Human neutrophil lactoferrin trans-activates the matrix metalloproteinase 1 gene through stress-activated MAPK signaling modules.

It has been proposed that human neutrophil lactoferrin (Lf) could be involved in gene expression as a DNA-binding protein after its translocation into the nucleus. However, the molecular basis of Lf action has not been defined, and Lf-regulated target genes have not been identified. We report here that overexpressed Lf functions as a specific trans-activator of matrix metalloproteinase 1 (MMP1) gene, and that induction of this AP-1-responsive gene is mediated via the stress-activated MAPK signaling modules. Transactivation of the MMP1 promoter by overexpressed Lf requires the presence of an AP-1 binding site. In gel shift experiments, Lf did not interact directly with AP-1-containing fragments of the MMP1 promoter. However, nuclear extracts from Lf-expressing cells contained increased levels of proteins that bound to AP-1 elements. This Lf-induced AP-1 DNA binding activity was reduced by a p38 MAPK inhibitor. Inhibitors of the MEK kinases had little effect on Lf-induced AP-1. However, expression of dominant-negative MKK4 or JNK1 inhibited Lf-induced gene expression. The JNK activity stimulated by Lf correlates with the enhanced AP-1 binding ability. These findings demonstrate that the Lf-induced activation of AP-1 is mediated via JNK and p38 MAPK pathways.

lating levels of Lf increase significantly. On the other hand, Lf has been reported to act as an inhibitor of cytokine response in vitro at a very low concentration by suppressing the release of cytokines such as interleukins and tumor necrosis factor (3). Lf also may act as a potent activator of natural killer cells and a direct inhibitor of tumor cell growth (4). Therefore, the balanced synthesis and exocrine secretion of the Lf in the cell suggest its precise roles in various biological systems.
A previous report revealed that Lf enters the cell from the serum and is transported into the nucleus where it binds to DNA (5). Recently random DNA sequences capable of binding to Lf have been demonstrated, which may imply an Lf-mediated regulation of gene expression (6). However, little information on naturally occurring target genes that are directly responsive to Lf is available. Therefore, we have investigated one of the roles of Lf as a multifunctional regulatory protein. To do so, the exogenous human neutrophil Lf gene was overexpressed in various mammalian cells. There is considerable interest in the role of Lf because it acts as a trans-activator for the expression of a subset of responsive genes, particularly AP-1-inducible genes such as fibroblast collagenase-1, matrix metalloproteinase 1 (MMP1), and monocyte chemoattractant protein 1 (MCP-1). The MMP1 exerts important effects on physiological and pathological conditions of cells including tumor metastasis (7), although its exact mechanism remains unclear.
In this study, we have focused on the elucidation of the molecular mechanism by which the overexpressed Lf mediates the stimulation of MMP1 production. Nested deletion and point mutations have allowed us to narrow down the Lf-responsive promoter region of the MMP1 gene to a specific element. By performing EMSA, we have attempted to see whether Lf could interact directly or indirectly with the promoter of the target gene. To test potential roles for specific signaling pathways in mediating Lf-induced gene expression, we have used several specific kinase inhibitors and dominant-negative mutant forms of kinase cDNAs. We demonstrate here that the activation of stress-activated MAPK signaling pathways by Lf plays a crucial role in the up-regulation of the MMP1 gene.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection Assay-Balb/c-3T3, HeLa, Jurkat, and COS-1 cells were purchased from American Type Culture Collection (Manassas, VA). The cells were grown and transfected using DEAEdextran (8), an equal amount (5 g) of the human neutrophil Lf expression vector pLf, and CAT construct. Cotransfection with the pRSV-␤galactosidase plasmid (0.5 g per transfection) and cytofluorometric analysis of ␤-galactosidase activity were performed to normalize for transfection efficiency. CAT activities were quantitated with a liquid scintillation counter. CAT assay data are averages of at least three separate experiments. Long term Lf-expressing cell lines, B-hLf2 and H-hLf6, were cloned by stably cotransfecting Balb/c and HeLa cells with either pLf or pCI-neo, respectively, by the calcium phosphate method (9) and followed by selecting desired clones with medium containing geneticin (G418). Both B-hLf2 and H-hLf6 cells were maintained in the complete medium supplemented with 0.5 mg/ml G418. Geneticin, TPA, poly(dI-dC), acetyl-CoA, SB203580, PD98059, and U0126 were purchased from Sigma.
