Effect of Human C-Reactive Protein on Chemokine and Chemotactic Factor-Induced Neutrophil Chemotaxis and Signaling¹

The synergistic action of the inflammatory cytokines IL-1β, TNF-α, and IL-6 reorchestrates the pattern of gene expression in hepatocytes and greatly enhances the synthesis of a group of acute phase blood proteins, more broadly termed acute phase reactants (APR)⁵ (1, 2, 3). These inflammatory cytokines also rapidly elicit the production of the chemokine IL-8 by monocytes and endothelial cells (4, 5). CRP is the prototype of the APR in humans and is widely used as a gauge for the presence and extent of systemic inflammation (1). Circulating CRP is a pentamer of noncovalently associated subunits with identical amino acid sequences, each of which displays Ca²⁺-dependent binding to phosphocholine (PC) and other monophosphate esters (6). CRP does serve as the ligand for specific receptors on leukocytes and has been reported to activate monocytes and macrophages, but inhibits the respiratory burst of PMN (reviewed in 7 . Nonetheless, CRP is thought to contribute to innate host resistance. Recently, CRP has been shown to modify inflammation in transgenic animal models (8, 9, 10). Like other APR, CRP has recently been classified as an innate recognition lectin, not only because the three-dimensional structure of each of the five identical subunits of CRP is similar to Con A, but also because of its potential to couple nonspecific host responses with specific immunity (11, 12). Despite extensive studies of CRP activities in vitro, a unique role for CRP has yet to be determined.

IL-8 is a chemokine of the CXC family that mediates the recruitment and transmigration of neutrophils across the endothelium into inflammatory sites (4, 5). CRP selectively accumulates at sites of tissue damage (13, 14) and is capable of being digested by PMN into peptides that influence the activities of both monocytes and neutrophils (15, 16, 17, 18, 19, 20, 21, 22). High blood levels of IL-8 are associated with acute inflammatory conditions, and thus correlate with the extent of neutrophil infiltration, as well as increased blood levels of CRP (1, 2, 23). Recruitment of granulocytes to inflamed sites is a crucial event for both host resistance and eventual tissue repair; it is dependent on both chemokines as well as a group of poorly defined extrinsic factors that control the cell’s response to chemokines (4, 5, 23, 24). Since PMN possess specific receptors for CRP (18, 19, 20, 21, 22), we tested the influence of CRP on the response of granulocytes to both IL-8 and the bacterial chemotactic peptide, FMLP-phenylalanine (FMLPP). The studies described in this work show that CRP inhibits the IL-8- and FMLPP-induced neutrophil chemotactic response, but augments phosphatidylinositol-3 kinase (PI-3K) signaling and the mitogen-activated protein kinase (MAPK) pathway, suggesting that CRP influences PMN responses via a distinct signaling pathway.

FMLP, FMLPP, PMA, DMSO, and ferricytochrome C (horse heart) were obtained from Sigma (St. Louis, MO). The PMA and FMLPP were stored concentrated in DMSO at −20°C and diluted with buffer just before use. The p-aminophenyl PC-Sepharose for CRP purification was purchased from Pierce (Rockford, IL). rIL-8 isoforms of both 72 and 77 amino acids were purchased from Sigma and Harlan Bioproducts (Madison, WI), respectively. ¹²⁵I-labeled IL-8 (¹²⁵I-IL-8) was obtained from Amersham Life Sciences (Arlington Heights, IL). The PI-3K inhibitor, wortmannin, was purchased from Sigma. The specific ERK kinase inhibitor PD098059 was obtained from Calbiochem (La Jolla, CA).

The promyelocytic cell line HL-60 was obtained from American Type Culture Collection (ATCC, Manassas, VA) and grown in RPMI 1640 plus 4% defined FBS (HyClone, Logan, UT) and 6% defined bovine calf serum (HyClone). HL-60 cells were differentiated into granulocytic (G) cells by incubating 2 × 10⁵ cells/ml with 1.2% DMSO for 6 or 7 days, after which 85 to 95% of the cells rapidly reduced nitroblue tetrazolium dye and stained positively for cytoplasmic esterase (25). The HL-60(G) cells were washed twice into Earle’s balanced salt solution (EBSS) containing 10 mM HEPES (pH 7.4), and adjusted to 10⁷ cells/ml and kept on ice for functional or binding assays.

