Apicidin, an Inhibitor of Histone Deacetylase, Prevents H-Ras-Induced Invasive Phenotype

Cancer Letters 157 (2000) 23±30 www.elsevier.com/locate/canlet

Apicidin, an inhibitor of histone deacetylase, prevents H-ras-induced invasive phenotype

Mi-Sung Kima, Mi-Won Sonb, Won-Bae Kimb, Young In Parkc, Aree Moona,*

aCollege of Pharmacy, Duksung Women's University, Seoul 132-714, South Korea bResearch Institute of Dong-A Pharm. Co., Kyunggido 449-900, South Korea cGraduate School of Biotechnology, Korea University, Seoul 136-701, South Korea Received 25 January 2000; received in revised form 1 May 2000; accepted 3 May 2000

Abstract Cancer metastasis represents the most important cause of cancer death and agents that may inhibit tumor cell invasion have been extensively pursued. In the present study, we have examined the anti-invasive effect of apicidin [cyclo(N-O-methyl-l- tryptophanyl-l-isoleucinyl-d-pipecolinyl-l-2-amino-8-oxodecanoyl)], a fungal metabolite that was identi®ed as an antiprotozoal agent known to inhibit parasite histone deacetylase (HDAC). We show that apicidin signi®cantly inhibits H-ras-induced invasive phenotype of MCF10A human breast epithelial cells in parallel with a speci®c downregulation of matrix metalloproteinase (MMP)-2, but not MMP-9. We also show that apicidin induces a morphological reversal and growth inhibition of H-ras MCF10A cells similar to that induced by other HDAC inhibitors. Taken in conjunction with the fact that uncontrolled ras activation is probably the most common genetic defect in human cancer cells, our data showing the anti-invasive and detransforming activities of apicidin in H-ras-transformed MCF10A cells may suggest a potential use of HDAC inhibitors for treatment of cancer. q 2000 Elsevier Science Ireland Ltd. All rights reserved.

Keywords: Apicidin; H-ras; Histone deacetylase inhibitor; Invasion; Matrix metalloproteinase; Detransformation

1. Introduction mammalian systems, little is known about the cellular effects of apicidin in oncogene-transformed Apicidin [cyclo(N-O-methyl-l-tryptophanyl-l-iso- mammalian cells. Accumulating evidence suggests leucinyl-d-pipecolinyl-l-2-amino-8-oxodecanoyl)] is that histone acetylation and deacetylation play a fungal metabolite that has been identi®ed as an important roles in the regulation of transcription in antiprotozoal agent that inhibits parasite histone eucaryotic cells [4]. Detransforming activities of deacetylase (HDAC) [1]. Shown in Fig. 1 are the naturally occurring HDAC inhibitors such as structures of apicidin and trapoxin, a structurally trapoxin, trichostatin A, and depudecin have been related HDAC inhibitor [2]. Although some biologi- elucidated [2,5±7]. Anti-angiogenic activity of depu- cal activities of apicidin, including anti-proliferative decin has been found both in vitro and in vivo [8]. [1] and toxic effects [3], have been shown in Cellular responses by HDAC inhibitors including cell cycle arrest, alteration of gene expression, and induction of apoptosis have been demonstrated [9±11]. * Corresponding author. Tel.: 182-2-901-8394; fax: 182-2-901- Cancer metastasis represents the most important 8386. cause of cancer death and antitumor agents that may E-mail address: [email protected] (A. Moon).

