Two major pathways for apoptosis have been identified, involving either mitochondria (intrinsic) or tumor necrosis factor (TNF)-family death receptors (extrinsic) as initiators of caspase protease activation and cell death. Because tumor resistance to TNF-family death receptor ligands is a common problem, helping malignant cells evade host immune defenses, we sought to identify compounds that selectively sensitize resistant tumor cells to death receptor ligands. We screened a 50,000-compound library for agents that enhanced anti-FAS antibody–mediated killing of FAS-resistant PPC-1 prostate cancer cell, then did additional analysis of the resulting hits to arrive at eight compounds that selectively sensitized PPC-1 cells to anti-FAS antibody (extrinsic pathway agonist) without altering sensitivity to staurosporine and etoposide (VP-16; intrinsic pathway agonists). These eight compounds did not increase Fas surface levels and also sensitized PPC-1 cells to apoptosis induced by TNF-family member TNF-related apoptosis-inducing ligand, consistent with a post-receptor mechanism. Of these, two reduced expression of c-FLIP, an intracellular antagonist of the extrinsic pathway. Characterization of the effects of the eight compounds on a panel of 10 solid tumor cell lines revealed two structurally distinct compounds that frequently sensitize to extrinsic pathway agonists. Structure-activity relation studies of one of these compounds revealed a pharmacophore from which it should be possible to generate analogues with improved potency. Altogether, these findings show the feasibility of identifying compounds that selectively enhance apoptosis via the extrinsic pathway, thus providing research tools for uncovering resistance mechanisms and a starting point for novel therapeutics aimed at restoring sensitivity of tumor cells to immune effector mechanisms. (Cancer Res 2006; 66(4): 2367-75)
Interest in tumor vaccines is increasing with recent announcements of promising clinical trials. Vaccine strategies rely, in part, on CTLs and natural killer cells to eliminate malignant cells by inducing rapid apoptosis. In part, these immune cells use death receptor ligands such as tumor necrosis factor α (TNFα), FAS-L, and TNF-related apoptosis-inducing ligand (TRAIL) to stimulate specific TNF-family death receptors on tumor target cells, resulting in activation of caspase-family proteases and triggering apoptosis (1–4). Attempts to exploit these immune effector molecules as anticancer agents have resulted in early-stage clinical trials using a recombinant soluble fragment of TRAIL and agonistic monoclonal antibodies targeting TRAIL receptors (reviewed in ref. 5). A limitation of such therapies, however, is acquired or intrinsic resistance to TNF-family death ligands and death receptors, which commonly occurs in advanced malignancies (6). Therefore, small molecules that restore sensitivity of tumor cells to TNF-family death receptors could be useful therapeutic adjuncts to new biological agents such as recombinant TRAIL and tumor vaccines.
TNF-family death receptors trigger apoptosis through a mechanism involving recruitment of certain caspase-family proteases to their cytosolic domains (e.g., caspases 8 and 10 in humans), resulting in formation of a death-inducing signaling complex. Upstream initiator caspases activated at the death-inducing signaling complex then enter the cytosol where they cleave and activate downstream effector caspases, resulting in apoptosis. This mechanism for achieving caspase activation has been dubbed the “extrinsic” pathway, standing in contrast to another apoptosis pathway that involves mitochondria and which has been termed the “intrinsic pathway” (7). Stimuli that activate the intrinsic pathway include DNA-damaging anticancer drugs, γ-irradiation, hypoxia, and growth factor deprivation, causing mitochondria to release cytochrome c and other apoptogenic proteins into the cytosol, resulting in caspase activation (8).
Diverse mechanisms can create roadblocks to apoptosis within the extrinsic or intrinsic pathways, occurring commonly in many cancers during tumor progression and thus creating impediments to successful treatment. Documented resistance mechanisms relevant to the extrinsic pathway include reduced expression of TNF-family death receptors, shedding of soluble death receptors and expression of ligand-binding decoy receptors, reduced expression of caspases 8 and 10, and overexpression of intracellular caspase inhibitors (reviewed in ref. 9). Among the endogenous caspase inhibitors affecting the extrinsic pathway is c-FLIP, a protein resembling caspases 8 and 10, which can bind and prevent their activation at the death-inducing signaling complex (10, 11).