Synthetic peptides, numbered according to the derived sequence for the 206 amino acids present in each of the human CRP subunits (26), were synthesized by Ohio State University Biochemistry Instrument Center using t-boc synthesis on a model 9500 peptide synthesizer (Milligen/Biosearch, Millipore, Burlington, MA): Pep 1–15, Q-T-D-M-S-R-K-A-F-V-F-P-K-E-S; Pep 27–38, T-K-P-L-K-A-F-T-V-C-L-H; Pep 47–63, R-G-Y-S-I-F-S-Y-A-T-K-R-Q-D-N-E-I; Pep 134–148, I-L-G-Q-E-Q-D-S-F-G-G-N-F-E-G; Pep 152–176, L-V-G-D-I-G-N-V-N-M-W-D-F-V-L-S-P-D-E-I-N-T-I-Y-L; Pep 191–205, K-Y-E-V-Q-G-E-V-F-T-K-P-Q-L-W.

CRP was purified as described elsewhere (27, 28). Briefly, serum amyloid P-component was removed from CRP-containing pleural or ascitic fluids by passage through a column of agarose beads. The eluted protein was then passed through a 10-ml (20-mm-diameter) column of p-aminophenyl PC-Sepharose (Pierce) and washed extensively with TBS (Tris-buffered saline) + 2 mM Ca²⁺, and the bound protein was eluted in TBS + 10 mM EDTA. A second round of affinity purification on the PC-bearing matrix was used to remove trace amounts of other proteins. The concentration of CRP was determined by ELISA or by radial immunodiffusion with sheep anti-human CRP. The protein was >99% CRP based on reactivity in the competitive ELISA and by SDS-PAGE. The concentration of endotoxin in the purified CRP was 0.1 to 0.2 endotoxin U/mg protein (Chromogenic Limulus assay; M. A. Bioproducts, Walkersville, MD), corresponding to an LPS concentration of <0.05 ng/mg of CRP.

The human promonocytic U937 cell line served as a source for the CRP-R protein (27). The cells (5 × 10⁸) were washed extensively in D-PBS and solubilized in 10 ml of a lysis buffer of 20 mM Tris, pH 7.5, 110 mM NaCl, 10 mM EDTA, 1% Nonidet P-40, 2 mM PMSF, 10 μg/ml chymostatin, 10 μg/ml leupeptin, 10 μg/ml pepstatin A, and 10 μg/ml antipain by tumbling at 4°C. After centrifugation at 12,000 × g, the supernatant was applied to a ligand-affinity column of purified CRP covalently coupled to Sepharose-4B at 5 mg CRP/ml of beads and recycled through the 5-ml column overnight. After extensive washing with cold D-PBS containing 1% Nonidet P-40 and 2 mM PMSF, the bound proteins were eluted with 100 mM glycine-HCl, pH 2.6, and the fractions (1 ml) immediately were neutralized with 50 μl of 2 M Tris, pH 7.6. The fractions with protein were pooled, passed through a detergent-removal gel (Extracti-Gel D; Pierce), and concentrated on a Centricon-10 membrane (Amicon, Beverly, MA). The proteins were separated by SDS-PAGE on a 10% resolving gel that was preelectrophoresed with running buffer containing 1 mM glutathione. A single-stained band of a glycosylated protein of ∼40 kDa (27) was dissected from the gel, soaked in 1 ml of 0.125 M Tris/HCl, pH 6.8, and 1 mM EDTA, and electroeluted to serve as a source of immunogen. BALB/c mice were injected i.p. with ∼100 μg of the purified CRP-binding membrane protein from U937 cells emulsified 1:1 by volume in CFA. The mice were subsequently injected both s.c. and i.p. once 3 wk later. A final i.p. injection of the protein alone was given 30 days later. After 3 days, spleen cells were fused with the P3x63.AG853 nonsecreting mouse myeloma variant (ATCC). Supernatants from Ig-secreting hybrids were screened for reactivity with the immunogen in a direct ELISA and for reactivity with HL-60 cells, which do not share HLA with U937 cells, but possess the CRP-R. Selected hybridomas were subcloned twice, rescreened, and expanded for the production of ascites. The clones were all found to be of the mouse IgM (κ) isotype, which was purified from ascites using a mannan-binding IgM affinity purification kit (Pierce). The IgM concentration was determined by absorbance at 280 nm and by radial immunodiffusion versus the mouse IgM myeloma MOPC-104e. The RC10.2 clone was used for the anti-CRP-R mAb in these studies since it inhibits ligand (¹²⁵I-labeled CRP) binding to HL-60 and U937 cells (27).

