Meprin A impairs epithelial barrier function, enhances monocyte migration, and cleaves the tight junction protein occludin

Abstract

Meprin metalloproteases are highly expressed at the luminal interface of the intestine and kidney and in certain leukocytes. Meprins cleave a variety of substrates in vitro, including extracellular matrix proteins, adherens junction proteins, and cytokines, and have been implicated in a number of inflammatory diseases. The linkage between results in vitro and pathogenesis, however, has not been elucidated. The present study aimed to determine whether meprins are determinative factors in disrupting the barrier function of the epithelium. Active meprin A or meprin B applied to Madin-Darby canine kidney (MDCK) cell monolayers increased permeability to fluorescein isothiocyanate-dextran and disrupted immunostaining of the tight junction protein occludin but not claudin-4. Meprin A, but not meprin B, cleaved occludin in MDCK monolayers. Experiments with recombinant occludin demonstrated that meprin A cleaves the protein between Gly¹⁰⁰ and Ser¹⁰¹ on the first extracellular loop. In vivo experiments demonstrated that meprin A infused into the mouse bladder increased the epithelium permeability to sodium fluorescein. Furthermore, monocytes from meprin knockout mice on a C57BL/6 background were less able to migrate through an MDCK monolayer than monocytes from their wild-type counterparts. These results demonstrate the capability of meprin A to disrupt epithelial barriers and implicate occludin as one of the important targets of meprin A that may modulate inflammation.

meprins, members of the “astacin” family of metalloproteases, are highly expressed normally in the rodent intestine and kidney and are localized to the apical (brush-border) membrane of polarized epithelial cells lining in these tissues (9). Meprins have also been detected in macrophages of mesenteric lymph nodes and in leukocytes from lamina propria of human inflammatory bowel tissue and are found at high concentrations in human urine in women with urinary tract infections (13, 53). There are three meprin isoforms, which are composed of α- and/or β-subunits. Heteromeric meprin A (α₂β₂ or α₃β₁) and meprin B (β₂) are membrane bound, whereas homomeric meprin A (α₂), which forms large complexes, is secreted into the lumen of the intestine and urinary tract.

Meprins are capable of cleaving a wide range of substrates in vitro, including extracellular matrix (ECM) proteins such as laminins, collagen, and gelatin, adherens junction proteins (E-cadherin), and cytokines (9, 25, 46). Meprins have been implicated in several inflammatory diseases, for example, meprin-β has been reported to be a candidate gene for diabetic nephropathy in the Pima Indian population (38). In an experimental model of urinary tract infection, meprin A has been shown to enhance renal damage and bladder inflammation after LPS challenge (53). In addition, polymorphisms of the human MEP1A gene have been correlated with inflammatory bowel diseases (5). In an experimental model of ulcerative colitis, meprin-α knockout (KO) mice were more susceptible to injury and inflammation than their wild-type (WT) counterparts (5). Moreover, meprins have also been implicated in cancer invasion and metastasis (30). Previous studies led to the proposal that meprins have an important role in leukocyte transmigration. For instance, Crisman et al. (13) reported that leukocytes lacking meprin-β have been impaired in migration through the ECM. Sun et al. (47) found that meprin-αβ deficiency in mice altered the dissemination of monocytes, with decreased egression from the bone marrow to peripheral blood.

The present study tested the hypothesis that meprins weaken epithelial barrier function by cleaving tight junction (TJ) proteins and thereby facilitate monocyte migration during inflammation. Two meprin isoforms, homomeric meprin A (α₂) and meprin B (β₂), were examined for their effects on epithelial barrier function and the degradation of TJ proteins in Madin-Darby canine kidney (MDCK) cell monolayers and extracts. Epithelial barrier function was determined by the permeability to FITC-dextran and transepithelial electrical resistance (TER). Since TJ proteins, such as occludin, zona occludens (ZO), and claudins, are essential for optimal epithelial barrier function (15, 42), further experiments were conducted to study the degradation of those proteins after meprin treatment by immunocytochemistry and Western blot assays. To relate the in vitro model to in vivo results, meprin A was infused into the mouse bladder, and permeability to sodium fluorescein was measured. Furthermore, monocytes from meprin-α KO mice were compared with those from WT mice to determine whether monocyte migration through a MDCK monolayer was compromised by the lack of meprin A. This work is the first to demonstrate interactions between meprins and the TJ protein occludin and to show that meprin A allows enhanced migration of inflammatory cells through epithelial barriers.

MATERIALS AND METHODS

Materials.

Minimum essential medium (MEM) was purchased from GIBCO. All other chemical reagents were obtained from Fisher or Sigma. Monoclonal mouse anti-occludin, anti-claudin-4, anti-ZO-1 antibodies and polyclonal rabbit anti-occludin antibodies were purchased from Zymed/Invitrogen. Goat anti-mouse Alexa fluor 488 was a gift from Dr. W.B Reeves [Penn State University College of Medicine (PSU-COM)], and goat anti-rat FITC and Hoeschst nuclear stain were purchased from Jackson ImmunoResearch. The EasySep mouse monocyte enrichment kit was purchased from STEMCELL Technologies.

Animal model.

