A Chemotactic Peptide from Laminin 5 Functions as a Regulator of Inflammatory Immune Responses via TNF-mediated Signaling
http://www.100md.com
免疫学杂志 2005年第3期
Abstract
Tissue injury triggers inflammatory responses that may result in release of degradation products or exposure of cryptic domains of extracellular matrix components. Previously, we have shown that a cryptic peptide (AQARSAASKVKVSMKF) in the -chain of laminin-10 (511), a prominent basement membrane component, is chemotactic for both neutrophils (PMNs) and macrophages (Ms) and induces matrix metalloproteinase-9 (MMP-9) production. To determine whether AQARSAASKVKVSMKF has additional effects on inflammatory cells, we performed microarray analysis of RNA from RAW264.7 Ms stimulated with AQARSAASKVKVSMKF. Several cytokines and cytokine receptors were increased >3-fold in response to the laminin 5 peptide. Among these were TNF- and one of its receptors, the p75 TNFR (TNFR-II), increasing 3.5- and 5.7-fold, respectively. However, the peptide had no effect on p55 TNFR (TNFR-I) expression. Corroborating the microarray data, the protein levels of TNF- and TNFR-II were increased following stimulation of RAW264.7 cells with AQARSAASKVKVSMKF. In addition, we determined that the production of TNF- and TNFR-II in response to AQARSAASKVKVSMKF preceded the production of MMP-9. Furthermore, using primary Ms from mice deficient in TNFR-I, TNFR-II, or both TNF- receptors (TNFRs), we determined that AQARSAASKVKVSMKF induces MMP-9 expression by Ms through a pathway triggered by TNFR-II. However, TNF- signaling is not required for AQARSAASKVKVSMKF-induced PMN release of MMP-9 or PMN emigration. These data suggest that interactions of inflammatory cells with basement membrane components may orchestrate immune responses by inducing expression of cytokines, recruitment of inflammatory cells, and release of proteinases.
Introduction
Basement membranes (BMs)3 are thin sheets of specialized extracellular matrix (ECM) which lay beneath epithelial and endothelial cells and surround other cell types, including adipocytes, muscle, and nerve cells. Besides providing tissue boundaries and structural support, BMs influence cell proliferation, differentiation, and migration (1, 2, 3, 4). BMs are predominantly composed of laminins, type IV collagen, heparan-sulfate proteoglycans, and entactin/nidogen. In tissue remodeling during inflammation, increased ECM turnover can result in release of degradation products or exposure of cryptic domains of ECM components that can elicit biological responses distinct from the parent molecules (3, 4).
Although the role of ECM in affecting inflammatory cell functions has been examined, the predominant focus has been the effect of ECM components on adhesion, migration (5, 6, 7, 8, 9), and proteinase production (5, 6, 7, 8, 9, 10). Little is known about the effects of ECM components on cytokine production. However, fragments of fibronectin (11) and hyaluronan stimulate macrophage (M) cytokine production (12, 13, 14), suggesting that ECM components may have regulatory roles in inflammatory responses. Although fragments of BM components, such as laminins, type IV collagen, and entactin, have been detected in biological fluids in association with various inflammatory diseases (15, 16, 17, 18, 19, 20, 21), it is unknown whether BM components are active in the regulation of inflammation.
The inflammatory response consists of the release of proinflammatory cytokines, followed by the recruitment of circulating leukocytes, which become activated at the inflammatory site and release further mediators. TNF-, a powerful mediator of inflammation, can be produced by a variety of cell types, including activated monocytes/Ms (22, 23). Synthesized as a membrane-anchored precursor, active TNF- can be proteolytically released by the TNF- converting enzyme (TACE)/a disintegrin and metalloproteinase 17 (24, 25). The biological activities of TNF- are mediated by two structurally related but functionally distinct receptors, termed TNFR-I (TNFR p55/p60) and TNFR-II (TNFR p75/p80). Although binding of activated TNF- to its receptors can result in activation of proapoptotic signaling events (26, 27, 28), signaling triggered by TNFR-I and TNFR-II usually causes activation of the AP-1 and NF-B transcription factors, which then induce numerous genes involved in inflammatory responses (22, 23).
Laminins are a family of integral BM glycoproteins, each containing an -, -, and -chain (29). To date, five , four , and three laminin chains have been identified, which assemble into at least 15 laminin isoforms (30, 31, 32, 33, 34). Immunohistochemical data indicate that laminin-10 (511) is one of the most abundant laminin heterotrimers in normal adult tissues (35, 36, 37). Because laminin 5 is widely expressed in fetal and adult tissues, we are characterizing the function of specific domains of the laminin 5 chain. We previously reported that the laminin 5 peptide, AQARSAASKVKVSMKF, induces matrix metalloproteinase-9 (MMP-9) production and is chemotactic for leukocytes in vitro and in vivo (8). Furthermore, our data indicated that AQARSAASKVKVSMKF is cryptic in the intact laminin-10 molecule. In the present study we investigated whether AQARSAASKVKVSMKF has additional effects on inflammatory cells. These studies indicate that AQARSAASKVKVSMKF induces M expression of several cytokines, including TNF-. In addition, AQARSAASKVKVSMKF increases expression levels of TNFR-II, but not TNFR-I. Using primary Ms isolated from mice deficient in TNFR-I, TNFR-II, or both TNFRs, we determined that AQARSAASKVKVSMKF induces MMP-9 through a TNFR-II-dependent pathway. Taken together, our results indicate a novel regulatory role for a cryptic peptide of laminin 5, AQARSAASKVKVSMKF, in inflammatory responses and suggest that BM components play a role in orchestrating immune responses triggered by tissue damage.
Materials and Methods
Peptides
The YIGSR and VGVAPG peptides, which are derived from the laminin 1 chain and elastin, respectively, were obtained from Sigma-Aldrich. Other laminin-derived peptides were synthesized and purified as previously described (8). The sequences of these peptides are as follows: 1, SRARKQAASIKVAVSADR; 3, QQARDAANKVAIPMRF; 5, AQARSAASKVKVSMKF; 5f, AQARSAA; 5, ASKVKVSMKF; and 5b-sc, KAKSFVMVSK.
Mouse strains
Six- to 8-wk-old mice deficient in TNFR-I, TNFR-II, or both TNFRs were obtained from The Jackson Laboratory. Wild-type 129 SvEv and C57BL/6 mice (Taconic Farms) were housed in a barrier animal facility in the veterinary care of the Department of Comparative Medicine at Washington University School of Medicine (St. Louis, MO). All procedures using mice were approved by the Washington University School of Medicine Animal Studies Committee and were performed in accordance with the Animal Welfare Act and the Guide for the Care and Use of Laboratory Animals.
Cell culture
Murine RAW264.7 Ms, obtained from American Type Culture Collection, were maintained in low bicarbonate (1.5 g/L) DMEM, supplemented with 10% FBS, penicillin (100 U/ml), streptomycin (100 μg/ml), and 4 mM L-glutamine. Primary mouse Ms were harvested by peritoneal lavage as previously described (8). To isolate human monocytes, heparinized, human venous blood from normal volunteers was diluted 1/2 in sterile saline, layered on one-half volume of Histopaque-1077 (Sigma-Aldrich), and centrifuged at 750 x g for 30 min at 4°C. The mononuclear layer was collected and applied to a density gradient. Supernatant was removed and cells were suspended in media and counted. Procedures using human subjects were approved by the Washington University School of Medicine Human Studies Committee.
RNA isolation, biotinylated target synthesis, and microarray hybridization
RAW264.7 Ms were incubated in M serum-free medium (MSFM; Invitrogen Life Technologies) alone or containing 100 μg/ml AQARSAASKVKVSMKF for 24 h. Total RNA was extracted by the ToTally RNA isolation kit (Ambion), and equal amounts of total RNA from three separate experiments were pooled to minimize variability. Biotinylated cRNA targets were generated from 20 μg of total RNA using the Bioarray High Yield RNA Transcript Labeling kit (Enzo Life Sciences) and purified using RNeasy spin columns (Qiagen). Twenty micrograms of each biotinylated cRNA target was fragmented and then hybridized to identical Affymetrix Murine Genome Mu74Av2 GeneChips in the Siteman Cancer Center GeneChip Facility (Washington University School of Medicine).
Analysis of microarray data
The images from the scanned microarrays were processed using Affymetrix Microarray Analysis Suite 4.0 and analyzed using GeneSpring 6.1 software (Silicon Genetics). Normalization was performed using the parameters recommended by the software as follows: 1) per chip, by dividing the raw data by the 50th percentile of all measurements, and 2) per gene to a specific sample, by dividing the normalized data of all samples to the untreated control sample. From a total of the 12,000 probe sets on the array, those representing control sequences and genes that were scored as "absent" by the Affymetrix Detection criterion in both samples were excluded from analysis. Genes that increased or decreased expression at least 3-fold relative to the untreated control sample were organized into functional categories using the Simplified Gene Ontology, which is based on the Affymetrix annotated database for the Mu74Av2 chip. The Signaling Ligand and Receptor genes reported here were identified from the Simplified Gene Ontology lists as Signal Transducers in the Molecular Functions category, which contains chemokines and cytokines. The complete set of data and sample metadata may be downloaded at http://bioinformatics.wustl.edu.
RT-PCR analysis of TNF- and TNFRs
RAW264.7 Ms were incubated 24 h in MSFM alone or containing 100 μg/ml AQARSAASKVKVSMKF. One microgram of total RNA extracted from these cells was reverse transcribed using random hexamers and the GeneAmp RNA PCR Core kit (Roche). After reverse transcription, PCR were performed using primers specific for TNF-, TNFR-I, and TNFR-II. The primers for mouse TNF- were 5'-TTCTGTCTACTGAACTTCGGGGTGATCGGTCC and 5'-GTATGAGATAGCAAATCGGCTGACGGTGTGGG (38). The primers for mouse TNFR-I were 5'-CCATCTTCGGTCCTAGTAACTG and 5'-CAGGTTCATCTTGGAAAGCAC (39). The primers for mouse TNFR-II were 5'-GATGCCAAGGTGCCTCATG and 5'-GAGCTGCTACAGACGTTCACG (39). The primers for mouse GAPDH were 5'-CCCCTTCATTGACCTCAACTACATGG and 5'-GACATCAAGAAGGTGGTGAAGCAGGC (40). Products of 354, 391, 290, and 689 bp were predicted for TNF-, TNFR-I, TNFR-II, and GAPDH, respectively. The resulting amplification products were electrophoresed on a 1.5% agarose gel, stained with ethidium bromide, and visualized under UV light. Following scanning of the gels, quantification for each reaction product was performed using the public domain NIH Image program (National Institutes of Health, http://rsb.info.nih.gov/nih-image/). The mRNA levels of TNF-, TNFR-I, and TNFR-II were normalized to GAPDH expression. Data represent at least three independent experiments.
