Induction of P16ink4a Transcription and of Cellular Senescence by Aclacinomycin-Derivatives and Cardiac Glycosides

April 2002 Biol. Pharm. Bull. 25(4) 461—465 (2002) 461

Induction of p16INK4a Transcription and of Cellular Senescence by Aclacinomycin-Derivatives and Cardiac Glycosides

a a b b a Kiyoshi EGAWA, Yosuke KURIHARA, Tatsuya ITO, Masahiko MATSUMOTO, and Kiyoshi NOSE Department of Microbiology, Showa University School of Pharmaceutical Sciences,a 1–5–8 Hatanodai, Shinagawa-ku, Tokyo 142–8555, Japan and Pharmaceutical Technology Laboratory, Chugai Pharmaceutical Co., Ltd.,b 1–135 Komakado, Gotemba, Shizuoka 412–8513, Japan. Received September 17, 2001; accepted December 24, 2001

Stable transformants of Saos-2 cells that contain the luciferase reporter gene under the control of the human p16INK4a transcriptional regulatory region were established, and were used to identify growth-inhibiting substances from culture broths of actinomycetes and extracts of plants. Among the active substances so far identified were an aclacinomycin-derivative, cenerubin B, and a cardiac glycoside, periplocin. These substances inhibited growth of normal human fibroblasts, and induced senescent phenotypes including a flattened morphology and increased acidic b-galactosidase activity, although the activities of their derivatives to induce p16CDKN2 and b-galactosidase did not coincided with each other. These results suggest that the reporter system using the p16CDKN2 transcriptional regulatory region is a useful means for screening growth inhibiting substances that are potential anti-tumor agents. Key words p16; luciferase; senescence; cinerubin B; periplocin

The p16INK4a, also called CDKN2, and MST1, is a specific upstream region (Ϫ869 to Ϫ1 bases from the cap site) lig- inhibitor of cyclin-dependent kinases CDK4 and CDK6, and ated to the luciferase gene (pGL2 basic, Promega) was kindly a putative tumor suppressor whose gene is often deleted in provided by Dr. E. Hara (Imperial Cancer Research Fund different types of human cancers.1—4) CDKs have critical Laboratories, London). The plasmid was transfected by a roles in cell-cycle progression by phosphorylating the conventional calcium phosphate-precipitation method to- retinoblastoma gene product (pRB),5,6) and their inhibitors gether with pSV2neo, and G418-resistant clones were iso- play important roles in regulation of the cell-cycle. mRNA lated. One of the clones, #13, that stably harbored the re- levels of p16INK4a are elevated in senescent human fibrob- porter and responded well to n-butyrate was used for further lasts,7,8) and forced expression of p16INK4a induced cellular studies. Normal human fibroblasts TIG-7 and transformed senescence in immortalized fibroblasts.9,10) These findings in- 293 cells were cultured in DMEM supplemented with 10% dicate that p16INK4a is involved in cellular senescence in fetal bovine serum. The p21-inducing activity was measured which cell growth is irreversibly arrested. as described previously using a stable transformant, clone Expression of p16INK4a is