two interaction segments, properdin binding. Importantly, CFHR212-CFHR512,the bound to CHO cells with 2.5-fold higher intensity and upon truncated variant that represents the duplicated interacting NHS challenge, C3b deposition was augmented almost 2 fold segment CFHR2-CFHR5Hyb of the German hybrid and which (1.83). CFHR2-CFHR5Var, with one interaction segment formed higher-ordered protein complexes, bound to proper- bound with a similar intensity as CFHR5, and both CFHR2- din with rather high intensity. Thus multimer formation of CFHR5Var and CFHR5 had a comparable surface-activating both CFHR5 mutants augmented properdin binding and effects (Supplemental Figure 5, B and C). These findings show these effects highlight the pathologic role of enhanced pro- that mutant proteins with duplicated interaction regions aggra- perdin interaction for C3G pathogenesis. vate local complement activation and surface opsonization. Upon NHS challenge CFHR2-CFHR5Hyb–loaded CHO C3G-Mutants Aggravate C3 Convertase Action cells showed enhanced C3b deposition, as compared with cells Surface matrices formed by either CFHR2-CFHR5Hyb or which had CFHR5 attached or to unloaded cells (Figure 4C, CFHR5Dup allowed efficient C3b convertase formation Supplemental Table 7). This increase in effect by 86% is pro- (Figure 4A). Both disease mutants enhanced convertase formation perdin mediated. In the absence of properdin, upon challenge about 2.5 fold compared with CFHR5 or CFHR2-CFHR5Var. with properdin-depleted human serum HSDp C3b deposition Similarly C3b deposition was increased on CFHR5-C3G-mutant:: was reduced and CFHR2-CFHR5Hyb did not influence surface

1416 Journal of the American Society of Nephrology J Am Soc Nephrol 27: 1413–1425, 2016 www.jasn.org BASIC RESEARCH

Figure 3. CFHR2-CFHR5Hyb forms large multimeric complexes. (A) Size exclusion chromatography was used to determine the size of selected CFHRs and mutants in solution. The elution profile of the marker proteins of 667, 443, 150 and 29 kDa on a Superdex 200 column is indicated. (B, C) Recombinant CFHR2-CFHR5Hyb and the duplicated N-terminal interaction segments (CFHR212-CHFR512) were separated on the calibrated column, collected fractions were separated by SDS-PAGE and the proteins were identified by

Western blotting. Recombinant CFHR2-CFHR5Hyb eluted in fractions #9 to #12, corresponding to a mass of .700–440 kDa. CFHR212- CHFR512 appeared in fractions #9 to #15, corresponding to molecular masses ranging from .700 to ca. 120 kDa. (D, E) NHS and serum from patient #635 (HS#635) were separated on the same column. CFHR5 was identified in fractions #12 and #13, corresponding to amassofca.350–150 kDa. The CFHR2-CFHR5Hyb in HS#635 was identified in fractions #8 to #14, corresponding to mass ranging from .800 to ca. 150 kDa. Results are representative of three independent experiments. (F) Binding of the N-terminal fragments, i.e.,

CFHR512-CFHR512,CFHR512, CFHR212 to immobilized properdin was evaluated by SPR. CHFR512,butnotCFHR212, bound to pro- perdin. CFHR512-CFHR512, which forms multimers, bound more strongly to properdin. Full-length CFHR5 and CFHR2 were analyzed for comparison (dashed lines). activity (Figure 4D, Supplemental Table 7). However, when pathology, deposition of diseased CFHR5 and of properdin properdin was reconstituted (HSDp+P) again, C3b opsoniza- was analyzed in an explanted kidney of patient #638. Clinical tion increased 2 fold (Figure 4E). CFHR5 attached to intact information and the development of disease of the German nonhuman CHO cells did not influence complement action in C3G patient was described in detail previously.5 Presence of this set up. Thus the disease mutant enhance C3 convertase CFHR2-CFHR5Hyb was first evaluated in kidney biopsies of formation and complement activation on the cell surface. patient #638 (Figure 5A) and of the sibling #635 (Supplemen- tal Figure 6A), which were taken in the years 1999 and 2000, Diseased CFHR5 Protein, Properdin and Complement respectively. In addition, CFHR5 reactivity as well as proper- Components Deposition in Glomeruli of C3G Patient din reactivity was evaluated in the explanted renal transplant #638 of patient #638, which was removed after 8 months. Biopsies To analyze whether the CFHR2-CFHR5Hyb::properdin inter- revealed strong reactivity of CFHR5 antiserum, indicating action has relevance in vivo and is of interest for disease deposition of the hybrid protein CFHR2-CFHR5Hyb in the

J Am Soc Nephrol 27: 1413–1425, 2016 CFHR5 Anchors Properdin 1417 BASIC RESEARCH www.jasn.org

Figure 4. CFHR5-associated disease mutants aggravate complement activation on surfaces. (A) Immobilized CFHR2-CFHR5Hyb, CFHR5Dup, CFHR2-CFHR5Var, CFHR5, CFHR2 or BSA were used to anchor properdin, then C3b, factor B and factor D were added and C3 convertase formation was evaluated by determining attached Bb fragments (filled columns). In the presence of properdin, disease- associated mutant proteins enhanced Bb generation (columns 1 and 3). In the absence of properdin (open columns), the two disease- associated CFHR5 mutant proteins had a minor effect on Bb formation (columns 2 and 4). (B) C3 convertase-mediated C3b opsonization was measured after adding C3 as the substrate. Matrices consisting of one disease-associated protein together with properdin aggravate C3b deposition (columns 1 and 3). Results represented mean6SEM, n=3. (C) C3b opsonization of untreated CHO cells challenged with

NHS (left panel) is shown. When CFHR2-CFHR5Hyb was attached to the CHO surface, C3b deposition was enhanced (middle panel). (D) When CHO cells were challenged with properdin-depleted human serum (HSDp), C3b surface deposition was reduced (left panel, gray line versus NHS [dotted]). In this case, CFHR2-CFHR5Hyb (middle panel) had no enhancing effect. (E) Untreated CHO cells, when challenged with properdin reconstituted serum (HSDp+P), had more C3b deposited onto their surface (left panel, gray line versus HSDp [dotted]).

