76) as a Natural Chemokine Inhibitor - The Journal of Immunology

Posttranslational Modifications Affect the Activity of the Human Monocyte Chemotactic Proteins MCP-1 and MCP-2: Identification of MCP-2(6 –76) as a Natural Chemokine Inhibitor1 Paul Proost,2* Sofie Struyf,2* Mikae¨l Couvreur,* Jean-Pierre Lenaerts,* Rene´ Conings,* Patricia Menten,* Peter Verhaert,† Anja Wuyts,* and Jo Van Damme3* Chemokines are important mediators in infection and inflammation. The monocyte chemotactic proteins (MCPs) form a subclass of structurally related C-C chemokines. MCPs select specific target cells due to binding to a distinct set of chemokine receptors. Recombinant and synthetic MCP-1 variants have been shown to function as chemokine antagonists. In this study, posttranslationally modified immunoreactive MCP-1 and MCP-2 were isolated from mononuclear cells. Natural forms of MCP-1 and MCP-2 were biochemically identified by Edman degradation and mass spectrometry and functionally characterized in chemotaxis and Ca21-mobilization assays. Glycosylated MCP-1 (12 and 13.5 kDa) was found to be two- to threefold less chemotactic for monocytes and THP-1 cells than nonglycosylated MCP-1 (10 kDa). Natural, NH2-terminally truncated MCP-1(5–76) and MCP-1(6 –76) were practically devoid of bioactivity, whereas COOH-terminally processed MCP-1(1– 69) fully retained its chemotactic and Ca21inducing capacity. The capability of naturally modified MCP-1 forms to desensitize the Ca21 response induced by intact MCP-1 in THP-1 cells correlated with their agonistic potency. In contrast, naturally modified MCP-2(6 –76) was devoid of activity, but could completely block the chemotactic effect of intact MCP-2 as well as that of MCP-1, MCP-3, and RANTES. Carboxylterminally processed MCP-2(1–74) did retain its chemotactic potency. Although comparable as a chemoattractant, natural intact MCP-2 was found to be 10-fold less potent than MCP-1 in inducing an intracellular Ca21 increase. It can be concluded that under physiologic or pathologic conditions, posttranslational modification affects chemokine potency and that natural MCP-2(6 –76) is a functional C-C chemokine inhibitor that might be useful as an inhibitor of inflammation. The Journal of Immunology, 1998, 160: 4034 – 4041.

P

osttranslational modifications (e.g., glycosylation, NH2- or COOH-terminal truncation) of natural cytokines and chemokines affect their biochemical and biologic characteristics. Chemokines are a family of chemotactic cytokines, classified as C-X-C and C-C chemokines based on the positioning of the first two cysteine residues (1–3). For some C-X-C chemokines, such as platelet basic protein, NH2-terminal processing is essential to yield a biologically active chemokine, i.e., neutrophil-activating protein-2 (4, 5). In the case of the C-X-C chemokine IL-8, NH2terminal truncation results in an increased specific activity (5– 8). However, deletion of a restricted number of residues from human granulocyte chemotactic protein-2 from the NH2 terminus does not affect the potency of this neutrophil chemoattractant (9). Under natural conditions, no cleavage of C-X-C chemokines beyond the

*Rega Institute for Medical Research, Laboratory of Molecular Immunology, and † Zoological Institute, University of Leuven, Leuven, Belgium Received for publication September 9, 1997. Accepted for publication December 15, 1997. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by the Fund for Scientific Research of Flanders (FWO), the Concerted Research Actions (GOA) of the Regional Government of Flanders, and the InterUniversity Attraction Pole (IUAP) initiative of Belgian Federal Government. P.P., S.S., P.M., P.V., and A.W. hold fellowships from the FWO. 2

P.P. and S.S. have contributed equally to this study.

3

Address correspondence and reprint requests to Dr. J. Van Damme, Rega Institute for Medical Research, Laboratory of Molecular Immunology, University of Leuven, Minderbroedersstraat 10, B-3000 Leuven, Belgium. Copyright © 1998 by The American Association of Immunologists

conserved ELR motif (in front of the first pair of cysteines and important for neutrophil activation) has been reported (1–3). For the C-C chemokines, such as RANTES4 and the monocyte chemotactic proteins (MCPs), the picture is completely different. Modification or removal of amino acids in the NH2-terminal region can completely inactivate these chemokines. RANTES becomes a potent antagonist of chemokine binding and retains its anti-HIV effect when the NH2 terminus has been artificially modified (10 –12). RANTES and the macrophage-inflammatory proteins MIP-1a and MIP-1b were reported previously as major HIVsuppressive factors (13). Chemical synthesis of MCP-1 analogues has revealed that the amino-terminal residues are functionally important for receptor recognition and signaling (14). However, the existence and role of naturally modified human MCP forms have not yet been reported. Glycosylation is known to modulate the specific biologic activities of some cytokines (15). Depending on the producer cell type, heterogeneity due to posttranslational modification has been described for human and murine MCP-1/JE (16 –19). However, the effect of glycosylation on the biologic activity of human MCP-1 has not been determined. No posttranslational modifications of MCP-2 have yet been described. On the basis of their chemotactic

4 Abbreviations used in this paper: RANTES, regulated upon activation, normal T cell expressed and secreted; [Ca21]i, intracellular calcium concentration; CCR, C-C chemokine receptor; CI, chemotactic index; FPLC, fast protein liquid chromatography; MALDI/TOF-MS, matrix-assisted laser desorption ionization/time-of-flight mass spectrometry; MCP, monocyte chemotactic protein; RP-HPLC, reversed phase high performance liquid chromatography.

0022-1767/98/$02.00

The Journal of Immunology activity for monocytes, authentic human MCP-1 and MCP-2 have been isolated and identified from leukocyte- or tumor cell-derived conditioned medium (20, 21). Under these circumstances, MCP forms that are only weakly chemotactic or completely inactive in the chemotaxis assay were not detected. In addition, modified MCPs that retain their biologic activity may have escaped during the isolation procedure if produced in minute amounts. In view of the possible agonistic or antagonistic effects of posttranslationally modified MCPs during inflammation or infection, we have searched for such physiologically produced chemokine variants. In this study, posttranslationally modified natural forms of MCP-1 and MCP-2 have been isolated from mononuclear cells and have been biochemically and functionally characterized as agonist or inhibitor of inflammatory responses.

