Macrophage stimulating protein: purification, partial amino acid sequence, and cellular activity.

Macrophage stimulating protein (MSP) was purified to homogeneity from human blood plasma by selection of biologically active fractions obtained by sequential immunoaffinity and high pressure liquid ion exchange chromatography. By sodium dodecyl sulfate-polyacrylamide gel electrophoresis the molecular mass of MSP was 70 kilodaltons (kD); under reducing conditions two gel bands were seen, at 47 and 22 kD. The disulfide-linked two-chain structure of MSP was confirmed by separation of reduced and alkylated MSP chains. A computer search comparison of six partial sequences of MSP digests showed that MSP has not been recorded in data banks of protein sequences. Two MSP fragments had greater than 80% identity in overlaps of 12-16 residues to sequences in the protein family that includes human prothrombin, plasminogen, and hepatocyte growth factor. The concentration of purified MSP required for half-maximal biological activity was the order of 10(-10) M. In addition to making mouse resident peritoneal macrophages responses to chemoattractants, MSP caused the appearance of long cytoplasmic processes and pinocytic vesicles in freshly plated macrophages. MSP also caused phagocytosis via the C3b receptor, CR1. Whereas resident peritoneal macrophages bind but do not ingest sheep erythrocytes opsonized with IgM anti-Forssman antibody and mouse C3b, addition of MSP caused ingestion. Thus, MSP causes direct or indirect activation of two receptors of the mouse resident peritoneal macrophage, CR1 and the C5a receptor.

Identification of the ron gene product as the receptor for the human macrophage stimulating protein.

Macrophage-stimulating protein (MSP) is a member of the hepatocyte growth factor-scatter factor (HGF-SF) family. Labeled MSP bound to Madin-Darby canine kidney (MDCK) cells transfected with complementary DNA encoding Ron, a cell membrane protein tyrosine kinase. Cross-linking of 125I-labeled MSP to transfected cells (MDCK-RE7 cells) and immunoprecipitation by antibodies to Ron revealed a 220-kilodalton complex, a size consistent with that of MSP (80 kilodaltons) cross-linked to the beta chain of Ron (150 kilodaltons). The binding of 125I-labeled MSP to MDCK-RE7 cells was inhibited by unlabeled MSP, but not by HGF-SF. MSP caused phosphorylation of the beta chain of Ron and induced migration of MDCK-RE7 cells. These results establish the ron gene product as a specific cell-surface receptor for MSP.

Proteolytic cleavage and activation of pro-macrophage-stimulating protein by resident peritoneal macrophage membrane proteases.

Macrophage stimulating protein (MSP), which is secreted as biologically inactive pro-MSP, is activated to MSP by cleavage at a single peptide bond. Our objectives were to determine the form of MSP in circulating blood and to study proteolytic activation of pro-MSP by its target cell. Western blot of immunoaffinity-purified serum MSP showed that all the protein was pro-MSP, without detectable MSP. The circulating form of the protein is therefore pro-MSP, and conversion to MSP does not occur when blood is shed. Incubation of radiolabeled pro-MSP with murine peritoneal macrophages caused proteolytic cleavage to predominantly inactive fragments. Among several protease inhibitors, soybean trypsin inhibitor was one of two that inhibited nonspecific cleavage and revealed a macrophage proteolysis of pro-MSP, and certain concentrations enhanced cleavage to mature MSP. Macrophage membranes had nonspecific and specific pro-MSP proteolytic activity, which was not present in macrophage culture fluids. The results suggest that control of MSP activity can occur at the level of the target cell by proteolytic cleavage of pro-MSP to mature MSP or to inactive fragments.

Two independent signaling pathways mediate the antiapoptotic action of macrophage-stimulating protein on epithelial cells.

In addition to its effects on macrophage function, macrophage-stimulating protein (MSP) is a growth and motility factor for epithelial cells. The growth and survival of epithelial cells generally require two signals, one generated by interaction with extracellular matrix via integrins, the other initiated by a growth factor. Therefore we investigated the effect of MSP on epithelial cell survival. Survival of epithelial cells cultured overnight in serum-free medium was promoted by adhesion, which activated both the phosphatidylinositol 3'-kinase (PI3-K)/AKT and mitogen-activated protein kinase (MAPK) pathways, operating independently of one another. The number of apoptotic cells resulting from inhibition of either pathway alone was approximately doubled by simultaneous inhibition of both pathways. This shows that each pathway made a partial contribution to the prevention of apoptosis. In the presence of an inhibitor of either pathway, MSP increased the activity of the other pathway so that the single uninhibited pathway alone was sufficient to prevent apoptosis. In contrast to the results with adherent cells, although MSP also prevented apoptosis of cells in suspension (anoikis), its effect was mediated only by the PI3-K/AKT pathway. Despite activation of MAPK by MSP, anoikis was not prevented in suspended cells with a blocked PI3-K/AKT pathway. Thus, activation of MAPK alone is not sufficient to mediate MSP antiapoptotic effects. Cell adhesion generates an additional signal, which is essential for MSP to use MAPK in an antiapoptotic pathway. This may involve translocation of MSP-activated MAPK from the cytoplasm into the nucleus, which occurs only in adherent cells. Our results suggest that there is cross talk between cell matrix adhesion and growth factors in the regulation of cell survival via the MAPK pathway. Growth factors induce MAPK activation, and adhesion mediates MAPK translocation from the cytoplasm into the nucleus.