Congenic C57BL/6 meprin-α KO mice and corresponding WT mice were used at 8–9 wk of age for all experiments. All mice were maintained in the PSU-COM Animal Facility and were allowed free access to water and rodent chow. The derivation of mixed-background (C57BL/6 × 129/Sv) meprin-α KO mice has been previously described by Banerjee et al. (5). Congenic meprin-α KO mice were generated by crossing mice on the mixed background with C57BL/6 mice for 10 generations. Mouse tail samples were sent to Charles River Laboratory to assess the level of genetic homogeneity. The results showed 99.07% homogeneity. All animal protocols were approved by the Institutional Animal Care and Use Committee of PSU-COM.

Mice were anesthetized by isoflurane inhalation. For some experiments, ketamine (100 mg/kg) and xylazine (10 mg/kg) were administered intraperitoneally to mice.

Cell culture.

MDCK cells [Accession No. CCL-34, American Type Culture Collection (ATCC)] were grown in MEM (GIBCO) and supplemented with Earl's salts, l-glutamine, sodium bicarbonate, and 10% FBS (Atlas Biologicals) at 37°C with 5% CO₂ in air.

A confluent monolayer was defined as a group of cells that cover the surface of the culture (e.g., plate) completely but without overlaying each other.

Meprin purification, activation, and assays.

Recombinant homomeric mouse meprin A and rat meprin B were purified from stably transformed human embryonic kidney (HEK)-293 cells (Accession No. CRL 1573, ATCC) (11). Latent forms of meprins were activated with trypsin at a molar ratio of 1:6 (trypsin-meprin) for 30–60 min at 37°C. Trypsin was removed by filtration through a Sephadex G-25 column. Activities of meprin A and B were determined using the fluorogenic substrates BK+ [Abz-Arg-Pro-Pro-Gly-Phe/Ser-Pro-Phe-Arg-Lys(Dnp)-Gly-OH] and OCK+ [Abz-Met-Gly-Trp-Met/Asp-Glu-Ile-Asp-Lys(Dnp)-Ser-Gly-OH], respectively (8).

Recombinant occludin expression and purification.

Recombinant human occludin (residues 48–290) was cloned into a maltose-binding protein (MBP)-fusion plasmid, pMTTH (50), with HindIII and XhoI restriction endonucleases sites. The recombinant plasmid was introduced into Esherichia coli BL21 (DE3) RIPL competent cells. Cells were grown in 5 ml LB medium supplemented with ampicillin (100 μg/ml) at 37°C overnight and then inoculated in 400 ml M9 medium in the presence of antibiotics. Cells were induced with 0.8 mM isopropyl thiogalactoside (IPTG) at an optical density value at 600 nm of 0.9. After induction, cells were grown at 16°C overnight. Cells were harvested, resuspended in buffer containing 70 mM Tris and 300 mM NaCl (pH 8.0), and lysed by sonication. Cell debris was removed by low-speed centrifugation. The supernatant fluid was then subjected to ultracentrifugation at 100,000 g at 4°C for 2 h. The membrane pellet was extracted with 1% dodecylphosphocholine (DPC; Anatrace) for 2 h at room temperature. The soluble fraction was then loaded onto a Ni-nitriloacetic acid (NTA) column (Qiagen) preequilibrated with a solution of 20 mM Tris·HCl (pH 8.0), 200 mM NaCl, and 0.2% DPC. The column was washed with 40 ml wash buffer containing 20 mM Tris (pH 8.0), 30 mM imidazole, and 0.2% DPC. The column was eluted with an elution buffer containing 20 mM Tris (pH 8.0), 200 mM NaCl, 250 mM imidazole, and 0.5% DPC. The eluted protein was dialyzed against buffer containing 70 mM NaCl, 1.35 mM KCl, 5 mM Na₂HPO₄, 0.9 mM KH₂PO₄ (pH 7.3), and 1.3 mM DPC at 4°C overnight. NH₂-terminal fusion MBP was cleaved by thrombin (no. 27-0846-01, lot no. 049k7540-11, GE). After digestion, the solution was reloaded onto a Ni-NTA column and washed with wash buffer. The eluted protein was concentrated and further purified with a Superdex-200 (16/60) column.

MBP-conjugated occludin loop 1 and loop 2.

Cloned DNA sequences expressing human occludin extracellular loop 1 or loop 2 were conjugated, respectively, into pMTTH (50) using KpnI and EcoRI endonucleases sites. The ligated vectors were transformed into E. coli DH5α strain, and the transformants were selected in the presence of ampicillin on LB agar plates. The clones thus obtained were sequenced in the PSU-COM Core Sequencing Facility to confirm the DNA sequence of inserted extracellular loops. Loops containing pMTTH vectors were transformed into E. coli BL21(DE3) competent cells, and the transformants were selected in the presence of ampicillin. The transformants were inoculated into LB broth with suitable antibiotics and grown to an optical density of 0.6–0.8 at 600 nm, at which point the expression of protein was induced by the addition of isopropyl β-d-1-thiogalactopyranoside to a final concentration of 1 mM. The culture was collected after 24 h of incubation at 16°C. For large-scale purification of the protein, the culture was centrifuged at 10,000 rpm for 25 min at 4°C. Purification of the histidine-tagged fusion protein was carried out using the Ni-NTA column.

Preparation of membrane-enriched fractions of MDCK cells.

