TNF-alpha stimulates activation of pro-MMP2 in human skin through NF-(kappa)B mediated induction of MT1-MMP

Abstract

Tumor necrosis factor-alpha (TNF-(alpha)) is an important mediator during the inflammatory phase of wound healing. Excessive amounts of pro-inflammatory cytokines such as TNF-(alpha) are associated with inflammatory diseases including chronic wounds. Matrix metalloproteinases (MMPs) are involved in matrix re-modeling during wound healing, angiogenesis and tumor metastasis. As with pro-inflammatory cytokines, high levels of MMPs have been found in inflammatory states such as chronic wounds. In this report we relate these two phenomena. TNF-(alpha) stimulates secretion of active MMP-2, a type IV collagenase, in organ-cultured full-thickness human skin. This suggests a mechanism whereby excess inflammation affects normal wound healing. To investigate this observation at the cellular and molecular levels, we examined TNF-(alpha) mediated activation of pro-MMP-2, induction of MT1-MMP, and the intracellular signaling pathways that regulate the proteinase in isolated human dermal fibroblasts. We found that TNF-(alpha) substantially promoted activation of pro-MMP-2 in dermal fibroblasts embedded in type-I collagen. In marked contrast, collagen or TNF-(alpha) individually had little influence on the fibroblast-mediated pro-MMP-2 activation. One well-characterized mechanism for pro-MMP-2 activation is through a membrane type matrix metalloproteinase, such as MT1-MMP. We report that TNF-(alpha) significantly induced MT1-MMP at the mRNA and protein levels when the dermal fibroblasts were grown in collagen. Although the intracellular signaling pathway regulating mt1-mmp gene expression is still obscure, both TNF-(alpha) and collagen activate the NF-(kappa)B pathway. In this report we provide three sets of evidence to support a hypothesis that activation of NF-(kappa)B is essential to induce MT1-MMP expression in fibroblasts after TNF-(alpha) exposure. First, SN50, a peptide inhibitor for NF-(kappa)B nuclear translocation, simultaneously blocked the TNF-(alpha) and collagen mediated MT1-MMP induction and pro-MMP-2 activation. Secondly, TNF-(alpha) induced I(kappa)B to breakdown in fibroblasts within the collagen lattice, a critical step leading to NF-(kappa)B activation. Lastly, a consensus binding site for p65 NF-(kappa)B (TGGAGCTTCC) was found in the 5'-flanking region of human mt1-mmp gene. Based on these results and previous reports, we propose a model to explain TNF-(alpha) activation of MMP-2 in human skin. Activation of NF(kappa)B signaling in fibroblasts embedded in collagen induces mt1-mmp gene expression, which subsequently activates the pro-MMP-2. The findings provide a specific mechanism whereby TNF-(alpha) may affect matrix remodeling during wound healing and other physiological and pathological processes.

Fig. 1

TNF-α stimulates secretion of active MMP-2 from the organ-cultured human skin. The biopsies of full-thickness human skin were processed as described in Materials and Methods. The biopsies (0.5×0.5 cm) were floated in 2 ml DMEM with or without 10 ng/ml TNF-α. After culture for 64 hours at 37°C with 5% CO₂ the conditioned media were collected for gelatinolytic activity assay and biopsies were processed for RT-PCR analysis as shown in Fig. 5. (A) The gelatinolytic activities from the conditioned media were examined by zymogram. Arrows indicate the 72 kDa latent and 62 kDa active MMP-2. (B) The identity of the gelatinolytic activity was confirmed by western blot with polyclonal anti-MMP-2, which recognizes both 72 and 62 kDa MMP-2. The gelatinase activities from the conditioned media were enriched by gelatin-Sepharose-4B. The bound protein was resolved by a reducing SDS-PAGE followed by a western blot. (C) The time course of TNF-α induced activation of pro-MMP-2 in human skin organ culture. The conditioned medium was sampled at the indicated time points and assayed by gelatinolytic zymogram. The skins from three independent donors were analyzed and gave similar results.

Fig. 2

TNF-α stimulates activation of pro-MMP-2 by human dermal fibroblasts embedded in collagen lattice. Primary human dermal fibroblasts were isolated from the normal human skin and the early passage cells were used for experiments. The fibroblasts were plated either as monolayer or embedded in collagen lattices at the same cell density and were stimulated by TNF-α at the indicated concentration. After 68 hours the conditioned media were analyzed by gelatinolytic zymogram. (A) The conditioned media from fibroblasts embedded in collagen lattice were analyzed for gelatinolytic activities. The zymogram was developed for 16 hours as standard procedure. (B) Identity of MMP-2 was confirmed by western blot with anti-MMP-2 antibodies. The conditioned medium was enriched for gelatinase by binding to gelatin-conjugated Sepharose. The bound protein was resolved by SDS-PAGE followed by western blot. (C) The conditioned media from fibroblasts grown in monolayer were assayed for gelatinolytic activities. This zymogram was developed for 40 hours in order to reveal the weak 62 kDa MMP-2. (D) The relative efficacy of collagen and TNF-α on activation of pro-MMP-2 was determined by densitometric scanning of the 72 and the 62 kDa MMP-2. The percentage of the 62 kDa MMP-2 at 0 and 50 ng/ml TNF-α were plotted. These data are the average of three independent experiments.