Fig. 1. Chemical structures of apicidin and trapoxin. inhibit this process have been pursued. Uncontrolled 2. Materials and methods degradation of the extracellular matrix and basement membrane, an essential part of the metastatic process, 2.1. Materials is thought to be associated with tumor cell invasion [12]. A role for members of the matrix metalloprotei- Apicidin was provided by Dr Y.-W. Lee (Seoul nase (MMP) family on tumor invasion and metastasis National University, Suwon, Korea) [3] and resus- formation has been suggested, especially, MMP-2 (72 pended in dimethyl sulphoxide (DMSO). All other kDa type IV collagenase, gelatinase A) and MMP-9 chemicals were of the highest quality commercially (92 kDa type IV collagenase, gelatinase B), which available. degrade type IV collagen, the major structural collagen of the basement membrane [13]. We have 2.2. Cell lines and culture condition previously shown that H-ras, but not N-ras, induces MCF10A is a spontaneously immortalized an invasive phenotype in MCF10A human breast `normal' breast epithelial cell line [16]. The estab- epithelial cells with which the expression of MMP-2 lishment of H-ras MCF10A cells was described is more closely associated [14], rather than that of previously [14]. MCF10A and H-ras MCF10A cells MMP-9 shown previously in rat embryonic ®broblasts were cultured in DMEM/F12 supplemented with 5% [15]. horse serum, 0.5 mg/ml hydrocortisone, 10 mg/ml In the present study, we attempted to examine the insulin, 2 ng/ml EGF, 0.1 mg/ml cholera enterotoxin, effect of apicidin, a HDAC inhibitor, on the invasive 2mMl-glutamine, 100 units/ml penicillin±strepto- phenotype of MCF10A cells transformed with v-H- mycin and 0.5 mg/ml fungizone. The cells were ras (H-ras MCF10A). Here, we show that apicidin cultured in a humidi®ed atmosphere of 5% CO at signi®cantly inhibits H-ras-induced invasiveness of 2 378C MCF10A cells in parallel with a speci®c downregulation of MMP-2, but not MMP-9. We also show that 2.3. In vitro HDAC activity apicidin induces a morphological reversal and growth inhibition of H-ras MCF10A cells similar to that The in vitro activity of human HDAC was induced by other HDAC inhibitors. Taken together, measured as the amount of [3H]acetic acid released apicidin exerted anti-invasive and detransforming from the [3H]acetylated histones as described by activities in H-ras-transformed MCF10A cells, Inoue and Fujimoto [17]. Brie¯y, cells (7 £ 108) suggesting a potential use of HDAC inhibitors for were harvested and suspended in HDAC buffer treatment of cancer. containing 15 mM potassium phosphate (pH 7.5), M.-S. Kim et al. / Cancer Letters 157 (2000) 23±30 25 5% glycerol and 0.2 mM EDTA. Nuclei were 3. Results collected and resuspended in HDAC buffer containing

1 M (NH4)2SO4. Protein (enzyme fraction) was preci- We examined whether apicidin, a known inhibitor pitated and dissolved in HDAC buffer followed by of parasite HDAC, indeed inhibited HDAC of H-ras dialysis. [3H]Acetyl-labeled histones were prepared MCF10A cells. As shown in Fig. 2, apicidin treatment from 1 £ 108 cells and added to the enzyme fraction ef®ciently inhibited HDAC activity of H-ras in the absence or presence of apicidin, and the mixture MCF10A cells in a dose-dependent manner with the was incubated at 378C for 10 min. The reaction was IC50 value of 102 nM. Potency of apicidin was much stopped by addition of 10 ml of concentrated HCl. The higher than that of sodium butyrate [18], which was released [3H]acetic acid was extracted with 1 ml of used as a positive control. Given that apicidin was ethyl acetate, and the radioactivity was determined. shown to display potent low nanomolar inhibitory

activity on parasitic HDAC with the IC50 value of 0.7 nM [1], its inhibitory activity on mammalian 2.4. Growth inhibition HDAC was approximately 150-fold less potent. Malignant transformation often causes a dramatic 5 Cells (1.5 £ 10 ) in a 48-well plate were cultured in morphological change in cells [19]. Many HDAC the presence of various concentrations of apicidin for inhibitors have been shown to revert the morphology 48 h. Control cells were treated with DMSO alone. of transformed cells [5±7]. To test whether apicidin Following treatment with apicidin, cells were trypsi- induces detransformation in oncogene-transformed nized and cell viability was determined by Trypan mammalian cells, we investigated the effect of apici- Blue exclusion assay. Number of cells were deter- din on morphology of H-ras MCF10A cells in which mined by counting with hemacytometer. constitutive activation of H-ras signaling results in a transformed phenotype [14]. As shown in Fig. 3, treatment with 480 nM apicidin for 48 h reversed the 2.5. In vitro invasion assay morphology of the H-ras-transformed cells to appar- In vitro invasion assay was performed as described ently normal cell counterpart, MCF10A. The spindle- previously [14] using a 24-well transwell unit with like, foci-forming morphology of H-ras MCF10A polycarbonate ®lters (Corning Costar, Cambridge, cells was changed to a ¯attened morphology that is MA). Fifty thousand cells pretreated with various similar to the parental MCF10A cells. Apicidin treat- concentrations of apicidin were resuspended in 100 ml of DMEM/F12 containing the corresponding concentrations of apicidin and placed in the upper part of a transwell plate. Cells were incubated for 17 h and the invasive phenotypes were determined by counting the cells that migrated to the lower side of the ®lter with microscopy at £ 400. Thirteen ®elds were counted for each assay. Each sample was assayed in triplicate.