1418 Journal of the American Society of Nephrology J Am Soc Nephrol 27: 1413–1425, 2016 www.jasn.org BASIC RESEARCH kidney. Reactivity was found in the mesangial region also in represent emerging complement inhibitors, however the exact the glomerular basement membrane (Figure 5A). In the ex- role of each CFHR protein in the complement cascade needs planted transplant, prominent staining of the mesangial re- to be defined. CFHR1, CFHR2 and CFHR3 inhibit comple- gion was revealed for CFHR5 (Figure 5B). Thus, it strongly ment action at the level of C3 convertase and C5 convertase, 30–32 demonstrates CFHR2-CFHR5Hyb deposition in the glomeru- and they also block C5b-9 formation. Here, we identify lar mesangia. No reactivity to mAb-CFHR5 was detected in CFHR5 as a new pattern recognition protein, which attaches kidney biopsies prepared from nonneoplastic renal paren- to necrotic human endothelial cells. Surface-bound CFHR5 chyma from tumor-nephrectomies (Supplemental Figure 6B). anchors properdin and thereby enhances local complement Similarly properdin was detected mainly in the mesangial re- activation. gion when cryopreserved material from the explanted kidney CFHR5 is expressed in plasma at levels of ca. 3–10 mg/ml. was used (Figure 5C), and not in a control kidney (Supplemental CFHR5 deposition increased in the damaged kidney,4 was lo- Figure 6B, bottom panel). These data show CFHR2-CFHR5Hyb calized in glomeruli of C3G patients and was a prominent and properdin deposition in the mesangial region of diseased component in glomerular deposits.4,33 Based on these aspects glomeruli. Furthermore, C3c and C5b-9 deposition occurred of we proposed a regulatory role of this protein at surfaces. Here the explanted kidney (Figure 5, D and E) confirming local C3- we show that CFHR5 binds to nonhuman CHO cells, to ne- and C5-convertase activity and formation of the terminal path- crotic human endothelial cells and to TNF-a–stimulated way complexes. Taken together, CFHR2-CFHR5Hyb and properdin MMCs, but not to intact, nondamaged human cells. reactivity together with positive C3b and C5b-9 staining in the CFHR5 has a modular structure. The N-terminal SCRs, i.e., 29 explanted kidney of patient #638 reveals a role of CFHR2- CFHR512, include the dimerization segment. Heparin- and CFHR5Hyb in disease pathology that shows enhanced properdin CRP-binding regions have previously been localized within deposition in the diseased kidney as well as local complement the middle region i.e., SCR4567,34 and may also be contained activation in vivo. This strong complement activation in the kid- in the C-terminal region, i.e., CFHR589. We show here for the ney biopsies and in the explanted transplant can explain the fact first time that the N-terminal dimerization region binds to that deposition of the hybrid protein and interaction with pro- properdin. Interaction with properdin seems specific for the perdin can cause glomerular damage that leads to loss of renal N-terminal region of CFHR5, as the highly related region function and ESRD. CFHR212 does not bind properdin. Binding of the disease-associated CFHR5 mutant proteins, Properdin, the only known complement activator, forms 26,35 i.e., CFHR2-CFHR5Hyb and CFHR5Dup, to mesangial cells was dimers, trimers and tetramers. Similarly CFHR5 also evaluated in vitro using murine mesangial cells (MMCs) stim- forms higher-ordered oligomeric complexes in plasma that ulated with recombinant human TNF-a for 24 hours. Both are composed of three to six units. Thus, CFHR5::properdin CFHR2-CFHR5Hyb and CFHR5Dup bound with about 2 fold oligomers have a rather complex stoichiometry. The strength higher intensity to the stimulated cells, as compared with un- of the CFHR5::properdin complex is highly influenced by the treated MMCs (Figure 5F, Supplemental Figure 7C). In addition multimeric state of the CFHR5 protein. This was evidenced as CFHR5 bound to stimulated and to nonstimulated MMCs, but the C3G-associated German and Cypriot CFHR5 mutants, CFHR2-CFHR5Hyb and CFHR5Dup bound with almost 80% which both have tandem-arranged N-terminal domains that higher intensity. NHS challenge of TNF-a–stimulated MMCs bind properdin with high intensity. The duplicated N-terminal resulted in prominent C3b- as well as C5b-9 deposition. For dimerization region is relevant for the interaction, as is the TNF-a–stimulated MMCs loaded with CFHR2-CFHR5Hyb or CFHR212-CFHR512 fragment, which exclusively represents the CFHR5Dup NHS treatment increased surface complement acti- two interaction segments of the German hybrid CFHR2- vation about 2 fold. In this case more C3b, as well as C5b-9, was CFHR5Hyb, for forming large multimers and the complex deposited onto the cell surface, as compared to cells with at- exposes multiple contact sites for properdin. The fact that multi- tached CFHR5. This in vitro complement surface activation is mer formation is relevant for enhanced properdin recruitment is in good agreement with the enhanced C3b and C5b-9 deposition highlighted by the engineered variant CFHR2-CFHR5Var.This observed in the kidney of patient #638. variant, with the second interaction segment deleted, bound to properdin with similar intensity as CFHR5 and with lower intensity compared with both disease mutants. DISCUSSION Properdin recruitment is essential for CFHR5 disease mutant-mediated complement deregulation on target surfaces. CNVs in the human CFHR gene locus have been linked to CFHR5::properdin as well as CFHR5 disease mutant::pro- C3G, as well as to other human disorders. CFHR proteins perdin complexes allow C3 convertase assembly, initiating

For cells loaded with CFHR2-CFHR5Hyb C3b opsonization was enhanced (middle panel). In these three settings CFHR5 had no enhancing effects (C–E, right panels).

