UNIVERSITÀ DI PISA
Dipartimento di Farmacia
Corso di Laurea Specialistica in Chimica e Tecnologia
Farmaceutiche
Tesi di Laurea
:
Synthesis and inhibitory properties of potent and
selective inhibitors of ADAM-17 and MMP-8
Relatori: Candidata:
Prof. Armando Rossello Alessandra Paolini
Dott.ssa Elisa Nuti
SSD: CHIM/08
2
Secretazione dell’elaborato scritto della Tesi di Laurea
“Il contenuto di questa Tesi di Laurea è strettamente riservato, essendo presenti argomenti tutelati dalla legge come segreti. Pertanto tutti coloro che ne prendono conoscenza sono soggetti all’obbligo, sanzionato anche penalmente dagli articoli 325 e 623 del codice penale, di non divulgare e di non utilizzare le informazioni acquisite.” *
Clause Secret
“The contents of this report is strictly confidential, as arguments protected by law as confidential. Therefore all those who take knowledge are required, even sanctioned fees to Articles 325 and 623 of the Criminal Code, and not to disclose not use the information gained.”
(These terms are in accordance with regulations of the University of Pisa in relation to the protection of intellectual property patents.)
Identificazione per presa visione *
3
Introduction
Chapter 1: The MMPs... 6
1.1 Introduction ... 6
1.2 Classification ... 7
1.3 Synthesis and activity regulation ... 8
1.4 Structure ... 8
1.5 Substrate specificity ... 10
1.6 Reaction mechanism ... 12
1.7 MMP-8 ... 12
Chapter 2: MMP-8 inhibitors ... 14
2.1 Design and synthesis ... 14
Chapter 3: The ADAMs ... 21
3.1 Introduction ... 21
3.2 Classification ... 22
3.3 Synthesis and structure ... 22
3.4 Substrates and activity ... 23
3.5 ADAM-17 ... 25
3.6 ADAM-10 ... 29
Chapter 4: ADAM-17 inhibitors ... 31
4
Experimental Part
Chapter 5: Experimental part introduction ... 37
5.1 Design and synthesis of selective MMP-8 inhibitors ... 37
Scheme 1. Synthesis of compounds 1 and 2 ... 40
5.2 Design and synthesis of selective ADAM-17 inhibitors ... 42
Scheme 2. Synthesis of compound 9 ... 45
5.3 Attempts to obtain ADAM-10 selective inhibitors ... 47
Scheme 3. Synthesis of compound 18 ... 48
Scheme 4. Synthesis of compound 24 ... 50
Scheme 5. Synthesis of compound 25 ... 52
Chapter 6: Synthesis ... 53
6.1 Materials and Methods ... 53
6.2 Synthesis of MMP-8 inhibitors. Scheme 1 ... 54
6.3 Synthesis of ADAM-17 inhibitors. Scheme 2 ... 59
6.4 Synthesis of ADAM-10 inhibitors. Scheme 3 ... 62
6.5 Synthesis of ADAM-10 inhibitors. Scheme 4 ... 65
6.6 Synthesis of ADAM-10 inhibitors. Scheme 5 ... 67
5
Introduction
Metalloendopeptidases (MEPs) are hydrolytic enzymes distributed among all kingdoms of life, they are ubiquitous and partecipate in proteolytic processing. Their targets are peptide bonds of proteins or oligopeptides. They catalyze processes like digestion or degradation of proteins and development, maintenance and remodeling of tissues. In particular they are also involved in specific and necessary proteolytic processes such as activation or inactivation of themselves or other (pro)enzymes, bioactive peptides or DNA repressors. For this reason is crucial the regulation of their activity, through the control at the transciptional level and via post-traslational modifications. If this control fails it can give rise to pathologies such as inflammation, tissue destruction, neurological diseases, cardiovascular disorders and even cancer. In particular, it has been underlined the crucial role of MEPs in tumorigenesis and tumorprogression events (angiogenesis, tissue invasion and metastasis). The majority of metalloendopeptidases are zinc-dependent, and hence are named zincins; their general structure shows: i) a short consensus sequence, HEXXH; ii) two histidines that are linked with the catalityc zinc and iii) the glutamate wich acts as a general base. Depending on the position of the third protein zinc ligand the zincins can be divided in: i) gluzincins, ii) aspzincins and iii) metzincins. The latter are a large family of MEPs and they include adamalysins (ADAMs), matrix metalloproteinases (MMPs), serralysins, snapalysins, and leishmanolysins. The metzincins are multidomain proteins with 130–260-residue globular catalytic domain showing a common core architecture characterized by zinc-binding consensus sequence (HEXXHXXXGXXH/D), which comprises three zinc ligands and glutamate. Morover, they display a conserved methionine-containing 1,4-β-turn, called Met-turn.1
6
Figure 1. Classification scheme of MEPs
Chapter 1: Matrix metalloproteinases (MMPs)
1.1 Introduction
Matrix metalloproteinases (MMPs), also called matrixins or vertebrate collagenases, are secreted or membrane-bound calcium dependent zinc-containing endopeptidases. They are involved in tissue remodeling and degradation of the extracellular matrix (ECM). The first MMP was discovered about 50 years ago as the agent implicated in tail resorption during the frog metamorphosis. Subsequently different mammalian enzymes were purified and identified through the techniques of molecular biology. Under normal physiological conditions they are minimally expressed and partecipate in tissue resorption, remodelling and repair, as observed during embryogenesis and development, organ morphogenesis and angiogenesis. Recently it has been demostrated their activity in sophisticated processes that go beyond mere ECM turnover. This action is linked to activation or inactivation of other proteins through a specific proteolysis. Substrates of this proteolysis consist of (pro-)proteases, protease inhibitors, clotting factors, antimicrobial peptides, chemotactic and adhesion molecules, growth factors, hormones, cytokines, like so their receptors and binding proteins.
7
Usually their potent proteolytic capacity is under fine-control but if this fails, it can give rise to pathologies as inflammation, ulcers, rheumatoid arthritis and ostearthritis, periodontitis, heart failure and cardiovascular disease, fibrosis, emphysema, cancer and metastasis. In the last few years, it has appeared their involvement in other diseases such as stroke, HIV-associated dementia, atherosclerosis, multiple sclerosis, bacterial meningitis and Alzheimer. Table 1
1.2 Classification
Until now, 26 human MMPs are known, they are numbered 1 to 3, 7 to 17, 19 to 21, and 23 to 29 for historical reasons and there are two identical forms for MMP-23, encoded by two genes, mmp-23a and mmpP-23b. On the basis of their specificity, these MMPs can be divided into: i) collagenases (MMP-1; -8; -13 and -18), which cut triple-helical collagen, ii) gelatinases (MMP-2 and -9), which target denatured collagens and gelatins, iii) stromelysins (MMP-3; -10; -11 and -27), which have broad specificity and may degrade proteoglycans and iv) matrilysins (MMP-7 and -26), which degrade
8
proteoglycans, fibronectin, elastin and casein. Moreover there are the membrane-type MMPs (MT-MMPs) and other enzymes that couldn’t be included in the previous classes.
