Background
The extracellular signal-regulated kinases-1 and 2 (ERK1/2) proteins are members of the mitogen activated protein (MAP) kinase superfamily that regulate cell proliferation and survival. ERK1/2-mediated cell survival occurs through protection against apoptosis by inactivating pro-apoptotic proteins. For example, ERK proteins promote cell survival by inhibiting caspase-9 [
1,
2] or Bim (Bcl-2-interacting mediator of cell death) through direct phosphorylation [
3]. Indirect inhibition of apoptosis occurs through ERK phosphorylation and activation of p90Rsk-1, which phosphorylates the pro-apoptotic Bad (Bcl-xL/Bcl-2 associated death promoter) protein and causes 14-3-3-mediated sequestering that prevents interactions with the pro-survival protein Bcl-2 [
4,
5]. Thus, constitutive activation of the ERK1/2 pathway through mutations in upstream receptors, Ras G-proteins, and kinases, such as B-Raf, provides transformed cancer cells with a survival advantage [
6‐
8].
Significant effort has gone into developing molecules that inhibit proteins in the ERK1/2 pathway [
9,
10]. These drug discovery efforts include monoclonal antibodies and small molecules that inhibit receptor tyrosine kinases, Ras G-proteins, Raf, or MEK proteins [
9,
11‐
13]. Although some of these therapies have shown promising clinical results, toxicity to skin, cardiac, and gastrointestinal tissue has been reported [
14,
15]. The toxicity associated with upstream inhibition of ERK1/2 signaling is likely due to the effects on the ERK pathway in normal tissue and the various ERK1/2 substrates that regulate cellular functions [
6,
16]. Thus, inhibition of specific ERK functions, such as regulation of pro-apoptotic proteins, may be an alternative approach to alleviating toxic side effects resulting from complete inhibition of ERK signaling by compounds targeting upstream proteins. To test this, we have identified molecules that act independent of the ATP binding site and are predicted to be selective for ERK1/2 substrate docking domains [
17,
18]. By developing compounds that are substrate selective, our goal is to inhibit ERK functions that are associated with cancer cell survival but preserve ERK functions in normal non-cancerous cells.
ERK1/2 are proline-directed serine/threonine kinases that phosphorylate substrate protein sequences containing, at minimum, a proline in the +1 position (S/T
P site). Proline in the -2 position (
PXS/TP sequence) may also determine phosphorylation specificity [
19]. While this consensus sequence is shared by the other MAP kinases proteins, including p38 MAP kinases, c-Jun N-terminal kinases (JNKs), and ERK5, each MAP kinase retains substrate specificity suggesting that other determinants of kinase-substrate interactions are involved. Currently, two distinct docking domains on substrates have been identified to mediate interactions between protein substrates and MAP kinases [
19‐
22]. The D-domain or DEJL site (
docking site for
ERK or
JNK,
LXL), consists of two or more basic residues, a short peptide linker, and a cluster of hydrophobic residues. ERK1/2 substrates containing D-domains include ELK-1, p90Rsk-1, MKP-3, and caspase-9 [
1,
23,
24]. D-domains have been found on substrates for ERK, JNK, and p38 MAP kinases [
25,
26]. MAP kinase substrates may also contain an F-site or DEF (
docking site for
ERK,
FXF) motif, which contains the consensus FXFP motif. The F-site is 6-20 amino acids C-terminal to the phosphorylation site [
19] is also found on ELK-1 as well as substrates like KSR and nucleoporins [
27].
Specific residues on MAP kinases form docking domains that determine binding specificity with substrate proteins. ERK1/2 and other MAP kinases contain a common docking (CD) domain, which includes aspartate residues 316 and 319 (labeled for ERK2) that are located on the side opposite of the TXY activation loop [
25] and mediates interactions with the substrate D-domains [
27,
28]. While the CD domain shares common features among MAP kinases, differences in the CD domains and adjacent residues of ERK1/2 and p38 MAP kinases may be responsible for determining the specificity of substrate interactions [
29]. The F-site containing substrates are thought to form hydrophobic interactions with ERK1/2 on a F-site recruitment site (FRS), which consists of leucines 198, 232, 235 and tyrosines 231 and 261 residues (labeled for ERK2) located near the TXY activation site [
27].