Detection of soluble TNF- and TNFRs by ELISA
RAW264.7 or primary peritoneal Ms were incubated for 24 h in MSFM alone or containing 100 μg/ml various peptides. At designated times, conditioned medium were harvested and assayed for soluble TNF-, TNFR-I, or TNFR-II in triplicate wells by ELISA according to the manufacturers’ protocols (OptEIA Mouse TNF- set, BD Pharmingen; DuoSets for Mouse sTNFR-I and sTNFR-II, R&D Systems). Data represent at least three independent experiments.
Gelatin zymography
Gelatinolytic activity was determined by gelatin zymography as described previously (8). Briefly, conditioned media were electrophoresed in a 7.5% acrylamide gel containing 1 mg/ml gelatin, and gels were washed in 2.5% Triton X-100 and incubated for 12 h in buffer containing 50 mM Tris (pH 8.0), 10 mM CaCl2, and 1 mM ZnCl2. Following incubation, gels were stained with Coomassie blue, destained, and gelatinolytic activity was detected as clear bands on a blue background. Mouse neutrophil (PMN) lysate was run as positive control for MMP-9.
Collection of bronchoalveolar lavage (BAL) fluid
Intrapulmonary instillation of the laminin 5 peptide and collection of BAL fluid were performed as previously described (8). Briefly, mice were anesthetized, placed supine, and dropwise inoculated intranasally with 50 μl of PBS, AQARSAASKVKVSMKF (4 mg/ml), or LPS from Escherichia coli (4 mg/ml) (Sigma-Aldrich). At various times, the mice were asphyxiated with carbon dioxide, and BAL fluid was retrieved by injecting PBS through the trachea. Total and leukocyte differential cell counts were performed. The BAL fluid was assayed for MMP-9 by gelatin zymography and soluble TNF- by ELISA.
Collection of peritoneal lavage fluid
Wild-type mice or mice deficient in TNFR-II or both TNFRs were inoculated with 100 μl of PBS or AQARSAASKVKVSMKF (2 mg/ml) into the peritoneal cavity (i.p.). At various times, the mice were asphyxiated with carbon dioxide, and the peritoneal cavity was lavaged once with 10 ml of PBS. Total and leukocyte differential cell counts were performed as previously described (8). Total RNA was isolated from these cells and the expression of MMP-9 was assayed by RT-PCR as previously described (40). The peritoneal fluid was assayed for MMP-9 by gelatin zymography and soluble TNF by ELISA.
Results
AQARSAASKVKVSMKF affects M gene expression
Previously, we reported that a laminin 5-derived peptide, AQARSAASKVKVSMKF, induces MMP-9 expression by Ms and is chemotactic for both PMNs and Ms in vivo (8). To examine whether AQARSAASKVKVSMKF regulates inflammation by inducing genes in immune cells, RNA isolated from RAW264.7 Ms that had been stimulated with the laminin 5 peptide for 24 h was analyzed by microarray. Genes whose expression levels increased or decreased at least 3-fold in response to AQARSAASKVKVSMKF were sorted into functional categories using the Affymetrix annotated database for the Mu74Av2 chip. Of the 70 known genes with increased expression of at least 3-fold in response to M stimulation with the laminin 5 peptide (data not shown), 13 were ligands or receptors involved in signal transduction pathways (Table I). These included several members of the CC or CXC chemokine families, as well as the proinflammatory cytokine, TNF-, and one of its receptors, TNFR-II. The expression levels of TNF- and TNFR-II were increased 3.5- and 5.7-fold, respectively, in response to AQARSAASKVKVSMKF. However, the laminin 5 peptide had no effect on the expression of TNFR-I or TACE (data not shown). Consistent with our previous findings (8), MMP-9 expression increased 2.9-fold in response to stimulation of Ms with the laminin 5 peptide (data not shown). These results indicate that AQARSAASKVKVSMKF induces the expression of distinct proinflammatory genes in immune cells and suggest that it can regulate inflammatory responses.
Table I. Microarray detection of signaling ligand and receptor mRNAs increased at least 3-fold by RAW264.7 Ms in response to AQARSAASKVKVSMKF
Confirmation of microarray data
To confirm the microarray data, RT-PCR analyses were performed using RNA isolated from RAW264.7 cells stimulated with AQARSAASKVKVSMKF. Because TNF- is a key modulator of inflammation (22, 23), we focused on the AQARSAASKVKVSMKF-induced expression of TNF- and its receptors. Following a 24-h stimulation with the laminin 5 peptide, an increase in the mRNA levels of TNF- and TNFR-II above the levels in control cells was detected (Fig. 1). In contrast, TNFR-I expression did not increase in response to AQARSAASKVKVSMKF. These findings confirm the microarray data, indicating that AQARSAASKVKVSMKF induces expression of TNF- and TNFR-II, but not TNFR-I, by Ms.
FIGURE 1. AQARSAASKVKVSMKF triggers expression of TNF- and TNFR-II, but not TNFR-I, by RAW264.7 Ms. RNAs isolated from RAW264.7 M that were untreated or stimulated with 100 μg/ml laminin 5 peptide for 24 h were subjected to RT-PCR using primers specific for murine TNF-, TNFR-I, TNFR-II, or GAPDH. The resulting agarose gels were scanned, digitized, and quantified using the NIH Image program. A, The mRNA levels of TNF-, TNFR-I, and TNFR-II were normalized to GAPDH and the mean relative expression ± SEM from at least three independent experiments is shown. B, A representative gel showing increased expression of TNF- and TNFR-II mRNAs in response to AQARSAASKVKVSMKF. Arrow, GAPDH; c, no template control.
AQARSAASKVKVSMKF induces TNF- protein production by mouse and human monocytes/Ms
Because AQARSAASKVKVSMKF increased the steady-state level of TNF- mRNA, we examined whether the laminin 5 peptide increased the production of TNF- protein by Ms. Because monocytes/Ms exhibit constitutive production of active TACE that can release TNF- (41, 42), the level of soluble TNF- in the conditioned media of RAW264.7 Ms was determined as an indicator of the quantity of TNF- protein produced by the cells. An increase in soluble TNF- was detectable 3 h after stimulation with AQARSAASKVKVSMKF and continued to increase over the 24-h period (Fig. 2A). To determine whether the AQARSAASKVKVSMKF-stimulated TNF- production by RAW264.7 Ms extends to primary mouse Ms, peritoneal Ms were stimulated with the laminin 5 peptide. An increase in soluble TNF- was detected in the conditioned media as early as 1 h and continued to increase over the 24-h period (Fig. 2B). These data indicate that AQARSAASKVKVSMKF induces the production of TNF- by primary mouse Ms.
FIGURE 2. AQARSAASKVKVSMKF induces TNF- protein production by mouse and human monocytes/Ms. RAW264.7 cells (A), mouse peritoneal Ms (B), and human blood monocytes (C) were incubated in MSFM alone () or containing 100 μg/ml laminin 5 peptide (?) for various times up to 24 h. Following incubation, the conditioned media were assayed for the presence of TNF- by ELISA. The concentration of TNF- per milliliter of conditioned media ± SEM is shown. Data are representative of at least three separate experiments.
Because this region of mouse laminin 5 has a high degree of sequence identity with the corresponding region of human laminin 5 (35), we examined whether AQARSAASKVKVSMKF could induce TNF- by human monocytes. Within 3 h of stimulation with the laminin 5 peptide, an increase in the production of TNF- by primary human blood monocytes was detected, which continued to rise over the 24-h period (Fig. 2C). These data indicate that the response to the laminin 5 peptide is not limited to mouse and suggests a regulatory role of laminin 5 in inflammatory responses in humans as well.
AQARSAASKVKVSMKF induces TNF- production in vivo
To test whether cryptic peptides of laminin 5 play a role in inflammatory responses in vivo, AQARSAASKVKVSMKF was instilled into the lungs of mice and the lungs were lavaged at various times. At 4 h postinstillation of the laminin 5 peptide, a marked increase in TNF- in BAL fluid was observed as compared with the BAL fluid from control mice treated with PBS (Fig. 3A). The potency of AQARSAASKVKVSMKF for the induction of TNF- was comparable to that of E. coli LPS. In addition, this response to the laminin 5 peptide was similar to LPS in that it was a rapid, transient response, as the levels dropped to near baseline levels by 12 h postinstillation. This is consistent with the production of TNF- following intrapulmonary instillation of Pseudomonas aeruginosa LPS (43).
FIGURE 3. Administration of AQARSAASKVKVSMKF in vivo results in increased TNF- levels in lavage fluid. Mice were inoculated in lungs (A) or peritoneum (B) with PBS alone or PBS containing 200 μg of 5 peptide or LPS. At various times postinstillation, lavage fluids were collected and assayed for the presence of TNF- by ELISA. The concentration of TNF- per milliliter of lavage fluid ± SEM is shown. Data are representative of three separate experiments using at least three mice per condition.
Because inflammatory responses can differ based on site of inflammation (44), we compared the response to AQARSAASKVKVSMKF stimulation in the lung and the peritoneal cavity. Similar to the response in the lung, a rapid, transient production of TNF- in response to the laminin 5 peptide was detected in peritoneal lavage fluid (Fig. 3B). These data suggest that TNF- production is a generalized response to AQARSAASKVKVSMKF stimulation in vivo.
Identification of residues in AQARSAASKVKVSMKF required for induction of TNF- expression by Ms
Previously, we and others have shown that a peptide derived from the laminin 1 chain induces MMP-9 production by monocytes/Ms (5, 6, 8). Our studies further indicated that a peptide derived from a corresponding region of laminin 5 chain also up-regulates M MMP-9 production, whereas the corresponding region of laminin 3 has no effect (8). ASKVKVSMKF was identified as the active domain of the laminin 5 peptide for MMP-9 production and a scrambled version of this minimal peptide, KAKSFVMVSK, failed to induce MMP-9 production by Ms. Therefore, we examined the effects of these peptides and other matrix peptides on TNF- production by Ms. Similar to their effects on MMP-9 expression, the laminin 1 peptide, SRARKQAASIKVAVSADR, increased the production of TNF- to levels comparable to the laminin 5 peptide, AQARSAASKVKVSMKF, whereas incubation with the laminin 3 peptide, QQARDAANKVAIPMRF, had no effect (Fig. 4). To further define the functional domains of the laminin 5 peptide responsible for TNF- induction, RAW264.7
Ms were treated with peptides derived from the N-terminal or C-terminal region of AQARSAASKVKVSMKF. AQARSAA did not affect the production of TNF-, but stimulation with ASKVKVSMKF did induce TNF- expression by Ms. However, the scrambled version of this peptide, KAKSFVMVSK, did not induce M TNF- production. Interestingly, the laminin 1-derived YIGSR peptide, which triggers biological activities such as cell adhesion (45, 46) and induction of MMP expression (7), had no effect on M TNF- expression. Furthermore, a peptide derived from the ECM component elastin (VGVAPG), which is known to modulate inflammatory cell functions (47), failed to increase TNF- production by Ms. These data indicate that the effects of AQARSAASKVKVSMKF on Ms are sequence-specific and, as with MMP-9 production, ASKVKVSMKF was sufficient to induce TNF- production by RAW264.7 cells.
