In the current study, we explored for the first time, the mechanism of action of the new Casein kinase 1 $\varepsilon$ (CK1$\varepsilon$) selective inhibitor GSD0054. Although GSD0054 behaved as a selective CK1$\varepsilon $ inhibitor in enzymatic assays, we studied whether this inhibitory activity also occurred inside the cells. The effects of GSD0054 on $\beta $-catenin expression and disruption of cell cycle progression were studied in the human breast cancer cell lines MDA-MB-453 ($\beta $-catenin negative) and T-47D ($\beta $-catenin positive). We also performed molecular modeling studies using computational docking against CK1$\varepsilon $ to explain and predict the mechanism of action of this compound. Moreover, the commercially available CK1$\varepsilon $ inhibitor PF-4800567 and the CK1$\delta $/$\varepsilon $ inhibitors PF-670462 and IC261 were also studied for comparison purposes.GSD0054 showed anti-proliferative activity against MDA-MB-453 and T-47D cells despite the fact that MDA-MB-453 cells do not possess active $\beta$-catenin. However, selective cell killing occurred in the more resistant, $\beta $-catenin active, T-47D cells. CK1$\varepsilon $ was confirmed as a cellular target, although other targets or alternative mechanisms of action could possibly explain the anti-proliferative activity in MDA-MB-453 cells.
Casein kinase 1 $\varepsilon $ (CK1$\varepsilon$) belongs to a highly conserved and ubiquitously expressed family of serine/threonine protein kinases. At least, seven CK1 isoforms ($\alpha $, $\beta $, $\gamma $ 1-3, $\delta $ and $\varepsilon$) have been identified in vertebrates [1]. With the exception of CK1$\beta $ which has been found solely in cows, all the other CK1 isoforms are expressed in humans. Besides other physiological implications, CK1 isoforms are important in cancer progression in diverse types of tumors [2]. In particular, changes in the activity or expression levels of CK1$\varepsilon $ can contribute to the development of cancer [3]. Moreover, other CK1 isoforms did not induce growth arrest, demonstrating the role of CK1$\varepsilon $ in promoting tumor cell proliferation [4].
CK1$\varepsilon $ is the protein encoded by the CSNK1E gene and behaves as an essential modulator of the Wnt/$\beta $-catenin signaling pathway, which is implicated in embryonic development, stem cell biology and diseases like cancer. This pathway is regulated at different levels by a broad range of effectors including CK1 members CK1$\alpha $, $\delta $ and $\varepsilon $ [5] and its deregulation has been implicated in a variety of human cancers. CK1$\varepsilon $ is required for the survival of breast cancer subtypes that rely on aberrant $\beta $-catenin activity. Active myristoylated CK1$\varepsilon $ is able to induce malignant transformation by stabilizing $\beta $-catenin and activating the transcription of Wnt target genes [6], which include several proto-oncogenes like colon cancer related c-myc, n-myc, ovarian adenocarcinoma related FGF9 and breast cancer-related LBH. The list of target genes gives an idea of the relevance of this pathway and the potential of its regulation. Therefore, targeting key mediators of the Wnt signaling pathway such as CK1$\varepsilon $ may inhibit tumor progression [7].
As part of our interest in the discovery of new anti-proliferative agents, we have reported a new CK1$\varepsilon $ selective inhibitor (ICK1$\varepsilon$) [8]. Thus, a small and structure-biased library of enantiopure anti-$\beta $-amino alcohols was prepared in a straightforward manner by a simplified version of the Reetz protocol [9]. For most active compounds, the antiproliferative activity against a panel of human solid tumor cell lines yielded GI$_{50}$ values in the range 1-20 $\mu $M. We designed a computational approach to predict the molecular target(s) based on the chemical structure and the antiproliferative activity. This method pointed to kinases as plausible candidates. The experimental determination of the interaction with 456 kinases indicated that these anti-$\beta $-amino alcohols behave as selective ICK1$\varepsilon $. The anti-$\beta $-amino alcohol GSD0054 (1) (Fig.1) was selected as the lead drug.
