MEDICAL ONCOLOGY (1997) 14, I - I0

Matrix metalloproteinase inhibitors in the treatment of cancer

PETER D. B R O W N

Department of Clinical Research, British Biotech Pharvnacezlticals Lid, WaHin~on Road, Oxford, UK

MatrLx metalloproteinases are a family of zinc-containing proteolytic enzymes that break down extracellular matrix proteins in normal physiological processes such as embryogenesis, tissue growth, and wound healing. The family includes collagenases, gelatinases, strometysins and metalloelastase. Observational and experimental data from studies of human malignancy indicate that these proteinases are induced by the turnout in order to reconstruct adjacent normal tissue to allow neovascularisation, tumour growth and spread. Tumours have been shown to overexpress certain matrix metalloproteinases relative to normal tissue and recent studies have shown an association between high levels of expression and poor prognosis. A large series of synthetic inhibitors have been developed using the structure of a principal substrate, collagen. The inhibitors contain a chemical group that binds the zinc atom in the active site of the metatloenzyme. Inhibition is specific for the known ma~'ix metaltoproteinase family and is reversible. Studies with these inhibitors and native tissue inhibitors of matrix metalloproteinases have shown that they can prevent the growth and spread of experimental turnouts. In other studies, the inhibitors have been shown to be directly anti-angiogenic. Synthetic matrix metalloproteinase inhibitors have now reached the stage of clinical testing and preliminary results indicate that the compounds may be effective in slowing tumour growth. Trials currently underway should reveal whether this approach will become a standard part of anti-neoplastic therapy in the future. Keywords: matrix metalloproteinase inhibitors; angiogenesis; metastasis.

INTRODUCTION The d e v e l o p m e n t of matrix metalloproteinase (MMP) inhibitors as a potential treatment for cancer b e g a n in the early 1980s following research b y several groups showing that neoplastic transformation is a c c o m p a n i e d by a marked increase in the expression of degradative enzymes, including the MMP, collagenase [1-3]. A second MMP active in degrading the type IV collagen of basal laminae was also isolated f r o m metastatic sarcoma cells at about this time [4,5]. The h u m a n M M P family is To whom correspondence should be addressed at: Department of Clinical Research, British Biotech Pharmaceuticals Ltd, Watlington Road, Oxford, OX4 5LY, UK.

n o w k n o w n to include at least 15 enzymes, which collectively are capable of degrading all components of the extracellular matrix (Table 1). Three 'collagenases" have b e e n identified, interstitial collagenase (MMP1), neutrophil collagenase (MMP8) and coltagenase 3 (MMPI3)o These enz y m e s can degrade the generally proteolytic resistant fibrillar collagens m a k i n g a characteristic 314 length b r e a k in the c~-chain [6,7]. There are two type IV collagenases [8,9], n o w t e r m e d gelatinase A (MMP2) and gelatinase B (MMPg) which, as described by Liotta [5], can degrade type IV collagen of basal laminae as welt as other "nonhelical" collagen domains a n d proteins such as fibronectin and laminin. The gelatinases have also b e e n s h o w n to degrade native insoluble elasfin [10]. Three enzymes h a v e b e e n classified as 0736-0118 :~, 1997 Chapman & Hall

2

Table 1.

MATRIX METALLOPROTEINASE tNHtBITORS

Human matrix metailoproteinase family

Enzyme

Number

Principal substrates ~

interstitial collagenase Neutrophil collagenase Cotlagenase-3

MMP-1 MMP-8 MMP-13

Fibrillar collagens, types I, II, III Fibritlar collagens, types t, tl, Ill Type I collagen

Gelatinase A Gelatinase B

MMP-2 MMP-9

Non-fibrillar collagens, fibronectin, taminin Non-fibrillar collagens, types IV and V

Stromelysin-I Stromelysin-2 Stromelysin-3

MMP-3 MMP-10 MMP-11

Non-fibriltar collagens, proteoglycan, laminin Nonfibrillar collagens, proteoglycan, laminin Serine protease inhibitors (serpins)

Matrilysin Metailoelastase

MMP-7 MMP-12

Non-fibritlar collagens, fibronectin, laminin Elastin, non-fibrillar collagen

MT1-MMP b MT2-MMP MT3-MMP MT4-MMP

MMP-14 MMP-15 MMP-16 MMP-17

Progelatinase A Not defined Progelatinase A Not defined

MMP-18

Not defined

aThe principal substrates listed are only helpful as a guide; in practice the substrate specificity shown in vitro is broad with considerable overlap between MMPs. bMT stands for membrane-type.

'stromelysins" although only stromelysin I (MMP3) and stromelysin 2 (MMP10) are closely related functionally, degrading various proteoglycan components of the extracetlular matrix as well as fibronectin and laminin [11,12]. Stromelysin 3 (?vLMPll) was identified relatively recently in the tissue surrounding invasive breast carcinoma [13]. Its preferred substrate remains a matter of debate, however, it is effective in degrading the serpin :~-1 antitrypsin and in doing so may potentiate the action of serine proteinases such as urokinase [14]. Two enzymes have been identified which, on the basis of sequence homology, do not belong in the three subgroups described above. These are matrilysin (MMPT, formerly known as Pump) [15] and metalloelastase (MMTq2) [16]. Matrilysin is a short 'truncated' proteinase which can degrade non-fibrillar collagen, fibronecfin and laminin. Metalloelastase, as the name suggests, is capable of degrading elastin. In the past two years, the MMP family has grown by the addition of a new subgroup, the membrane-type-, or NIT-, MMPso Currently four members have been identified (MMP14-17) [17-20]. These proteinases have a c-terminal transmembrane domain that allows them to be anchored in the cell membrane. The substrates for most of these enzymes have yet to be established, however, MT1-MMP and MT3MMP (MMP14, MMP16) do appear to be specific activators of latent gelatinase A [19,21] (see below). More recently, MMP18, an enzyme with some homology to the stromelysins, has been added to the family [22] and it is likely that other members ~ l l be discovered in the next few years. It is now generally accepted that MMPs play an important role in normal processes of tissue