FIGURE 4. Identification of the sequence in AQARSAASKVKVSMKF required for TNF- production. A, The amino acid sequences of the tested peptides. B, RAW264.7 cells were incubated in either MSFM alone (c) or containing 100 μg/ml the indicated peptides. Following 24 h stimulation, the conditioned media were assayed for the presence of TNF- by ELISA. The concentration of TNF- per milliliter of conditioned media ± SEM is shown. Data are representative of at least three separate experiments.
AQARSAASKVKVSMKF increases the production of TNFR-II, but not TNFR-I, by RAW264.7 cells
Because AQARSAASKVKVSMKF increased the steady-state level of TNFR-II mRNA, we examined whether this peptide affects protein levels of the TNFRs by Ms. Because TNFRs are shed from the cell surface by TACE (48, 49), we assayed the conditioned media for soluble TNFRs by ELISA. Similar to previous studies that TNFR-I and TNFR-II are constitutively released by monocytes (42), both soluble TNFR-I and TNFR-II were detected in the conditioned media of untreated RAW264.7 Ms (Fig. 5). However, following stimulation with the laminin 5 peptide, a marked increase in the amount of soluble TNFR-II to levels comparable to that stimulated by LPS was detected. In contrast, the amount of soluble TNFR-I did not change in response to either AQARSAASKVKVSMKF or LPS. Taken together, these data suggest that AQARSAASKVKVSMKF selectively induces the production of TNFR-II by Ms.
FIGURE 5. AQARSAASKVKVSMKF triggers production of TNFR-II, but not TNFR-I, by RAW264.7 Ms. RAW264.7 cells were incubated in either MSFM alone (c) or containing 100 μg/ml laminin 5 peptide or LPS for 24 h. Following incubation, the conditioned media were assayed for the presence of TNFR-I or TNFR-II by ELISA. The concentration of TNFR per milliliter of conditioned media ± SEM is shown. Data are representative of at least three separate experiments.
AQARSAASKVKVSMKF induces production of TNF- and TNFR-II before MMP-9
Because MMP-9 production by Ms can be stimulated by TNF- (50), we compared the time course production of TNF-, TNFR-II, and MMP-9 in response to the laminin 5 peptide. Like TNF- (Fig. 2A), an increase in soluble TNFR-II was detectable in the conditioned media 3 h after stimulation with AQARSAASKVKVSMKF and continued to increase over the 24-h period (Fig. 6A). However, the increase in MMP-9 production in response to the laminin 5 peptide was not detectable until 12 h of stimulation (Fig. 6B). These data suggest that TNF- signaling is required for AQARSAASKVKVSMKF-stimulated production of MMP-9 by Ms.
FIGURE 6. AQARSAASKVKVSMKF induces production of TNFR-II before MMP-9. RAW264.7 cells were stimulated with 100 μg/ml laminin 5 peptide in MSFM for up to 24 h. Following incubation, the conditioned media were assayed for the presence of TNFR-II by ELISA (A) and MMP-9 by gelatin zymography (B). The concentration of TNFR-II per milliliter of conditioned media ± SEM is shown. Data are representative of at least three separate experiments.
TNFR-II signaling triggers AQARSAASKVKVSMKF-mediated induction of MMP-9 expression by Ms
To examine whether the AQARSAASKVKVSMKF-induced MMP-9 production by Ms is TNF--dependent, primary peritoneal Ms from wild-type and double TNFR knockout mice were isolated and stimulated with the laminin 5 peptide. Consistent with previously reported findings (8), MMP-9 was increased in the conditioned media of primary wild-type Ms after stimulation with AQARSAASKVKVSMKF (Fig. 7A). However, peritoneal Ms isolated from mice deficient in both TNFRs failed to produce MMP-9 in response to the laminin 5 peptide, which indicates that AQARSAASKVKVSMKF-induced MMP-9 production by Ms requires TNF- signaling. This induction of MMP-9 production by Ms was not due to binding of AQARSAASKVKVSMKF to the TNFRs, because specific binding of the laminin 5 peptide to double TNFR-deficient Ms was comparable to binding to wild-type Ms, as detected in binding studies using 125I-labeled peptide (data not shown).
FIGURE 7. TNFR-II-induced signaling is required for AQARSAASKVKVSMKF-mediated induction of MMP-9 production by Ms. Primary peritoneal Ms from wild-type and TNFR-I, TNFR-II, or double TNFR-deficient mice were incubated in either MSFM alone (c) or containing 100 μg/ml laminin 5 peptide (5) for 24 h. Following incubation, the conditioned media were assayed for MMP-9 by gelatin zymography. Mouse PMN lysate containing MMP-9 was run as a positive control. Gels are representative of multiple experiments using independent isolates of Ms.
Because TNF- can activate NF-B via TNFR-I and TNFR-II (22, 23), we investigated whether AQARSAASKVKVSMKF-induced MMP-9 production by Ms is influenced by the individual TNFRs. Peritoneal Ms isolated from mice deficient in either TNFR-I or TNFR-II, were stimulated with AQARSAASKVKVSMKF for 24 h, and gelatinase activity in the conditioned media was analyzed by zymography (Fig. 7B). MMP-9 production triggered by the laminin 5 peptide in TNFR-I-deficient Ms was comparable to Ms from wild-type mice. However, Ms lacking TNFR-II had minimal MMP-9 induction. These findings indicate that AQARSAASKVKVSMKF-induced MMP-9 production by Ms in vitro is mediated through TNFR-II signaling.
TNF- signaling is not required for AQARSAASKVKVSMKF-induced PMN emigration or release of MMP-9, but is required for M production of MMP-9 in vivo
Because AQARSAASKVKVSMKF-induced production of MMP-9 by Ms required TNF- signaling, we examined the TNF--dependence of other responses to AQARSAASKVKVSMKF by inflammatory cells. To examine whether the AQARSAASKVKVSMKF-induced emigration of PMNs was TNF--dependent, wild-type and double TNFR-deficient mice were inoculated i.p. with the laminin 5 peptide and the recruitment of PMNs and Ms into the peritoneal cavity was evaluated. At 4 h postinoculation of the laminin 5 peptide, an increase in PMNs in the lavage fluid was observed as compared with that from PBS-treated mice (Fig. 8A). Although the number of resident Ms in the peritoneal cavity of double TNFR-deficient mice is greater than that of wild-type mice (data not shown), the percent of PMNs to total cells in the peritoneal lavage fluid of mice deficient in both TNFRs (31%) was similar to that of wild-type mice (30%). These data suggest that a M-derived cytokine(s), independent of TNF-, is responsible for the PMN recruitment response to the laminin 5 peptide. This speculation fits with the finding that recruitment of neutrophils to the peritoneal cavity in sterile peritonitis in the rat is dependent on peritoneal Ms (51).
FIGURE 8. TNF- signaling is not required for AQARSAASKVKVSMKF-induced PMN emigration or release of MMP-9, but is required for M production of MMP-9 in vivo. Wild-type and double TNFR-deficient mice were inoculated i.p. with PBS alone or containing 200 μg of laminin 5 peptide. Following 4 h of incubation, peritoneal lavage fluids were collected, and the cells were separated from the lavage fluids and characterized as indicated in Materials and Methods. A, The number of PMNs and Ms per milliliter of lavage fluid ± SEM is shown. B, Gelatin zymography for the MMP-9 in peritoneal lavage fluid. C, RT-PCR for murine MMP-9 (arrow) or GAPDH mRNAs from cells in peritoneal lavage fluid. Data are representative of three separate experiments using at least three mice per condition.
Coinciding with an increase in PMNs in the peritoneal lavage fluid, AQARSAASKVKVSMKF led to increased MMP-9 in the peritoneal lavage fluid in both wild-type and double TNFR-deficient mice (Fig. 8B). Because AQARSAASKVKVSMKF-induced production of MMP-9 by Ms requires TNF--signaling (Fig. 7), the MMP-9 detected in the lavage fluid likely represents preformed MMP-9 protein released from PMNs. Therefore, AQARSAASKVKVSMKF-induced MMP-9 release from PMNs does not require TNF- signaling. Alternatively, some of the MMP-9 detected could be from Ms if, in vivo, the induction of M-derived MMP-9 by the laminin 5 peptide does not require TNF- signaling.
To determine whether AQARSAASKVKVSMKF-induced production of MMP-9 by Ms is TNF--dependent in vivo as it is in vitro (Fig. 7), RNAs isolated from cells in the peritoneal lavage fluids were subjected to RT-PCR for MMP-9. Because PMNs do not synthesize MMP-9 outside of the bone marrow (52), the MMP-9 RNA detected from cells in the peritoneal lavage fluid is from Ms. An increase in M MMP-9 mRNA was detected after i.p. inoculation of wild-type mice with AQARSAASKVKVSMKF, but peritoneal Ms isolated from double TNFR-deficient mice failed to increase MMP-9 mRNA in response to the laminin 5 peptide (Fig. 8C). Thus, TNF- signaling is required for AQARSAASKVKVSMKF-induced MMP-9 production by Ms in vivo.
Discussion
We previously have shown that the laminin 5-derived peptide AQARSAASKVKVSMKF is chemotactic in vitro and in vivo (8). However, the chemotactic response detected in vivo following intrapulmonary instillation of the peptide was greater than that seen in in vitro studies. This led us to hypothesize that AQARSSASKVKVSMKF has additional effects on inflammatory cells that could amplify the chemotactic response. To test this, RNA from RAW264.7 Ms that had been stimulated with AQARSAASKVKVSMKF was analyzed by microarray. A major finding of the present study is that AQARSAASKVKVSMKF can induce the expression of several proinflammatory genes, suggesting that the AQARSAASKVKVSMKF-induced M cytokine expression could be responsible for the increased chemotactic response in vivo. To our knowledge, this is the first description of a BM component regulating M cytokine gene expression.
Intrapulmonary instillation of AQARSAASKVKVSMKF induces PMN infiltration (8). Because TNF- induces PMN migration (53), we examined whether the AQARSAASKVKVSMKF-induced emigration of PMNs was TNF--dependent. Other investigators obtained varying results in studies of the role of TNF- in PMN emigration in response to lung injury or infection (54, 55, 56, 57). We found that the number of neutrophils in the BAL fluid of mice deficient in both TNFRs were similar to that of wild-type mice at 1 day postinstillation of AQARSAASKVKVSMKF (data not shown). In addition, PMN recruitment into the peritoneal cavity in response to the laminin 5 peptide was the same in both wild-type and double TNFR-deficient mice (Fig. 8B). Therefore, the emigration of PMNs in response to the laminin 5 peptide does not require TNFR-triggered signaling. Combined with the microarray data presented in Table I, these data suggest that in the absence of TNF- signaling, other cytokines induced by AQARSAASKVKVSMKF are sufficient to induce PMN emigration.
