Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Ubiquitin isopeptidases (ubiquitin specific proteases) are a family of cysteine proteases that salvage ubiquitin for its reuse by the 26S proteasome system (Hochstrasser, 1996; Hershko and Ciechanover, 1998) and regulate the activity of a variety of substrates by altering their ubiquitination status. The ubiquitin salvage activity of the isopeptidases maintains a cellular pool of monomeric ubiquitin by cleaving the isopeptide bond between the C-terminal carboxyl of ubiquitin and the -amino group of a lysine residue on an adjacent protein, thereby disassembling ubiquitin oligomers, ubiquitin-protein conjugates, and ubiquitin-peptide conjugates. Few inhibitors of isopeptidases have been identified, other than analogs based on ubiquitin itself. These include nonhydrolyzable ubiquitin dimer analogs (~16 kDa) (Yin et al., 2001) and ubiquitin aldehyde (~8.5 kDa) (Dang et al., 1998), which are suitable for investigating isolated enzymes. Recently, we reported that 12-prostaglandin J₂ (12-PGJ₂) is a novel isopeptidase inhibitor with activity in intact cells. Results with cyclopentenone prostaglandins (PG) prompted our hypothesis that isopeptidase inhibition depends on nuances of olefin-ketone conjugation. For example, 12-PGJ₂, with its cross-conjugated ,-unsaturated dienone substituent and two sterically accessible -carbons (Rodriguez et al., 1997), was a potent inhibitor of isopeptidase activity. PGA₁, PGA₂, and 15-keto-PGs with a simple ,-unsaturated ketone and only one accessible -carbon were significantly less potent. PGB₁ with an ,-unsaturated ketone and a sterically hindered -carbon was inactive (Mullally et al., 2001). In its current formulation, our pharmacophore hypothesis predicts that compounds chemically unrelated to PG, but with a cross-conjugated ,-unsaturated ketone and two sterically accessible -carbons, will also inhibit ubiquitin isopeptidases. Although it disrupts ubiquitin salvage and impairs protein turnover, 12-PGJ₂ induces apoptosis independently of tumor suppressor p53 trans-activation (Mullally et al., 2001). Therefore, our corresponding pharmacological mechanism hypothesis predicts that such isopeptidase inhibitors will also induce cell death independently of tumor suppressor p53 function. Herein, we report the identification of dibenzylideneacetone (DBA), curcumin, and, via the National Cancer Institute's Developmental Therapeutics Database, a diterpene, shikoccin (NSC-302979), as agents that fulfill these predictions and reinforce our pharmacophore and mechanism hypotheses. As isopeptidase inhibitors, DBA, curcumin, and NSC-302979 inhibit the proteasome pathway in a manner chemically and mechanistically distinct from lactacystin (Fenteany et al., 1995), eponemycin (Meng et al., 1999), and peptide-aldehyde or boronate inhibitors (Adams et al., 1999), which all covalently inhibit the 20S catalytic subunit of the proteasome. Our results support the hypothesis that the sterically accessible, cross-conjugated ,-unsaturated dienone is a pharmacophore that confers inhibitory activity toward isopeptidases. Furthermore, these results lend support to the proposition that the ubiquitin-dependent proteasome pathway contains molecular targets suitable for antineoplastic drug discovery (Kisselev and Goldberg, 2000).

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. We used 12-PGJ₂ and PGB₁ (Cayman Chemicals, Ann Arbor, MI); dibenzylideneacetone, etoposide, paclitaxel, Curcumin, and 2,6-diphenyl-4H-thiopyran-4-one (Sigma, St. Louis, MO); NSC-302979 and NSC-156236 (Drug Synthesis and Chemistry Branch, Developmental Therapeutics Program, Division of Cancer Treatment, National Cancer Institute); complete protease inhibitor mixture (Roche Applied Science, Indianapolis, IN); enhanced chemiluminescence reagents (Amersham Biosciences, Piscataway, NJ); antibodies directed against p53 (DO-1), horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology, Santa Cruz, CA), and ubiquitin (Zymed Laboratories, Inc., San Francisco, CA); MG115 (Peptides International, Louisville, KY); z-LLVY-MCA, z-LRGG-MCA (Biomol Research Laboratories, Plymouth Meeting, PA); (3-(4,5-dimethylthiazo)-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Molecular Probes Inc., Eugene, OR); ubiquitin-PEST (Ub-PEST; a gift of Dr. Martin Rechsteiner, Department of Biochemistry, University of Utah, Salt Lake City, UT). Centricon YM-30 centrifugal filters (Amicon Bioseparations; Millipore, Bedford, MA).