We sought to identify compounds that selectively modulate the extrinsic pathway, sensitizing resistant tumor cells to TNF-family death receptors and death ligands. To this end, we established a cell-based screen using the prostate cancer cell line PPC-1. PPC-1 cells are resistant to apoptosis induced by TNF-family death ligand TRAIL and to agonistic antibodies targeting TNF-family death receptor FAS (CD95) despite expressing FAS and TRAIL receptors on their surface and despite expressing the requisite intracellular caspase activation machinery, including adapter protein Fas-associated death domain and procaspase-8 (12). We reasoned therefore that PPC-1 cells are suitable for identification of compounds that restore sensitivity to the extrinsic pathway and devised a primary screen and subsequent secondary and tertiary screens to arrive at compounds that selectively modulate the extrinsic pathway.
Materials and Methods
Reagents. The 50,000-compound Diversa chemical library was obtained from Chembridge (San Diego, CA). The anti-FAS monoclonal antibody CH-11 was purchased from MBL Co. Ltd. (Nagoya, Japan). TRAIL was obtained from Alexis (San Diego, CA). VP-16 and staurosporine were purchased from Sigma, Inc. (Milwaukee, WI). 2-Cyano-3,12-dioxooleana-1,9-dien-28-oic acid (CDDO) was a generous gift from Michael Sporn (Dartmouth University, Hanover, NH).
Cell lines. Cell lines were maintained in RPMI 1640 supplemented with 2.5% to 10% FCS (Hyclone, Tulare, CA), 1 mmol/L l-glutamine, and antibiotics (streptomycin/penicillin). Cells were cultured at 37°C in a humid atmosphere with 5% CO2.
High-throughput screening. Screens were done using a fully integrated, programmable robotic liquid handling system (Biomek FX, Beckman-Coulter, Inc., Fullerton, CA) with integrated plate reader (LJL analyst HT 96-384, Sunnyvale, CA) and environmentally controlled plate carousel set at 37°C and 5% CO2:95% air. PPC-1 cells (1 × 104) were seeded overnight into 96-well flat-bottomed plates (Costar, Cambridge, MA) in 100 μL of medium containing 2.5% FCS. The next day, aliquots from the 50,000-compound Diversa library were added at a final concentration of 7.5 μg/mL (∼25 μmol/L) in a final concentration of 0.5% (v/v) DMSO. CH-11 antibody (100 ng/mL) was then added and the cells were incubated for 24 hours before assessing cell viability by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) dye reduction assay (Sigma).
Cell death assays. Cell viability was measured by MTT and 3-(4,5-dimethyl-thiazol-2yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) assays essentially as previously described (13). Absorbance readings were plotted against a standard curve to derive the corresponding cell number and cell viability was expressed as a percentage relative to untreated cells. Apoptosis was measured by flow cytometric analysis of Annexin V surface expression after staining cells with FITC-anti-Annexin V and propidium iodide (Biovision, Mountain View, CA) as previously described (14).
Cell transfections. PPC-1 cells (2 × 105) were seeded in 35-mm diameter plates in RPMI with 10% FCS. The next day, the cells were cotransfected using Lipofectamine Plus (Invitrogen, Carlsbad, CA) with 0.5 μg green fluorescent protein (GFP)–encoding plasmid pEGFP (Invitrogen) in combination with 1.5 μg of plasmids encoding Bcl-xL, the viral caspase-8 inhibitor CrmA, or empty vector. At 2 days posttransfection, cells were incubated with various concentrations of CH-11 antibody and the test compounds for 24 hours; then percentage apoptosis was scored by UV-microscopic analysis of the GFP-positive cells, counting a minimum of 200 cells. Cells that had rounded up and were floating in the medium were counted as nonviable whereas cells that remained adherent to the plate with normal morphologic features were counted as viable.