Oncogenic mutants of RON and MET receptor tyrosine kinases cause activation of the beta-catenin pathway.

beta-Catenin is an oncogenic protein involved in regulation of cell-cell adhesion and gene expression. Accumulation of cellular beta-catenin occurs in many types of human cancers. Four mechanisms are known to cause increases in beta-catenin: mutations of beta-catenin, adenomatous polyposis coli, or axin genes and activation of Wnt signaling. We report a new cause of beta-catenin accumulation involving oncogenic mutants of RON and MET receptor tyrosine kinases (RTKs). Cells transfected with oncogenic RON or MET were characterized by beta-catenin tyrosine phosphorylation and accumulation; constitutive activation of a Tcf transcriptional factor; and increased levels of beta-catenin/Tcf target oncogene proteins c-myc and cyclin D1. Interference with the beta-catenin pathway reduced the transforming potential of mutated RON and MET. Activation of beta-catenin by oncogenic RON and MET constitutes a new pathway, which might lead to cell transformation by these and other mutant growth factor RTKs.

Hepatic catabolism of intravenously administered pro-macrophage-stimulating protein in mice.

We injected 125I-pro-macrophage-stimulating protein (pro-MSP) intravenously into normal mice to determine its clearance from the circulation and to test for conversion of pro-MSP to the biologically active heterodimer in the absence of inflammation or tissue injury. Pro-MSP was cleared from the circulation with a half-life of approximately 100 min. This rapid clearance was not peculiar to 125I-pro-MSP, since clearance rates of unlabeled pro-MSP and of 125I-bovine serum albumin were comparable. The liver was the major locus of radioactivity 10-20 min after the intravenous injection of 125I-pro-MSP. By 90 min, over 60% of total recovered radioactivity was in the small intestine. Reflecting gastrointestinal transit, counts decreased in the small intestine and appeared in the colon by 180 min. Essentially all counts in urine and feces obtained at later times were soluble in trichloracetic acid. These findings reflected rapid hepatic proteolysis of pro-MSP to fragments undetectable by antibody to pro-MSP; within 20 min after intravenous administration, immunoprecipitable counts were only 22% of the total liver extract radioactivity. Comparison of sodium dodecyl sulfate-polyacrylamide gel electrophoresis and radioautography data for immunoprecipitated plasma and liver extract revealed no evidence for hepatic conversion of pro-MSP to MSP. Thus, the hepatic catabolic pathway of pro-MSP is degradative and does not yield mature MSP. The results support our view that MSP is not released into the circulation but is generated at specific extravascular loci by pro-MSP convertases.

Analysis of human monocyte chemoattractant binding by flow cytometry.

The chemotactic peptide fMet-Leu-Phe-Lys is a potent chemoattractant for human blood monocytes. However, only one-third of the monocytes respond. To determine whether or not lack of response reflected absence of attractant receptors, we equilibrated peripheral blood mononuclear cells with fMet-Leu-Phe-Lys-FITC and analyzed binding by flow cytometry. The fluoresceinated peptide bound rapidly at 0 degree C, and the amount bound approached saturation with increasing concentration. The percentage of blood monocytes that bound the peptide was 60 +/- 8 (SEM for seven experiments). In contrast, only 36 +/- 3% (SEM for 16 experiments) of monocytes responded to the attractant by directed migration. It follows that, among the 64 nonmigrating monocytes per 100 total monocytes, approximately 40, or two-thirds of them, fail to bind attractant; the remaining one-third bind attractant but do not respond with directed movement.

Purification of human blood basophils by single step isopycnic banding on Percoll.

