Viewpoint 5

During the last decade, scientists have learned that proteases are more than just agents of destruction in the extracellular space. Here, we argue that proteases deserve much more careful scrutiny in dealing with the problem of itch pathogenesis and improved itch management. Proteases have become appreciated as signaling molecules. They signal by activating cell-surface receptors, thereby inducing a cascade of intracellular signals. Nature has developed several strategies to signal to cells via proteases, by (a) conventional binding and activation (plasmin-activator receptor), (b) cleavage of a protein which subsequently activates a receptor, or (c) nibbling off a part of the extracellular domain of a receptor which activates the same (autoactivation) or the neighbored (transactivation) receptor (proteinase-activated receptor) (PAR) (1). The fact that approximately up to 1200 genes (4.5% of the human genome) encode proteases in the human gene and 475 proteases are described in the human body reflects the impact of proteases as signaling molecules in the body under physiological and pathophysiological conditions. We differentiate between serine proteases, cysteine proteases, aspartic proteases, and metalloproteinases. They are generated by many skin cells, including epithelium, endothelium, leukocytes (T cells, mast cells, eosinophils, and neutrophils), bacteria, fungi, and viruses, some of them with a potential contribution to nerval stimulation (2, 3). A role of proteolytic enzymes for the induction of pruritus in humans was already suggested about 50 years ago4-6). In 1955, Arthur and Shelley demonstrated that cutaneous injection of one spicule with friction or pressure led to a burning itch lasting up to 30 min even when the spiculae were immediately removed from the skin by washing or by applying an adhesive tape (1, 2). Mucunain, a protein with endopeptidase and dipeptidase activity, was identified as the active pruritic agent. The protein was extractable only from the spicules using aqueous solutions and could be inactivated by autoclaving, changing the pH, or using a similar denaturing process that did not change the spicule structure. Further investigations confirmed the biochemical nature of this pruritic agent, identified as a thermolabile protein with a molecular weight of 40 000 (7). Depletion of mast cell histamine prior to the injection of trypsin or chymase markedly abrogated the itch response suggesting a role of trypsin and chymase as histamine liberators (4, 6). Thereafter, a variety of proteases were identified to be pruritogenic in humans in vivo including trypsin 8-10). Moreover, Hagermark and co-workers (11) postulated from their findings that serin proteases such as trypsin and mast cell chymase provoke itching and visible changes (edema, flare) when injected intracutaneously. Interestingly, injection of the cystein protease papain also provoked itch. Moreover, pretreatment with compound 40/80 did not influence papain-elicited itch indicating that papain-induced itch is independent from the release of histamine (12). Thus, a direct role of proteases on primary afferent nerves could not be excluded although the proteolytic effects were merely interpreted as ‘toxic agents’ on ‘free’ sensory nerve endings. Later, from human studies with patients suffering from atopic dermatitis (AD), we learned that protease inhibitors such as epsilon-amino-caprinoic acid have beneficial effects for the treatment of pruritus in AD patients, although more efficient when combined with antihistamines than alone (13). In support of this idea, a specific mast cell chymase inhibitior, Y-40613 was tested in a mouse pruritus model. Y-40613 dose-dependently suppressed the scratching response, strengthening the hypothesis that chymase directly contributes to the development of pruritus (14). (Table 1). On a receptor level, most evidence for a receptor-mediated effect of serine proteases on pruritus comes from studying PARs. PARs are G protein-coupled receptors with seven transmembrane domains. These receptors are activated by a unique mechanism where proteinases hydrolyze, at a specific cleavage site, the extracellular amino terminus of the receptor. This cleavage exposes a new amino terminus that acts as a tethered ligand, which binds intramolecularly to initiate cellular signals (15, 16). Short synthetic peptides based on the tethered ligand sequences of the different PARs (PAR-activating peptides: PAR-APs) mimic the effects of proteinases, activating selectively the different members of the PAR family. Four PARs have been cloned thus far: PAR1, PAR3, and PAR4 are targeted by thrombin and cathepsin G (PAR4) while PAR2 is activated by trypsin, mast cell tryptase, and other endogenous or exogenous serine proteases of different origin (3). To a lesser extent, other proteinases like factor Xa, cathepsin G, plasmin, and granzyme A are also capable of activating PARs. The development of receptor-selective PAR-APs has allowed accurately discerning the physiological consequences of PAR activation in vivo, describing a role for PARs in platelet activation, vascular functions, inflammatory or even nociceptive responses. PAR4 was the most recent member of the PAR family that has been described. With exception of a clear role in thrombin-induced human platelet activation, very little is known about the physiological and pathophysiological importance of PAR4. PARs are highly expressed in the nervous system. Both PAR1 and PAR2 have been described in neurons and the brain 17-21). Only recently, expression of PAR1 and PAR2 on neurons of the peripheral nervous system has been revealed. Both PAR1 and PAR2 were localized in guinea-pig enteric neurons (22). Activation of isolated myenteric neurons by thrombin, trypsin, tryptase, PAR1-AP and PAR2-AP resulted in calcium signal in more than 50% of isolated neurons. More than 60% of guinea-pig enteric neurons