J. Med. Chem. 2005, 48, 4367-4377
4367
Pyrrolo[1,5]benzoxa(thia)zepines as a New Class of Potent Apoptotic Agents. Biological Studies and Identification of an Intracellular Location of Their Drug Target Margaret M. Mc Gee,# Sandra Gemma,‡,| Stefania Butini,‡,| Anna Ramunno,‡,| Daniela M. Zisterer,# Caterina Fattorusso,†,| Bruno Catalanotti,†,| Gagan Kukreja,‡,| Isabella Fiorini,‡,| Claudio Pisano,§ Carla Cucco,§ Ettore Novellino,†,| Vito Nacci,‡,| D. Clive Williams,# and Giuseppe Campiani*,‡,| Department of Biochemistry, Trinity College, Dublin 2, Ireland, Dipartimento Farmaco Chimico Tecnologico, via Aldo Moro, and European Research Centre for Drug Discovery and Development, Universita` di Siena, 53100 Siena, Italy, Dipartimento di Chimica delle Sostanze Naturali and Dipartimento di Chimica Farmaceutica e Tossicologica, Universita` di Napoli Federico II, via D. Montesano 49, 80131 Napoli, Italy, and Sigma-Tau Industrie Farmaceutiche Riunite, via Pontina Km 30,400, 00040 Pomezia, Italy Received July 27, 2004
We have recently developed five novel pyrrolo-1,5-benzoxazepines as proapoptotic agents. Their JNK-dependent induction of apoptosis in tumor cells suggested their potential as novel anticancer agents. The core structure of the apoptotic agent 6 was investigated, and the SARs were expanded with the design and synthesis of several analogues. To define the apoptotic mechanism of the new compounds and the localization of their drug target, two analogues of 6 were designed and synthesized to delineate events leading to JNK activation. The cell-penetrating compound 16 induced apoptosis in tumor cells, while its nonpenetrating analogue, 17, was incapable of inducing apoptosis or activating JNK. Plasma membrane permeabilization of tumor cells resulted in 17-induced JNK activation, suggesting that the pyrrolo-1,5-benzoxazepine molecular target is intracellular. Interestingly, compound 6 displayed cytotoxic activity against a panel of human tumor cell lines but demonstrated negligible toxicity in vivo with no effect on the animals’ hematology parameters. Introduction The regulation of apoptosis or programmed cell death is critical to the normal development and maintenance of tissue homeostasis. The molecular machinery of apoptosis has been studied extensively, and many components have been identified. Two main pathways of cell death have been elucidated: a death-receptormediated pathway and a mitochondrial pathway. During death-receptor-mediated apoptosis, ligand interaction at the cell surface initiates signal transduction cascades by adaptor complexes that lead to activation of downstream substrates such as caspases.1,2 Activation of the mitochondrial apoptotic pathway occurs in response to various stimuli including cellular stress, DNA damaging agents, and UV irradiation and results in the release of apoptogenic proteins, which activate downstream substrates.1,3,4 Signaling cascades that regulate mitochondrial apoptotic pathways are incompletely defined, but some have been shown to involve mitogen activated protein (MAP) kinases and members of the Bcl-2 protein family.3,5 MAP kinases are a group of serine/threonine protein kinases that control a wide variety of cellular processes * To whom correspondence should be addressed. Address: Dipartimento Farmaco Chimico Tecnologico, Universita` di Siena, via Aldo Moro, 53100 Siena, Italy. Phone: 0039-0577-234172. Fax: 0039-0577234333. E-mail:
[email protected]. # Trinity College. ‡ Dipartimento Farmaco Chimico Tecnologico, Universita ` di Siena. | European Research Centre for Drug Discovery and Development, Universita` di Siena. † Universita ` di Napoli Federico II. § Sigma-Tau Industrie Farmaceutiche Riunite.
including cell growth and differentiation. In mammalian cells, three MAP kinases have been characterized in detail, the c-Jun N-terminal kinase (JNK), the p38 MAP kinase, and the extracellular signal regulated kinase (ERK). The JNK and p38 kinases are activated primarily by cellular stress such as heat shock, osmotic shock, and DNA damaging drugs, whereas the ERK subgroup is primarily activated by mitogens such as growth factors. These kinases are activated in response to their phosphorylation by upstream MAP kinase kinases (MAPKK). The MAPKK group are phosphorylated and activated by upstream MAP kinase kinase kinases (MAPKKK).6 The JNK subgroup comprises at least eight isoforms derived by differential splicing from three or more genes to generate proteins of 54 and 46 kDa.7 The consequence of JNK activation in cells is dependent on a number of factors including the death-inducing stimulus and the cell type, and these have been shown to mediate both pro- and antiapoptotic responses.8 Proapoptotic JNK signaling cascades are mediated in response to a variety of stimuli such as the TNFR death receptor and the DNA damaging agent etoposide9 and in some cases have been shown to regulate apoptosis by a direct interaction with a number of regulators such as members of the Bcl-2 family.5,10 We have previously synthesized a novel series of pyrrolo-1,5-benzoxazepine compounds as high-affinity ligands to the peripheral-type benzodiazepine receptor (PBR), a receptor hypothesized to be implicated in cell growth, apoptosis, and other functions, in an attempt to probe the function of the receptor, which remains
10.1021/jm049402y CCC: $30.25 © 2005 American Chemical Society Published on Web 05/27/2005
4368
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 13
Mc Gee et al.