We have previously utilized computer aided drug design (CADD) to identify low-molecular weight (LMW) compounds that are predicted to interact near the ERK2 CD domain and selectively disrupt ERK2 interactions with substrate proteins [
17,
18]. These studies identified several compounds that inhibited ERK1/2's ability to phosphorylate selected substrate proteins but not affect the overall activity of the ERK enzyme supporting the goal of using small molecules to inhibit productive protein-protein interactions. Moreover, some of the compounds tested did not appear to affect the ability for the related p38 MAP kinase to phosphorylate substrate proteins [
18]. In addition to inhibiting ERK substrate phosphorylation, several compounds identified inhibited cell proliferation. In the current studies, we have identified a class of structurally similar compounds that inhibit cell proliferation and cause rapid induction of apoptosis in transformed cancer cell lines. Apoptosis appeared to occur through an intrinsic mechanism involving the prevention of ERK2-mediated inhibition of caspase-9 and p90Rsk-1 phosphorylation of the pro-apoptotic protein Bad. In addition, transformed cells appeared more sensitive to the growth inhibitory effects of the test compounds as compared to non-transformed epithelial cells. These findings support the identification of novel LMW compounds that target ERK1/2 proteins and restore apoptosis responses in cancer cells by preventing phosphorylation events that inactivate pro-apoptotic proteins.
Methods
Cells and reagents
HeLa S3 (human cervical carcinoma) and retinal pigment epithelial cells that stably express human telomerase (hTERT-RPE) were purchased from American Type Culture Collection (ATCC, Manassas, VA) and maintained in complete medium consisting of Dulbecco's modified Eagle medium (DMEM) plus 10% fetal bovine serum (FBS, Atlanta Biologicals, Lawrenceville, GA) and antibiotics (Penicillin, 100 U/ml; Streptomycin, 100 μg/ml) (Invitrogen, Carlsbad, CA). Hygromycin B, 0.01 mg/mL, (Roche, Indianapolis, IN) was used to maintain selection for the hTERT-RPE cells. Epidermal growth factor (EGF) and etoposide were purchased from Sigma (St. Louis, MO) and used at final concentrations of 25 ng/ml and 50 μM, respectively. The ERK (pT183/pY185) and α-tubulin antibodies were purchased from Sigma. Antibodies against phosphorylated Bad (pS112 or pS136), total Bad, phosphorylated Rsk-1 (pT573), phosphorylated Akt substrates (RXRXXpS/T), Mcl-1, phosphorylated caspase-9 (pT125), and total caspase-9 were purchased from Cell Signaling (Beverly, MA). Antibodies against poly ADP-ribose polymerase (PARP), total Rsk-1, and ERK5 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). U0126, SB203580, and Akt Inhibitor I were purchased from Calbiochem and used at final concentrations of 10 μM, 20 μM, and 25 μM, respectively. LY294002 was purchased from Cell Signaling and used at a final concentration of 25 μM. Test compounds were purchased from Chembridge (San Diego, CA) and stored as 25 mM stock solutions in DMSO. The general caspase inhibitor, Z-VAD-FMK, was purchased from BD Biosciences (San Jose, CA) and used at a final concentration of 20 μM.
Protein expression
Plasmids for mammalian and bacterial expression of caspase-9 (catalytically inactive C287A mutant) were purchased from Addgene (catalog #11819 and 11830). The His
6-tagged caspase-9 (C287A) was purified from BL21(DE3) cells as described [
30]. Briefly, BL21(DE3) cells were induced with 0.2 mM isopropyl β-D-1-thiogalactopyranoside for 4 hours and harvested with BugBuster protein extraction reagent (EMD Biosciences, San Diego, CA). Lysates were loaded onto a Talon Co
2+- IMAC affinity resin column (BD Biosciences, San Jose, CA) and eluted with imidazole. GST-p90Rsk-1 was expressed and purified as previously described [
31]. His
6-tagged ERK was expressed and purified as previously described [
17]. The plasmid for Bad was provided by Dr. Michael Greenberg (Harvard University). Transient expression of Bad and caspase 9 (C287A) in HeLa cells was done using Lipofectamine (Invitrogen).
Immunoblotting
Cells were washed with cold phosphate buffered saline (PBS, pH 7.2; Invitrogen) and protein lysates were collected with 2× SDS-PAGE sample buffer (4% SDS, 5.7 M β-mercaptoethanol, 0.2 M Tris pH 6.8, 20% glycerol, 5 mM EDTA) or cold tissue lysis buffer (TLB; 20 mM Tris-HCl pH 7.4, 137 mM NaCl, 2 mM EDTA, 1% Triton X-100, 0.1% SDS, 25 mM β-glycerophosphate, 2 mM sodium pyrophosphate, 10% glycerol, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine). Lysates collected in TLB were centrifuged at 20,000 (× g) to remove insoluble material and then diluted with an equal volume of 2× SDS-sample buffer. Proteins were separated by SDS-PAGE and analyzed by immunoblotting using enhanced chemiluminesence (ECL, GE Healthcare, United Kingdom). The relative protein levels were determined by densitometry scanning (Alpha Innotech), keeping the pixel intensity within the linear range of detection.