J Am Soc Nephrol 27: 1413–1425, 2016 CFHR5 Anchors Properdin 1419 BASIC RESEARCH www.jasn.org

Figure 5. CFHR5 and properdin deposition in the explanted kidney of patient #638 and effect of CFHR5 disease variants on com- plement activation on the surface of TNF-a–treated mesangial cells in vitro. (A) Deposited CFHR2-CFHR5Hyb and CFHR5 were detected on basement membrane and in mesangium by immunohistochemistry (4003) with mAb-CFHR5 in kidney biopsy of patient #638 taken in 2000. (B) Deposition of CFHR2-CFHR5Hyb and CFHR5 was detected in mesangium by immunohistochemistry (4003) with

1420 Journal of the American Society of Nephrology J Am Soc Nephrol 27: 1413–1425, 2016 www.jasn.org BASIC RESEARCH a role in local complement activation at surfaces. CFHR5 convertase formation and local complement action. Mutant surface interaction is specificformodified self-surfaces or CFHR5 proteins equipped with two interaction segments form tissues. CFHR5 did not bind to intact human cells, thus not large multimeric complexes, expose multiple interaction sites allowing inappropriate complement activation on such and attract properdin more efficiently, thereby aggravating local intact self-surfaces. This activation on damaged cells may complement action which ultimately causes pathology. These direct a further immune response of attracting immune results define a novel mechanism for CFHR5-associated C3G. effector cells for clearance. Mutant CFHR5 proteins form large multimers, bind efficiently Both CFHR5 disease variants harbor two copies of the N- to surfaces and enhance properdin recruitment. This effect is terminal interaction segments. The hybrid protein has the related to that described for properdin and factor H-targeted dimerization region of CFHR2 (CFHR2-SCRs1–2) linked to mice.36,37 Disrupting the multimeric CFHR5 mutant com- that of CFHR5 (CFHR5-SCRs1–2) and the duplicated CFHR5 plexes provides a new therapeutic option that may reduce or variant has two tandem copies of the CFHR5 dimerization even block surface-targeted complement activation in CFHR5- segment (i.e., CFHR5-SCRs1–2-SCRs19–29). Apparently associated C3G patients. such a duplication of the dimerization segments allows for- mation of large multimeric protein complexes and these large protein multimers attach strongly or more efficiently to cellu- CONCISE METHODS lar surfaces, in particular to modified cell surfaces. Thereby the attached proteins recruit higher amounts of properdin, thus Protein Expression and Purification explaining stronger complement activation at the cell surface, The constructs of CFHR5 and the German hybrid CFHR2-CFHR5Hyb as demonstrated by C3b and C5b-9 surface deposition. Thus in the expression vector PcDNA4/To/myc-His B (Invitrogen) were multimer formation enhances properdin recruitment and generated as described previously.5 The Cypriot construct CFHR5Dup oligomeric CFHR5 mutants/complexes substantially ampli- was generated by insertion of CFHR5 exons II–III into the CFHR5

fied the activating effect and C3b surface deposition. construct with primers CFHR512-F/CFHR512-R. A truncated con- Deposition of the mutant protein, i.e., CFHR2-CFHR5Hyb struct representing the German hybrid protein with only one inter- and properdin, was also identified in renal biopsies from the action segment, the interaction segment of CFHR2 and lacking the two German C3G patients. Deposited CFHR2-CFHR5Hyb and segment of CFHR5, i.e., CFHR2-CFHR5Var, was generated by deleting properdin were mainly located in the mesangium and the exon IV–VintheCFHR2-CFHR5Hyb construct with primers CFHR2– similar staining patterns support the in vitro findings that 5del34-F/CFHR2–5del34-R. CFHR2-CFHR5Hyb binds to modified cell surfaces and re- The N-terminal fragments, CFHR212-CFHR512 and CFHR512, cruits properdin at sites of damage. This allowed C3b and were amplified and inserted into the PcDNA4/To/myc-His B expres-

C5b-9 deposition in the mesangial region as shown in the bi- sion vector using the primers CFHR5N-F/CFHR5N-R. Restriction opsy, as well as increased C3b and C5b-9 deposition on the sites for endonucleases KpnI and XbaI (New England Biolabs) are surface of mesangial cells in vitro.Thein vivo evidence dem- underlined. Primers are listed in Supplemental Table 8. onstrates that CFHR5 mutants assist in glomerular surface Recombinant CFHR5 and CFHR5-associated mutants, and the damage and amplify local C3 convertase, and likely also C5 N-terminal fragments, were expressed in HEK293 cells by transient convertase action, which ultimately causes severe kidney transfection with polyethylenimine (jetPEI, Polyplus).5 Secreted pro- damage and injury. This provides novel insights into surface teins were purified from the culture supernatant by affinity chroma- dysregulation in C3G pathology and explains why both patients tography. Supernatants were applied onto a HiTrap TALON crude lost renal function and reached ESRD. (GE Healthcare) with buffer A (10 mM sodium phosphate, 10 mM In summary, CFHR5 is a new pattern recognition pro- sodium hydrogen phosphate, 500 mM NaCl, 5 mM imidazole, tein and a novel properdin ligand. CFHR5 binds to modified pH 7.8); after gradient washes, bound proteins were eluted with self-surfaces and anchors properdin, which results in targeted C3 buffer B (10 mM sodium phosphate, 10 mM sodium hydrogen

mAb-CFHR5 in the explanted kidney of patient #638 taken in 2011. (C) Properdin deposition in the mesangium, identified in cryopreserved material that was stained with pAb- or mAb-properdin by immunofluorescence (bottom panel). (D) C3c immunohistochemistry (4003) reveals intense mesangial reactivity in the glomerulus. (E) C5b-9 reactivity in the explanted kidney (4003) with mesangial positivity. Original magnifications are indicated in brackets. (F) Binding of CFHR5 disease variants to the surface of TNF-a–treated MMCs. TNF- a–treated MMCs that were loaded with CFHR2-CFHR5Hyb,CFHR5Dup, CFHR5 or CFHR2 and after extensive washing the bound proteins were detected by CFHR5- or CFHR2-reacting mAbs by flow cytometry. Antibody binding in the absence of proteins was used as the negative control (dashed lines, buffer). (G, H) CFHR5 disease variants loaded onto the surface-activated MMCs enhance C3b/C3c and

C5b-9 deposition. TNF-a activated MMCs loaded with CFHR2-CFHR5Hyb,CFHR5Dup, CFHR5 or CFHR2 were challenged with comple- ment active NHS. As controls, cells were treated with BSA. Following incubation for 20 minutes, C3b/C3c deposited onto the cell surface was evaluated by flow cytometry using a C3b/C3c antibody and Alexa 488 rabbit antiserum (G). C5b-9 deposited onto the cell surface was evaluated by flow cytometry, using a C5b-9 reacting antibody and Alexa 647 mouse antiserum (H). Data represented median fluorescence intensity 6SEM, n=3.