Preparation of Nuclear Extracts and EMSAs-Nuclear extracts of Balb/c-pCI-neo and B-hLf2 cells were prepared by the method of Andrews and Faller (15). EMSAs were performed as described previously (12). Oligonucleotides with an AP-1 consensus binding site (5Ј-CGCT-TGATGAGTCAGCCGGAA-3Ј) were end-labeled with [␥-32 P]ATP using T4 polynucleotide kinase (Promega) and reannealed by heating and cooling them down to the room temperature. A DNA fragment of the MMP1 gene promoter cut with PvuII/BamHI was used as either a probe or an unlabeled competitor. A DNA oligonucleotide specially designed for Lf binding (5Ј-CGCTTGAGGCACTTGCGCCGGAA-3Ј) contained the artificial Lf binding sequence (6) and was prepared as described above. Iron-saturated Lf protein was purchased from Sigma.
Zymography Assay-After washing cells with cold phosphate-buffered saline, cells were harvested and solubilized in a SDS sample buffer. An 8% polyacrylamide gel containing 0.1% type A gelatin (Sigma) was used for gelatin zymograms. After running the gel at 4°C, the gels were washed twice in 100 ml of 2.5% Triton X-100 for 1 h and incubated overnight at 37°C in substrate buffer (100 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 10 mM CaCl 2 ). Gels were visualized by staining with Coomassie Blue.
Protein Kinase Activity Assays-JNK activity in cell lysates was determined by immunocomplex protein kinase assays using the substrate glutathione S-transferase-c-Jun (16). Immunoblot analysis was performed according to the instructions of the manufacturer (New Eng-land Biolabs). Cellular lysates of the BALB/c-pCI-neo and B-hLf2 cells were prepared, and each 100 g of cellular protein was run on a 10% SDS-polyacrylamide gel, electrotransferred to an Immobilon-P membrane (Millipore), and probed with the anti-phospho-c-Jun (Ser-73) antibody (New England Biolabs). The membrane was then incubated with horseradish peroxidase-conjugated secondary antibody and checked with a chemiluminescence kit (ECL, Amersham Pharmacia Biotech). The blots were exposed to Kodak x-ray film for 3-5 min. Anti-JNK1 antibody was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA).

RESULTS
To identify Lf-responsive promoters, a set of putative gene promoters linked to the reporter gene CAT was cotransfected with an Lf expression vector (pLf). Transcriptional activities of the MMP1 and MCP-1 gene promoters (Fig. 1A, compare pCol-lCAT-517 and MCP-1-CAT, respectively) were stimulated at least 30 -40-fold, whereas the interleukin-2 gene promoter was not affected by the Lf expression in Balb/c fibroblast cells. To examine the influences of Lf on the expression of endogenous cellular genes, Balb/c and HeLa cells were stably transfected with pLf. As expected, after neomycin selection, the resulting stable transfectants (designated B-hLf2 and H-hLf6) were found to express Lf protein (data not shown). Expression of MMP1 mRNA was elevated 15-fold in H-hLf6 as compared with the expression of the control gene glyceraldehyde-3-phosphate dehydrogenase as shown by RT-PCRs, which generated a 438base pair product with a primer pair for the MMP1 coding region (Fig. 1B). In COS-1 cells transiently transfected with pLf, the level of the MMP1 enzyme activity was also increased as measured by the zymography (Fig. 1C). These results suggest that the Lf acts as a transcriptional activator of the MMP1 gene in intact cells and functions in trans.
The promoter of the TPA-inducible MMP1 gene was further analyzed to identify a cis-acting element for the Lf. A putative responsive region on the MMP1 gene promoter was identified by deletion analysis. Nested deletion of the MMP1 gene promoter linked to the reporter gene CAT up to the TPA-responsive element (TRE) (Ϫ73 to Ϫ67, TGAGTCA) preserved a moderately high inducibility of the chimeric gene by Lf (Fig. 2A). The pCollCAT-73 construct spanning the Ϫ73/ϩ63 region, which contains an active TRE, showed the weakest Lf-respon- trans-Activation of Matrix Metalloproteinase 1 Gene by Lf sive activity as compared with other constructs with longer promoters. However, deletion of the first three nucleotides in the TRE (Ϫ70/ϩ63) resulted in the reduction by 80% in Lfresponsive trans-activation (Fig. 2B). Removal of the entire TRE immediately upstream of the TATA box (Ϫ60/ϩ63) almost abolished all trans-activation by Lf. These observations indicate that the TRE or TRE-like sequence is required for the trans-activation of the MMP1 gene promoter by Lf.