1.3 Synthesis and activity regulation
MMPs are proteins excreted by several connective tissues and pro-inflammatory cells, such as fibroblats, osteoblasts, enothelial cells, macrophages, neutrophils and lymphocytes. They are synthesized as zymogens and processed by other proteolytic enzymes (serine proteases, furin, plasmin, etc.) to be activated. Their activity is regulated via modulation of gene expression, compartimentalization, removal of pro-domain and inhibition by protein inhibitors. After that they are released in extracellular space or anchored to the membrane and activated, they are held in check by α2 -macroglobulin, a general endopeptidase inhibitor, and by their endogenous tissue inhibitors (TIMPs). In particular, it has been discovered four TIMPs (numbered from 1 to 4) that bind MMPs in a 1:1 stoichiometry.2
They show a double action, they inhibit both the active form of MMPs and their activation. Recently it has been demostrated that the TIMPs partecipate in complex biologic functions, such as changes in cell morphologis and stimulation of growth in different cell. They are implicated even in cancer, their expression has been associated with a less aggressive tumor behavior.3
1.4 Structure
MMPs are mosaic proteins costituted by a sequence of inserts and domains. Most of them are costituted by four principal domains preceded by a short signal peptide for secretion. These domains are, from N- to C-terminus, a ∼80-residue zymogenic pro-peptide; a ∼165-residue zinc- and calcium-dependent catalytic domain; a ∼(15-65)-residue linker region and a ∼200-∼(15-65)-residue hemopexin-like domain. While MT-MMPs show an additional transmembrane domain that anchors them to the cell surface. The structure of pro-domain (known for MMP-2; -3 and -9) show three α-helices and connecting loops. After the last α-helix there is a long peptide region which contains the conserved cysteine switch, necessary to keep the zymogen inactive.
9
The structure proceeds with the catalytic domain, which has a sphere shaped with a diameter of ∼40 Å. On the front surface there is a shallow active-site cleft which accept the substrates, in an extended conformation. The orientation of the bond occurs horizontally from left (N-terminal side) to right (C-terminal side) of the catalytic metal ion. The fissure divide the molecule asymetrically in two parts: an upper N-terminal (NTS, 127 residues on average) and a lower C-terminal sub-domain (CTS; 37 residue on average). There aren’t disulfide bonds in any MMPs but some of them, like MMP-8, show a salt-bridge between the N-terminal α-amino group of the mature enzymes and the Asp232, which is at the beginning of α-helix αC, and this make the MMP in a “superactive” form. The chain consist of: i) five-stranded twisted β-sheet; ii) three α-helices and iii) connective loops. As shown in figure 2, all strands are parallel except one and the two α-helices (αA, the “backing helix”, and αB, the “active-site” helix) are into the concave side of the sheet. The loop that connect strands βIII and βIV is called the “S-loop” and meanders around two ion-binding sites: the catalytic zinc and a calcium ion. While the structural zinc ion is tetrahedrally co-ordinated by three histidines and, monodentately, by an aspartate (His147, His162, His175, and Asp149). After the last strand there is a large loop, which is characteristic for each MMPs and establishes to substrate specificity. After, the chain continues with the active-site helix, αB, which includes the first half of the zinc-binding consensus sequence, HEXXHXXGXXH, that is conserved among all metzincins. The last part of the protein is formed by CTS, called the hemopexin-like domain,4 that consists of a tight 1,4-turn of type I, called the “Met-turn” and the helix αC.
10
Other possible insertions include: fibronectin type-II-related domains, only found in gelatinases; a collagen type-V-like insert; a vitronectin-like insertion domain; a cysteine-rich, proline-rich and interleukin-1 receptor-like domain; an immunoglobulin-like domain; a glycosyl hosphatidylinositol linkage signal; a type-I or type-II trans-membrane domain; a stem or “linker 2” region downstream of the hemopexin-like domain; and a cytoplasmic tail.5
In particular, the active-site cleft is formed by strand βIV, the S-loop, α-helix B and the extended loop after αB. The active-site zinc is coordinated by three histidines (His197, His201 and His207) and a solvent molecule that, generally, is a water molecule. Other elements are present to support the catalysis as the Met-turn-residue and a glutammate which assistes the hydrolysis by its carboxyl group.
On the right to this cleft there is a pocket, called specificity pocket or S1’ pocket, that is important for the specificity of the bond with the substrate and it changes among the MMPs. Infact, the volume of the pocket varies from a small hydrophobic site in MMP-7 to a large site in MMP-8.
1.5 Substrate specificity
It has been demostrated that MMPs are able to recognise the exact substrate to cleave. Most of them hydrolyse a peptide bond before a residue with a hydrophobic side chain, such as Leu, Ile, Met, Phe or Tyr. Additionally, the S1’ pocket and other subsite (S1, S2, S3, S2’ and S3’) play an important role on the substrate specificity among MMPs. The S1’ pocket is highly selective for the P1’ residue of the substrate wich is immediatelly after the scissile bond. For example in MMP-1 and MMP-7 the small S1’ pocket interacts preferentially with small hydrophobic residues at P1’. While, other MMPs, as MMP-8 and MMP-9, show a large S1’ pocket that can receive different substrates. Further there are susbstrate binding sites located outside the catalytic domain, the “Exosites”, even them implicated of substrate recognition. They are involved in the preparation of the substrate prior to cleavage, increasing the affinity of MMPs for this and presenting the substrate scissile bond in the optimal orientation to the active site.6
11
12
1.6 Reaction mechanism
That reaction consists of nucleophilic attack (figure 3). The scissile amide carbonyl, coordinated to the active-site zinc(II) ion, is attacked by a solvent molecule (in general water, that is polarizated by interactions both with a conserved glutamic acid and with the zinc(II) ion).
Figure 3. Reaction mechanism for the proteolysis by MMPs.26
The water molecule acts as donator of proton to the nitrogen of the scissile amide and both the zinc(II) ion and the conserved alanine residue help to stabilize the charges that form.
1.7 MMP-8
Matrix metalloproteinase 8 is also known as collagenase-2 or neutrophil collagenase. It is a member of the collagenases because it cleaves the triple helix structure of native collagen. Originally MMP-8 was considered to be released only by neutrophils in a highly glycosylated form. After, other MMP-8 species have been identified with a smaller molecular mass and these are produced from different cell types and are secreted in the extracellular environment after synthesis. MMP-8 is essentially coinvolted in a exstensive range of inflammatory disorders, such as asthma, acute lung injury, periodontal disease, atherosclerosis, keratitis, bacterial meningitis, rheumatoid arthritis, liver fibrosis, hepatitis and cancer. It has been underlined as the progression of these diseases is associated with the upregulation of the collagenase-2.
13
But this MMP has a double action, infact it has been described both a pro-inflammatory and anti-inflammatory action. MMP-8 activity is linked to some lung diseases because it is the most potent collagenase that degrades collagen type I, highly present in lung ECM. At the begging of lung injury it has a pro-inflammatory action but during the last stages of disease it could be involved in clearing inflammatory cells from the site of infection. Even in the case of periodontal disease it has been underlined the MMP-8 action. Periodontitis is a common chronic infectious of the oral cavity that involves the non-reversible soft tissue and bone destruction, and recently it has been linked to systemic illnesses such as cardiovascular disease. An other example in which MMP-8 is involded, is the development of atherosclerotic plaques. The collagen type I is the major structural element of the plaque, an uncrontrollable destruction of it could be leaded to the plaque rupture. Further, the levels of MMP-8 could be used to predict the progression of above diseases. For example measuring its concentration in human fluid is a non-invasive and sensitive biomarker.7 The same dual role of this protease is present in tumourgenic and antitumourgenic properties linked to MMP-8 and, in particular, depending on the type of tumour and its stage of development and progression. Just before 1970, MMPs were associated with cancer and different studies shown the presence of this enzymes in the metastasis. Unfortunately the first clinical trials with MMP inhibitors have produced negative results, because they block unspecifically MMPs. Thus, it has been necessary re-evaluated the role of proteases in cancer which is more complex. It is known as the degradation of ECM is the first step in tumour invasion and metastasis, and it is correlated with the MMP activity. But Balbin et al. have demostrated the anti-tumour properties of MMP-8, using a MMP-8-/-mice for the experiments, and after, other studies have been underlined the protective role of collagenase-2 in cancer. Molecular and histopathological studies have indicated that the loss of this protease causes a wrong inflammatory response that creates a favorable microenvironment for cancer progression. Recent experiment using MMP-8-/-mice have demostrated that this collagenase acts as a metastasis suppressor via the modulation of tumor cell adhesion and invasion.8 Recent studies on human genome suggest that MMP-8 is a tumor suppressor gene, in the experiment it has been compared the effects of wild type MMP-8 on melanoma metastasis with control cells.
