Inhibition of MMPs and ADAM/ADAMTS

Abstract

Matrix metalloproteinases (MMPs), A Disintegrin and Metalloproteinase (ADAM) and A Disintegrin and Metalloproteinase with Thrombospondin Motif (ADAMTS) are zinc-dependent endopeptidases that play a critical role in the destruction of extracellular matrix proteins and, the shedding of membrane-bound receptor molecules in various forms of arthritis and other diseases. Under normal conditions, MMP, ADAM and ADAMTS gene expression aids in the maintenance of homeostasis. However, in inflamed synovial joints characteristic of rheumatoid arthritis and osteoarthritis. MMP, ADAM and ADAMTS production is greatly increased under the influence of pro-inflammatory cytokines. Analyses based on medicinal chemistry strategies designed to directly inhibit the activity of MMPs have been largely unsuccessful when these MMP inhibitors were employed in animal models of rheumatoid arthritis and osteoarthritis. This is despite the fact that these MMP inhibitors were largely able to suppress pro-inflammatory cytokine-induced MMP production in vitro. A focus on ADAM and ADAMTS inhibitors has also been pursued. Thus, recent progress has identified the “sheddase” activity of ADAMs as a viable target and the development of GW280264X is an experimental ADAM17 inhibitor. Of note, a monoclonal antibody, GLPG1972, developed as an ADAMTS-5 inhibitor, entered a Phase I OA clinical trial. However, the failure of many of these previously developed inhibitors to move beyond the preclinical testing phase has required that novel strategies be developed that are designed to suppress both MMP, ADAM and ADAMTS production and activity.

Graphical Abstract

1. Introduction

Matrix metalloproteinases (MMPs), A Disintegrin and Metalloproteinase (ADAM) and A Disintegrin and Metalloproteinase with Thrombospondin Motif (ADAMTS) are zinc-dependent endopeptidases that play an integral role in the maintenance of normal organ and tissue homeostasis [1]. MMPs, ADAMs and ADAMTS are also intimately involved in several pathologic conditions, including, various forms of cancer [2–6], rheumatoid arthritis (RA) [7–9], and osteoarthritis (OA) [10–12]. In that regard, RA is characterized by a state of chronic inflammation of large and small synovial joints that is accompanied by defects in both innate and adaptive arms of the immune system [13–16]. Abnormal proliferation of activated synovial fibroblasts is perpetuated by the migration of activated T-cells and B-cells, dendritic cells, macrophages and neutrophils into synovial tissue [17, 18]. The migration of these cell types into the synovial space leads to elevated levels of pro-inflammatory cytokines, most prominently, IL-1ß, IL-6, IL-15, IL-16,IL-17, IL-18, IL-22, IL-23 and TNF-α [15, 19] which, in turn, are responsible for increasing MMP, ADAMs/ADAMTS gene expression in RA and psoriatic arthritis (PsA). By contrast, the early stages of OA likely occurs as a function of ageing and through abnormal mechanical stressors which alter chondrocyte function [20]. However, as OA progresses over time (time is probably measured in decades) a state of what has been termed, “non-classical inflammation” ensues. This involves immune cell dysregulation, heightened prostaglandin production and synovial tissue activation the latter considered a factor responsible for elevating pro-inflammatory cytokines that results in elevated levels of MMP, ADAMs and ADAMTS gene expression [11, 21–23]. Thus, MMPs and ADAMTS mediate the degradation of cartilage extracellular (ECM) proteins resulting in the loss of cartilage integrity [22–26] whereas ADAMS act a “sheddases” and the removal of membrane-bound receptors (e.g. mIL-6R) [27, 28].