Fig. 3

The time-course study of TNF-α and collagen on the activation of pro-MMP-2. The early passages of human dermal fibroblasts were seeded as monolayers or cast into collagen lattices and stimulated with or without 10 ng/ml TNF-α. (A) The monolayer fibroblasts were stimulated or not with TNF-α and conditioned medium was sampled at the indicated time points. Gelatinolytic activities were examined by zymogram. (B) Fibroblasts embedded in collagen lattices were stimulated with or without TNF-α. (C) The relative amount of the 62 kDa MMP-2 from A and B was determined by densitometric scanning. The percentage of the 62 kDa active MMP-2 from was plotted against the time. These data are the average of three independent experiments.

Fig. 4

TNF-α and collagen synergistically induce MT1-MMP protein. The dermal fibroblasts were grown either as monolayer (A) or in collagen lattice (B). The cells were stimulated or not with 10 ng/ml TNF-α. Monolayer and lattice were collected at the indicated time and cell lysates were subjected to western blot with anti-MT1-MMP antibodies. Two forms of MT1-MMP of molecular mass 65 and 63 kDa were detected. A nonspecific band (NS) is indicated.

Fig. 5

TNF-α increases MT1-MMP mRNA in human skin and the dermal fibroblasts within collagen lattice. (A) Normal adult human skin was cultured in DMEM and stimulated with or without 10 ng/ml TNF-α as described in Fig. 1. The total RNA from the skin piece was extracted. The mRNA levels of human MT1-MMP and β-actin were measured by RT-PCR. One third of the PCR product for β-actin and the whole product for MT1-MMP were resolved by agarose gel and stained by ethidium bromide. (B) Human dermal fibroblasts were embedded in collagen lattice. The lattices were stimulated with TNF-α at 10 ng/ml or not for 24 hours. The mRNA levels of human β-actin, MT1-MMP, MMP-2, uPA, PAI-1, TIMP-1 and GAPDH were analyzed by RT-PCR. The size of each PCR product matches the predicted size (see Materials and Methods).

Fig. 6

NF-κB inhibitor, SN50, blocks TNF-α and collagen mediated MT1-MMP expression and pro-MMP-2 activation. The dermal fibroblasts were embedded in collagen lattice. The cell-permeable peptide inhibitor for NF-κB, SN50 and the control mutant peptide, SN50M, were added to the lattices at the indicated concentrations followed by stimulation or not with TNF-α at 10 ng/ml in DMEM. Conditioned media and lattices were collected after culture for 48 hours. (A) Cell lysates from the treated lattices were subjected to western blotting with anti-MT1-MMP. (B) Conditioned medium was analyzed by zymogram. The latent, an intermediate and the active MMP-2 with molecular mass 72, 64 and 62 kDa, respectively, are indicated.

Fig. 7

TNF-α stimulates the breakdown of IκB in fibroblasts populated in collagen lattices. Activation of the NF-κB pathway was monitored by the breakdown of IκBα. The fibroblasts were embedded in collagen and stimulated with or without 10 ng/ml TNF-α. The lattices were collected at the indicated time and the cytosolic fractions were subjected to western blot with anti-IκBα antibodies. (A) A representative western blot of two independent experiments is shown. (B) The IκBα protein level from the cytosolic fraction was determined by densitometric scanning.

Fig. 8

Mechanism of the TNF-α and collagen mediated activation of pro-MMP-9 in human dermal fibroblasts. Based on our results and those of others, we propose the following pathway for pro-MMP-2 activation, showing the role of TNF-α and the interplay between molecules involved in the pro-MMP-2 activation cascade. Binding of TNF-α and type I collagen to their receptors activates the NF-κB signaling pathway by inducing the breakdown of IκB, the cellular inhibitor of NF-κB. NK-κB then translocates from the cytosolic compartment into the nucleus. Binding of NF-κB to the NF-κB responsive cis-element in the promoter of the mt1-mmp gene triggers the transcription of mt1-mmp. Elevated levels of MT1-MMP mRNA generate MT1-MMP protein, which then translocates to the plasma membrane and forms a functional complex with TIMP-2. In contrast, mmp-2 gene transcription is constitutively active in fibroblasts and pro-MMP-2 accumulates in the extracellular environment. Increased MT1-MMP protein on the cellular membrane promotes cleavage of the N-terminal pro-domain of the 72 kDa pro-MMP-2, increasing the amount of the 62 kDa active form of MMP-2.