Materials and Methods Chemokines and immunoassays rMCP-1 was a gift of Dr. J. J. Oppenheim (National Cancer Institute, Frederick, MD); MCP-2 and MCP-3 were synthesized and purified as described earlier (22); and recombinant RANTES was purchased from Peprotech (Rocky Hill, NY). Specific polyclonal Ab against rMCP-1 were raised in rabbits and purified by protein G-Sepharose chromatography. MCP-1 was detected with a classical sandwich ELISA using polyclonal rabbit anti-human MCP-1 Ab for coating and mouse anti-human MCP-1 mAb (R&D Systems, Abingdon, U.K.) for capturing. The detection was performed with peroxidase-labeled goat anti-mouse Ab and 3,39,5,59tetramethylbenzidine dihydrochloride hydrate (TMB; Aldrich Chemical, Milwaukee, WI). Specific anti-human MCP-2 Ab were obtained from mice and affinity purified on a Sepharose column to which synthetic MCP-2 was coupled using the conditions provided by the manufacturer (CNBr-activated Sepharose 4B; Pharmacia, Uppsala, Sweden). ELISA plates were coated with the affinity-purified anti-human MCP-2, and biotinylated antiMCP-2 was used as the capturing Ab. The detection was performed with peroxidase-labeled streptavidin and TMB. The detection limit for both MCP ELISAs was about 0.1 ng/ml.

Production and purification of MCPs MCPs were purified from PBMC-derived conditioned medium from 132 blood donations (Blood Transfusion Centers of Antwerp and Leuven) (23). Erythrocytes and granulocytes were removed by sedimentation in hydroxyethyl starch (Fresenius, Bad Homburg, Germany) and by gradient centrifugation in a sodium metrizoate solution (Lymphoprep; Nyegaard, Oslo, Norway). Mononuclear cells (60 3 109 cells) were incubated (5 3 106 cells/ml) with 10 mg/ml Con A and 2 mg/ml LPS. After 48 to 120 h, conditioned medium was collected and kept at 220°C until purification. Natural MCP was purified in a four-step purification procedure, as previously described (23). Briefly, the conditioned medium was concentrated on controlled pore glass or silicic acid and partially purified by affinity chromatography on a heparin-Sepharose column (Pharmacia). Fractions containing MCP immunoreactivity were further purified by Mono S (Pharmacia) cation-exchange chromatography and eluted in an NaCl gradient at pH 4. Natural MCPs were purified to homogeneity through RP-HPLC on a C-8 Aquapore RP-300 column (Perkin-Elmer, Norwalk, CT) equilibrated with 0.1% trifluoroacetic acid. Proteins were eluted in an acetonitrile gradient.

Biochemical characterization of MCP forms by SDS-PAGE, amino acid sequence analysis, and mass spectrometry The purity of column fractions was examined by SDS-PAGE under reducing conditions on Tris/tricine gels (23). Proteins were stained with silver, and the following relative molecular mass (Mr) markers were used: OVA (Mr 45,000), carbonic anhydrase (Mr 31,000), soybean trypsin inhibitor (Mr 21,500), b-lactoglobulin (Mr 18,400), lysozyme (Mr 14,400), and aprotinin (Mr 6,500). The NH2-terminal sequence of purified chemokines was determined by Edman degradation on a pulsed liquid 477A/120A protein sequencer (Perkin-Elmer) with N-methylpiperidine as a coupling base. Blocked proteins were cleaved between Asp and Pro in 75% formic acid for 48 h (24). The formic acid digest was sequenced without further purification. The Mr of MCPs was determined by matrix-assisted laser desorption ionization/time-of-flight mass spectrometry (MALDI/TOF-MS) (Micromass TofSpec, Manchester, U.K.). a-Cyano-4-hydroxycinnamic acid and cytochrome c were used as matrix and internal standard, respectively.

4035 Detection of chemotactic activity MCPs were tested for their chemotactic potency on freshly purified monocytes (2 3 106 cells/ml) or monocytic THP-1 cells (0.5 3 106 cells/ml; 2 days after subcultivation) in the Boyden microchamber using polyvinylpyrrolidone-treated polycarbonate membranes with 5 mm pore size. Samples and cells were diluted in HBSS (Life Technologies, Paisley, Scotland) supplemented with 1 mg/ml human serum albumin (Red Cross, Leuven, Belgium). After 2-h incubation at 37°C, the cells were fixed and stained with Diff-Quick staining solutions (Harleco, Gibbstown, NJ), and the cells that migrated through the membranes were counted microscopically in 10 oil immersion fields at 3500 magnification. The chemotactic index (CI) of a sample (triplicates in each chamber) was calculated as the number of cells that migrated to the sample over the number of cells that migrated to control medium (20). For desensitization experiments, THP-1 cells were incubated with biologically inactive chemokine variants for 10 min at 37°C before they were added (without washing) to the upper well of the Boyden microchamber. Only weak random motility was observed with THP-1 cells, and no chemokinesis was induced by the chemokine forms tested. The percentage of inhibition of chemotaxis toward the test sample was calculated using the CI of HBSS pretreated cells as a reference value.

Detection of intracellular Ca21 concentrations Intracellular calcium concentrations ([Ca21]i) were measured as previously described (25). Purified mononuclear cells or THP-1 cells (107 cells/ml) were incubated in Eagle’s MEM (EMEM; Life Technologies) 1 0.5% FCS with the fluorescent indicator fura-2 (fura-2/AM 2.5 mM; Molecular Probes Europe, Leiden, The Netherlands) and 0.01% Pluronic F-127 (Sigma Chemical, St. Louis, MO). After 30 min at 37°C, the cells were washed twice and resuspended at 106 cells/ml in HBSS with 1 mM Ca21 and 0.1% FCS (buffered with 10 mM HEPES/NaOH at pH 7.4). The cells were equilibrated at 37°C for 10 min before fura-2 fluorescence was measured in an LS50B luminescence spectrophotometer (Perkin-Elmer). Upon excitation at 340 and 380 nm, fluorescence was detected at 510 nm. The [Ca21]i was calculated from the Grynkiewicz equation (26). To determine Rmax, the cells were lysed with 50 mM digitonin. Subsequently, the pH was adjusted to 8.5 with 20 mM Tris, and Rmin was obtained by addition of 10 mM EGTA to the lysed cells. The Kd used was 224 nM. For desensitization experiments, monocytes or THP-1 cells were first stimulated with buffer or chemokine at different concentrations. As a second stimulus, MCPs were used at a concentration inducing a significant increase in the [Ca21]i after prestimulation with buffer. The second stimulus was applied 2 min after addition of the first stimulus. The percentage inhibition of the [Ca21]i increase in response to the second stimulus was calculated comparing the signal after prestimulation with chemokine with the signal after addition of buffer.