Fig.1.Chemical structures of ICK1$\varepsilon $ GSD0054 (1) and PF-4800567 (2); and ICK1$\delta $/$\varepsilon $ PF-670462 (3) and IC 261 (4).
In the current work, we describe the first findings on the mechanism of action of GSD0054 (1). Although we have unveiled its role as a growth inhibitor in cell cultures and CK1$\varepsilon $ selective inhibition in kinetic assays, we attempted to find out if these activities correlate. Herein, we report the preliminary studies on the mechanism of action of GSD0054 (1) and compare the results with those of the ICK1$\varepsilon $ PF-4800567 (2); and the CK1$\delta/\varepsilon $ dual inhibitors (ICK1$\delta/\varepsilon$) PF-670492 (3) and IC 261 (4) (Fig.1).
GSD0054 (1) was prepared as described earlier [8]. PF-4800567 (2), PF-670462 (3) and IC 261 (4) were purchased from Sigma-Aldrich.
The breast cancer cell line MDA-MB-453 was provided by CEAMED S. A. (Tenerife, Canary Islands). The breast cancer cell line T-47D, the non-small human lung cancer cell line SW1573 and its P-gp overexpressing variant (SW1573/P-gp) were kindly provided by Prof. G. J. Peters (Cancer Center Amsterdam, Vrije Universiteit, Amsterdam, The Netherlands). Cells were grown in RPMI-1640 medium containing 5% heat- inactivated FCS, 2 mM L-glutamine, 100 U/mL of penicillin G and 0.1 mg/mL of streptomycin in a 37$^\circ$C, 95% humidified atmosphere of 5% CO$_{2}$. Cells were maintained in culture in 100 mM cell culture dishes in growth medium (10 mL) and passaged twice weekly.
Cell line suspensions were counted with Moxi Z and diluted to reach the appropriate cell densities (SW1573 2,500 cells/well; T-47D, MDA-MB-453 and SW1573/P-gp 5,000 cells/well) for inoculation onto 96-well plates. After 24 hours, pure compounds were added. Each agent was initially dissolved in DMSO at 400 times the desired final maximum test concentration (100 $\mu $M). Control cell samples were exposed to an equivalent concentration of DMSO (0.25% v/v, negative control). Compounds 1-4 were tested at serial dilutions (0.01-100 $\mu $M). Paclitaxel was used as a positive control and was initially dissolved in DMSO at a stock solution of 4 mM. Verapamil was used as a P-gp transport inhibitor and the stock solution was prepared in DMSO at 40 mM. Cell culture medium containing verapamil was prepared by adding the final concentration of 10 $\mu $M verapamil.
Drug incubation times were 48 hours, after which cells were precipitated with 25 $\mu $L ice-cold trichloroacetic acid solution (50% w/v) and fixed for 60 min at 4$^\circ$C. Then, plates were rinsed with water, following which 25 $\mu $L of a sulforhodamine B (SRB) solution (0.4% w/v in 1% acetic acid) was added for 15 min. Unbound SRB was rinsed with 1% acetic acid. Bound SRB was dissolved with 150 $\mu $L of a 10 mM Tris solution (pH 10.5). The optical density of each well was determined at 530 and 620 nm using PowerWave XS microplate reader. Dual wavelength was used to reduce optical interference caused by scratches, fingerprints or other matter that absorb light equally at both wavelengths. The anti-proliferative activity, expressed as GI$_{50}$ (50% growth inhibition), was calculated according to NCI formulas [10].