Previously, we reported that AQARSAASKVKVSMKF can induce MMP-9 expression by Ms (8). Here we found that AQARSAASKVKVSMKF also induces expression of TNF- by Ms. Because TNF- can induce MMP-9 via activation of NF-B (50), we tested the hypothesis that AQARSAASKVKVSMKF activates M production of MMP-9 via TNF--dependent signaling. Our results show that AQARSAASKVKVSMKF increases production of both TNF- and TNFR-II before production of MMP-9 by Ms and that TNF- signaling, specifically through TNFR-II, is required for MMP-9 production (Fig. 7). Our previous studies also showed that AQARSAASKVKVSMKF induces MMP-9 release by PMNs (8). PMNs express surface TNFRs and respond to TNF-. However, MMP-9 release by PMNs does not require new protein synthesis (52), suggesting that AQARSAASKVKVSMKF-induced release of MMP-9 by PMNs is not dependent on TNF--induced NF-B activation. This hypothesis is supported by the presence of both PMNs and MMP-9 in the peritoneal lavage fluid of mice deficient in both TNFRs following i.p. administration of the laminin 5 peptide (Fig. 8). Whether PMNs have a different receptor for the laminin 5 peptide than Ms or whether the signaling mechanism for MMP-9 production is different between the two cell types is currently being investigated.
The first line of defense against infections is the epithelial cell layer itself. Tissue injury may result in the loss of this epithelial layer, exposing the underlying BM. Unlike the interstitial matrix which provides cell support, tensile strength, and elasticity, the BM also serves as a selective and protective barrier. Signals exposed in BM components during times of loss of BM integrity could be recognized by cells of the innate immune system as a breach in host defense. These signals would recruit leukocytes, including PMNs and monocytes, and coordinate their actions to prevent infection. Our studies have shown that a synthetic peptide derived from the -chain of laminin-10 (511), a prominent BM component, is chemotactic for both PMNs and Ms (8) and can induce M cytokine production. We further demonstrate that the corresponding region from the -chain of laminin-1 (111), but not of laminin-5 (332), can also induce TNF-. Interestingly, the elastin-derived peptide, VGVAPG, has chemotactic activity (47), but failed to increase TNF- production by Ms, which indicates that the mechanism of recruitment of inflammatory cells by VGVAPG is different from that of AQARSAASKVKVSMKF, and suggests specificity in signals triggered by ECM components. Fragments or peptides derived from other ECM components have also been shown to be chemotactic (58, 59, 60, 61, 62), but whether these peptides have the capability to induce inflammatory cell cytokine production remains to be investigated.
Hallmarks of chronic inflammation include the accumulation of inflammatory cells and increased degradation of ECM components. During tissue remodeling, increased turnover of ECM can result in release of degradation products of ECM components. The presence of laminin or laminin fragments has been demonstrated in several chronic inflammatory diseases, including chronic hepatitis (17, 18), diffuse interstitial lung diseases (19), renal diseases (16), and inflammatory bowel diseases (21). TNF- has also been shown to be increased in many of these chronic inflammatory diseases (63, 64, 65, 66). Although TNF- has a central role in the innate immune system for protection against infections, when produced in excess, TNF- may also orchestrate a chronic inflammatory response that leads to severe tissue damage. Thus, the regulation of inflammatory cell functions by interactions with BM components may provide important information in determining how inflammatory responses are perpetuated or resolved.
The biological activities of TNF- are mediated by two structurally related, but functionally distinct receptors, TNFR-I and TNFR-II. The majority of inflammatory responses classically attributed to TNF- are mediated by TNFR-I. Studies using TNFR-deficient mice reveal that TNFR-I plays a critical role in mediating endotoxic shock, and LPS- and TNF--induced cytotoxicity. However, there is increasing evidence for an important independent role of TNFR-II signaling in chronic inflammatory conditions. For example, an increase in TNFR-II has been demonstrated to aggravate experimental colitis and contributes to the chronicity of inflammation in vivo (67). Our data indicates that the AQARSAASKVKVSMKF-induced expression of MMP-9 by Ms is predominantly mediated through TNFR-II. Whether inflammatory activities of AQARSAASKVKVSMKF affect tissue remodeling and repair or contribute to the pathophysiology of inflammatory diseases through TNFR-II signaling requires further investigation.
In conclusion, the laminin 5 peptide AQARSAASKVKVSMKF is a clear example of a cryptic matrix epitope that has biological activity. AQARSAASKVKVSMKF induces diverse inflammatory cell activities, including chemotaxis and expression of proteinases, cytokines, and cytokine receptors. These data suggest that exposure of cryptic BM motifs at sites of injury may be an important mechanism for inducing expression of cytokines, expression and release of proteases, and recruitment of inflammatory cell. Together, these processes would serve to initiate repair and remodeling of the matrix and restoration of tissue integrity.
Acknowledgments
We thank Mark Watson for assistance in the analysis of the Affymetrix GeneChips and Robert Mecham and Thomas Broekelmann for providing the laminin-derived peptides.
Footnotes
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 National Institutes of Health, National Heart, Lung and, Blood Institute Grants IF32 HL071423-01 (to T.L.A.-K.), HL29594 and HL47328, the Alan A. and Edith L. Wolff Charitable Trust (to R.M.S.), and a Research Fellowship from the American Lung Association (to J.J.A.).
2 Address correspondence and reprint requests to Dr. Robert M. Senior, Department of Medicine, Washington University School of Medicine, 902 Yalem, Box 8052, 660 South Euclid Avenue, St. Louis, MO 63110. E-mail address: rsenior@im.wustl.edu
3 Abbreviations used in this paper: BM, basement membrane; ECM, extracellular matrix; MMP, matrix metalloproteinase; PMN, neutrophil; M, macrophage; TACE, TNF- converting enzyme; BAL, bronchoalveolar lavage; MSFM, M serum-free medium.
Received for publication May 13, 2004. Accepted for publication November 2, 2004.
References
Tani, T., V. P. Lehto, I. Virtanen. 1999. Expression of laminins 1 and 10 in carcinoma cells and comparison of their roles in cell adhesion. Exp. Cell Res. 248:115.
Ignatius, M. J., T. H. Large, M. Houde, J. W. Tawil, A. Barton, F. Esch, S. Carbonetto, L. F. Reichardt. 1990. Molecular cloning of the rat integrin 1-subunit: a receptor for laminin and collagen. J. Cell Biol. 111:709.
Lee, E. C., M. M. Lotz, G. D. Steele, Jr, A. M. Mercurio. 1992. The integrin 64 is a laminin receptor. J. Cell Biol. 117:671.
von der Mark, H., I. Williams, O. Wendler, L. Sorokin, K. von der Mark, E. Poschl. 2002. Alternative splice variants of 71 integrin selectively recognize different laminin isoforms. J. Biol. Chem. 277:6012.
Corcoran, M. L., M. C. Kibbey, H. K. Kleinman, L. M. Wahl. 1995. Laminin SIKVAV peptide induction of monocyte/macrophage prostaglandin E2 and matrix metalloproteinases. J. Biol. Chem. 270:10365.
Khan, K. M., D. J. Falcone. 1997. Role of laminin in matrix induction of macrophage urokinase-type plasminogen activator and 92-kDa metalloproteinase expression. J. Biol. Chem. 272:8270.
Kanemoto, T., R. Reich, L. Royce, D. Greatorex, S. H. Adler, N. Shiraishi, G. R. Martin, Y. Yamada, H. K. Kleinman. 1990. Identification of an amino acid sequence from the laminin A chain that stimulates metastasis and collagenase IV production. Proc. Natl. Acad. Sci. USA 87:2279.
Adair-Kirk, T. L., J. J. Atkinson, T. J. Broekelmann, M. Doi, K. Tryggvason, J. H. Miner, R. P. Mecham, R. M. Senior. 2003. A site on laminin 5, AQARSAASKVKVSMKF, induces inflammatory cell production of matrix metalloproteinase-9 and chemotaxis. J. Immunol. 171:398.
Jacob, S. S., P. Shastry, P. R. Sudhakaran. 2002. Monocyte-macrophage differentiation in vitro: modulation by extracellular matrix protein substratum. Mol. Cell. Biochem. 233:9.
Xie, B., A. Laouar, E. Huberman. 1998. Fibronectin-mediated cell adhesion is required for induction of 92-kDa type IV collagenase/gelatinase (MMP-9) gene expression during macrophage differentiation. The signaling role of protein kinase C-. J. Biol. Chem. 273:11576.
Beezhold, D. H., C. Personius. 1992. Fibronectin fragments stimulate tumor necrosis factor secretion by human monocytes. J. Leukocyte Biol. 51:59.
McKee, C. M., M. B. Penno, M. Cowman, M. D. Burdick, R. M. Strieter, C. Bao, P. W. Noble. 1996. Hyaluronan (HA) fragments induce chemokine gene expression in alveolar macrophages: the role of HA size and CD44. J. Clin. Invest. 98:2403.
Horton, M. R., C. M. McKee, C. Bao, F. Liao, J. M. Farber, J. Hodge-DuFour, E. Pure, B. L. Oliver, T. M. Wright, P. W. Noble. 1998. Hyaluronan fragments synergize with interferon- to induce the C-X-C chemokines mig and interferon-inducible protein-10 in mouse macrophages. J. Biol. Chem. 273:35088.
Horton, M. R., M. D. Burdick, R. M. Strieter, C. Bao, P. W. Noble. 1998. Regulation of hyaluronan-induced chemokine gene expression by IL-10 and IFN- in mouse macrophages. J. Immunol. 160:3023.
Walker, P. D., G. P. Kaushal, S. V. Shah. 1998. Meprin A, the major matrix degrading enzyme in renal tubules, produces a novel nidogen fragment in vitro and in vivo. Kidney Int. 53:1673.
Horikoshi, S., H. Koide. 1991. Serum laminin P1 fragment concentration in renal diseases. Clin. Chim. Acta 196:185.
Walsh, K. M., A. Fletcher, R. N. MacSween, A. J. Morris. 2000. Basement membrane peptides as markers of liver disease in chronic hepatitis C. J Hepatol. 32:325.
Gabrielli, G. B., F. Capra, M. Casaril, R. Corrocher, R. Colombari, G. De Sandre. 1996. Serum laminin P1 in chronic viral hepatitis: correlations with liver histological activity and diagnostic value. Clin. Chim. Acta 252:171.
Perez-Arellano, J. L., M. J. Pedraz, A. Fuertes, J. L. de la Cruz, J. M. Gonzalez de Buitrago, A. Jimenez. 1993. Laminin fragment P1 is increased in the lower respiratory tract of patients with diffuse interstitial lung diseases. Chest 104:1163.
Horslev-Petersen, K., K. D. Bentsen, P. Halberg, P. Junker, K. I. Kivirikko, K. Majamaa, L. Risteli, J. Risteli, I. Lorenzen. 1988. Connective tissue metabolites in serum as markers of disease activity in patients with rheumatoid arthritis. Clin. Exp. Rheumatol. 6:129.
Koutroubakis, I. E., E. Petinaki, P. Dimoulios, E. Vardas, M. Roussomoustakaki, A. N. Maniatis, E. A. Kouroumalis. 2003. Serum laminin and collagen IV in inflammatory bowel disease. J. Clin. Pathol. 56:817.