Human venous blood, anticoagulated with EDTA, was layered onto a discontinuous Percoll gradient, made from solutions of density 1.088, 1.079, and 1.070 gm/ml. After centrifugation at 700g for 15 min at 22 degrees C, the majority of the blood basophils was found in a narrow band at the density 1.070-1.079 interface (Percoll band 2). For 15 normal donors, mean total basophil number recovered from all locations in the gradient was 3.8 +/- 1.2 (SD) X 10(4) basophils per ml of blood applied. Thus, 95% of the values ranged from 1.5 to 6 X 10(4), which compares favorably with the reported range of 1 to 8 X 10(4) basophils per ml for normal subjects. In the basophil-rich Percoll band 2, 2.8 +/- 0.8 X 10(4) basophils were recovered per ml of blood applied. The mean percentage of basophils in Percoll band 2 was 19%, with a range of 5 to 53%. Monocytes and neutrophils were present in very small numbers; the majority of accompanying cells were small lymphocytes.

Antibodies to macrophage stimulating protein (MSP): specificity, epitope interactions, and immunoassay of MSP in human serum.

Macrophage stimulating protein (MSP) is a member of a family of proteins characterized by a triple disulfide loop structure (kringle). We developed antibodies to human MSP for detection in Western blots, quantification in biological fluids, and neutralization of activity. Immunogens included native MSP, reduced and alkylated alpha and beta chains, and peptides of MSP regions with minimal sequence similarity to other kringle proteins. We found three antibody categories based on interaction with the following types of epitope: primary sequence, discontinuous (dependent on disulfide bonds), and cryptic (not exposed in native MSP). None of the antibodies reacted with related kringle proteins. A specific sandwich ELISA was developed for measuring human MSP. The mean serum concentration was 4 nM. Serum MSP did not increase over a 24-h period in response to intravenous lipopolysaccharide, indicating that MSP is not an acute phase protein. These findings are consistent with the hypothesis that regulation of MSP activity is by conversion of pro-MSP to MSP rather than by rapid changes in rates of synthesis.

A high molecular weight chemoattractant generated from C5 by ultracentrifugation of mouse serum without activation of complement.

After ultracentrifugation of normal mouse serum, we found chemoattractant activity in the high molecular weight protein region at the bottom of the tube, which was comparable in amount and potency to the attractant in endotoxin-activated serum. This was not a pre-formed attractant, but was generated from serum reactants at least one of which was inactivated by heating at 56 degrees C. Analysis of sera from 10 different mouse strains for hemolytic C5 activity and for capacity to generate chemoattractant on ultracentrifugation showed that the 4 strains without C5 were the only strains that failed to generate the attractant. Thus, the attractant precursor is C5. Since the activity was generated in the presence of 0.01 M EDTA, classical or alternative complement activation was not required. The chemoattractant product had a mol. wt of approximately 170,000; it was therefore not free C5a. These results, and data recently published on digestion of purified human C5 by trypsin, suggest that limited proteolysis of C5 can produce a chemoattractant molecule without release of free C5a.

Three forms of monocyte-derived neutrophil chemotactic factor (MDNCF) distinguished by different lengths of the amino-terminal sequence.

Human monocyte-derived neutrophil chemotactic factor (MDNCF) was purified from culture supernatant of lipopolysaccharide-stimulated human peripheral blood mononuclear leukocytes on a column of Sepharose-bound murine monoclonal anti-MDNCF. About 65% of the culture fluid chemotactic activity was bound to the column. The unbound 35% probably represents chemotactic activity of other cytokines in the culture fluid. More than 85% of the bound activity was eluted by pH 2.5 glycine buffer. When this material was applied to an HPLC-CM column, gradient elution produced four well-separated A280 peaks, each of which had chemotactic activity. N-terminal amino acid analysis of the four peaks revealed three different sequences. One (MDNCF-c) was identical to the sequence that we reported previously. The other two (MDNCF-a and -b) had seven and five additional amino acids, respectively, at the N-terminus. MDNCF-a, -b and -c accounted for 8, 47 and 45% of the total MDNCF peptide. Alignment with the MDNCF cDNA sequence shows that MDNCF-a results from cleavage of a 20 residue signal peptide. MDNCF-c results from culture fluid proteolytic cleavage of the N-terminal sequences of MDNCF-a and -b at an R-S bond. The three peptides occurred in the four HPLC-CM peaks in different ratios. The bulk of any one peptide was distributed in two adjacent HPLC-CM peaks. This suggests that each peptide exists in a minimum of two states. In contrast to our previous multi-step purification, the immunoaffinity and HPLC-CM column sequence resulted in complete purification of MDNCF in two steps and led to identification of two additional MDNCF peptides, one of which has not heretofore been detected.

Proteolytic cleavage and activation of pro-macrophage-stimulating protein and upregulation of its receptor in tissue injury.