regulated primarily at the tran- #22.17) scriptional level, the signal for which includes the Ras-Raf- Luciferase Assay Cells were plated into 96-well tissue MEK kinase cascade.11,12) The targets of this signaling to in- culture plates (Falcon) at a density of 3ϫ103 cells/well, and duce p16INK4a have recently been shown to be transcription treated with chemicals 16 h later. Cells were incubated for an factors, Ets1/Ets2, that bind and activate the p16INK4a pro- additional 48 h, and harvested for determination of luciferase moter.13) The Ets family of transcription factors are nega- activity. Assay was carried out using a kit provided by tively regulated by a helix-loop-helix protein Id,14) and in Promega Corp. (Madison, WI) following the manufacturer’s senescent cells the level of Ets1 increased, whereas the Id manual. AP-1/luciferase that contained seven TRE was pur- level was reciprocally decreased.13) Some chemicals such as chased from Promega. sodium butyrate15) and N-acetylcysteine16) are reported to in- SA b-Galactosidase Assay TIG-7 cells were fixed in duce the p16INK4a gene. From the above findings, it could be 2% formaldehyde and were stained in a solution as described expected that agents that activate the p16INK4a promoter previously.18) Blue-stained positive and negative cells were would induce cellular senescence in immortalized cells, and counted under a microscope. are potential anti-tumor agents. We previously reported a Test Samples and Preparation of Extracts Fungi and screening system that used the luciferase reporter gene under actinomycetes isolated from soil samples were cultivated for control of the transcriptional regulatory region of the p21Cip1 5 d at 27 °C in a nutrient medium (glucose 2%, starch 2%, gene, and the system worked to identify several anti-tumor soy bean meal 1%, NaCl 0.3%, dry bouillon 0.1%, CaCO3 agents.17) In the present communication, we extended our 0.3%). To an aliquot of whole broth three volumes of ethanol previous system using a reporter system that contained a were added and the mixture was shaken. After centrifugation, human p16INK4a promoter, and identified several active agents the upper phase was concentrated under vacuum and ex- that induced cellular senescence. tracted with ethyl acetate. Dried plant materials were pulverized with a spice mill. MATERIALS AND METHODS Approximately 10 g of the pulverized plant parts was allowed to stand in 50 ml of 75% aqueous ethanol for 24 h at room Cell Lines Human osteosarcoma cells (Saos-2) were temperature. Extracts of interest were separated by silica gel cultured in a-MEM supplemented with 10% heat-inactivated chromatography using a guide to identify the active material. fetal bovine serum under a humidified atmosphere of 95% Cinerubin B19) and periplocin20) were isolated and purified Ј air/5% CO2. The reporter that contained the human p16 5 - from the culture broth of an actinomycete, strain No. I-60, ∗ To whom correspondence should be addressed. e-mail: [email protected] © 2002 Pharmaceutical Society of Japan 462 Vol. 25, No. 4