Fluorescence Quenching Assay
Fluorescence quenching of 1 μM ERK2 was evaluated by spectra analysis (SpectraMax5, Molecular Devices) utilizing an excitation wavelength of 295 nm and the emission spectra was monitored from 300 to 500 nm in the presence or absence of 25 μM of the indicated test compound. Relative changes in ERK2 fluorescence intensity in presence of test compounds were compared to ERK2 fluorescence in the presence of DMSO vehicle only.
Cell proliferation assays
Cell proliferation was evaluated by colony formation or by water soluble tetrazolium-1 (WST-1) assays. For colony formation, cells (~250 cells/mL) were plated and allowed to recover for 24 hours before treatment with test compounds. Cells were grown for 10-14 days and colonies that formed (approximately 40 cells or more) were fixed for 10 minutes in 4% paraformaldehyde and stained with 0.2% crystal violet in 20% methanol for 1-2 minutes. The number of colonies in the treated samples was expressed as a percentage of the controls, which consisted of 75, 146, or 154 colonies from 3 separate experiments, respectively. WST-1 assay was done according to manufacturer's instructions and cleavage of WST-1 to formazan by cellular mitochondrial dehydrogenases is used as an indicator of viable cells. Briefly, cells (~250/mL) were seeded in 96 well plates and allowed to recover overnight followed by test compound treatment for 7 days. WST-1 reagent was added and absorbance was read at 450 nm with background readings taken at 650 nm. After background subtraction, values were normalized to the control (DMSO only) cells.
Kinase assays
Active ERK2 (2 ng, New England BioLabs, Ipswich, MA) was incubated with 0.5 μg His6-tagged caspase 9 (C287A) or p90RSK-1 for 60 minutes at 30°C in 50 mM Tris-HCl, 10 mM MgCl2, 1 mM EGTA, 2 mM DTT, 0.01% Brij 35 (pH 7.5) containing 20 μM ATP and 2 μCi γ-32P-ATP. Reactions were stopped with an equal volume of 2× SDS-PAGE sample buffer and the proteins were resolved by SDS-PAGE. The gels were stained with coomassie blue, dried, and 32P incorporation into substrate was determined by phosphoimager analysis.
Caspase activity assays
HeLa cells were seeded at equal density and treated with indicated compound for 5 hours. Cells were harvested and frozen at -20°C. Fluorometric kits (Calbiochem) were used according to manufacturer's instructions to determine the activity of caspases-8 and 9 in cell lysates. Briefly, cell pellets were resuspended with 100 μL sample buffer and incubated on ice for 10 minutes. Samples were centrifuged at 10,000 × g for 10 minutes at 4°C; 50 μL of clarified lysate was transferred to a black walled 96 well plate and 50 μL of assay buffer was added. Fluorescent caspase substrate was added and a baseline reading was taken utilizing 400 nm/505 nm excitation/emission filters. The plates were incubated at 37°C for two hours and read again utilizing the same filters as baseline. Data shown reflects the mean difference ± SEM of relative fluorescence units (RFU) for three independent experiments.
Annexin-V/Propidium Iodide Staining
HeLa cells were plated at equal density and treated with 50 μM of indicated test compound. Cells were washed and stained according to manufacturers protocol (Calbiochem). Briefly, ~5×105 cells were incubated for 15 minutes with AnnexinV-FITC. Cells were centrifuged at 1,000 × g for 5 minutes and resuspended in 0.5 mL of cold 1× binding buffer. Propidium Iodide (PI) was added and samples were immediately analyzed by flow cytometry (FACScan Analyser; Becton Dickinson, Franklin Lakes, NJ)
Statistical Analysis
Comparisons between control or EGF and treated samples were performed utilizing an unpaired students t-test with equal variance using KaleidaGraph software (Synergy Software, Reading, PA). Comparisons within treated groups were performed utilizing a one-way analysis of variance (ANOVA) followed by a Tukey's post hoc analysis with an α-value of 0.05. Statistical significance was indicated with an asterisk (*) if the p value was equal to or less than 0.05 or a number sign (#) if the p value was equal to or less than 0.01.
Discussion
The goal of these studies was to characterize a class of novel ERK inhibitor compounds being developed in our laboratory that inhibit cell proliferation. Our findings demonstrate the induction of an apoptotic response by structurally-related compounds that inhibit ERK regulation of pro-apoptotic proteins. Aberrant activation of ERK signaling has been well-documented and provides a survival advantage in a number of cancers by inactivating pro-apoptotic proteins [
40]. Thus, selective inhibition of ERK anti-apoptotic functions is a potential approach to sensitize cancer cells to chemotherapeutic agents. One important feature of the active test compounds is that some appear to be more selective for decreasing cell proliferation and inducing apoptosis in transformed cells as compared to non-transformed cells (Table
2 and Figure
3). This finding is in agreement with our previous studies indicating that the test compounds did not affect the life span or cause general toxicity in a
C. elegans whole organism model [
32]. Interestingly, we also have not observed overt toxicity when treating mice harboring B-cell lymphomas with compound
76 [
41]. Coincidentally, these
in vivo studies also demonstrated that B-cell lymphoma tumor burden was significantly reduced by
76. However, we predict that the significance of the ERK-targeted inhibitors in reducing tumor burden will best be realized using potentially more potent compounds, such as
76.3, in combination with other chemotherapeutic agents.