Immunoblot analysis. Protein extracts were obtained by washing cells with PBS (pH 7.4) and suspending in lysis buffer [10 mmol/L Tris (pH 7.4), 150 mmol/L NaCl, 0.1% Triton X-100, 0.5% sodium deoxycholate, and 5 mmol/L EDTA] containing protease inhibitors (Complete tablets; Roche, Indianapolis, IN). Immunoblot assays were done as previously described (15). Briefly, equal amounts of protein, as determined by Bradford assay (16), were subjected to SDS-PAGE (4-20% gradient gels; ISC BioExpress, Kaysville, UT), followed by transfer to nitrocellulose membranes. Membranes were incubated with 1:500 (v/v) mouse monoclonal anti-human FLIP (NF6 clone; Alexis), 1:1,000 (v/v) anti–caspase-8 clone 5F7 (Upstate), and 1:2,000 (v/v) mouse monoclonal antitubulin (Sigma). Secondary antibodies consisted of horseradish peroxidase–conjugated goat anti-rabbit IgG or goat anti-mouse IgG (Bio-Rad, Hercules, CA). Detection was done by the enhanced chemiluminescence method (Pierce, Rockford, IL).
Transfection of siRNA oligonucleotides. Double-stranded SMARTPOOL siRNA oligonucleotides targeting c-FLIP mRNA and double-stranded firefly luciferase control siRNA (Dharmacon Research, Lafayette, CO; 10 nmol/L) were transfected into cells with Lipofectamine according to the instructions of the manufacturer.
Quantitative reverse trancription-PCR. The cDNAs encoding the long isoform of FLIPL and GAPDH were amplified using the following primer pairs: 5′-CCTAGGAATCTGCCTGATAATCGA-3′ (forward primer for FLIP), 5′-TGGGATATACCATGCATACTGAGATG-3′ (reverse primer for FLIP), 5′-GAAGGTGAAGGTCGGAGTC-3′ (forward primer for GAPDH), and 5′-GAAGATGGTGATGGGATTTC-3′ (reverse primer for GAPDH). Equal amounts of cDNA for each sample were added to a prepared master mix (SYBR Green PCR Master mix, Applied Biosystems, Foster City, CA). Real-time quantitative PCR reactions were done on an ABI Prism 7700 sequence detection system (Applied Biosystems). The relative abundance of a transcript was represented by the threshold cycle of amplification (CT), which is inversely correlated to the amount of target RNA/first strand cDNA being amplified. To normalize for equal amounts of the latter, we assayed the transcript levels of the putative housekeeping gene GAPDH. The comparative CT method was calculated as per instructions of the manufacturer. The normalization of the CT of FLIP for each sample was calculated as CT(FLIP) − CT(GAPDH). The expression level of FLIP relative to the baseline level was calculated as 2−ΔΔCT(FLIP), where ΔCT is (average FLIP CT − average GAPDH CT) and ΔΔCT is (average ΔCT untreated sample − average ΔCT treated sample).
Statistics. Cytotoxicity induced by compounds used in combination with conventional agents (e.g., CH-11, TRAIL, VP-16, and staurosporine) was evaluated for evidence of synergy toxicity by comparing the slopes of the dose-response curves. If the combination of the potential sensitizing compound with the conventional agent increased the slope of the dose-response curve compared with the slopes of either the sensitizer or the conventional agent alone, then the interaction was considered synergistic. If the slopes of the curves were not significantly different, we concluded that the enhanced toxicity was additive but not synergistic. Statistical significance was defined as P < 0.01 using two-sided tests. Synergy was confirmed by multiple drug dose-effect calculations using the Median Effect methods as previously described (17) Combination index plots were generated using the Calcusyn software (Biosoft, Ferguson, MO). Combination index < 1.0 indicates a more than expected additive effect (synergism).
Identification of small-molecule FAS sensitizers. Resistance to death receptor ligands may permit malignant cells to escape immune surveillance and limit the clinical efficacy of recombinant death receptor ligands such as TRAIL. To identify small molecules that restore sensitivity to death receptor ligands, a cell-based high-throughput screen was done using the FAS- and TRAIL-resistant prostate cancer cell line PPC-1 and a commercially available 50,000-compound library. The screens were done in 96-well plates to which compounds were added at 7.5 μg/mL (∼25 μmol/L), followed by agonistic anti-FAS monoclonal antibody CH-11 (100 ng/mL). Cell viability was measured 24 hours later by MTT assay. Each plate included controls of untreated cells, cells treated only with CH-11, and cells treated with a positive control compound, CDDO, previously determined to sensitize PPC-1 cells to TNF-family death receptors and ligands (12). The coefficient of variation for PPC-1 cells treated with CH-11 alone was determined to be 5% based on 90 replicate determinations. A 50% decrease in cell viability was used as a cutoff for scoring hits.
