Expression and function of matrix metalloproteinase (MMP)-28

Abstract

Matrix metalloproteinase-28 (MMP-28, epilysin) is highly expressed in the skin by keratinocytes, the developing and regenerating nervous system and a number of other normal human tissues. In epithelial cells, over-expression of MMP-28 mediates irreversible epithelial to mesenchymal transition concomitant with loss of E-cadherin from the cell surface and an increase in active transforming growth factor beta. We recently reported the expression of MMP-28 in both cartilage and synovium where expression is increased in patients with osteoarthritis. In human chondrosarcoma cells MMP-28 was activated by proprotein convertases and the active form of the enzyme preferentially associated with the extracellular matrix in a C-terminal independent manner. over-expression of MMP-28 in chondrosarcoma cells led to altered cell morphology with increased organisation of actin. Adhesion to type II collagen and fibronectin was increased, and migration across the former was decreased. MMP-28 was localised to the cell surface, at least transiently, in a C-terminal dependent manner. Heparin prevented both extracellular matrix association and cell surface binding of MMP-28 suggesting that both are via heparan sulphate proteoglycans. Over-expression of activatable MMP-28, but not catalytically inactive EA mutant increased the expression and activity of MMP-2, and all forms of MMP-28 tested increased expression of MMP19 and TIMP3 mRNA. These data demonstrate that expression of MMP28 alters cell phenotype towards a more adhesive, less migratory behaviour. Further, MMP-28 activity may reside predominantly in the extracellular matrix, and we are currently searching for substrates in this compartment.

Fig. 1

Expression of MMP-28 in transiently transfected HeLa cells. HeLa cells were transiently transfected with the wild-type (WT) MMP28 cDNA, previously cloned into a modified pcDNA4FLAG vector. Vector-only (VO) transfects were included as a negative control. Protein expression in conditioned medium (A), cell lysate (B) and extracellular matrix (C) was detected by western blotting, using an anti-FLAG antibody. The pro and active forms of MMP-28 are indicated. CTD = C-terminal domain.

Fig. 2

Furin activity is required for processing of proMMP-28. HeLa cells were transiently transfected with the wild-type MMP28 gene or the pcDNA4FLAG vector as a negative control. Cells were treated with 100 μM of the furin inhibitor Decanoyl-Arg-Val-Lys-Arg-chloromethyl ketone prior to transfection. The same dose was repeated 24 h after transfection. Untreated cells were included as controls. Protein expression in the extracellular matrix was detected by western blotting, using an anti-FLAG antibody.

Fig. 3

Expression of MMP-28 in stably transfected SW1353 cells. SW1353 cells were stably transfected with wild-type (A), EA mutant (B) or pro-cat mutant (C) MMP28 gene constructs. Cells stably transfected with pcDNA4FLAG vector only were included as a negative control (D). In a separate experiment, WT cells were treated with 0.1 mg/ml heparin, or left untreated as a negative control (E). Protein expression in conditioned medium (CM), cell lysate (CL) and extracellular matrix (ECM) was detected by western blotting, using an anti-FLAG antibody. The pro and active species are indicated for the WT and pro-cat forms of MMP-28. The pro and processed species are indicated for the catalytically inactive EA mutant. CTD = C-terminal domain.

Fig. 4

Detection of MMP-28 expression in SW1353 cells using immunofluorescence. SW1353 cells stably transfected with MMP28 constructs were probed for protein expression with an anti-FLAG primary antibody and an Alexa 488-conjugated secondary antibody. Cells were permeabilised (A to D) or non-permeabilised (E to H) to detect cell surface localisation of MMP-28. Wild-type cells treated with an IgG isotype control primary antibody were also included as a control, both permeabilised (I) and non-permeabilised (J). Cells were counterstained with DAPI and viewed under 20× magnification.

Fig. 5

Heparin competes for cell surface-associated MMP-28. SW1353 cells stably transfected with the wild-type MMP28 expression construct were probed, non-permeabilised, for protein expression with an anti-FLAG primary antibody and an Alexa 488-conjugated secondary antibody. Cells were either untreated (A, D) or treated with 0.1 mg/ml heparin (B, E). Wild-type cells treated with an IgG isotype control primary antibody were also included as a control (C). Cells were counterstained with DAPI and viewed under 20× magnification.

Fig. 6

Effect of over-expression of MMP-28 on SW1353 morphology. Phase contrast images of live SW1353 cells stably transfected with MMP28 cDNA constructs were obtained when the cells were at low confluence, 24 h after plating (A–D). Fixed cells were stained with phalloidin (E–H). Cells were viewed under 20× magnification.

Fig. 7

Effect of over-expression of MMP-28 on adhesion and migration of SW1353 cells. For adhesion experiments, two clones each of cells stably transfected with MMP28 cDNA constructs (wild-type (WT), EA mutant (EA), pro-cat, (PC)) or vector-only control (VO) were plated in wells of a 96-well plate coated with either 100 µl/well of 50 µg/ml fibronectin (A) or 100 µg/ml of collagen II (B). To study the role of the integrin β₁ subunit in mediating adhesion, cells were treated with anti-β₁ antibody for 30 min prior to plating on fibronectin (C) or collagen II (D). Cells were allowed to adhere for 1 h, followed by washing, fixing and staining with methylene blue. Absorbance of lysed cells was measured at 630 nm. Data were corrected for blank readings and control adhesion to 1% BSA, and are presented as fold-change relative to the mean of the empty vector controls. For migration experiments, stably transfected cells were plated in DMEM/10% serum in wells of a 24-well plate coated with 100 μg/ml collagen II (E). Media was switched to DMEM/0.5% serum after 24 h, and cell migration measured at 15 minute intervals over a 13 hour period. Data are expressed as μm per hour. Values for adhesion and migration experiments were compared to mean values for the empty vector control. Experiments were performed in triplicate, data are plotted as mean +/− s.e.m. Statistical significance indicated by: ⁎⁎⁎⁎, p < 0.0001; ⁎⁎⁎, p < 0.001: ⁎⁎, p < 0.01 and ⁎, p < 0.05.

Fig. 8

Effect of over-expression of MMP28 on expression of other MMP and TIMP family members. Expression of all of the members of the MMP and TIMP family, including MMP28, was measured using qRT-PCR and normalised to 18S rRNA in two clones each of cells stably transfected with MMP28 cDNA constructs (vector only (VO), wild-type (WT), EA mutant (EA) and pro-cat (PC)). Expression of each of the MMP28 constructs was verified (A) with MMP28 undetectable in the vector-only transfects. The expression of three genes altered in response to MMP28 over-expression: MMP2 (B), MMP19 (C) and TIMP3 (D). Data are plotted as mean +/− s.e.m. (n = 6). Statistical significance indicated by: ⁎⁎⁎⁎, p < 0.0001; ⁎⁎⁎, p < 0.001: ⁎⁎, p < 0.01 and ⁎, p < 0.05.

Fig. 9

Analysis of MMP-2 activity. Activity of MMP-2 was analyzed by gelatin zymography. Two clones each of cells stably transfected with empty vector or MMP28 cDNA constructs were plated in 6-well dishes. Medium was exchanged to serum-free 24 h after plating. Conditioned medium was collected for a further 48 h later and concentrated. Protein concentration was determined and equal amounts of protein for each sample were loaded (vector only (VO), wild-type (WT), EA mutant (EA) and pro-cat (PC)).