Discovery, in-vitro and in-vivo ecacy of an antiinammatory small molecule inhibitor of C-reactive protein

C-reactive protein (CRP) is an acute phase protein. We recently identied a novel mechanism that leads to a conformational change from the native, pentameric structure (pCRP) to a pentameric intermediate (pCRP*) and ultimately to the monomeric form, mCRP, both being highly pro-inammatory. This ‘CRP activation’ is mediated by binding of pCRP to activated/damaged cell membranes via exposed phosphocholine (PC) lipid head groups. We designed a low molecular weight pCRP – PC inhibitor, C10M. Binding assays and X-ray crystallography revealed direct, competitive binding of C10M to pCRP, blocking interaction with PC and thereby inhibiting formation of pCRP*/mCRP and their pro-inammatory effects. The anti-inammatory potential of C10M was conrmed in-vitro by various measures of leukocyte and endothelial cell activation and in-vivo in rat models of acute ischemia/reperfusion injury and hindlimb transplantation. In conclusion, inhibition of pCRP*/mCRP generation via the PC-mimicking compound C10M represents a promising, potentially broadly applicable anti-inammatory therapy. therapeutic potential of allograft rejection via aggravation of IRI and activation of the innate In both animal models, we establish C10M’s unique benets in reducing CRP-mediated tissue damage. In the model the acceleration of acute allograft rejection by is reversed by C10M. In the IRI model, renal function is signicantly improved and histological signs reduced. The deposition of in the tissue of renal IRI is signicantly reduced after C10M, conrming our in-vitro ndings that pCRP activated subsequent


Introduction
C-reactive protein (CRP) is an acute phase protein that is synthesized in the liver under the regulation of the cytokine interleukin 6 (IL-6) and circulates as a disc-shaped homo-pentamer (pCRP) 1 . Whilst pCRP itself is not pro-in ammatory when injected into healthy individuals 2 , it ampli es tissue injury in the context of in ammation and ischemia 3 . Animal models of myocardial infarction 4,5 , stroke 6 and ischemia/reperfusion injury (IRI) 7,8 have demonstrated that pCRP administration can signi cantly increase tissue injury 3 . Previously, we identi ed a novel mechanism that can be viewed as a proin ammatory 'CRP-activation' process 9,10 . In the context of in ammation and tissue damage, membrane changes on activated cell membranes mediated by phospholipase A2 11 lead to the exposure of bioactive lipids. This results in the binding of circulating pCRP and subsequent changes in CRP conformation from the pentameric structure (pCRP) to a partially dissociated pentamer (pCRP*) and ultimately in dissociation to its monomeric form (mCRP) 9,12,13 . pCRP* and mCRP are strong pro-in ammatory agents, and can induce IL-8 secretion in neutrophils 14 and human coronary artery endothelial cells 15 , promote neutrophil-endothelial cell adhesion 16 , and delay apoptosis of human neutrophils 17 . pCRP* can also bind and activate complement C1q 13 , which further contributes to aggravation of pre-existing in ammation and detrimental tissue damage 9,11 . Phosphocholine (PC) and phosphoethanolamine (PE) head groups of bioactive lipids are exposed on the surface of activated and damaged cells. Both head groups are known to bind to a shallow groove containing two Ca 2+ cations located on one face of the monomeric subunit of pCRP, hence there are 5 PC/PE binding pockets on the same face of the CRP pentamer. The concept of targeting the PC/PE binding pockets of pCRP with a synthetic ligand was rst explored by Pepys et al 4 , when they synthesized and evaluated palindromic compounds comprising two molecules of PC covalently linked through one of the phosphate oxygen atoms by a exible carbon-based linker; one such compound was 1,6-bis(phosphocholine)-hexane (bis-PC, molecular weight ~ 450 Da). It was demonstrated by X-ray crystallography that each of the PC head groups of bis-PC binds to two separate CRP pentamers, and when multiple bis-PC molecules bind they bring the PC-binding surfaces of the two pentamers together in a parallel fashion. By binding in this manner, palindromic compounds like bis-PC prevent pCRP from interacting with bioactive lipids on activated/damaged cell surfaces thereby blocking the formation of pCRP* and mCRP. As CRP has a central role in many in ammatory reactions and diseases it represents an attractive therapeutic target. Helical polypeptides with covalently attached (via a carbon-based linker) PC molecules 18 , peptide mimetics (molecular weight > 700 Da) 19 and anti-sense oligonucleotides 20 targeting CRP have also been developed, with only the latter reaching phase II clinical trials (ISIS-329993 and ISIS-353512, NCT01710852, NCT01414101 and NCT00734240) 21 . Furthermore, reducing circulating pCRP levels via CRP-apheresis using PC-linked resins is currently being investigated as an adjunct therapy to minimize cardiac injury in patients with myocardial infarction 22 . The therapeutic approaches discussed above target general pCRP inhibition and reduction of circulating pCRP levels. However, such approaches come at the cost of some level of immune suppression, which can potentially lead to signi cant side effects.
We hypothesized that we could design a low molecular weight compound that targets the PC/PE binding pocket on pCRP and thereby prevent the formation of the pro-in ammatory pCRP* and mCRP species. As proof of concept we present the data for C10M, (3-(dibutylamino)propyl)phosphonic acid, a pCRP*/mCRP inhibitor with a molecular weight of just 250 Da. After con rming that C10M bound to the PC/PE binding pocket of pCRP using X-ray crystallography and biophysical techniques, we investigated the therapeutic potential of this compound in various in-vitro, ex-vivo and in-vivo experimental models of in ammation. We analyzed its mode of action and demonstrated that C10M prevents pCRP from binding to activated or damaged cell membranes, thereby blocking the formation of the pro-in ammatory pCRP* and mCRP species. Furthermore, we tested the immunosuppressive impact of C10M on the CRP-dependent innate immune response to bacterial pathogens to preserve protective capacities against Streptococcus pneumoniae (S. pneumoniae) in order to present a targeted therapy against exacerbated in ammation. Overall, we generated and characterized a PC-mimicking compound that inhibits the generation of proin ammatory pCRP*/mCRP and represents a promising, potentially broadly applicable therapeutic for the many patients suffering from in ammatory diseases.

