Neutrophil Elastase and Cathepsin G

Neutrophil elastase and cathepsin G are serine proteinases that are synthesized as precursors in promyelocytes in bone marrow and subsequently stored in the azurophil granules of polymorphonuclear leukocytes as active enzymes. Mature leukocytes do not synthesize elastase, but they mobilize azurophil granules to the cell surface and release the proteinases in response to various stimuli. Monocytes have low levels of elastase but lose the enzyme during the differentiation into macrophages. Neutrophil elastase and cathepsin G are basic glycoproteins with isoelectric points larger than 9 (neutrophil elastase) and about 12 (cathepsin G). They can be readily trapped in cartilage matrix that has a negative charge.

Neutrophil elastase and cathepsin G cleave elastin; the telopeptide region of fibrillar collagen types I, II, and III; other collagen types IV, VI, VIII, IX, X, and XI; and other ECM components such as fibronectin, laminin, and aggrecan at neutral pH. These serine proteinases also can be involved indirectly in the breakdown of ECM by activating the zymogen of matrix metalloproteinases (proMMPs)⁶ and by inactivating endogenous proteinase inhibitors such as α₂-antiplasmin, α₁-antichymotrypsin, and tissue inhibitors of metalloproteinases (TIMPs).

Mast Cell Chymase and Tryptase

Chymase and tryptase are packaged in secretory granules together with histamine and other mediators in mast cells, which are infiltrated in rheumatoid synovium. Chymase is a chymotrypsin-like proteinase with a broad spectrum of activity against ECM components such as type VI collagen⁷ and aggrecan. It also activates proMMPs such as proMMP-1, proMMP-3, and proMMP-9.⁶ Although prochymase is activated intracellularly and stored in the granules, the activity in the granules is limited at low pH and becomes fully active when released extracellularly. Tryptase is a trypsin-like proteinase that degrades collagen type VI⁷ and fibronectin; it also activates proMMP-3.⁶

Plasmin and Plasminogen Activators

Plasminogen is synthesized in liver and secreted to plasma. It can bind to fibrin and to cells, and after activation by plasminogen activators, plasmin readily digests fibrin. Membrane-bound plasmin also degrades many ECM components including proteoglycan, fibronectin, type IV collagen, and laminin.⁸ Other important functions of plasmin are to initiate the activation of proMMPs,⁶ activate latent cell-associated transforming growth factor (TGF)-β1, and act as proenzyme convertase. Plasmin is generated through activation of plasminogen mainly by plasminogen activators, principally two serine proteinases, tissue-type plasminogen activator (tPA) and urokinase-type plasminogen activator (uPA).

The tPA is synthesized as a proenzyme of 70 kD and is secreted into the circulating blood primarily by endothelial cells, fibroblasts, chondrocytes, and tumor cells.⁸ In addition to being a major activator of plasminogen for fibrinolysis, tPA plays a key role in the clearance of fibrin from the circulation.

The uPA molecule was first purified from urine as a proenzyme of 54 kD.⁸ It is converted to the active form of two chains of 30 kD and 24 kD linked by a disulfide bond. Another fully active form of 33 kD is generated by plasmin. Although the expression of uPA is limited to certain cells such as renal tubules and bladder urothelium under physiologic conditions, it is more widely expressed in various cells including invasive cancer cells, migrating keratinocytes, and activated leukocytes in pathologic situations. Pro-uPA and two-chain uPA bind to a specific uPA receptor, a single-chain glycoprotein with a glycosylphosphatidylinositol (GPI) moiety expressed on fibroblasts, macrophages, and tumor cells. Receptor-bound uPA preferentially activates cell membrane-bound plasminogen into plasmin. Cell membrane-bound plasmin can activate receptor-bound pro-uPA. Among its specificities, uPA has a limited action on fibronectin.