Membrane-enriched fractions of MDCK cells were prepared according to a modified protocol of the Dr. David Antonetti laboratory (PSU-COM). Briefly, MDCK cells were washed with ice-cold PBS and scraped into 500 μl buffer A [0.5 M sucrose, 2 M Tris, 0.5 M EDTA, protease inhibitor tablet (41), benzamidine (1 mM), 10 mM Na₃PO₄, 0.5 M NaF, and 0.5 M Na-pyrophosphate]. EDTA (0.5 M) was added to buffers A and B for the preparation of membrane-enriched fractions of MDCK cells except for those experiments that included meprin treatments to avoid the inhibition of protease activity in these experiments. The suspension was homogenized with a Dounce homogenizer (6 strokes with a pestle). The preparation was centrifuged at 39,000 rpm for 20 min at 4°C. The sediment was suspended in 2 ml buffer B (2 M Tris and 0.5 M EDTA), homogenized again with a pestle, and centrifuged at 39,000 rpm for 20 min at 4°C. The sediment was suspended in 500 μl Stuart's lysis buffer (with a protease inhibitor tablet) and sonicated for 30 s for three times at 50% pulse (22).

Detection of cleaved product of occludin or claudin-4 by meprin A or meprin B.

MDCK cells were cultured until they reached confluence. Meprin A or B was added to cells or to membrane-enriched fractions of cells for various amounts of time. Reactions were stopped by the addition of sample buffer [60 mM Tris (pH 6.8), 25% glycerol, 2% SDS, 0.1% bromophenol blue, and 14.4 mM 2-mercaptoethanol] and an incubation in boiling water for 5 min. Samples were then subjected to SDS-PAGE. Monoclonal mouse anti-occludin or anti-claudin-4 antibodies or polyclonal rabbit anti-occludin antibodies were used to detect potential cleavage products by meprins using Western blot analyses.

Determination of occludin cleavage site(s).

To determine the cleavage site(s) on occludin cleaved by meprin A, the products of digestion of extracellular loop 1 or loop 2 were excised from SDS-PAGE gels. Samples were digested with trypsin and subjected to analysis by C18 nanoflow followed by tandem mass spectroscopy (MS/MS) analysis. MS analysis was performed in the Core Facility of PSU-COM and Proteomics and Mass Spectrometry Core Facility of Cornell University. The amino acid sequence of each fragment was identified by further subjecting it to collision-induced dissociation followed by MS/MS. The masses of ions thus formed were determined and compared with the theoretical masses of ions predicted.

Immunocytochemistry and confocal microscopy.

MDCK cells were grown in 12-well plates on round glass coverslips until they reached confluence. Cells were treated with 4 μg/ml (47 nM) active or latent recombinant meprin A or B for 5 h. MDCK cells were then fixed in 1% paraformaldehyde, permeabilized with 0.2% Triton X-100, and blocked in 10% goat serum with 0.1% Triton X-100 with shaking. Cells were incubated with monoclonal mouse anti-occludin (1:50 dilution), anti-ZO-1 (1:4 dilution), and anti-claudin-4 (1:150) antibodies. After five washes in 0.1% Triton X-100, monolayers were incubated with fluorescence-labeled secondary antibodies for 1 h and rinsed as indicated above. Coverslips were mounted on slides and examined using a confocal microscope.

Permeability assay and TER.

MDCK cells were cultured on filters with 0.4-μm pores (Millipore) until they reached confluence. MEM was changed to serum-free MEM before treatment with meprin. Active or latent meprins (meprin A or B; 47 nM) were added to the apical chamber of insets and incubated at 37°C for the desired amount of time.

For permeability assays, the apical medium was replaced with fresh medium containing 10 or 15 μg/ml FITC-dextran (10 kDa). After a 2-h incubation at 37°C, basal medium was collected, and the fluorescence of permeated FITC-dextran was measured with a fluorescence spectrophotometer (F-2000, Hitachi) at an excitation wavelength of 492 nm and an emission wavelength of 520 nm (45).

For TER measurements, the degree of resistance of the TJ to ions was assessed using voltohmmeter EVOM with a STX2 electrode (World Precision Instruments) (21).

Assessment of bladder permeability.

Transurethral catheterization was performed to instill exogenous reagents, such as meprins and sodium fluorescein, into the bladder. Homomeric mouse meprin A was instilled into the bladder via transurethral catheterization using a 0.5-mm polyethylene catheter (Intramedic PE-10) attached to the hub of a 50-μl Hamilton no. 705 syringe with a 30-gauge blunt-tipped needle. After 2 h, bladder permeability was determined as previously described (18, 53). Briefly, 100 μl of 20 mg/ml sodium fluorescein were instilled into the bladder via catheterization. After 15 min, blood samples were collected from the interior vena cava, and plasma fluorescein concentrations were measured with a fluorescence spectrophotometer (F-2000, Hitachi) using a standard curve of 0.1–100 μg/ml (excitation: 494 nm, emission: 516 nm).

Isolation of monocytes from the mouse bone marrow.

Monocytes from WT and meprin-α KO mice were obtained by negative selection from the bone marrow using a mouse monocyte enrichment kit (no. 19761, StemCell Technologies) following the manufacturer's instructions. Briefly, bone marrow cells were flushed from femurs and tibias with PBS. Mouse enrichment cocktail was added to the single cell suspension, mixed well, and incubated for 15 min. Biotin selection cocktail was added, mixed well, and incubated for 15 min. EasySep D magnetic particles were added, and the tube was then placed into the magnet and set aside for 5 min. The magnet and tube were inverted in one continuous motion to collect the desired fraction to a new tube. The magnetically labeled unwanted cells remained bound inside the original tube (52).

Transmigration of monocytes through MDCK monolayers.