We were interested in elucidating the mechanism of MMP1 trans-activation by Lf. Other workers have shown that Lf can bind to DNA. Therefore, we hypothesized that Lf-induced trans-activation of MMP1 resulted from direct interaction of Lf with TRE or TRE-like sequences. To address this hypothesis, EMSAs were performed using radiolabeled TRE probes in the presence of highly purified iron-saturated Lf proteins (Fig. 3A). However, neither the TRE consensus oligonucleotides nor the DNA fragment containing the Ϫ73/ϩ63 region of the MMP1 gene promoter bound to Lf under the condition in which the artificial DNA oligonucleotides specially designed for Lf binding were efficiently bound (6). The lack of binding to the TRE consensus sequences suggested that Lf-mediated trans-activation of the MMP1 promoter did not involve direct binding of Lf to the TRE. However, it was observed that AP-1 binding activity was greatly enhanced in B-hLf2 cells expressing Lf (Fig.  3B). Lf-induced AP-1 DNA binding activity almost disappeared with the addition of a 50-fold molar excess of either the unlabeled TRE consensus oligonucleotides or the fragment of minimal MMP1 gene promoter. This experiment implies that upregulation of MMP1 by Lf may be mediated through an alteration of either the activity or the level of AP-1 instead of a direct binding of Lf to DNA and that Lf may influence the AP-1 activation pathway.
Many transcription factors including c-Jun, c-Fos, and ATF2 are effectors of three distinct groups of MAPKs: extracellular signal-regulated kinase, p38 MAPK, and JNK. To determine whether Lf-induced AP-1 activation possibly resulted from activation of any one of the MAPK pathways, the effect of specific kinase inhibitors (SB203580, PD98059, and U0126) was tested. As measured by EMSAs using nuclear extracts from B-hLf2 cells treated with SB203580, which is a specific inhibitor of p38␣ (SAPK2a) and its isoform p38␤ (SAPK2b) but which has no inhibitory effect on SAPK3 and SAPK4 (17), AP-1 DNA binding activity was reduced by about 50% as compared with control extracts (from cells that received only Me 2 SO) (Fig. 4A). This result suggests that Lf-induced AP-1 activation is, at least in large part, p38 MAPK-mediated. In contrast, both PD98059 (18) and U0126 (19), which specifically inhibit phosphorylation of MEK1 and MEK2, had little effect on the AP-1 activity. MKK4 is an essential component of the JNK signal transduction pathway and also activates p38 MAPK (20). Therefore, it was anticipated that interfering with the MKK4 activity by overexpression of the dominant-negative MKK4 (SEK1) might cause the inhibition of downstream pathways implicated in both JNK and p38 MAPK signaling pathways. Fig. 4B demonstrates that the transient cotransfection of Balb/c cells with pLf and dominant-negative SEK1 selectively blocked AP-1 binding, indicating the involvement of stress-activated MAPK signaling modules in Lf-induced AP-1 activation. Furthermore, the expression of dominant-negative JNK1 inhibited the ability of Lf to trans-activate the MMP1 (Fig. 4C). To confirm that the activated AP-1 observed in cells expressing Lf resulted from the activated JNK, the phosphorylation levels of c-Jun in lysates were determined. In B-hLf2 cells, JNK strongly phosphorylated Ser-73 of the c-Jun NH 2 -terminal domain (Fig. 4D,  upper panel). Moreover, Lf expression in B-hLf2 cells caused a marked increase in JNK1 protein kinase activity (Fig. 4D,  lower panel).