Cell Culture. We used RKO and RKO-E6 colon cancer cells (gift from Dr. Mark Meuth, Institute for Cancer Studies, University of Sheffield, Sheffield, UK). We used HCT 116 colon cancer cells with varying degrees of p53 haplosufficiency (gift of Dr. Bert Vogelstein, Johns Hopkins School of Medicine, Baltimore, MD). We maintained RKO and RKO-E6 cells in DMEM [supplemented with 2 mM L-glutamine, 1 mM sodium pyruvate, 50 units/ml penicillin and streptomycin, and 10% (v/v) fetal bovine serum] in a humidified incubator with 5% CO₂. We maintained HCT 116 cells in McCoy's 5A medium [supplemented with 1 mM sodium pyruvate, 50 units/ml penicillin and streptomycin, and 10% (v/v) fetal bovine serum] in a humidified incubator with 5% CO₂.

Immunochemical Detection of Proteins. We removed the medium and lysed cells in 50 mM Tris, pH 7.4, 100 mM NaCl, 2 mM EDTA with 0.1% SDS, 0.1% deoxycholate, 1× complete protease inhibitor mixture. We measured protein concentration by the method of Bradford (1976). We fractionated equal portions of the total cell lysate from each sample (12.5 µg of protein) by SDS-PAGE. We transferred proteins to polyvinylidene difluoride blocked with 5% (w/v) nonfat dry milk in Tris-buffered saline [20 mM Tris-HCl, pH 7.5, 100 mM sodium chloride, 0.1% (v/v) Tween 20]. We detected proteins immunochemically by using primary antibodies directed against p53 (1:4000) or ubiquitin (1:1000), followed by horseradish peroxidase-conjugated secondary antibodies (1:4000). We detected antigen-antibody complexes with enhanced chemiluminescence reagents. We scanned gels and quantified intensities using Kodak 1D Image Analysis Software (Eastman Kodak, Rochester, NY)

Cell Viability Assay. We determined cell viability by the MTT assay. Briefly, we incubated 1 × 10⁵ cells per well of a sterile, 96-well assay plate with 0-60 µM test compounds for 48 h. We added MTT reagent to each well (final concentration, 0.5 mg/ml) and incubated for an additional 3 h. We aspirated the media and remaining MTT reagent from each well and added 100 µl of HCl/isopropanol (1:24). We measured the absorbance of each sample at 405 nm.

Ubiquitin Isopeptidase Activity Assays. We measured cellular isopeptidase enzymatic activity with two assays that used different substrates. In one assay we used Ub-PEST, a full-length ubiquitin molecule with an 18-amino acid C-terminal peptide extension (total mass, 10.5 kDa). Ubiquitin isopeptidases specifically cleave the 18-amino acid peptide extension, releasing full-length ubiquitin (8.5 kDa). Briefly, we incubated 6 × 10⁵ cells with 0-60 µM test compounds for 12 h. We lysed cells in 50 µl of 25 mM HEPES, 5 mM EDTA, 0.1% CHAPS, 5 mM ATP, pH 7.5. We adjusted the protein concentration of each sample to 0.3 mg/ml and incubated with 50 µg/ml Ub-PEST for 45 min at 25°C. Under these conditions Ub-PEST hydrolysis occurs at a linear rate. We mixed 20-µl samples with 20 µl of 2× Laemmli buffer, boiled briefly, and fractionated by SDS-PAGE. We monitored isopeptidase activity by determining the amount of product (8.5-kDa ubiquitin) formation.