From the primary screen of 50,000 compounds, 313 reproducible hits were obtained. Figure 1 shows the overall workflow plan used to evaluate hits. To determine whether these 313 compounds were toxic molecules as opposed to FAS sensitizers, the 313 compounds were evaluated in secondary screens where PPC-1 cells were treated with increasing concentrations of the compounds in the presence or absence of CH-11 antibody. Through these secondary screens, nine sensitizers were identified that increased CH-11-mediated killing above the cell death produced by treatment of the cells with the compound alone (Supplementary Table S1 and Fig. 2). In contrast, the remaining 304 compounds displayed toxicity as single agents and did not potentiate CH-11 killing.
The molecules identified in this screen included compounds 4-(4-chloro-2-methylphenoxy)-N-hydroxybutanamide (5809354) and N-[4-chloro-3-(trifluoromethyl)phenyl]-3-oxobutanamide (6094911), which displayed little direct toxicity at concentrations up to 80 μmol/L but which sensitized PPC-1 cells to 100 ng/mL CH-11 with LD50 of 20 ± 2 and 35 ± 4 μmol/L, respectively (Fig. 2). In contrast, other compounds such as 1-[(4-chlorophenyl)acetyl]-4-[(2E)3-phenylprop-2-enyl]piperazine (5703229) were directly toxic with LD50 of 50 ± 5 μmol/L but sensitized to CH-11 at lower concentrations with an LD50 of 34 ± 4 μmol/L (Fig. 2). We confirmed that compounds 5809354 and 6094911 showed synergy when combined with CH-11 in PPC-1 cells using the median-dose combination index (ref. 17; Fig. 3). The mean combination index at LD50, LD75, and LD90 of 5809354 and 6094911 when combined with CH-11 was 0.1 ± 0.04 and 0.01 ± 0.01, respectively, where combination index < 1 denotes synergy. However, compound 5569100 did not show synergy with CH-11 in PPC-1 cells with this method, likely reflecting the narrow range of concentrations over which 5569100 enhanced CH-11 killing.
Effects of small-molecule sensitizers on CH-11 killing of a spectrum of malignant cell lines. To assess the spectrum of activity of the small-molecule FAS sensitizers, an additional nine malignant cell lines derived from breast, ovarian, and prostate carcinomas were treated with increasing concentrations of the sensitizers in the presence or absence of CH-11. Compound 6094911 sensitized 4 of 10 tumor cell lines to CH-11, including OVCAR-3 ovarian, T47D breast, HT29 colon, and PPC-1 prostate cancer cells. Compound 5809354 sensitized the same tumor lines with the exception of HT29 (Fig. 4A and data not shown). None of the other compounds sensitized more than 4 of 10 tumor lines to CH-11. Compounds 6094911 and 5809354 failed to sensitize SKOV-3 ovarian, Colo 205 colon, MDA-MB-468 breast, and DU 145 and LNCAP prostate cancer cells. Consistent with the difference among tumor cell lines in sensitivity to 6094911 and 5809354, median-dose combination index plots showed synergy of these compounds with a sensitive cell line (PPC-1) but not an insensitive cell line (DU 145; Fig. 3). Except for MDA-MB-468, all of the nonresponding cell lines expressed the FAS receptor on the cell surface, as measured by flow cytometry using a specific fluorescinated antibody, and all of the nonresponding cell lines, except Colo 205, had detectable levels of FLIP by immunoblotting (data not shown).
Identification of small molecules that specifically sensitize to extrinsic pathway. The nine compounds identified through this screen using anti-FAS antibody theoretically could nonspecifically sensitize tumor cells to apoptotic stimuli or they could be selective for the extrinsic pathway. To distinguish between these two possibilities, we compared the effects of compounds on apoptosis induced by extrinsic pathway stimuli (e.g., CH-11 antibody and TRAIL) versus intrinsic pathway stimuli (e.g., VP-16 and staurosporine). Accordingly, PPC-1 cells were treated with the various concentrations of compounds with or without CH-11, TRAIL, VP-16, or staurosporine, and 24 hours later, cell viability was measured by MTT assays. Of the nine candidate compounds, eight sensitized PPC-1 cells to FAS and TRAIL (death receptor pathway stimuli) but not to VP-16 or staurosporine (intrinsic pathway stimuli), suggesting they selectively modulate the extrinsic pathway. Representative experiments with compounds 6094911, 5809354, 5703229, and 5569100 are shown in Fig. 4B. In contrast, compound 5362611 sensitized to both the death receptor (extrinsic) and mitochondrial pathway (intrinsic) stimuli (Fig. 1), suggesting it operates downstream at the point of convergence of these two apoptotic pathways. Similar experiments were done for compounds 6094911, 5809354, and 5569100 using OVCAR-3 and T47D tumor lines, confirming that the results are not unique to PPC-1 cells (Fig. 4C and data not shown).