J Am Soc Nephrol 27: 1413–1425, 2016 CFHR5 Anchors Properdin 1421 BASIC RESEARCH www.jasn.org phosphate, 500 mM NaCl, 250 mM imidazole, pH 7.4). Purified antiserum by flow cytometry. Cells that were incubated with proteins were concentrated with Amicon Ultra-15 filters (Millipore MgEGTA buffer in the absence of a CFHR proteinwere used as antibody Corporation) and dialyzed against Dulbecco’s phosphate buffered control. A similar approach was used for intact CHO cells. The effects of saline (DPBS; Lonza). Recombinant CFHR2, and the N-terminal frag- the CHO cells loaded with German and Cypriot CFHR5 mutants on 31 ments (CFHR212), were expressed in the Pichia system. C3b deposition upon incubation with NHS were tested in the same way,

comparing with CFHR2-CFHR5Var and CFHR5. – Binding of Serum-Derived CFHR Proteins to Cell Similarly, C3b deposition on CFHR2-CFHR5Hyb loaded CHO Surfaces cells was measured when cells were challenged with complement CHO (ACC 110) cells, untreated, and also necrotic HUVE cells, active, properdin-depleted human serum (HSDp). CFHR2-

American Type Culture Collection CRL-1730) were incubated either CFHR5Hyb or CFHR5 (each at 200 nM) was bound to the surface of in NHS (pooled from five healthy donors) or in serum from patient CHO cells as described above. Following incubation in HSDp(10%), #635 (each at 25%) at 37°C for 30 minutes. After intensive washing, C3b deposition was quantitated with C3b mAb by flow cytometry. the cells were lysed, the lysates were separated by SDS-PAGE, trans- A similar experiment was performed with properdin reconstituted ferred to nitrocellulose membrane (Waterman) and bound CFHR serum (HSDp+P), i.e., HSDp to which purified properdin (5 mg/ml, proteins were visualized via Western blotting with a factor H/CFHR Comptech) was added. reacting-serum (polyclonal CFHR1 antiserum). Necrotic HUVE cells TNF-a–stimulated MMCs were loaded with CFHR2-CFHR5Hyb, were produced by heating the cell suspension at 65°C for 1 hour. CFHR5Dup, CFHR5 or CFHR2 (each at 50 and 200 nM). NHS (10%) diluted in MgEGTA buffer was added to protein-coated, TNF- Binding of Recombinant CFHR Proteins to Cell Surfaces a–stimulated MMCs. The cells were incubated at 30°C for 20 minutes. Recombinant CFHR5 and CFHR2, each at 50 and 200 nM, were After intensive washing, deposited C3b/C3c, as well as C5b-9, were incubated with CHO cells in 0.53 PBS buffer at 30°C for 30 minutes. detected with a C3b/C3c reacting pAb (Dako) in combination with After washing, bound CFHR proteins were detected by flow cytom- Alexa 488 rabbit antiserum or a C5b-9 reacting mAb (Dako) in com- etry with appropriate mAbs: CFHR5 mAb (R&D Systems) or CFHR2 bination with Alexa 647 mouse antiserum by flow cytometry. Cells mAb (generated in-house) followed by Alexa 647 mouse antiserum that were incubated with MgEGTA buffer in the absence of any CFHR (Life Technologies). Binding of CFHR5 or CFHR2 (50 and 200 nM) protein were used as antibody control. to necrotic HUVE cells was prepared using a similar approach and detected with CFHR5 mAb or CFHR2 mAb. The binding assay for CFHR Binding to Properdin by ELISA CFHR5-associated disease proteins CFHR2-CFHR5Hyb and To assay binding of CFHR proteins to immobilized properdin: properdin

CFHR5Dup, the truncated variant CFHR2-CFHR5Var to CHO cells, (5 mg/ml) was coated onto MaxiSorp microtiter plates (F96 Maxisorb, as well as to untreated and TNF-a–stimulated (25 ng/ml, 24 hours) Nunc-Immuno Module), after blocking with DPBS containing 2% BSA, MMCs was performed similarly. recombinant CFHR5, CFHR2 or BSA (Sigma-Aldrich), each at 50 nM, Properdin expression on the surface of premonocytic human were added to each well and the mixture was incubated at 37°C for THP1 cells was detected using a Zeiss LSM 710 confocal microscope 1 hour. Following washing, bound CFHR5 or CFHR2 were quantitated and ZEN 2009 program (Carl Zeiss) with properdin pAB (Comptech), with CFHR5 mAb or CFHR2 mAb in combination with the appropriate together with Alexa 488 goat antiserum (Life Technologies). Binding secondary antiserum. OD values were measured at 450 nm. assays with recombinant CFHR5 and CFHR2 (1000 nM) to THP1 cells To assay properdin binding to immobilized CFHR proteins: were performed as described above and following washing, bound recombinant CFHR5, CFHR2 or BSA, each at 50 nM, was coated CFHR2 or CFHR5 was identified with appropriate mAb CFHR2 or onto MaxiSorp microtiter plates. After blocking with DPBS containing mAb CFHR5 together with Alexa 647 mouse antiserum by confocal 2% BSA, properdin (0–200 nM) was added to each well and the mixture microscope. Binding of CFHR5 mutants, CFHR2-CFHR5Var and was incubated at 37°C for 1 hour. Following washing, bound properdin CFHR5 (each at 50 and 200 nM) was analyzed with CFHR5 mAb was quantitated with properdin pAB in combination with the appro- by flow cytometry as described above. Colocalization of CFHR2- priate secondary antiserum. OD values were measured at 450 nm. To

