Pharmaceutical agents that are recognized for efficacy in one clinical application occasionally find new life in another. One well known example is aspirin which, originally used for its analgesic, antipyretic and anti-inflammatory actions, now also plays a major role in management of vaso-occlusive diseases due to its antithrombotic actions. A similar set of discoveries has occurred for several commonly used drugs whose anti-collagenase properties have only recently been explored. A significant advantage of these types of drugs is that there is already a considerable body of basic scientific data and clinical experience in using them, particularly with respect to side effects and idiosyncratic reactions. The most successful example has come from the tetracycline family, shown in Figure 9.2. Tetracyclines were originally derived from the soil bacteria, Streptomyces, and introduced clinically in 1948.
Since then, they have been an important class of antibiotics which treat a wide variety of infections in a number of different clinical settings. Other commercially available, semisynthetic members of the family include doxycycline and minocycline. The mechanism of their anti-bacterial action is by inhibition of protein synthesis through preferential and reversible binding to the 30S bacterial ribosomal subunit and subsequent inhibition of aminoacyl tRNA binding and polypeptide elongation.11,12 In therapeutic use, the tetracyclines are well tolerated overall. The principal side effects are gastrointestinal irritation, particularly with oral administration, and vestibular symptoms in 30-90% of patients receiving minocycline. Rarely, hepatic and renal toxicity can also occur at high doses of the drug and with intravenous administration, often in the setting of pre-existing renal and hepatic insufficiency. Dermatological side-effects and complications include photosensitivity, onycholysis (particularly with demeclocycline), hyperpigmentation associated with minocycline, and gramnegative folliculitis. Other superinfections include Candidal overgrowth, and pseudomem-branous colitis as a result of overgrowth of Clostridium difficile. Deposition of tetracyclines in teeth and bone, as well as pseudotumor cerebri with minocycline, are complications occurring primarily in infants and young children.11,12
Evidence has accumulated that the tetracyclines have anti-inflammatory actions which are distinct from their anti-microbial mechanisms and has been reviewed in detail else-where.4,8,13,14 The periodontal and rheumatological literature has pioneered the documentation of the tetracyclines' anti-collagenase properties and efficacy in periodontal disease, and in rheumatoid and osteoarthritis.4,15,16 Early evidence of a nonantimicrobial mechanism of action came with the reduction of alveolar bone loss in a diabetic rat model system and the associated inhibition of collagenase activity in the gingiva and skin of rats treated with minocycline, even under aseptic conditions.17 Other nontetracycline antibiotics were unable to match the effect. Subsequently, minocycline was shown to inhibit human synovial collagenase and numerous reports of anticollagenase activity of the tetracycline family followed in diverse biochemical, cell culture, and animal model systems,18-21 including studies focusing on inhibition of osteoclast-secreted collagenase,22 cancer invasion and metastasis,23 and aneurysmal degeneration.24 Further evidence for a distinct nonantimicrobial mechanism of action came with a series of chemically modified tetracyclines (CMT) that possessed inhibitory effects on collagenase yet lacked antibacterial activity.25 This effect also extended to other Ca2+ and Zn2+ dependent MMPs.
While it is clear that tetracyclines do possess distinct and even potent antiinflammatory activities apart from their antibiotic properties, the principal mechanisms of action are still unclear. A priori, all of the processes diagrammed in Figure 9.1 are possible targets, and so far there is some evidence to support the existence of multiple direct targets of inhibition along the protease expression pathway. An early hypothesis suggested that the well-known ability of the tetracyclines to chelate metal ions could be involved in preventing proper calcium or zinc ion binding to collagenase, especially since addition of calcium ions reversed the inhibitory effect of minocycline on neutrophil collagenase in vitro.17 In further support that tetracyclines inhibit cationic binding, a series of CMTs which typically lack the dimethylamino moiety at position 4 of tetracycline (4-dedimethylamino tetracycline) has been synthesized; the loss of anticollagenase activity in one, a pyrazole analog named CMT-5, was explained by the loss of moieties on the molecule that could chelate a divalent cation such as Zn2+ or Ca2+.26 Crystallographic studies may help decide whether all or some divalent cation binding sites in collagenases are affected.