remodelling such as trophoblast implantation, embryogenesis, mammary, involution, tissue growth and wound repair. In these processes, the activity of these degradative enzymes can be kept in check at three levels. Firstly, gene expression is often tightly controlled with most, but not all, MMPs being expressed only when their activity is required. Secondly, MMPs are first synthesized as latent proenzymes that require the proteolytic removal of a 10 kDa amino-terminal domain in order to become proteolytically active. Thirdly, the activated MMPs can be inhibited by various proteinase inhibitors such as ~2-macroglobutin and more importantly the family of tissue inhibitors of metalloproteinases, T!MPs 1 - 4 [23-26]. The secretion and activation of gelatinase A has been studied in some detail. This MMP is secreted as the pro-enzyme with its native inhibitor TIMP-2 bound at a non-inhibitory site. The pro-gelatinaseTIMP-2 complex is believed to be activated by cell surface MT1-MMP. TIMP-2 released from the complex is presumably in a good position to rebind to activated gelafinase A, this time blocking the active site [21]. This abundance of negative control mechanisms may explain how gelatinase A can be expressed constitutively in some tissues, such as arteriolar smooth muscle, without causing extensive degradation. Its activation in this tissue is likely to be tightly controlled and the active species short lived. More detailed reviews on the matrix metalloproteinase family and their inhibitors are given in [27,281. In human malignancy, there is now a large body of observational and experimental data which indicates that MMPs are important in the processes of tumour invasion and metastasis. FollowMEDICAL ONCOLOGY (1997) 14 (1)

BROWN

ing the early observations of elevated cotlagenolytic activity in human tumours, the expression of individual MMPs has been studied, and native and synthetic MMP inhibitors have been applied in models of tumour invasion, tumour growth and metastasis. The aim of this review is to provide an overview of the latest developments in this field, focusing on the clinical application of MMP inhibitors in the treatment of cancer.

MATRIX METALLOPROTEINASE EXPRESSION IN CANCER Studies of the expression of specific MMPs and their native inhibitors, TIMPs, in samples of human cancer have provided valuable correlative evidence in support of a role for these enzymes in tumour growth and metastatic spread. It is hoped that such studies will, in the future, identify individual MaMPs or groups of MMPs as key factors in the progression of particular cancer types. This will help to focus the currently rather broad clinical programmes and will also provide a rationale for the development of synthetic MMP inhibitors with a narrower range of activity. However, as already mentioned, there are now at least 15 members of the human MMP family and multiple levels at which their activity can be regulated. Understanding the part played by each of these enzymes in a particular tumour type is a complex and arduous task. Generally, studies of MMP expression in human cancer employ only one or two techniques to reveal one facet of the overall pattern of activity. Studies of mRNA expression by Northern blot analysis can provide information about the quantity of expression relative to tumour grade or matched normal tissue but in situ hybridization is required to identify the cellular source of the MMP or TIIVIP. Also, mRNA expression does not always equate with protein expression. Equally, many techniques (ELISA, immunohistochemistry) that are used to study expression at the protein level are unable to estimate the proportion of enzyme that is in the activated, as opposed to the latent, state. Substrate gel electrophoresis (zymography) and conventional enzyme-substrate assays can provide information on the extent of activation and net proteolytic activity, respectively. Unfortunately, although sensitive, the technique of zymography is in practice restricted to the gelafinases. The conventional enzyme-substrate assays are relatively insensitive and provide little information on which/vlaMPs are contributing to the proteolysis. Consequently, the information provided by these studies must be viewed with a critical eye in order to assess the relative importance of the findings. MEDICAL ONCOLOGY (1997) 14 (1)

Rather than give a broad review of MMP expression in human cancer it is probably more informative to review the body of research into MMP expression in one particular malignancy, colorectal cancer. Collagenases One of the earliest studies on MMP expression in colorectal cancer examined the degradation of fibrillar type I collagen by tumour extracts [29]. This activity was shown to be increased in poorly differentiated compared to well differentited tumours, and also in tumours invading the serosa compared to tumours whose invasion was confined to the muscularis propria. The specific MMP responsible was not identified but is likely to have been one of the collagenases in view of the substrate and the fact that the activity could be blocked by metalloproteinase inhibitors. Immunohistochemical staining for interstitial collagenase has shown that expression of this MMP is elevated in colorectal adenocarcinoma relative to both adenoma and normal mucosa. Staining was most intense over stromal fibroblasts and collagen fibres with the strongest staining in stromal tissue adjacent to islands of turnout cells. Staining was generally absent over the tumour cells themselves [30]~Interestingly, the same study showed staining for the inhibitor TIMP-1 to be very similar to that for interstitial collagenase although there was no intensification in staining at the tumour-stroma interface. TIMP-1 staining was also more marked on the epithelial basement membrane and blood vessels, suggesting a possible protective role (see discussion below). More recently Murray et al. have reported that interstitial collagenase expression is associated with poor prognosis in colorectal cancer [31]. In a series of 64 colorectal tumours, immunostaining for interstitial collagenase was noted in 5/38 Dukes' B turnours and 5125 Dukes' C tumourso The prognosis of patients with collagenase positive tumours was significantly poorer than that of patients with no collagenase staining and this factor appeared to be independent of tumour stage and patient age. The incidence of collagenase positivity is much lower than in the previous report where 20/20 colorectal samples showed staining for this MMP [30]. The significance of this difference is not clear although other studies have also reported a lower frequency of cotlagenase expression in this tumour type [32,33]. Gelatinases

Formerly known as type 1V collagenases, these MMPs have been widely studied in human malignancy in part because of the early interest

MATRIX METALL OPROTEINASE INHIBITORS

4

in their ability to degrade the non-helical collagens of basal iaminae. Analysis of gelatinase A mRNA and protein expression revealed a marked increase in expression in colorectal tumour relative to normal mucosa and a significant correlation between the intensity of immunohistochemical staining and Dukes" classification [34]. The expression and extent of activation of gelatinases in human tissue can also be assayed by the relatively simple technique of zymography. In the largest reported series of human colorectal samples, Liabakk and colleagues used this technique to show that gelatinase B, pro-gelatinase A and activated gelatinase A are significantly elevated in colorectal carcinoma relative to adenoma and normal colorectal tissue [35]. The particular technique used in this study did not resolve pro- and activated forms of gelatinase B although this has been achieved in other zymographic series [36,37}. Interestingly, Dukes' B tumours showed much lower levels of gelatinase B and activated gelatinase A than either Dukes" A or C tumours, which the authors suggest may reflect differences in the tumour/stroma interaction at the invasive front. In situ mRNA hybridization indicated that gelatinase A was produced by fibrobtasts or myofibroblasts of the tumour stroma rather than the tumour cells themselves. There was no correlation between gelafinase A or B expression and survival. Zymography has also been applied to samples of colorectal carcinoma obtained by microdissection. Pro-gelatinase B and pro- and activatedgelatinase A were shown to be elevated in invasive regions of the tumours by factors of 12-, 3- and 10-fold respectively over matched normal tissue [37]. Activated-gelatinase B was absent from these invasive regions but was detected in areas of necrosis and inflammation. This is consistent with in situ mRNA hybridization studies s h o ~ n g that macrophages and neutrophils are the primary source for gelatinase B [33,38]. A recent Northern blot study of gelatinase B mRNA expression in a series of 71 colorectal samples and matched normal tissue showed a marked increase in expression in the turnout samples with significant correlations with presence or absence of distal metastasis and Dukes" stage. Furthermore, a high tumour/normal gelatinase B mRNA ratio was associated with a significantly shorter disease-free and overall survival. This ratio was found to be an independent predictor of diseaseflee survival [39]. Finally, getatinase B is one of the few MMPs to have been assayed in the plasma of cancer patients. Concentrations of this MMP in the plasma of patients with colorectal cancer were found to be significantly greater than concentrations in either healthy subjects or hospitalized noncancer patients [40]. Further studies on circulating