Fig. 1. Induction of p16/ and p21/Luciferase Activity by Aclacinomycin-Derivatives Saos-2 stable transformants, #13 (filled bars) and #22 (open bars), were used to measure p16/ and p21/luciferase activity, respectively. Cells were treated with cinerubin B, aclacinomycin A, daunorubicin, or doxorubicin at 1.6 ng/ml (1), 8 ng/ml (2), 40 ng/ml (3), 200 ng/ml (4) or 1 mg/ml (5) for 48 h. Relative luciferase activity based on the untreated control was calculated. The values are the means of three independent experiments. and the root bark of Periploca sepium BUNGE, (Asclepi- Another active substance isolated from plant extracts, adaceae), which is called “Hokugokahi” in Japanese, respec- periplocin, activated p16/luciferase activity at 1 to 100 mg/ml tively. Bufotalin and gincenoside Rgi and Rbi were purchasd (Fig. 2a). As periplocin is a cardiac glycoside, the p16-induc- from Wako Pure Chemical Industries Ltd. (Tokyo), and ing activity was examined using other cardiac glycosides. digoxin, digitoxin and digitoxigenin were from Tokyo Kasei Figure 2a and b show that periplocin as well as bufotalin, gi- Kogyo Co., Ltd. (Tokyo). The other cardiac glycosides (bu- toxin, digitoxin, and digitoxigenin also strongly induced p16 falin, proscillaridin A, oleandrin, lanatoside C, convallatoxin, promoter activity at concentrations as low as 0.1—1 mg/ml. ouabain, gitoxin), aclacinomycin A, daunorubicin and doxo- These agents, however, did not induce the p21/WAF1 pro- rubicin were products of the Sigma Chemical Co. (St. Louis, moter (Fig. 2b) that was activated by histone deacetylase in- MO, U.S.A.). hibitor, n-butyrate, and actinomycin D.17) These cardiac glycosides are an inhibitor of Naϩ/Kϩ-ATPase,25) and induce RESULTS differentiation of leukemia cells26,27) and melanoma cells.27) The relationship between the structure of cardiac glycosides Activation of p16/Reporter by an Anthracycline-Deriv- and the differentiation-inducing activity has not been exten- ative, Cinerubin B, and a Cardiac Glycoside, Periplocin sively determined, but in the case of melanoma differentia- From the screening of about 1000 plant extracts and broths tion, ouabain was more potent than bufalin. In p16-induction, of fungi or actinomycetes, we identified one active substance these two agents had similar activity (Fig. 2a, b), and the in each extract, respectively. Each active substance was puri- mechanisms underlying these two processes may be differ- fied, and determination of its chemical structure revealed ent. Digitalis was proposed to be an anticancer agent.29) periplocin20) from 75% ethanol extracts of P. sepium, and Periplocin and bufotalin are thus expected to be potential cinerubin B (1-hydroxy-aclacinomycin B)19) from the culture chemotherapeutic agents against cancer. broth of an actinomycetes, strain No. I-60. To examine the specificity of induction, the AP-1/lu- As shown in Fig. 1, cinerubin B is a potent inducer of ciferase that is controlled by the TPA-responsive element was p16/reporter at concentrations of 40 to 200 ng/ml, and also used as a reporter at optimal doses to induce p21/luciferase activated p21/luciferase activity. At higher doses, it was toxic reporter. The results shown in Fig. 3 suggest that periplocin to the cells, and hence luciferase activity was decreased. and bufotalin strongly activated expression of AP-1/lu- Other anthracycline-derivatives, doxorubicin, daunorubicin, ciferase but cenerubin B did not. The latter induced p21/lu- and aclacinomycin A, were not as potent as cinerubin B in ciferase but the former did not (Figs.1, 2). These characteris- inducing p16 promoter (Fig. 1). These antibiotics are known tics suggest that the mode of action of these agents is differ- to have anti-tumor activity and are used clinically,21) and also ent, even though they induced p16/luciferase expression. induce differentiation of leukemia cells.22,23) They appear to Induction of Senescent Phenotypes p16 expression is be inhibitors of topoisomerase II,24) but the mode of actions known to be elevated during cellular senescence of normal of cinerubin B has not yet been characterized. From the find- human fibroblasts,9,10) but levels of mRNA of p16 were not ings shown in Fig. 1, it can be expected that cinerubin B has significantly increased in TIG-7 cells treated with either a unique biological action among anthracycline-derivatives. periplocin or cinerubin B as revealed from Northern blotting April 2002 463