Tetracyclines have been shown to exert inhibitory effects earlier in MMP gene expression. They may inhibit proper activation of the neutrophil procollagenase, either by suppressing normal oxidative cleavage of the propeptide or increasing degradation of the proenzyme,26-29 although this has not been observed by some workers.20 Transcription can be
affected, as seen when keratinocytes treated with doxycycline and CMTs had lower levels of MMP-2 mRNA, though some nonspecific cytotoxicity and suppression of total RNA were also noted, possibly indicating a chelation of free divalent cations necessary for general cellular function.30 More recently, tetracycline has been shown to inhibit induction of stromelysin mRNA by IL-1 while not affecting glycerol aldehyde phosphate dehydrogenase RNA levels, and this effect has been shown to require the presence of the AP-1 enhancer site upstream of the gene.31
Although tetracyclines have a well-documented direct inhibitory effect on collagenases and other MMPs, evidence also exists that these drugs have other mechanisms of action which may contribute to their in vitro, in vivo and clinical effects. Tetracyclines have been shown to inhibit production of inducible nitric oxide synthase, partially at the level of transcription and translation of this enzyme which produces nitric oxide, a molecule with pleio-tropic biological actions, including up-regulation of MMP activity.32 Inhibition of inducible nitric oxide synthase is in proportion to the tetracyclines' inhibitory potency against the collagenases.33 Some other reported cellular effects of tetracyclines which could explain their anti-inflammatory and anti-collagenolytic properties include variable modulation of phagocytic activity of leukocytes,34 inhibition of neutrophil motility,35 modulation of intracellular calcium concentrations during T-cell activation,36 and inhibition of type X collagen synthesis in chondrocytes that could not be reversed by addition of excess extracellular calcium ions.37 In some cases such as angiogenesis, the role of tetracyclines in inhibiting an-giogenesis by interacting with MMPs38 has been questioned since these drugs inhibited angiogenesis in an in vitro model, while a synthetic peptide-mimetic inhibitor with activity against a broad spectrum of MMPs did not.39
Regardless of the precise mechanism of action, the tetracyclines have already been employed to treat a variety of inflammatory diseases. Although there was anecdotal clinical evidence for therapeutic efficacy of tetracyclines in periodontal disease40 and arthritis,13 a small double-blind study using a relatively low-dose of tetracycline failed to show any improvement in patients with rheumatoid arthritis.41 However, an uncontrolled study using minocycline in conjunction with nonsteroidal anti-inflammatory drugs led to improvement in patients with rheumatoid arthritis.42 Subsequently, two larger, randomized, double-
blind, placebo-controlled studies in Europe and the United States have shown modest to moderate improvement in patients with long-standing rheumatoid arthritis who were treated with minocycline 100 mg twice daily.43,44 Most recently, a randomized, double-blind, placebo-controlled study of patients with early rheumatoid arthritis who had not been previously treated with so-called disease-modifying antirheumatic drugs (methotrexate, hydroxychloroquine, sulfasalazine or gold) showed improvement with minocycline 100 mg twice daily.45
A number of cutaneous disorders have been reported to respond to the tetracyclines, including acne vulgaris, acne rosacea, pyoderma gangrenosum, telangiectasias, pityriasis lichenoides et varioliformis acuta, panniculitides, pustulosis palmaris et plantaris, confluent and reticulated papillomatosis, mycosis fungoides, white sponge nevus, dystrophic epider-molysis bullosa, and bullous pemphigoid.46-48 Of these, the response of dystrophic epidermolysis bullosa and bullous pemphigoid to tetracycline therapy may in part be rationalized in terms of collagenase inhibition. Dystrophic epidermolysis bullosa is a severe, subepidermal blistering skin disease with autosomal recessive and dominant inheritance types which has been reported to respond to minocycline in two patients.49 Although interstitial collagenase was once considered to be a candidate for the primary defect in this disease,5,50 the disease is now known to be due to defects in Type VII collagen which forms the anchoring fibrils which help in mediating dermal-epidermal adhesion.51 However, Type VII collagen is a substrate for interstitial collagenase and gelatinase,52 and it is conceivable that