Design of the pCRP*/mCRP inhibitor C10M
To design a proof of concept inhibitor, we felt it was important to retain the interactions made by the PC/PE head groups of bioactive lipids to pCRP. The PC/PE binding pockets are located on what is designated as the B-face of the disc-shaped CRP pentamer 13 . Once bound, the lipid PC/PE head groups anchor the pCRP protein to the damaged cell surface, which then initiates the conformational change to pCRP* and ultimately mCRP 13 . The face of the CRP pentamer not containing the PC/PE binding pockets is known as the A-face and is not involved in binding to activated/damaged cell membranes. PC is anchored to the shallow groove on the B-face of pCRP monomers via (1) a salt bridge interaction between the positively charged quaternary amine and the negatively charged carboxyl group of Glu 81 and (2) coordination of the negatively charged phosphate moiety to the positively charged Ca 2+ cations (Fig. 1A).
In addition, pCRP residues Asn 61 and Gln 150 lie on opposite sides of the binding pocket and their sidechains participate in hydrogen bonds with oxygen atoms of the PC phosphate moiety, while the other residues lining the binding pocket contribute hydrophobic interactions only (the full list of interacting pCRP residues is given in Table S1). The low molecular weight compound C10M (250 Da, Fig. 1A) was intended to mimic the anchoring interactions of PC. To retain the critical coordination to the Ca 2+ cations, whilst removing susceptibility to serum nuclease activities, the PC phosphate moiety was replaced with a phosphonate. To take advantage of the space in the binding pocket near the PC quaternary amine (indicated by the vectors R1 and R2, Fig. 1A), we replaced the methyl substituents of PC with n-butyl groups and reverted to a tertiary amine to mitigate any steric hindrance to accessing Glu 81 that the longer n-butyl groups may cause. C10M was synthesized by a two-step synthetic method from commercially available precursors in reasonable yield (Supplementary Materials and Methods) and evaluated for its binding a nity for pCRP.