MDCK cells were cultured on Transwell permeable supports (8-μm polyester membrane, Costar) until they reached confluence. Isolated monocytes were added to the top of the MDCK monolayer, and monocyte chemotactic protein (MCP)-1 (1.5 nM) was added to the bottom chamber. After an incubation at 37°C for 3 h, the underside of the supports was flushed with trypsin, and cells in the bottom chamber were collected by centrifugation at 1,500 rpm for 10 min. Suspended cells were labeled with fluorescence antibodies. CD11b⁺/NK1.1⁻/Ly6g⁻/Ly6c⁺ monocyte subsets (inflammatory monocytes) were detected with a flowcytometer (16).

Statistical analysis.

Microsoft Excel and PRISM GraphPad software were used to plot data and for data analysis. Results are expressed as means ± SE. P values of <0.05 were considered significantly different.

RESULTS

Meprin A and B impaired the barrier function of MDCK monolayers.

The permeability of MDCK monolayers to FITC-dextran and resistance to ionic flux (TER) were measured to evaluate the epithelial barrier function in response to meprin A or B challenge. The concentration of 47 nM was chosen on the basis of preliminary experiments and the fact that previous work had established that the concentration of meprin A monomers in mouse urine is in the range of 20–120 nM (Ref. 6 and PSU thesis of R. Yura, 2008). After the exposure to active meprin A for 5 h, the permeability of MDCK monolayers to 40-kDa FITC-dextran showed a tendency to increase, but the difference was not significant (Fig. 1A). After the exposure to active meprin A for 5 h or longer, the permeability of MDCK monolayers to 10-kDa FITC-dextran significantly increased compared with cells incubated with medium only or those treated with latent meprin A (Fig. 1B). Exposure to active meprin B for 9 h significantly increased the permeability of MDCK cell monolayers to 10-kDa FITC-dextran compared with untreated samples or those treated with latent meprin B (Fig. 1C).

Fig. 1.

Mabin-Darby canine kidney (MDCK) monolayer permeability to FITC-labeled dextran. MDCK cell monolayers were cultured on filters. Active or latent meprin A or B (47 nM) were added to the apical chamber of filters and incubated at 37°C for various periods of time. Controls were MDCK cell monolayers incubated with medium only. After meprin treatment, apical medium was replaced with FITC-dextran (40 or 10 kDa). After a 2-h incubation at 37°C, the basal medium was collected, and fluorescence was measured by a spectrophotometer at an excitation wavelength of 492 nm and emission wavelength of of 520 nm. A: there was a tendency for active meprin A to increase permeability to 40-kDa FITC-dextran after 5 h of treatment, but the increase was not significant. B: active meprin A significantly increased permeability to 10-kDa FITC-dextran after 5 h of treatment compared with medium controls. Active meprin A also significantly increased permeability to FITC-dextran after 7 and 9 h of treatment compared with medium controls and latent meprin A. *P < 0.05; **P < 0.01. C: active mepirn B significantly increased permeability to 10-kDa FITC-dextran after 9 h of treatment compared with medium controls and latent meprin B. *P < 0.05.

The resistance to ionic flux (TER) was also measured to determine the epithelial barrier function in response to meprin A or B. Active meprin A decreased the electrical resistance of MDCK monolayers slightly after 9 h of incubation compared with cells incubated with medium only or those treated with latent meprin A (Fig. 2A). No statistical significance in resistance to ionic flux was observed between monolayers treated with active meprin B and untreated monolayers or those treated with latent meprin B, even after 9 h of incubation (Fig. 2B). These TER data indicate that meprin A slightly impaired epithelial barrier function of MDCK monolayers, whereas meprin B had no effect.

Fig. 2.

Transepithelial electrical resistance (TER) across MDCK cell monolayers. MDCK cell monolayers were cultured on filters. Active or latent meprin A or B (47 nM) were added to the apical chamber of filters and incubated at 37°C for various periods of time. Controls were MDCK cell monolayers incubated with medium only. TER across the monolayer was measured by voltohmmeter EVOM with a STX2 electrode (World Precision Instruments). A: active meprin A significantly decreased the TER value after 9 h of treatment. The inset shows the same graph with an expanded y-axis. NS, not significant. *P < 0.05. B: active meprin B did not significantly decrease the TER value after treatment. The inset shows the same graph with an expanded y-axis.

Meprin A and B disrupted TJs on MDCK monolayers.

TJs in the epithelium are essential for the maintenance of barrier function; therefore, MDCK monolayers were treated with meprin A or B to assess whether the immune staining of TJs was disrupted. In monolayers treated with active meprin A, immune stainings of occludin (Fig. 3A) and ZO-1 (Fig. 3B) were disrupted. In monolayers treated with active meprin B, the immune staining of ZO-1 (Fig. 3C) was disrupted. In untreated or latent meprin-treated monolayers, immune stainings of tight junctions were continuous at cell borders. The immune staining of claudin-4 was not disrupted by active meprin A treatment (Fig. 3D). These observations indicated that meprin A and B disrupted TJs between MDCK cells and affect certain types of TJ proteins.

Fig. 3.