Baud, V., M. Karin. 2001. Signal transduction by tumor necrosis factor and its relatives. Trends Cell Biol. 11:372.(Tracy L. Adair-Kirk, Jeff)
Tissue injury triggers inflammatory responses that may result in release of degradation products or exposure of cryptic domains of extracellular matrix components. Previously, we have shown that a cryptic peptide (AQARSAASKVKVSMKF) in the -chain of laminin-10 (511), a prominent basement membrane component, is chemotactic for both neutrophils (PMNs) and macrophages (Ms) and induces matrix metalloproteinase-9 (MMP-9) production. To determine whether AQARSAASKVKVSMKF has additional effects on inflammatory cells, we performed microarray analysis of RNA from RAW264.7 Ms stimulated with AQARSAASKVKVSMKF. Several cytokines and cytokine receptors were increased >3-fold in response to the laminin 5 peptide. Among these were TNF- and one of its receptors, the p75 TNFR (TNFR-II), increasing 3.5- and 5.7-fold, respectively. However, the peptide had no effect on p55 TNFR (TNFR-I) expression. Corroborating the microarray data, the protein levels of TNF- and TNFR-II were increased following stimulation of RAW264.7 cells with AQARSAASKVKVSMKF. In addition, we determined that the production of TNF- and TNFR-II in response to AQARSAASKVKVSMKF preceded the production of MMP-9. Furthermore, using primary Ms from mice deficient in TNFR-I, TNFR-II, or both TNF- receptors (TNFRs), we determined that AQARSAASKVKVSMKF induces MMP-9 expression by Ms through a pathway triggered by TNFR-II. However, TNF- signaling is not required for AQARSAASKVKVSMKF-induced PMN release of MMP-9 or PMN emigration. These data suggest that interactions of inflammatory cells with basement membrane components may orchestrate immune responses by inducing expression of cytokines, recruitment of inflammatory cells, and release of proteinases.
Introduction
Basement membranes (BMs)3 are thin sheets of specialized extracellular matrix (ECM) which lay beneath epithelial and endothelial cells and surround other cell types, including adipocytes, muscle, and nerve cells. Besides providing tissue boundaries and structural support, BMs influence cell proliferation, differentiation, and migration (1, 2, 3, 4). BMs are predominantly composed of laminins, type IV collagen, heparan-sulfate proteoglycans, and entactin/nidogen. In tissue remodeling during inflammation, increased ECM turnover can result in release of degradation products or exposure of cryptic domains of ECM components that can elicit biological responses distinct from the parent molecules (3, 4).
Although the role of ECM in affecting inflammatory cell functions has been examined, the predominant focus has been the effect of ECM components on adhesion, migration (5, 6, 7, 8, 9), and proteinase production (5, 6, 7, 8, 9, 10). Little is known about the effects of ECM components on cytokine production. However, fragments of fibronectin (11) and hyaluronan stimulate macrophage (M) cytokine production (12, 13, 14), suggesting that ECM components may have regulatory roles in inflammatory responses. Although fragments of BM components, such as laminins, type IV collagen, and entactin, have been detected in biological fluids in association with various inflammatory diseases (15, 16, 17, 18, 19, 20, 21), it is unknown whether BM components are active in the regulation of inflammation.
The inflammatory response consists of the release of proinflammatory cytokines, followed by the recruitment of circulating leukocytes, which become activated at the inflammatory site and release further mediators. TNF-, a powerful mediator of inflammation, can be produced by a variety of cell types, including activated monocytes/Ms (22, 23). Synthesized as a membrane-anchored precursor, active TNF- can be proteolytically released by the TNF- converting enzyme (TACE)/a disintegrin and metalloproteinase 17 (24, 25). The biological activities of TNF- are mediated by two structurally related but functionally distinct receptors, termed TNFR-I (TNFR p55/p60) and TNFR-II (TNFR p75/p80). Although binding of activated TNF- to its receptors can result in activation of proapoptotic signaling events (26, 27, 28), signaling triggered by TNFR-I and TNFR-II usually causes activation of the AP-1 and NF-B transcription factors, which then induce numerous genes involved in inflammatory responses (22, 23).
Laminins are a family of integral BM glycoproteins, each containing an -, -, and -chain (29). To date, five , four , and three laminin chains have been identified, which assemble into at least 15 laminin isoforms (30, 31, 32, 33, 34). Immunohistochemical data indicate that laminin-10 (511) is one of the most abundant laminin heterotrimers in normal adult tissues (35, 36, 37). Because laminin 5 is widely expressed in fetal and adult tissues, we are characterizing the function of specific domains of the laminin 5 chain. We previously reported that the laminin 5 peptide, AQARSAASKVKVSMKF, induces matrix metalloproteinase-9 (MMP-9) production and is chemotactic for leukocytes in vitro and in vivo (8). Furthermore, our data indicated that AQARSAASKVKVSMKF is cryptic in the intact laminin-10 molecule. In the present study we investigated whether AQARSAASKVKVSMKF has additional effects on inflammatory cells. These studies indicate that AQARSAASKVKVSMKF induces M expression of several cytokines, including TNF-. In addition, AQARSAASKVKVSMKF increases expression levels of TNFR-II, but not TNFR-I. Using primary Ms isolated from mice deficient in TNFR-I, TNFR-II, or both TNFRs, we determined that AQARSAASKVKVSMKF induces MMP-9 through a TNFR-II-dependent pathway. Taken together, our results indicate a novel regulatory role for a cryptic peptide of laminin 5, AQARSAASKVKVSMKF, in inflammatory responses and suggest that BM components play a role in orchestrating immune responses triggered by tissue damage.
Materials and Methods
Peptides
The YIGSR and VGVAPG peptides, which are derived from the laminin 1 chain and elastin, respectively, were obtained from Sigma-Aldrich. Other laminin-derived peptides were synthesized and purified as previously described (8). The sequences of these peptides are as follows: 1, SRARKQAASIKVAVSADR; 3, QQARDAANKVAIPMRF; 5, AQARSAASKVKVSMKF; 5f, AQARSAA; 5, ASKVKVSMKF; and 5b-sc, KAKSFVMVSK.
Mouse strains
Six- to 8-wk-old mice deficient in TNFR-I, TNFR-II, or both TNFRs were obtained from The Jackson Laboratory. Wild-type 129 SvEv and C57BL/6 mice (Taconic Farms) were housed in a barrier animal facility in the veterinary care of the Department of Comparative Medicine at Washington University School of Medicine (St. Louis, MO). All procedures using mice were approved by the Washington University School of Medicine Animal Studies Committee and were performed in accordance with the Animal Welfare Act and the Guide for the Care and Use of Laboratory Animals.
Cell culture
Murine RAW264.7 Ms, obtained from American Type Culture Collection, were maintained in low bicarbonate (1.5 g/L) DMEM, supplemented with 10% FBS, penicillin (100 U/ml), streptomycin (100 μg/ml), and 4 mM L-glutamine. Primary mouse Ms were harvested by peritoneal lavage as previously described (8). To isolate human monocytes, heparinized, human venous blood from normal volunteers was diluted 1/2 in sterile saline, layered on one-half volume of Histopaque-1077 (Sigma-Aldrich), and centrifuged at 750 x g for 30 min at 4°C. The mononuclear layer was collected and applied to a density gradient. Supernatant was removed and cells were suspended in media and counted. Procedures using human subjects were approved by the Washington University School of Medicine Human Studies Committee.
RNA isolation, biotinylated target synthesis, and microarray hybridization
RAW264.7 Ms were incubated in M serum-free medium (MSFM; Invitrogen Life Technologies) alone or containing 100 μg/ml AQARSAASKVKVSMKF for 24 h. Total RNA was extracted by the ToTally RNA isolation kit (Ambion), and equal amounts of total RNA from three separate experiments were pooled to minimize variability. Biotinylated cRNA targets were generated from 20 μg of total RNA using the Bioarray High Yield RNA Transcript Labeling kit (Enzo Life Sciences) and purified using RNeasy spin columns (Qiagen). Twenty micrograms of each biotinylated cRNA target was fragmented and then hybridized to identical Affymetrix Murine Genome Mu74Av2 GeneChips in the Siteman Cancer Center GeneChip Facility (Washington University School of Medicine).
Analysis of microarray data
The images from the scanned microarrays were processed using Affymetrix Microarray Analysis Suite 4.0 and analyzed using GeneSpring 6.1 software (Silicon Genetics). Normalization was performed using the parameters recommended by the software as follows: 1) per chip, by dividing the raw data by the 50th percentile of all measurements, and 2) per gene to a specific sample, by dividing the normalized data of all samples to the untreated control sample. From a total of the 12,000 probe sets on the array, those representing control sequences and genes that were scored as "absent" by the Affymetrix Detection criterion in both samples were excluded from analysis. Genes that increased or decreased expression at least 3-fold relative to the untreated control sample were organized into functional categories using the Simplified Gene Ontology, which is based on the Affymetrix annotated database for the Mu74Av2 chip. The Signaling Ligand and Receptor genes reported here were identified from the Simplified Gene Ontology lists as Signal Transducers in the Molecular Functions category, which contains chemokines and cytokines. The complete set of data and sample metadata may be downloaded at http://bioinformatics.wustl.edu.
RT-PCR analysis of TNF- and TNFRs
RAW264.7 Ms were incubated 24 h in MSFM alone or containing 100 μg/ml AQARSAASKVKVSMKF. One microgram of total RNA extracted from these cells was reverse transcribed using random hexamers and the GeneAmp RNA PCR Core kit (Roche). After reverse transcription, PCR were performed using primers specific for TNF-, TNFR-I, and TNFR-II. The primers for mouse TNF- were 5'-TTCTGTCTACTGAACTTCGGGGTGATCGGTCC and 5'-GTATGAGATAGCAAATCGGCTGACGGTGTGGG (38). The primers for mouse TNFR-I were 5'-CCATCTTCGGTCCTAGTAACTG and 5'-CAGGTTCATCTTGGAAAGCAC (39). The primers for mouse TNFR-II were 5'-GATGCCAAGGTGCCTCATG and 5'-GAGCTGCTACAGACGTTCACG (39). The primers for mouse GAPDH were 5'-CCCCTTCATTGACCTCAACTACATGG and 5'-GACATCAAGAAGGTGGTGAAGCAGGC (40). Products of 354, 391, 290, and 689 bp were predicted for TNF-, TNFR-I, TNFR-II, and GAPDH, respectively. The resulting amplification products were electrophoresed on a 1.5% agarose gel, stained with ethidium bromide, and visualized under UV light. Following scanning of the gels, quantification for each reaction product was performed using the public domain NIH Image program (National Institutes of Health, http://rsb.info.nih.gov/nih-image/). The mRNA levels of TNF-, TNFR-I, and TNFR-II were normalized to GAPDH expression. Data represent at least three independent experiments.
Detection of soluble TNF- and TNFRs by ELISA
RAW264.7 or primary peritoneal Ms were incubated for 24 h in MSFM alone or containing 100 μg/ml various peptides. At designated times, conditioned medium were harvested and assayed for soluble TNF-, TNFR-I, or TNFR-II in triplicate wells by ELISA according to the manufacturers’ protocols (OptEIA Mouse TNF- set, BD Pharmingen; DuoSets for Mouse sTNFR-I and sTNFR-II, R&D Systems). Data represent at least three independent experiments.
