sign in sign up
<<
Home , Matrix metalloproteinase

OPEN TRANSPARENT Closeup ACCESS PROCESS TNFa cleavage beyond TACE

TNFa cleavage beyond TACE/ADAM17: 13 is a potential therapeutic target in sepsis and colitis

Christoph Becker-Pauly, Stefan Rose-John*

Keywords: ADAM ; inflammatory bowel disease; MMP13; protease web; sepsis See related article in EMBO Molecular Medicine http://dx.doi.org/10.1002/emmm.201202100

For decades, matrix and turnover but little is the intestinal epithelium. Moreover, ele- (MMPs) have been described as rather known about other functions. vated levels of bioactive TNFa affect the unspecific matrix‐ and ‐degrad- In this issue, Libert and coworkers intestinal permeability through endocy- ing . Accordingly, this group of report on a pro‐inflammatory activity of tosis of tight junction , thereby was mainly associated with MMP13 not by ‘simply’ degrading extra- increasing systemic inflammation. All and pathologic con- cellular matrix but through the release of these events were abolished in MMP13 ditions such as tumour development and biologically active soluble TNFa from its deficient mice, resulting in increased . However, broad spectrum membrane bound precursor form. Ecto- survival of MMP13 knockout animals, MMP inhibitors completely failed in domain shedding of TNFa is a paracrine when treated with LPS or DSS. The cancer therapy. authors conclude that MMP13 is a Recent studies with gene deficient potential therapeutic target for the treat- mice revealed more specific functions » …pro-inflammatory activi- ment of colitis (Vandenbroucke et al, of MMPs, e.g. the cleavage of chemo- ty of MMP13 not by ‘simply’ 2013). kines, thereby stimulating inflammatory TACE/ADAM17 has been identified by responses (Dufour & Overall, 2013). degrading extracellular ma- its ability to cleave TNFa but later on it Additionally, mass spectrometry‐based trix but through the release of turned out that many more proteins are proteomic techniques allowed for the biologically active soluble processed by this . TACE/AD- identification of hundreds of new sub- TNFa from its membrane AM17 knockout mice are not viable and strates of MMPs, giving rise to unexpect- bound precursor form. « show a phenotype strikingly similar to ed biological activity of these proteases in mice lacking ligands of the EGF . health and disease. All these ligands are transmembrane MMP13 is known to be one of few proteins and need to be shed from the capable of cleaving rigid signaling event in inflammation per- cell surface in order to be systemically collagen fibrils. This ability is based on a formed by the metalloprotease tumour active. More than 74 TACE/ADAM17 complex unfolding mechanism of the necrosis factor a converting enzyme substrates have been identified (Scheller triple‐helical collagen structure. There- (TACE, also known as ADAM17) (Schel- et al, 2011). Up to now, TACE/ADAM17 fore, MMP13 biology is well studied in ler et al, 2011). Although TACE/ADAM17 was considered to be the only biologically is believed to be the major sheddase of relevant TNFa cleaving enzyme in vivo TNFa, Vandenbroucke et al demonstrat- although it was clear that also other Biochemical Institute, Christian-Albrechts- ed that after LPS‐induced sepsis and DSS‐ proteases such as proteinase‐3 could University Kiel, Germany induced colitis, MMP13 is up‐regulated generate biologically active TNFa (Rob- Corresponding author: Tel: þ49 431 880 3336; a ‐ ‐ Fax: þ49 431 880 5007; and cleaves TNF . This leads to ER stress ache Gallea et al, 1995). Mice, in which E-mail: [email protected] and mucus depletion resulting in an the TACE/ADAM17 gene was inactivated DOI 10.1002/emmm.201302899 increased interaction of bacteria with only in neutrophils and monocytes/

ß 2013 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. This is an open access article under the terms of the Creative Commons Attribution License (CC BY 3.0), which permits use, distribution and reproduction 970 in any medium, provided the original work is properly cited. EMBO Mol Med (2013) 5, 970–972 www.embomolmed.org Closeup Christoph Becker-Pauly and Stefan Rose-John