A mouse mAb to IL-8RA designated 9H1 was obtained from Genentech (So. San Francisco, CA) and has been shown to bind to the NH₂ terminus of the IL-8R and neutralized IL-8 binding (29).

Human peripheral blood PMNs were obtained from heparinized blood by centrifugation on a one-step polymorph purification solution (Accurate Chemical & Scientific, Westbury, NY). The purity of the PMNs was always >95% with 2 to 3% contaminating lymphocytes, as judged by morphology. The neutrophils were washed twice and brought up to 5 × 10⁶/ml in cold Gey’s balanced salt solution (GBSS) containing 0.1% BSA and 20 mM HEPES buffer at pH 7.2. Chemotaxis was assessed using modified Boyden chambers with polyvinyl propylene-free polycarbonate Nucleopore membranes with 5-μm pores. The chemotactic substance was added to the lower chamber, and 2.5 × 10⁵ PMNs were added to the upper chamber. The chambers were incubated at 37°C for 90 min, and the filters were removed, fixed, and stained with Wright’s Giemsa stain. A single-blind assessment of chemotaxis was conducted by coding the slides before the number of cells/oil-immersion field (×450) were counted by an unbiased observer. The data from at least 20 fields/filter were obtained from two independent examinations of the slides, and the mean value of cells/high power field was calculated. The mAb HD2-4 was used to neutralize CRP, and is described elsewhere (30).

¹²⁵I-IL-8 was added to HL-60(G) cells (10⁶/0.1 ml) at 1 nM in D-PBS with 1% BSA. The competing proteins, CRP and unlabeled IL-8, were added at the same time as ¹²⁵I-IL-8, and the binding was allowed to continue for 60 min on ice. The cells were centrifuged at 14,000 × g for 1 min at 4°C through a phthalate oil mixture (0.5 ml), and the cell pellet was counted immediately (27, 28) in a Beckman Gamma 4000 counter. Results were expressed as percentage of control binding (5–20,000 cpm/10⁶ cells), and the experiment was repeated three times with triplicate samples.

HL-60(G) cells at 5 × 10⁶/sample were lysed by TN-1 (1% Nonidet P-40; 20 mM Tris, pH 8; 150 mM NaCl; 10 mM each EDTA, NaF, Na₃VO₄, and NaP∼P; 10 μg/ml aprotinin; and 10 μg/ml leupeptin) solution after stimulation with IL-8 or FMLPP. MAPK was immunoprecipitated at 4°C for 3 h by 20 μl protein A/G-Sepharose beads precoated with 1 μg of IgG rabbit anti-ERK-2 (Santa Cruz Biochemicals, Santa Cruz, CA). The beads were washed four times in TN-1 and then twice in the kinase assay buffer (MgCl₂, 20 mM; sodium orthovanadate, 0.1 mM; β-glycerophosphate, 20 mM; and HEPES, 30 mM, pH 7.6). Myelin basic protein (MBP) at 6 μg was used as the substrate, and 5 μCi [γ-³²P]ATP (DuPont, Wilmington, DE) was added per sample. The γ-³²P phosphorylated MBP was detected by autoradiography after separation by SDS-PAGE and transfer to a nitrocellulose membrane. MAPK protein was detected by immunoblotting the same membrane.