2.6. Gelatin zymography

The conditioned media of cells pretreated with various concentrations of apicidin for 48 h were Fig. 2. Apicidin inhibits in vitro HDAC activity in H-ras MCF10A subjected to gelatin zymography [14]. After staining cells. In vitro HDAC activity was determined as the amount of [3H]acetic acid released from 3H-acetylated histones which were with 0.1% Coomassie brilliant blue followed by extracted from H-ras MCF10A cells and added to the enzyme frac- destaining with 10% acetic acid, areas of lysis were tion in the absence or presence of apicidin. Sodium butyrate was observed as white bands against a blue background. used as a positive control. 26 M.-S. Kim et al. / Cancer Letters 157 (2000) 23±30 apicidin while 95% of non-transformed MCF10A

cells survived after the same treatment. The IC50 value of apicidin for inhibiting H-ras MCF10A cell growth was 190 nM. The growth inhibitory concen- tration of apicidin appears to vary depending on the cell type: apicidin has been shown to inhibit HeLa cell proliferation at 50±100 nM for a 50% inhibitory effect [1]. These results indicate that apicidin induces detransformation in H-ras-transformed human breast epithelial cells, as evidenced by the reversal of cell morphology and the selective growth inhibition of the transformed cells. An essential part of metastatic process includes tumor cell invasion. We have previously shown that non-invasive MCF10A cells obtained invasiveness by activation of H-ras signaling [14]. In this study, we attempted to examine the effect of apicidin on the H- ras-induced invasive phenotype of MCF10A cells. After pretreatment with apicidin for 48 h, viable H- ras MCF10A cells (5 £ 104) were resuspended in DMEM/F12 media containing apicidin and subse- quently incubated in a transwell chamber coated with Matrigel for 17 h to allow cell invasion. As shown in Fig. 5, apicidin treatment signi®cantly reduced the number of invaded cells through a recon- Fig. 3. Apicidin induces a morphological change in the H-ras stituted basement membrane in a dose-dependent MCF10A cells. Subcon¯uent H-ras MCF10A cells were treated with 480 nM apicidin for 48 h and the morphology of the cells manner. The IC50 value of anti-invasive activity of was examined microscopically (100 £ ). apicidin in H-ras MCF10A cells was 450 nM. No change in invasiveness was observed in non-invasive ment did not alter the morphology of the parental MCF10A cells (data not shown). Several other HDAC inhibitors, including trapoxin, trichostatin A and depudecin are known to revert from a spindle- like, transformed to a normal cell morphology [5±7], suggesting that the detransforming activity of apicidin may also be due to the inhibition of HDAC. The broad spectrum of detransforming activities displayed by HDAC inhibitors implies a role for HDAC in cell growth control [20]. Antiproliferative activity of trapoxin was observed as a result of HDAC inhibition [21]. We examined the effect of apicidin on the proliferation of H-ras MCF10A cells. As shown in Fig. 4, treatment of apicidin for 48 h inhibited H-ras MCF10A cell proliferation in a dose-dependent manner Fig. 4. Apicidin inhibits growth of H-ras MCF10A cells. Cells (1.5 £ 105) were treated with apicidin (0.16±1.6 mM) for 48 h. The growth inhibitory effect of apicidin was selec- The number of live cells was determined by Trypan Blue exclusion tive: approximately 50% of H-ras MCF10A cells assay. The percentage of cell survival was normalized to the control remained viable following treatment with 160 nM cells. The results presented are means ^ SE of triplicates. M.-S. Kim et al. / Cancer Letters 157 (2000) 23±30 27