macrophages, failed to shed TNFa upon ty of endogenous activators and/or of TACE/ADAM17 within the protease LPS challenge and were largely resistant inhibitors, e.g. TIMP’s, resulting in al- web signalizes no change in cell homeo- against LPS mediated endotoxin shock tered shedding of TNFa when MMP13 or stasis, while the lack of TACE/ADAM17 is (Horiuchi et al, 2007). TACE/ADAM17 is deleted. Therefore, it is recognized to stimulate/recruit compen- Since TNFa is involved in the induc- important to study MMP13 and TACE/ satory proteases. Nevertheless, Blobel and tion and maintenance of many inflamma- ADAM17 expression and activity in vivo coworkers provided evidence that long‐ tory diseases and blockade of TNFa is an in corresponding knockout animals with term inhibition of TACE/ADAM17 in- effective treatment for these conditions, regard to cleavage. This can duced compensation by ADAM10. At this TACE/ADAM17 has been considered a be challenging, but it is known that point, however, one should keep in mind therapeutic target. However, due to the protease activity does not necessarily that mice lacking ADAM10 or TACE/ severe phenotypes of mice lacking TACE/ correlate with mRNA or levels. ADAM17 are not viable demonstrating ADAM17, treatment of patients seemed to For e.g. ADAM17 mRNA is expressed that compensation is somewhat limited. be critical with regard to side effects. virtually in all cells and is not subject to A protease network does not explain Indeed, knock‐in mice with 95% de- major transcriptional regulation. The why MMP13 does not compensate for the creased levels of TACE/ADAM17 showed protease is, however, mostly located loss of TACE/ADAM17 and vice versa. increased susceptibility to inflammatory within the cell but is transported to the Proteases might be organized in different bowel disease, which again is critical for cell surface mainly in inflamed tissues sub‐clusters, meaning that MMP13 and medical treatment (Chalaris et al, 2010). and in cancer (Scheller et al, 2011). TACE/ADAM17 are not directly linked in Unexpectedly, Blaydon et al identified a Are proteolytic enzymes lone fighters the protease web. To clarify this issue, single patient with no functional TACE/ or do they have guidance to their identification of regulatory molecules that ADAM17 due to a mutation in the substrates in a molecular network that orchestrate such protease networks is Adam17 gene demonstrating that loss builds the protease web? Could it be important. The tetraspanins and iRhoms, of TACE/ADAM17 activity can be com- that groups of proteases are physically the latter being catalytically inactive pensated in humans (Blaydon et al, connected in ‘clusters’, either directly rhomboid proteases, might be key players 2011). or through linker molecules and that in this scenario. It was shown that these Libert and coworkers additionally these networks rule substrate cleavage? transmembrane proteins influence locali- found MMP7 to be up‐regulated in DSS‐ Such a molecular microenvironment has zation and biologic activity of ADAMs induced colitis in MMP13 knockout mice the capacity to register the loss of single and it is likely that other proteases are (Vandenbroucke et al, 2013). MMP7 had players and might then compensate the involved as well (Adrain et al, 2012; previously been shown to induce TNFa‐ lack of activity (Fig 1). Yanez‐Mo et al, 2011). Additionally, for release in a model of herniated disc numerous secreted proteases tethering to resorption. However, it is not clear the was demonstrat- whether this is a direct proteolytic effect » …analysing proteolytic ed, which further supports the hypothesis on TNFa or if it is mediated through events in health and disease of clusters of proteolytic enzymes. MMP7‐mediated stimulation of other with regard to localization, Interestingly, only few full knockouts proteases. Along the same line, expres- contributing activators, of proteases in animal models lead to sion of ADAM19, which was also identi- severe phenotypes or even lethality. For fied as a TNFa‐releasing enzyme, is inhibitors, or other regulatory example, all MMP knockout mice, except significantly increased in the mucosa of molecules, in one word, the of MMP14, reveal fairly mild phenotypic patients with colitis. These studies pro- protease web, is abnormalities, pointing to compensatory vide evidence for other proteolytic important. « mechanisms on mRNA‐ and/or protein enzymes that might contribute to TNFa levels, implying activators, endogenous cleavage and progression of colitis and inhibitors, enhancers and other regulato- that are possible candidates to compen- An example is the partial compensation ry molecules. This makes it challenging in sate lacking ADAM17 activity. of TACE/ADAM17 activity by ADAM10. these animal models to distinguish be- The crucial point in understanding Blobel and coworkers demonstrated that tween molecular differences that are protease networks in terms of compensa- in embryonic fibroblasts, deletion of specifically due to the loss of the deleted tion is, why MMP13 does not cleave TACE/ADAM17 recruits ADAM10 as protease or based on the regulation of TNFa in TACE/ADAM17 negative mice compensatory sheddase for many sub- compensatory enzymes. and why TACE/ADAM17 does not do so strates (Le Gall et al, 2009). This could be Proteolytic systems are obviously flex- in MMP13 negative mice. The most due to the accumulation of uncleaved ible in terms of substrate‐cleavage‐ obvious reason is an activation of substrates, which are then processed by compensation as demonstrated by Libert MMP13 by ADAM17 and vice versa, other proteases. However, at least in short and coworkers for the ectodomain shed- which is rather unlikely due to their term treatment, pharmacological inhibi- ding of TNFa (Vandenbroucke et al, 2013). activation mechanisms and the different tion of TACE/ADAM17 in cells also Next to the shedding of TNFa by TACE/ phenotypes of knockout mice. However, expressing ADAM10 led to no compensa- ADAM17 and MMP13, there are other both enzymes might influence the activi- tion. Consequently, the physical presence proteins fitting into this scenario. The