14
Finally it has been demostrated that MMP-8 has a protective role. For all these reasons, there is great attention on the synthesis of specific inhibitors which could potentially be used in therapy.
Chapter 2: MMP-8 inhibitors
2.1 Desing and synthesis
In the last years, great interest has been focused on the development of matrix metalloproteinase inhibitors (MMPIs), since MMPs participate in the remodelling of the extracellular matrix (ECM) and their proteolytic action can lead to diseases such as arthritis, multiple sclerosis, cancer, etc. These multidomain proteins have some parts in common (the signal domain, the propeptide domain and the catalytic domain). All MMPs contain a Zn ion coordinated by a tris(histidine) motif, in that catalytic site, which is important both substrate binding and cleavage. Near the active site there are subsite pockets designated as S1, S2, S3, S1’ and S2’, in particular S1’ pocket is important for the recognition of substrate and provides the basis of selectivity for many MMPIs. Unfortunately, first inhibitor (for example marimastat and prinomastat figure 19) have had no fortune and the use of broad-spectrum MMPIs in clinical trials has caused undesired side effects. In addition, it is important to underline that not all MMPs have the same role in the pathological diseases. Recently, it has been demonstrated that some MMPs could have a protective action toward the cancer (MMP-8). For this is most important to synthesize selective inhibitors. In general, most MMPIs consist of two parts: a zinc-binding group (ZBG) to bind the catalytic metal ion and a peptidomimetic backbone, usually an elongated hydrophobic substituent as byphenyl, to interact noncovalently with specific subsites neighbouring the active site of the protein. Coehn et al. have demonstrated that both the ZBG and the backbone contribute to selective MMP inhibition, indeed small changes in the ZBG modulate the selectivity against deep S1’ pocket MMPs. Moreover from their studies emerged the different size of S1’ pocket, some MMPs show a shallow pocket while other, as MMP8 and MMP12, have a deep subsite.
15
In addition, they illustrate that the residues surrounding the Zn ion in the above MMPs are identical, and this could explain the similar inhibition trends.9,10
Hydroxamate is considered the most effective ZBG because it is able to efficiently chelate the catalytic zinc ion, forming five-membered chelates and two additional H-bonds with the enzyme. First of these inhibitors, however, were lacking of selectivity, showed poor pharmacokinetic properties and caused toxicity resulting from metabolic activation of hydroxylamine.11 Because of these problems, alternative ZBG have been investigated, in particular compounds that show a non-hydroxamate zinc-chelating groups. Tucker et al., in 1999, published the crystal structure of human neutrophil collagenase (HNC) complexed with a prime-side inhibitor, the (R)-2-(biphenyl-4-ylsulfonyl)-1,2,3,4-tetrahydroisochinolin-3-carboxylic acid (figure 4).
Figure 4. (R)-2-(biphenyl-4-ylsulfonyl)-1,2,3,4-tetrahydroisochinolin-3-carboxylic acid.
This simple D-Tic derivate (TetrahydroIsochinolin carboxylic acid, TIC) attracted their attention because of its nanomolar inhibitory activity against HNC. After that, the docking calculations had been validated, they have performed a large series of Tic derivates in the catalytic domain of MMP-8. The crystal structure of the complex MMP8:TIC shows that the axial carboxylate group of the inhibitor ligates the catalytic zinc ion in an asymmetric bidentate mode.
16
Figure 5. Schematic representation of the binding interactions (dashed lines) of the TIC inhibitor at the
active site of MMP-8. The distances are reported in Å.
The D-Tic residue protrudes out of the binding region and is mainly exposed to the solvent, it covers the cavity of the catalytic zinc ion and the biphenyl group is properly directed by the sulfonamide junction into the deep primary specificity pocket S1’ where it is completely buried. Moreover, from the data, it shows how the sulfonamide inhibitors induce conformational changes of the S1’ entrance which allow to accommodate the biphenyl group into the S1’ pocket. Even the cation-aromatic interactions are important because they contribute to make stable the complex MMP8:inhibitor and could be taken into account in the design of MMP inhibitors.12 Mazza et al. have reported the activity and mode of binding of two nonzinc chelating inhibitors of human neutrophil collagenase, compounds 1 and 2.
17
Figure 6. Chemical structure of 1 and 2.13
Both ligands inhibit MMP-8 with activity in the low nM range and are highly selective versus several other MMPs. Their high selectivity profile can be attributed to the conformational change induced to the S1′ loop, which is known to present a considerable variation in the size, shape and composition amongst the MMPs, while the data confirm that the Zn coordination is unchanged.
Table 3. IC50 [nM] of 1 and 2 against MMPs.
Compound MMP1 MMP2 MMP8 MMP9 MMP13 MMP14 1 >10000 >2500 57 >10000 <25 >10000 2 >10000 >10000 7.4 >10000 <25 >10000
The central scaffold of the inhibitors is formed by two fused-ring systems separated by a methylencarboxamide group, and it presents similar size and shape. The main differences are given by the squaramide moiety bound to one terminal end of 1 and by the carbamoylmethyl substituent, attached to the asymmetric carbon atom at the opposite end of 2. The mode of binding is similar for both inhibitors.
18
They are deeply inserted into the primary specificity pocket S1′, where they adopt an extended conformation with the mean planes of the two fused-ring systems almost perpendicular to each other, whereas the different binding interactions involve the terminal parts of the two inhibitors. Even in this case their high selectivity profile can be attributed to the conformational change induced to the S1’ pocket.13
As said above, the size of the S1’ pocket varies among MMPs and this is one of the main determining factors of substrate specificity, furthermore, the S1’ loop differs from other MMPs with two residues Arg222 and Tyr227, which are useful to design specific MMP8 inhibitors. Starting from the compounds 1 and 2, Kalva-Vinod-Saleena have generated a common structure-based pharmacophore model, based on the receptor-ligand interactions exhibited by these two ligands in the receptors’ active site.
The docking analysis of this compound shows the intermolecular interactions with the S1’ loop of MMP8, in particular the carbonyl oxygen and secondary nitrogen groups made strong H-bond interactions with Ala 220 and Arg 222. The presence of aromatic rings at both sides of the compounds assisted strong hydrophobic interactions with Tyr 219 and Tyr 227, which favour deep binding in the S1’ pocket of MMP-8. All these interactions are important to promote the selective inhibition of the MMP-8 enzyme. Moreover there is no interaction with the zinc atom. Between all compounds generated the 1-[2-(3-Bromo-4-anisoyl)hydrazino]-2-(p-nitrophenoxy)-1-ethanone is resulted the best one.
19
Throughout the simulation, the H-bond with Arg 222 allows to this compound to interact with the S1’ loop region of MMP-8 protein, which indicates that the complex is stable.14 Finally, a series of nonpeptidic 2-(arylsulfonyl)-1,2,3,4-tetrahydroisoquinoline-3-carboxylates and –hydroxamates and other bicyclic rigidified analogues were synthesized as MMP8 inhibitors.13 As zinc complexing functionalities hydroxamates and carboxylates were selected, although the latter do not provide the optimal geometry for strong zinc interactions. This led to the design of rigid 2-(arylsulfonyl)-
1,2,3,4-tetrahydroisoquinoline-3-carboxylates and corresponding hydroxamates.
Figure 8. Schematic interaction (biphenilsulfonyl)-1,2,3,4-tetrahydroisoquinoline-3-carboxylates . (RCBS)
A hydrogen bond involving Leu160 and Ala161 NH’s is a common motif, where a sulfonamide involving the tetrahydroisoquinoline nitrogen provides ideal complementarities plus a convenient synthetic route to a number of variations to probe S1’ requirements. The incorporation of the aromatic side chains of Tyr219 and His197 at opposite sides of the S1’ wall suggested that another favourable contribution to binding energy could result from protein-ligand aromatic-aromatic interactions. The final design consideration was to fill the space occupied by crystallographic water molecules in MMP-8 structures in S1’ by bulky hydrophobic substituents.