Gelatin zymography
Gelatinolytic activity was determined by gelatin zymography as described previously (8). Briefly, conditioned media were electrophoresed in a 7.5% acrylamide gel containing 1 mg/ml gelatin, and gels were washed in 2.5% Triton X-100 and incubated for 12 h in buffer containing 50 mM Tris (pH 8.0), 10 mM CaCl2, and 1 mM ZnCl2. Following incubation, gels were stained with Coomassie blue, destained, and gelatinolytic activity was detected as clear bands on a blue background. Mouse neutrophil (PMN) lysate was run as positive control for MMP-9.
Collection of bronchoalveolar lavage (BAL) fluid
Intrapulmonary instillation of the laminin 5 peptide and collection of BAL fluid were performed as previously described (8). Briefly, mice were anesthetized, placed supine, and dropwise inoculated intranasally with 50 μl of PBS, AQARSAASKVKVSMKF (4 mg/ml), or LPS from Escherichia coli (4 mg/ml) (Sigma-Aldrich). At various times, the mice were asphyxiated with carbon dioxide, and BAL fluid was retrieved by injecting PBS through the trachea. Total and leukocyte differential cell counts were performed. The BAL fluid was assayed for MMP-9 by gelatin zymography and soluble TNF- by ELISA.
Collection of peritoneal lavage fluid
Wild-type mice or mice deficient in TNFR-II or both TNFRs were inoculated with 100 μl of PBS or AQARSAASKVKVSMKF (2 mg/ml) into the peritoneal cavity (i.p.). At various times, the mice were asphyxiated with carbon dioxide, and the peritoneal cavity was lavaged once with 10 ml of PBS. Total and leukocyte differential cell counts were performed as previously described (8). Total RNA was isolated from these cells and the expression of MMP-9 was assayed by RT-PCR as previously described (40). The peritoneal fluid was assayed for MMP-9 by gelatin zymography and soluble TNF by ELISA.
Results
AQARSAASKVKVSMKF affects M gene expression
Previously, we reported that a laminin 5-derived peptide, AQARSAASKVKVSMKF, induces MMP-9 expression by Ms and is chemotactic for both PMNs and Ms in vivo (8). To examine whether AQARSAASKVKVSMKF regulates inflammation by inducing genes in immune cells, RNA isolated from RAW264.7 Ms that had been stimulated with the laminin 5 peptide for 24 h was analyzed by microarray. Genes whose expression levels increased or decreased at least 3-fold in response to AQARSAASKVKVSMKF were sorted into functional categories using the Affymetrix annotated database for the Mu74Av2 chip. Of the 70 known genes with increased expression of at least 3-fold in response to M stimulation with the laminin 5 peptide (data not shown), 13 were ligands or receptors involved in signal transduction pathways (Table I). These included several members of the CC or CXC chemokine families, as well as the proinflammatory cytokine, TNF-, and one of its receptors, TNFR-II. The expression levels of TNF- and TNFR-II were increased 3.5- and 5.7-fold, respectively, in response to AQARSAASKVKVSMKF. However, the laminin 5 peptide had no effect on the expression of TNFR-I or TACE (data not shown). Consistent with our previous findings (8), MMP-9 expression increased 2.9-fold in response to stimulation of Ms with the laminin 5 peptide (data not shown). These results indicate that AQARSAASKVKVSMKF induces the expression of distinct proinflammatory genes in immune cells and suggest that it can regulate inflammatory responses.
Table I. Microarray detection of signaling ligand and receptor mRNAs increased at least 3-fold by RAW264.7 Ms in response to AQARSAASKVKVSMKF
Confirmation of microarray data
To confirm the microarray data, RT-PCR analyses were performed using RNA isolated from RAW264.7 cells stimulated with AQARSAASKVKVSMKF. Because TNF- is a key modulator of inflammation (22, 23), we focused on the AQARSAASKVKVSMKF-induced expression of TNF- and its receptors. Following a 24-h stimulation with the laminin 5 peptide, an increase in the mRNA levels of TNF- and TNFR-II above the levels in control cells was detected (Fig. 1). In contrast, TNFR-I expression did not increase in response to AQARSAASKVKVSMKF. These findings confirm the microarray data, indicating that AQARSAASKVKVSMKF induces expression of TNF- and TNFR-II, but not TNFR-I, by Ms.
FIGURE 1. AQARSAASKVKVSMKF triggers expression of TNF- and TNFR-II, but not TNFR-I, by RAW264.7 Ms. RNAs isolated from RAW264.7 M that were untreated or stimulated with 100 μg/ml laminin 5 peptide for 24 h were subjected to RT-PCR using primers specific for murine TNF-, TNFR-I, TNFR-II, or GAPDH. The resulting agarose gels were scanned, digitized, and quantified using the NIH Image program. A, The mRNA levels of TNF-, TNFR-I, and TNFR-II were normalized to GAPDH and the mean relative expression ± SEM from at least three independent experiments is shown. B, A representative gel showing increased expression of TNF- and TNFR-II mRNAs in response to AQARSAASKVKVSMKF. Arrow, GAPDH; c, no template control.
AQARSAASKVKVSMKF induces TNF- protein production by mouse and human monocytes/Ms
Because AQARSAASKVKVSMKF increased the steady-state level of TNF- mRNA, we examined whether the laminin 5 peptide increased the production of TNF- protein by Ms. Because monocytes/Ms exhibit constitutive production of active TACE that can release TNF- (41, 42), the level of soluble TNF- in the conditioned media of RAW264.7 Ms was determined as an indicator of the quantity of TNF- protein produced by the cells. An increase in soluble TNF- was detectable 3 h after stimulation with AQARSAASKVKVSMKF and continued to increase over the 24-h period (Fig. 2A). To determine whether the AQARSAASKVKVSMKF-stimulated TNF- production by RAW264.7 Ms extends to primary mouse Ms, peritoneal Ms were stimulated with the laminin 5 peptide. An increase in soluble TNF- was detected in the conditioned media as early as 1 h and continued to increase over the 24-h period (Fig. 2B). These data indicate that AQARSAASKVKVSMKF induces the production of TNF- by primary mouse Ms.
FIGURE 2. AQARSAASKVKVSMKF induces TNF- protein production by mouse and human monocytes/Ms. RAW264.7 cells (A), mouse peritoneal Ms (B), and human blood monocytes (C) were incubated in MSFM alone () or containing 100 μg/ml laminin 5 peptide (?) for various times up to 24 h. Following incubation, the conditioned media were assayed for the presence of TNF- by ELISA. The concentration of TNF- per milliliter of conditioned media ± SEM is shown. Data are representative of at least three separate experiments.
Because this region of mouse laminin 5 has a high degree of sequence identity with the corresponding region of human laminin 5 (35), we examined whether AQARSAASKVKVSMKF could induce TNF- by human monocytes. Within 3 h of stimulation with the laminin 5 peptide, an increase in the production of TNF- by primary human blood monocytes was detected, which continued to rise over the 24-h period (Fig. 2C). These data indicate that the response to the laminin 5 peptide is not limited to mouse and suggests a regulatory role of laminin 5 in inflammatory responses in humans as well.
AQARSAASKVKVSMKF induces TNF- production in vivo
To test whether cryptic peptides of laminin 5 play a role in inflammatory responses in vivo, AQARSAASKVKVSMKF was instilled into the lungs of mice and the lungs were lavaged at various times. At 4 h postinstillation of the laminin 5 peptide, a marked increase in TNF- in BAL fluid was observed as compared with the BAL fluid from control mice treated with PBS (Fig. 3A). The potency of AQARSAASKVKVSMKF for the induction of TNF- was comparable to that of E. coli LPS. In addition, this response to the laminin 5 peptide was similar to LPS in that it was a rapid, transient response, as the levels dropped to near baseline levels by 12 h postinstillation. This is consistent with the production of TNF- following intrapulmonary instillation of Pseudomonas aeruginosa LPS (43).
FIGURE 3. Administration of AQARSAASKVKVSMKF in vivo results in increased TNF- levels in lavage fluid. Mice were inoculated in lungs (A) or peritoneum (B) with PBS alone or PBS containing 200 μg of 5 peptide or LPS. At various times postinstillation, lavage fluids were collected and assayed for the presence of TNF- by ELISA. The concentration of TNF- per milliliter of lavage fluid ± SEM is shown. Data are representative of three separate experiments using at least three mice per condition.
Because inflammatory responses can differ based on site of inflammation (44), we compared the response to AQARSAASKVKVSMKF stimulation in the lung and the peritoneal cavity. Similar to the response in the lung, a rapid, transient production of TNF- in response to the laminin 5 peptide was detected in peritoneal lavage fluid (Fig. 3B). These data suggest that TNF- production is a generalized response to AQARSAASKVKVSMKF stimulation in vivo.
Identification of residues in AQARSAASKVKVSMKF required for induction of TNF- expression by Ms
Previously, we and others have shown that a peptide derived from the laminin 1 chain induces MMP-9 production by monocytes/Ms (5, 6, 8). Our studies further indicated that a peptide derived from a corresponding region of laminin 5 chain also up-regulates M MMP-9 production, whereas the corresponding region of laminin 3 has no effect (8). ASKVKVSMKF was identified as the active domain of the laminin 5 peptide for MMP-9 production and a scrambled version of this minimal peptide, KAKSFVMVSK, failed to induce MMP-9 production by Ms. Therefore, we examined the effects of these peptides and other matrix peptides on TNF- production by Ms. Similar to their effects on MMP-9 expression, the laminin 1 peptide, SRARKQAASIKVAVSADR, increased the production of TNF- to levels comparable to the laminin 5 peptide, AQARSAASKVKVSMKF, whereas incubation with the laminin 3 peptide, QQARDAANKVAIPMRF, had no effect (Fig. 4). To further define the functional domains of the laminin 5 peptide responsible for TNF- induction, RAW264.7
Ms were treated with peptides derived from the N-terminal or C-terminal region of AQARSAASKVKVSMKF. AQARSAA did not affect the production of TNF-, but stimulation with ASKVKVSMKF did induce TNF- expression by Ms. However, the scrambled version of this peptide, KAKSFVMVSK, did not induce M TNF- production. Interestingly, the laminin 1-derived YIGSR peptide, which triggers biological activities such as cell adhesion (45, 46) and induction of MMP expression (7), had no effect on M TNF- expression. Furthermore, a peptide derived from the ECM component elastin (VGVAPG), which is known to modulate inflammatory cell functions (47), failed to increase TNF- production by Ms. These data indicate that the effects of AQARSAASKVKVSMKF on Ms are sequence-specific and, as with MMP-9 production, ASKVKVSMKF was sufficient to induce TNF- production by RAW264.7 cells.