Statistics. We used analysis of variance for statistical calculations.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

SubStructure Analysis of the NCI Cancer Screening Database. The Developmental Therapeutics Program (DTP) of the National Cancer Institute (NCI) has systematically evaluated >70,000 compounds for cytostatic and cytotoxic activity against human cell lines in vitro (Monks et al., 1997). The cell lines typify cancers of the colon, blood (leukemia), brain, breast, kidney, lung, ovary, prostate, and skin (melanoma). Intramural NCI investigators, who have access to the entire database, have applied this information-intensive approach with promising results (Weinstein et al., 1997; Shi et al., 1998). Extramural investigators have access to a restricted portion of the database, last released in August 2000. Using these available data, we conducted substructure searches to test our hypothesis that the cross-conjugated ,-unsaturated dienone with two sterically accessible -carbons is the primary molecular determinant that confers the inhibition of isopeptidases. Specifically, we sought nonprostanoid compounds with this feature that varied in the accessibility of their olefinic -carbons (e.g., -carbons with -H versus with -CH₃ substituents).

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Chemical structures of the pharmacophore test panel. We used 2-cyclopenten-5-methylene-1-one (i) as a substructure to search the NCI-DTP database. We identified, and the NCI provided, NSC-156236 (achillin, ii) and NSC-302979 (shikoccin, v). We also identified and used PGB1 (iii), DPTP (iv), 12-PGJ2 (vi), DBA (vii), and curcumin (viii) to test our pharmacophore hypothesis.

Electrophilic Cross-Conjugated Dienones Inhibit Cellular Ubiquitin Isopeptidases. Little is known about the substrate specificity of the individual isopeptidase family members. To investigate total cellular isopeptidase activity, we use two simple substrates, Ub-PEST and z-LRGG-AMC, which most isopeptidases use as substrates. Figure 2A shows the effect of the test panel compounds, 12-PGJ₂ (lanes 3-5), DBA (lanes 6-8), NSC-302979 (lanes 9-11), PGB₁ (lane 12), NSC-156236 (lane 13), curcumin (lanes 15-17), and DPTP (lane 18) on cleavage of the Ub-PEST substrate by isopeptidases in HCT 116 colon cancer cell lines. Similar results were obtained for RKO cells (raw data not shown). Consistent with our pharmacophore hypothesis, the compounds with cross-conjugated ketones and sterically accessible -carbons, 12-PGJ₂, DBA, NSC-302979, and curcumin, each inhibited isopeptidase activity in a concentration-dependent manner (Fig. 2B). Compounds with sterically hindered -carbons (PGB₁, NSC-156236, and DPTP) did not inhibit isopeptidase activity. The rank order of potency for inhibition of ubiquitin-PEST hydrolysis by isopeptidases was DBA  NSC-302979  12-PGJ₂ > curcumin NSC-156236  PGB₁  DPTP.

View larger version (33K): [in this window] [in a new window]   Fig. 2.   Effect of the pharmacophore test panel on ubiquitin isopeptidase activity in colon cancer cells: Ub-PEST as substrate. A, inhibition of isopeptidase proteolysis of Ub-PEST substrate. We treated RKO and HCT 116 cells with DMSO vehicle, 6, 20, or 60 µM 12-PGJ2, DBA, NSC-302979, or curcumin, 60 µM PGB1, 60 µM NSC-156236, or 60 µM DPTP for 12 h at 37°C. We incubated cell lysates from each treatment with Ub-PEST as described under Materials and Methods. We fractionated samples by SDS-PAGE and detected proteins with ubiquitin epitopes immunochemically (HCT 116+/+ Western blot shown as example of raw data). Lane 1 shows the substrate before incubation with lysate. B, inhibition of isopeptidase activity by test panel compounds. We determined isopeptidase activity by measuring, via densitometry, the amount of Ub generated by isopeptidase cleavage of Ub-PEST. The amount of Ub generated in the vehicle treated sample (A, lane 2) is arbitrarily designated 100%.