Either IL-8- or FMLPP-stimulated HL-60(G) cells (5 × 10⁶ in 0.1 ml of RPMI 1640) were lysed with TN-1 solution, and immunoprecipitation of the p85 regulatory subunit of PI-3K was done overnight at 4°C using protein A/G-Sepharose beads precoated with a rabbit polyclonal IgG Ab to the p85 subunit of PI-3K. The Ab was raised by injecting a purified glutathione-S-transferase fusion protein containing the N-terminal SH2 domain of p85 (obtained from B. Schaffhausen, Department of Biochemistry, Tuft’s University School of Medicine, Boston, MA). The Ab did not cross-react with other SH2 domain-containing proteins (data not shown). The beads were washed in TN-1 and three times in 10 mM HEPES solution (pH 7.4). One-half of the beads were boiled in 2× DTT-SDS-PAGE sample buffer for 5 min before running the supernatant on SDS-PAGE. Proteins were transferred to a polyvinylidene difluoride membrane (MSI, Westborough, MA) and blotted with the same rabbit Ab at 1:1000. Horseradish peroxidase-labeled anti-rabbit IgG (Kirkegaard & Perry Laboratories, Gaithersburg, MD) was used as a secondary Ab, and a commercial chemoluminescence solution (KPL) was used as the substrate. The remaining half of the beads were used for a PI-3K activity assay. The beads were suspended in an assay buffer (30 mM HEPES, pH 7.4, 30 mM MgCl₂, 1 mM EDTA, and 50 μM ATP), using 100 μg of exogenous phosphatidylinositol (Sigma) and [γ-³²P]ATP (2 μCi) as substrates. The reaction was run for 15 min at room temperature and stopped with 0.1 M HCl. Lipids were extracted with a mixture of methanol:chloroform at 1:1. Detection of γ-³²P-labeled phosphatidylinositol 5-monophosphate was done by running TLC Silica G-60 plastic-backed plates, followed by autoradiography.

Since the chemotactic activity of blood PMNs is an indicator of the cellular inflammatory response, the effects of purified human CRP on the neutrophil chemotactic response to the chemokine IL-8 and the bacterial chemotactic peptide FMLPP were examined. Brief exposure of isolated neutrophils to CRP under ligand-binding conditions (on ice) and subsequent examination of their chemotactic responsiveness to rhIL-8 (1 to 10 nM) revealed that CRP significantly inhibited the response at a concentration of >10 μg/ml CRP, as shown in a representative experiment (Fig. 1,A). Using the same experimental approach with FMLPP at 10 nM, CRP inhibited neutrophil chemotaxis at concentrations >20 μg/ml (Fig. 1,B). In these experiments, addition of the CRP alone to the lower chamber had no effect on chemotaxis, nor did CRP alone, or in combination with IL-8 or FMLPP, have any effect on cell viability, as judged by dye exclusion. The addition of the high affinity mAb to human CRP, HD2-4, neutralized its inhibitory activity on chemotaxis (Fig. 1 B). Thus, CRP affects neutrophil chemotaxis in response to both IL-8 and FMLPP, but required a greater concentration of CRP to inhibit the response to the more potent chemotaxin.

Since human CRP is digested by activated PMNs into biologically active peptides (15, 31), the activity of CRP-derived peptides generated by incubation with PMA-activated PMNs on neutrophil chemotaxis was examined. The products of CRP digestion at 800 pmol/ml (20 μg/ml) inhibited chemotaxis in a manner similar to intact CRP (Table I). Approximately 60 to 80% of the CRP was cleaved into polypeptide fragments of <25 kDa by the PMA-activated PMNs, as judged by SDS-PAGE under reduced conditions (31). The recovered CRP polypeptide fragments, from which intact CRP was removed, by themselves exerted only very weak chemotactic activity (data not shown). Therefore, the net effect of the digested CRP on chemotaxis is one of inhibition.

The synthetic cell-binding peptide of residues 27–38 within each CRP subunit, which binds to the CRP-R on leukocytes (28, 32), was tested along with additional synthetic CRP peptides for chemotactic activity at 800 pmol/ml, the equivalent of 20 μg/ml of the intact CRP pentamer. Of the peptides listed in Table II, only the cell-binding peptide of residues 27 through 38 displayed weak chemotactic activity when compared with FMLPP. Several of the CRP-synthetic peptides were also tested for their effects on neutrophil chemotaxis mediated by IL-8. Of the peptides tested, only two of them inhibited the chemotactic response: the macrophage-activating peptide of residues 174–185 (33), and the C-terminal peptide of amino acids 191–205 (Table II). Additional experiments at peptide concentrations up to fivefold higher did not alter the results (data not shown). When FMLPP (10 nM) was used as the chemotaxin, synthetic peptides corresponding to residues 174–185 and the C-terminal residues 191–205 of CRP significantly inhibited the induced PMN chemotactic response (Table II). Thus, two of the CRP peptides, Pep 174–185 and Pep 191–205, inhibited the chemotactic response triggered by concentration gradients of either FMLPP or IL-8.