C10M binds to pCRP in-vitro
The binding of C10M to pCRP was investigated using surface plasmon resonance (SPR). The pCRP was immobilized on the sensor chip using amine coupling and different concentrations of the C10M compound were injected (Fig. 1B, top panel). The relative responses (RU), after corrections for a blank surface and buffer, indicated a binding a nity (K D ) of 105 ± 12 µM for C10M to pCRP (Fig. 1B, bottom panel). In comparison, PC (positive control compound) reported a K D of 4.0 ± 0.6 µM using the same SPR protocol (data not shown), which is consistent with previous studies by Christopeit et al. 23 . These biophysical data con rmed a direct binding of C10M to pCRP. The ability of C10M to inhibit the binding of pCRP to the physiological ligand PC was analyzed using immobilized PC-KLH by ELISA. In this assay C10M had an IC 50 of 1.5 ± 0.3 mM (data not shown). The inhibitory effect of C10M was further investigated using PC immobilized on agarose beads. C10M (43.5 µg/ml) reduced binding of pCRP (200 µg) to these beads by ~ 70 % (Fig. 1C).
To determine the mode of binding of compound C10M to pCRP, co-crystallization experiments were undertaken and the structure of the complex was solved using X-ray crystallography at a resolution of 3.5 Å (see Table S2 for statistics and Fig. S1). The asymmetric unit consisted of four CRP pentamers with the A-faces stacked against each other (Fig. 1D). Each pCRP monomeric unit adopted a similar fold to that previously published for the PC:pCRP complex (PDB ID: 1B09 24 ), with a root mean square deviation of 0.5 Å over all Cα atoms upon alignment. Only ve of the pCRP monomers within the asymmetric unit had a C10M molecule bound. The location of the C10M molecules con rmed that the compound binds to the PC/PE binding pocket, with the phosphonate moiety being the main mediator of the interaction with pCRP via the Ca 2+ cations (Fig. 1D -F Table S1). In contrast to the positively charged quaternary amine of PC, the tertiary amine of C10M does not appear to interact with Glu 81 via a hydrogen bond, rather the interactions with this residue are hydrophobic (Table S1). The n-butyl amine substituents of C10M make numerous hydrophobic contacts with pCRP residues (a full list of putative interacting pCRP residues is given in Table S1). Upon alignment of all ve pCRP monomers with a bound C10M molecule, it is evident that while the phosphonate moiety anchors the compound into the pocket by interacting with the Ca 2+ cations, there is a wide variation in the binding orientation of the C10M alkyl backbone and the n-butyl amine substituents (Figs. 1G and S1).
C10M inhibits pCRP binding to activated cell membranes Activated platelet membranes expose the PC head groups of lysophosphatidylcholine (LPC) and thereby mediate the conformational changes of CRP 9,12,25,26 . We have previously demonstrated that pCRP binds to activated cell membranes but not to membranes of healthy cells 13 . After 120 min bound to activated cell membranes, pCRP begins to dissociate to the pro-in ammatory pCRP* species; however, the proportion of pCRP* is greatest on the surface of microvesicles released by damaged cells. To investigate whether C10M binding to pCRP is su cient to inhibit the pro-in ammatory effects of CRP, we determined the ability of C10M to prevent pCRP binding to activated platelets. Binding of uorescently labeled (Atto 594, λ ex 601nm λ em 627nm, Sigma-Aldrich) human pCRP on isolated and ADP-stimulated platelets was investigated by ow cytometry (Fig. 2A). When incubated with C10M, signi cantly less pCRP-Atto 594 bound to activated platelets. Controls showed Ca 2+ -dependent binding of pCRP-Atto 594 to platelets (Fig.  S2). Flow cytometric results were visualized and con rmed by confocal laser scanning microscopy ( Fig. 2B). This was further con rmed by Western blotting of activated platelets that were incubated with either pCRP alone or pCRP + C10M. After washing steps, only a small fraction of CRP could be detected by Western blot in the presence of C10M, whereas signi cant amounts of CRP were detectable in the platelet lysates without C10M. Data were quanti ed by densitometry ( Fig. 2C and D).
C10M inhibits pCRP*/mCRP-induced monocyte adhesion and pro-in ammatory cytokine production pCRP* and mCRP have previously been demonstrated to enhance leukocyte adhesion, transmigration, and subsequent cytokine production, all of which are key events in the in ammatory cascade. Therefore, we next investigated the inhibitory effects of C10M in abrogating these pro-in ammatory properties of CRP. The potential of C10M to inhibit CRP-induced monocyte adhesion, expression of pro-in ammatory cytokines, and formation of platelet-leukocyte aggregates was evaluated in-vitro. pCRP*/mCRP was generated by incubation of ADP-stimulated platelets with pCRP as described previously 9,13 . mCRP leads to the formation of platelet-leukocyte aggregates, which can be inhibited by C10M (Fig. 3A -C). In a static monocyte adhesion assay analyzing binding of monocytes to a brinogen matrix, pCRP* induces monocyte adhesion, which can be inhibited by C10M (Fig. 3D). Notably, C10M is only reducing the pCRP*induced exacerbation of in ammation, which is represented by the increase in monocyte adhesion, but not the underlying increase in adhesion of monocytes that is induced by the ADP-activated platelets.
We further analyzed the interaction of platelet-bound pCRP*/mCRP with monocytes by intracellular staining (ICS) and ow cytometry. Cytokine expression levels of pro-in ammatory cytokines, as measured by ICS, were found upregulated in monocytes (Fig. 3E). Tumor necrosis factor (TNF), IL-6 and IL-1β were expressed at a low level when whole blood of healthy donors was incubated for six hours or longer without stimulating agent (control). When incubated with pCRP, expression levels did not differ signi cantly from control. In contrast, ADP-stimulated platelets cause increased expression of all three cytokines in monocytes, which importantly was signi cantly increased further by addition of pCRP. These exacerbating effects were blunted by C10M. To con rm the ow cytometric data, we performed confocal uorescence microscopy (Fig. 3F) and found TNF expression upregulated in cells incubated with ADPactivated platelets and pCRP, resulting in pCRP* formation (middle row). In contrast, cells incubated with activated platelets, pCRP and C10M were not expressing more TNF than the control group (bottom row) due to the inhibition of pCRP* formation (Fig. 3F).
C10M inhibits pCRP binding to activated platelets and microvesicles and reduces pCRP*/mCRP induced ICAM-1 and VCAM-1 expression on human endothelial cells and activation of leukocytes The interaction of immune cells with activated endothelial cells, leading to cell adhesion and transmigration, is a crucial event in localized tissue in ammation. We investigated the effects of pCRP* on endothelial cells and leukocytes in order to test the therapeutic anti-in ammatory potential of C10M. Circulating pCRP binds to activated cells and is shed on microvesicles largely as pCRP*; this mechanism is crucial in transporting and mediating pCRP* in circulation in-vivo 13,27 . We therefore examined the effects of C10M on pCRP binding to human umbilical vein endothelial cell (HUVEC) monolayers and binding of ADP-activated platelets and microvesicles derived from mononuclear cell lines (THP-1) after LPS stimulation. pCRP on platelets and microvesicles, respectively, bind to HUVEC cell monolayers (Figs. 4A, rst row and S3A, upper panel). C10M signi cantly reduces this initial step of pCRP*-mediated aggravation of in ammation (Figs. 4A, second row and S3A, lower panel). In order to test the functional relevance of this process we investigated the expression of ICAM-1 and VCAM-1 in the HUVEC cells. C10M signi cantly reduces ICAM-1 and VCAM-1 expression induced by pCRP* on platelets and microvesicles, respectively (Figs. 4B, C and S3A). Upregulation of adhesion receptors represents the in ammatory response of the endothelial cells in the initial phase of in ammation 28,29 . ICAM-1 and VCAM-1 are crucial endothelial ligands for receptors of the integrin family, essential for the adhesion and tissue in ltration of leukocytes. In order to look at this mechanism further, we examined the effects of pCRP* on neutrophil (Fig. 4D) and monocyte ( Fig. 4E) activation, as determined by CD11b expression 30 . C10M reduces pCRP*-induced monocyte and neutrophil activation ( Fig. 4D and E). Generation of reactive oxygen species (ROS) measured by redox-indicator dihydroethidium in ow cytometry served as another pro-in ammatory readout in both monocytes and neutrophils (Fig. S3B). A novel nding of our investigation into the mode of action of CRP-regulated in ammation is that pCRP*/mCRP induces formation of neutrophil extracellular traps (NETs) (Fig. S3C), a process called NETosis. NETosis is a mediator of sterile in ammation, a key event that modulates tissue and organ damage 31 . Our nding that C10M reduces pCRP*/mCRP-induced NETosis further highlights the relevance of CRP regulation at an important immune checkpoint and the therapeutic potential of C10M.
C10M inhibits CRP-induced aggravation of renal ischemia/reperfusion injury IRI represents the prototypic sterile in ammation in which an exacerbated immune response leads to unwanted tissue damage. We have previously demonstrated that IRI-associated tissue damage is induced by the pro-in ammatory forms of CRP (pCRP* and mCRP) and that the palindromic inhibitor bis-PC can largely prevent this tissue damage by stabilizing the non-in ammatory pCRP form 8 . Thus, an IRIinduced acute renal injury model in rats represents an ideal in-vivo model to evaluate the therapeutic potential of C10M 8 . Firstly, the pharmacokinetic plasma half-life (t 1/2 ) of C10M was determined by mass spectrometry to be 90 min in the rat ( Fig. 5A and B). C10M is cleared by the kidneys after i.v. administration. After IRI, the rat kidneys were examined for CRP deposits by immunohistochemistry and Western blotting ( Fig. 5C and D). Staining for CRP using the conformation-speci c anti-pCRP*/mCRP antibody 9C9, which targets an epitope exposed in the pro-in ammatory forms of CRP but not in pCRP 7,13 , demonstrated deposition of the pro-in ammatory forms of CRP speci cally localized to the IRIexposed renal tissue (Fig. 5C). After i.v. administration of C10M, CRP deposition could not be detected in the tissue. This was further con rmed by Western blots of tissue lysates separated by SDS-PAGE (Fig. 5D). The bene cial effects of C10M in IRI was re ected by the signi cant improvement of excretory renal function as analyzed by blood urea levels (Fig. 5E). To obtain further mechanistic data we assessed the CD68 + monocytic cell in ltration (Fig. 5F) in renal tissue and performed PAS staining of renal tissue ( Fig. 5G). In these assays, administration of pCRP leads to signi cant increase of IRI-associated in ammatory cell in ltration and tissue damage that can be blunted by the administration of C10M.
C10M inhibits pCRP*/mCRP-induced aggravation of allograft rejection in a hindlimb transplantation model IRI is a major aggravating factor in organ damage and allograft rejection after allograft transplantation 32,33 . To further con rm the therapeutic potential of pCRP*/mCRP inhibition in-vivo, we performed hindlimb transplantation on fully mismatched rat strains (Lewis and Brown-Norway) as a model for acute allograft rejection of vascularized composite allografts (VCA) and clinically assessed graft survival (Fig. 6A). We found human pCRP to strongly promote the diapedesis of monocytes and tissue degradation, and thereby to accelerate VCA-graft loss signi cantly compared to a transplanted control group (control vs pCRP, 7.83 vs 4.83 days; n = 4, each group; P value 0.0005, log-rank (Mantel-Cox) test) (Fig. 6B, top, black vs red broken line). Most importantly, premature graft loss driven by pCRP was prevented by i.v. C10M application during the rst two days after transplantation ( Fig. 6A and B, top, blue broken line). C10M effects are attributable to the inhibition of pCRP* and mCRP, as without extrinsic pCRP application C10M did not show protective effects (Fig. 6B, bottom, blue dotted vs black line).
Transplanted rat hindlimbs showed signi cant clinical signs of rejection (edema, erythema, and blistering) on day three after transplantation post CRP administration (Fig. 6C) that were not present in the control group or when formation of pCRP*/mCRP was blocked with C10M. Skin and muscle biopsies were taken at day 3 and analyzed histologically. Monocyte in ltration was detected by immuno uorescence microscopy, which revealed signi cantly more monocyte in ltrates in VCA-tissue of rats treated with pCRP compared to control rats. To investigate whether these exacerbating effects were speci c for CRP, C10M was used to block the formation of the pCRP* and mCRP species. In rats treated with both pCRP and C10M, no CRP deposits were detected in either muscle (Fig. 6D) or skin tissue (Fig. 6E). We found the number of transmigrated CD68 + cells to be reduced to control levels when C10M was administered in the pCRP group. Further, we analyzed the amount of depleted CRP in the tissue by Western blotting and found CRP signi cantly reduced in both muscle and skin (Fig. 6F). These results indicate that compound C10M inhibits the CRP-dependent activation and transmigration and thereby abrogates the CRP-mediated local in ammatory exacerbation in transplant rejection.