Fig. 5. Apicidin inhibits invasive phenotype of H-ras MCF10A cells. The cells were pretreated with various concentrations of apicidin for 48 h. Upper panel, light microscopic examinations (100£) of invaded cells treated with 0.16 mM (B), 0.48 mM (C) and 1.6 mM (D) of apicidin. Control cells are shown in A. Bottom panel, quantitative analysis of the anti-invasive activity of apicidin. The number of invaded cells per ®eld were counted (400£) in thirteen ®elds and the mean values were determined. Experiments were performed triplicate. Bar is mean number of invaded cells per ®eld, and line represents the SE for triplicate determinations. **Statistically different (P , 0:05) from control cell by the independent two sample Student's t-test.

MCF10A cells upon apicidin treatment (data not was observed by gelatin zymograph while the activity shown). of MMP-9 was not changed (Fig. 6). These results are Since members of the MMP family are known to consistent with our previous ®nding that MMP-2 is play an essential role in tumor invasion, we deter- more closely associated with the H-ras-induced inva- mined the possible association of MMP-2 and/or sive phenotype than MMP-9 in human breast epithe- MMP-9 in the anti-invasive effect of apicidin. We lial cells [14]. Upon treatment of 1.6 mM apicidin, examined the enzymatic activities of MMP-2 and where invasion was almost completely blocked (see MMP-9 secreted by H-ras MCF10A cells treated Fig. 5), MMP-2 activity remained about 50% of with various concentrations of apicidin for 48 h. A control, suggesting that MMP-2 is not the only signi®cant downregulation of MMP-2 by apicidin element responsible for the anti-invasive effect of 28 M.-S. Kim et al. / Cancer Letters 157 (2000) 23±30 We showed that apicidin reverted the morphology of H-ras MCF10A cells from a spindle-like, foci- forming to a ¯attened morphology similar to that induced by other HDAC inhibitors including not only structurally related trapoxin but also trichostatin A and depudecin, which are structurally unrelated. Given that other structurally unrelated HDAC inhibitors have similar effects on cells, these results suggest that HDAC enzyme may be the biologically relevant target protein of the cellular effects of apicidin. Our data showing that apicidin inhibits HDAC activity

(IC50 value of 102 nM) at concentrations lower than that required for detransformation (480 nM for morphological reversal; 190 nM for inhibiting cell

growth by 50%) and anti-invasion (IC50 value of 450 nM) do not apparently support the above sugges- tion. However, it should be considered that the dose response would be different between the in vitro Fig. 6. Apicidin downregulates MMP-2 activity, but not MMP-9, in H-ras MCF10A cells. Gelatin zymogram assay was performed to HDAC activity, which was measured in the isolated detect the enzymatic activities of MMP-2 (72 kDa) and MMP-9 (92 enzyme from the cell, and the detransforming/anti- kDa) secreted by H-ras MCF10A cells treated with various concen- invasive activities, for which apicidin had to penetrate trations of apicidin for 48 h. Relative activities of MMP-2 and into the cell in order to exert any cellular responses MMP-9 were determined by densitometry after gelatin zymograph. including reversal of cell morphology and inhibition Bars represent the SE for triplicate determinations. of invasive phenotype. apicidin in these cells and different mechanisms may Cancer metastasis is a complex, multistep process, possibly be involved. The data demonstrate that apici- which includes invasion of tumor cells through the din inhibits invasive phenotype of H-ras MCF10A surrounding tissues. Tumor cell invasion involves human breast epithelial cells in which MMP-2 is cell motility and extracellular matrix-degrading more likely involved, at least in part, rather than proteinase activity. Undoubtedly, multiple gene MMP-9. products are required for invasive activity. In the current work, we have con®rmed our previous ®nding [14] regarding the association of MMP-2 in H-ras- 4. Discussion induced invasive phenotype by showing that apicidin inhibits invasion and suppresses MMP-2 secretion in A growing number of studies have focused on the H-ras MCF10A cells. In agreement with these results, biological role and therapeutic application of histone an association of MMP-2 activation potential with acetylation in mammalian cells. Mounting evidence metastatic progression in human breast cancer cell suggests a role for acetylation/deacetylation of lines has also been reported [22]. In contrast, associa- histones on the regulation of transcription in eucaryo- tion of MMP-9 with invasive phenotype has been tic cells [4]. Our data clearly demonstrate that apici- shown in some cell systems, including rat ®broblasts din, a fungal product isolated from the fermentations [15], suggesting that the role of MMPs in invasive of Fusarium spp. [1,3], inhibits the H-ras-induced phenotype may be cell type-speci®c. malignant phenotypes including invasion and trans- Members of the activated ras oncogene family can formation in MCF10A human breast epithelial cells. induce complex signaling pathways leading to diverse We suggest that these cellular effects may possibly be cellular responses including proliferation, differentia- through HDAC inhibition. The molecular basis for the tion, apoptosis, transformation, and invasion [23]. Our downstream effects of HDAC inhibition by apicidin present results, along with the morphological rever- remains to be determined. sion of v-ras-transformed ®broblasts by depudecin M.-S. Kim et al. / Cancer Letters 157 (2000) 23±30 29