Results Isolation of posttranslationally modified MCP-1 forms A specific and sensitive ELISA was used to trace different MCP-1 forms produced by PBMC stimulated with mitogen and endotoxin. The conditioned medium was purified according to a standard isolation procedure (23), including adsorption to controlled pore glass and heparin-Sepharose chromatography (data not shown). Subsequent purification by FPLC mono S cation-exchange chromatography (Fig. 1A) yielded, in addition to the previously identified biologically active MCP-1 (fraction 35) (20, 21), a second peak of immunoreactivity corresponding to a substantial amount (650 mg) of MCP-1 (fractions 37 and 38). Further purification of the authentic MCP-1 (FPLC fractions 33 to 36) by C-8 RP-HPLC allowed to dissociate the activity in at least three Mr forms on SDSPAGE (13.5, 12, and 10 kDa). Molecular mass determinations using MALDI/TOF-MS revealed that the different mononuclear cell-derived MCP-1 forms have molecular masses of 9643, 9343, and 8661 Da, respectively (Table I). The latter Mr corresponds to the theoretical mass of intact unmodified MCP-1 (8664 kDa). These three MCP-1 forms were NH2-terminally blocked for Edman degradation. Their differences in molecular mass and retention time on RP-HPLC must be explained by posttranslational modifications. Because MCP-1 contains both N- and O-glycosylation sites, these MCP-1 forms most probably represent different

4036

NATURALLY MODIFIED MCP-1 AND MCP-2 AS CHEMOKINE INHIBITORS

FIGURE 1. Isolation of natural MCP-1 forms produced by mononuclear cells. MCP-1 immunoreactivity (ELISA) was fractionated by FPLC cation-exchange chromatography (A) and subsequently purified to homogeneity by RP-HPLC (B). Proteins were eluted by an NaCl and an acetonitrile gradient, respectively

glycovariants. A comparison of the theoretical and experimental (SDS-PAGE and MALDI/TOF-MS) molecular mass of the various MCP-1 forms (Table I) confirms the presence of glycosylation. Further purification of the second FPLC peak of MCP-1 immunoreactivity (FPLC fractions 37 to 38 in Fig. 1A) by RP-HPLC resulted in two MCP-1 peaks corresponding to several 8- to 12kDa protein bands (Figs. 1B and 2, upper panel). One major protein (12 kDa, HPLC fraction 38) was identified by amino acid sequence analysis as an NH2-terminally truncated form of MCP-1 missing five residues (MCP-1(6 –76)) and eluted 6 min ahead of intact MCP-1 on HPLC. Since its molecular mass (as measured by MALDI/TOF-MS) was found to be 9175 Da (Table I), which is about 1 kDa more than the theoretical molecular mass of MCP-

Table I. Biochemical characterization of natural forms of MCP-1 and MCP-2a Mr (Da) MCP Form

MCP-1(1–76) MCP-1(1–76) MCP-1(1–76) MCP-1(1–69) MCP-1(5–76) MCP-1(5–76) MCP-1(6–76) MCP-2(1–76) MCP-2(6–76)

NH2-Terminal Theoretical unglycosylated SDS-PAGE MALDI/TOF-MS Sequence

Blocked Blocked Blocked Blocked INAPVTCC INAPVTCC NAPVTCC Blocked SIPITCC

8,664 8,664 8,664 7,879 8,270 8,270 8,156 8,893 8,384

13,500 12,000 10,000 8,500 11,500 9,500 12,000 7,500 7,000

9,643 9,343 8,661 7,896 ND 8,265 9,175 8,881 8,365

a NH2-terminal amino acid sequence analysis and comparison of the experimental (SDS-PAGE and MALDI/TOF-MS) and theoretical Mr of C-8 RP-HPLC purified natural MCP-isoforms.

FIGURE 2. SDS-PAGE of posttranslationally modified MCP-1 and MCP-2 forms. Upper panel, Gelelectrophoresis of purified MCP-1 (RPHPLC fractions 34 –50 of Fig. 1B), 20 ml/lane were loaded. Lower panel, SDS-PAGE of homogeneous MCP-2: natural MCP-2(1–76) at 100 and 30 ng in lanes 1 and 2, respectively; natural MCP-2(6 –76) at 30 ng in lane 3; and synthetic MCP-2(1–76) at 60 ng in lane 4. Gels were run under reducing conditions, and proteins were stained with silver.

1(6 –76), this truncated MCP-1 form is expected to be glycosylated. In HPLC fraction 41 (Figs. 1B and 2), an NH2-terminally truncated form (11.5 kDa) of glycosylated MCP-1 missing four residues (MCP-1(5–76)) was identified by Edman degradation. Fractions 39 and 40 contained mixtures of MCP-1(5–76) and MCP-1(6 –76) and also included an unidentified NH2-terminally blocked 9-kDa protein. These proteins were not further investigated. From HPLC fraction 44 (Figs. 1B and 2), intact 10-kDa MCP-1(1–76) was recovered, confirming the faster elution pattern of MCP-1(5–76) and MCP-1(6 –76) upon HPLC. Its authenticity as nonglycosylated MCP-1 was proven by MALDI/TOF-MS and by internal sequence analysis after digestion with formic acid. In fraction 47 of the same HPLC column (Fig. 2), a minute amount of an 8.5-kDa MCP-1 form with a blocked NH2 terminus was eluted. This MCP-1 isoform (7896 Da by MALDI/TOF-MS) corresponds to unglycosylated MCP-1 (1– 69) missing seven COOH-terminal amino acids (theoretical Mr of 7879 Da). Finally, another peak of MCP-1 immunoreactivity, corresponding to a 9.5-kDa protein, was purified by RP-HPLC from the second FPLC peak (fraction 39, Fig. 1A). Edman degradation revealed that this MCP-1 form is also missing the four NH2-terminal residues. In contrast to the glycosylated 11.5-kDa MCP-1(5–76), the Mr of this protein (8265 Da on MALDI/TOF-MS) corresponded to the theoretical mass (8270 Da) of unglycosylated MCP-1(5–76).

The Journal of Immunology

4037

FIGURE 4. Calcium mobilization by naturally modified MCP-1 forms. The increase of the [Ca21]i in THP-1 cells was compared for the following MCP-1 forms: rMCP-1(1–76); 10-, 12-, and 13.5-kDa glycoforms of natural intact MCP-1(1–76) and truncated MCP-1(5–76); natural NH2-terminally truncated MCP-1(6 –76); and COOH-terminally truncated MCP-1 (1– 69). Results are the mean from two or more independent experiments. The lowest detectable increase of the [Ca21]i was 10 nM.

FIGURE 3. Comparison of the chemotactic potency of natural MCP-1 forms. Chemotactic activity on monocytic THP-1 cells (A) and on freshly isolated monocytes (B) of natural MCP-1 forms: glycosylated variants of intact MCP-1(1–76) (10- and 13.5-kDa); NH2-terminally truncated MCP1(5–76) and MCP-1(6 –76); and COOH-terminally truncated MCP-1(169). MCP-1 was tested for chemotactic activity in the microchamber assay. Results represent the mean CI 6 SEM from three or more independent experiments.