MMP levels are likely" as MMP ELISA methodology is refined and commercialized. Other MMPs Expression of matrilysin has been shown to be elevated in colorectal cancer at the level of mRNA [41] with the highest expression in Dukes" C/D carcinomas and liver metastases [42,43]. Stromelysin-1 was localized by immunohistochemistry to the extracellular matrix surrounding clumps of invasive tumour cells in 19 out of 40 colorectal tumours studied [33]. Stromelysin-3 mRNA has also been shown to be elevated in colorectal primaries and metastases relative to normal mucosa, with the mRNA localized to stromal fibrobtasts adjacent to neoplastic foci [44]. TIMPs An important observation of recent years is that not ordy are MMPs overexpressed in malignant disease but so too are their native inhibitors, TIMPs. In a study of 56 colorectal cancers and 10 liver metastases, with matched normal mucosa, Zeng et aI. showed that TIMP-1 mRNA levels are increased 12- and 10-fold in the primary tumours and metastases respectively. T1MP-1 expression did not correlate with tumour size, location or differentiation but was significantly higher in patients with either local or distant metastases compared to patients without [45]. In situ mRNA hybridisation showed the elevated TIMP-1 expression to be limited to stromal fibroblasts or myofibroblasts surrounding the cancer ceils. Elevated expression was not seen in more distant mesenchymal or mucosal cells. TIMP-1 protein expression in colorectal carcinoma was reported to be relatively infrequent in one study [33], but was detected in all 20 colorectal carcinomas stained by Hewitt et aI. [30]. In the latter study, staining for ThMP-1 was continuous in the epithelial basement membranes of 6/7 adenomas, but was only continuous in 2/20 carcinomas with focal staining in a further 15 samples. Overexpression of a native metalloproteinase inhibitor is not unique to colorectal cancer and has been observed in other malignancies [46-48].

TUMOUR STROMA-'DEFENDER" OR "COLLABORATOR'? The previous section demonstrates the relative complexity of analysing the pattern of MMP/TIMP expression in human malignancy. Several of the published studies give conflicting results particularly with respect to the relative incidence of expression of specific M-MPs. Two studies [31,39] MEDICAL ONCOLOGY (1997) 14 (1)

5

BROWN

intriguingly report a prognostic significance for the overexpression of interstitial collagenase and gelatinase B, respectively, but when analysed at a different level the prognostic value of one of these MMPs is not seen [35]. However, certain common themes do emerge, perhaps most importantly in the contribution of the tumour stroma. Stromal fibroblasts appear to be the primary source for elevated expression of interstitial collagenase, gelatinase A, stromelysin-3 and TIMP-1, while macrophages and neutrophils in inflammatory tissue within the t u m o u r appear to contribute gelatinase B. In m a n y cases MMP expression is restricted to the t u m o u r - s t r o m a interface. If stromal fibroblasts are the source for m a n y of the MMPs considered important in tumour progression, this raises the question of the role of the stroma in t u m o u r growth. Is it a 'defender" or a 'collaborator" in the conflict between the b o d y and the malignancy? It is likely that it has the potential to be both. Clearly, to 'succeed' a malignant neoplasm must do more than degrade its immediate environment. In effect the tumour must 'remodel' the local tissue to suit its own needs. The generation of a modified and increased vasculature is perhaps the most obvious feature of this "remodelling', but associated with this must be the generation of supportive stromal tissues. This may explain the high levels of TIMP observed in invasive lesions. The invasive process is not simply an upregulation of tissue degradation, but rather an upregulation of tissue remodelling, requiring inhibitors as well as proteinases for the generation of new tissues, even if they are imperfectly formed. To this extent the stroma acts as 'collaborator" with induction of proteinases and inhibitors by the adjacent tumour cells resulting in angiogenesis and invasive growth. However, it also seems likely that the fibrotic reaction to a t u m o u r is part of a host defence mechanism, an attempt by the body to 'wall off' the invasive lesion. The significance of the extent and form of t u m o u r stroma has long been a matter of debate. Studies on h u m a n invasive breast cancer have suggested that a marked stromal reaction with intense fibrosis is an indicator of poor prognosis [49,50], however, a recent study of over 1200 cases of infiltrating ductal carcinoma found no relationship between fibrosis and outcome [51]. Fibrosis was also found not to be of prognostic significance in a series of 105 infiltrating breast adenocarcinomas [52]. By contrast, the presence of fibronectin in invasive breast cancer was found to be an indicator of good prognosis [53]. In a study of over 500 colorectal carcinoma samples, pronounced fibrosis at the t u m o u r edge was found to indicate a poor prognosis [54] while in pancreatic cancer an increased stromal element was associated with a low incidence of metastasis [55]. MEDICAL ONCOLOGY (1997) 14 (1)

The significance of an inflammatory cell reaction has also been questioned. In a morphometric study of h u m a n invasive breast cancer, high macrophage counts in the t u m o u r were associated with reduced relapse-free and overall survival [56]. Another recent study has found the presence of an inflammatory infiltrate in high grade and c-erB-2 positive breast cancers to be indicative of good prognosis [57]. Similarly, in a study of colorectal cancer, lymphocytic infiltration with perivascular lymphocytic cuffs was associated with increased overall survival [54]. Since m a n y studies indicate that inflammatory cells are the primary source of MMP-9 in tumours this also raises questions about the role of this MMP in turnour invasion~ Again, the inflammatory element of the stromal reaction probably has the ability to act as both 'collaborator" and 'defender', with the balance being decided by the nature of the t u m o u r - s t r o m a interaction. Therefore, if the stromal tissue can be viewed as being 'ambivalent' in its response to the tumour, then perhaps MMP inhibitor therapy can be thought of as a means of stiffening its resolve to restrict turnout growth; to act as a 'defender'. As can be seen from the simplified diagram (Fig. 1) it is theoretically possible that MMP inhibitor treatm e n t might assist some tumours of very high degradative potential by restoring the enzymeinhibitor balance to a point at which tissue remodelling becomes possible (a transition from A to B). However, this does not seem to occur in practice, as observed through cancer models and the first clinical studies. Moreover, there are early indications that a therapeutic effect (a transition from B to C) m a y be attainable in patients.