The MMP superfamily is comprised of the classical types of MMPs. These include MMP-1 (collagenase-1), MMP-8 (neutrophil collagenase), MMP-13 (collagenase-3), and gelatinases (MMP-2) [29] and (MMP-9), the stromelysins exemplified by MMP-3, MMP-10, MMP-11, the matrilysins (MMP-7/PUMP-1 and MMP-26), the membrane-type MMPs (MMP-14, MMP-17, MMP-24/−25), [30–32], the ADAMs, also referred to as adamlysins, including, ADAM-2, −7, −−10, −11, −17, −18, −19, −22, −23, −29, −32, 33 [33–36], ADAM soluble variants, 9-S, 12-S, 28-S and DEC-1 [31] and the ADAMTS, most notably, ADAMTS-2, −3, 4, −5, −7, −10, −12, −13, −17, −20, SL-2 and SL-4 [37–40] As such, these enzymes have been of considerable interest for developing agents that curtail MMP gene expression and/or MMP production using classical medicinal chemistry paradigms designed to inhibit enzymes [41–47], blocking signal transduction pathways, initiated by TNF-α [48–50] after anti-TNF-α receptor blockade in Crohn’s patients [51] and in experimentally-induced arthritis with Type II collagen [52], IL-17 [53], and IL-6 [54, 55], by exploiting epigenetic mechanisms [56], by preventing enzyme-protein interactions [57], and by employing microRNA (miRNA) technology [58, 59] or small interfering RNAs [60] or by using viral vectors [61] or glycosylated and non-glycosylate substrates [62].

One of the more important aspects for curtailing inflammation involves ADAM-17 also known as Tumor necrosis factor-α converting enzyme [63]. Thus inhibition of TACE could serve as an effective way of blocking the involvement of TNF-α in inflammation [64–67]. In that regard, the development of effective inhibitors of ADAM17 is ongoing [68].

At present, the lack of selectivity for MMPs and ADAMs seriously compromise their use in the clinical setting [69–71]. Thus, in part, this narrative review focuses on the search for effective selective MMP inhibitors that could be added to various treatment modalities for RA, OA, and PsA.

1. Selection of Literature

The PubMed database (www.ncbi.nlm.nih.gov/pubmed) was employed to select papers for coverage in this narrative review. The search terms, “MMP Gene Expression/Drugs”, “ADAMS”, “ADAMTS”, “ADAMTS/Drugs”, “ADAMs/Drugs” “Synthetic Aggrecanase Inhibitors” and “Tumor necrosis factor-α converting enzyme” were employed for this purpose. Although a few cited references were published prior to 2004, the majority of the cited references from the PubMed database were from 2004 to 2018.

2. MMP-13, ADAMTS and ADAMs

2.1. MMP-13

MMP-13 (along with MMP-1, MMP-8) are interstitial collagen-degrading enzymes that have a particular relevance to the degradation of articular cartilage in RA and OA because they aggressively breakdown Type II collagen [72]. Of note, the expression of MMP-13 (along with MMP-1) is increased in response to IL-1 and TNF-α suggesting transcriptional regulation of both MMP-13 and MMP-1 genes. Moreover, high levels of MMP-1 and MMP-13 have been found in arthritic tissues [72].

2.2. MMP-13 Inhibitors

As indicated above, the degradation of Type II collagen and aggrecan constitute major cellular events in the progression of RA and OA to joint failure. Several findings have also implicated MMP-13 as a suitable target for the development of selective MMP-13 inhibitors [73, 74]. Thus, medicinal chemistry produced an MMP-13 inhibitor, PF152 (N-(4-fluoro-3-methoxybenzyl)-6-(2-(((2S,5R)-5-(hydroxymethyl)-1,4-dioxan-2-yl)methyl)-2H-tetrazol-5-yl)-2-methylpyrimidine-4-carboxamide), which was shown to decrease human cartilage degradation ex vivo as well as possessing the capacity to reduce the severity of articular cartilage lesions in dogs with OA induced by partial medial meniscectomy [75]. However, additional preclinical testing of PF152 indicated significant nephrotoxicity which was believed to have been mediated by human organic anion transporter 3. Thus, a follow-up analysis produced a compound lacking this nephrotoxic property [76]. As of this writing, a search of the PubMed data base using the search term, “MMP-13 inhibitors/Osteoarthritis Clinical Trials” failed to reveal any human OA trials as yet in which PF152 or its successor was evaluated for clinical efficacy.

2.3. ADAMTS

ADAMTS and ADAMTS-like proteins are members of a superfamily of 26 secreted enzyme molecules comprising 2 related, but distinct families. ADAMTS are zinc-metalloproteinases with a thrombospondin motif, whereas ADAMTS-like molecules lack the thrombospondin motif [77]. ADAMTS-5 is the principle “aggrecanase” found in animal [78] and human OA articular cartilage [77]. In that regard, the degradation and diffusion of Type II collagen and aggrecan fragments from OA articular cartilage without the compensatory synthesis of these macromolecules to replace those lost through degradation significantly compromises the biomechanical properties of articular cartilage [79].