Fig. 2. Induction of p16/ and p21/Luciferase Activity by Cardiac Glycosides (a) Cells of clone #13 were treated for 48 h with the indicated agents at 10 ng/ml (1), 100 ng/ml (2), 1 mg/ml (3), 10 mg/ml (4) or 100 mg/ml (5). (b) Clone #22 cells were treated for 48 h with the indicated agents at 10 ng/ml (1), 100 ng/ml (2), 1 mg/ml (3), or 10 mg/ml (4). Relative luciferase activity based on the untreated control was calculated. The values are the means of three independent experiments. 464 Vol. 25, No. 4

(data not shown). The cells treated with these agents, how- rivatives activate signaling pathways that contain stress-acti- ever, showed a flattened morphology that is associated with vated protein kinases, but cardiac glycosides work through senescent cells. As a marker of senescence, expression of distinct pathways. Activated ras is known to induce senes- acidic b-galactosidase activity (SA-Gal) was examined in cence and p16 in normal fibroblasts,11) but the MAP kinase TIG-7 cells treated with cardiac glycosides and aclacino- cascade that is downstream of ras did not seem to be in- mycin-derivatives (Fig. 4). Periplocin, bufotalin, and cineru- volved in p16 induction by cardiac glycosides or anthracy- bin B that were potent inducers of p16/luciferase increased clins (Fig. 5). These results indicate that senescent pheno- the fractions of SA-Gal-positive cells. Induction of this types induced by these chemicals are caused by different sig- senescent phenotypes suggested that these chemicals act as nals from the normal senescent process. anti-tumor agents by inducing cellular senescence. Activities The real targets of transcriptional regulators of the p16 to induce p16/ or p21/luciferase were, however, not directly gene that respond to these agents are unclear at present, but related: aclacinomycin A did not activate p16/reporter but in- our preliminary results suggest that a Ϫ167 bp upstream re- duced induced senescence-like phenotypes, whereas oubain gion of the human p16 gene was sufficient to drive the induc- was a potent inducer of p16/reporter but failed to induce tion of the luciferase reporter (data not shown). Whatever the senescent phenotypes (Fig. 1 and Fig. 4). Even though p16 mechanism, the assay system described in this manuscript gene expression is critical for cellular senescence,7 —10) the will be useful for large-scale screening of potential antitumor mechanism underlying the reporter-activation by drugs may agents. be different from senescent processes. Effect of MAP Kinase Inhibitors on p16-Induction In Acknowledgements This study was supported in part by the case of the induction of p21 by actinomycin D or tricho- a Grant-in-Aids for Cancer Research from the Ministry of statin, an inhibitor of stress-activated protein kinase p38 Education, Sports, Culture, and Science, and a Grant of Re- (SB203580) suppressed the induction. The signaling path- search on Health Sciences focusing on Drug Innovation from ways to induce p16/luciferase were estimated by the addition the Japan Health Sciences Foundation. of specific inhibitors. The results shown in Fig. 5 suggest that SB203580 significantly inhibited the induction caused by REFERENCES cinerubin B or aclacinomycin A, but did not significantly affect the induction by cardiac glycosides. The MAP kinase in- 1) Kamb A., Gruis N. A., Weaver-Feldhaus J., Liu Q., Harshman K., hibitor, PD98059 did not affect the induction by either cate- Tavtigian S. V., Stockert E., Day R. S., III, Johnson B. E., Skolnick M. gory of agents. These results suggest that aclacinomycin-de-

Fig. 3. Induction of AP-1/Luciferase Activity Fig. 4. Induction of Acidic b-Galactosidase Activity HEK 293 cells were transfected with AP-1/luciferase plasmid, and were either un- TIG-7 cells were treated for 4 d with 5 ng/ml of periplocin, bufotalin, bufalin, treated (Ϫ), treated with 50 ng/ml TPA, 250ng/ml of periplocin, bufotalin, or 200 ng/ml proscillaridin A, ouabain, digitoxin ginsenoside, cinerubin B or aclacinomycin A. Cells cinerubin B. Luciferase activity was measured 24 h later. Values are means of three in- were ﬁxed and stained for acidic b-galactosidase activity. Values are means of three independent experiments. dependent experiments.

Fig. 5. Effect of Kinase Inhibitors on p16-Induction Cells of clone #13 were either untreated (1), or treated with 200 ng/ml cinerubin B (2), 1 mg/ml aclacinomycin A (3), 40 ng/ml bufotalin (4) or 200 ng/ml periplocin (5) in the ab- sence (ﬁlled bars) or presence (open bars) of 10 m M SB203580 (a) or PD98059 (b) for 48 h. Luciferase activity was measured, and the values are the means of three independent experiments. April 2002 465