PRECLINICAL STUDIES WITH M M P INHIBITORS An important early series of synthetic MMP inhibitors were those based on the structure of the collagen molecule at the site of cleavage by interstitial cotlagenase. Peptide mimetics from the right-hand side of this cleavage site containing a hydroxamate zinc-binding group have been tested extensively in animal models and at least three such compounds are now in clinical trial. The structure of one of these compounds, marimastat (BB-2516, British Biotech), is shown in Fig. 2. The inhibitor fits into the active site of the MMP in a stereospecific m a n n e r and the co-ordination of the zinc atom in the M ~ active site by the hydroxamate group results in a potent (low nanomolar) but reversible inhibition [58]. Like most of the compounds currently u n d e r development, marimastat is a "broad spectrum' MMP inhibitor showing inhibitory activity against all classes of

MA TRtX METALLOPROTEINASE INHtBITORS

MMP, but little or no activity against more distantly related metalloproteinases such as angiotensin converting enzyme and enkephalinase. The activity of these compounds against the recently

Fi0. 1 a, Excessive degradation, By breaking down the extracellular matrix in its immediate environment the turnout (T) is unable to establish new blood vessels and other structures essential for growth, b. Balanced remodelling, Proteinase activity induced by the turnout is countered to a sufficient extent to allow the spread of turnout and the developmefit of new tissue, c. Excess proteinase inhibitor. Breakdown of the extracellular matrix is prevented leading to encapsulation of the tumour by avascutar fibrotic stroma.

described cytokine-processing metalloproteinase activities has yet to be fully defined [59]. Early studies with synthetic MMP inhibitors, such as the hydroxamate SC'44463 (G.D. Searle), and recombinant TIMPs demonstrated that these inhibitors could block tumour cell invasion through matrix barriers in vitro and inhibit organ colonisation by tumour cells (experimental metastasis) in vivo [60,61]. At this time MMP inhibitors were primarily thought of as antimetastatic drugs; that is drugs to block the passage of metastatic cells "into' and 'out of' lymphatic and vascular channels. However, as further studies -were carried out it was realized that these inhibitors might have the potential to inhibit the ~ o w t h of established tumours, either at their primary or secondary site. In a model of rat mammary carcinoma, the hydroxamate batimastat (BB-94, British Biotech) was shown to reduce both the spread of metastatic cells and the growth of established metastases [62]. Batimastat was also shown to inhibit loco-regional regrowth and metastasis of a human breast carcinoma, MDA-MB-435, following resection in

Fig. 2. Structure of marimas[at. Many of the inhibitors currently being studied are derived from the peptide structure of the ~-chain of type I collagen at the point at which collagenase first cleaves the molecule. Marimastat is based on the right hand side of this cleavage site although the chemical groups at the Pl', P2' and P3' positions are different from the original amino acid residues. The hydroxamate {-CONHOH) group binds the zinc atom in the active site of the MMP enzyme.

MEDICAL ONCOLOGY (1997) 14 (1)

BROWN

athymic nude mice [63]. In another series of experiments, increased expression of TIMP-2, in clones of rTIMP-2 transfected metastatic 4R cells, was associated with marked suppression of tumour growth and local invasion follo~ng subcutaneous implantation. Clones showing suppressed growth also showed a greater degree of fibrotic stroma, although the ability of these clones to colonise lung tissue in experimental metastasis assays was largely undiminished [64]~ In these models it seems likely that MMP inhibitors are able to suppress tumour growth both by blocking tumour-mediated extracellular matrix degradation and, as a consequence, by inhibiting tumour neovascularisation. Indeed, antiangiogenic activity has been demonstrated in various angiogenesis assays for both TIMPs and the synthetic inhibitors [65,66]. Recent studies have also shown that MMP inhibitors may be effectively combined with established cytoreductive cancer treatments. In a study of the synthetic inhibitor CT1746 (Celltech) and cyclophosphamide the two compounds combined were shown to be significantly more effective in inhibiting the growth and metastasis of the murine Lewis lung carcinoma than either agent used alone [67]. In a model of human ovarian carcinoma, bafimastat and cisplatin were shown to be significantly more effective in prolonghlg survival than either single agent and, as in the study with CT1746, the additive therapeutic effects did not appear to be accompanied by additive toxicity (R. Giavazzi, personal communication). As noted earlier, many of the synthetic inhibitors currently being developed have a broad spectrum of activity for MMPs. Since it is likely that certain tumours will rely on particular MMPs or classes of MMPs for their invasive potential attempts have been made to develo p more specific or 'selective' inhibitors. The main obstacle to the development of selective inhibitors has been the attainment of oral bioavailability. This has been a general problem with the whole class of synthetic MMP inhibitors and it has taken several years to achieve good oral bioavailability with broad spectrum compounds [58]. CDP-845 (Celltech), a potent selective gelatinase inhibitor, has recently been withdrawn from development in part because of poor oral bioavailability. Ro 32-3555 (Roche) is a hydroxamate based inhibitor with relatively weak activity against gelafinase A and stromelysin-1, but good activity against interstitial coUagenase. This compound shows good activity when given orally in animal models of arthritis and is currently in clinical development as a treatment for rheumatoid arthritis [68]. It is expected that other orally bioavaflable selective MMP inhibitors will follow Ro 32-3555 into the clinic. MEDICAL ONCOLOGY (1997) 14 (1)