A series of mouse mAb was generated against isolated membrane proteins from U937 monocytic cells that were obtained by affinity chromatography on CRP-Sepharose. These mAbs were screened for specificity and their ability to inhibit specific ligand (CRP) binding, as described elsewhere (27, 32). The IgM mAb RC10.2, which inhibited labeled CRP binding, was tested for its ability to alter the chemotatic response of PMNs to FMLPP. Chemotaxis was inhibited at concentrations >0.1 μg/ml of purified IgM mAb per 10⁶ PMN (Fig. 2). This concentration of the mAb is sufficient to occupy >70% of the CRP-R calculated at 10⁴ receptor sites/PMN (27). Thus, the mAb to the putative CRP-R mimicked the action of CRP itself on neutrophil chemotaxis.

Since one possibility for the inhibitory action of CRP on the chemotactic response is that CRP blocks the chemokine-receptor binding interaction, the effect of CRP on labeled IL-8 binding was examined. Human CRP at concentrations from 1 to 100 μg/ml (800 nM) failed to significantly inhibit binding of 1 nM ¹²⁵I-IL-8 to HL-60(G) cells; however, at 200 μg/ml, CRP inhibited only ∼25% of the IL-8 binding at a 1600-fold molar excess (Fig. 3). In the same experiments, rhIL-8 itself inhibited ¹²⁵I-IL-8 binding by approximately 70% at a 100-fold molar excess of the unlabeled ligand (Fig. 3). Thus, it seems unlikely that CRP inhibits IL-8 mediated chemotaxis by inhibiting ligand-receptor binding.

In preliminary experiments, when the pattern of all of the tyrosine-phosphorylated proteins from HL-60(G) cells stimulated with IL-8 or FMLPP after exposure to human CRP was examined, the intensity of two prominent phosphorylated proteins of 35 to 40 kDa was increased. Since MAPK is ∼38 to 44 kDa and its activity was reported to be inducible by IL-8 over a 30-min interval, reaching maximum levels at 3 min (34), CRP was tested for its influence on MAPK activity measured as ERK-2. ERK-2 was increased by twofold in response to CRP at concentrations of 200 μg/ml at 1, 3, and 10 min after stimulation with 50 nM of IL-8 (Fig. 4,A). A plot of the cpm of ³²P-labeled MBP present in each lane is shown for comparison (Fig. 4,B). ERK-2 activity induced by FMLPP was also enhanced by preincubation with CRP (Fig. 5,A). The same blotted membrane was probed with anti-ERK-2 Ab to demonstrate that approximately the same amount of ERK-2 protein was present in each of the lanes (Fig. 5 B). CRP appears to augment MAPK activity by increasing its phosphorylation.

Since bacterial chemotactic peptides have been shown to activate the two isoforms of mitogen-activated extracellular signal-regulated kinase kinases (MEK or MAPK kinase) present in neutrophils (35, 36), we tested the effect of the MEK-specific inhibitor, PD98059, on FMLPP-induced chemotaxis. This inhibitor at its ED₅₀ of 5 μM and higher concentrations potentiated the chemotactic response (Table III). Since PI-3K is upstream of MAPK kinase (MEK) and has also been implicated in the regulation of the neutrophil chemotactic response, we examined the effects of the PI-3K-specific inhibitor, Wortmannin, and found that it inhibited the chemotactic response at its ED₅₀ of 10 nM (Table III). Therefore, PI-3K appeared to be a more logical target for the signaling triggered by CRP.