expressed PAR1 and PAR2. A large portion of these neurons also expresses neuropeptides such as substance P or vasoactive intestinal peptide and neurotransmitters such as nitric oxide and respond to adenosine triphosphate (22). Moreover, agonists of PAR1, PAR2, and PAR4 evoke depolarizing responses in Dogiel multipolar morphologic type II neurons with AH-type electrophysiologic behavior and Dogiel uniaxonal morphologic type I neurons with S-type electrophysiology (2). Importantly, functional PAR1 and PAR2 are present on primary spinal afferents, where their activation causes rapid intracellular calcium mobilization (21, 23, 24). In serial sections of human renal peripheral nerves, differential PAR expression was observed in healthy donors. Interestingly, PAR4 appears to be the most abundant PAR present in this particular peripheral nerve (2). PAR expression patterns in human and murine skin have not been fully explored yet, under normal and pathophysiological conditions. Recently, PAR2 has been localized on rat sensory neurons (25). Moreover, functional data strongly support the idea that the peripheral nervous system is directly regulated by PAR2 during neurogenic inflammation, pain, and pruritus (21, 26, 27). In rat skin, PAR2 is localized on approximately 63% of primary afferent sensory neurons, 30–40% of them colocalizing with SP or calcitonin-gene related peptide (CGRP). Later, PAR1 was described to contribute to neurogenic inflammation in murine skin (24). In rat skin, neuropeptides such as calcitonin gene-related peptide and substance P from primary spinal afferent neurons are known as important mediators of neurogenic inflammation and pain and probably contribute to pruritus. We speculated that serine proteinases including PAR2 agonists may activate PAR2 on sensory neurons to trigger itch in AD patients based on the following facts: there exists a close proximity of tryptase-containing mast cells to spinal afferent fibers and agonists of PAR2 cause effects similar to those of tryptase in many tissues, comprising many of the characteristics of itching. Itching is one of the most frequent symptoms in dermatologic diseases and accompanies inflammatory and immune responses of many diseases such as AD, hypersensitivity reactions, or urticaria, for example. Indeed, neuronal PAR2 appears to be involved in the induction of pruritus in human skin (28). Importantly, the endogenous PAR2 agonist tryptase was increased up to fourfold in AD patients, and PAR2 expression was markedly enhanced on primary afferent nerve fibers in lesional skin biopsies of AD patients. In contrast, no significant differences in histamine concentrations were observed between AD patients and healthy controls. On this background, one may speculate that tryptase may be more important for the transmission of itch responses in AD than histamine. Intracutaneous injection of specific PAR2 agonists provoked enhanced and prolonged itch when applied intralesionally. These effects were not diminished when cetirizine, a histamine type-1 blocker, was used indicating the specificity of these effects. Additionally, this observation may also explain why antisedative histamines are poorly effective in AD patients (29). Thus, PAR2 activation on cutaneous sensory nerves may be a novel pathway for the transmission of itch and inflammatory responses during AD and probably other skin diseases. PAR2 antagonists may be promising therapeutic targets for the treatment of pruritus (28). We postulate that different qualities of itch exist among various diseases based on the inducing ‘itchy’ molecule (histamine, tryptase, cytokine, protons, etc.) and the corresponding receptor on primary afferent neurons. In other words, in order to treat pruritus, future investigations have to focus on the characterization of the crucial molecules/receptors in each itchy disease. After the unexpected discovery of proteases as mediators of G protein-coupled receptors and their role during inflammation, immunity, tumor growth, and now itch, our view about proteases is more astonishing than ever. Besides neuropeptides, cytokines, amines, and kinins, serine proteases enter the stage as mediators of neuronal regulation. Using specific synthetic agonist peptides, PAR2 has been identified as a novel receptor for itch responses on primary afferent sensory nerves of patients with AD. This observation expands our knowledge of potential classes of itch mediators and supports the idea of a complex and multidimensional itch system. In particular, this implies that distinct itch mediators may be of differing relative importance during defined pruritic skin diseases (prurigo, AD, urticaria, renal pruritus, etc.). Thus, future research will have to focus on the different molecules that regulate the itch responses in a particular disease and will have to work out the underlying mechanisms that are crucial for the transmission of itch on the molecular level. With regard to PARs, further studies are necessary to fully explain the underlying direct or indirect effects of PAR2-induced itch. For example, neuropeptides released from neurons upon PAR2 stimulation may activate the release of nociceptive mast cell mediators such as kinins or prostanoids (30). We conclude that, in certain pruritic skin diseases and under defined circumstances, endogenous serine proteinases (including tryptase) as well as exogenous proteases may activate PAR2 on cutaneous sensory nerves, thereby mediating itch. Beyond this scenario, proteases may also induce itch responses by other mechanisms or as yet unknown ‘protease-receptors’ by activating specific intracellular signal transduction pathways (e.g. chymase). This will be a fascinating field within the ‘itch research community’ in the future. Supported by IZKF Münster, SFB293, DFG, CERIES, Paris; Rosacea foundation & Novartis Research foundation (M.S). The contribution of Martin Schmelz is gratefully acknowledged.