20 Table 4. Compound R1 R2 R3 R4 IC50 1 NHOH 4-Phenoxybenzene H H 2 2 NHOH Hexadecil H H 3000 3 NHOH 4-Methoxybenzene H H 4 4 NHOH 4-Biphenyl H H 2 5 NHOH 4-Methoxybenzene OH H 1 6 COOH 4-Phenoxybenzene H H 9 7 COOH 4-Biphenyl H H 10
From the table, it is proved that the carboxylic acids are less active than the hydroxamates. A steric bulk in the inhibitor molecules increases biological activity but bulky substituents could lead to negative results.13
21
Chapter 3: The ADAMs
3.1 Introduction
In addiction to MMPs, another subclass of the metzincins is represented by adamalysins. The ADAMs (A Disintegrin And Metalloproteases), also called reprolysins or MDCs (metalloprotease-like, disintegrin-like and cysteine-rich proteins), are transmembrane and secreted proteins of about 750 amino acids in leght, and most of them are extracellular type-I transmembrane glycoproteins. Their acronym evokes their modular structure and makes a fascinating allusion to the initially characterization in both hemorrhagic and non-hemorrhagic reptilian poisons and as sperm proteins associated with fertility (Wolfsberg et al. 1995). Later studies have led to an additional subdivision of the adamalysin family in: i) snake venom metalloproteinases (SVMPs), ii) ADAMs and iii) ADAMS with thrombospondin motifs (ADAMTS).15,16 The domain structure of ADAMs is responsible for their proteolytic, adhesive, and putative signalling activities. Some ADAMs are found predominantly at the cell surface whereas others appear to reside primarily in an intracellular compartment (presumably the Golgi apparatus).
ADAMs with metalloprotease activity are unique among membrane, and especially cell surface, and they have two seemingly antithetical domains: an adhesive domain (disintegrin domain) and a degradative domain (metalloprotease domain). In general, their activity is related to cell adhesion and proteolytic ectodomain release, also known as “shedding”, of several cell surface receptors and signaling molecules.15,16
The biological processes in which ADAMs are present, include sperm-egg interactions, cell fate determination in the nervous system, cell migration, axon guidance, muscle development, different aspects of immunity and inflammatory responses. And they are linked to pathological states, such as cancer, cardiovascular disease, asthma and Alzheimer’s disease.
22
3.2 Classification
To date 40 members of ADAM family have been found in the mammalian genome, and 21 of them have been described in the human genome. But only 12 of these human ADAM genes (ADAM 8, 9, 10, 12, 15, 17, 19, 20, 21, 28, 30 and 33) encode proteins which have the Zn-binding active site and show a proteolytic activity and this confirms the additional non-proteolytic roles of some ADAMs.15
3.3 Synthesis and structure
Most probably, ADAMs are synthesized in the rough endoplasmaic reticulum and matured in the Golgi compartment. As the MMPs, even ADAMs are produced as zymogens and, next to complex glycosylation, they are activated through removal prodomain by the convertase or furin.15
ADAMs present a modular structure in which it is possible to identify the prodomain; the metalloprotease domain; the disintegrin domain; the cysteine-rich and EGF-like domains; a transmembrane region and the cytoplasmic tail.
All members of ADAMs family, even MMPs, at their N-termini possess a signal sequence that direct them into a secretory pathway, and a prodomain which acts as an intramolecular chaperone facilitating the correct protein folding and maintaining the proteinase in a latent state via a cysteine-switch mechanism. And as told above, during the transit in the Golgi system the pro-domain is generally removed.
23
After this domain, there is the metalloprotease (MP) domain, common in the other metalloproteases. It is a globular structure and consits of five-stranded β-sheet, multiple α-helices and a conserved methionine residue, called Met-turn. Into this domain is located the catalytic Zn-binding site, with the conserved consequence HEXGHXXGXXHD and the three histidine residue.
At this point, the adamalysins show a disintegrin-like domain, which contains a 14-amino acid implicated in interactions between ADAMs and integrins, and it can be divided in two subdomains, described as the “shoulder” and “arm” domains, respectively Ds and Da. The consecutive Cys-rich domain show a N-terminus region called the “wrist” (Cw) domain that, with Ds and Da, make to the whole molecule the form of a C (Takeda et al. 2006). After, there is a “hand” Ch segment, which include the hypervariable region, involved in target recognition.
And the last domain is the cytoplasmic tail that is present only in the transmembrane ADAMs, it varies in length and sequence and can contain binding site motifs for SH3 domain-contining proteins and potential site for phosphorylation. It may play important roles in the regulation of protease function.
Moreover the cytoplasmic domain does indeed play a key role in coupling ADAM-17 and other ADAMs to specific signalling events such as GPCR activated ADAM-mediated EGFR ligand release, termed ‘‘triple membrane-passing signal” (TMPS).
3.4 Substrates and activity
In general, the main substrates of these proteases are the ectodomains of other transmembrane proteins. These include precursor forms of growth factors, cytokines, growth factor receptors, cytokine receptors and several different types of adhesion molecules.19
The most important fuction of some ADAMs are the proteolytic “shedding” of membrane-associated proteins, they are capable of cleaving transmembrane protein ectodomains adjacent to the cell membrane.
24
Figure 10. Ectodomain proteolytic release ADAM mediated.
Shedding of a membrane-anchored cytokine or growth factor by an ADAM proteinase could be relevant to various modes of signal transduction, including paracrine, autocrine and juxtacrine.16
Moreover, it has been fuond that this “shedding” activity can solubilize a number of cell adhesion molecules, including L-selectin, E-cadherin and N-cadherin, leading profound effects on cell-cell interactions and cell migrations, events most important in tumor development. These molecules are RIPing substrates, RIP means regulated intramembrane proteolysis, it consist of a sequential proteolytic cascade that starts with an ectodomain sheddase action on a transmembrane protein, followed by a subsequent cleavage within the membrane itself, which in many cases generates a signal-transducing intracellular domain fragment.
It has been found that ADAM-17 is involved in Notch signalling, cleveage by extracellular ADAMs creates a truncated transmembrane stub that becomes accessible for cleavege by an intramembrane-cleaving protease (I-CLiP), in particular the presenilins, a proteolytic component of γ-secretase. These cleavages generate an intracellular domain (ICD) that moves off to the nucleus and influences gene transcription (figure 11).
25
Figure11. The “shedding” and RIP.(18)
As for the MMPs, a major level of control occurs via their interaction with specific tissue inhibitors of metalloproteinases (TIMPs), of which there are four in mammals, and the inhibition occurs when it is formed a 1:1 non-covalent complex with the active enzyme. In general, several ADAMs are inhibited exclusively by TIMP-3, for example ADAM10, 12, 17, 28 and 33.18 Even the G-protein-coupled recepetors (GPCRs) regulate the ADAMs activity and hence act as mediators of EGFR transactivation.
Other levels of regulation can be the trafficking and cell membrane compartmentalization of the ADAMs.
Certainly, the last studies have been focused on development of inhibitor of these proteases because a dysregulation of them contribute to pathogenetic mechanisms of human disease. It has known their role in neurological and cardiovascular diseases, cancer, inflammation and immunologically response.
3.5 ADAM-17
ADAM-17 or TACE (tumor necrosis factors-converting enzyme), is a zinc-endopeptidase composed by a multidomain extracellular part, an apparent transmembrane helix and an intracellular C-terminal tail. The extracellular part comprises an N-terminal pro domain, a 259-residue catalytic domain, and a Cys-rich moiety that is composed of a disintegrin-like, and a crambin-like domain.
26
In comparison to other metzincins, the polypeptide chain of the TACE catalytic domain is clearly longer and is stable in the absence of calcium. Further, in contrast to the MMPs, TACE is relatively insensitive to the tissue inhibitor of metalloproteinases-1 (TIMP-1), but it is inhibited by TIMP-3.