View larger version (28K): [in this window] [in a new window]   Fig. 3.   Effect of the pharmacophore test panel on ubiquitin isopeptidase activity in colon cancer cells: z-LRGG-AMC as substrate. We treated RKO (upper) and HCT 116+/+ cells (lower) with DMSO vehicle, 6, 20, or 60 µM 12-PGJ2, DBA, NSC-302979, or curcumin, 60 µM PGB1, 60 µM NSC-156236, or 60 µM DPTP for 12 h at 37°C. We incubated cell lysates (0.5 mg/ml) from each treatment with the isopeptidase substrate, z-LRGG-AMC, for 3 h at 37°C. We determined the amount of AMC cleaved by isopeptidase fluorometrically, as described under Materials and Methods.

View larger version (57K): [in this window] [in a new window]   Fig. 4.   Effect of the pharmacophore test panel on ubiquitin-dependent proteolysis in colon cancer cells. A, accumulation of polyubiquitin. We treated RKO-E6 cells with DMSO vehicle for 24 h, 37°C, or with the concentrations of test compounds that gave 80% cell death (as determined by cell viability assay at 48 h) for 6, 12, or 24 h at 37°C [EC80, 12-PGJ2 (13.2 µM), DBA (12.0 µM), NSC-302979 (3.5 µM), or curcumin (17 µM)]. For compounds that were inactive in the cell viability assay, we treated with the highest concentrations examined (60 µM PGB1, NSC-156236, and DPTP for 24 h at 37°C). We fractionated cell lysates by SDS-PAGE and detected proteins with ubiquitin epitopes immunochemically. B, accumulation of p53, a protein targeted for proteasomal degradation. i, Schematic depicting p53 degradation via the proteasome pathway in RKO cells versus RKO-E6 cells. The activity of the HPV E6 oncoprotein hastens p53 degradation in RKO-E6 cells. ii, We treated RKO and RKO-E6 cells with vehicle, etoposide (50 µM), MG115 (20 µM), 12-PGJ2 (60 µM), NSC-302979 (20 µM), DBA (20 µM), PGB1 (60 µM), NSC-156236 (60 µM), curcumin (60 µM), or DPTP (60 µM) for 6 h, 37°C. We fractionated lysates by SDS-PAGE and determined their p53 content immunochemically. iii, We measured p53 levels by densitometry and calculated the ratio of p53 protein in RKO-E6 cells/RKO cells.

The Prototype Isopeptidase Inhibitor, 12-PGJ₂, Inhibits Ubiquitin Isopeptidases Irreversibly. Although we have not yet identified a covalent complex between an isopeptidase and one of the isopeptidase inhibitors, this proposed mechanism of action is consistent with data that we have obtained from analyzing cell lysates treated with 12-PGJ₂. Treatment of cell lysates (versus treatment of whole cells) should exclude the likelihood of transcriptional/translational events due to the nature of the lysate preparation (i.e., sonication of cell lysates probably shears all polynucleotides). 12-PGJ₂ inhibited isopeptidase activity in treated cell lysates (Fig. 5, predialysis). Furthermore, isopeptidase inhibition by 12-PGJ₂ could not be reversed by dialysis of treated lysates (Fig. 5, postdialysis).

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Reversibility of isopeptidase inhibition by 12-PGJ2. We treated cell lysates (0.5 mg/ml) with DMSO vehicle or 100 µM 12-PGJ2 for 1 h at 25°C. We divided each of the treated lysates into two aliquots. We immediately analyzed one aliquot for isopeptidase activity by incubating with z-LRGG-AMC as described previously (PreDialysis). We dialyzed the second aliquot by washing with 3 volumes of assay buffer on a centricon YM-30 column, followed by analysis for isopeptidase activity by incubating with z-LRGG-AMC (PostDialysis).