Immunocytochemistry of tight junctions (TJs) on MDCK cell monolayers. MDCK cells were treated with active or latent meprin A or B (47 nM) for 5 h. Cells were fixed, permeabilized, and then incubated with anti-occludin (A), anti-zonula occludens (ZO)-1 (B and C), and anti-claudin-4 (D) antibodies. Hoeschst staining was used (top rows) to show nuclei. For each image, three to five random fields of monolayers were observed by microscopy and showed similar disruptions of TJs; one representative field was captured and is shown. A: occludin immunostaining was disrupted in monolayers treated with active meprin A (b), whereas it was continuous in untreated monolayers and monolayers treated with latent meprin A (a and c). Magnification in all images: ×20. Scale bar = 50 μm. B: ZO-1 immunostaining was disrupted in monolayers treated with active meprin A (b), whereas it was continuous in untreated monolayers and monolayers treated with latent meprin A (a and c). Magnification in all images: ×20. Scale bar = 50 μm. C: ZO-1 immunostaining was disrupted in monolayers treated with active meprin B (b), whereas it was continuous in untreated monolayers and monolayers treated with latent meprin B (a and c). Magnification in all images: ×40. Scale bar = 50 μm. D: claudin-4 immunostaining was not disrupted in monolayers with active meprin A (b) treatment, similar to untreated monolayers and monolayers treated with latent meprin A (a and c). Magnification in all images: ×63. Scale bar = 50 μm.

Occludin, but not claudin-4, in MDCK membrane fractions was cleaved by meprin A and B.

To determine whether occludin or claudin-4 was cleaved by meprin A or B, membrane-enriched fractions from MDCK cell lysates were incubated with meprin A or B. Preliminary experiments have shown that ZO-1 protein abundance did not change as measured by Western blot analysis after meprin A treatment. The images shown in Fig. 4 are representative images from two to three Western blots of each experiment. The percentage of hydrolysis from measurements of the optical density of occludin bands was averaged from these two blots. Western blot analysis showed the degradation of occludin by both active meprin A and B but not by latent meprins or active meprins with the metalloprotease inhibitor EDTA (Fig. 4, A and B). The percentage of hydrolysis after 1 h of active meprin A or B incubation was 83% and 84%, respectively. The antibody used in our experiments has an epitope on the COOH-terminus of occludin. Computational prediction has revealed multiple cleavage sites for meprin A on the COOH-terminal of occludin. Thus, any cleavage products will likely be further degraded by meprin A and difficult to detect. Western blot analysis did not show degradation of claudin-4 by either meprin A or B (Fig. 4, C and D).

Fig. 4.

Detection of degradation of occludin but not claudin-4 in membrane-enriched fractions of MDCK cells. Membrane-enriched fractions (80 mg) of MDCK cells were incubated with active or latent meprins or EDTA-inhibited meprins for various periods of time. Occludin and claudin-4 were detected by Western blot analyses. Each experiment was repeated two to three times, and a representative example is shown here. Average percentages of band densities over control were calculated and are shown in the bar graphs. Error bars represent SEs in A and B and ranges in C and D. A: membrane fractions were incubated with 20 nM meprin A. Active meprin A degraded occludin. B: membrane fractions were incubated with 10 nM meprin B. Active meprin B degraded occludin. C: membrane fractions were incubated with 47 nM meprin A. Claudin-4 was not degraded by meprin A. D: membrane fractions were incubated with 47 nM meprin B. Claudin-4 was not degraded by meprin B.

Occludin in MDCK monolayers was degraded by meprin A.

To determine whether occludin in MDCK monolayers was cleaved by meprins, MDCK monolayers were directly treated with meprin A or B. The images shown in Fig. 5 are representative of three Western blots of each experiment. The percentage of hydrolysis from measurements of the optical density of occludin bands was averaged from these three blots. Western blot analysis showed that the amount of occludin in MDCK cells treated with active meprin A was decreased (Fig. 5A). The percentage of hydrolysis after treatment with 47 nM active meprin A was 86%. However, meprin B had little effect on occludin in MDCK monolayers (Fig. 5B).

Fig. 5.

Detection of degradation of occludin but not claudin-4 in MDCK monolayers. MDCK cell monolayers were incubated with exogenous active or latent meprin A or meprin B for 5 h. Occludin and claudin-4 were detected by Western blot analyses. Each experiment was repeated two to three times, and representative blots are shown here. Average percentages of band densities over control were calculated and are shown in the bar graphs. Error bars represent SEs for A and B and ranges for C. A: occludin in MDCK monolayers was degraded by active meprin A. B: occludin in MDCK monolayers was not degraded by active meprin B. C: claudin-4 in MDCK monolayers was not degraded by active meprin A.

Claudin-4 in MDCK monolayers was not degraded by meprin A.

To determine whether claudin-4 is cleaved by meprin A, MDCK monolayers were incubated with increasing concentrations of meprins (4.7, 23, and 47 nM) for 5 h. Western blot analysis did not show degradation of claudin-4 by meprin A (Fig. 5C).

Recombinant occludin in micelles was cleaved by meprin A.

To identify the site(s) where meprin A cleaves occludin, a sample of 3 μM recombinant occludin reconstituted in micelles was incubated with 47 nM of active or latent meprin A. The major recombinant occludin band was at 28 kDa (Fig. 6A). Other observed bands might reflect different oligomerization states of occludin or residual E. coli proteins. Based on the size of protein and Western blot analysis, the 28 kDa-band is occludin. Incubation with active meprin A up to 4 h led to a 23% and 48% decrease of the occludin band intensity, as shown by Coomassie blue staining and Western blot analysis, respectively (Fig. 6, A and B). A potential cleavage product (∼22 kDa) was observed with Coomassie blue staining but not with Western blot analysis. The probable reason is that the cleavage product observed by Coomassie blue staining does not contain the histidine tag.