CFHR5Hyb (200 nM) with properdin on THP1 cells was detected directly compare the binding intensity of properdin to the two CFHR5 by confocal microscopy using a similar approach, and CFHR5 and mutants, CFHR2-CFHR5Var, and to CFHR5, each of the four proteins CFHR2 (each at 200 nM) were analyzed for comparison. was coated at 50 nM on the microtiter plates, after addition of pro- perdin (5 mg/ml) and incubation as above, bound properdin was Complement Activation on the Cell Surface detected with properdin pAb. RecombinantCFHR5andCFHR2,eachat50and200nM, were loaded onto necrotic HUVECells as described above. NHS (10%) diluted in Real-Time Interaction by Surface Plasmon Resonance 1 HEPES buffer containing EGTA and Mg2 (MgEGTA buffer: 20 mM Interactions of CFHR proteins with properdin were analyzed using HEPES, 144 mM NaCl, 7 mM MgCl2, 10 mM EGTA, pH 7,4) was the BiaCore 3000 instrument (BiaCore AB). Briefly, purified properdin added to CFHR5-, or CFHR2-loaded necrotic, as well as to untreated was immobilized via standard amine coupling (in 10 mM acetate buffer, necrotic HUVECells. The cells were incubated at 30°C for 30 minutes. pH 5.0) to the flow cells of a CMD 500M sensor chip (Xantec) until an After intensive washing deposited C3b was detected with a C3b re- appropriate level of coupling was reached. Ligands (CFHR5, CFHR2, acting mAb (Fitzgerald) in combination with Alexa 647 mouse CFHR5 associated disease mutants and CFHR2-CFHR5Var,eachwas

1422 Journal of the American Society of Nephrology J Am Soc Nephrol 27: 1413–1425, 2016 www.jasn.org BASIC RESEARCH calculated as a monomer of 400 nM) in DPBS were flowed onto October 2010. He reached his best creatinine level of 1.13 mg/dl properdin-coupled chip or onto a blank control chip at 5 ml/min at 5 months after transplantation. Six months after transplantation

25°C. Interactions of N-terminal segments (CFHR212-CFHR512, transplant failure occurred, when the first rejection episode was ver-

CFHR512, CFHR212) with properdin were also analyzed using the ified upon biopsy, which was formerly classified as cellular, type same approach. CFHR512 and CFHR212 were calculated as dimers, BANFF IA. Despite aggressive rejection therapy including several and CFHR212-CFHR512 was calculated as a tetramer based on the cali- courses of pulse prednisolone, antithymocyte globuline and plasma- brated minimum size. pheresis over the following months, the rejection was refractory to all interventions. The patient had to return to regular dialysis 7 months Surface-Bound C3 Convertase Assembly and after transplantation. Eight months after transplantation (June 2011) Amplification the transplant was explanted because of continued inflammation and Recombinant disease mutants, CFHR2-CFHR5Var, CFHR5, CFHR2 exceeding immunosuppressive therapy. or BSA, each at 50 nM, were coated onto the MaxiSorp microtiter Native renal biopsies (formalin-fixed paraffin-embedded) were plates at 4°C overnight. After blocking, properdin (5 mg/ml) was performed in kidney biopsies of patients #635 and #638, obtained in added to each well and incubated for 1 hour at 37°C. C3b (10 mg/ml), 1999 and 2000, and tissue of a renal transplant (formalin-fixed factor B (5 mg/ml), factor D (2 mg/ml) mixed in MgEGTA buffer were paraffin-embedded and cryopreserved) which was explanted in 2011 added to each well and incubated for 20 minutes at 37°C. After from patient #638 were investigated. Two or three kidney biopsies with intensive washing, attached Bb fragments were quantitated with factor dense deposit disease and four samples of nonneoplastic renal B pAb (Comptech). OD values were measured at 450 nm. The same parenchyma from tumor nephrectomies served as controls. Primary assays were performed in the absence of properdin. antibodies used for immunohistochemistry were as follows: CFHR5 Tomeasure the effect of the CFHR5 on C3 convertase amplification (1:500, mouse, monoclonal; R&D Systems), C3c (1:75,000, rabbit and action, disease mutants, CFHR2-CFHR5Var, CFHR5, CFHR2 or polyclonal; Dako), C5b-9 (1:1000, mouse, monoclonal, clone BSA, each at 50 nM, were immobilized onto the plates as described. C5b9; Qidel). Stainings for C3c and C5b9 were performed on a Ven- Next properdin (5 mg/ml) was attached to the immobilized proteins, tana BenchMark Ultrastainer immunohistochemistry device (Roche, then C3b (2 mg/ml), factor B (10 mg/ml), factor D (2 mg/ml), together Grenzach-Wyhlen, Germany). For the detection of CFHR5 Vector with C3 (20 mg/ml), all mixed in MgEGTA buffer, were added to ABC-Kit (Vector Laboratories) was applied and ImmPACT DAB the wells and the mixture was incubated for 30 minutes at 37°C. Vector (Vector Laboratories) was used as a chromogen. Properdin Following washing, deposited C3b was quantitated using C3b mAb. immunofluorescence was performed on cryopreserved tissue of the OD values were measured at 450 nm. above-mentioned explanted renal transplant and on normal renal parenchyma of a tumor nephrectomy used as a control. Two primary Size Exclusion Chromatography antibodies specific for properdin (1:200, mouse monoclonal, 134038 Constructs of CFHR2-CFHR5Hyb and the N-terminal truncated frag- and 1:200, goat, polyclonal) were used. A donkey anti-mouse Alexa ment CFHR212-CHFR512 were transfected to HEK293 cells and ex- 568 and a donkey anti-goat Alexa 488 antibody (both Life Technol- pressed transiently. Supernatants (20 ml) of CFHR2-CFHR5Hyb or ogies), respectively, were applied as secondary antibodies.