DISCUSSION
Human Lf has been implicated in various biological processes including the regulation of cell growth and inflammation as well as the primary defense against bacterial infection (21). Lf has been considered to exert a control function in the balance of cellular components as a nuclear factor (6,22). It has been demonstrated that sequence-specific binding of Lf to DNA occurs under stringent conditions. Here we demonstrate that overexpressed Lf is capable of regulating the expression of the MCP-1 and MMP1 genes, which are implicated in inflammation and metastasis (23,24). Moreover, we found that Lf specifically trans-activates MMP1, an AP-1-inducible gene, by altering the DNA binding properties of AP-1 in both fibroblasts and T lymphocytes. The biological significance of the up-regulation of the MMP1 gene by Lf is unclear. Circulating levels of Lf increase significantly during the inflammatory process (2). Additionally activated leukocytes and many tumor cells secrete large amounts of the metalloproteinases (25). Therefore, it is likely that the activation of MCP-1 and MMP1 genes in response to Lf may be involved in physiological and pathological processes such as inflammation and tumorigenesis (3,26).
In this study, we also found that Lf required a TRE site for the trans-activation of the MMP1 gene and that AP-1 binding activity was substantially enhanced in cells expressing Lf. AP-1 activity in cells is known to be regulated by upstream signal transduction pathways such as the MAPK cascades (20). Therefore, it seemed likely that Lf-mediated AP-1 activation might be mediated via known signaling networks. Our studies with specific kinase inhibitor drugs and dominant-negative components of the MKK4 pathway indicated that both JNK and p38 MAPK might contribute to Lf-induced AP-1 activation. Consistent with a role for MKK4 in Lf-responsive AP-1 activation, the MKK4 effector kinase JNK was activated by Lf. Together these data strongly suggest that SAPK and c-Jun may mainly mediate the Lf-induced signaling pathway. We do not rule out the possibility that the SAPK can also activate other transcription factors such as Elk1, MEF2c, and ATF2 so that their subsequent transcriptional and/or posttranslational targets in turn may contribute indirectly to the activation of AP-1 in cells expressing Lf (27).
The transcriptional targets of the Lf-mediated SAPK signaling pathway have not been fully established. However, we have observed that human neutrophil Lf greatly up-regulated AP-1inducible inflammatory mediator cytokines and some particular genes in the immunoglobulin gene superfamily. 2 Although the biological functions of Lf have been controversial (3,22,28), our new results suggest an important role for Lf in gene regulation and eventually cellular activity. We demonstrated that the constitutively overexpressed Lf induced the MKK4-SAPK signaling pathway leading to AP-1 activation and MCP-1 expression. The increased expression of chemoattractants such as MCP-1 by the AP-1 causes leukocyte recruitment, resulting in the enhancement of the inflammatory response by releasing the protease MMP1 (25). This AP-1-induced expression of MMP1 is likely to have a number of consequences, including remodeling of matrix, oxidative burst, and cell injury (29). It is likely that the circulating level of Lf is an important factor for the regulation of Lf-induced genes. Indeed, the effect of Lf on natural killer cell cytotoxicity depends on the level of Lf (30). Furthermore, multihormone signaling pathways may be involved in modulating Lf gene activity and its circulating level, and the activated human blood lymphocytes may express Lf receptors (31). The autocrine property of the Lf molecule and its role in the specific gene activation imply the involvement of Lf receptors in further upstream signaling immediately following binding of Lf to its receptor. Therefore, studies are under way to investigate whether the expression level of Lf and its translocation through the receptor will be crucially related with the cellular activity in the response to a variety of stresses. Nuclear extracts for EMSA were prepared from Balb/c-pCI-neo (Ϫ) or B-hLf2 cells (ϩ) treated with each inhibitor or only its vehicle Me 2 SO (DMSO). Before the treatment, cells were starved for 24 h in Dulbecco's modified Eagle's medium containing 0.5% fetal bovine serum. B, AP-1 binding activity and effects of dominantnegative SEK1. An EMSA was performed as described above. Either dominant-negative SEK1 (SEK1-DN) or pCDNA3 backbone plasmid was used for the transient co-transfection of Balb/c cells with an equal amount of pLf (ϩ) or pUC19 (Ϫ). C, effects of dominant-negative JNK1 (JNK1-DN) on trans-activation. Relative CAT activity was determined after cotransfection of HeLa cells with the indicated plasmids. Shown is the mean Ϯ S.E. for three independent experiments in duplicates. D, JNK activity in cells expressing Lf. Immunoblot analysis (upper panel) using phospho-c-Jun (p-c-Jun) (Ser-73) antibody and in vitro kinase assay (lower panel) for JNK1 were performed as described under "Experimental Procedures." Total cell lysates obtained from Balb/c-pCI-neo (Ϫ) or B-hLf2 (ϩ) cells were used for each assay.