The TACE catalytic domain (figure 12) has the shape of an oblate ellipsoid, notched at its flat side to give a relatively small active-site cleft. As in the other ADAMs, it consists of five-stranded β-pleated sheet, flanked on its convex side by a-helices hB and hB2 and on its concave side by helices hA and hC. The active-site cleft of TACE is relatively flat on the left-hand (nonprimed) side but becomes notched toward the right. The catalytic zinc residing in its center is coordinated with three histidines (His-405, His-409, and His-415) and with the Glu-406, which acts as a general base during catalysis. Moreover there are the conserved zinc-binding consensus motif HEXXHXXGXXH and the Met-turn. Immediately to the right of the catalytic zinc there is the medium-sized, essentially hydrophobic, S1’ specificity pocket and additionally to the right opens through a polar entrance a second hydrophobic (S3’) pocket, which merges inside the molecule with the S1’ pocket. The polypeptide topology and in particular the surface presentation of the catalytic zinc reveal the catalytic domain of TACE to be a typical metzincin. But some structural features of TACE are in common exclusively with the adamalysins: the long helix hB and the preceding multiple-turn loop arranged on top of the β-sheet; the typically arranged and shaped C-terminal helix hD; and the extended C terminus placed on the backside surface. However, the catalytic domain of TACE also differs from adamalysin in several aspects: i) with 259 residues, its chain is much longer and most of the additional residues of TACE are clustered, ii) the lack of an adamalysin-like calcium-binding site, iii) the S3’ pocket merged with S1’ and iv) the inverted charge pattern in and around the primed subsite.20
27
Figure 12.The ADAM-17 catalytic domain.20
TACE is one of the enzymes responsible for some processing events that include the release of cytokine factors, shedding of cell surface molecules, release of growth factors and cleavage of amyloid precursor protein (APP). Through several experiments has been demostrated that ADAM-17 is the major physiological TNF-α converting enzyme. Whereas from further studies it has been underlined the involvement of this enzymes in the processing of multiple growth factors of the EGF family of ligands, such as epidermal growth factor (EGF), (HB-EGF), amphiregulin and even TGF-α. Preliminary biochemical data suggests that different constructs of TACE are necessary for turnover of substrate to product, and this could demostrate how it chooses its target between all substrates. Another possibility could be that processing of certain substrates is dependent on where it is localized.23
All of the above ligands interact with the epidermal growth factor receptor (EGFR)/human EGFR (HER) family of receptors, and they are implicated in cancer development and progression. In particular the ADAM proteolytic activity involved in the formation and/or progression of cancer is related to the shedding of TNF-α and EGFR ligands, by ADAM-17. As previous enunciated ADAM-17 is a TNF-α converting enzyme. TNF-α is a pluripotent peptide with multiple activities potentially important in cancer, infact it is involved in the initiation or progression of skin, liver, intestinal, and ovarian malignancies.
28
It is able to up-regulate MMP expression, induce angiogenic factors, enhance cell migration, promote the epithelial-to-mesenchymal transition, induce expression of the transcription factor nuclear factor κB and induce reactive oxygen species that damage DNA.
Figure 13. Shedding of EGFR ligands.34
TACE is involved even in the shedding of EGFR ligands. In general the EGFR proteins are tyrosine kinases that mediate cell growth, cell survival, cell migration, angiogenesis and invasion. Their ligands (EGF, TGF-α and amphiregulin) are initially sinthesized as transmembrane precursors and can be released by ADAM-17 and this release can be activated by physiologic and pharmacologic stimuli.15 Recent investigations have demostrated that GPCRs (G-Protein-Coupled receptors) are able to utilize the EGFR as a downstream signalling partner in the generation of mitogenic signals.16 This cross-talk process is know as Triple Membrane-Passing Signal (TMPS) and it involves a metalloproteinase to release EGFR ligand. The broad relevance of EGFR signal transactivation in the development and progression of human cancer has been demonstrated and extended towards other pathophysiological diseases, including cystic fibrosis and cardiovascular diseases.23
29
Figure 14. Triple Membrane-Passing Signal (TMPS).35
3.6 ADAM-10
It is a Zn-metalloprotease with proteolytic activity, also know as mammalian Kuz from Kuzbanian (its Drosophila homolog). It is the first ADAM family member to be know for its proteolytic function. Its structure is similar to other ADAMs and contains the two principal domains: the metalloprotease and the disintegrin domain, with function of degradation and adhesion, respectively. High levels of this enzyme have been found in bone, cartilagine and brain.21 As metalloprotease, ADAM-10 contributes to ectodomain shedding, it has a lot of substrates, such as N-cadherin, E-cadherin, vascular endothelial VE-cadherin, type IV collagen, amyloid precursor protein (APP) and ephrin-A2.
ADAM-10 itself is also subject to regulated intramembrane proteolysis by two other ADAMs (ADAM-9 and -15) which were identified as the proteases responsible for releasing its ectodomain, and presenilin (gamma-secretase) responsible for the release of the ADAM-10 intracellular domain (ICD), that then translocates to the nucleus and is involved in gene regulation.25 It has been demostrate that ADAM-10 plays an essential role during neuronal development, by regulating L1-dependent neuronal cell adhesion, cell migration and neurite outgrowth.26
30
ADAM-10 is, also, the principal ADAM sheddase involved in Regulated Intramembrane Proteolysis (RIP), with important actions on Notch/Delta signalling and APP processing. From the studies it has been demostrated that the cleavage of APP by ADAM-10 happens to α-site, for this it is considered as α-secretase. This actions could be considered to be a protective factor in the etiology of Alzheimer’s disease.16,17
Figure 15. ADAM function in the brain. ADAM10 has been identified as the α-secretase which releases
the neuroprotective ectodomain of APP.24
As above mentioned ADAM-10, and also ADAM-17, modulates tumor progression because it’s able to regulate the activation of the EGFR tyrosine kinase family. In particular, it releases the EGF, betacellulin and even the Her2/ErbB2 (the human epidermal growth factor receptor-2) from the cell membrane, an event necessary for Her2 positive tumour cells to proliferate.24
31
Chapter 4: ADAM-17 inhibitors
4.1 Design and synthesis
Tumour Necrosis Factor (TNF)-α is a pleitropic, pro-inflammatory cytokine produced by monocytes, macrophages, neutrophils, T-cells, mast cells, epithelial cells, osteoblasts and dendritic cells. Over-expression of TNF-α is responsible for a number of pathological conditions like Crohn’s disease, ulcerative colitis, diabetes, multiple sclerosis, atherosclerosis, stroke and rheumatoid arthritis (RA). For this several diseases, some efforts have been made to develop inhibitors of TNF-α, actually there are three protein-based drugs approved by USFDA in the treatment of RA (Etanercept, Infliximab and Adalimumab). Successes of these biological agents as TNF-α blockers have inspired the researches to develop small molecules orally bioavailable TACE inhibitors, moreover it has also been demonstrated that inhibition of TACE by these molecules could be more effective than the biological agents cited above.
Figure 16. Action of anti-TNF biological agents and TACE inhibitors.27
TACE and MMP show a similar structure, both are zinc endopeptidases and are involved in the proetolityc ectodomain release, in other words the ectodomain 'shedding'.
32
Because of this, the first drugs tested on TACE are initially identified as MMP inhibitors. But, unluckily, they failed in clinical trials. For example marimastat, prinomastat and GCS 27023A (figure 19).
The difference in shape and size of the S1’ pocket of TACE and MMPs might be exploited to design selective TACE inhibitors devoid of any MMP activity. From data, it is quite evident that the S1’ pocket (right hand side of the zinc atom corresponding to P1’, P2’ and P3’ residues of the substrate) of TACE is larger and narrower than MMPs.