Macrophage stimulating protein (MSP) exists in blood as inactive pro-MSP. Cleavage yields active MSP, the ligand for a membrane receptor (RON) that is expressed on keratinocytes as well as macrophages. Because both cells have roles in tissue injury, we looked for active MSP and expressed RON in wounds. Concentration of pro-MSP + MSP in wound exudates was in the range for optimal activity. Western blot showed that MSP comprised about half the total, in contrast to less than 10% of the total in blood plasma. The presence of MSP was attributed to an exudate pro-MSP convertase that had an inhibitor profile consistent with a trypsin-like serine protease. Exudate evoked morphologic changes in macrophages in vitro like that of MSP. Removal of this activity by an anti-MSP column shows that exudate stimulation of macrophages is due to MSP. RON was infrequently detected in normal skin. RON protein was markedly upregulated in burn wound epidermis and accessory structures, in proliferating cells or differentiated cells, or both. RON was also detected on macrophages and capillaries. Tissue injury leads to cleavage of pro-MSP to MSP, which has potential to act on keratinocytes, macrophages, and capillaries, all components of the wound healing response.

Disposable microliter immunoabsorbent columns: construction and operation.

Disposable microliter immunoabsorbent columns were constructed from pasteur pipets. The bed support was a cube of gelatin surgical sponge, which was tamped into the pipet tip. Column dead space, represented by the compressed volume of the sponge, was 5 microliters. The columns were used with protein A-Sepharose; settled bed volumes were 50 microliters. It was possible to pour columns that functioned as immunoabsorbents with bed volumes as small as 10 microliters. There was no gravity flow through these columns. Flow was achieved by touching column tips to absorbent paper if liquid was to be discarded or to 50 microliters capillary tubes for fluid collection. A simple capillary collection tube assembly was designed for operation of a row of 10 columns at a time. In a test system of [3H]methotrexate and IgG anti-methotrexate, 90% of applied antigen was bound to antibody columns, whereas 90% was recovered in the eluates from control columns. The columns were used in the initial step of screening uncloned hybridoma culture fluids for anti-MSP.

A multiwell cell settling and adherence chamber for morphology and differential counting.

A simple multiwell chamber is described that can be used to prepare randomly distributed cells on a microscope slide, suitable for morphological identification and differential counting. To the eight wells of the chamber are added 50-microliter volumes of cell suspension at concentrations of 10(3)-10(6) cells/ml. As the cells settle, fluid is slowly wicked away by a damp filter paper sandwiched between the microscope slide and the acrylic top plate of the multiwell chamber. Within 20-40 minutes, the cell monolayers on the slide are completely dry. The combined settling and bulk fluid removal results in a distribution of adherent cells that are sufficiently spread to exhibit excellent morphology after staining. If the chamber is centrifuged for 30 seconds at 50x g immediately after addition of cells, recovery of cells in the monolayer is virtually 100%, and as few as 50 input cells per 50 microliters can be detected. Agreement between predicted and observed differential counts of cell mixtures indicates that cells in the monolayer were distributed randomly.

Dynamics of chemotactic peptide-induced superoxide generation by human monocytes.

Addition of the chemotactic peptide, f-Met-Leu-Phe, to human monocytes induced a burst of superoxide release, which ceased after approximately 3 min. Diminished responsiveness to f-Met-Leu-Phe, but not to phorbol myristate acetate (PMA), was induced by 1- to 3-h storage at 0 degrees C or by 2 min in 40 microM adenosine (ADO). Reversal of the ADO block was achieved by addition of adenosine deaminase (ADA) as little as 15 sec before the f-Met-Leu-Phe stimulus; ADA had no effect when added poststimulus. The ADO experiments suggest that there are a minimum of two sequentially produced intermediates in the f-Met-Leu-Phe stimulus-response pathway. The first intermediate persists for less than 30 sec. The second, formation of which is stimulated by the first, persists for the duration of the response and is the target of ADO inhibition. The ADO target is apparently not protein kinase-C, since the response of inhibited cells to PMA was unimpaired. The maximal inhibition by adenosine of f-Met-Leu-Phe-induced superoxide generation was approximately 50%. It is possible that f-Met-Leu-Phe stimulates two pathways of NADPH activation, only one of which is inhibited by adenosine.

Biological aspects of monocyte chemoattractant protein-1 (MCP-1).

In this communication, we have asked if MCP-1 is the mediator of cellular infiltration in DCH, outlining the criteria in Table 3. Preliminary data suggest that PHA-stimulated lymphocytes secrete MCP-1, and that MCP-1 can be produced in response to antigen stimulation. MCP-1 attracts monocytes and basophils, but not neutrophils. The question of a lymphocyte response to MCP-1 requires further study. We have emphasized that the discovery of leukocyte-specific NAP-1 and MCP-1 should now be followed by exploration of conditions in which one agonist is secreted without the other. This would be expected, for example, in DCH, which is characterized by mononuclear leukocyte infiltration without neutrophils.