H., Science, 264, 436—440 (1994). 188, 21—30 (1998). 2) Nobori T., Miura K., Wu D. J., Lois A., Takabayashi K., Carson D. A., 16) Liu M., Wikonkal N. M., Brash D. E., Carcinogenesis, 20, 1869— Nature (London), 368, 753—756 (1994). 1872 (1999). 3) Serrano M., Hannon G. J., Beach D., Nature (London), 366, 704—707 17) Egawa K., Nishigori H., Kunimoto S., Takeuchi T., Nose K., Biol. (1993). Pharm. Bull., 21, 899—904 (1998). 4) Liu Q., Neubausen S., McClure M., Frye C., Weaver-Felhaus J., Gruis 18) Dimri G. P., Lee X., Basile G., Acosta M., Scott G., Roskelley C., N. A., Eddington K., Allalunis-Turner M. J., Skolnick M. H., Fujimura Medrano E. E., Linskens M., Rubeli I., Pereira-Smith O., Peacocke F. K., Kamb A., Oncogene, 10, 1061—1067 (1995). M., Campisi J., Proc. Natl. Acad. Sci. U.S.A., 92, 9363—9367 (1995). 5) Sherr C. J., Cell, 73, 1059—1065 (1993). 19) Oki T., Shibamoto N., Matsuzawa Y., Ogasawara T., Yoshimoto A., Ki- 6) Weinberg W. A., Cell, 81, 323—330 (1995). tamura I., Inui T., Naganawa H., Takeuchi T., Umezawa H., J. An- 7) Alcorta D. A., Xiong Y., Hannon G., Beach D., Barrett J. C., Proc. tibiot., 30, 683—687 (1977). Natl. Acad. Sci. U.S.A., 93, 13742—13747 (1996). 20) Stoll A., Renz J., Helv. Chim. Acta, 22, 1193—1205 (1939). 8) Hara E., Smith R., Parry D., Tahara H., Stone S., Peters G., Mol. Cell. 21) Westendorf J., Marquardt H., Ketkar M. B., Marquardt H., Cancer Biol., 16, 859—867 (1996). Res., 43, 5248—5251 (1983). 9) Kato D., Miyazawa K., Ruas M., Starborg M., Wada I., Oka T., Sakai 22) Morin M. J., Sartorelli A. C., Cancer Res., 44, 2807—2812 (1984). T., Peters G., Hara E., FEBS Lett., 427, 203—208 (1998). 23) Schaefer A., Dressel A., Lingelbach K., Schmidt C. A., Steinheider G., 10) Vogt M., Haggblom C., Yeargin J., Christiansen-Weber T., Haas M., Marquardt H., Leukemia, 6, 828—833 (1992). Cell Growth Differ., 9, 139—146 (1998). 24) Rappa G., Lorico A., Sartorelli A. C., Cancer Res., 50, 6723—6730 11) Serrano M., Lin A. W., McCurrach M. E., Beach D., Lowe S. W., Cell, (1990). 88, 593—602 (1997). 25) Sweadner K. J., Biochim. Biophys. Acta, 988, 185—220 (1989). 12) Lin A. W., Barradas M., Stone J. C., van Aelst L., Serrano M., Lowe S. 26) Jing Y., Ohizumi H., Kawazoe N., Hashimoto S., Masuda Y., Nakajo W., Genes Dev., 12, 3008—3019 (1998). S., Yoshida T., Kuroiwa Y., Nakaya K., Jpn. J. Cancer Res., 85, 645— 13) Ohtani N., Zebedee Z., Huot T. J. G., Stinson J. A., Sugimoto M., 651 (1994). Ohashi Y., Sharrocks A. D., Peters G. Hara E., Nature (London), 409, 27) Numazawa S., Shinoki M., Ito H., Yoshida T., Kuroiwa Y., J. Cell. 1067—1070 (2001). Physiol., 160, 113—120 (1994). 14) Yates P. R., Atherton G. T., Deed R. W., Norton J. D., Sharrocks A. D., 28) Zhang L., Yoshida T., Kuroiwa Y., Life Sci., 51, 17—24 (1992). EMBO J., 18, 968—976 (1999). 29) Haux J., Medical Hypothesis, 53, 543—548 (1999). 15) Schwartz B., Avivi-Green C., Polak-Charcon S., Mol. Cell. Biochem.,