2.4. ADAMTS Inhibitors

ADAMTS-5 was validated as a drug target for OA and experimental ADAMTS-5 inhibitors were shown to reduce synovial joint damage in OA animal models. Thus, an active ADAMTS-5 drug development program has been established with the lead compound, GLPG1972, being assessed in a Phase I OA clinical trial (NCT03311009). In addition to GLPG1972, a humanized anti-ADAMTS-5 monoclonal antibody, GSK2394002 [80] which was shown to inhibit ADAMTS-5 catalytic activity with a K_i 0.08nM, has been earmarked as a potential OA therapeutic agent. However, as of April, 2018, GSK2394002 did not appear to have progressed beyond preclinical evaluation. ADAMTS-4 and −5 also abolish cartilage integrity in RA by degrading aggrecan [81].

Additional novel ADAMTS-5 inhibitors are in the process of development. In one such study a bias-selection of antibodies analysis targeting ADAMTS-5 was shown to block the catalytic site of ADAMTS-5 [82] resulting in selective “aggrecanase” inhibition.

2.5. ADAMs and ADAM Inhibitors

We previously proposed a biological role for soluble IL-6 receptor (IL-6R) in OA [28]. In that regard, sIL-6R was shown to stimulate MMP synthesis by activating the JAK-STAT and ERK-MAPK signaling pathways in human chondrocyte cultures [83]. The sIL-6R is generated by “ectodomain shedding” [84–86] mediated by the ADAM class of metzincin proteases [87]. In the present view, dysregulation of “ectodomain shedding” mediated by ADAM proteases has been associated with autoimmune and cardiovascular diseases, neurodegeneration, cancer, infection, and inflammation [85]. Regarding the removal of the membrane form of the IL-6 receptor (mIL-6), this is carried out either by ADAM10 or ADAM17 [88], where ADAM17 is mostly associated with sIL-6R arising from neutrophils during acute and chronic inflammation [89]. In the course of recognizing the role played by ADAM17, an inhibitor, GW280264X was developed wherein this agent was shown to block the constitutive release of mIL-6R in addition to blocking the release of chemokines CX3CL1/fractalkine, and chemokine C-X-C ligand 16 [90]. This finding was consistent with a previous report showing that ADAM17, and not ADAM10 was responsible for removing TNF-α and L-selectin from leukocyte membranes [89].

3. Signal Transduction Pathways: Pro-inflammatory Cytokines, NF-κB, MMPs and Apoptosis

Early on the JAK-STAT pathway was identified as a critical inducer of inflammation in RA and PsA because several of the pro-inflammatory cytokines (e.g. IL-6) and other soluble mediators (e.g. interferon-γ) activate STAT proteins via their interaction with specific receptors [91–93].

However, more recently, other pro-inflammatory cytokines exemplified by the activity of IL-29 [94] was shown to increase the production of other pro-inflammatory cytokines (e.g. IL-1ß, IL-6, TNF-α), growth factors (e.g. fibroblast growth factor) that are intimately associated with the maintenance of chronic inflammation (e.g. elevated production of IL-8, CRP), and MMP-3/ADAMTS gene expression in synovial tissue obtained from OA patients [95]. This response also involved the activation of the NF-κB complex (Figure 1). Although this and other evidence have implicated the MAPK pathway in MMP and ADAMTS gene expression as well as apoptosis (Figure 1), the finding that activation of JAK-STAT signaling also leads to MMP gene expression indicated that more than one signaling pathway participated in regulating pro-inflammatory cytokine-induced MMP gene expression. Thus, in that scenario one envisions that more than one signaling pathway may have to be inhibited for cytokine-induced MMP gene expression to be completely dampened. We had also previously contended that this was likely to be the case for preventing the progression of tissue destruction in various autoimmune and inflammatory diseases [96]. Despite this conjecture, the small molecule inhibitor tofacitinib [97, 98] was developed with the goal of inhibiting the progression of RA mainly through its ability to reduce the capacity of pro-inflammatory cytokine [99] and MMP gene expression [100–102] both of which sustain the chronic inflammatory state characteristic of RA and PsA.