C10M does not suppress CRP-independent host defense against pathogens
Phagocytosis of bacteria is a crucial protective mechanism of the innate immune response and CRPmediated phagocytosis has been previously described 34 . To demonstrate that phagocytosis of S. pneumoniae is not abrogated by C10M we performed a ow cytometry-based phagocytosis assay. pCRP leads to a moderate increase in phagocytosis of S. pneumoniae in monocytes and neutrophils (Fig. 7A), which is reduced by addition of C10M (Fig. 7B). Baseline phagocytosis is not affected by C10M, suggesting that innate immune mechanisms that have protective functions are not inhibited by C10M.

Material And Methods
Detailed experimental material and methods are described in the Supplementary Information.

Reagents and antibodies
For the biophysical assays and crystallization studies, pCRP was commercially acquired from Merck (Product number: 236608). For the cell assays and animal studies, preparation of human pCRP was performed as described previously by our group 13 . In brief, pCRP puri ed from human ascites was purchased from Calbiochem (Nottingham, UK) and was thoroughly dialyzed twice (1:500 v/v) against Dulbecco's phosphate buffered saline (DPBS) supplemented with 0.9 mM CaCl 2 and 0.49 mM MgCl 2 .
Monomeric CRP (mCRP) was generated by treating pCRP with 8 M urea for 1 hour at 37°C and following dialysis against 25 mM Tris-HCl (pH 8.5) overnight at 4°C as described by Bíro et al. 35 . The protein concentration was determined after each dialysis and dissociation procedure by a benchtop uorometer (Qubit® 3.0 Fluorometer, Invitrogen™ by life technologies™, Carlsbad, CA, USA).

Synthesis of C10M
Synthesis of the compound C10M was achieved using standard synthetic methods (refer to Supplementary Information for full details). Dibutylamine was reacted with diethyl-(3bromopropyl)phosphonate in dimethylformamide in the presence of a catalytic amount of sodium iodide at elevated temperature. The diethyl protected key intermediate was puri ed by column chromatography on silica gel and isolated in modest yield (28 %). The diethyl protecting groups on the phosphonate were removed using trimethylsilylbromide in dichloromethane. After removal of the volatiles under vacuum and trituration with hexane, the target material was isolated as the bromide salt in modest yield (44 %) in better than 95% purity as determined by HPLC.

Surface plasmon resonance
A Biacore S200 instrument (GE Healthcare) was used to perform the experiments at 25°C, 30 µl/sec in 10 mM HEPES, 100 mM NaCl, 10 mM CaCl 2 , 0.005 % Tween pH 7.4 as running buffer. The active ow cell of the CM5 chip (GE Healthcare) was activated using NHS/EDC and the pCRP protein in 50 mM MES pH 6 was immobilized on the chip by amine coupling before blocking with ethanolamine. A range of concentrations of either C10M (0-1600 µM) or PC (0-100 µM) were injected both on the active and reference ow cells. All binding curves were double referenced, with the responses from the blank ow cell and buffer being subtracted from the active cell response during analysis using the S200 Biacore Evaluation software. The RU of each concentration of compound was plotted against the compound concentration to determine the a nity of binding (K D ). The data were expressed as mean ± SEM, each experiment was performed in triplicate.
Crystallization and X-ray crystallography The pCRP protein was prepared for crystallization by adding 1 µl 100 mM CaCl 2 to 50 µl pCRP at 8 mg/ml in 20 mM Tris, 140 mM NaCl, 2 mM CaCl 2 . Compound C10M was dissolved in water and added to pCRP to achieve a 1:3 molar ratio of pCRP:C10M. Drops of 2 µl size (1:1 volume ratio of well buffer and protein + C10M) were set up over 500 µl reservoir solution at 20 °C in a 24 well Linbro plate using hanging drop vapor diffusion methods. Crystals of pCRP in complex with C10M were obtained with 100 mM Tris pH 9, 10 % PEG 4000, 50 mM LiCl and 200 mM MgCl 2 as the well buffer. The crystals were frozen using liquid nitrogen after application of 20% ethylene glycol as cryoprotectant. Diffraction data were collected on the MX2 beamline at the Australian Synchrotron. The structure was solved from the collected dataset by molecular replacement using the previously published structure of the PC:pCRP complex as the input model (PC molecules were removed, PDB ID: 1B09 24 ). The data were processed with XDS 36 and CCP4 37 followed by re nement using PHENIX 38 . The PDB coordinates have been deposited in the Protein Data Bank under PDB ID: 7L9V.
Human ex-vivo studies Whole blood and cells isolated from peripheral venous blood used in the assays described hereafter were taken from healthy human volunteers after informed consent. All human studies were approved by the ethics committee of the University of Freiburg Medical Center (# 112/17) and conducted in accordance with the declaration of Helsinki.
In-vitro and ex-vivo testing of C10M All cell cultures were tested for mycoplasma contamination on a regular basis. HUVECs were purchased from PromoCell (Heidelberg, Germany) and cultured in supplemented Endothelial Cell Basal Medium (SupplementMix, 10 % FCS, 50 U/ml penicillin, 50 µg/ml streptomycin; PromoCell). The acute monocytic leukemia cell line THP-1 (DKMZ, Braunschweig, Germany) was used for microvesicle preparation and was cultured in RPMI 1640 medium supplemented with 10 % FCS, 2 mM L-glutamine and 50 U/ml penicillin and 50 µg/ml streptomycin. Identity of the utilized cell line was con rmed by Multiplex Human Cell Line Authentication Test (Multiplexion, Heidelberg, Germany). Peripheral blood human primary monocytes were isolated from peripheral whole blood using density gradient centrifugation. Blood was taken from normal and healthy volunteers ( Ischemic acute kidney injury model in rats The experimental protocol has been described previously by our group 24 and was conducted with minor modi cations. In brief, male Wistar rats (six weeks old, body weight 180-220 g) were anesthetized with 1.5-2 vol % iso urane (Abbott, Wiesbaden, Germany). Both renal pedicles were dissected via two ank incisions and clamped for 45 min followed by a 24 hour reperfusion period. 25 µg pCRP per ml serum volume was injected intraperitoneally at the end of the ischemia and 12 hours later. Rats intravenously received either DPBS (IRI, IRI + pCRP, sham + pCRP) or C10M (IRI + pCRP + C10M, IRI + C10M) in DPBS (1:100 in molar ratio, pCRP to C10M) four times every six hours starting with the beginning of the reperfusion period. After 24 hours, rats were killed and tissue prepared as described previously 24 .
Renal excretory function in acute ischemic kidney injury was assessed as described previously 24 by blood urea nitrogen (BUN) concentration in serum. Thus, blood samples were taken at given time points from the tail vein into micro tubes with clotting activator (Micro tube 1.3 ml Z, Clotting Activator/Serum, Sarstedt) and centrifugated after clotting. BUN was measured using a cobas 8000 modular analyzer (Roche, Basel). Hemolytic samples were discarded.