CFHR212-CHFR512 expressing cells were collected 96 hours post- transfection, and concentrated to 1 ml with Amicon Ultra-15 filters. Responses to Cytokine Stimulation of the MMCs 500 ml of each supernatant was applied on a calibrated Superdex 200 The MMCs39 were seeded into 24-well plates in RPMI medium sup- 10/300 GL column (GE Healthcare) and separated against DPBS plemented with 10% heat inactivated FBS (F7524; Sigma-Aldrich) using a flow rate of 0.5 ml/min. Fractions were collected and aliquots and incubated for 24 hours to reach 60–70% confluence. Then the of each fraction were separated by SDS-PAGE, then transferred to a medium was replaced by RPMI FBS (10%) supplemented with TNF-a membrane and proteins were visualized by Western blotting using (ImmunoTools, 0–100 ng/ml).40 Following incubation for 24 hours CFHR5 pAb (R&D Systems). The column was calibrated with the at 37°C, the culture supernatants were collected and IL-6 levels were filtration calibration kit (GE Healthcare) to evaluate the elution pro- determined by ELISA (eBioscience). The kinetic for IL-6 expression file of the four marker proteins and similarly to evaluate the mass of was evaluated by incubating MMCells at 37°C in RPMI medium sup- the sample proteins. plemented with FBS (10%) and TNF-a (25 ng/ml) for varying times In addition NHS and also serum from patient #635 (500 ml ranging from 0.5 to 72 hours (Supplemental Figure 7, A and B). straight) was applied to the same calibrated column, separated and processed as described above. A combination of factor H/CFHR an- Statistical Analyses tiserum and CFHR5 pAb was used to detect factor H, CFHR1, Statistical analysis of the data was done by t-test for unpaired data.

CFHR2, CFHR5 and CFHR2-CFHR5Hyb. Data in Figure 1, A–C, E and F), Figure 2C, Figure 3, Figure 4, C–E) and Figure 5F) are representative of three independent experiments. Detection of CFHR5, Properdin and Complement Components in Paraffin-Embedded and Cryopreserved Study Approval Human Renal Specimens This study was approved by the Ethical Board of the Friedrich Schiller Patient #638 described recently5 reached ESRD at age 23, in 2006, and University, Jena, Germany. The patients were informed on this study performed peritoneal dialysis until being successfully transplanted in and they provided written consent on participation.

J Am Soc Nephrol 27: 1413–1425, 2016 CFHR5 Anchors Properdin 1423 BASIC RESEARCH www.jasn.org