Secretion of monocyte chemoattractant protein-1 (MCP-1) by human mononuclear phagocytes.

Concentrations of MCP-1 and NAP-1 in culture fluids of human leukocytes were measured by sandwich ELISA. PPD caused PBMC's from tuberculin-sensitive subjects to secrete MCP-1 and NAP-1. PPD did not stimulate secretion by cells from a tuberculin-negative subject. Since the amounts secreted were more than could be produced by the few PPD-sensitized lymphocytes in the culture, we postulate that other cells were stimulated to secrete these chemoattractants. This study evaluated secretory capacity of one of the cell types in the PBMC culture. Unstimulated monocytes did not secrete MCP-1 or NAP-1. In order of increasing effect, IL-2 + IFN gamma, IL-1 alpha, and LPS caused monocyte secretion of MCP-1. The rank order for NAP-1 secretion was the same. TNF alpha did not cause secretion of MCP-1, but caused about the same amount of NAP-1 secretion as IL-2 + IFN gamma. Composition of the culture medium was especially critical for LPS-induced secretion of MCP-1, which was greatly enhanced by FCS and by Iscove's DMEM compared to RPMI 1640. IL-4 inhibited LPS-induced secretion of both MCP-1 and NAP-1. Secretory patterns were also a function of mononuclear phagocyte phenotype. LPS-induced secretion of MCP-1 was much greater for monocytes cultured several days in CSF-1 than for freshly isolated monocytes. LPS stimulation of bronchoalveolar macrophages caused NAP-1 secretion, but no secretion of MCP-1 above a relatively low baseline level.

Separation of human basophils into two fractions with different density and histamine content.

We designed experiments in this study to test the hypothesis suggested by recent purification data that blood basophils comprise two populations of different density, which circulate in numbers characteristic for each human subject. Basophils were separated into two density bands by single step centrifugation on a discontinuous Percoll gradient. Band 1 cells were at the interface between plasma and Percoll of density 1.070 gm/ml. Band 2 cells were at the Percoll 1.070 to 1.080 interface. When the number of band 1 basophils was expressed as a percentage of the total in bands 1 and 2, this relative amount generally remained in a narrow range for blood obtained from the same donor on 3 successive days but differed markedly in different individuals. In a series of leukapheresis experiments, we demonstrated that the percentage of band 1 basophils in postleukapheresis venous blood was strikingly similar to the preleukapheresis value. If basophils that repopulated the leukapheresis-depleted circulation came from the bone marrow, we can conclude that blood levels of basophils in bands 1 and 2 are under physiologic control and that the two types of basophils are released in amounts characteristic for each human subject. Additional evidence for two distinct blood basophil populations was provided by histamine measurements. The histamine content per basophil was consistently higher in cells from band 1 than from band 2, the mean difference between pairs of values for 30 subjects being 0.3 +/- 0.04 pg or about 27% of the band 1 basophil histamine content of 1.1 pg.

Action and target cell specificity of human macrophage-stimulating protein (MSP).

Macrophage-stimulating protein (MSP) induces mouse resident peritoneal macrophages to become responsive to the chemoattractant C5a and to ingest C3bi-coated erythrocytes. We now show that MSP action is not limited to complement-induced responses, because it also induced responsiveness to the noncomplement chemoattractant casein. In addition to stimulating responsiveness to attractants, MSP functioned alone as a chemoattractant for resident peritoneal macrophages, with an optimal concentration of approximately 0.2 nM. A critical difference between MSP and C5a is that resident macrophages did not migrate to C5a without an additional stimulus such as MSP in the cell suspension, whereas macrophages suspended in medium alone migrated to MSP in the attractant well. Thus, in contrast to C5a, MSP seems capable of a dual role, both activator and attractant. MSP had no effect on responsiveness of mouse peritoneal exudate macrophages to C5a; nor could it attract exudate macrophages or human blood monocytes. Absorption studies showed that resident macrophages have a receptor for MSP, but exudate macrophages do not. In view of these findings, it seems that the biological role of MSP is not as a recruiter of blood monocytes to sites of inflammation, but as an activator of mature macrophages. The MSP-induced activated state for responsiveness to C5a or C3bi was transient, and decayed at a first order rate with a t 1/2 of approximately 1 h. This is a new example of the transience of activation induced in macrophages by proinflammatory stimuli.