Immunostaining and histomorphological evaluation
Immunohistochemistry and histomorphological evaluation of the renal tissue was performed on formalin-xed para n-embedded renal tissue sections (5 µm thick serial sections). Para n-embedded sections were de-para nized in xylol, rehydrated, and boiled for 20 min in concentrated citric acid (pH 6.0). Antigen unmasking for anti-monocyte detection was done by application of pepsin solution (Digest-All™ 3, life technologies) at room temperature for 20 min. Previously, both kidneys were ushed with DPBS followed by xation in 4 % PFA. Evaluation was performed as described previously 24 : histomorphological changes were evaluated in a blinded fashion by two researchers using a Zeiss microscope (Carl Zeiss Microscopy Axio Imager.M2, Germany) on Periodic acid-Schiff stained sections by quantitative measurement of tubulointerstitial injury, which was assessed by loss of tubular brush border and cast formation. The morphological assessment was scaled in ve steps: not present (0), mild (1), moderate (2), severe (3) to very severe (4). Transmigrated leukocytes were detected by anti-rat CD68 antibody (clone ED-1) in a 1:100 dilution and renal in ammation was evaluated by counting ED-1 positive cells in 20 randomized areas of interest of the renal cortex at ×200 magni cation. Sections were counterstained with Mayer's hematoxylin. Unspeci c isotype matched primary antibodies served as negative control. Detection of human CRP on the renal tissue sections was performed using anti-pCRP*/mCRP antibody 9C9 (1:10 dilution).

Hind limb transplant rejection model
The two inbred stains Brown-Norway (BN, recipient) and Lewis (LW, donor) show a strong antigenic mismatch 25 and were used for the acute rejection model. The method was rst described by Doi 26 and was performed with minor modi cations. In brief, two experimenters performed transplantation together: while one is working on the donor rat, another was preparing the recipient rat. In both rats, hind limbs are shaved and thoroughly disinfected, then a circumferential skin incision was performed at mid-thigh level. The donor limb is rst xed by femoral bone osteosynthesis, which was achieved by using an intramedullary rod made from a 0.8 mm Kirschner wire. Muscles are then sutured with 4/0 nylon running sutures with adaption of the according functional groups (thigh extensors, adductors, gluteal muscles and hamstrings). This model of acute rejection was set as a non-functional hind limb transplant, so no suturing of the nerves was performed. Revascularization was performed using 9/0 nylon sutures. Both vessels were sutured under the microscope with 8-10 single stiches. The inguinal fat ap (containing the super cial epigastric artery) from the donor hind limb is used to cover the anastomoses to prevent major bleeding. The wounds were rinsed with 0.9 % saline solution and the transplantation is completed by a Penrose drain including skin closure with running sutures (4/0 nylon). The skin is cleaned with nonalcoholic disinfectant (Octenisept®, Schülke, Germany) after the skin suture is completed. The total operative time was on average 90 min. All rats received postoperative subcutaneous injections of 100 µg/100 g bodyweight of carprofen for pain relief and 1 ml/100 g bodyweight saline solution for volume compensation. A plastic collar was used to prevent auto-mutilation and hind limbs with self-in icted wounds were excluded from further evaluation. For postoperative management after the completion of the skin suture, su cient reperfusion of the transplanted hind limb was assessed again. Rats then received a rst intraperitoneal bolus of 25 µg pCRP per ml serum volume and 500 µl of DPBS supplemented with calcium and magnesium (control), respectively. The second bolus was administered 24 hours after the rst. An intravenous catheter (Abbocath-T 26 G, 0.6 x 19 mm) in the tail vein was used to inject C10M (1:100 molar ratio to pCRP) in DPBS and DPBS, respectively, every six hours for the rst 42 hours (eight applications in total, starting with the rst intraperitoneal bolus) (Fig. 6A). Rats are allowed to awake from anesthesia and cared for until fully awake and warmed. All rats showed slight edema of the transplanted hind limb within the rst post-operative day. Rejection of the hind limb graft was assessed by clinical control every 8 hours and graded as described previously 27 according to an established clinical classi cation for allograft rejection, from 0 (no clinical signs of rejection), 1 (edema), 2 (erythema), and 3 (epidermolysis and desquamation) to 4 (necrosis). Four experimental groups were included in this study (n = 4). In the control group (n = 4, LW◊BN), Brown-Norway recipient rats received intraperitoneal DPBS administration. Rats in the pCRP group received two intraperitoneal boli of 25 µg pCRP (BD Micro-Fine™ +Demo, 30G insulin syringes) per ml serum volume directly following to the surgical procedure and after 24 hours. Serum volume was estimated as described previously as a function of the body weight 28 . Immediately after surgery, subcutaneous saline supplementation was given to avoid dehydration of the rats. In the C10M treatment group (n = 4, LW◊BN), rats were treated as in the pCRP group. Additionally, rats received intravenous compound C10M (1:100 molar ratio) via a 26G catheter (Abbocath-T, ICU Medical B.V., Netherlands) in the lateral tail vein every 6 hours for the rst two postoperative days. Biopsies were taken on day three after transplantation of skin and muscle tissue and immunohistochemistry performed on formalin-xed and para n embedded samples. After incubation with primary antibody anti-CD68 (clone ED1, 1:100) and anti-human CRP (clone 8, 1:200) for 1 hour at room temperature, slides were incubated with secondary antibody anti-mouse-conjugated CF488 (green) following the manufacturer's protocol.