Fig. 6.

Recombinant occludin is cleaved by homomeric meprin A. Recombinant occludin (3 μM) was incubated with active or latent meprin A (47 nM) for various periods of time. Samples were subjected to SDS-PAGE, and occludin was detected by Coomassie blue staining or Western blot assay. A: Coomassie blue stain. Active meprin A treatment for 4 h decreased occludin band intensity (arrow 1). A cleavage product (∼22 kDa) was also observed (arrow 2). B: Western blot analysis. Active meprin A treatment for 2 and 4 h decreased occludin band intensity (arrow).

Occludin extracellular loops were cleaved by meprin A.

To demonstrate cleavage of occludin extracellular loops, 6 μM of MBP-conjugated occludin loop 1 and loop 2 were incubated with meprin A (47 nM) for various periods of time for up to 4 h (Fig. 7A). Cleavage of occludin loop 1 was observed after active meprin A treatment but not after latent meprin A treatment. Similar results were observed in occludin loop 2, for example, the 28-kDa band density decreased after active meprin A treatment but not after latent meprin A treatment (Fig. 7B). To determine whether MBP, a protein of 42 kDa, is cleaved by meprin A, 6 μM MBP was incubated with meprin A for various periods of time for up to 4 h (Fig. 7C). No cleavage was observed after meprin A treatment. These results confirmed that there were cleavage sites of meprin A on the extracellular loops of occludin.

Fig. 7.

Occludin extracellular loops are cleaved by meprin A. maltose-binding protein (MBP)-conjugated loop 1 and loop 2 (6 μM) were incubated, respectively, with active or latent meprin A (47 nM) for the times shown. After the incubation, samples were subjected to 10% SDS-PAGE and stained with Coomassie blue. Each experiment was repeated two times, and a representative gel is shown. A: cleavage of occludin loop 1 was observed after active meprin A treatment (arrow) but not after latent meprin A treatment. B: cleavage of occludin loop 2 was observed after active meprin A treatment (arrow) but not after latent meprin A treatment. C: when MBP itself was treated with meprin, no cleavage was observed after active meprin A treatment up to 4 h (lane 4).

The cleavage site(s) was determined by MS analysis.

The cleavage products of extracellular loop 1 and loop 2 were excised from the SDS-PAGE gel, and samples were trypsin digested and subjected to analysis by C18 nanoflow followed by MS/MS analysis. These data allowed for the identification of consecutive ions and thereby the peptide fragment. The cleavage site identified between Gly¹⁰⁰ and Ser¹⁰¹ is on the first extracellular loop of occludin (Fig. 8). This cleavage site is consistent with a computational prediction using Prediction of Protease Specificity, a computer program designed by Sarah Boyd (Monash University), based on meprin substrate specificities. The computational program predicted that meprin A has several cleavage sites on occludin extracellular loop 1 and loop 2. For example, there were G-Y (92–93), G-S (100–101), and G-Y (103–104) on loop 1, and G-S (100–101) was the most potent predicted site. The predicted fragment sizes of MBP-loop 1 were 4 and 43 kDa, which is the correct size product. On loop 2, the potential cleavage sites included V-N (197–198), G-S (209–210), and V-D (239–240). However, due to the low ion density of the sample, MS analysis was not able to identify the cleavage site on loop 2.

Fig. 8.

Mass spectrometry (MS) analysis of occludin cleavage products. The cleavage products of recombinant extracellular loop 1 and loop 2 were excised from the SDS-PAGE gel, and samples were subjected to MS analysis. A: shaded peptides were identified from loop 1 with >95% confidence. The peptide closest to the COOH-terminus is HHHHHR. The rest of the COOH-terminus has an excess of glycine residues; thus, it was difficult to obtain more accurate identification. B: based on the predicted cleavage site, the precursor ions of the possible COOH-terminal peptides (GYGTSLLGG, GYGTSLLGGS, GYGTSLLGGSV, GYGTSLLGGSVG, GYGTSLLGGSVGY, etc.) were analyzed. Based on the XIC profile, the measured mass [mass-to-charge ratio (m/z): 824.41217] was very close to the theoretical mass m/z (824.4148) of GYGTSLLGG. The mass difference was 3.2 ppm, which indicates that the cleavage site is very likely to be Gly³⁸⁵ (arrow in A). It corresponds to Gly¹⁰⁰ on the first extracellular loop of full-length occludin.

Meprin A increased mouse bladder permeability.

The bladder wall serves as a good in vivo system to study alterations in the integrity of the bladder epithelial lining in response to a meprin A challenge. Preliminary studies in our lab showed that the concentration of soluble meprin A in mouse urine is ∼120 nM (6, 33). In addition, the infiltrating leukocytes that accumulate at inflammation sites also express meprins (13, 47). Thus, meprin A concentration in the bladder during inflammation could be in the high nanomolar or micromolar range. To mimic in vivo conditions under inflammation, a high concentration of meprin A (4.7 μM) was used to challenge the bladder. Bladder permeability was determined by the measurement of sodium fluorescein leakage from the bladder into the serum. Active meprin A increased bladder permeability to sodium fluorescein compared with Tris buffer, actinonin-inactivated meprin A, or actinonin-only treatments (Fig. 9). These data indicate that active meprin A was able to impair the epithelial barrier in vivo.