On the base of this size difference between the two metalloproteinases, Bristol–Myers– Squibb compared the compound 4 with the compound 5, and they demonstrated as the replace of the 4-methoxyphenyl group with a quinolinyl moiety increased the selectivity towards TACE. (4: IC50= >1000 nM; 5: IC50=3.7 nM). The quinolinyl moiety would fit into the larger S1’ pocket of TACE while it would clash with the smaller S1’ pocket of MMPs.
Recently, it has been discovered a 4-4’piperidine β-sulfone hydroxamate inhibitor (E) that shows good selectivity for TACE over MMPs 1, 2, 9, 13, and 14 due to its sterically bulky isopropyl sulfonamide P1 group. (IC50= 1.5 TACE; 8780 MMP-1; 355 MMP-2; 1670 MMP-9; 230 MMP-13).27
33
To increase the selectivity towards TACE, Condon et al. have synthesized a series of compounds varying P1’ substituents at the piperidine nitrogen, starting from compound
E. Table 5. IC50 nM C. R TACE MMP 1 MMP 2 MMP 9 MMP 13 MMP 14 F CH2 -4-Pyridine 2.1 7250 77 - 114 784 G C(O)-4-Pyperidine 2.6 3010 46 191 21.5 583 14 2.2 37300 664 5500 2277 24000 27 1.6 2680 555 920 259 14800 28 1.7 - 183 449
The X-ray structure of compound 14 bound to TACE (PDB entry 2I47) illustrates how the dimethyl isoxazole ring is accommodated by TACE, while an X-ray structure of 14-MMP-13 (1ZTQ22) shows how the isoxazole ring emerging from the S1 pocket, leaving it largely exposed to solvent. This is due to the somewhat larger and more hydrophobic MMP-13 Tyr 151 residue replacing Lys 315 of TACE. As such, Tyr 151 cannot readily adopt a conformation that allows the isoxazole ring to effectively bind in the MMP-13 S1 pocket.
34
Having et al. identified β,β’-disubstitution of the substituent borne by the P1 sulfonamide of the piperidine ring as a motif that supplied selectivity, they have been synthesized the compounds 27 and 28 to increase the selectivity for TACE over MMP-2 and MMP-13, by the incorporation of steric bulk adjacent to the attachment point of the P1 group to the piperidine nitrogen.28
Additional efforts to improve the TACE selectivity of this series of compounds has been made focusing on the S1’ specificity pocket. The distinctively bent shape of the TACE S1’/S3’ pockets, as compared to the MMPs, appeared to offer an opportunity for designing P1’ groups that would confer selectivity for TACE.
The replacement of the butynyloxy group with a butynylamino (in particular -NHCH2CCCH3) group might provide enhanced selectivity for TACE.
Table 6. IC50 nM. Compound R1 R2 TACE MMP-2 MMP-13 MMP-14 E SO2-i-Pr OCH2CCCH3 3 280 157 3642 7 Boc NHCH2CCCH3 1.6 120 860 2370 17 SO2-i-Pr NHCH2CCCH3 2 97 2075 NT 22 COCH(CH3)2 NHCH2CCCH3 2.3 513 3199 4891
The ‘bent’ conformation required for binding this P1’ moiety in the TACE S1’–S3’ pocket is 0.68 kcal/mol lower in energy than the ‘extended’ conformation needed to bind in the MMP-13 S1’ pocket. Thus, the relative cost for extending the tail is significantly higher for the butynylamine analogue than for the butynyloxy analogue and 7 would therefore be predicted to be more TACE selective.
35
Between the analogues prepared for this study, compound 22 has the best combination of potency against TACE enzyme and a desirable selectivity profile.
Levin et al. have explored thiomorpholine hydroxamates, bearing butynyloxy moiety at P1’ site. Amongst the compounds then developed, the most promising was TMI-1 (Figure 18).27
Figure 18. TMI-1
Efforts were also made to explore whether the enhanced selectivity for TACE afforded by the butynyl amine P1’ group could be extended to the thiomorpholine sulfonamide series that had previously provided potent but non-selective TACE/MMP inhibitors. This butynylamine 26 analogue was slightly more active and selective than butynyloxy analogue 6 against TACE.
Table 7. IC 50 nM Compound TACE MMP-2 MMP-13 6 8 4.7 2.4 26 2.3 56 30
36
It was also reported that substituted benzylic ether as P1’ group can also produce good TACE inhibitors. One of these active compounds is 29 (IC50=110 nM). The 2-position of the terminal phenyl group of 5-hydroxy pipecolic acid has to be substituted with alkyl or halo group for the TACE inhibitory activity.
Compound 30 has been developed by Letavic et al., it is piperizine-based dual inhibitor of TACE and MMP-13 that shows minimal MMP-1 activity.27
Table 8. IC50 nM
Compound TACE MMP-13 MMP-1
37
Chapter 5: Experimental part Introduction
5.1. Design and synthesis of selective MMP-8 inhibitors.
Matrix metalloproteinases (MMPs) are a large family of zinc-dependent endopeptidases. They are widely involved in the regulation of metabolism through both extensive protein degradation and selective peptide-bond hydrolysis. Moreover they are the main processors of extracellular matrix (ECM) components. As a result of their activity, MMPs participate in physiological processes such as tissue turnover, embryogenesis, angiogenesis but even processes that go beyond the mere ECM turnover. These comprise the activation or inactivation of other proteins by proteolysis of selected bonds, i.e. the “shedding” of membrane-anchored forms into circulation. It has been discovered several substrates that include other proteases, growth factors, hormones, cytokines, chemotactic and adhesion molecules. Anyway their activity is substrate, spatial and temporal-specific, but if it isn’t under control, this can lead up to different pathologies such as arthritis, inflammation and cancer.5 Therefore they were identified as promising targets for the design and development of low-molecular-weight inhibitors potentially useful as innovative therapeutic agents for treating the above diseases. Most synthetic inhibitors of MMPs are formed by a zinc-binding function and a peptide or peptidomimetic backbone, assembled to ensure cooperative binding interactions with both the catalytic zinc ion and the adjacent specificity subsites of the enzyme. The majority of the most powerful synthetic inhibitors incorporate a hydroxamate group as the zinc-binding function (ZBG).12 However in MMPs, the main subsites for substrate recognition are the specificity pocket S1′ and to a lesser extent, S2’. The S1’ pocket is shaped by elements from helix αB, the Met-turn and diverges in length among MMPs, for example it is shorter in MMP-1, -9 and -11, with respect to MMP-8. The vast majority of the first-generation inhibitors were peptido-derived compounds and bound MMPs on their primed side (right-hand of the zinc). Examples include: marimastat, prinomastat and CGS27023A(Figure 19).
38
Figure 19. First generation MMP inhibitors.
The first inhibitors were used for the treatment of arthritis, and after promising results in preclinical studies, their design was rapidly optimized and pointed towards other diseases such as cancer. But unfortunately they were not successful, probably because they were not specific and acted on several MMPs. In fact, the use of broad-spectrum MMPIs (MMP inhibitors) in clinical trials was greatly limited by muskoloskeletal syndrome (MSS) side effects, which required a drug dose reduction and the consequent reduction of beneficial effects.9 This has made necessary to re-evaluate the role of these proteases in cancer. The work of Balbin et al., which was based on the use of mouse models deficient in specific MMPs, has revealed that these enzymes play many roles in cancer and some MMPs could have also an anti-tumorigenic role (for example MMP8).8 On the basis of these results, researches have focused their interest in finding out more selective MMPIs.
MMP-8, also known as neutrophil collagenase, is considered as an antitarget for cancer therapy, but it has been claimed to have a key role in heart disease, osteoarthritis, and various inflammatory conditions such as hepatitis and ulcerative colitis. For these reasons, several research groups have focused their attention on the development of low-molecular-weight inhibitors of MMP-8 as new therapeutic agents for the above diseases.
In my thesis project I have synthesized two hydroxamates 1 and 2 (Figure 20) as potential MMP-8 selective inhibitors.