unclear.11 We provided evidence demonstrating that three pyrrolobenzoxazepines inhibited the proliferation of rat C6 glioma and 1321N1 astrocytoma cells,12 but the antiproliferative effect of these compounds was independent of binding to the peripheral benzodiazepine receptor.12 While examining the effect of these compounds and of the commonly used PBR ligand PK11195 on the proliferation of human acute myelocytic leukemia HL-60 cells, we observed that some pyrrolobenzoxazepines could induce a low degree of apoptosis (1020%) in these cells. Starting from the original series of benzoxazepines,12 we initially synthesized a small set of heterocycles, based on an oxazepine structure, characterized by significant apoptotic activity on HL-60 acute myelocytic, Jurkart T lymphoma, and Hut-78 subcutaneous T-cell lymphoma cells.13 This study led to the identification of the prototypic apoptotic agent 6. Compound 6 was also found to induce apoptosis in a variety of cancerous cells including three chemotherapyresistant chronic myelogenous leukemia (CML) cell lines (K562, KYO.1, and LAMA 8414) and the human breast carcinoma (MCF-7 cells15), indicating its potential as a novel anticancer therapeutic. Although the molecular drug target of pyrrolo-1,5benzoxazepines remains unidentified, we have dissected part of the biochemical signaling pathway initiated by the representative compound 6. It has been found that 6 induces apoptosis in CML cells by bypassing the apoptotic suppressor Bcr-Abl.14 In addition, the early activation of JNK, which occurs within 15 min, was found to be essential in the apoptotic signaling cascade initiated by 6 in CML cells,16 whereas the activation of caspase 3 is dispensable. This is in agreement with recent reports supporting the existence of caspaseindependent cell death.17 These results stimulated our interest in the development of a series of novel apoptotic agents as potential antitumorals, and by an integrated effort involving synthesis, molecular modeling, and biological studies, we describe herein a number of apoptotic agents equal to or with increased apoptotic activity over our lead 6. Several of the new heterocyclic analogues showed apoptotic activity on HL-60 cells and on the K562 cells. K562 cells, which are derived from the pleural effusion of a CML patient in terminal blast crisis and express the p210 Bcl-Abl fusion protein, are particularly resistant to the induction of apoptosis by various agents including camptothecin, Ara-C, etoposide, paclitaxel, staurosporine, and anti-Fas antibodies. In an attempt to delineate the signaling events leading to JNK activation in K562 cells and to determine whether 6, representative of the whole set, initiates a signaling cascade by binding to a death receptor located on the plasma membrane or by binding to an intracellular target, a penetrating derivative and a nonpenetrating derivative (16 and 17, respectively) of 6 were synthesized. In the present work we report the design, synthesis, and biological evaluation of analogues of 6 and discuss how they have been used to increase our understanding of the mechanism of action of these new series of apoptotic compounds. Several compounds, differently substituted at C-6 and C-7, were used as “molecular yardsticks” to probe the spatial dimension of the lipophilic pockets L1-3 (Chart 1) at a putative receptor binding site, and
Chart 1. Schematic Representation of the Hypothesized Main Areas of Interaction with the Molecular Target
a number of structure-activity relationships trends have been identified. A possible explanation is advanced in order to account for the differences on apoptotic activity. Furthermore, a thorough pharmacological investigation of our lead 6 is discussed. Chemistry The synthesis of compounds 1-4, 6, 7, and 18 was accomplished as previously described (see refs 11 and 18), and the synthesis of the new compounds 5 and 8-17 is reported in Scheme 1. 4-Amino-3-hydroxybenzoic acid 19 underwent a Clauson-Kaas reaction followed by an esterification to afford 20a in high yield. 