PI-3K plays a pivotal role for both IL-8 signaling (37), as well as signaling by FMLP (38) in PMN; therefore, the effects of CRP on PI-3K were tested. PI-3K activity is sensitive to the inhibitor wortmannin, which down-regulates the activity of Raf-1, B-Raf, and MAPK, but not Ras in PMNs stimulated by IL-8, C5a, and FMLP (37, 38, 39). Preliminary experiments with FMLPP at 0.1 to 100 nM revealed that maximum PI-3K activity was induced at 25 nM. Therefore, the effect of CRP on FMLPP-induced PI-3K activity was initially evaluated and found to be increased at CRP concentrations >5 μg/ml; however, the controls containing CRP in the absence of any FMLPP also displayed enhanced PI-3K activity. Therefore, the effect of treating the PMNs to different concentrations of CRP alone was tested. PI-3K activity was increased significantly at CRP levels ≥5 μg/ml (Fig. 6,A). When an immunoblot of the precipitated PI-3K was probed with an anti-p85 Ab specific for the regulatory chain of PI-3K, the relative amount of p85 per lane was the same (Fig. 6,B). Exposing the HL-60(G) cells to CRP alone for different intervals from 1 to 60 min indicated that an exposure time of only 1 min was sufficient to elevate PI-3K activity (Fig. 7,A). The optimal PI-3K response takes 3 min with FMLPP. Thus, the kinetics of PI-3K activation by CRP are rapid. The immunoblot of the samples in Figure 7,A indicated that similar amounts of p110 catalytic activity of PI-3K were examined based on the presence of the p85 subunit (Fig. 7 B). Thus, CRP may attenuate neutrophil responses via PI-3K-dependent reaction(s) in a time- and dose-dependent fashion.

Since one of the downstream effects of CRP on the G-protein-linked IL-8R that might account for inhibition of chemotaxis is desensitization by phosphorylation of one or more Ser residues within the cytoplasmic C terminus (29, 39), the IL-8RA isoform within neutrophils was immunoprecipitated and probed with an anti-phosphoserine Ab. The outcome of several such assays was that a range of CRP concentrations failed to alter the extent of Ser phosphorylation in the presence or absence of IL-8 (data not shown).

The results described in this work demonstrate that purified human CRP inhibits agonist-induced neutrophil chemotaxis in vitro by altering ligand-induced signaling. Furthermore, CRP itself initiated signaling via the specific CRP-R and did not interfere with chemokine (IL-8)-binding. CRP peptide fragments generated by neutrophil-mediated proteolysis, as well as certain synthetic CRP peptides, inhibited PMN chemotactic responses. The significance of these findings is that CRP may function as a physiologic regulator of leukocyte infiltration within the context of an acute inflammatory insult. CRP, as an inducible APR, is an ideal candidate regulator of leukocyte activities since CRP blood levels in humans become elevated as early as 6 h after tissue injury, but do not reach a maximum until 24 to 36 h, with a t_1/2 of only 19 h (2, 6, 40). During inflammation, the recruitment of particular leukocyte populations to a local site by chemokines is a critical event not only for innate host defense, but also governs the severity of any untoward tissue damage (5, 23, 41). Since leukocyte infiltration is a complex, multistage series of events, the mechanisms regulating the early events of inflammation, when there is greatly enhanced expression of proteins such as CRP, are of considerable interest because of the potential for the development of therapeutics.

Reports of the influence of human CRP on a variety of leukocyte activities in vitro are consistent with our findings of inhibition of chemotaxis. The earliest reports emphasized the ability of CRP to function as an opsonin (21, 42, 43, 44); subsequent studies with neutrophils reported that aggregated CRP potentiated the generation of intracellular reactive O₂ intermediates triggered via IgG-FcRs (22). By contrast, native pentameric CRP inhibited both degranulation and the respiratory burst of neutrophils, as detected by the extracellular release of reactive O₂ intermediates in response to a variety of agonists, including FMLP, platelet activating factor, C5a, as well as the protein kinase C-activator PMA (18, 19, 20, 45). The attenuation of the respiratory burst was proposed as a means to minimize tissue damage in which CRP was invested (43). These pronounced effects of CRP on neutrophil functions recently have been extended to in vivo experiments using mice expressing either the human or rabbit CRP transgene that confers complement-dependent resistance to pneumococci (8), or protects against both neutrophil alveolitis initiated by C5a (46) and lethal endotoxemia (10), respectively. These findings emphasize a multifunctional role for CRP during inflammation.