CLINICAL TRIALS WITH MATRIX METALLOPROTEINASE INHIBITORS IN CANCER PATIENTS The first two MMP inhibitors to be tested in patients were ilomastat (GM6001, Glycomed) [69] and batimastat (BB-94, British Biotech) [70]. Neither compound showed good oral bioavailability. Ilomastar was administered as a topical agent in patients with corneal ulceration while batimastat was given as an intraperitoneal or intrapleural suspension in patients with malignant effusions. Currently, at least four MMP inhibitors are believed to be in clinical trial in cancer patients as oral treatments; AG 3340 (Agouron), CGS-27023A (Novartis), 12-9566 (Bayer), and marirnastat (BB-2516, British Biotech). To date results have only been presented for marirnatstat. Marimastat displays limited oral bioavailability in rodents and systemic delivery by mini-pump is required for therapeutic blood concentrations of the drug to be maintained~ These are believed to be in the range of 40-80 #g 1-~, However, preliminary results from healthy volunteer studies showed high blood concentrations of marimastat follov~g oral administration [58]. Marimastat is currently being tested in a series of trials in cancer patients. Results from the first of these trials were presented at a recent European Society for Medical Oncology meeting and provide the first indications that the therapeutic potential seen in animal cancer models may be realized in the clinic. A series of studies in patients with advanced malignancy examined the effect of different doses of marimastat on the serum cancer antigens CA125, CEA, PSA and CA19-9. There was a dose-related reduction in the rate of rise of these markers, with a proportion of patients showing a fall in the absolute cancer antigen serum concentration over the 28-day study period [71- 73]. In a separate study in patients with advanced gastric cancer, treatment with marimastat was associated with changes in the macroscopic and histological appearance of the tumours consistent with an increase in the quantity of fibrotic stromal tissue. The changes were very similar to those seen in various cancer models and several of the patients appeared to benefit from these alterations in tumour-stroma ratio [74]. Preliminary indications are that MMP inhibitors are generally well tolerated when given for periods of 3 - 6 months. The nature of trials in patients with advanced malignancy complicates the analysis of potential side-effects and a clearer picture must await randomised placebo-controlled trials. Musculo-skeletal pain has emerged as the principal treatment related side effect with marimastat. The severity and rate of onset of symptoms were found to be dose related and the effects were considered manageable at the dose range selected for future

/VIATRIX METALLOPROTEINASE INHIBITORS

studies. The condition generally resolved rapidly on discontinuation of m a r i m a s t a t a n d several patients restarted t r e a t m e n t after an interruption of 2 - 4 weeks [71-74], T h e m e c h a n i s m responsible for the musculo-skeletat p a i n has n o t b e e n established b u t it s e e m s likely that it is related to inhibition of metalloproteinase activity in the n o r m a l physiologic remodelling of t e n d o n s a n d joints.

PROMISE AND EXPECTATIONS In the p a s t 50 y e a r s m u c h e n e r g y has b e e n d e v o t e d to t e c h n i q u e s to eradicate metastatic disease a n d yet overall mortality rates h a v e c h a n g e d little. A m a l i g n a n t t u m o u r s h o w s a r e m a r k a b l e ability to a d a p t to the various t r e a t m e n t s directed towards it, driven b o t h b y its genetic instability a n d high proliferative rate. M M P inhibitors a n d other n e w approaches, s u c h as anti-angiogenic c o m p o u n d s , differ f r o m cytoreductive t r e a t m e n t s in that they are essentially targeting the other c o m p o n e n t of malignancy, the stroma. As this review has h o p e fully illustrated, the t u m o w " s t r o m a plays a vital role in the growth, invasion a n d s p r e a d of m a l i g n a n t disease. P e r h a p s this target will be less able t h a n the t u m o u r to e v a d e treatment. In blocking the ability of the t u r n o u t to utilize the adjacent tissue for its o w n p u r p o s e s these n e w treatments m a y reveal a n "Achilles heel' in the m a l i g n a n t p h e n o t y p e ; n a m e l y its reliance o n the 'collaboration" of n o n - m a l i g n a n t tissue. M a r i m a s t a t is c u r r e n t l y b e i n g tested in a series of r a n d o m i z e d controlled trials with survival as the p r i m a r y endpoint. T h e s e will p r o v i d e the m o s t i m p o r t a n t test of this n e w a p p r o a c h . Several other M M P inhibitors are also m o v i n g into later p h a s e trials. If significant inhibition of t u m o u r growth a n d s p r e a d can be a c h i e v e d it will alter the w a y both s u r g e o n s a n d oncologists view their respective m e a n s of intervention. Resection of m o r e w i d e s p r e a d disease m i g h t b e c o m e worthwhile if it is k n o w n that the r e s i d u a l t u m o u r can be held in check. Equally, the u s e of r a d i o t h e r a p y a n d c h e m o t h e r a p y in patients w h e r e the responses are short lived s h o u l d b e c o m e m o r e worthwhile if the time to relapse can b e significantly extended. The next two years are likely to reveal w h e t h e r these b r o a d s p e c t r u m inhibitors will b e c o m e part of a n e w generation of anti-neoplastic agents.

REFERENCES 1 Taylor,A.C., Levy,, B.M. and Simpson, J.W. (1970) Collagenolytic activity of sarcoma tissues in culture. Nahr 228, 366-7. 2 Dresden, M.H., Heilman, S.A. and Schmidt, J.D. (1972) Collagenolyfic enzymes in human neoplasms. Cancer Res. 32, 993- 6.

