New links for meprin β within the protease web

Abstract

Proteases are organised in interconnected networks, together forming the protease web whose disturbance can have detrimental consequences for tissue homeostasis and response to environmental insults. Membrane‐anchored sheddases are proteases that themselves can be released into the pericellular space by ectodomain shedding. Werny et al. have uncovered unexpected promiscuity in ectodomain shedding of meprin β, a metalloprotease with critical functions in inflammation and fibrosis. These findings suggest new links within complex proteolytic networks like the epidermal protease network with potential implications for skin homeostasis, inflammation and response to injury.

Comment on: https://doi.org/10.1111/febs.16586

Proteases do not act alone but in cascades, circuits and networks that together form the protease web. Disturbances of the protease web can result in disease, but its complex connections have been only poorly unravelled. Werny et al. have identified new links for meprin β within the protease web, helping to better understand the high connectivity of this sheddase with other proteases and its roles in tissue inflammation and fibrosis.

Comment on: https://doi.org/10.1111/febs.16586

Abbreviations

AD

Alzheimer's disease

ADAM

a disintegrin and metalloproteinase

CASP

caspase

CD109

cluster of differentiation 109

CST

cystatin

CTS

cathepsin

ELA

elastase

IL‐6R

interleukin‐6 receptor

KLK

kallikrein

LGMN

legumain

MEP

meprin

MMP

matrix metalloproteinase

MT1‐MMP

membrane‐type‐I matrix metalloproteinase

PLG

plasmin

PRSS8

prostasin

SASP

senescence‐associated subtilisin protease

SERPIN

serine protease inhibitor

SPINK

serine protease inhibitor Kazal‐type

ST14

suppressor of tumorigenicity 14

TMPRSS6

trans‐membrane serine protease 6

TREM2

triggering receptor expressed on myeloid cells 2

Proteases play pivotal roles in development, cellular signalling, tissue homeostasis and responses to environmental stimuli in health and disease. Either by degradation or by specific processing of target substrates, they irreversibly modulate extracellular environments or orchestrate intracellular processes to critically determine cell behaviour. Together with their inhibitors, proteases do not act alone but in cascades, circuits and complex networks, mutually affecting activity, localization and substrate repertoires. Classic examples of interconnected protease networks are blood coagulation and caspase‐mediated programmed cell death, which control seminal tissue and cellular responses and whose disturbance is inevitably associated with life‐threatening diseases. However, they are only parts of a higher order network, termed the protease web, that integrates all proteases (> 550 in humans) and inhibitors (> 150 in humans) and their complex interactions [1]. Despite rapid advances in high‐throughput degradomics technologies that have significantly contributed to deconvolution of the protease web [2], only a tiny fraction of its complexity has been unravelled. Hence, many missing links need to be established to devise effective strategies for personalised treatment for aberrant proteolysis in devastating diseases, such as cancer, neurodegeneration and chronic inflammatory disorders.

In this issue of The FEBS Journal, Werny et al. [3] provide new insight into the intricate interplay of the membrane‐anchored proteases meprin β, MT1‐MMP (membrane‐type‐I matrix metalloproteinase) and ADAMs 10/17 (a disintegrin and metalloproteinase) that have all been associated with carcinogenesis, inflammation and Alzheimer's disease (AD). Acting in the pericellular space, meprin β, MT1‐MMP and ADAMs 10/17 are sheddases, which can modulate extracellular environments by releasing bioactive ligands from the cell surface. Interestingly, the catalytically active ectodomains of all these sheddases are also shed themselves, not only controlling their activity in close proximity to the plasma membrane, but also extending their proteolytic activity further into the protease‐rich extracellular space. As shedding is mediated by proteolysis, these shedding events are critical links within the protease web and significantly contribute to dynamics of protease signal propagation. Thus, they must be tightly regulated in space and time to prevent uncontrolled proteolysis, potentially causing, or resulting from disease.

High flexibility in controlling ectodomain shedding can be conferred by multiple proteases cleaving the same substrate, allowing individual spatiotemporal regulation of expression and activity in complex tissue responses. Werny et al. now provide evidence for a striking overlap of ADAM10/17 and MT1‐MMP activity towards meprin β ectodomain shedding, using the exact same cleavage site. Consistently, all three sheddases can only release inactive but not active meprin β from the cell surface, indicating a similar effect of the meprin β propeptide on susceptibility to proteolysis by ADAM10/17 and MT1‐MMP. With MT1‐MMP, a new meprin β sheddase has been discovered, which allows fine‐tuning both the cell surface sheddase activity of meprin β to cleave bioactive mediators, such as IL6R (interleukin‐6 receptor), CD109 (cluster of differentiation 109) and TREM2 (triggering receptor expressed on myeloid cells 2), and the amount of soluble pro‐meprin β that eventually is activated by secreted tryptic proteases to functionally modulate the extracellular matrix [4].