Fig. 9.

Bladder permeability to sodium fluorescein was increased after exposure to active meprin A. C57BL/6 mice were anesthetized, and active or actinonin-inhibited meprin A (4.7 μM) or actinonin alone was then injected into the bladder via transurethral catheterization. Mice in the control group were instilled with Tris buffer. After 2 h, 100 ml of 20 mg/ml sodium fluorescein was injected into bladder. After 15 min, blood samples were collected from the interior vena cava, and plasma fluorescein concentrations were measured at an excitation wavelength of 494 nm and emission wavelength of 516 nm. Results are an average of two independent experiments and are expressed as means ± SE; n = 3 per group. *P < 0.05.

Meprin A regulated monocyte transmigration.

A goal of this investigation was to test the hypothesis that epithelium integrity compromised by meprin A allows enhanced monocyte transmigration. Accordingly, bone marrow-derived monocytes from WT and meprin-α KO mice were cultured on MDCK monolayers. A group of monocytes was added on the apical side of MDCK monolayers grown on an inserted filter without MCP-1 in the bottom chamber as a chemoattractant, another group of monocytes was added to the apical side of the inserted filter without MDCK cell monolayers but with MCP-1 in the bottom chamber, and the last group of monocytes was added on MDCK monolayers with MCP-1 in the bottom chamber. Transmigration of monocytes was measured using a flowcytometry assay (Fig. 10). The results showed that the migration of monocyte was driven by the chemoattractant, because when no MCP-1 was present, neither genotype of monocytes transmigrated efficiently through MDCK monolayers. More importantly, the results showed that in the presence of both chemokine and MDCK monolayers, there were significantly fewer meprin-α KO monocytes that transmigrated through compared with WT monocytes. Since the genotypic difference depends on the presence of MDCK monlayers, the results support the hypothesis that the effects of meprin A on epithelial barrier lead to enhanced monocyte migration.

Fig. 10.

Meprin A regulates monocyte transmigration. The same number of monocyte cells were isolated from wild-type and meprin-α knockout (KO) mouse bone marrow. Monocytes were added to the apical side of MDCK monolayers grown on inserted filters; the lower chamber contained monocyte chemotactic protein (MCP)-1 (1.5 nM) as a chemoattractant. Control groups included monocytes added to the apical side of the inserted filter without MDCK cell monolayers but with MCP-1 (1.5 nM) in the bottom chamber and a group of monocytes added on MDCK monolayers without MCP-1 in the bottom chamber. After an incubation at 37°C for 3 h, monocytes in the bottom chamber were collected and measured by flowcytometry for the CD11b⁺/NK1.1⁻/Ly6g⁻/Ly6c⁺ subsets, which represent the infiltrating inflammatory monocytes. Significantly fewer meprin-α KO monocytes transmigrated through MDCK monolayers than wild-type monocytes. Results are expressed as means ± SE; n = 3 per genotype per group for two independent experiments. *P < 0.05.

DISCUSSION

In this study, the hypothesis that meprins are determinative factors in disrupting the barrier function of epithelium was tested. The results demonstrated that meprin A impairs epithelial barrier function in vivo, with experiments of bladder permeability to sodium fluorescein, as well as in vitro, with experiments of permeability of an MDCK monolayer to FITC-dextran. A consequence of the activity of meprin A is to enable monocytes to migrate through an epithelial barrier more readily, and in vivo this would allow inflammatory molecules such as cytokines and monocytes to gain access to sites of injury.

Meprin A was more effective than meprin B in impairing MDCK epithelial barrier function. The difference may be due to the fact that while meprin B is able to cleave occludin in cell fractions, it had little proteolytic activity on occludin in MDCK monolayers. One interpretation of these results is that meprin B is unable to gain access to the vulnerable peptide bonds in the paracellular space of occludin, as meprin A does. Meprin A and B have different peptide bond and substrate specificities, and this may explain their differential effects (7, 8). Previous studies have shown that meprin B can cleave E-cadherin at the an extracellular site and weaken the intercellular contacts. The cleavage site was localized in the extracellular domain adjacent to the plasma membrane (25). Thus, a possible explanation for the observed increased permeability of MDCK monolayers to FITC-dextran after meprin B treatment is that meprin B cleaves other junctional proteins, such as E-cadherin.

The meprin A-induced permeability change to FITC-dextran was greater than the TER change to ionic flux. There is evidence that specific TJ proteins play different roles in epithelial barrier functions. Whereas occludin is important for the paracellular transport/diffusion of small molecules, claudins are regulators of cation/anion exchange and charge selectivity through the epithelium (23, 40, 44, 48). Thus, the observed difference in the extent of increased FITC-dextran permeability and decreased TER after treatment with meprin A can be explained by the fact that meprin A cleaves occludin, a key factor, in permeability to FITC-dextran but not claudin-4, which is more critical for ionic flux.