ACKNOWLEDGMENTS Skerka C: Deletion of complement factor H-related genes CFHR1 and CFHR3 is associated with atypical hemolytic uremic syndrome. PLoS We especially acknowledge the patients and their families for par- Genet 3: e41, 2007 11. Eyler SJ, Meyer NC, Zhang Y, Xiao X, Nester CM, Smith RJ: A novel ticipating in this study and for contributing valuable specimens. We hybrid CFHR1/CFH gene causes atypical hemolytic uremic syndrome. thank Ina Loeschmann, Nadine Flach and Steffi Hälbich for excellent Pediatr Nephrol 28: 2221–2225, 2013 technical support, for protein expression and protein analysis by 12. Venables JP, Strain L, Routledge D, Bourn D, Powell HM, Warwicker P, Dr. Hannes U. Eberhardt for discussion and Dr. Hans Martin Dahse Diaz-Torres ML, Sampson A, Mead P, Webb M, Pirson Y, Jackson MS, for support with mesangial cell work (all Department of Infection Hughes A, Wood KM, Goodship JA, Goodship TH: Atypical haemolytic uraemic syndrome associated with a hybrid complement gene. PLoS Biology, Leibniz Institute for Natural Product Research and Infection Med 3: e431, 2006 Biology). 13. Krid S, Roumenina LT, Beury D, Charbit M, Boyer O, Frémeaux-Bacchi V, Niaudet P: Renal transplantation under prophylactic eculizumab in DISCLOSURES atypical hemolytic uremic syndrome with CFH/CFHR1 hybrid protein. Am J Transplant 12: 1938–1944, 2012 This work was supported by the German Research Foundation (DFG) Zi432 14. Francis NJ, McNicholas B, Awan A, Waldron M, Reddan D, Sadlier D, and SK 46. Q.C. is supported by the International Leibniz Research School Kavanagh D, Strain L, Marchbank KJ, Harris CL, Goodship TH: A novel (Jena, Germany) and the Jena School for Microbial Communication, which hybrid CFH/CFHR3 gene generated by a microhomology-mediated is supported by the excellence initiative of the DFG. The research leading to deletion in familial atypical hemolytic uremic syndrome. Blood 119: – these results has received funding from the European Community’s Seventh 591 601, 2012 15. Józsi M, Heinen S, Hartmann A, Ostrowicz CW, Hälbich S, Richter H, Framework Programme under grant agreement no. 2012-305608 “European Kunert A, Licht C, Saunders RE, Perkins SJ, Zipfel PF, Skerka C: Factor H Consortium for High-Throughput Research in Rare Kidney Diseases and atypical hemolytic uremic syndrome: mutations in the C-terminus (EURenOmics)”. The funders had no role in study design, data collection cause structural changes and defective recognition functions. JAmSoc and analysis, decision to publish, or preparation of the manuscript. Nephrol 17: 170–177, 2006 16. Józsi M, Licht C, Strobel S, Zipfel SL, Richter H, Heinen S, Zipfel PF, Skerka C: Factor H autoantibodies in atypical hemolytic uremic syn- REFERENCES drome correlate with CFHR1/CFHR3 deficiency. Blood 111: 1512– 1514, 2008 17. Dragon-Durey MA, Loirat C, Cloarec S, Macher MA, Blouin J, Nivet H, 1. Zipfel PF, Jokiranta TS, Hellwage J, Koistinen V, Meri S: The factor H Weiss L, Fridman WH, Frémeaux-Bacchi V: Anti-factor H autoanti- protein family. Immunopharmacology 42: 53–60, 1999 bodies associated with atypical hemolytic uremic syndrome. JAmSoc 2. Skerka C, Chen Q, Fremeaux-Bacchi V, Roumenina LT: Complement Nephrol 16: 555–563, 2005 factor H related proteins (CFHRs). Mol Immunol 56: 170–180, 2013 18. Józsi M, Strobel S, Dahse HM, Liu WS, Hoyer PF, Oppermann M, Skerka 3. Zipfel PF, Skerka C: Complement regulators and inhibitory proteins. C, Zipfel PF: Anti factor H autoantibodies block C-terminal recognition Nat Rev Immunol 9: 729–740, 2009 function of factor H in hemolytic uremic syndrome. Blood 110: 1516– 4. Murphy B, Georgiou T, Machet D, Hill P, McRae J: Factor H-related 1518, 2007 protein-5: a novel component of human glomerular immune deposits. 19. Fremeaux-Bacchi V, Fakhouri F, Garnier A, Bienaimé F, Dragon-Durey AmJKidneyDis39: 24–27, 2002 MA, Ngo S, Moulin B, Servais A, Provot F, Rostaing L, Burtey S, Niaudet 5. Chen Q, Wiesener M, Eberhardt HU, Hartmann A, Uzonyi B, Kirschfink M, Amann K, Buettner M, Goodship T, Hugo C, Skerka C, Zipfel PF: P, Deschênes G, Lebranchu Y, Zuber J, Loirat C: Genetics and outcome Complement factor H-related hybrid protein deregulates complement of atypical hemolytic uremic syndrome: a nationwide French series com- – in dense deposit disease. J Clin Invest 124: 145–155, 2014 paring children and adults. Clin J Am Soc Nephrol 8: 554 562, 2013 6. Gale DP, de Jorge EG, Cook HT, Martinez-Barricarte R, Hadjisavvas A, 20. Moore I, Strain L, Pappworth I, Kavanagh D, Barlow PN, Herbert AP, McLean AG, Pusey CD, Pierides A, Kyriacou K, Athanasiou Y, Schmidt CQ, Staniforth SJ, Holmes LV, Ward R, Morgan L, Goodship Voskarides K, Deltas C, Palmer A, Frémeaux-Bacchi V, de Cordoba SR, TH, Marchbank KJ: Association of factor H autoantibodies with dele- Maxwell PH, Pickering MC: Identification of a mutation in complement tions of CFHR1, CFHR3, CFHR4, and with mutations in CFH, CFI, CD46, factor H-related protein 5 in patients of Cypriot origin with glomeru- and C3 in patients with atypical hemolytic uremic syndrome. Blood – lonephritis. Lancet 376: 794–801, 2010 115: 379 387, 2010 7. Athanasiou Y, Voskarides K, Gale DP, Damianou L, Patsias C, Zavros M, 21. Fritsche LG, Lauer N, Hartmann A, Stippa S, Keilhauer CN, Oppermann Maxwell PH, Cook HT, Demosthenous P, Hadjisavvas A, Kyriacou K, M, Pandey MK, Köhl J, Zipfel PF, Weber BH, Skerka C: An imbalance of Zouvani I, Pierides A, Deltas C: Familial C3 glomerulopathy associated human complement regulatory proteins CFHR1, CFHR3 and factor H with CFHR5 mutations: clinical characteristics of 91 patients in 16 influences risk for age-related macular degeneration (AMD). Hum Mol pedigrees. Clin J Am Soc Nephrol 6: 1436–1446, 2011 Genet 19: 4694–4704, 2010 8. Tortajada A, Yébenes H, Abarrategui-Garrido C, Anter J, García- 22. Zipfel PF, Lauer N, Skerka C: The role of complement in AMD. Adv Exp Fernández JM, Martínez-Barricarte R, Alba-Domínguez M, Malik TH, Med Biol 703: 9–24, 2010 Bedoya R, Cabrera Pérez R, López Trascasa M, Pickering MC, Harris CL, 23. Hughes AE, Orr N, Esfandiary H, Diaz-Torres M, Goodship T, Sánchez-Corral P, Llorca O, Rodríguez de Córdoba S: C3 glomerulopathy- Chakravarthy U: A common CFH haplotype, with deletion of CFHR1 associated CFHR1 mutation alters FHR oligomerization and complement and CFHR3, is associated with lower risk of age-related macular regulation. J Clin Invest 123: 2434–2446, 2013 degeneration. Nat Genet 38: 1173–1177, 2006 9. Malik TH, Lavin PJ, Goicoechea de Jorge E, Vernon KA, Rose KL, Patel 24. Gharavi AG, Kiryluk K, Choi M, Li Y, Hou P, Xie J, Sanna-Cherchi S, Men MP, de Leeuw M, Neary JJ, Conlon PJ, Winn MP, Pickering MC: CJ, Julian BA, Wyatt RJ, Novak J, He JC, Wang H, Lv J, Zhu L, Wang W, A hybrid CFHR3-1 gene causes familial C3 glomerulopathy. JAmSoc Wang Z, Yasuno K, Gunel M, Mane S, Umlauf S, Tikhonova I, Beerman I, Nephrol 23: 1155–1160, 2012 Savoldi S, Magistroni R, Ghiggeri GM, Bodria M, Lugani F, Ravani P, 10. Zipfel PF, Edey M, Heinen S, Józsi M, Richter H, Misselwitz J, Hoppe B, Ponticelli C, Allegri L, Boscutti G, Frasca G, Amore A, Peruzzi L, Coppo Routledge D, Strain L, Hughes AE, Goodship JA, Licht C, Goodship TH, R, Izzi C, Viola BF, Prati E, Salvadori M, Mignani R, Gesualdo L,