Caspase-dependent induction of apoptosis by combination treatment with CH-11 and FAS-sensitizing compounds. To further assess the mechanism of the eight compounds that selectively modulated tumor sensitivity to extrinsic pathway stimuli, we tested the effects of benzoyl-Val-Ala-Asp-fluoromethylketone (zVAD-fmk), a broad-spectrum irreversible inhibitor of caspase-family proteases (Enzyme Systems, Dublin, CA). Accordingly, PPC-1 cells were treated with CH-11 and sensitizers such as 6094911, 5809354, and 5569100 with and without zVAD-fmk (100 μmol/L) for 12 hours. Apoptosis was then measured by Annexin V staining. Consistent with a caspase-dependent mechanism of action, zVAD-fmk blocked sensitization to CH-11 (Fig. 5A). Likewise, the caspase-8 inhibitory compounds acetyl-isoleucinyl-glutamyl-threoninyl-aspartyl-fluoromethylketone (Ac-IETD-fmk; Calbiochem, San Diego, CA) also inhibited apoptosis induced by CH-11 in combination with these FAS-sensitizing compounds (data not shown).
To further address the specificity of the compounds for the extrinsic pathway, the effects of selective apoptosis inhibitory genes were examined. For these experiments, PPC-1 cells were transfected with plasmids encoding CrmA, a viral protein that blocks the extrinsic pathway by inhibiting caspase-8 (18) or encoding Bcl-xL, a mitochondria-targeting protein that blocks the intrinsic pathway by inhibiting release of cytochrome c (19, 20). After transfection, cells were treated with CH-11 antibody and compounds 6094911, 5809354, or 55619100, scoring the percentage of apoptotic cells among the successfully transfected cells as determined by cotransfection of a GFP-marker plasmid.
CrmA almost completely protected against apoptosis induced by the combination of CH-11 and either 6094911, 580935, or 5569100 whereas Bcl-xL had no protective effect (Fig. 5B). As a positive control, transfection with Bcl-xL, but not with CrmA, protected PPC-1 cells from staurosporine-induced apoptosis. Taken together, these results indicate that molecules such as 6094911, 5809354, and 5569100 specifically target the death receptor (extrinsic) pathway proximal to its convergence with the mitochondrial (intrinsic) pathway at the level of downstream effector caspases. Furthermore, because these compounds sensitize to both FAS and TRAIL, they presumably act distal to death receptors. Supporting this hypothesis, treatment of PPC-1 cells with 6094911, 5809354, or 5569100 did not change the surface expression of the FAS or TRAIL receptors as measured by flow cytometry using specific fluorescinated antibodies (data not shown).
Effects of sensitizing compounds on FLIP protein. Because FAS-sensitizing compounds 6094911 and 5809354 modulate the extrinsic pathway downstream of TNF-family death receptors but upstream of effector caspases, we interrogated the effects of our compounds on expression of FLIP, an intracellular antiapoptotic protein that binds caspases 8 and 10 and that is capable of suppressing death receptor signaling at a proximal point within the extrinsic pathway. For these experiments, PPC-1 cells were treated with sensitizing compounds in the presence or absence of CH-11 antibody; cell lysates were prepared 24 hours later and then analyzed by immunoblotting for FLIP (Fig. 6A). Comparisons were made with the triterpenoid CDDO, which has previously been shown to reduce FLIP protein levels and restore sensitivity of tumor cells to FAS and TRAIL (12), thus serving as a positive control. The levels of the short isoform of FLIP were <5% of total FLIP protein as measured by immunoblotting and quantitation with densitometry. Therefore, given the very low levels of the short isoform, we restricted our analysis to the effects of the compounds on the long isoform of FLIP. Compounds 5809354 and 5569100 decreased levels of FLIP protein. Pretreatment with z-VAD-fmk did not prevent reductions in FLIP protein, indicating that the changes in FLIP were not a secondary event mediated by caspase activation (data not shown). In contrast, the other FAS-sensitizing compounds did not alter FLIP expression, indicating that they act through different mechanisms. Similar reductions in FLIP protein were observed after treating OVCAR-3 and T47D cells with 5809354. Likewise, levels of FLIP were decreased in the nonresponding cell lines DU 145 and MDA-MB-468 after treatment with 5809354 (Fig. 6B). Whereas these compounds reduced levels of FLIP protein, no change in levels of Bcl-2, Fas-associated death domain, caspase-8, or Mcl-1 was observed after treatment with these compounds (data not shown).