Kallikreins

Two types of kallikreins, plasma and tissue kallikreins, are known. Plasma kallikrein, with two disulfide-linked chains (36 kD and 52 kD), is generated from prokallikrein of 88 kD by coagulation factor XIIa or by kallikrein itself. It activates kininogens to bradykinin and activates proMMP-1 and proMMP-3.⁶ Tissue kallikrein is synthesized in glandular tissues. It releases Lys-bradykinin from kininogen and activates proMMP-8.⁶

Matrix Metalloproteinases

The human MMP family comprises 23 different members, which have MMP designations (numbered according to a sequential numbering system) and common names coined by the authors of the published reports (see Table 8-2). On the basis of the biochemical properties provided by the domain structures and on their substrate specificity, these family members are classified into two major subgroups: secreted-type MMPs and membrane-anchored MMPs. MMP-4, MMP-5, and MMP-6 are excluded from the list because they are identical to other known MMPs (i.e., MMP-3 and MMP-2). MMP-18 and MMP-22 also are missing in Table 8-2 because they are assigned to Xenopus collagenase-4 and chicken MMP.

Most secreted-type MMPs including collagenases, stromelysins, and other MMPs are composed of three basic domains—the propeptide, catalytic, and hemopexin-like domains—that are preceded by hydrophobic signal peptides (Figure 8-1). The N-terminal propeptide domain has one unpaired cysteine in the conserved sequence of PRCGXPD. The cysteine residue in the sequence interacts with the essential zinc atom in the catalytic domain to prevent it from binding the catalytic water molecule, maintaining the proenzyme in an inactive state. The catalytic domain has the zinc-binding motif HEXGHXXGXXH, in which three histidines bind to the catalytic zinc atom. The four-blade, C-terminal hemopexin-like domain, which is connected to the catalytic domain by a proline-rich hinge region, interacts with ECM components and can play a role in determining the substrate specificity in some MMPs. Gelatinases have these domains with additional insertions of collagen-binding type II repeats of fibronectin in the catalytic domain (see Figure 8-1); this provides them with collagen-binding properties. Matrilysins are the shortest, lacking the hemopexin-like domains. Furin-activated MMPs contain insertions of a basic motif with a cleavage site by proprotein convertases including furin at the end of the propeptide domains (see Figure 8-1).

Figure 8-1 Domain structures of two types of matrix metalloproteinases (MMPs) (secreted-type MMP and membrane-anchored MMP) and two types of a disintegrin and metalloproteinases (ADAMs) (a disintegrin and metalloproteinase with thrombospondin motifs [ADAMTS] and membrane-type ADAM). The typical domain structure of most secreted-type MMPs (collagenases, stromelysins, and other MMPs) is composed of a prodomain, catalytic domain, hinge, and hemopexin-like domain. Gelatinases (MMP-2 and MMP-9) have additional insertions of collagen-binding type II repeats of fibronectin in the catalytic domain, whereas matrilysins (MMP-7 and MMP-26) lack a hemopexin-like domain. Furin-activated MMPs (MMP-11 and MMP-28) contain an RKRR sequence (furin recognition site = F) at the end of the propeptide. Membrane-anchored MMPs are composed of type I transmembrane-type MMPs (MT1-, MT2-, MT3-, MT5-MMPs); GPI-linked MMPs (MT4, MT6-MMPs); and type II transmembrane-type MMP (MMP-23). All of them have furin recognition sites. ADAMTS has a prodomain, a furin recognition site (F), a catalytic domain, a hinge, a disintegrin domain (D), thrombospondin motifs (Ts), and a spacer domain (S). Membrane-type ADAM is composed of a prodomain, furin recognition site (F), catalytic domain, hinge, disintegrin domain (D), cysteine-rich domain (C), EGF-like domain (E), and transmembrane domain (Tm).