Pharmacokinetic studies in rats
Plasma concentrations of C10M were measured by an LC-MS method after a single intravenous injection into the tail vein of male Wistar rats (250-350 g). Rats were anesthetized as described above and temperature controlled. 100 µg of C10M was injected and blood samples were taken at given time points (1,5,10,15,30,45, 60 and 90 min after bolus injection). EDTA-anticoagulated (1.6 mg/ml EDTA) blood samples were centrifugated for 10 min at 2,000 x g and 4°C to remove the cellular portion. The resulting supernatant was snap frozen and stored at − 80°C until further sample preparation with solid-phase extraction.
Renal excretion of C10M was measured by a model previously described 29 with marginal modi cations.
Male Wistar rats (300-350 g bodyweight) were anesthetized as described above and the urinary bladder was carefully exposed and externalized under sterile conditions. Sterile urine was drawn from the bladder after a single intravenous application of C10M and C10M + pCRP 15, 30, 45, 60, and 90 min after the i.v. application, respectively. The urine samples were immediately snap frozen and stored at -80°C until measurements taken. Pharmacokinetic parameters were calculated using PKSolver add-in for Microsoft Excel 30 .

SDS-PAGE and Western blotting
For SDS-PAGE, tissue lysates from rat kidneys, muscle and skin tissue samples were precipitated on ice with same volumes of 10 % trichloroacetic acid after homogenization with a disperser tool on ice (Ultra Turrax® IKA®, Germany). Pelleted protein was denaturated in SDS loading dye supplemented with DTT at 95°C, 5 min and then separated on 10-12 % SDS-polyacrylamide gels. After Western blot, nitrocellulose membranes were blocked in 5 % BSA in TBS-T and incubated with mouse anti-human CRP antibody (Clone CRP-8, 1:2,000). HRP-conjugated goat anti-mouse antibody 1:5,000 v/v in 1 % BSA-TBS-T was used to detected bound CRP antibodies after washing steps. GAPDH served as loading control and was detected with anti-human GAPDH-HRP (1:1,000 in 1 % BSA in TBS-T). Protein bands were visualized using ECL™ Western blotting analysis system (GE Healthcare, United Kingdom), medical x-ray lm (Fuji lm, Japan), and developed on a CURIX 60 developer (AGFA).
For detection of CRP in IRI kidneys, snap frozen tissue was homogenized on ice using a high-power disperser in lysis buffer with added protease inhibitors. After centrifugation of the homogenized tissue, the supernatant was transferred, and protein concentrations were determined with BCA protein assay, and processed as described above.
For semiquantitative analysis of CRP binding to activated human platelets, we performed SDS-PAGE and Western blotting as described previously 10 . Brie y, human platelets were isolated and washed from citrate-anticoagulated whole blood by differential centrifugation in sequestrene buffer. pCRP (100 µg/ml) was incubated with ADP-activated platelets and C10M at different concentrations (10 mM and 100 mM,