39
These are sulfone-based benzoic hydroxamate analogues of 3 (Figure 20), a nanomolar inhibitor of MMP-8 previously described.33 The para-methoxy substituent on the biphenylsulfonyl moiety of 3 was replaced by a para-methylthio and a para-N-morpholino group respectively, with the aim to improve the selectivity of the inhibitors for MMP-8 over the other MMPs.
Figure 20.
In fact, previous studies had shown that the presence of an elongated hydrophobic substituent in P1’ could ensure a good inhibition of deep-pocket MMPs, such as MMP-8, thanks to its optimal interaction with the S1’ pocket. The tricky point was to obtain also a good selectivity for MMP-8 over the other deep-pocket MMPs, like MMP-2, -9 or -13, given the high level of homology among the various MMP catalytic sites.
40
Scheme 1. Synthesis of hydroxamic acids 1 and 2.
The benzyl 2-(4-bromophenylthio)benzoate 4 was oxidized by treatment with Oxone in methanol, THF and water to the corresponding sulfone 5. Palladium-catalyzed cross-coupling (Suzuki conditions) of protected arylbromide 5 with 4-methylthiophenylboronic or 4-morpholinophenylboronic acid afforded biphenyl derivatives 6a,b that were hydrolyzed to the corresponding carboxylic acid 7a and 7b by saponification with KOH in good yields. Finally, the corresponding hydroxamates, 1 and
2, were obtained from carboxylates 7a,b via condensation with
O-(tert-butyldimethylsilyl)hydroxylamine, followed by acid hydrolysis with trifluoracetic acid in dichlormethane.
41
The newly synthesized compounds 1 and 2 have been tested on the target enzyme, MMP-8, and on other MMPs (MMP-1,-2, -3, -9 and -14) in order to assess their selectivity profile. The assays have been conducted in vitro on human recombinant MMPs by a fluorometric assay which uses a fluorogenic peptide as the substrate (Table 9). The previously described inhibitor 3 has been used as reference compound.
Table 9. In Vitro Enzymatic Activity (IC50 nM values) of MMP inhibitors. Compound MMP-1 MMP-2 MMP-3 MMP-8 MMP-9 MMP-14
3 28000 100 4000 87 930 1300
1 >5000 37 2400 98 1200 670
2 167000 24 12000 57 2300 1140
As shown in Table 1, the substitution of the 4-methoxy group of 3 with 4-N-morpholino-group (2) caused just a small increase in inhibitory activity against MMP-8 but an interesting increase in selectivity over MMP-1, MMP-3, MMP-9 and -14 with respect to
3. These compounds will be tested in a mice model of corneal perforation by Dr. Cintia
42
5.2. Design and synthesis of selective ADAM-17 inhibitors.
In the second part of this thesis, my studies have focused on another subclass of metzincins, the adamalysins.
Ectodomain shedding, or rather proteolytic ectodomain release, is essential for several biological processes (cell fate determination, cell migration, wound healing, neurite and axon guidance, heart development, immunity, cell proliferation and angiogenesis). ADAMs (A Disintegrin And Metalloproteinases) have emerged as the major proteinase family that mediates ectodomain shedding and when this process runs away the control mechanisms it is associated with autoimmune and cardiovascular diseases, neurodegeneration, infection, inflammation and cancer.18 ADAM-10 and ADAM-17 (or TACE, TNF-alpha converting enzyme) are the two principal metalloproteases of ADAMs family able to modulate inflammatory responses. TNF-α is the key player in inflammation, rheumatoid arthritis (RA), anorexia and septic shock, and in the last few years many pharmacological approaches focused on the development of ADAM-17 inhibitors to treat this diseases.18 As regards neurodegeneration, it has been proven that some ADAMs such as ADAM-10, can act as α-secretases, involved in the non-amyloidogenic processing of APP. For this reason, ADAM-10 could be a promising therapeutic target to treat neurodegenerative Alzheimer’s disease. Furthermore, over the last few years the ADAM family has been associated with the tumor development. Extracellular proteases, such as ADAMs, are key regulators of both cell–cell and cell– extracellular matrix (ECM) interactions that are associated with all aspects of tumorigenesis and metastasis, and there is interest in their potential as targets for the development of anticancer agents. In particular ADAM-17 is a major sheddase for the EGFR and other Erbb receptor ligands, including the proforms of heparin-binding EGF-like growth factor (HbEGF), transforming growth factor-α (TGFα), amphiregulin and epiregulin. This has implications for many cancers in which signalling from EGFR affects cell proliferation and the migration both of tumour cells and stromal cells.18 In view of the fact that specific ADAMs have been shown to promote cancer initiation and progression, it is reasonable to hypothesize that blocking their actions would slow or prevent progression. In recent years, a number of selective synthetic inhibitors against a small number of ADAMs have been described.
43
The majority of these use hydroxamate as the zinc-binding group (ZBG) and are designed to interact with the prime site subsites (S1’-S3’) of the MMP-like catalytic site. Whereas most of them are broad inhibitors, acting both MMPs and ADAM-17, other are relatively selective for specific ADAMs, especially ADAM-10 and ADAM-17.15 To date, TACE inhibitors reported in the literature are almost exclusively of the ‘right-hand-side design’ previously described for MMPs inhibitors and are of either the succinyl hydroxamate (exemplified by INCB3619) or the right-hand side peptidomimetic/sulfonamide hydroxamate variety.21
INCB3619 is one of the most widely investigated selective ADAM inhibitors and it is selective for ADAM-10 and ADAM-17 with IC50 values of 14 and 22 nM/l, respectively (Incyte corporation).15 Regarding the sulfonamide hydroxamate variety, several structural models have been studied.
In particular, Prof. Rossello’s group is currently studying the properties of dimeric ADAM-17 inhibitors,that are constituted of two identical units linked by a spacer in P1 position. In fact, similar studies carried out on MMP-9 and MMP-12 (unpublished results from this group) have shown that bi-functional ligands obtained by linking two identical heads (MMP binding moieties) with a proper spacer is able to chemically induce an unnatural MMP dimerization process which seems to arrestcancer cell invasiveness in a human glioblastoma cell line.
In order to prove a similar effect of bi-functional ligands on ADAM-17 dimerization, in the second part of my thesis project I have synthesized a dimeric compound (9, Figure 21) as analogue of a previously obtained dual-inhibitor (10, Figure 21) endowed with a good selectivity for ADAM-17.
44
Figure 21. Compounds 9 and 10.
The new derivative is a di-hydroxamate with a spacer between the two symmetrical arylsulfonamido units shorter than the one originally present in 10. As arylsulfonamido moiety (P1’ substituent) was chosen a 4-(3,5-dibromobenzyloxy)benzene, a group previously discovered31 by Prof. Rossello’s group as a suitable one to properly interact with the “L-shaped” S1’ subpocket characteristic of ADAM-17 catalytic site.
45
Scheme 2. Synthesis of compound 9.
Sulfonyl chloride 12 was synthesized starting from sodium 4-hydroxybenzenesulfonate dihydrate, which was alkylated with 3,5-dibromobenzyl bromide using sodium hydroxide as base. Reaction of sodium salt 11 with oxalyl chloride in the presence of
46
N,N-dimethylformamide in dichloromethane afforded sulfonyl chloride 12, which was coupled with the commercial amino acid H-D-Orn(Boc)OH in water and 1,4-Dioxane in the presence of N,N,N-Triethylamine to give the sulfonamide 13. Carboxylate 13 was converted into protected hydroxamate 14 by condensation with O-(tetrahydro-2Hpyran-2-yl)hydroxylamine using of N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC) as coupling reagent and then hydrolyzed with HCl 4N to obtained the salt 15. Reaction of the isophtaloyl dichloride with N-hydroxy succinimmide afford the activated linker 16 which is reacted with the hydrochloride salt 15 in presence of DIPEA leading to the formation of the dimeric compound 9.