inhibition of these enzymes by a tetracycline could explain some of the effects, albeit in just two cases.49 Bullous pemphigoid is an acquired blistering disease typically associated with au-toantibodies against two hemidesmosomal proteins. In contrast to the single case report of tetracycline therapy in dystrophic epidermolysis bullosa, there have been numerous isolated reports of tetracyclines with or without nicotinamide used to treat bullous pemphig-oid.46 More recently, two unblinded clinical studies with twenty and seven patients,53,54 respectively, reported that tetracycline plus nicotinamide was effective in the treatment of bullous pemphigoid, and comparable to systemic corticosteroids.53 Since one of the bullous pemphigoid autoantigens (BP180) has multiple collagen-like domains and both 92 kD gelatinase (MMP-9) and collagenase (MMP-1) are elevated in blister fluid, and BP180 is a substrate for gelatinase in vitro,51,55,56 it is tempting to speculate that collagenases are more intimately involved in the pathogenesis of bullous pemphigoid than conventionally thought, and that part of tetracycline's effect arises from its ability to inhibit these enzymes. So far, no attempt has been made to stratify patients based on their type of autoantibody. Both tetracycline and nicotinamide are also potential free-radical scavengers and electron acceptors and conceivably could prevent oxidative cleavage of procollagenase.28 Moreover, recently reported studies in transgenic mice support the notion that matrix metalloproteinases mediate not only connective tissue metabolism, but also are somehow involved in communicating with overlying epithelial tissues and directing their properties.57 Thus it is interesting to speculate that inhibition of collagenase may either directly help to prevent disruption of hemidesmosomes in the basement membrane, or indirectly aid by modulating dermal-epidermal interactions.
The tetracyclines have also been reported to have efficacy in a variety of other human inflammatory diseases which have no clear relationship to bacterial infection, such as persistent corneal ulcerations58 and recurrent aphthous ulcers.59 In addition, inhibition of MMPs has been postulated as one mechanism to explain the successful use of tetracycline as a sclerosing agent in chemical pleuredesis.60
Other drugs have also been shown to inhibit collagenase in vitro. Cephalothin, but not doxycycline, tetracycline or gentamicin, was able to inhibit MMP activity in extracts from tissue around loose total hip arthroplasty prostheses.61 This curious result was explained by the relative absence of neutrophil collagenase in such tissue extracts, though further studies and confirmation are required. Structurally related to the tetracyclines, anthracyclines such as aranciamycin have been reported as inhibitors,62 as have anthracene carboxylic acids, anthraquinones and coumarins,9 suggesting that not all rings of the tetracycline template are essential for activity. Of the nonantibiotic agents, phenytoin, a commonly used anti-epileptic, has been the most carefully studied. Following the initial observation that collagenase is elevated from fibroblasts of patients with recessive dystrophic epidermolysis bullosa, and is inhibited in vitro by phenytoin at the level of synthesis,63,64 anecdotal case reports and an open study reporting efficacy in patients with RDEB were reported.65,66 However, a multicenter, placebo-controlled, double-blinded study subsequently showed no improvement in clinical outcome of RDEB patients as measured by numbers of blisters and ero-sions.67 Since collagenase levels in blister fluid from these patients were not assayed, it is unclear if there is a subgroup of RDEB patients which may benefit from such therapy. Moreover, since increased collagenase is now known to be a secondary phenomenon in the patho-genesis of RDEB which is caused by defects in type VII collagen, the role of phenytoin in treating this particular disease may be minimal, although type VII collagen is a substrate for collagenase.52 Phenytoin may yet have beneficial uses in other diseases in which over-expressed collagenase plays a more prominent if not primary role.
Finally, while our discussion has focused on pharmacological inhibition of events downstream from gene activation, it is worth noting that various pharmacological agents such as Tenidap and methotrexate are capable of intercepting the IL-1 stimulated cascade,68,69 specifically inhibiting collagenase induction while not affecting TIMP-1 or stromelysin mRNA levels,69 and deserve further exploration in the future.
Was this article helpful?