Influence of posttranslational modifications on the specific biologic activity of various human MCP-1 forms: effects on receptor desensitization Homogeneous preparations of naturally occurring MCP-1 forms, which were found to differ in either NH2-terminal truncation (MCP-1(5–76) or MCP-1(6 –76)), COOH-terminal truncation (MCP-1(1– 69)), and/or glycosylation (8.5- to 13.5-kDa MCP-1), were compared for their chemotactic potency on THP-1 monocytic cells and on freshly isolated peripheral blood monocytes in the microchamber migration assay (Fig. 3, A and B). In addition, the effect of posttranslational modifications on the capacity of MCP-1 to induce a rise in the [Ca21]i was studied in THP-1 cells (Fig. 4). Although a similar sp. act. for the different MCP-1(1–76) glycoforms (10, 12, and 13.5 kDa on SDS-PAGE) was observed in the Ca21 assay (minimal effective concentration of 1 ng/ml), glycosylated 13.5-kDa MCP-1 was somewhat less efficient than 10-kDa MCP-1(1–76) to induce monocyte chemotaxis (minimal effective concentration of 7 and 3 ng/ml, respectively). Furthermore, a relatively higher sp. act. was reached with nonglycosylated MCP1(5–76) when compared with glycosylated MCP-1(5–76) in the Ca21 assay (Fig. 4). Natural MCP-1 was about three times more potent than nonglycosylated rMCP-1 in Ca21 assays. NH2-terminally truncated MCP-1 forms (MCP-1(5–76) and MCP-1(6 –76)) were practically devoid of chemotactic activity when tested on THP-1 cells and monocytes at concentrations between 3 and 1000 ng/ml. A weak but significant effect on monocytes was only observed with MCP-1(5–76) at 100 ng/ml (Fig. 3). Both glycosylated (11.5-kDa) and certainly unglycosylated (9.5kDa) MCP-1(5–76) showed a detectable activity in the Ca21 as-

say, but their sp. act. was 5- to 50-fold lower than that of intact MCP-1(1–76). MCP-1(6 –76) remained inactive in the Ca21 assay at concentrations up to 100 ng/ml (Fig. 4). In contrast to NH2terminal truncation, COOH-terminal processing had little or no effect on the chemotactic or Ca21-releasing potency of MCP-1, because MCP-1(1– 69) was found to be equally active as intact MCP-1. To verify whether the truncated MCP-1 forms could function as natural inhibitors of intact MCP-1, receptor desensitization experiments were done by measuring the [Ca21]i in THP-1 cells. Similar to intact MCP-1(1–76), the biologically active COOH-terminally truncated MCP-1(1– 69) form could fully desensitize 3 ng/ml of intact natural 10-kDa MCP-1(1–76) when added at a 10-fold higher concentration (Table II). Inactive MCP-1(6 –76) could not desensitize intact MCP-1. Indeed, when THP-1 cells were first stimulated with 100 ng/ml of MCP-1(6 –76), only 10% inhibition of the [Ca21]i rise by 1 ng/ml of intact MCP-1(1–76) was observed, whereas a threefold excess of MCP-1(1–76) as a first stimulus was sufficient to yield 50 to 70% desensitization of intact MCP-1(1–76). When applied as the first stimulus at a 15-fold excess, glycosylated or unglycosylated MCP-1(5–76) only partially (30%) inhibited a second increase of the [Ca21]i induced by intact MCP-1(1–76). Reversibly, the effect of 15 ng/ml of MCP-1(5–76) could be blocked almost completely by pretreatment of THP-1 cells with 10 ng/ml of intact MCP-1. Thus, whereas MCP-1(5–76) is still active and able to inhibit receptor activation by intact MCP-1, MCP-1(6 –76) is inactive and not able to affect the rise in the [Ca21]i after subsequent stimulation with MCP-1(1–76). It can be concluded that the MCP-1 residue isoleucine at the fifth position is crucial for receptor interaction. Identification and functional comparison of NH2- and COOHterminally modified forms of human MCP-2 Based on the same purification and detection strategy as used for natural MCP-1, different forms of MCP-2 were isolated from mitogen- and endotoxin-stimulated mononuclear cells. However, 10 times less MCP-2 than MCP-1 was obtained. In addition to authentic 7.5-kDa MCP-2(1–76), an NH2-terminally truncated 7-kDa form of MCP-2 missing five residues (MCP-2(6 –76)) was purified

4038

NATURALLY MODIFIED MCP-1 AND MCP-2 AS CHEMOKINE INHIBITORS Table II. Desensitization of Ca21 responses induced in THP-1 cells by natural MCP-1 forms First Stimulus

Chemokine

Buffer MCP-1(1–76)

MCP-1(1–69) MCP-1(6–76) MCP-1(5–76)b MCP-1(5–76)c MCP-1(1–76)

Second Stimulus

Concentration (ng/ml)

Increase in [Ca21]i (nM)a

3 3 10 30 0.4 4 40 10 30 100 1.5 5 15 5 15 50 10

,10 64 50 100 159 ,10 43 111 ,10 ,10 15 ,10 31 61 ,10 20 39 75

Chemokine

MCP-1(1–76) MCP-1(1–76)

MCP-1(1–76) MCP-1(1–76) MCP-1(1–76) MCP-1(1–76) MCP-1(5–76)


Increase in [Ca21]i (nM)a

% Inhibition of the second stimulus by the first stimulus

3 1 3 3 3 3 3 3 1 1 1 1 1 1 3 3 3 15

53 16 32 13 ,10 45 36 ,10 35 32 30 33 34 25 49 44 34 18

53 23 67 100 8 27 100 0 4 10 0 0 24 0 10 31 70

a

Results shown are the mean of two or more experiments. b Unglycosylated 9.5-kDa MCP-1(5–76). c Glycosylated 11.5-kDa MCP-1(5–76).

to homogeneity by RP-HPLC and identified by amino acid sequence analysis (Fig. 2). MALDI/TOF-MS (Table I) yielded a molecular mass of 8881 Da for intact MCP-2 (theoretical Mr of 8893 Da), whereas for MCP-2(6 –76) a molecular mass of 8365 Da was measured, confirming the deletion of the five NH2-terminal amino acids (theoretical Mr of 8384 Da). Functional comparison of these natural MCP-2 forms in the THP-1 chemotaxis assay showed that intact MCP-2 is still active at 5 ng/ml, whereas truncated MCP2(6 –76) remains devoid of chemotactic activity when tested at a concentration range from 0.6 to 60 ng/ml (Fig. 5). Intact natural MCP-2 was also compared in potency with the synthetical MCP2(1–76) and a COOH-terminally truncated synthetical form (22) missing two residues (MCP-2(1–74)). The minimal effective chemotactic concentration of these forms was also found to be 5 ng/ml (Fig. 5). Although in chemotaxis assays the sp. act. of natural

FIGURE 5. Comparison of the chemotactic potency of modified MCP-2 forms. Intact natural and synthetic MCP-2(1–76), NH2-terminally truncated natural MCP-2(6 –76), and COOH-terminally truncated synthetic MCP-2(1–74) were tested for chemotactic activity on THP-1 cells. Results represent the mean CI 6 SEM from four or more independent experiments.