Inhibitors of Ubiquitin Isopeptidases Cause Cell Death Independently of Tumor Suppressor p53 Function. Previously, we showed that electrophilic prostaglandins, typified by 12-PGJ₂, inhibit p53-mediated transcription under the same conditions in which they cause cell death, suggesting that cell death occurs independently of p53 (Mullally et al., 2001). Therefore, integration of our pharmacophore and molecular mechanism hypotheses predicts that various isopeptidase inhibitors will cause cell death independently of tumor suppressor p53 function. Analysis of the NCI 60 cell line cancer screening data, according to O'Connor et al. (1997), showed that NCI-302979 and curcumin (NSC-32982) act independently of p53 (Table 1). NSC-156236 had no appreciable cytotoxic activity. Thus, data on NSC-302979, NSC-32892, and NSC-156236, available from the DTP public database, fulfill the minimal, initial prediction of our hypotheses.

/

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Cytotoxicity of the pharmacophore test panel in HCT 116 and RKO cell lines with different p53 status. We incubated RKO and RKO-E6 cells (left column), and HCT116 p53+/+ and HCT116 p53/ cells (right column) with vehicle or 0.5-60 µM NSC-302979, DBA, 12-PGJ2, NSC-156236, PGB1, curcumin, or DPTP for 48 h at 37°C. We measured cell viability with MTT reagent as described under Materials and Methods. Data are percentage of control viability, mean ± S.D. (n = 4).

/

/

/

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Cytotoxicity of the calibration set in HCT 116 and RKO cell lines with different p53 status. We incubated RKO and RKO-E6 cells (left column), and HCT116 p53+/+ and HCT116 p53/ cells (right column) with vehicle, 0.01 to 100 µM etoposide, 0.001 to 1 µM paclitaxel, or 0.02 to 2 µM MG115 for 48 h at 37°C and determined their effect on cell viability, as described under Materials and Methods.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Our results indicate that nonprostanoid classes of compounds, with ,-unsaturated ketones and two sterically accessible -carbons, will inhibit ubiquitin isopeptidase activity. We further demonstrate that these compounds cause cell death independently of p53 tumor suppressor function in vitro. Specifically, the diterpene NSC-302979, the synthetic compound DBA, the prostaglandin 12-PGJ₂, and the curcuminoid curcumin all cause cell death with efficacy and potency that is indistinguishable (p > 0.05) between HCT 116 p53^+/+ and HCT 116 p53^/ or RKO and RKO-E6 cells. Furthermore, cell death correlated with inhibition of isopeptidase activity. Regression analysis (IC₅₀ for inhibition of z-LRGG-AMC hydrolysis by isopeptidase versus IC₅₀ for cytotoxicity) yields a straight line with a correlation coefficient r² = 0.93 (n = 7). Regression analysis (IC₅₀ for inhibition of ubiquitin-PEST hydrolysis by isopeptidase versus IC₅₀ for cytotoxicity) also yields a straight line with r² = 0.73 (n = 13).

Inhibition of ubiquitin isopeptidase activity probably propagates cell death by shifting the polyubiquitin chain length equilibrium to one of greater molecular mass. As a consequence of unfettered polyubiquitin chain growth, the pool of monoubiquitin diminishes. Alteration of monoubiquitin/polyubiquitin dynamics inevitably affects several transcription factors other than p53 (Desterro et al., 2000). Furthermore, with depleted monoubiquitin pools, cells are hampered in their efforts to rid themselves of damaged/toxic proteins, eventually affecting protein-protein or protein-DNA interactions that modulate cell survival and apoptosis. Although our data support a covalent mechanism, we are presently investigating whether ,-unsaturated dienones covalently inhibit isopeptidases, specifically, via their electrophilic -carbons (e.g., Michael adduct formation between an isopeptidase cysteine residue and the -carbon of a dienone). Note that compounds with sterically inaccessible or inert -carbons (NSC-156236, PGB₁, and DPTP) were inactive as isopeptidase inhibitors.