FIGURE 4. Identification of the sequence in AQARSAASKVKVSMKF required for TNF- production. A, The amino acid sequences of the tested peptides. B, RAW264.7 cells were incubated in either MSFM alone (c) or containing 100 μg/ml the indicated peptides. Following 24 h stimulation, the conditioned media were assayed for the presence of TNF- by ELISA. The concentration of TNF- per milliliter of conditioned media ± SEM is shown. Data are representative of at least three separate experiments.
AQARSAASKVKVSMKF increases the production of TNFR-II, but not TNFR-I, by RAW264.7 cells
Because AQARSAASKVKVSMKF increased the steady-state level of TNFR-II mRNA, we examined whether this peptide affects protein levels of the TNFRs by Ms. Because TNFRs are shed from the cell surface by TACE (48, 49), we assayed the conditioned media for soluble TNFRs by ELISA. Similar to previous studies that TNFR-I and TNFR-II are constitutively released by monocytes (42), both soluble TNFR-I and TNFR-II were detected in the conditioned media of untreated RAW264.7 Ms (Fig. 5). However, following stimulation with the laminin 5 peptide, a marked increase in the amount of soluble TNFR-II to levels comparable to that stimulated by LPS was detected. In contrast, the amount of soluble TNFR-I did not change in response to either AQARSAASKVKVSMKF or LPS. Taken together, these data suggest that AQARSAASKVKVSMKF selectively induces the production of TNFR-II by Ms.
FIGURE 5. AQARSAASKVKVSMKF triggers production of TNFR-II, but not TNFR-I, by RAW264.7 Ms. RAW264.7 cells were incubated in either MSFM alone (c) or containing 100 μg/ml laminin 5 peptide or LPS for 24 h. Following incubation, the conditioned media were assayed for the presence of TNFR-I or TNFR-II by ELISA. The concentration of TNFR per milliliter of conditioned media ± SEM is shown. Data are representative of at least three separate experiments.
AQARSAASKVKVSMKF induces production of TNF- and TNFR-II before MMP-9
Because MMP-9 production by Ms can be stimulated by TNF- (50), we compared the time course production of TNF-, TNFR-II, and MMP-9 in response to the laminin 5 peptide. Like TNF- (Fig. 2A), an increase in soluble TNFR-II was detectable in the conditioned media 3 h after stimulation with AQARSAASKVKVSMKF and continued to increase over the 24-h period (Fig. 6A). However, the increase in MMP-9 production in response to the laminin 5 peptide was not detectable until 12 h of stimulation (Fig. 6B). These data suggest that TNF- signaling is required for AQARSAASKVKVSMKF-stimulated production of MMP-9 by Ms.
FIGURE 6. AQARSAASKVKVSMKF induces production of TNFR-II before MMP-9. RAW264.7 cells were stimulated with 100 μg/ml laminin 5 peptide in MSFM for up to 24 h. Following incubation, the conditioned media were assayed for the presence of TNFR-II by ELISA (A) and MMP-9 by gelatin zymography (B). The concentration of TNFR-II per milliliter of conditioned media ± SEM is shown. Data are representative of at least three separate experiments.
TNFR-II signaling triggers AQARSAASKVKVSMKF-mediated induction of MMP-9 expression by Ms
To examine whether the AQARSAASKVKVSMKF-induced MMP-9 production by Ms is TNF--dependent, primary peritoneal Ms from wild-type and double TNFR knockout mice were isolated and stimulated with the laminin 5 peptide. Consistent with previously reported findings (8), MMP-9 was increased in the conditioned media of primary wild-type Ms after stimulation with AQARSAASKVKVSMKF (Fig. 7A). However, peritoneal Ms isolated from mice deficient in both TNFRs failed to produce MMP-9 in response to the laminin 5 peptide, which indicates that AQARSAASKVKVSMKF-induced MMP-9 production by Ms requires TNF- signaling. This induction of MMP-9 production by Ms was not due to binding of AQARSAASKVKVSMKF to the TNFRs, because specific binding of the laminin 5 peptide to double TNFR-deficient Ms was comparable to binding to wild-type Ms, as detected in binding studies using 125I-labeled peptide (data not shown).
FIGURE 7. TNFR-II-induced signaling is required for AQARSAASKVKVSMKF-mediated induction of MMP-9 production by Ms. Primary peritoneal Ms from wild-type and TNFR-I, TNFR-II, or double TNFR-deficient mice were incubated in either MSFM alone (c) or containing 100 μg/ml laminin 5 peptide (5) for 24 h. Following incubation, the conditioned media were assayed for MMP-9 by gelatin zymography. Mouse PMN lysate containing MMP-9 was run as a positive control. Gels are representative of multiple experiments using independent isolates of Ms.
Because TNF- can activate NF-B via TNFR-I and TNFR-II (22, 23), we investigated whether AQARSAASKVKVSMKF-induced MMP-9 production by Ms is influenced by the individual TNFRs. Peritoneal Ms isolated from mice deficient in either TNFR-I or TNFR-II, were stimulated with AQARSAASKVKVSMKF for 24 h, and gelatinase activity in the conditioned media was analyzed by zymography (Fig. 7B). MMP-9 production triggered by the laminin 5 peptide in TNFR-I-deficient Ms was comparable to Ms from wild-type mice. However, Ms lacking TNFR-II had minimal MMP-9 induction. These findings indicate that AQARSAASKVKVSMKF-induced MMP-9 production by Ms in vitro is mediated through TNFR-II signaling.
TNF- signaling is not required for AQARSAASKVKVSMKF-induced PMN emigration or release of MMP-9, but is required for M production of MMP-9 in vivo
Because AQARSAASKVKVSMKF-induced production of MMP-9 by Ms required TNF- signaling, we examined the TNF--dependence of other responses to AQARSAASKVKVSMKF by inflammatory cells. To examine whether the AQARSAASKVKVSMKF-induced emigration of PMNs was TNF--dependent, wild-type and double TNFR-deficient mice were inoculated i.p. with the laminin 5 peptide and the recruitment of PMNs and Ms into the peritoneal cavity was evaluated. At 4 h postinoculation of the laminin 5 peptide, an increase in PMNs in the lavage fluid was observed as compared with that from PBS-treated mice (Fig. 8A). Although the number of resident Ms in the peritoneal cavity of double TNFR-deficient mice is greater than that of wild-type mice (data not shown), the percent of PMNs to total cells in the peritoneal lavage fluid of mice deficient in both TNFRs (31%) was similar to that of wild-type mice (30%). These data suggest that a M-derived cytokine(s), independent of TNF-, is responsible for the PMN recruitment response to the laminin 5 peptide. This speculation fits with the finding that recruitment of neutrophils to the peritoneal cavity in sterile peritonitis in the rat is dependent on peritoneal Ms (51).
FIGURE 8. TNF- signaling is not required for AQARSAASKVKVSMKF-induced PMN emigration or release of MMP-9, but is required for M production of MMP-9 in vivo. Wild-type and double TNFR-deficient mice were inoculated i.p. with PBS alone or containing 200 μg of laminin 5 peptide. Following 4 h of incubation, peritoneal lavage fluids were collected, and the cells were separated from the lavage fluids and characterized as indicated in Materials and Methods. A, The number of PMNs and Ms per milliliter of lavage fluid ± SEM is shown. B, Gelatin zymography for the MMP-9 in peritoneal lavage fluid. C, RT-PCR for murine MMP-9 (arrow) or GAPDH mRNAs from cells in peritoneal lavage fluid. Data are representative of three separate experiments using at least three mice per condition.
Coinciding with an increase in PMNs in the peritoneal lavage fluid, AQARSAASKVKVSMKF led to increased MMP-9 in the peritoneal lavage fluid in both wild-type and double TNFR-deficient mice (Fig. 8B). Because AQARSAASKVKVSMKF-induced production of MMP-9 by Ms requires TNF--signaling (Fig. 7), the MMP-9 detected in the lavage fluid likely represents preformed MMP-9 protein released from PMNs. Therefore, AQARSAASKVKVSMKF-induced MMP-9 release from PMNs does not require TNF- signaling. Alternatively, some of the MMP-9 detected could be from Ms if, in vivo, the induction of M-derived MMP-9 by the laminin 5 peptide does not require TNF- signaling.
To determine whether AQARSAASKVKVSMKF-induced production of MMP-9 by Ms is TNF--dependent in vivo as it is in vitro (Fig. 7), RNAs isolated from cells in the peritoneal lavage fluids were subjected to RT-PCR for MMP-9. Because PMNs do not synthesize MMP-9 outside of the bone marrow (52), the MMP-9 RNA detected from cells in the peritoneal lavage fluid is from Ms. An increase in M MMP-9 mRNA was detected after i.p. inoculation of wild-type mice with AQARSAASKVKVSMKF, but peritoneal Ms isolated from double TNFR-deficient mice failed to increase MMP-9 mRNA in response to the laminin 5 peptide (Fig. 8C). Thus, TNF- signaling is required for AQARSAASKVKVSMKF-induced MMP-9 production by Ms in vivo.
Discussion
We previously have shown that the laminin 5-derived peptide AQARSAASKVKVSMKF is chemotactic in vitro and in vivo (8). However, the chemotactic response detected in vivo following intrapulmonary instillation of the peptide was greater than that seen in in vitro studies. This led us to hypothesize that AQARSSASKVKVSMKF has additional effects on inflammatory cells that could amplify the chemotactic response. To test this, RNA from RAW264.7 Ms that had been stimulated with AQARSAASKVKVSMKF was analyzed by microarray. A major finding of the present study is that AQARSAASKVKVSMKF can induce the expression of several proinflammatory genes, suggesting that the AQARSAASKVKVSMKF-induced M cytokine expression could be responsible for the increased chemotactic response in vivo. To our knowledge, this is the first description of a BM component regulating M cytokine gene expression.
Intrapulmonary instillation of AQARSAASKVKVSMKF induces PMN infiltration (8). Because TNF- induces PMN migration (53), we examined whether the AQARSAASKVKVSMKF-induced emigration of PMNs was TNF--dependent. Other investigators obtained varying results in studies of the role of TNF- in PMN emigration in response to lung injury or infection (54, 55, 56, 57). We found that the number of neutrophils in the BAL fluid of mice deficient in both TNFRs were similar to that of wild-type mice at 1 day postinstillation of AQARSAASKVKVSMKF (data not shown). In addition, PMN recruitment into the peritoneal cavity in response to the laminin 5 peptide was the same in both wild-type and double TNFR-deficient mice (Fig. 8B). Therefore, the emigration of PMNs in response to the laminin 5 peptide does not require TNFR-triggered signaling. Combined with the microarray data presented in Table I, these data suggest that in the absence of TNF- signaling, other cytokines induced by AQARSAASKVKVSMKF are sufficient to induce PMN emigration.