intact MCP-1 and MCP-2 is comparable (20), the calcium-mobilizing capacity of MCP-2 is still a matter of debate. However, in Ca21-mobilization experiments, the minimal effective dose for both natural or synthetic MCP-2(1–76) was about 10-fold higher compared with that of natural intact MCP-1(1–76) (Fig. 6), whereas MCP-2(6 –76) remained inactive (data not shown). Nevertheless, intact MCP-2 (50 ng/ml) was capable of desensitizing for MCP-2 (15 ng/ml) and MCP-3 (10 ng/ml), yielding 52 and 45% inhibition, respectively (data not shown). Due to this lower sp. act. of MCP-2 in Ca21 assays, desensitization of chemotaxis by MCP-2(6 –76) was performed in the Boyden microchamber. Since intact MCP-2 is reported to cross-desensitize with active MCP-1, MCP-2, and MCP-3 in the monocyte chemotaxis assay (27), we investigated whether natural, inactive MCP-2(6 –76) could also desensitize for MCP-1, MCP-2, MCP-3, and RANTES (Table III). Preincubation of THP-1 cells with 100 ng/ml of inactive MCP-2(6 –76) could already significantly inhibit chemotaxis induced by 10 ng/ml MCP-1 (63%), 5 ng/ml MCP-2 (75%), 30 ng/ml MCP-3 (62%), and 100 ng/ml RANTES (75%). Moreover, chemotaxis in response to three times lower concentrations of the respective MCPs was completely (91–100%) inhibited by 100 ng/ml MCP-2(6 –76). Furthermore, at a concentration as low as 10 ng/ml, MCP-2(6 –76) was still able to significantly inhibit the chemotactic activity induced by MCP-1 (3 ng/ml), MCP-2 (1.5 ng/ml) or MCP-3 (10 ng/ml), and RANTES (30 ng/ml). Taken together, MCP-2(6 –76) is produced naturally, is inactive as a chemoattractant, and inhibits several C-C chemokines, the effect being most predominant for MCP-3. To confirm the inhibitory capacity of MCP-2(6 –76), the chemokine was also added to the lower wells of the microchamber together with authentic MCP-2 or MCP-3 as agonist. In this test system, the chemotactic effect of 1.5 ng/ml MCP-2 and 3 ng/ml MCP-3 was inhibited by 10 ng/ml MCP-2(6 –76) for 66% and for 84%, respectively (Table IV).

Discussion Although the structurally highly related MCPs form a branch of the C-C chemokine subfamily, they each attract distinct sets of

The Journal of Immunology

4039 Table IV. MCP-2(6 –76) inhibits THP-1 cell chemotaxis induced by MCP-2 and MCP-3 Inhibition of Chemotactic Responseb Chemokinea

MCP-2 MCP-3


Buffer

% Inhibition 10 ng/ml MCP-2(6–76) of Chemotaxis

1.5 3

6.9 6 1.3 5.5 6 1.3

2.8 6 0.3 2.8 6 1.5

66 6 5.6 84 6 9.9

a MCP-2 and MCP-3 were added as chemoattractants to the lower wells of the microchamber in the absence (buffer) or presence of 10 ng/ml MCP-2(6 –76). b Mean CI 6 SEM of four independent experiments.

FIGURE 6. Natural MCP-2 is a weaker agonist than MCP-1 to mobilize calcium in monocytes. Intact natural 10-kDa MCP-1 (3 and 30 ng/ml) and intact MCP-2 (15, 50, and 150 ng/ml) dose-dependently increase the [Ca21]i in THP-1 cells. The result of one representative experiment of two is shown.

leukocyte types (monocytes, lymphocytes, dendritic cells, NK cells, basophils, or eosinophils), which is explained by differential binding to unique and shared receptors expressed on these target cells. In addition to CCR-2b, which is a common receptor for all of the MCPs, MCP-2 and MCP-3 bind to CCR-1 and CCR-3 (3, 28 –30). Among chemokines, MCPs show the unique feature of an NH2-terminal pyroglutamate, which is important for biologic activity. Modification of chemokines by chemical synthesis or mutagenesis technology is considered as a tool to develop antagonists of inflammation. In this context, the natural occurrence of posttranslational modifications (glycosylation, NH2- and COOHterminal truncation) of C-X-C chemokines in vitro and in vivo has been reported extensively (1–9). Schro¨der et al. (1992) have isolated different forms of IL-8 and GRO (growth-related oncogene) from psoriatic scales (31). In contrast, for C-C chemokines, only artificially modified or recombinant molecules have been studied as antagonists (10, 11, 14, 32). Using immunoassays, a number of natural, posttranslationally processed MCP-1 forms were purified to homogeneity from the conditioned medium of mononuclear cells. Different glycoforms of intact natural MCP-1 (MCP-1(1–76)) with an Mr between 10 and 13.5 kDa were biochemically characterized by SDS-PAGE, amino acid sequence analysis, and MALDI/TOF-MS (Table I and Fig. 2). Similarly, one COOH-terminally and two NH2-terminally truncated natural MCP-1 forms (MCP-1(1– 69), MCP-1(5–76), and MCP-1(6 –76)) have been isolated. Although the MCP-1 protein

Table III. MCP-2(6 –76) desensitizes the monocytic THP-1 cell chemotactic responses of MCP-1, MCP-2, MCP-3, and RANTES in the microchamber Inhibition of Chemotactic Responseb,c Chemokinea

MCP-1 MCP-2 MCP-3 RANTES

MCP-1 MCP-2 MCP-3 RANTES a b c


Buffer

100 ng/ml MCP-2(6–76)

% Inhibition of Chemotaxis

10 3 5 1.5 30 10 100 30

22.3 6 7.9 15.0 6 8.0 36.0 6 12.6 6.7 6 1.4 13.2 6 0.4 3.0 6 1.5 6.3 6 0.8 4.0 6 0.8

8.3 6 3.8 1.3 6 0.3 10.8 6 6.1 1.5 6 0.3 6.0 6 4.0 ,1 2.6 6 1.3 1.5 6 0.3

63 6 21 99 6 1.0 75 6 8.0 91 6 7.0 62 6 31 100 6 0.0 75 6 19 77 6 16

Buffer

12.7 6 2.3 7.5 6 0.0 38.0 6 5.3 18.3 6 4.6 13.2 6 1.9 7.7 6 1.4 5.5 6 0.6 3.2 6 0.7

10 ng/ml MCP-2(6–76)

10 3 5 1.5 30 10 100 30

10.5 6 3.8 3.0 6 0.3 27.2 6 4.9 9.2 6 1.4 8.0 6 1.0 1.7 6 0.3 5.8 6 0.9 2.5 6 0.5

MCP-1, MCP-2, MCP-3, or RANTES were added as chemoattractants to the lower wells. The upper wells of the microchamber were filled with THP-1 cells preincubated with MCP-2(6 –76) or with buffer. Mean CI 6 SEM of three independent experiments.