3 Yamanishi, Y., Maeyens, E., Dabbous, M.K., Ohyama, H. and Hashimoto, K. (1973) Cotlagenolytic activity in malignant melanoma. Cancer Res. 33, 2790-801. 4 Liotta, L.A., Tryggvason, S., Garbisa, S., Hart, I., Foltz, C.M. and Shafie, S. (1980) Metastatic potential correlates with enzymatic degradation of basement membrane collagen. Nature 284, 67-8. 5 Liotta, L.A., Tryggvason, IC, Garbisa, S., Robey, P.G. and Abe, S. (1981) Partial purification and characterisation of a neutral protease which cleaves type IV collagen. BiochemLstry20, 100-4. 6 Wilhehn, S.C., Eisen, A.Z., Teter, M., Clark, S.D., Kronberger, A. and Goldberg, G. (1986) Human fibroblast collagenase: glyeosylation and tissue-specific levels of enzyme synthesis. Proc. Natl. Acad. Sci. USA 83, 3756-60. 7 Hasty, K.A., Pourmotabbed, T.F., Goldberg, G.I., Thompson, J.P., Spinella, D.G., Stevens, R.M. and Mainardi, C.L. (1990) Human neutrophil collagenase; a distinct gene product with homology to other matrix metailoproteinases. ]. Biol. Chem. 265, 11421-4. 8 Collier,I.E., Wilhelm, S.M., Eisen, A.Z., Manner, B,L., Grant, G.A., Seltzer, J.L., Kronberger, A., He, C., Bauer, E.A. and Goldberg, G.I. (1988) H-ras oncogene-transformed human bronchial epithelial cells (TBE-1) secrete a single metalloprotease capable of degrading basement membrane collagen. J. Biol. Chem, 263, 6579-87. 9 Wilhelm, S.M., Collier, LE., Marmer, B.L., Eisen, A.Z., Grant, G.A. and Goldberg, G.L (1989) SV40-transformed human tung fibroblasts secrete a 92-kDa Type IV collagenase which is identical to that secreted by normal human macrophages. J. Biol. Chem. 264, 17213-21. 10 Senior, ILM., Griffin, G.L., Fliszar, C.J., Shapiro, S.D., Goldberg, G.I. and Welgus, FLG. (t991) Human 92- and 72-kilodalton type IV collagenases are elastases. J. Biol. Chem. 266, 7870-5. 11 Wilhelm, S.M., Collier, I.E., Kronberger, A., Eisen, A.Z., Mariner, B.L., Grant, G.G., Bauer, E.A. and Goldberg, G.I. (1987) Human skin fibroblast stromelysin: Structure, glycosylation, subs~rate specificity, and differential expression in normal and tumourigenic cells. Proc. Natl. Acad. Sci. USA 84, 6725- 9. 12 Muller, D., Quantin, B., Gesnet~ M.C., Millon-CoUard, R., Abecassis, J. and Breathnach, R. (1988) The collagenase gene family in humans consists of at least four members. Biochem. J. 253, 187-92. 13 Basset, P., BeUocq,J.P., Wolf, C., Stoll, L, Hutin, P., Limacher, J.M., Podhajcer, O.L., Chenard, M.P., Rio, M.C. and Chambon, P. (1990) A novel metalloproteinase gene specifically expressed in stromal cell of breast carcinomas. Nature 348, 699-704. 14 Pei, D., Majmudar, G. and Weiss, S.J. (1994) Hydroly'dc inactivation of a breast carcinoma cell-derived serpin by human stromelysin-3. J. Biol. Chem. 269, 25849-55. 15 Quantin, B., Murphy, G. and Breathnach, R. (1989) Pump-1 cDNA codes for a protein with characteristics similar to those of classical collagenase family members. Biochemistry 28, 5325-34. 16 Shapiro, S.D. Kobayashi, D.K. and Ley, T.J. (1993) Cloning and characterisation of a unique elastolytic metalloproteinase produced by human alveolar macrophages. J. Biol. Chem. 268, 23824-9. 17 Sato, H., Takino, T., Okada, Y., Cao, J., Shinagawa, A~, Yamamoto, E. and Seiki, M. (1994) A matrix metalloproteinase expressed on the surface of invasive turnout ceils. Nature 370, 61-5. 18 Will, H. and Hinzmann, B. (1995) cDNA sequence and mRNA tissue distribution of a novel human matrix metaLloproteinase with a potential tTansmembrane segment. Eur. J. Biochem. 231, 602-8. 19 Takino, T., Sato, H., Shinagawa, Ao and Seiki, M. (1995) MEDICAL ONCOLOGY (1997) 14 (1)

BROWN

Identification of the second membrane-type matrix metalloproteinase (MT-MMP2) gene from a human placenta cDNA library. MT-MMPs form a unique membrane-type subclass in the MMP family. J. Biol. Chem. 270, 23013- 20. 20 Puente, X.S., Pendas, A.M., Llano, E., Velasco, G. and LopezOfin, C. (1996) Molecular cloning of a novel membranetype matrix metalloproteinase from a human breast carcinoma. Cancer Res. 56, 944--9. 21 Strongin, A., Collier, I., Bannikov, G., Mariner, B.L., Grant, G.A. and Goldberg, G.L (1995) Mechanism of cell surface activation of 72 kDa ~ p e 1V collagenase. J. Biol. Chem. 270, 5331 - 8. 22 Cossins, J., Dudgeon, TJ., Catlin, G., Gearing, AJ.H. and Clements, J.M. (1996) Identification of MMP-18, a putative novel human matrix metatloproteinase. Biochem. Biophys. Res. Commun. 228, 494-8. 23 Docherty, AJ.P., Lyons, A., Smith, BJ., Wright, E.M., Stephens, P.E. and Harris, T.J.R. (1985) Sequence of human tissue inhibitor of metalloproteinases and its identity to erythroid-potentiating activity. Nature 318, 66-9. 24 Stefler-Stevenson, W.G., Krutzsch, H.C. and Liotta, L.A, (1989) Tissue inhibitor of metalloproteinase (TIMP-2). A new member of the metalloproteinase inhibitor family, f. Biol. Chem. 264, 17374-8. 25 Apte, S.S., Matteir M.G, and Otsen, B.R. (1994) Cloning of the cDNA encoding human tissue inhibitor of metalloproteinases-3 (TfMP-3) and the mapping of the TIMP-3 gene to chromosome 22. Genomics 19, 86-90. 26 Greene, J., Wang, M.S., Liu, Y.L.E., Raymond, L.A., Rosen, C. and Shi, Y.N.E. (1996) Molecular cloning and characterisation of human tissue inhibitor of metalloproteinase 4. ]. Biol. Chem. 271, 30375-NL 27 Kleiner, DJ. and Stetler-Stevenson, W.G. (1993) Structural biochemistry and activation of matrix metalloproteinases. Curt. Op. Cell Biol. 5, 891- 7. 28 Powell, W.C. and Matrisian, L.M. (1996) Complex roles of matrix metalloproteinases in tumour progression. Curt. Top. Microbiol. ]mmun. 213, 1-21. 29 van der Stappen, J.W.J., Hendriks, T. and Wobbes, T. (1990) Correlation between collagenolytic activ@ and grade of histological differentiation in colorectal tumours. Int. J. Cancer 45, 1071- 8. 30 Hewitt, R.E., Leach, I.H, Powe, D.G., Clark, I.M., Cawston, T.E. and Turner, D.IL (1991) Distribution of collagenase and tissue inhibitor metatloproteinases (TIMP) in colorectal tumours, blL jr. Cancer 49, 666-72. 31 Murray, G.I., Duncan, M.E., O'Neil, P., Melvin, W.T. and Fothergill, J.E. (1996) Matrix metalloproteinase-1 is associated with poor prognosis in colorectal cancer. Nat~lre Med. 2, 461- 2. 32 Newell, ICJ., Witty, J.P,, Rodgers, W.H. and Matrisian, L.M. (1994) Expression and localisation of matrix-degrading metaUoproteinases during colorectal tumourigenesis. MoL Carcin. 10, 199-206. 33 Gallegos, N.C., Smales, C., Savage, F.G., Hembry, R.M. and Boulos, P.B. (1995) The distribution of matrix metaLloproteinases and tissue inhibitor of metalloproteinases in colorectal cancer. Surg. Oncol. 4, 21- 9. 34 Levy, A.T., Cioce, V., Sobel, M.E., Garbisa, S., Grigioni, W.F., Liotta, L.A. and Stefler-Stevenson, W.G. (1991) Increased expression of the Mr 72,000 type IV collagenase in human colonic adenocarcinoma. Cancer Res. 51, 439-44. 35 Liabakk, N-B., Talbot, I., Smith, R.A., Wilkinson, K. and Balkwill, F. (1996) Matrix metalloproteinase 2 (MMP-2) and matrix metalloproteinase 9 (MMP-9) type W collagenases in colorectal cancer. Cancer Res. 56, 190-6. 36 Brown, P.D., Bloxidge, R.E, Smart, N.S.A., Gaiter, K.C. and Carmichael, J. (1993) Correlation between expression of activated 72-kDa gelatinase and tumour spread in nonMEDICAL O N C O L O G Y (1997) 14 (1)