[24] and oxam¯atin [25] via the inhibition of HDAC, microbial metabolite containing two epoxide groups, exhibits suggests a possible involvement of HDAC in ras- anti-angiogenic activity in vivo, Biol. Pharm. Bull. 18 (1995) signaling pathway. Taken in conjunction with the 1305±1307. [9] Y.B. Kim, K.-H. Lee, K. Sugita, M. Yoshida, S. Horinouchi, fact that uncontrolled ras activation is probably the Oxam¯atinis a novel antitumor compound that inhibits mamma- most common genetic defect in human cancer cells, lian histone deacetylase, Oncogene 18 (1999) 2461±2470. our data showing the anti-invasive and detransform- [10] J. Arts, M. Lansink, J. Grimbergen, K.H. Toet, T. Kooistra, ing activities of apicidin in H-ras-transformed cells Stimulation of tissue type plaminogen activator gene expres- may support a potential use of HDAC inhibitors for sion by sodium butyrate and TSA in human endothelial cells involves histone acetylation, Biochem. J. 310 (1995) 171±176. treatment of cancer. [11] M. Vicente, E. Belinda, P.Y. Graeme, J. Robert, A. Sarah, D.Z. Peter, Induction of caspase-3 protease activity and apoptosis by butyrate and trichostatin A (Inhibitors of histone Acknowledgements deacetylase): dependence on protein synthesis and synergy with a mitochondrial/cytochrome c-dependent pathway, We are grateful to H.-M. Doh (Research Institute of Cancer Res. 57 (1997) 3697±3707. Dong-A Pharm. Co.) for technical assistance and Dr [12] L.A. Liotta, Tumor invasion and metastasis ± role of the extra- H.-R.C. Kim (Wayne State University, School of cellular matrix, Cancer Res. 46 (1986) 1±7. Medicine, MI) for helpful discussion. This work was [13] K. Tryggvason, Type IV collagenase in invasive tumors, Breast Cancer Res. 24 (1993) 209±218. supported in part by a Good Health R and D Grant [14] A. Moon, M.-S. Kim, T.-G. Kim, S.-H. Kim, H.E. Kim, Y.Q. (HMP-96-D-2-1034) and a 1999 research grant from Chen, H.-R. Kim Choi, H-ras, but N-ras, induces and invasive Duksung Women's University. phenotype in human breast ephedial calls: A role for MMP-2 in the H-ras-indused unvasive phenotype, Int. J. Cancer 85 (2000) 176±181. References [15] E.J. Bernhard, S.B. Gruber, R.J. Muschel, Direct evidence linking expression of matrix metalloproteinase 9 (92-kDa [1] S. Darkin-Rattray, A.M. Gurnett, R.W. Myres, P.M. Dulski, gelatinase/collagenase) to the metastatic phenotype in trans- K.M. Crumely, J.J. Allocco, C. Cannova, P.T. Meinke, S.L. formed rat embryo cells, Proc. Natl. Acad. Sci. USA 91 (1994) Colletti, M.A. Bednarek, S.B. Singh, M.A. Goetz, A.W. 4293±4297. Dombrowski, J.D. Polishook, D.M. Schimatz, Apicidin: a [16] H. Soule, T.M. Maloney, S.R. Wolman, W.D. Peterspm, R. novel antiprotozoal agent that inhibits parasite histone deace- Brenz, C.M. McGrath, J. Russo, R.J. Pauley, R.F. Jones, S.C. tylase, Proc. Natl. Acad. Sci. USA 93 (1996) 13143±13147. Brooks, Isolation and characterization of a spontaneously [2] H. Itazaki, K. Nagashima, K. Sugita, H. Yoshida, Y. Kawa- immortalized