To determine whether the reductions in FLIP protein occurred at the level of the mRNA or protein, we measured expression on FLIP mRNA by quantitative reverse trancription-PCR (RT-PCR) in cells treated with 5809354 or buffer control. Treatment with 5809354 reduced expression of FLIP mRNA in the responding and nonresponding cell lines (Fig. 6C). In contrast, compounds 5569100 and 6094911 did not reduce FLIP mRNA (data not shown).
To determine whether the compounds activate the death-inducing signaling complex in the presence of CH-11, processing of caspase-8 was analyzed by immunoblotting (Fig. 6D). Treatment of cells with the sensitizers and CH-11 activated caspase-8 as evidenced by a decrease in the proform of caspase-8 and an increase in the cleaved form. 5890354 and 6094911 did not sensitize DU 145 or MDA-MB 468 cells to CH-11 and caspase-8 activation was also not detected. In contrast, 5569100 activated caspase-8 in these cell lines, suggesting it acts through a mechanism distinct from the other sensitizers.
To test the functional importance of decreases in FLIP by 5809354, we examined whether knocking down FLIP could recapitulate the activity of the molecule and abrogate the ability of 5809354 to sensitize cells to CH-11. PPC-1 cells were transfected with double-stranded siRNA that targeted FLIP or luciferase as a control. At 24 hours after transfection, cells were treated with increasing concentrations of 5809354 with or without CH-11. FLIP siRNA, but not luciferase control, sensitized PPC-1, OVCAR-3, and T47D cells, but not MDA-MB-468 and DU 145 cells, to CH-11, thereby recapitulating the effects of 5809354 (Fig. 6E and F). Furthermore, in the presence of FLIP siRNA, 5809354 no longer enhanced CH-11-mediated killing (Fig. 6G). Thus, taken together, the data suggest that decreases in FLIP by 5809354 are functionally important. In contrast, reductions in FLIP by 5569100 did not seem to be functionally important in as much as 5569100 was directly toxic to DU 145 and MDA-MB 468 cells with an LD50 of 11 ± 0.5 and 7 ± 0.2 μmol/L, respectively. However, FLIP siRNA did not reduce the viability of these cells and did not sensitize them to CH-11.
Structure-activity relation analysis. To begin exploring the relationship between the structure and function of the FAS sensitizing compounds, a series of structurally related analogues of 6094911 were evaluated. PPC-1 cells were treated with the increasing concentrations of the analogues, with and without CH-11 (100 ng/mL), and cell viability was measured 24 hours later, determining approximate LD50 values (Table 2). Comparison of the analogues indicated that modifications of the 4-chloro-3-(trifluoromethyl)-phenyl group abolished or significantly reduced activity of the compounds. Conversely, various substituents are well tolerated in the R position (Supplementary Fig. S1 and Supplementary Table S2). These results suggest that the N-[4-chloro-3-(trifluoromethyl)-phenyl]-amide moiety contains a critical functional pharmacophore for the activity of 6094911 whereas the methyl-keto moiety is expendable (Supplementary Table S2).
Most high-throughput screens are designed to identify molecules that interact with specific protein targets. In contrast, the study described here used a chemical biology approach to identify molecules that reverse the phenotype of FAS resistance. With our cell-based, high-throughput assay, we identified nine compounds from a library of 50,000 that reversed resistance of PPC-1 cells to CH-11 anti-FAS antibody. The molecules differed in their dose-response curves, with some compounds displaying FAS-independent toxicity at higher doses while enhancing death receptor–mediated killing at lower concentrations; some compounds such as 5934859 reversed FAS resistance only for PPC-1 cells whereas other compounds such as 6094911 were more broadly acting, sensitizing 4 of 10 tumor lines to extrinsic pathway stimuli. These differences among compounds likely reflect different mechanisms of action and different cellular targets, combined with differences in FAS resistance mechanisms among tumor cell lines. Of course, variations in uptake and metabolism of compounds may also contribute to the heterogeneous responses among tumor lines.