24 6 1.8 69 6 4.0 30 6 6.0 45 6 23 37 6 19 90 6 6.0 217 6 7.0 39 6 18

4040

NATURALLY MODIFIED MCP-1 AND MCP-2 AS CHEMOKINE INHIBITORS

structure contains one consensus sequence for N-linked sugars in the NH2-terminal region (16), N-glycosylation could not be detected for natural MCP-1 (15, 19). Furthermore, MCP-1 is known to be sialylated and O-glycosylated (15, 19). Glycosylation did not protect MCP-1 against NH2-terminal truncation since both MCP1(5–76) and MCP-1(6 –76) were isolated as glycosylated molecules, as shown by MALDI/TOF mass determination and by their apparent relative molecular mass on SDS-PAGE. The sp. act. of unglycosylated MCP-1 was two- to threefold higher than that of glycosylated MCP-1. In this respect, a highly glycosylated 30-kDa form of murine MCP-1/JE was also found to be less chemotactic for monocytes and lymphocytes than the 7-kDa unglycosylated molecule (33). Although in one study it was concluded that human MCP-1 glycosylation did not significantly affect its capacity to induce monocyte migration (34), another report confirms a threefold lower sp. act. for the most glycosylated MCP-1 form (35). Taken together, although O-glycosylation may constitute a significant portion of the relative molecular mass of MCP-1, it seems to interfere only weakly with receptor binding and activation. In contrast, limited NH2-terminal truncation of MCP-1, but not COOH-terminal truncation, significantly reduced its chemotactic potential. Indeed, MCP-1(5–76) was only active at high concentrations, and the activity completely disappeared by the removal of one extra amino acid (MCP-1(6 –76)). This finding is in accordance with a study on monocytic cells using synthetic NH2-terminally truncated MCP-1 analogues (14), showing that the integrity of residues 1 to 6 is required for functional activity. In the calcium assay, MCP-1(5–76) and MCP-1(1– 69) could both desensitize for intact MCP-1, but only at concentrations at which these truncated molecules became active by themselves. Inactive MCP1(6 –76) was unable to desensitize for MCP-1(1–76). The biological significance of the natural MCP-1(5–76) and MCP-1(6 –76) forms isolated from mononuclear cells needs to be investigated in more detail, e.g., on cells other than monocytes. Indeed, it has been reported that deletion of the NH2-terminal residue converts MCP-1 into an eosinophil chemoattractant and dramatically decreases its basophil-activating potency (36). A possible antagonistic effect of truncated MCP-1 forms on these cell types has not been investigated in this study due to the lack of sufficient material. Although natural and synthetic MCP-2 are as active as MCP-1 and MCP-3 in chemotaxis assays (20), such intact MCP-2 was found to be 10-fold weaker compared with MCP-1 and MCP-3 in Ca21-mobilization tests on THP-1 cells. However, intact natural MCP-2 could desensitize the MCP-2- and MCP-3-induced Ca21 mobilization for about 50%, when applied as first stimulus at a three- to fivefold excess (data not shown). This is in agreement with the finding that a relatively low concentration of synthetic MCP-2 still desensitizes the MCP-1- or MCP-3-induced chemotaxis and calcium release (27, 37). Contrasting reports (28, 29, 38) on MCP-2 signaling through CCR-2 can be explained by the weak Ca21 response induced by MCP-2. Indeed, MCP-2 was able to bind to CCR-2-transfected cells and to inhibit adenylylcyclase, but only a very weak Ca21 response was detected in these transfected cells (28). In addition, although MCP-2 bound well to CCR-2 (and CCR-1)-transfected cells, MCP-2 only poorly competed for radioactive MCP-1 binding (28, 29). Thus, MCP-2, compared with MCP-1, differently interacts with CCR-2, which may partially explain different signal-transduction cascades used by the two chemokines. In contrast to monocyte chemotaxis toward MCP-1 and MCP-3, MCP-2-induced chemotaxis was cholera (but not pertussis) toxin sensitive (27). This finding implies that MCP-2 signal transduction may, as opposed to other chemokines, use Gas instead of Gai proteins. Coupling of one receptor to different Ga proteins is a common feature for this type of receptor, allowing spreading

of the response to the stimulus over several pathways (39). CCR-2 and CCR-1 possess the ability to associate with multiple G proteins (40). Finally, the different behavior of MCP-2 (Ca21 response and G protein coupling) can also be explained by the existence of a so far unidentified specific MCP-2 receptor. Posttranslationally modified MCP-2 forms have not yet been described. Natural NH2-terminally truncated MCP-2(6 –76) possesses no monocyte chemotactic activity, nor did it induce Ca21 mobilization in THP-1 cells. In contrast to inactive MCP-1(6 –76), inactive MCP-2(6 –76) desensitized for intact MCP-2 (Tables II and III). Indeed, MCP-2(6 –76) at 100 ng/ml inhibited chemotaxis, induced by a 3- to 30-fold lower dose of MCP-1, MCP-2, MCP-3, or RANTES, for more than 75%. Thus, although MCP-2 did not compete well for MCP-1 binding to CCR-2 (28), in terms of inhibition of chemotactic activity, inactive MCP-2(6 –76) is the first isolated natural inhibitor for MCP-2 and MCP-1, as well as MCP-3 and RANTES. This is in accordance with the observation that MCP-2 also signals through CCR-1 (29). Indirect evidence for usage of CCR-3 (the C-C chemokine receptor specifically expressed on eosinophils) by MCP-2 has been provided with the use of a mAb directed against CCR-3 (30). Whether this natural truncated MCP-2(6 –76) can also inhibit binding of MCP-3, MCP-4, eotaxin, and RANTES to CCR-3 remains to be elucidated. Finally, it remains at present not clear why MCP-1(6 –76) is a much weaker chemokine inhibitor compared with MCP-2(6 –76). The search for chemokine variants as receptor antagonists has increased since the discovery that chemokines function as suppressive factors for HIV infection (13). Recently, a number of artificially modified chemokines, e.g., RANTES derivatives with an altered NH2 terminus (through chemical synthesis or mutagenesis), have been described as potent antagonists of HIV-1 binding (10, 12). In this study, we demonstrate that potent chemokine inhibitors can be generated under natural conditions. Indeed, proteolytic cleavage of intact MCP-2 to MCP-2(6 –76) converts this chemokine into a potent inhibitor of C-C chemokine-induced chemotaxis. Such an endogenous negative feedback mechanism could be functional during disease, especially during acute inflammation. For example, septic shock results in dramatic increases in MCP-1 and MCP-2 levels in the circulation (41). Although a number of leukocyte-derived proteases are known to be secreted during such inflammatory responses, the molecules contributing to this complex posttranslational processing of MCPs are not well known. Very recently, we have found that RANTES, but not MCPs, is NH2-terminally truncated by dipeptidyl peptidase IV (CD26). Similar to MCP-2(6 –76), RANTES(3– 68) showed a dramatic reduction in chemotactic potency without losing its potential to inhibit chemotaxis (data not shown). In contrast to MCP-2, MCP-1 has already been investigated extensively in animal models (inflammation, transgenic mice) (42, 43). Recently, the synthetic chemokine antagonist MCP-1(9 –76) has been shown to inhibit chronic inflammatory arthritis in mouse (44). The exact mechanism of inhibition of the disease by this chemokine antagonist is not fully understood yet. It should be investigated whether the levels of naturally generated chemokine inhibitors, as described in this work, are effective during infection or inflammation. For that purpose, such naturally modified chemokines need to be present both at an early stage of the disease (acute infection) and in sufficient amounts (chronic inflammation). From this study, it can be deduced that the modified MCP-1 and MCP-2 forms were produced in vitro in amounts that are 3- to 10-fold lower than those of intact chemokine. If this represents the in vivo situation during disease, administration of excess exogeneous chemokine antagonist might be beneficial.