We used the substructure search capabilities of the NCI DTP database (60 cell-line screen) to identify NSC-302979 and NSC-156236, compounds used to test our pharmacophore and mechanism of action hypotheses. We think our results, along with the results by NCI scientists (O'Connor et al., 1997; Weinstein et al., 1997; Shi et al., 1998), exemplify the potential of this database and compound repository and the foresight of the NCI Developmental Therapeutics Branch. Others have suggested that the database content is misaligned with the goal to discover new anticancer drugs, based on a poor correlation between clonogenic survival and the NCI archival antiproliferative activity (Brown, 1997). However, direct extension of data acquired in vitro to clinical situations in vivo is rarely straightforward. Used prudently, to enable or to advance mechanistic and pharmacophore hypotheses, the database supports the quest for anticancer drugs with novel structures and mechanisms of action.

Our mechanistic and pharmacophore hypotheses are compatible with the structure-activity relationships reported by Kato et al. (1986), Sasaki et al. (1991) and Sasaki and Fukushima (1994). Kato et al. (1986) reported that 12-PGJ₂ and several related 7-PGA₁ derivatives (all of which are cross-conjugated dienones) increased the life span of Ehrlich ascites tumor-bearing mice: i.p. doses of 20 to 30 mg/kg/day for 5 consecutive days prolonged survival 66 to 111%. In addition, both 12-PGJ₂ and 7-PGA₁ exhibit little cross-resistance with cisplatin and doxorubicin in vivo (Sasaki et al., 1991; Sasaki and Fukushima, 1994). Despite these promising results, 7-PGA₁ is rapidly metabolized to an inactive compound (t_1/2 < 5 min) in serum (Suzuki et al., 1998). Therefore, our discovery of isopeptidase inhibitors among chemical classes other than PG might be advantageous in surmounting any difficulties intrinsic to the antineoplastic development of the PG class.

There exists considerable debate as to whether agents that inhibit the proteasome pathway cause cell death via a p53-independent process (Dietrich et al., 1996; Shinohara et al., 1996; Lopes et al., 1997; Adams et al., 1999; Wagenknecht et al., 1999; An et al., 2000). Our results with inhibitors of ubiquitin isopeptidase activity and with a representative catalytic subunit inhibitor of the 20S proteasome accord with those who conclude that proteasome inhibition causes apoptosis independently of p53. This debate may originate from faulty assumptions about the competence of p53 that accumulates in cells treated with proteasome pathway inhibitors. For instance, genetically wild-type p53 accumulates in the presence of the isopeptidase inhibitor 12-PGJ₂, but in a conformationally and functionally impaired state (Moos et al., 2000; Mullally et al., 2001). An et al. (2000) have also reported that accumulation of wild-type p53 protein and induction of apoptosis occur as independent markers of proteasome inhibition. Therefore, one must use caution when interpreting the consequences of p53 accumulation without first testing its functionality.

The response to chemotherapy is complex; focus on a single factor, no matter how prominent, may exaggerate its role. However, numerous investigations show that disruption of p53 impairs the potency and efficacy of drugs used in oncology [e.g., 5-fluorouracil (Lowe et al., 1994, 1995; Mueller and Eppenberger, 1996; O'Connor et al., 1997; Bunz et al., 1999; Pich, 1998; Weller, 1998; Karpf et al., 2001)]. It is notable that vinca alkaloids, one of the few drug classes that act independently of p53 (O'Connor et al., 1997; Fan et al., 1998), may target the proteasome in addition to tubulin (Piccinini et al., 2001). LDP-341, the first proteasome inhibitor to enter clinical trials, seems to have a favorable safety and efficacy profile (Dalton et al., 2001). Clinical studies to evaluate proteasome inhibition as an adjuvant to systemic chemotherapy are also currently in development (Cusack et al., 2001). Our results demonstrate that another component of the proteasome pathway, isopeptidase activity, warrants further investigation as a target for antineoplastic drug discovery.