1424 Journal of the American Society of Nephrology J Am Soc Nephrol 27: 1413–1425, 2016 www.jasn.org BASIC RESEARCH

Bertinetto F, Mesiano P, Amoroso A, Scolari F, Chen N, Zhang H, Lifton and FHR-4: binding to the C3d region of C3b and differential regulation RP: Genome-wide association study identifies susceptibility loci for IgA by heparin. FEBS Lett 462: 345–352, 1999 nephropathy. Nat Genet 43: 321–327, 2011 33. Sethi S, Gamez JD, Vrana JA, Theis JD, Bergen HR 3rd, Zipfel PF, 25. Zhao J, Wu H, Khosravi M, Cui H, Qian X, Kelly JA, Kaufman KM, Dogan A, Smith RJ: Glomeruli of dense deposit disease contain com- Langefeld CD, Williams AH, Comeau ME, Ziegler JT, Marion MC, Adler ponents of the alternative and terminal complement pathway. Kidney A, Glenn SB, Alarcón-Riquelme ME, Pons-Estel BA, Harley JB, Bae SC, Int 75: 952–960, 2009 Bang SY, Cho SK, Jacob CO, Vyse TJ, Niewold TB, Gaffney PM, Moser 34. McRae JL, Duthy TG, Griggs KM, Ormsby RJ, Cowan PJ, Cromer BA, KL, Kimberly RP, Edberg JC, Brown EE, Alarcon GS, Petri MA, Ramsey- McKinstry WJ, Parker MW, Murphy BF, Gordon DL: Human factor Goldman R, Vilá LM, Reveille JD, James JA, Gilkeson GS, Kamen DL, H-related protein 5 has cofactor activity, inhibits C3 convertase activity, Freedman BI, Anaya JM, Merrill JT, Criswell LA, Scofield RH, Stevens binds heparin and C-reactive protein, and associates with lipoprotein. AM, Guthridge JM, Chang DM, Song YW, Park JA, Lee EY, Boackle SA, JImmunol174: 6250–6256, 2005 Grossman JM, Hahn BH, Goodship TH, Cantor RM, Yu CY, Shen N, 35. Pangburn MK: Analysis of the natural polymeric forms of human pro- Tsao BP; BIOLUPUS Network; GENLES Network: Association of genetic perdin and their functions in complement activation. JImmunol142: variants in complement factor H and factor H-related genes with sys- 202–207, 1989 temic lupus erythematosus susceptibility. PLoS Genet 7: e1002079, 2011 36. Lesher AM, Zhou L, Kimura Y, Sato S, Gullipalli D, Herbert AP, Barlow 26. Smith CA, Pangburn MK, Vogel CW, Müller-Eberhard HJ: Molecular PN, Eberhardt HU, Skerka C, Zipfel PF, Hamano T, Miwa T, Tung KS, architecture of human properdin, a positive regulator of the alternative Song WC: Combination of factor H mutation and properdin deficiency pathway of complement. J Biol Chem 259: 4582–4588, 1984 causes severe C3 glomerulonephritis. J Am Soc Nephrol 24: 53–65, 27. Lesher AM, Nilsson B, Song WC: Properdin in complement activation 2013 and tissue injury. Mol Immunol 56: 191–198, 2013 37. Ruseva MM, Vernon KA, Lesher AM, Schwaeble WJ, Ali YM, Botto M, 28. Cortes C, Ohtola JA, Saggu G, Ferreira VP: Local release of properdin in Cook T, Song W, Stover CM, Pickering MC: Loss of properdin the cellular microenvironment: role in pattern recognition and ampli- exacerbates C3 glomerulopathy resulting from factor H deficiency. fication of the alternative pathway of complement. Front Immunol 3: JAmSocNephrol24: 43–52, 2013 412, 2012 38. Pauly D, Nagel BM, Reinders J, Killian T, Wulf M, Ackermann S, 29. Goicoechea de Jorge E, Caesar JJ, Malik TH, Patel M, Colledge M, Ehrenstein B, Zipfel PF, Skerka C, Weber BH: A novel antibody against Johnson S, Hakobyan S, Morgan BP, Harris CL, Pickering MC, Lea SM: human properdin inhibits the alternative complement system and Dimerization of complement factor H-related proteins modulates specifically detects properdin from blood samples. PLoS One 9: complement activation in vivo. Proc Natl Acad Sci U S A 110: 4685– e96371, 2014 4690, 2013 39. Wolf G, Haberstroh U, Neilson EG: Angiotensin II stimulates the pro- 30. Heinen S, Hartmann A, Lauer N, Wiehl U, Dahse HM, Schirmer S, Gropp liferation and biosynthesis of type I collagen in cultured murine K, Enghardt T, Wallich R, Hälbich S, Mihlan M, Schlötzer-Schrehardt U, mesangial cells. Am J Pathol 140: 95–107, 1992 Zipfel PF, Skerka C: Factor H-related protein 1 (CFHR-1) inhibits com- 40. Kamanna VS, Pai R, Bassa B, Kirschenbaum MA: Activation of mesangial plement C5 convertase activity and terminal complex formation. Blood cells with TNF-alpha stimulates M-CSF and monocyte 114: 2439–2447, 2009 proliferation: evidence for involvement of protein kinase C and protein 31. Eberhardt HU, Buhlmann D, Hortschansky P, Chen Q, Böhm S, Kemper tyrosine kinase. Biochim Biophys Acta 1313: 161–172, 1996 MJ, Wallich R, Hartmann A, Hallström T, Zipfel PF, Skerka C: Human factor H-related protein 2 (CFHR2) regulates complement activation. PLoS One 8(11): e78617, 2013 32. Hellwage J, Jokiranta TS, Koistinen V, Vaarala O, Meri S, Zipfel PF: This article contains supplemental material online at http://jasn.asnjournals. Functional properties of complement factor H-related proteins FHR-3 org/lookup/suppl/doi:10.1681/ASN.2015020212/-/DCSupplemental.

J Am Soc Nephrol 27: 1413–1425, 2016 CFHR5 Anchors Properdin 1425