Compound 9 was tested in vitro for its ability to inhibit human recombinant ADAM-17, MMP-1, -2, -9, and -14 by a fluorometric assay which uses a fluorogenic peptide as the substrate (Table 10) in comparison with 10.
Table 10. In Vitro Enzymatic Activity (IC50 nM values) of ADAM-17 inhibitors. Compound ADAM-17 MMP-1 MMP-2 MMP-9 MMP-14
9 24 >200000 3000 8800 50000
10 164 415000 4660 3320 57000
The introduction of a shorter spacer was able to increase the inhibitory activity towards ADAM-17 (IC50= 24nM) and also the selectivity over the tested MMPs with respect to
10. Further studies will be carried out with compound 9 to study its activity in cancer cell
lines and possibly elucidate the advantages coming from the use of a dimeric inhibitor instead of a monomeric one.
47
5.3. Attempts to obtain ADAM-10 selective inhibitors.
In the last part of my thesis I have started the synthesis of a new series of compounds as ADAM-10 selective inhibitors. Previous studies conducted by Prof. Rossello’s group have led to the discovery of a promising nanomolar ADAM-10 inhibitor, compound 17 (Figure 22).
Figure 22.
Compound 17 was tested in vitro for its ability to inhibit human recombinant ADAM-10, ADAM-17, MMP-1, -2, -8, -9, and -14 by a fluorimetric assay, which uses a fluorogenic peptide as the substrate (Table 11).
Table 11. In Vitro Enzymatic Activity (IC50 nM values) of 17.
Compound ADAM-10 ADAM-17 MMP-1 MMP-2 MMP-9 MMP-14
17 9.2 90 >200000 370 4500 50000
On the basis of these promising results, it was decided to further investigate this sulfonamide scaffold in order to possibly improve the inhibitory activity against ADAM-10.
In the first place, I tried to synthesize a dimeric analogue of 17, derivative 18 (Scheme 3) using the same spacer already present in compound 9 (see Scheme 2).
48
49
Commercially available 1,4-piperazine was mono-protected with di-tert-butyl dicarbonate (Boc2O) in CH2Cl2 to give compound 19. Arylpiperazine 20 was obtained by Buchwald-Hartwig cross coupling reaction of amine 19 with commercial 4-bromo-3-methylbenzonitrile, in the presence of palladium catalysts and potassium tert-butoxide in anhydrous tetrahydrofuran (THF). The cleavage of Boc protecting group with a solution of HCl/Dioxane 4 N gave arylamine 21 as hydrochloride. Sulfonyl chloride 22 was prepared by sulfonation of 21 with chlorosulfonic acid (ClSO3H) in the presence of triethylamine (Et3N) followed by treatment with phosphorus pentachloride (PCl5) in dry toluene at reflux. Sulfonyl chloride 22 was converted into sulfonamides 23 by reaction with the commercial amino acids H-D-Orn(Boc)OH in H2O and THF in the presence of triethylamine (Et3N). Unfortunately, the reaction gave sulfonamides 23 with a very low yield and the quantity of sulfonamide was not enough to continue the synthesis of 18. Moreover I tried to obtain a series of cyclic compounds of general formula A (Figure 23), as ADAM-10 selective inhibitors.
Figure 23. General structure A.
In Scheme 4 and 5 are described the synthetic pathways followed to obtain compound
50
Scheme 4. Synthesis of compound 24 (X=CO).
The commercially available D-thiazolidine-4-carboxylic acid was protected as tert-butyl ester 26 by reaction with isobutylene and H2SO4. The treatment with Et2O ∙ HCl gave the corresponding hydrochloride salt which was coupled with carbonyldiimidazole to give the compound 27.
51
The next step, necessary to obtain compound 29, via the reaction with arylpiperazine
28, did not give the expected results and the synthesis was interrupted. However, the
reaction conditions to obtain compound 28 were best optimized with respect to those reported in Scheme 3, affording 28 in one step by reaction of piperazine with 5-cyano-2-fluorotoluene in the presence of tetra-n-butylammonium bromide and potassium carbonate in DMSO.
52
Ethyl ester 30 was obtained via N-alkylation of arylpiperazine 28, described above, with ethyl bromoacetate in the presence of potassium carbonate as base. Basic hydrolysis of ethyl ester 30 with LiOH in THF and H2O gave the corresponding carboxylic acid 31. Unfortunately, also in this case, it was not possible to complete the synthetic pathway and thus to obtain the final compound 25 since the coupling reaction of the carboxylic derivate 31 with the tert-butyl thiazolidine-4-carboxylate 26 or with the sodium (R)-thiazolidine-4-carboxylate 33 did not afford the expected compounds 32 or 34.
53
Chapter 6: Synthesis
6.1 Materials and Methods
1H and 13C NMR spectra were determined with a Varian Gemini 200 MHz or a Bruker Avance III HD 400 MHz spectrometer. Chemical shifts (δ) are reported in parts per million downfield from tetramethylsilane and referenced from solvent references. Coupling constants J are reported in hertz. The following abbreviations are used: singlet (s), doublet (d), triplet (t), doubledoublet (dd), broad (br), and multiplet (m). Chromatographic separations were performed on silica gel columns by flash column chromatography (Kieselgel 40, 0.040-0.063 mm; Merck) or using ISOLUTE Flash Si II cartridges (Biotage). Reactions were followed by thin-layer chromatography (TLC) on Merck aluminum silica gel (60 F254) sheets that were visualized under a UV lamp, and hydroxamic acids were visualized with FeCl3 aqueous solution. Evaporation was performed in vacuo (rotating evaporator). Sodium sulfate was always used as the drying agent. Yields refer to isolated and purified products. D-Boc-ornithine (H-D-Orn(Boc)-OH) and D-thiazolidine-4-carboxylic acid were purchased from Bachem (Switzerland). All other commercially available chemicals were purchased from Sigma-Aldrich.
54
6.2 Scheme 1
Synthesis of benzyl 2-((4-bromophenyl)sulfonyl) benzoate (5).
A solution of oxone (12.32 g; 20 mmol) in H2O (48 ml) was added slowly to a solution of benzyl 2-(4-bromophenylthio)benzoate (1 g; 2.50 mmol) in THF/MeOH (36/12 ml). The reaction was stirred at room temperature for 5 days, and the organic solvent were evaporated under reduced pressure. The obtained suspension was diluted with H2O, and the product was extracted with EtOAc . The combined organic extracts were dried over anhydrous Na2SO4, filtered and evaporated under reduced pressure. The crude oil was purified by trituration with Et2O/Hexane to yield 5 (980 mg) as a white solid.
Yield 91%
1H NMR (CDCl
3) δ: 5.39(s, 2H), 7.36-7.44(m, 5H), 7.54(d, J=8.2 Hz, 2H), 7.57-7.65(m, 3H), 7.79(d, J= 8.2 Hz, 2H), 8.11-8.15(m, 1H).
General procedure to synthesize compounds 6a and 6b by Suzuki coupling.
A solution of Pd(OAc)2 (3%mol) and PPh3 (15%mol) in EtOH and Toluene was stirred at room temperature for 15’, under Argon atmosphere, to form the catalyst Pd(PPh3)4 in situ. Benzyl 2-((4-bromophenyl)sulfonyl) benzoate, Na2CO3 2M and phenylboronic acid were then added and the reaction was stirred under reflux (85°C) for 4h. After cooling, the obtained suspension was diluted with H2O and extracted with EtOAc. The combined organic extracts were dried over anhydrous Na2SO4, filtered and evaporated under reduced pressure.
Synthesis of benzyl 2-((4'-(methylthio)-[1,1'-biphenyl]-4-yl)sulfonyl)benzoate (6a).
The compound was prepared from benzyl 2-((4-bromophenyl)sulfonyl) benzoate (800 mg; 1.85 mmol) and phenylboronic acid (374 mg; 2.22 mmol) following the general procedure. The crude product was purified by silica gel flash chromatography to give the pure compound 6a (667 mg).