Membrane-anchored MMPs include three different types of MMPs: type I transmembrane-type MMPs, GPI-linked MMPs, and type II transmembrane-type MMP. MMP-14, MMP-15, MMP-16, and MMP-24 (MT1-, MT2-, MT3-, and MT5-MMPs, respectively) have type I transmembrane domains in the C-terminal region, but MMP-17 (MT4-MMP) and MMP-25 (MT6-MMP) contain GPI anchors in the C-terminal region without transmembrane domains (see Figure 8-1). MMP-23 is unique in that it has type II transmembrane domain, a cysteine array, and an immunoglobulin-like domain instead of a hemopexin-like domain (see Figure 8-1).

ADAM and ADAMTS Families

Two ADAM (a disintegrin and metalloproteinase) gene families exist: membrane-type ADAM with transmembrane domain (ADAM) and secreted-type ADAM with thrombospondin motifs (ADAMTS) (see Figure 8-1). The active sites in the catalytic domains of most members of both gene families contain a common sequence of HEXGHXXGXXHD with the “Met-turn,” which also is present in MMP members.

The ADAMTS family includes 19 members. Although information about substrates and biologic functions is still limited, ADAMTS1, ADAMTS2, ADAMTS3, ADAMTS4, ADAMTS5, ADAMTS8, ADAMTS9, ADAMTS14, ADAMTS15, ADAMTS16, ADAMTS18, and ADAMTS20 all are ECM-degrading proteinases (see Table 8-3). ADAMTS1,⁴⁹ ADAMTS4,⁵⁰ ADAMTS5,⁵¹ ADAMTS8, ADAMTS9, and ADAMTS15 can preferentially cleave aggrecan at the five Glu-X bonds including the Glu³⁷³-Ala³⁷⁴ bond (the aggrecanase site). Because of aggrecan-degrading activity, ADAMTS4 and ADAMTS5 are named aggrecanase-1 and aggrecanase-2^50,51; versican also is digested by these proteinases,⁵² and brevican is cleaved by ADAMTS4 (see Table 8-3).⁵³ The C-terminus-truncated ADAMTS4 also degrades fibromodulin and decorin.⁵⁴ ADAMTS16, ADAMTS18, and ADAMTS20 also appear to have weak aggrecanase activity. ADAMTS2 and ADAMTS3 process the N-terminal propeptides of type I and II collagens and are named procollagen N-proteinase. Activity of procollagen N-proteinase also is known with ADAMTS14. ADAMTS13 is a von Willebrand factor–cleaving proteinase, and its mutation causes thrombotic thrombocytopenic purpura. Proteinase activities of other ADAMTS species are still unknown.

The human genome contains 25 ADAM genes including 4 pseudogenes, and thus the human ADAM family is composed of 21 members (Table 8-4). Among the ADAMs, ADAM8, ADAM9, ADAM10, ADAM12, ADAM15, ADAM17, ADAM19, ADAM20, ADAM21, ADAM28, ADAM30, ADAM33, and ADAMDEC1 exhibit proteolytic activity (proteinase-type ADAMs) (see Table 8-4). Although ADAM10, ADAM12, and ADAM15 degrade type IV collagen, the main substrates of these ADAMs are various membrane proteins, which include precursors of cytokines and growth factors such as TNF, HB-EGF and neuregulin, IGF-BPs, receptors such as p75 TNF receptor, IL-1 receptor II, and other membrane proteins related to development such as Notch ligand and ephrin (see Table 8-4).^55–60 According to these data, a major function of the ADAMs is shedding of the membrane proteins. ADAM17 cleaves a proform of TNF at the physiologic processing site into the soluble form of TNF and is called TNF-converting enzyme. ADAM17 also is involved in release of L-selectin, TGF-α, and p75 TNF receptor.⁵⁵ ADAM9, ADAM12, and ADAM17 can shed HB-EGF from its precursor. ADAM12 and ADAM28 cleave IGF-BP-3 and IGF-BP-5.^60,61 CD23 is shed by ADAM8, ADAM15, and ADAM28.⁶² Other functions of ADAMs include binding to integrins, cell-cell interaction, cell migration, and signal transduction (see Table 8-4).⁶³