Another and particularly interesting new link that has been established for meprin β within the protease web is its vice versa activity on MT1‐MMP to shed the ectodomain of this major collagenase, sheddase of bioactive ligands and facilitator of MMP activation [5]. Interestingly, MT1‐MMP ectodomain shedding has been described as a physiological event [6], generating an active soluble protease with enhanced activity in the pericellular space [7]. Data suggest that the MT1‐MMP ectodomain sheddase is a membrane‐anchored metalloprotease, but a defined candidate has not been identified yet [8]. Although focussing on meprin β rather than MT1‐MMP ectodomain shedding, Werny et al. now present convincing data that suggest meprin β as a potent candidate to fill this void. Importantly, in addition to taking part in proMMP9 activation [9], meprin β now connects to the complex MMP network also through MT1‐MMP, one of whose major functions is the zymogen activation of proMMP2 that in turn can activate other MMPs, thereby further transmitting the proteolytic signal [5].

As the authors have established these new links within the protease web in cell‐based systems, major questions arise on whether, where and under which conditions they are relevant in vivo. Since MT1‐MMP, as well as ADAM10/17, are expressed with low tissue specificity in many cell types and under many conditions, there might be hardly situations where meprin β ectodomain shedding is exclusively mediated by either of the three sheddases. Thus, the tissue specificity of meprin β will guide the search for the most relevant tissues and responses to physiological or pathological stimuli. Meprin β is part of protease networks in the brain, intestine, kidney and skin with high expression in the epidermal compartment [4]. The latter is of special interest, since, in addition to MT1‐MMP and ADAM10/17, it includes several member proteases such as kallikreins that are involved in the regulation of meprin β activity [10, 11] (Fig. 1). Upon injury or in inflammatory skin diseases, the epidermal protease network undergoes significant changes in expressed proteases and proteolytic interactions. Wound healing phenotypes in mice with genetic ablation of either MT1‐MMP or meprin β in epidermal keratinocytes are mild [12, 13], but both proteases are increased in abundance during the redifferentiation phase, whereas ADAM17 also plays important roles in keratinocyte differentiation [14]. Activities with similar functional consequences have been assigned to matriptase‐1 (ST14 (suppressor of tumorigenicity 14)) [15] that is closely related and shares cleavage specificity with matriptase‐2 (TMPRSS6), a major activator of membrane‐anchored meprin β [4], making it a potential candidate for meprin β activation in the epidermal compartment. As a caveat, most of these studies have been performed using mouse models, whereas Werny et al. could only demonstrate mutual shedding of meprin β and MT1‐MMP using human proteins. However, as the authors mention, this might be related to dynamic O‐glycosylation and thus could add another level of control to the complex system. Together, current evidence and the high connectivity of meprin β within the epidermal protease network (Fig. 1) warrant further analyses of its contributions to this very robust network with interconnected and redundant proteolytic activities. These might be addressed using humanised model systems and multiplexed gene editing strategies to understand complex rewiring of epidermal protease signalling in healing impairments and inflammatory skin disorders.

Fig. 1.

Connectivity of meprin β within the epidermal protease network. Red connections indicate new links analysed by Werny et al. [3]. Figure inspired by [11]. Created with BioRender.com. ADAM, a disintegrin and metalloproteinase; CASP, caspase; CST, cystatin; CTS, cathepsin; ELA, elastase; LGMN, legumain; MEP, meprin; MMP, matrix metalloproteinase; MT1‐MMP, membrane‐type‐I matrix metalloproteinase; PLG, plasmin; PRSS8, prostasin; SASP, senescence‐associated subtilisin protease; SERPIN, serine protease inhibitor; SPINK, serine protease inhibitor Kazal‐type; ST14, suppressor of tumorigenicity 14; TMPRSS6, trans‐membrane serine protease 6.

Conflict of interest

The authors declare no conflict of interest.

Author contributions

VC and UadK wrote, reviewed and edited the manuscript.