The effects of CRP-derived peptides on leukocyte functions, as first described by Robey et al. (15), clearly demonstrated that peptides generated by digestion of CRP by PMNs, but not intact CRP, were chemotactic for monocytes. Subsequent work by Shephard et al. (16, 17, 31) revealed that CRP proteolysis by PMNs generated peptides that inhibited both neutrophil chemotaxis and O₂⁻ production. Indeed, one of the most active peptides consisted of residues 201 to 206 (16), contained within the synthetic peptide of residues 191–205 that we show, in this work, inhibits IL-8-induced chemotaxis. Another peptide that inhibited chemotaxis was composed of residues 174 to 185, which were characterized previously as a macrophage-activating agent for tumoricidal activity in mice (33). CRP peptides have also more recently been shown to inhibit neutrophil alveolitis (47). We did not observe chemotaxis with the intact CRP pentraxin, nor with most of the CRP peptides; however, the peptides including residues 174 to 185 and 191 to 205 displayed reproducible chemotactic activity. The mAb, RC10.2, which inhibits receptor binding of CRP, mimics the effects of CRP by inhibiting chemotaxis. The effects of CRP on the two phagocytic leukocyte populations are distinct with activation of the monocyte/macrophage population and inhibition of the granulocytic responses (7).

The signaling mechanism whereby CRP alters the nicotinamide-adenine dinucleotide phosphate (NADPH) oxidase complex activity in PMNs has not yet been determined; however, one study (19) does show an alteration in the pattern of protein phosphorylation in activated neutrophils that was very similar to our initial observations on changes in the pattern of protein tyrosine phosphorylation induced by CRP. The effects of CRP on PMN also correlated with a pronounced increase in cytosolic cAMP (45) and inhibition of intracellular Ca²⁺ mobilization (48). Whether the direct effect of CRP on the respiratory burst of activated PMNs involves the same signaling pathway(s) regulating their chemotactic response remains to be determined.

The chemoattractant receptors all belong to the seven-transmembrane-receptor superfamily coupled to heterotrimeric G proteins (4, 5). Both forms of the IL-8R (CXCR1 and CXCR2), as well as the FMLP-R, are regulated by a desensitization process that requires phosphorylation of Ser and Thr residues in the C-terminal cytoplasmic region, a process that is protein kinase C or protein kinase A dependent (49, 50, 51). It seems unlikely that CRP stimulates IL-8R degradation via homologous desensitization, since CRP did not bind or compete for IL-8 binding sites. Attempts to demonstrate that CRP altered the extent of IL-8-induced Ser phosphorylation of the IL-8RA isoform in neutrophils failed to detect a significant difference. Rather, the experiments described herein with IL-8 and FMLPP suggest that the CRP-R mediates inhibition of chemotaxis by a process dependent on PI-3K. Indeed, our results show that wortmannin blocks ligand-induced chemotaxis, suggesting that PI-3K is essential, yet CRP induces PI-3K activity. This apparent contradiction may have one or more of the following explanations. CRP may block chemotaxis via a different mechanism than through PI-3K, e.g., inhibition of the activities of the small G proteins, such as Rac and/or Rho, that regulate neutrophil cytoskeletal functions (52). CRP may activate a distinct isoform of PI-3K that is not sensitive to Wortmannin and also fails to generate the appropriate 3-phosphoinositide product (53). CRP may also activate an inositol phosphatase that consumes any nascent 3-phosphoinositides (54). We are presently examining these alternative mechanisms in neutrophils exposed to CRP under various conditions. The CRP-triggered PI-3K-dependent pathway should also propagate reactions that inhibit the phosphorylation and assembly of the components of the nicotinamide-adenine dinucleotide phosphate (NADPH) oxidase (55).

We thank Dr. Pravin Kaumaya (Ohio State University) for his help with the synthesis of the C-reactive protein peptides, and Dr. Barbara Barna (Cleveland Clinic, Cleveland, OH) for the gift of peptide 174–185. We thank David Eeles for his patient word processing.