Of the nine sensitizers to CH-11 identified by this screen, eight sensitized to the extrinsic pathway agonists CH-11 and TRAIL but not to cell death stimuli that trigger the intrinsic pathway, such as VP-16 and staurosporine. Furthermore, FAS sensitization by these eight compounds was inhibited by CrmA but not by Bcl-xL, consistent with a selective effect on the extrinsic pathway. These results indicate that these compounds act selectively on targets in the extrinsic pathway, operating distal to death receptors but proximal to downstream effector caspases.
Previous studies have shown that PPC-1 cell resistance to FAS and TRAIL can be reversed by decreasing the levels of the caspase-8 inhibitor FLIP using antisense oligonucleotides or the triterpenoid CDDO, which reduces FLIP expression (12). Therefore, we assessed the effects of our compounds on levels of FLIP protein. Two of the compounds identified decreased levels of FLIP protein. The decrease in FLIP was not secondary to caspase activation based on experiments using broad-spectrum caspase inhibitor zVAD-fmk. Reducing expression of endogenous FLIP using siRNA-based gene silencing recapitulated the ability of 5809354 to sensitize tumor cells to CH-11 and abrogated the ability of 5809354 to sensitize tumors further to CH-11. In contrast, in the nonresponding cell lines MDA-MB 468 and DU 145, FLIP siRNA and 5809354 reduced levels of FLIP protein but did not sensitize cells to CH-11, suggesting that FLIP is not an important mechanism of resistance to CH-11 in those cell lines or that it is not the only mechanism of resistance. This result is in keeping with our observation that MDA-MB 468 cells do not express the FAS receptor on their cell surface. Taken together, these data argue that FLIP is an important target of 5809354 because the absence of the target nullifies the actions of the 5809354 with respect to FAS sensitization. However, we cannot entirely exclude the possibility that 5809354 has additional mechanisms of action that promote sensitization to CH-11 independent of its effects on FLIP.
5809354 decreased FLIP mRNA and thus seems to act through a mechanism different than CDDO that reduces FLIP protein by promoting its ubiquitination (12). The mechanism by which 5809354 reduces FLIP mRNA is under investigation. Of note, other extrinsic pathway modulating compounds such as 6094911 did not alter levels of FLIP protein, indicating that they act through different mechanisms. 5569100 reduced FLIP protein but not FLIP mRNA. However, FLIP siRNA did not recapitulate the effects of 5569100. Thus, reductions in FLIP do not seem to be functionally important for the toxicity of 5569100 and this compound seems to act through a mechanism distinct from 5809354.
A preliminary analysis was conducted to determine functional groups important for the activity of one of the compounds, 6094911, which enhanced FAS-induced killing of several tumor cell lines. For compound 6094911, the N-[4-chloro-3-(trifluoromethyl)-phenyl]-amide moiety, but not the methyl-keto moiety, seemed to be necessary for activity. Accordingly, compound 6094911 could be derivatized at this methyl-keto moiety with affinity labels (i.e., biotin) for identifying its cellular targets. Furthermore, the initial structure-activity relation data reported here provide some guidance for the design and synthesis of second-generation compounds that may be more potent than the parent and potentially more amenable to clinical use.
In summary, we have identified compounds that sensitize a spectrum of resistant cancer cells to death receptor ligand stimulation. These compounds may serve as prototypes for development of novel therapeutic adjuncts for the treatment of malignancy based on immune-based treatments such as recombinant TRAIL, agonistic anti-TRAIL antibodies, or tumor vaccines. In addition, these compounds provide new research tools for chemical biological experiments aimed at understanding mechanisms of resistance to TNF-family death ligands and death receptors.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
A.D. Schimmer is a Canadian Institutes of Health Research Clinician Scientist.
Grant support: NIH, Canadian Institutes of Health Research, the Ontario Cancer Research Network, and the Elsa U. Pardee Foundation.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank J. Valois and P. DeLuca for manuscript preparation, M. Hanii for figures, and M. Sporn for CDDO.