Our data support a mechanism by which the ERK-targeted compounds prevent inactivation of caspase-9 and p90Rsk-1 phosphorylation of Bad (Figures
4 and
6). Both caspase-9 and p90Rsk-1 contain D-domains that are thought to form contacts with the CD domain of ERK proteins [
1,
24]; the region predicted to be targeted by the test compounds. Structural studies underway are determining the site of interactions between the compounds and ERK proteins to better understand the mechanisms of inhibition and identify chemical features that can be modified to improve inhibitor binding and efficacy. A region analogous to the CD domain of ERK can be found on other MAP kinases family members including p38 and JNK [
29,
42]. The subtle differences between these regions are determinants of substrate selectivity between the MAP kinases. In addition, differences in the residues associated with the D-domains of MAP kinase interacting proteins can confer selectivity. For example, the differences in D-domains on MAP kinase kinase (MKK) proteins appear to be important for determining what MAP kinase the MKK isoform will target [
43].
The p90Rsk-1 protein is a major regulator of ERK-mediated cell survival [
5] and, like Bad, other p90Rsk-1 substrates may be affected by the test compounds. For example, inhibition of p90Rsk-1 could also restore activity of the death associated protein kinase (DAPK), which is a pro-apoptotic tumor suppressor protein that is phosphorylated and inactivated by p90Rsk-1 [
44]. Moreover, it is recognized that other survival proteins may be affected by the test compounds and involved in mediating the apoptotic response observed. While some inhibition of PI3K/Akt signaling may have occurred with the test compounds (data not shown), these effects were not to the extent of inhibition of Akt substrates with the PI3K inhibitor, LY294002. In addition, the activation of other growth related proteins like ERK5 and EGF receptor (data not shown) were not affected, which further supports the mechanism of apoptosis induction by the test compounds involves inhibition of ERK1/2 signaling.
Another objective of this study was to determine whether improvements in potency and selectivity of the parent compound (
76) could be achieved by examining structurally similar compounds. Cell based and
in vitro kinase assays in figures
4 and
6 suggest that compound
76.3 has higher potency than
76. Thus, these studies have identified a class of structurally similar compounds that have similar biological effects and chemical features that can be modified to improve potency. Studies are now aimed at performing a more detailed structure activity relationship to identify the mode of the compound's interaction with ERK, improve potency, and determine whether other substrate proteins are affected. The present results also indicate the potential utility of using LMW compounds that inhibit protein-protein interactions in a therapeutic setting. While it has traditionally been assumed that small compounds would not be effective inhibitors of protein-protein interactions [
45,
46], recent studies involving the transcription factor BCL6 indicate that protein-protein interactions can be targeted
in vivo with LMW compounds [
47]. The efficacy of the present compounds as inhibitors of ERK-substrate interactions and their activity in a xenograft mouse model [
41] further indicate that protein-protein interactions are valid therapeutic targets of LMW compounds.
The specific targeting of the ERK1/2 pathway to treat cancer has been best studied using MEK1/2 inhibitors (reviewed in [
48]). Although a number of preclinical studies showed very good efficacy of MEK inhibitors for reducing cancer cell proliferation and inducing apoptosis, patient responses in clinical trials have been minimal with these compounds as single agents. Thus, MEK inhibitors in combination with other chemotherapeutic drugs will likely be more effective. A potential problem with targeting MEK for cancer therapy is the development of drug resistance, possibly through compensatory mechanisms involving elevated expression and activity of MEK/ERK proteins [
49]. Thus, the approach to completely inhibit MEK/ERK signaling may be counterproductive in the long term. In contrast, the use of compounds that selectively inhibit ERK functions involved in cell survival may have the potential to sensitize cancer cells to other chemotherapeutic agents without inducing the compensatory pathways that lead to acquired drug resistance.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
SB participated in study design, performed the experiments and statistical analysis as well as writing the manuscript. RD performed the colony formation assays. UP and AM designed the ERK targeted docking domain inhibitors. SS provided valuable insight and critically reviewed the manuscript. AM and PS participated in the conception and study design. PS coordinated the study and the writing of the manuscript. All authors approved the final manuscript.