human breast epithelial cell line, MCF-10, mura, Y. Yasuda, K. Matsumoto, K. Ishii, N. Uotani, H. Cancer Res. 50 (1990) 6075±6086. Nakai, A. Terui, S. Yoshimatsu, Y. Ikenishi, Y. Nakagawa, [17] A. Inoue, D. Fujimoto, Histone deacetylase from calf thymus, Isolation and structural elucidation of new cyclotetrapeptides, Biochim. Biophys. Acta 220 (1970) 307±316. trapoxins A and B, having detransformation activities as anti- [18] E.M. Peter, R. Reeves, J.R. Davie, Sodium butyrate inhibits tumor agents, J. Antibiot. 43 (1990) 1524±1532. histone deacetylation in cultured cells, Cell 14 (1978) 105±113. [3] J.-S. Park, K.-R. Lee, J.-C. Kim, S.-H. Lim, J.-A. Seo, Y.-W. [19] M.Z. Wahrman, S.E. Gagnier, D.R. Kobrin, P.J. Higgins, L.H. Lee, A hemorragic factor (apicidin) produced by a toxic Augenlicht, Cellular and molecular changes in 3T3 cells trans- Fusarium isolates from soybean seeds, Appl. Environ. Micro- formed spontaneously or by DNA transfection, Tumor Biol. 6 biol. 65 (1999) 126±130. (1985) 41±56. [4] M. Grunstein, Histone acetylation in chromatin structure and [20] Y. Hoshikawa, H.L. Kwon, M. Yoshida, S. Horinouchi, T. transcription, Nature 389 (1997) 349±352. Beppu, Trichostatin A induces morphological changes and [5] H. Yoshida, K. Sugita, A novel tetracyclic peptide, trapoxin, gelsolin expression by inhibiting histone deacetylase in human induces phenotypic change from transformed to normal in sis- carcinoma cell lines, Exp. Cell Res. 214 (1994) 189±197. oncogene-transformed NIH3T3 cells, Jpn. J. Cancer Res. 83 [21] M. Kijima, M. Yoshida, K. Sugita, S. Horinouchi, T. Beppu, (1992) 324±328. Trapoxin, an antitumor cyclic tetrapeptide, is an irreversible [6] K. Sugita, K. Koizumi, H. Yoshida, Morphological reversion inhibitor of mammalian histone deacetylase, J. Biol. Chem. of sis-transformed NIH3T3 cells by trichostatin A, Cancer 265 (30) (1993) 22429±22435. Res. 52 (1992) 168±172. [22] H. Azzam, G. Arland, M. Lippman, E. Thompson, Association [7] H.J. Kwon, T. Owa, C.A. Hassig, J. Shimada, S.L. Schreiber, of MMP-2 activation potential with metastatic progression in Depudecin induces morphological reversion of transformed human breast-cancer cell lines independent of MMP-2 ®broblasts via the inhibition of histone deacetylase, Proc. production, J. Nat. Cancer Inst. 85 (1993) 1758±1764. Natl. Acad. Sci. USA 95 (1998) 3356±3361. [23] M. Barbacid, ras genes, Annu. Rev. Biochem. 56 (1987) 779± [8] T. Oikawa, C. Onozawa, M. Inose, M. Sasaki, Depudecin, a 827. 30 M.-S. Kim et al. / Cancer Letters 157 (2000) 23±30

[24] K. Sugita, H. Yoshida, M. Matsumoto, S. Matsutani, A novel [25] H. Sonoda, K. Nishida, T. Yoshida, M. Ohtani, K. Sugita, compound, depudecin, induces production of transformation Oxam¯atin: a novel compound which reverses malignant to the ¯at phenotype of NIH3T3 cells transformed by ras- phenotype to normal one via induction of JunD, Oncogene oncogene, Biochem. Biophys. Res. Commun. 182 (1992) 13 (1996) 143±146. 379±387.