Figure 1. Pro-Inflammatory Cytokine Activation of NF-κB in Synovial Fibroblasts.

Figure 1 shows that pro-inflammatory cytokines (e.g. TNF-α) activates the NF-κB complex (IκB/ReLa/p50). The activation of the NF-κB complex can cause an enhancement of MMP gene expression [136]. However, as noted, TNF-α/TNFR does not alter ADAMTS activity [95]. Other factors such as Fas Ligand (FasL) produced by macrophages can lead to apoptosis through pro-caspase activation [137]. Growth factors can also activate the PI3K/AKT1/mTor pathway and “cross-talk” with the TNF-α/TNFR pathway [138]. This, and other evidence suggests a complex network of interacting ligand/receptor activities that modulates downstream synovial fibroblast pathways associated with inflammatory arthritis [136].

Figure 1 was previously published: C. J. Malemud. Intracellular signaling pathways in rheumatoid arthritis. J Clin Cell Immunol-Open Access 4 (2013) 160. DOI: 10.4172/2155-9899.1000160

4. Which If Any MMP Inhibitors Should Be Considered for Further Development?

4.1. In Vitro Cell Cultures Appear to be Integral for Assessing the Action of MMP Inhibitors

The association of periodontal inflammation and the development of RA has been reported [103–106] as well the crucial involvement of MMPs in the progression of periodontal disease [107, 108]. In that regard using periodontal ligament fibroblasts [107] was influential in assigning significance to MMPs in this disease process. Thus, these cells proved useful for identifying the main gelatinolytic activity in a band that migrated on gelatin-impregnated gels at 72kDa (MMP-2) whereas a less plentiful gelatinolytic band was observed at 92kDa (this was likely to be MMP-9). Moreover, several agents, namely, H7, staurosporine, cycloheximide and TGF-ß suppressed MMP-2 production, indicative of the utility of using these cells to potentially identify MMP-2 inhibitors. Platelets have also been identified as playing a role in arthritis and as such, the types of MMPs secreted by these cells is fundamentally important to developing inhibitors of platelet MMP activity [109].

Fibroblast-like synoviocytes [110], and cartilage “constructs” comprised of chondrocytes derived from pluripotential stem cells [111] have also been employed. In one such study [111], a readout of specific MMP activity in response to IL-1 was employed to identify agents that could neutralize this effect. In another study immortalized human chondrocytes [55] were employed to show that tocilizumab, a humanized monoclonal antibody that neutralizes the action of IL-6, suppressed MMP-9 production. Importantly, authentic mature chondrocytes [112] and other cell types were identified as target cells for determining the role of MMPs in disease processes. Additional studies used tumor cells [113] where a novel MMP-2 inhibitor 3-azidowithaferin A (3-azidoWA) was tested. Tumor cells that include human metastatic breast cancer cells [114–116] were employed to determine the effect of an MMP inhibitor reversion-inducing cysteine-rich protein with Kazal motifs (RECK) on levels of MMP-9, MMP-2, and membrane Type 1 [(MT1) MMP (MMP-14)] secretion [116]. Human cardiac smooth muscle cells [117], monocyte/macrophages [118], and rat astrocytes and microglia cultures [119] were among others that were employed as appropriate for testing experimental MMP inhibitors or for providing data on other strategies designed to limit the effects of MMPs. As previously mentioned, monoclonal antibodies [120] are also among several additional experimental MMP/ADAMs inhibitors undergoing pre-clinical testing in various inflammatory disorders and cancer. These include, MMP-14 [121], MMP-9 [122], and, ADAM17 [123].

5.2. In Vivo Studies

Many of these experimental strategies have emerged as a novel approach in basic and translational research that are designed to either inhibit MMP/ADAMTS gene expression or to reduce the pathologic changes brought about by these enzymes in various disorders where these cell types play an intimate role. A few of these MMP inhibition strategies have also been employed in preclinical animal models of arthritis [124, 125] and persistent inflammatory pain [126], the effect(s) of which are summarized in Table 1.

Table 1.