Previously, we reported that AQARSAASKVKVSMKF can induce MMP-9 expression by Ms (8). Here we found that AQARSAASKVKVSMKF also induces expression of TNF- by Ms. Because TNF- can induce MMP-9 via activation of NF-B (50), we tested the hypothesis that AQARSAASKVKVSMKF activates M production of MMP-9 via TNF--dependent signaling. Our results show that AQARSAASKVKVSMKF increases production of both TNF- and TNFR-II before production of MMP-9 by Ms and that TNF- signaling, specifically through TNFR-II, is required for MMP-9 production (Fig. 7). Our previous studies also showed that AQARSAASKVKVSMKF induces MMP-9 release by PMNs (8). PMNs express surface TNFRs and respond to TNF-. However, MMP-9 release by PMNs does not require new protein synthesis (52), suggesting that AQARSAASKVKVSMKF-induced release of MMP-9 by PMNs is not dependent on TNF--induced NF-B activation. This hypothesis is supported by the presence of both PMNs and MMP-9 in the peritoneal lavage fluid of mice deficient in both TNFRs following i.p. administration of the laminin 5 peptide (Fig. 8). Whether PMNs have a different receptor for the laminin 5 peptide than Ms or whether the signaling mechanism for MMP-9 production is different between the two cell types is currently being investigated.
The first line of defense against infections is the epithelial cell layer itself. Tissue injury may result in the loss of this epithelial layer, exposing the underlying BM. Unlike the interstitial matrix which provides cell support, tensile strength, and elasticity, the BM also serves as a selective and protective barrier. Signals exposed in BM components during times of loss of BM integrity could be recognized by cells of the innate immune system as a breach in host defense. These signals would recruit leukocytes, including PMNs and monocytes, and coordinate their actions to prevent infection. Our studies have shown that a synthetic peptide derived from the -chain of laminin-10 (511), a prominent BM component, is chemotactic for both PMNs and Ms (8) and can induce M cytokine production. We further demonstrate that the corresponding region from the -chain of laminin-1 (111), but not of laminin-5 (332), can also induce TNF-. Interestingly, the elastin-derived peptide, VGVAPG, has chemotactic activity (47), but failed to increase TNF- production by Ms, which indicates that the mechanism of recruitment of inflammatory cells by VGVAPG is different from that of AQARSAASKVKVSMKF, and suggests specificity in signals triggered by ECM components. Fragments or peptides derived from other ECM components have also been shown to be chemotactic (58, 59, 60, 61, 62), but whether these peptides have the capability to induce inflammatory cell cytokine production remains to be investigated.
Hallmarks of chronic inflammation include the accumulation of inflammatory cells and increased degradation of ECM components. During tissue remodeling, increased turnover of ECM can result in release of degradation products of ECM components. The presence of laminin or laminin fragments has been demonstrated in several chronic inflammatory diseases, including chronic hepatitis (17, 18), diffuse interstitial lung diseases (19), renal diseases (16), and inflammatory bowel diseases (21). TNF- has also been shown to be increased in many of these chronic inflammatory diseases (63, 64, 65, 66). Although TNF- has a central role in the innate immune system for protection against infections, when produced in excess, TNF- may also orchestrate a chronic inflammatory response that leads to severe tissue damage. Thus, the regulation of inflammatory cell functions by interactions with BM components may provide important information in determining how inflammatory responses are perpetuated or resolved.
The biological activities of TNF- are mediated by two structurally related, but functionally distinct receptors, TNFR-I and TNFR-II. The majority of inflammatory responses classically attributed to TNF- are mediated by TNFR-I. Studies using TNFR-deficient mice reveal that TNFR-I plays a critical role in mediating endotoxic shock, and LPS- and TNF--induced cytotoxicity. However, there is increasing evidence for an important independent role of TNFR-II signaling in chronic inflammatory conditions. For example, an increase in TNFR-II has been demonstrated to aggravate experimental colitis and contributes to the chronicity of inflammation in vivo (67). Our data indicates that the AQARSAASKVKVSMKF-induced expression of MMP-9 by Ms is predominantly mediated through TNFR-II. Whether inflammatory activities of AQARSAASKVKVSMKF affect tissue remodeling and repair or contribute to the pathophysiology of inflammatory diseases through TNFR-II signaling requires further investigation.
In conclusion, the laminin 5 peptide AQARSAASKVKVSMKF is a clear example of a cryptic matrix epitope that has biological activity. AQARSAASKVKVSMKF induces diverse inflammatory cell activities, including chemotaxis and expression of proteinases, cytokines, and cytokine receptors. These data suggest that exposure of cryptic BM motifs at sites of injury may be an important mechanism for inducing expression of cytokines, expression and release of proteases, and recruitment of inflammatory cell. Together, these processes would serve to initiate repair and remodeling of the matrix and restoration of tissue integrity.
Acknowledgments
We thank Mark Watson for assistance in the analysis of the Affymetrix GeneChips and Robert Mecham and Thomas Broekelmann for providing the laminin-derived peptides.
Footnotes
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 National Institutes of Health, National Heart, Lung and, Blood Institute Grants IF32 HL071423-01 (to T.L.A.-K.), HL29594 and HL47328, the Alan A. and Edith L. Wolff Charitable Trust (to R.M.S.), and a Research Fellowship from the American Lung Association (to J.J.A.).
2 Address correspondence and reprint requests to Dr. Robert M. Senior, Department of Medicine, Washington University School of Medicine, 902 Yalem, Box 8052, 660 South Euclid Avenue, St. Louis, MO 63110. E-mail address: rsenior@im.wustl.edu
3 Abbreviations used in this paper: BM, basement membrane; ECM, extracellular matrix; MMP, matrix metalloproteinase; PMN, neutrophil; M, macrophage; TACE, TNF- converting enzyme; BAL, bronchoalveolar lavage; MSFM, M serum-free medium.
Received for publication May 13, 2004. Accepted for publication November 2, 2004.
References
Tani, T., V. P. Lehto, I. Virtanen. 1999. Expression of laminins 1 and 10 in carcinoma cells and comparison of their roles in cell adhesion. Exp. Cell Res. 248:115.
Ignatius, M. J., T. H. Large, M. Houde, J. W. Tawil, A. Barton, F. Esch, S. Carbonetto, L. F. Reichardt. 1990. Molecular cloning of the rat integrin 1-subunit: a receptor for laminin and collagen. J. Cell Biol. 111:709.
Lee, E. C., M. M. Lotz, G. D. Steele, Jr, A. M. Mercurio. 1992. The integrin 64 is a laminin receptor. J. Cell Biol. 117:671.
von der Mark, H., I. Williams, O. Wendler, L. Sorokin, K. von der Mark, E. Poschl. 2002. Alternative splice variants of 71 integrin selectively recognize different laminin isoforms. J. Biol. Chem. 277:6012.
Corcoran, M. L., M. C. Kibbey, H. K. Kleinman, L. M. Wahl. 1995. Laminin SIKVAV peptide induction of monocyte/macrophage prostaglandin E2 and matrix metalloproteinases. J. Biol. Chem. 270:10365.
Khan, K. M., D. J. Falcone. 1997. Role of laminin in matrix induction of macrophage urokinase-type plasminogen activator and 92-kDa metalloproteinase expression. J. Biol. Chem. 272:8270.
Kanemoto, T., R. Reich, L. Royce, D. Greatorex, S. H. Adler, N. Shiraishi, G. R. Martin, Y. Yamada, H. K. Kleinman. 1990. Identification of an amino acid sequence from the laminin A chain that stimulates metastasis and collagenase IV production. Proc. Natl. Acad. Sci. USA 87:2279.
Adair-Kirk, T. L., J. J. Atkinson, T. J. Broekelmann, M. Doi, K. Tryggvason, J. H. Miner, R. P. Mecham, R. M. Senior. 2003. A site on laminin 5, AQARSAASKVKVSMKF, induces inflammatory cell production of matrix metalloproteinase-9 and chemotaxis. J. Immunol. 171:398.
Jacob, S. S., P. Shastry, P. R. Sudhakaran. 2002. Monocyte-macrophage differentiation in vitro: modulation by extracellular matrix protein substratum. Mol. Cell. Biochem. 233:9.
Xie, B., A. Laouar, E. Huberman. 1998. Fibronectin-mediated cell adhesion is required for induction of 92-kDa type IV collagenase/gelatinase (MMP-9) gene expression during macrophage differentiation. The signaling role of protein kinase C-. J. Biol. Chem. 273:11576.
Beezhold, D. H., C. Personius. 1992. Fibronectin fragments stimulate tumor necrosis factor secretion by human monocytes. J. Leukocyte Biol. 51:59.
McKee, C. M., M. B. Penno, M. Cowman, M. D. Burdick, R. M. Strieter, C. Bao, P. W. Noble. 1996. Hyaluronan (HA) fragments induce chemokine gene expression in alveolar macrophages: the role of HA size and CD44. J. Clin. Invest. 98:2403.
Horton, M. R., C. M. McKee, C. Bao, F. Liao, J. M. Farber, J. Hodge-DuFour, E. Pure, B. L. Oliver, T. M. Wright, P. W. Noble. 1998. Hyaluronan fragments synergize with interferon- to induce the C-X-C chemokines mig and interferon-inducible protein-10 in mouse macrophages. J. Biol. Chem. 273:35088.
Horton, M. R., M. D. Burdick, R. M. Strieter, C. Bao, P. W. Noble. 1998. Regulation of hyaluronan-induced chemokine gene expression by IL-10 and IFN- in mouse macrophages. J. Immunol. 160:3023.
Walker, P. D., G. P. Kaushal, S. V. Shah. 1998. Meprin A, the major matrix degrading enzyme in renal tubules, produces a novel nidogen fragment in vitro and in vivo. Kidney Int. 53:1673.
Horikoshi, S., H. Koide. 1991. Serum laminin P1 fragment concentration in renal diseases. Clin. Chim. Acta 196:185.
Walsh, K. M., A. Fletcher, R. N. MacSween, A. J. Morris. 2000. Basement membrane peptides as markers of liver disease in chronic hepatitis C. J Hepatol. 32:325.
Gabrielli, G. B., F. Capra, M. Casaril, R. Corrocher, R. Colombari, G. De Sandre. 1996. Serum laminin P1 in chronic viral hepatitis: correlations with liver histological activity and diagnostic value. Clin. Chim. Acta 252:171.
Perez-Arellano, J. L., M. J. Pedraz, A. Fuertes, J. L. de la Cruz, J. M. Gonzalez de Buitrago, A. Jimenez. 1993. Laminin fragment P1 is increased in the lower respiratory tract of patients with diffuse interstitial lung diseases. Chest 104:1163.
Horslev-Petersen, K., K. D. Bentsen, P. Halberg, P. Junker, K. I. Kivirikko, K. Majamaa, L. Risteli, J. Risteli, I. Lorenzen. 1988. Connective tissue metabolites in serum as markers of disease activity in patients with rheumatoid arthritis. Clin. Exp. Rheumatol. 6:129.
Koutroubakis, I. E., E. Petinaki, P. Dimoulios, E. Vardas, M. Roussomoustakaki, A. N. Maniatis, E. A. Kouroumalis. 2003. Serum laminin and collagen IV in inflammatory bowel disease. J. Clin. Pathol. 56:817.
Baud, V., M. Karin. 2001. Signal transduction by tumor necrosis factor and its relatives. Trends Cell Biol. 11:372.(Tracy L. Adair-Kirk, Jeff)