In multicellular organisms, the epithelia and endothelia delineate the borders between different compartments. This demarcation relies on the establishment of cell-cell contacts, such as TJs between adjacent cells to withstand mechanical stress and prevent paracellular flux (19). TJs are important for maintaining the epithelium barrier functions, enabling adjacent cell communications as well as influencing the migration of mobile cells (17, 39). TJs are maintained through a complex network of interacting proteins, such as occludin, ZO, and E-cadherin. Occludin, one of the first TJ proteins to be identified, has a cytoplasmic NH₂-terminus, four transmembrane domains, and two extracellular loops (20). It is well conserved in the human, mouse, rat, and dog (2). Previous studies by others have demonstrated that occludin contributes to barrier function and leukocyte migration. For example, murine epithelial cells with mutant occludin lacking the NH₂-terminus and extracellular domains were unable to form strong TJs (4). Elicited expression of occludin in transformed epithelial cells rescues epithelial morphology and promotes reacquisition of the epithelial phenotype (12, 51). The first loop of occludin is required for TJ resealing resulting from intercellular occludin interactions of adjacent cells (28), whereas the second loop is required for occludin to localize to cell membrane and assemble the junctional complex with other TJ proteins (31). Occludin has also been shown to mediate neutrophil migration across MDCK cell monolayers (24, 36). In addition, occludin has a critical regulatory function in that it interacts with several other junctional proteins, including ZO-1, actin, and other cytoskeleton proteins (32).

Several lines of evidence have shown that downregulation of occludin via either proteolysis or factors such as VEGF results in elevated permeability (14). For example, proteolysis of occludin by metalloproteases led to increased permeability to FITC-dextran in endothelial cells and disruption of immunostaining, while other junctional proteins (ZO-1 and cadherin) were intact (49). Knockdown of occludin by small interfering RNA led to increased permeability to FITC-dextran and decreased TER in epithelial monolayers without affecting ZO-1 or various claudins (1, 36). These findings are consistent with the results of the present study with meprin A.

The results of the present study implicate occludin as one of the important targets of meprin A in the disruption of epithelial barriers. However, the role of occludin in barrier function is controversial, especially in the light of the fact that occludin-null mice do not exhibit deficiencies in barrier function (43). It has been suggested that other proteins (such as tricellulin, a TJ protein localized at tricellular junctions) provide the functional redundancy that perserves barrier function in occludin-null mice (37). In addition, meprin A is capable of cleaving many proteins, and thus the functional effects observed in the present study may be due to other targets.

Previous studies (26, 35, 43) have indicated that TJs regulate leukocyte transmigration. The present study demonstrates that meprin A enhances leukocytes transmigration. We suggest that the cleavage of junctional proteins is the molecular basis for increased meprin-mediated leukocytes transmigration.

Meprin metalloproteases are involved in inflammatory processes of both acute and chronic disease conditions (5, 34). The present findings that meprin A impairs epithelial barrier functions can explain previous observations in mouse models of inflammation. In an acute model of bladder inflammation, the host response to intravesicular LPS challenge of meprin-α KO mice was less severe than their WT counterparts, as it was shown that there was less bladder edema, less leukocyte infiltration, and less bladder permeability in meprin-α KO mice than in WT mice (53). Serum cytokine profiles showed that TNF-α, IL-1β, and MCP-1 levels were significantly lower in meprin-α KO mice than in WT mice after an intraperitonel LPS challenge. Moreover, meprin-β KO and WT mice showed similar hypothermia and similar serum nitrate/nitrite levels. Serum cytokine levels in meprin-β KO mice were not significantly different from their WT counterparts (53). These data indicate that meprin A has a determinative and proinflammatory role in the inflammatory response to an acute urinary infection. The present results are consistent with the LPS findings. Both meprin-β KO and WT mice produce meprin A in the kidneys and migratory monocytes. The meprin A released by the kidney is predominantly inactive. Thus, the source of active meprin A is likely monocytes attracted to the site of injury. Therefore, the breakdown in bladder barriers and elevated leukocyte infiltration can be attributed to the ability of active meprin A to disrupt cellular TJs.

Monocytes and their interactions with endothelial/epithelial cells have been implicated in several kidney disorders, such as diabetic nephropathy and ischemia-reperfusion injury (3, 29). The recruitment of monocytes/macrophages to sites of injury correlates with the progression of kidney damage in diabetic nephropathy. The ability of meprins to affect the movement of monocytes from the bone marrow to peripheral sites as well as to enhance transmigration through epithelial barriers may be factors in the progression of kidney damage. The high concentration of meprin A that is normally present in apical membranes of proximal tubules of rodent kidneys is another factor that may result in kidney damage in response to an acute injury. Studies (10, 34) of kidney ischemia-reperfusion in mice have shown that the redistribution of meprin A to the cytosol and other cellular compartments in response to an insult such as ischemia-reperfusion results in renal injury and inflammation. Another study (27) in rats has shown that meprin-deficient mice are less vulnerable to kidney damage in response to acute kidney injury and that actinonin, an excellent inhibitor of meprin A, prevents renal pathology in acute kidney injury. Taken together, these studies indicate that blockade of meprin activity could be a promising therapeutic approach for treatment of kidney injuries.

GRANTS

This work was supported by National Institutes of Health Grants DK-19691 and R01GM094526 and the Carlino Gift Fund of Penn State Hershey.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: J.B., R.E.Y., G.L.M., S.G.B., and J.S.B. conception and design of research; J.B., G.L.M., P.S., and F.T. performed experiments; J.B., S.G.B., and J.S.B. analyzed data; J.B., G.L.M., S.G.B., and J.S.B. interpreted results of experiments; J.B. prepared figures; J.B. drafted manuscript; J.B., R.E.Y., G.L.M., S.G.B., P.S., F.T., and J.S.B. edited and revised manuscript; J.B., R.E.Y., G.L.M., S.G.B., P.S., F.T., and J.S.B. approved final version of manuscript.