The Journal of Immunology The data presented in this work must also be considered in relation to the recent finding that Kaposi’s sarcoma-associated herpesvirus encodes chemokine-like proteins, which bind to, but do not induce signal transduction through, known C-C and C-X-C chemokine receptors. Similar to naturally modified chemokines, these viral chemokines can block binding of endogenous chemokines or HIV and function as inhibitors of chemotaxis or infection (45, 46).

Acknowledgments The technical support of Willy Put is greatly appreciated. We thank Ghislain Opdenakker for the critical evaluation of the manuscript.

References 1. Baggiolini, M., B. Dewald, and B. Moser. 1994. Interleukin-8 and related chemotactic cytokines—CXC and CC chemokines. Adv. Immunol. 55:97. 2. Taub, D. D., and J. J. Oppenheim. 1994. Chemokines, inflammation and the immune system. Ther. Immunol. 1:229. 3. Proost, P., A. Wuyts, and J. Van Damme. 1996. The role of chemokines in inflammation. Int. J. Clin. Lab. Res. 26:211. 4. Walz, A., B. Dewald, V. von Tscharner, and M. Baggiolini. 1989. Effects of the neutrophil-activating peptide NAP-2, platelet basic protein, connective tissue activating peptide III and platelet factor 4 on human neutrophils. J. Exp. Med. 170:1745. 5. Van Damme, J., J. Van Beeumen, R. Conings, B. Decock, and A. Billiau. 1989. Purification of granulocyte chemotactic peptide/interleukin-8 reveals N-terminal sequence heterogeneity similar to that of b-thromboglobulin. Eur. J. Biochem. 181:337. 6. He´bert, C. A., F. W. Luscinskas, J.-M. Kiely, E. A. Luis, W. C. Darbonne, G. L. Bennett, C. C. Liu, M. S. Obin, M. A. Gimbrone, Jr., and J. B. Baker. 1990. Endothelial and leukocyte forms of IL-8: conversion by thrombin and interactions with neutrophils. J. Immunol. 145:3033. 7. Nourshargh, S., J. A. Perkins, H. J. Showell, K. Matsushima, T. J. Williams, and P. D. Collins. 1992. A comparative study of the neutrophil stimulatory activity in vitro and pro-inflammatory properties in vivo of 72 amino acid and 77 amino acid IL-8. J. Immunol. 148:106. 8. Clark-Lewis, I., C. Schumacher, M. Baggiolini, and B. Moser. 1991. Structureactivity relationships of interleukin-8 determined using chemically synthesized analogs: critical role of NH2-terminal residues and evidence for uncoupling of neutrophil chemotaxis, exocytosis, and receptor binding activities. J. Biol. Chem. 266:23128. 9. Proost, P., A. Wuyts, R. Conings, J.-P. Lenaerts, A. Billiau, G. Opdenakker, and J. Van Damme. 1993. Human and bovine granulocyte chemotactic protein-2: complete amino acid sequence and functional characterization as chemokines. Biochemistry 32:10170. 10. Arenzana-Seisdedos, F., J.-L. Virelizier, D. Rousset, I. Clark-Lewis, P. Loetscher, B. Moser, and M. Baggiolini. 1996. HIV blocked by chemokine antagonist. Nature 383:400. 11. Proudfoot, A. E., C. A. Power, A. J. Hoogewerf, M. O. Montjovent, F. Borlat, R. E. Offord, and T. N. Wells. 1996. Extension of recombinant human RANTES by the retention of the initiating methionine produces a potent antagonist. J. Biol. Chem. 271:2599. 12. Simmons, G., P. R. Clapham, L. Picard, R. E. Offord, M. M. Rosenkilde, T. W. Schwartz, R. Buser, T. N. C. Wells, and A. E. I. Proudfoot. 1997. Potent inhibition of HIV-1 infectivity in macrophages and lymphocytes by a novel CCR5 antagonist. Science 276:276. 13. Cocchi, F., A. L. DeVico, A. Garzino-Demo, S. K. Arya, R. C. Gallo, and P. Lusso. 1995. Identification of RANTES, MIP-1a and MIP-1b as the major HIV-suppressive factors produced by CD81 T cells. Science 270:1811. 14. Gong, J.-H., and I. Clark-Lewis. 1995. Antagonists of monocyte chemoattractant protein 1 identified by modification of functionally critical NH2-terminal residues. J. Exp. Med. 181:631. 15. Opdenakker, G., P. M. Rudd, M. Wormald, R. A. Dwek, and J. Van Damme. 1995. Cells regulate the activities of cytokines by glycosylation. FASEB J. 9:453. 16. Rollins, B. J., P. Stier, T. Ernst, and G. G. Wong. 1989. The human homolog of the JE gene encodes a monocyte secretory protein. Mol. Cell. Biol. 9:4687. 17. Rollins, B. J., and M. E. Sunday. 1991. Suppression of tumor formation in vivo by expression of the JE gene in malignant cells. Mol. Cell. Biol. 11:3125. 18. Rollins, B. J., E. D. Morrison, and C. D. Stiles. 1988. Cloning and expression of JE, a gene inducible by platelet-derived growth factor and whose product has cytokine-like properties. Proc. Natl. Acad. Sci. USA 85:3738. 19. Jiang, Y., A. J. Valente, M. J. Williamson, L. Zhang, and D. T. Graves. 1990. Post-translational modification of a monocyte-specific chemoattractant synthesized by glioma, osteosarcoma, and vascular smooth muscle cells. J. Biol. Chem. 265:18318. 20. Van Damme, J., P. Proost, J.-P. Lenaerts, and G. Opdenakker. 1992. Structural and functional identification of two human, tumor-derived monocyte chemotactic proteins (MCP-2 and MCP-3) belonging to the chemokine family. J. Exp. Med. 176:59. 21. Van Damme, J., B. Decock, J.-P. Lenaerts, R. Conings, R. Bertini, A. Mantovani, and A. Billiau. 1989. Identification by sequence analysis of chemotactic factors for monocytes produced by normal and transformed cells stimulated with virus, double-stranded RNA or cytokine. Eur. J. Immunol. 19:2367.