small cell lung carcinoma. J. Natl. Cancer Inst. 85, 574-8. 37 Emmert-Buck, M.R., Roth, M.J., Zhuang, Z., Campo, E., Rozhin, J., Sloane, B.F., Liotta, L.A. and Stetler-Stevenson, W.G. (1994) increased gelatinase A (MMP-2) and cathepsin B activity in invasive tumor regions of human colon cancer samples. Am. J. Pathol. 145, 1285-90. 38 Zeng, Z.S. and Guillem, J.G. (1995) Distinct pattern of ma~Lx metalloproteinase 9 and tissue inhibitor of metalloproteinase 1 mRNA expression in human colorectal cancer and liver metastases. Br. J. Cancer 72, 575- 82. 39 Zeng, Z.S., Huang, Y., Cohen, A.M. and Guillem, J.G. (1996) Prediction of colorectal cancer relapse and survival via tissue RNA levels of matrix metalloproteinase-9, f. Clin. Oncol. 14, 3133-40. 40 Zucker, S, Lysik, R.M~ Zarrabi, M.H. and Moll, U. (1993) Mr92,000 ~Tpe IV collagenase is increased in plasma of patients with colon cancer and breast cancer. Cancer Res. 53, 140-6. 41 Yoshimoto, M., Itoh, F., Yamamoto, H., Hinoda, Y., Imai, K. and Yachi, A. (1993) Expression of MMP-7 (Pump-l) mRNA in human colorectal cancers. Int. J. Cancer 54, 614- 8. 42 Mori, M., Barnard, G.F., Mimori, K., Ueo, H., Akiyoshi, T. and Sugimachi, K. (1995) Overexpression of matrix metalloproteinase-7 mRNA in human colon carcinomas. Cancer 75, 1516- 9, 43 Ishikawa, T., Ichikawa, Y., Mitsuhashi, M., Momiyama, N., Chishima, T., Tanaka, K., Yamaoka, H., Miyazakic, K~ Nagashima, Y., Akitaya, T. and Shimada, H. (1996) Matrilysin is associated with progression of colorectal tumor. Cancer Lett. 107, 5-10. 44 Porte, H., Chastre, E., Prevot, S., Nordlinger, B., Empereur, S., Basset, P., Chambon, P. and Gespach, C. (1995) Neoplastic progression of human colorectal cancer is associated with overexpression of the stromelysin-3 and BM-40/SPARC genes. Int. ]. Cancer 64, 70-5. 45 Zeng, Z, Cohen, A.M., Zhang, Z., Stetler-Stevenson, W.G. and Guillem, J.G. (1995) Elevated tissue inhibitor of metalloproteinase 1 RNA in colorectal cancer stroma correlates with lymph node and distant metastases. Clino Cancer Res. 1, 899-906. 46 Kossakowska, A.E., Huchcroft, S.A., Urbanski, S.J. and Edwards D.R. (1996) Comparative analysis of the expression patterns of metalloproteinases and their inhibitors in breast neoplasia, sporadic colorectal neoplasia, pulmonary carcinomas and malignant non-Hodgkin's lymphomas in humans. Br. J. Cancer 73, 1401- 8. 47 Yoshiji, H., Gomez, D.E. and Thorgeirsson, U.P. (1996) Enhanced RNA expression of tissue inhibitor of metalloproteinases-1 (TIMP-1) in human breast cancer. Int. J. Cancer 69, 131-4. 48 Grignon, D.J., Sakr, W., Toth, M, Ravery, V., Angulo, J., Shamsa, F., Pontes, J.E., Crissman, J.C. and Fridman, R. (1996) High levels of tissue inhibitor of metalloproteinase2 (TIMP-2) expression are associated with poor outcome in invasive bladder cancer. Cancer Res. 56, 1654-9. 49 Alderson, M.R., Hamlin, L and Staunton, M.D. (1971) The relative significance of prognostic factors in breast carcinoma. Br. [. Cancer 25, 646- 56. 50 Anastassiades, O.T. and Pryce, P.M. (1974) Fibrosis as an indication of time in infiltrating breast cancer and its importance in prognosis. Br. J. Cancer 29, 232-9. 51 Carlomagno, C., Perrone, F., Lauria, R, De Laurentiis, M., Gallo, C., Morabito, A, Pettinato, G., Panico, L., Bellelli, T., Apicella, A., Petrella, G., Bianco, A.R. and De Placido, S. (1995) Prognostic significance of necrosis, elastosis, fibrosis and inflammato~ cell reaction in operable breast cancer. Oncology 52, 272-7. 52 Parham, D.M., Hagen, N. and Brown, R.A. (1992) Simplified method of grading primary, carcinomas of the breast. ]. Clin. Pathol. 45, 517-20.