Statistical analyses
All statistical analyses were performed using GraphPad Prism v9.0 for Mac (GraphPad Software, La Jolla, California, USA). Experiments were performed at least three times. The data are shown as mean and standard error of the mean (SEM) as indicated. One-way analysis of variance (ANOVA) and post-hoc Tukey's test was used to compare more than two groups. If only two groups were compared, a two-tailed Student's t-test was employed. A P value of < 0.05 was considered statistically signi cant. Discussion C-reactive protein is an evolutionary highly conserved and central player in in ammatory diseases 1 . The circulating isoform of CRP, pCRP, binds to PC and PE head groups of bioactive lipids exposed on the membranes of damaged cells and microvesicles, which subsequently leads to the formation of the proin ammatory CRP isoforms, pCRP* and mCRP 3,12 . This CRP 'activating' mechanism has only recently been identi ed 3,9,11,13 . It transforms a relatively inert molecule, pCRP, to highly pro-in ammatory molecules, pCRP* and mCRP, both contributing to and aggravating tissue damage 8,11,12 . Disrupting the interaction between pCRP and PC/PE, and thereby inhibiting this 'CRP activation', represents a novel and attractive anti-in ammatory strategy. Targeting the direct interaction between PC/PE and pCRP, we employed a combination of medicinal chemistry and computational modeling to develop a novel low molecular weight inhibitor that binds to the PC/PE binding pocket of pCRP thereby blocking pCRP binding to exposed PC/PE head groups and consequently blocking 'CRP activation'. Utilizing SPR and X-ray crystallography as direct protein binding assays, in combination with in-vitro binding assays, we demonstrate that the small molecule, C10M, binds to the PC/PE binding pocket of pCRP in a competitive manner. C10M inhibits pCRP*/mCRP-induced monocyte adhesion, cytokine and ROS production, NET formation as well as pCRP*/mCRP-mediated upregulation of endothelial ICAM-1 and VCAM-1. Most importantly, C10M affords signi cant protection from CRP-mediated tissue injury in two distinct preclinical models of in ammation, a model of renal IRI and hindlimb transplantation.
Our approach contrasts to the previously described mode of action for bis-PC, a compound that combines two PC moieties into one bivalent molecule. Bis-PC prevents the formation of pCRP* and mCRP by each of its PC portions binding to the PC/PE binding pocket on two separate pCRP molecules, thereby bringing the two pentamers together with opposing B-faces in a doughnut-like decameric arrangement 5 . Within this decamer structure, the pCRP B-faces are no longer available to bind to exposed PC/PE head groups of bioactive lipids and the 'CRP activation' step is blocked 5 . It had been suggested that the CRPinhibitory effect of bis-PC was primarily due to its bivalency (i.e., two functional PC head groups), however we hypothesized that a compound with just one head group would be able to elicit an antiin ammatory response. Here, we explored this hypothesis by designing a low molecular weight (~ 250 Da) molecule, C10M, which mimics PC by binding to the PC/PE binding pocket on pCRP, thereby stabilizing the inert, non-in ammatory form of CRP by competitively inhibiting the binding of pCRP to exposed PC/PE head groups of bioactive lipids. The binding mode of C10M to pCRP was con rmed by Xray crystallography, showing that the phosphonate moiety was the main anchor point of the compound to pCRP. The mode of action of C10M was further investigated and the therapeutic potential of C10M is supported by our following ndings (as summarized in the schematic drawing depicted in Fig. 8): (1) Biophysical assays and X-ray crystallography reveal a speci c, stable and competitive binding of C10M to the PC/PE binding pocket of pCRP. (2) C10M binds pCRP and prevents pCRP binding to activated/damaged cell membranes. (3) In turn, this prevents the conformational change from pCRP to pCRP*/mCRP and their tissue deposition in the area of in ammation. (4) C10M is therefore inhibiting the exacerbation of in ammation by CRP, rather than being a general anti-in ammatory drug. (5) C10M reduces tissue damage in a renal IRI model and reduces CRP-mediated acceleration of allograft rejection.
(6) C10M does not inhibit CRP-independent phagocytosis suggesting that protective innate immune properties remain intact. pCRP binding to membrane phospholipids of apoptotic cells has been described previously 40,41 . The phospholipase A2-dependent membrane changes in apoptosis and cell activation that lead to the generation of LPC appear to be crucial in mediating pCRP binding to cell membranes, as pCRP neither binds to non-activated leukocytes 13 nor the ubiquitous PC head groups in the plasma membrane of healthy living cells. We recently showed that the membrane curvature of healthy cells prevents access to the PC/PE binding pocket of pCRP 13 . Once cell membranes are damaged, the membrane curvature increases, as is the case in microvesicles that avidly bind and transport pCRP*/mCRP 13,27 . Therefore, we chose the PC/PE binding pocket of pCRP as the target for our drug discovery approach. As PC/PE head groups are only accessible for pCRP binding after an in ammatory stimulus or in apoptosis, this therapeutic targeting concept aims at the inhibition of the binding events critical to the generation of 'activated' CRP to minimize any side effects compared to general pCRP inhibition and reduction of circulating pCRP levels such as achievable by anti-sense oligonucleotides 21 or by CRP-apheresis via binding to a PC resin 22 . This was considered crucial in our inhibitor design, as in ammation can clearly be bene cial as a defense and repair mechanism of the organism. Thus, anti-in ammatory strategies are often considered "two-edged swords" as most anti-in ammatory therapies come at the cost of some level of immune suppression which can lead to signi cant side effects. Moreover, a small molecule inhibitor is likely to offer signi cant advantages compared to either anti-sense oligonucleotides or CRP-apheresis given the potential ability to administer small molecule inhibitors via the oral or parenteral route, in addition to having an immediate onset. In contrast, the use of anti-sense oligonucleotides to inhibit CRP expression requires pre-treatment over weeks making this approach unsuitable for acute applications such as IRI. Whilst CRP-apheresis is effective at depleting CRP acutely and thus highly attractive in the acute/emergency setting of tertiary hospitals, the use of apheresis is time consuming and contingent upon the availability of the necessary infrastructure. Therefore, a small molecule CRP inhibitor is highly desirable given the potential ease of administration and suitability for acute and chronic indications.
It is important to emphasize that we adopted an approach that provides a selective therapeutic strategy, only inhibiting the uncontrolled exacerbation of in ammation by pCRP*/mCRP but not in ammation per se. We con rm this in our in-vitro monocyte assays, in which we analyze the mode of action of our compound. In these assays we use pCRP*/mCRP on the surface of ADP-stimulated platelets to stimulate monocyte adhesion and cytokine expression. Our data con rms our hypothesis that C10M is inhibiting the effects caused by pCRP*/mCRP, but not the increase in in ammation that is induced by ADPstimulated platelets, by stabilizing the non-in ammatory pCRP species. Further supporting the concept that the basic innate immune response is not affected by C10M, we demonstrate that phagocytosis is not inhibited by C10M. Indeed, neutrophil phagocytosis and killing of bacteria are essential for host defense against bacteria such as pneumococci 42 and pCRP mediates an increased resistance to bacterial infections 43 via binding to bacterial PC and opsonization of bacteria 44 . We demonstrate that C10M does not inhibit phagocytosis of S. pneumonia, zymosan or E. coli by monocytes and neutrophils, suggesting that the protective capacities of these innate immune cells are maintained. This is crucial for an antiin ammatory treatment with reduced side effects. These are important observations since a major aim of our therapeutic approach is to target the uncontrolled exacerbation of in ammation rather than in ammation or innate immunity in general.
To validate the effect of C10M in-vivo, we used two distinct animal models of in ammation. Firstly, an established renal IRI model 8 in which we previously demonstrated that pCRP*/mCRP lead to enhanced leukocyte activation, tissue in ltration and generation of ROS resulting in aggravation of tissue injury 8 .
This is an ideal model to investigate the anti-in ammatory properties of C10M as IRI represents the prototypic, sterile in ammatory setting that results in increased tissue damage. Secondly, the relevance of IRI for allograft survival after tissue transplantation was investigated. The initial in ammatory stage after transplantation is characterized by IRI and has a crucial impact on long term allograft survival. Indeed, the in ltration of kidney allografts by macrophages within 10 days of transplantation is associated with worse clinical outcome 45,46 . Furthermore, episodes of acute allograft rejection in the rst posttransplantation period have a severe negative impact on long-term allograft survival 47 . Therefore, we used a well described allograft rejection model (hindlimb transplantation as a model of vascularized composite tissue allotransplantation -VCA) 48 to test the therapeutic potential of C10M. We demonstrate that CRP accelerates allograft rejection via aggravation of IRI and activation of the innate immune response. In both animal models, we establish C10M's unique bene ts in reducing CRP-mediated tissue damage. In the VCA model the acceleration of acute allograft rejection by CRP is reversed by administration of C10M. In the IRI model, renal function is signi cantly improved and histological signs of kidney injury are markedly reduced. The deposition of CRP in the tissue of renal IRI is signi cantly reduced after administration of C10M, con rming our in-vitro ndings that pCRP binding to activated cell membranes, which we have demonstrated to be a prerequisite for subsequent tissue deposition in the area of in ammation, is reduced 11 . For our in-vivo experiments we utilized rat models, as although rats have abundant pCRP (300-600 µg/ml in normal healthy pathogen-free rats), rat CRP is not utilized as an acute phase protein and rat complement is not activated by rat CRP 49 . This is in contrast to human CRP that activates both rat and human complement, but not mouse complement 50 . Therefore, rats supplemented with human pCRP are a suitable animal model for CRP research 8,10 .
In our experiments, intermittent C10M i.v. injections were su cient to obtain protective effects in two animal models of localized in ammation. The development of C10M as a novel small molecule inhibitor of CRP provides important proof of concept that such a therapeutic strategy holds signi cant promise as a targeted anti-in ammatory approach and paves the way for future compound design, potentially further optimizing pharmaceutical properties. In conclusion, competitively blocking the PC/PE binding pocket on pCRP with C10M is a successful and attractive strategy to reduce CRP-mediated aggravation of in ammation. Signi cantly, given the wide range of clinical conditions where CRP-mediated tissue damage has been demonstrated, the therapeutic targeting of CRP with small molecule inhibitors is likely to be of broad clinical relevance and can potentially be used in many in ammatory diseases. Figure 1 Compound (colored sticks) con rmed that the compound binds to the same pocket as PC/PE. The asymmetric unit consisted of 4 stacked pCRP pentamers. C10M was bound to ve CRP monomeric subunits, four in one pentamer and one in another (Table S1)  activated platelets. Platelets were isolated, washed and stained (anti-CD62P FITC, green) from 3.8 % sodium-citrated human whole blood. 50 µg/ml pCRP-Atto594 (red) was incubated with isolated platelets (green) with (bottom) or without (top) C10M in a 1:100 molar ratio. pCRP localizes on the plasma membrane, which can be inhibited by C10M. Scale bar: 20 µm. (C) Western blot of CRP binding to activated platelets. Platelets bind less CRP when incubated with C10M as detected with the anti-CRP antibody (clone CRP-8), compared to the anti-GAPDH antibody (clone 0411) as a control. Washed platelets isolated from human citrated whole blood were incubated with pCRP and C10M, respectively, lysed and separated on SDS-PAGE. (D) Densitometric quanti cation of protein bands of Western blots (n = 3) from the experiment described in (C) using ImageJ. P value was calculated by Student's t-test. **P <0.01.   Urine was sampled by sterile puncture of the urinary bladder 15, 30, 45, 60, and 90 min after intravenous application of C10M. Over 80 % of the applied C10M mass was excreted after 90 min. (C) Immunohistochemistry of rat kidneys subjected to IRI and i.p. pCRP application revealed distinct staining by anti-pCRP*/mCRP-9C9 antibody (green). C10M reduces the deposition of total CRP in the impaired tissue. No deposits in the non-ischemic tissue (sham). Exemplary stainings out of at least three are shown. Scale bars, 100 µm. (D) Tissue lysates of rat kidney were separated on SDS-PAGE and total CRP was identi ed with anti-CRP antibody (Clone 8). A band at the size of mCRP (~23 kDa) was detected in kidneys subjected to IRI and pCRP, but not in animals treated additionally with C10M. Representative results are shown. (E) Renal excretion is impaired by pCRP*/mCRP-driven tissue damage. Blood urea nitrogen (BUN) was utilized as surrogate marker for the excretion function of the kidney. Blood samples were taken before the surgical procedure (preoperative) and 24 hours after the procedure (harvest). (F) Immunohistochemical detection of transmigrated CD68+ cells in IRI kidneys. Representative results are shown. pCRP (25 µg/ml) increased the number of CD68+ cells transmigrated into injured renal tissue signi cantly, while C10M abolished these effects. Scale bars indicate 100 µm. Quanti cation of immunohistochemical results is shown on the right. Presented are mean cell counts per ROI in each animal (n = 6 per group). One-way ANOVA was used to compare treatments. # not signi cant, ***P <0.001. (G) Periodic acid-Schiff (PAS) stained kidney sections show increased damage after renal IRI in rats when pCRP was injected i.v. The tubulointerstitial injury was quanti ed by the loss of tubular brush border and by cast formation following an established protocol 52,53. Scale bars indicate 100 µm.