4041 22. Proost, P., P. Van Leuven, A. Wuyts, R. Ebberink, G. Opdenakker, and J. Van Damme. 1995. Chemical synthesis, purification and folding of the human monocyte chemotactic proteins MCP-2 and MCP-3 into biologically active chemokines. Cytokine 7:97. 23. Proost, P., A. Wuyts, R. Conings, J.-P. Lenaerts, W. Put, and J. Van Damme. 1996. Purification and identification of natural chemokines. Methods: Companion Methods Enzymol. 10:82. 24. Piszkiewicz, D., M. Landon, and E. L. Smith. 1970. Anomalous cleavage of aspartyl-proline peptide bonds during amino acid sequence determinations. Biochem. Biophys. Res. Commun. 40:1173. 25. Wuyts, A., N. Van Osselear, A. Haelens, I. Samson, P. Herdewijn, A. Ben-Baruch, J. J. Oppenheim, P. Proost, and J. Van Damme. 1997. Characterization of synthetic human granulocyte chemotactic protein 2: usage of chemokine receptors CXCR1 and CXCR2 and in vivo inflammatory properties. Biochemistry 36:2716. 26. Grynkiewicz, G., M. Poenie, and R. Y. Tsien. 1985. A new generation of Ca21 indicators with greatly improved fluorescence properties. J. Biol. Chem. 260:3440. 27. Sozzani, S., D. Zhou, M. Locati, M. Rieppi, P. Proost, M. Magazin, N. Vita, J. Van Damme, and A. Mantovani. 1994. Receptors and transduction pathways for monocyte chemotactic protein-2 and monocyte chemotactic protein-3: similarities and differences with MCP-1. J. Immunol. 152:3615. 28. Yamagami, S., H. Tanaka, and N. Endo. 1997. Monocyte chemoattractant protein-2 can exert its effects through the MCP-1 receptor (CC CKR2B). FEBS Lett. 400:329. 29. Gong, X., W. Gong, D. B. Kuhns, A. Ben-Baruch, O. M. Z. Howard, and J. M. Wang. 1997. Monocyte chemotactic protein-2 (MCP-2) uses CCR1 and CCR2B as its functional receptors. J. Biol. Chem. 272:11682. 30. Heath, H., S. Qin, P. Rao, L. Wu, G. LaRosa, N. Kassam, P. D. Ponath, and C. R. Mackay. 1997. Chemokine receptor usage by human eosinophils: the importance of CCR3 demonstrated using an antagonistic monoclonal antibody. J. Clin. Invest. 99:178. 31. Schro¨der, J.-M., H. Gregory, J. Young, and E. Christophers. 1992. Neutrophilactivating proteins in psoriasis. J. Invest. Dermatol. 98:241. 32. Masure, S., L. Paemen, P. Proost, J. Van Damme, and G. Opdenakker. 1995. Expression of a human mutant monocyte chemotactic protein 3 in Pichia pastoris and characterization as an MCP-3 receptor antagonist. J. Interferon Cytokine Res. 15:955. 33. Liu, Z.-G., A. Haelens, A. Wuyts, S. Struyf, X.-W. Pang, P. Proost, W.-F. Chen, and J. Van Damme. 1996. Isolation of a lymphocyte chemotactic factor produced by the murine thymic epithelial cell line MTEC1: identification as a 30 kDa glycosylated form of MCP-1. Eur. Cytokine Netw. 7:381. 34. Jiang, Y., L. A. Tabak, A. J. Valente, and D. T. Graves. 1991. Initial characterization of the carbohydrate structure of MCP-1. Biochem. Biophys. Res. Commun. 178:1400. 35. Ishii, K., S. Yamagami, H. Tanaka, M. Motoki, Y. Suwa, and N. Endo. 1995. Full active baculovirus-expressed human monocyte chemoattractant protein 1 with the intact N-terminus. Biochem. Biophys. Res. Commun. 206:955. 36. Weber, M., M. Uguccioni, M. Baggiolini, I. Clark-Lewis, and C. A. Dahinden. 1996. Deletion of the NH2-terminal residue converts monocyte chemotactic protein 1 from an activator of basophil mediator release to an eosinophil chemoattractant. J. Exp. Med. 183:681. 37. Uguccioni, M., M. D’Apuzzo, M. Loetscher, B. Dewald, and M. Baggiolini. 1995. Actions of the chemotactic cytokines MCP-1, MCP-2, MCP-3, RANTES, MIP-1a and MIP-1b on human monocytes. Eur. J. Immunol. 25:64. 38. Franci, C., L. M. Wong, J. Van Damme, P. Proost, and I. F. Charo. 1995. Monocyte chemoattractant protein-3, but not monocyte chemoattractant protein-2, is a functional ligand of the human monocyte chemoattractant protein-1 receptor. J. Immunol. 154:6511. 39. Neer, E. J. 1995. Heterotrimeric G proteins: organizers of transmembrane signals. Cell 80:249. 40. Arai, H., and I. F. Charo. 1996. Differential regulation of G-protein-mediated signaling by chemokine receptors. J. Biol. Chem. 271:21814. 41. Bossink, A. W. J., L. Paemen, P. M. Jansen, C. E. Hack, L. G. Thijs, and J. Van Damme. 1995. Plasma levels of the chemokines monocyte chemotactic proteins-1 and -2 are elevated in human sepsis. Blood 86:3841. 42. Chensue, S. W., K. S. Warmington, J. H. Ruth, P. S. Sanghi, P. Lincoln, and S. L. Kunkel. 1996. Role of monocyte chemoattractant protein-1 (MCP-1) in Th1 (mycobacterial) and Th2 (schistosomal) antigen-induced granuloma formation: relationship to local inflammation, Th cell expression, and IL-12 production. J. Immunol. 157:4602. 43. Rutledge, B. J., H. Rayburn, R. Rosenberg, R. J. North, R. P. Gladue, C. L. Corless, and B. J. Rollins. 1995. High level monocyte chemoattractant protein-1 expression in transgenic mice increases their susceptibility to intracellular pathogens. J. Immunol. 155:4838. 44. Gong, J.-H., L. G. Ratkay, J. D. Waterfield, and I. Clark-Lewis. 1997. An antagonist of monocyte chemoattractant protein 1 (MCP-1) inhibits arthritis in the MRL-lpr mouse model. J. Exp. Med. 186:131. 45. Boshoff, C., Y. Endo, P. D. Collins, Y. Takeuchi, J. D. Reeves, V. L. Schweickart, M. A. Siani, T. Sasaki, T. J. Williams, P. W. Gray, P. S. Moore, Y. Chang, and R. A. Weiss. 1997. Angiogenic and HIV-inhibitory functions of KSHV-encoded chemokines. Science 278:290. 46. Kledal, T. N., M. M. Rosenkilde, F. Coulin, G. Simmons, A. H. Johnsen, S. Alouani, C. A. Power, H. R. Lu¨ttichau, J. Gerstoft, P. R. Clapham, I. Clark-Lewis, T. N. C. Wells, and T. W. Schwartz. 1997. A broad-spectrum chemokine antagonist encoded by Kaposi’s sarcoma-associated herpesvirus. Science 277:1656.