10 53 Takei, H., Iino, Y., Horignchi, J. and Yakoe, T. (1995) Immunohistochemical fibronectin staining pattern and prognosis in invasive breast carcinoma. Ontology 52, 10611. 54 Halvorsen, T.B. and Seim, E. (1989) Association between invasiveness, inflammatory cell reaction, desmoplasia and survival in colorectal cancer. J. Clin. PathoL 42, 162-6. 55 Yamauchi, H., Kikuchi, Y. and Kakizaki, K. (1995) Relation between the stromal volume and liver metastasis in ducs cell carcinoma of the pancreas. Tohoku f. Krp. Med. 177, 179-81. 56 Leek R.D, Lewis, C.E., Whitehouse, R., Greenall, M., Clarke, J. and Harris, A.L. (1996) Association of macrophage infiltration with angiogenesis and prognosis in invasive breast carcinoma. Cancer Res. 56, 4625- 9. 57 Pupa, S.M., Bufallno, R., Invernizzi, A.M., Andreola, S., Rilke, F, Lombardi, L., Colnaghi, M.I. and Menard, S. (1996) Macrophage infiltrate and prognosis in c-erB-2overexpressing breast carcinomas. ]. Clin. Oncol. 14, 85- 94. 58 Becket-t, R.P., Davidson, A.H., Drummond, A.H., HtLxley, P. and Whittaker, M. (1996) Recent advances in matrix metalloproteinase inhibitor research. Drug. Dev. Today 1, 16-26. 59 Gearing, AJ.H., Beckett, P., Christodoulou, M., Churchill, M., Clements, J., Davidson, A.H., Drummond, A.H., Galloway, W.A., Gilbert, R, Gordon, J.L., Leber, T.M., Mangan, M., Miller, K., Nayee, P., (3wen, K., Patel, S., Thomas, W., Wells, G., Wood, L.M. and Woolley, K. (1994) Processing of tumour necrosis factor-.~ precursor by metalloproteinases. Nature 370, 555-7. 60 Reich, R., Thompson, E~ Iwamoto, Y., Martin, G.R., Deason, J.R., Fuller, G.C. and Miskin, R. (1988) Effects of inhibitors of plasminogen activator, serine proteinases, and collagenase IV on the invasion of basement membranes by metastatic cells. Cancer Res. 48, 3307-12. 61 Schultz, ICM, Silberman, S., Persky, B., Bajkowski, A.S. and Carmichael, D.F. (1988) Inhibition by human recombinant tissue inhibitor of metalloproteinases of human amnion invasion and lung colonization by murine B16-F10 melanoma cells. Cancer Res. 48, 5539-45. 62 Ecdes, S.A., Box, G.M., Court, W.J, Bone, E.A., Thomas, W. and Brown, P.D. (1996) Control of lymphatic and hematogenous metastases of a rat mammary, carcinoma by the matrix metalloproteinase inhibitor batimastat (BB94). Cancer Res. 56, 2815-22. 63 Sledge, G.W., Qulalg M., Goulet, R., Bone, E.A. and Fife, R. (1995) Effect of matrix metalloproteinase inhibitor batimastat on breast cancer regrowth and metastasis in athymic mice. f. Natl. Cancer Res. 87, 1546-50.

MATRIX METALLOPROTEINASE tNHIBITORS 64 DeClercK Y.A., Perez, N., Shimada, H., Boone, T.C., Langley, K.E. and Taylor, S.M. (1992) Inhibition of invasion and metastasis in cells transfected with an inhibitor of metalloproteinases. Cancer Res. 52, 701-8. 65 Johnson, M.D., Kim, H.R., Chesler, L., Tsao-Wu, G., Bouck, N. and Polverini, PJ. (1994) Inhibition of angiogenesis by tissue inhibitor of metalloproteinase. [. Cell Physiol. 160, 194-202. 66 Taraboletti, G., Garofalo, A., Belotti, D., Drudis, T., Borsotti, P., Scanziani, E., Brown, P. and Giavazzi, R. (1995) Inhibition of angiogenesis and murine hemangioma growth by batimastat, a synthetic inhibitor of matrix metalloproteinases. ]. Natl. Cancer Inst. 87, 293-8. 67 Anderson, I.C., Shipp, M.A., Docherty, AJ.P. and Teicher, B.A. (1996) Combination therapy including a gelatinase inhibitor and cytotoxic agent reduces local invasion and metastasis of routine Lewis lung carcinoma. Cancer Res. 56, 715 - 20. 68 Beckett, R.P. (1996) Recent advances in the field of matrix metalloproteinase inhibitors (patent update). Exp. (?pin. 77~c~r. Patents 6, 1305-15. 69 Galardy, R.E., Cassabone, M.E., Giese, C., Gilbert, J.H., Lapierre, F., Lopez, H. et aL (1994) Low molecular weight inhibitors in corneal ulceration. Ann. N.Y. Acad. Sci. 732, 31.5- 23. 70 Macaulay, V.M., O'Byrne, KJ., Saunders, M.P., Salisbury, A., Long, L., Gleeson, F., Ganesan, T.S., Harris, A.L. and Talbot, D.C. (1995) Phase I study of matrix metalloproteinase (MMP) inhibitor batimastat (BB-94) in patients with malignant pleural effusions. Br. J. Cancer 71, 11 (Abstract). 71 Primrose, J., Bleiberg, H,, Daniel, F., Johnson, P., Mansi, J., Neoptolemos, J., Seymour, M. and Van Belle, S. (1996) A dose-finding study of marimastat, an oral matrix metalloproteinase inhibitor, in patients with advanced colorectal cancer. Ann. Oncol. 7, 35 (Abstract). 72 Poole, C., Adams, M., Barley, V., Graham, J., Kerr, D., Louviaux, I., Perren, T., Piccart, M. and Thomas, H. (1996) A dose-finding study of marimastat, an oral matrix metalloproteinase inhibitor, in patients with advanced ovarian cancer. Ann. Oncol. 7, 68 (Abstract). 73 Millar, A. and Brown, P. (1996) 360 patient meta-analysis of studies of marimastat, a novel matrix metalloproteinase inhibitor. Anra Oncot. 7, 123 (Abstract). 74 Parsons, S.L., Watson, S.A., Griffin, N.R. and Steele, RJ.C. (1996) An open phase I/II study of the oral matrix metaUoproteinase inhibitor mafimastat in patients with inoperable gastric cancer. Ann. Oncol. 7, 47 (Abstract).

MEDICAL O N C O L O G Y (1997) 14 (1)