Figures
Quanti cation of immunohistochemical results is shown on the right. Statistical analysis was performed with ANOVA and Tukey's post hoc test. n = 6, **P <0.01, ***P <0.001, ns: not signi cant. were injected i.p. and i.v., respectively. (B) Kaplan-Meier plots for control, pCRP, and pCRP + C10M (above) and C10M vs control (bottom) treatment for total hindlimb allograft survival. Recipients were injected twice with PBS (control), 25 µg pCRP, or 25 µg pCRP + C10M per ml serum, respectively. C10M boli were applied i.v. at a 1:100 molar ratio pCRP/C10M postoperative every six hours for two days. Kaplan-Meier curves for different treatments were compared using log-rank test. Hindlimb survival interval was compared by Mantel-Cox log-rank test and was signi cantly reduced by pCRP administration (P = 0.0005, median survival control vs pCRP, 7.8 vs 4.8 days). C10M masks the CRP-accelerated hindlimb rejection (median survival 7.  Opsono-phagocytosis of S. pneumoniae serotype 27 is not compromised by C10M. (A) Phagocytosis of pCRP-opsonized heat-killed and FITC-labeled S. pneumoniae by monocytes and neutrophils serves as exemplary phagocytosis assay. Whole blood samples of healthy volunteers were incubated with targets with and without pCRP (100 µg/ml), processed as described in Supplementary Materials and Methods and then measured by ow cytometry. Scatter plot shows phagocytic index (percentage of target positive cells of subtype / all cells of subtype) of un-opsonized (control, light-grey dots) and CRP-opsonized targets (dark-grey triangles) after 5, 10, 15, and 20 min, respectively. Targets were incubated with pCRP in the presence of calcium for 30 min, 37° C. Mean ± SEM are indicated. P values were calculated using ANOVA and Tukey's post hoc test. n = 4, **P < 0.01, *P < 0.05. (B) Experiments as described in (A) were repeated but with targets incubated with pCRP (100 µg/ml) and C10M (1:100 molar ratio) for 30 min, 37°C . Mean ± SEM are indicated. n = 4, **P < 0.01, *P < 0.05. Figure 8