EMBO Mol Med (2013) 5, 970–972 ß 2013 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. 971 Closeup www.embomolmed.org TNFa cleavage beyond TACE

Figure 1. TNFa shedding by MMP13 and TACE/ ADAM17 in the protease web. Enhanced ectodomain a COMPENSATION NO COMPENSATION shedding of TNF leads to decreased epithelial barrier function, which promotes sepsis and colitis. Different proteases have been identified, however it is still ambiguous which factors under which SEPSIS and conditions guide these enzymes to their substrates. COLITIS Several regulatory proteins, such as tetraspanins (Tspans), inactive rhomboids (iRhoms), or tissue inhibitors of metalloproteinases (TIMPs), might be involved in protease-substrate-interactions. This Soluble TNFα physically linked protease web has - sometimes but not always — the ability to recognize the loss of Cellular ADAM17 single factors, e.g. in knock-out cells, which might knockout lead to molecular rearrangements capable of compensating the lack of proteolytic activity. ADAM MMP7 TOP VIEW (a disintegrin and metalloprotease); MMP (matrix metalloproteinase); P3 (proteinase-3). TNFα ADAM17 Proteases MMP13 ADAM19 Regulatory P3 proteins • Tspans • iRhoms • TIMPs • […]

amyloid precursor protein (APP) for in- References (2009) ADAMs 10 and 17 represent differen- stance fulfills distinct functions when Adrain C, Zettl M, Christova Y, Taylor N, Freeman tially regulated components of a general cleaved by different proteases, at worst M (2012) Tumor necrosis factor signaling shedding machinery for membrane proteins such as transforming growth factor alpha, leading to neurodegeneration and the requires iRhom2 to promote trafficking L-selectin, and tumor necrosis factor alpha. Mol development of Alzheimer’s disease. and activation of TACE. Science 335: 225-228 Blaydon DC, Biancheri P, Di WL, Plagnol V, Cabral Biol Cell 20: 1785-1794 In conclusion, analysing proteolytic RM, Brooke MA, van Heel DA, Ruschendorf F, Robache-Gallea S, Morand V, Bruneau JM, Schoot events in health and disease with regard Toynbee M, Walne A, et al (2011) Inflammatory B, Tagat E, Realo E, Chouaib S, Roman-Roman S to localization, contributing activators, skin and bowel disease linked to ADAM17 (1995) In vitro processing of human tumor inhibitors, or other regulatory molecules, deletion. N Engl J Med 365: 1502-1508 necrosis factor-alpha. J Biol Chem 270: 23688- in one word, the protease web, is impor- Chalaris A, Adam N, Sina C, Rosenstiel P, Lehmann- 23692 Scheller J, Chalaris A, Garbers C, Rose-John S tant. Anyhow, it will help to elucidate Koch J, Schirmacher P, Hartmann D, Cichy J, Gavrilova O, Schreiber S, et al (2010) Critical (2011) ADAM17: a molecular switch to control therapeutic strategies under certain patho- role of the disintegrin metalloprotease ADAM17 inflammation and tissue regeneration. Trends logical conditions and might even help to for intestinal inflammation and regeneration in Immunol 32: 380-387 develop personalized medical treatment. mice. J Exp Med 207: 1617-1624 Vandenbroucke RE, Dejonckheere E, Van Hau- Dufour A, Overall CM (2013) Missing the target: wermeiren F, Lodens S, De Rycke R, Van matrix metalloproteinase antitargets in in- Wonterghem E, Staes A, Gevaert K, Lopez-Otin Acknowledgements flammation and cancer. Trends Pharmacol Sci C, Libert C (2013) Matrix metalloproteinase 13 CB‐P and SR‐J were supported by the 34: 233-242 modulates intestinal epithelial barrier integrity in inflammatory diseases by activating TNF. Deutsche Forschungsgemeinschaft Bonn, Horiuchi K, Kimura T, Miyamoto T, Takaishi H, Okada Y, Toyama Y, Blobel CP (2007) Cutting EMBO Mol Med 5: 932-948 Germany (SFB877, projects A1, A9) and Yanez-Mo M, Gutierrez-Lopez MD, Cabanas C ‘ fl edge: TNF-alpha-converting enzyme (TACE/ by the Cluster of Excellence In ammma- ADAM17) inactivation in mouse myeloid cells (2011) Functional interplay between tetraspa- ’ tion at Interfaces . prevents lethality from endotoxin shock. J nins and proteases. Cell Mol Life Sci 68: 3323- Immunol 179: 2686-2689 3335 The authors declare that they have no Le Gall SM, Bobe P, Reiss K, Horiuchi K, Niu XD, conflict of interest. Lundell D, Gibb DR, Conrad D, Saftig P, Blobel CP

972 ß 2013 The Authors. Published by John Wiley and Sons, Ltd on behalf of EMBO. EMBO Mol Med (2013) 5, 970–972