20a and the pyrrolo derivatives 20b-d11,18 were then converted into the corresponding esters 21a-d by means of an O-alkylation with the appropriate R-bromoarylacetic ester. Basic hydrolysis of the esters 21b-d afforded acids 22b-d, while acid 22a was obtained by a palladium-catalyzed cleavage of the allyl ester 21a. The intramolecular Friedel-Crafts cyclization of the acids 22a-d then gave the corresponding cyclic ketones 23a-d. Treatment of the potassium or sodium enolates of ketones 23a-e with the appropriate N,N-dialkylcarbamyl chloride or acyl chloride provided the target compounds 5, 6, 8-17. The conversion of the ethyl ester 23a into the allyl ester 25 was accomplished by basic hydrolysis of the ester 23a, followed by treatment with allyl bromide and cesium fluoride of the free acid 24. Then treatment of the sodium enolate of the ketone 25 with N,N-dimethylcarbamyl chloride and subsequent deprotection of the ester function with tetrakis(triphenylphosphine)palladium(0) and phenylsilane afforded the desired final compound 17. Results and Discussion The capability of novel oxa(thia)zepines to induce apoptosis in HL-60 and in highly resistant chronic myelogenous leukaemia (CML) K562 cells, tested at 10 µM, is reported in Table 1. The characteristic morphological effects of apoptosis, i.e., cell shrinkage, extensive membrane blebbing, condensation of chromatin, and DNA fragmentation were observed in these cells. In addition, the degree of necrosis induced by these agents under the same conditions was determined by the presence of swollen cells and loss of cell membrane. The degree of necrosis induced by each agent was negligible, with values of less than 1.0% in the two cell lines tested. These results correlate with previous findings13,14 and indicate that the mode of cell death induced by this novel class of compounds is apoptosis. The pyrrolobenzoxazepine 6, representative of the new series of apoptotic agents, was further evaluated for its in vitro
Pyrrolo[1,5]benzoxa(thia)zepines
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 13 4369
Scheme 1
To rationalize the structural parameters responsible for apoptotic activity, we performed a systematic conformational search (Sybyl 6.6, Tripos, St. Louis, MO) on compounds 4, 6, 7, 12, 14, and 15. Each generated conformation was subjected to a geometry optimization (Discover, Accelrys, San Diego, CA). Conformers were energetically selected and then grouped into families by fitting the atoms of the tricyclic ring system. Structural and conformational features of the active compounds were compared to the ones of the inactive pyrrolobenzoxa(thia)zepines, allowing us to identify three main areas of interaction with the molecular target, schematically represented in Chart 1. The R′ group could be accommodated into the L1 region of the putative binding site. The complete loss of activity observed in compounds 1 and 2, in contrast to the potent apoptotic agents 5 and 6, demonstrated that an extension of the aromatic system is responsible for biological activity. Accordingly, the size reduction of the system at C-6 to p-MePh (3 vs 6) or to p-OMePh (18 vs 6) led to analogues with weak-to-negligible activity. Furthermore, the inactivity of the 2-naphthyl derivative 7 (the same result was found for its pyrido-fused counterpart; data not shown) clearly indicated a shape-dependent effect of the substituent placed at C-6 on the pyrrolooxa(thia)zepines-induced apoptotic event (Figure 1). In particular, the additional aromatic ring is able to favorably interact with L1 if it protrudes along the X direction of the Cartesian axes displayed in Figure 1 (1-naphthyl derivatives 6, 15; Figure 1A) while it is not tolerated along the Y direction (7, Figure 1B). The R′′ substituents interact with a second region in the putative binding site (L2, Chart 1). Two different series of substituents were investigated, esters (1, 5, 8-10, 13-15) and carbamates (2-4, 6, 7, 11, 12, 1618). In both series of compounds the introduction of an alkyl chain was detrimental for efficacy. Indeed, the methyl group in the ester series (5, 10, and 13) is welltolerated, while the introduction of longer alkyl groups (8 and 9) resulted in a complete loss of apoptotic activity. The ethyl and propyl derivatives are also devoid of apoptotic activity (data not shown). Similar behavior can be observed in the carbamate series; substitution with a diethylamino group caused a drop of activity with respect to the dimethylamino group (12 vs 11). The same trend was found for analogues in which the dimethylamino group of 6 was replaced by a methylamino (41% apoptosis on HL-60 cells at 10 µM) or an ethylamino group (15% apoptosis on HL-60 cells at 10 µM). As we discussed above, substitution of the 1-naphthyl group of oxa(thia)zepine derivatives by a phenyl (1, 2), p-MePh (3), and a p-OMePh (18) led to compounds characterized by weak-to-negligible activity. Interestingly, apoptotic activity may be partially restored by replacing the dimethylamino group at C-7 by a 4-pyridyl (18 (