Effects MMP Inhibitors in Preclinical Animal Models of Arthritis and Persistent Pain

Animal Model Animal	Molecule/Targeted MMP/Effect(s)	Reference
Collagen-induced Arthritis Mouse	TGF-ß-inducible gene h3/MMP-1-cleavable composite peptide, MFK24/Reduced Inflammatory Mediators	124
Adjuvant-Induced Arthritis Rat	PI3Kδ,γ inhibitor-IPI-145/MMP-13/Reduced Arthritis Severity; Reduced MMP-13 gene expression	125
Collagen-Induced Arthritis Mouse	PI3Kδ,γ inhibitor-IPI-145/MMP-3,-13/Reduced Arthritis Severity; Reduced MMP3/MMP13 gene expression Anti-CII IgG2a:Total IgG Modestly Reduced	125
Persistent Inflammatory Pain CFA1-Rat	NOV2/CCN3/MMP-2, -9 Abolished Induction of MMP-9 (Dorsal Root Ganglia); Abolished Induction of MMP-2 and MMP-9 (Dorsal Horn Spinal Cord)	126
Persistent Inflammatory Pain CFA1-Rat	Attenuation of Mechanically-Induced Allodynia	126

¹Complete Freund Adjuvant;

²Nephroblastoma Overexpressed (Also Known as CCN3)

5. Conclusions and Future Perspectives

From the foregoing narrative review we concluded that development of inhibitors to directly suppress MMP gene expression has given way to other novel pharmacological strategies designed to impair MMP gene expression and/or MMP activity. The reason for this development appears to be several-fold. Firstly, inhibition of MMP activity using administered synthetic MMP inhibitors generally failed in vivo because of poor individual MMP selectivity as well as unacceptable bioavailability [127]. Secondly, OA, cancer and cardiovascular clinical trials which were designed to test the efficacy of administering synthetic MMP inhibitors were not only poorly designed, but the use of these MMP synthetic inhibitors also caused undesirable side-effects [128]. Thirdly, restoring the balance between MMPs and Tissue Inhibitor of Metalloproteinases (TIMPs), which is skewed towards MMP in most inflammatory disorders and generally determines overall MMP activity [129] was also difficult to achieve. In fact, several studies have clarified the extent to which deficient TIMP activity provided the perfect microenvironment that allowed MMP activity to efficiently degrade neutral substrates [130–132]. Thus, strategies designed to determine the extent to which improving serum TIMP levels to regulate MMP activity [133] in pre-clinical models and thereafter in clinical trials remains critical for reestablishing the proper balance of MMP to TIMP.

With these issues in mind, it is now noteworthy that one of the experimental strategies shown in Table 1 utilized activated MMP-2 to develop a potentially novel approach to anti-arthritis therapy [124]. This formulation was designed to show clinical efficacy in a preclinical experimental model of arthritis and in the short run, resulted in a follow-up study. Thus, Nam et al [134] followed up their initial study of TGF-ß1-gene h3 (ßig-h3) [124] by designing ßig-h3 derivatives, which encompasses the 4^th fas-1 domain truncated for H1 and H2 sequences of mouse and MMP-2 cleavable peptides complex, termed MFK902. Thus, MFK902 was specifically cleaved by activated MMP-2. Furthermore, MFK902 reduced arthritis severity in collagen-induced arthritic (CIA) mice that was accompanied by a reduction in the histopathologic changes in synovial joints associated with CIA. These results indicated that an MMP-2 cleavable peptide complex based on ßig-h3 structure could eventually become be a useful adjunctive anti-arthritis therapy. Another recent strategy was one in which the experimental design involved inhibiting the activity of MMP-13 and ADAMTS-5 after surgically-induced destabilization of the medical meniscus led to changes consistent with OA in male C57/BL6 mice [135]. This study employed intra-articular injection of silencing RNA (siRNA) directed at each enzyme or at the combination of MMP-13 and ADAMTS-5 [135]. The combination of siRNAs directed against MMP-13 and ADAMTS-5 resulted in just about the same inhibitory effect as MMP-13 siRNA alone. Moreover, the histological score of the OA mice improved as well.

It may be hard to believe but at the present time the antibiotic, doxycycline, is the only inhibitor of MMP activity approved by the US Food and Drug Administration for the treatment of musculoskeletal and other disorders where elevated levels of MMPs play a significant role in their pathogenesis and progression. Thus, there is ample room for assessing novel pharmacological strategies such as the ones reviewed herein which may prove to be useful in treating RA, OA and other musculoskeletal conditions going forward.