Western blot analysis was used to determine the expression of $\beta $-catenin in T-47D cells. MDA-MB-453 cells were used as a negative control as these cells barely express detectable levels of nuclear (transcriptionally active) $\beta $-catenin [6]. Cell line suspensions of MDA-MB-453 and T-47D were counted with Moxi Z and diluted to reach the appropriate cell densities for the inoculation of 50,000 cells/well onto 6-well plates. After 24 hours, compounds 1-4 were added only to wells containing T-47D cells. Drug incubation times were 24 hours, after which time the growth medium was aspirated and cells were washed with PBS and treated with 1 mL of lysis buffer (25 mM Tris pH 8.0; 125 mM NaCl; 1% SDS; and 6.3 M urea). Lysates were collected in 1.5 mL microtubes and frozen until use. Cells floating in the growth medium were harvested in a different manner to prevent their loss; growth medium and PBS were collected together and then centrifuged at 1500 rpm for 2 minutes. The supernatant was removed before adding the lysis buffer to the cell pellet.
The amount of protein in cell lysates was quantified to ensure equal loading (20 $\mu $g of protein) onto the polyacrylamide gels. Samples were then mixed with 5 $\mu $L of loading buffer (4% SDS, 0.05% bromophenol blue, 20% glycerol and 2% $\beta $-mercaptoethanol in 100 mM Tris-HCl, pH 6.8), heated at 95$^\circ$C for 5 min and loaded onto 8% acrylamide gels. Electrophoresis was performed at a constant voltage of 160 V (120 V first 15 minutes) for one hour. A molecular weight marker was also loaded in order to facilitate identification of molecular weights. Electroblotting of the proteins from the gel to the membrane took place at 160 V-300 mA for one hour. The membrane was then incubated with agitation in 4 mL milk over a period of 45 minutes to block membrane pores and then 1 $\mu $L of primary (mouse) antibodies (anti-$\beta $-catenin and anti-$\beta$-actin) were added and left over night. Secondary antibodies conjugated to horseradish peroxidase were used to analyze protein bands and quantified using ImageQuant.
Changes in cell cycle were studied using the DNA dye propidium iodide (PI) and flow cytometry. Cells were seeded onto six well plates at a density of 50,000 cells/well. On the next day, compounds 1-4 were added to the respective wells and incubated for 24 hours. After treatment, cells were trypsinized, harvested, transferred to tubes and centrifuged at 300 for 10 min. Cell pellets were resuspended in 200 $\mu $L of cold PBS and fixed in 1 mL ice-cold 70% ethanol overnight at $-$20$^\circ$C. Fixed cells were centrifuged and resuspended in 500 $\mu $L of PBS containing 0.1 mg/mL of DNAse-free RNAse. The mixture was incubated in the dark at 37$^\circ$C for 30 min. After incubation, 5 $\mu $L of PI (0.5 mg/mL) were added at room temperature. Typically, 4500 events were collected using excitation/emission wavelengths of 488/670 nm for PI. Flow cytometric analysis of DNA content was undertaken using Accuri C6. Cell aggregates were gated out and viable cells were selected based on size, complexity and DNA content such that the fractions of the cells in G$_{0}$/G$_{1}$, S, and G$_{2}$/M phase were calculated using cell cycle analysis software.
The level of ROS was determined using the ROS-GloTM H$_{2}$O$_{2}$ Assay (Promega Corporation, USA). Cells were exposed to compounds at the indicated doses for 48 h, after which time the non-lytic assay was performed following the manufacturer's instructions. Luminescence was measured on a Synergy HTX multi-mode microplate reader (BioTek, USA).
CK1$\varepsilon $ structure (PDB ID: 4HNI) was downloaded from Protein Data Bank. We used HyperChem software to draw and optimize the molecular structures and OpenBabel application to turn the files into adequate format before exporting them to docking software Glide. We also performed docking calculations of compound IC261 against the crystal structure of $\beta$-tubulin (PDB ID: 3E22). For the purpose of the docking calculations, we removed ligands and non-essential water molecules present in the crystal structure and added polar hydrogen atoms to the protein chain before running any docking calculations.
In order to determine the antiproliferative activity (expressed as GI$_{50}$) of the inhibitors 1-4, we exposed MDA-MB-453 and T-47D cells to these compounds for 48 h. All inhibitors suppressed cell proliferation in both tumor cell lines (Table 1). Compounds 1-3 exhibited GI$_{50}$ values in the range 2.4-79 $\mu $M, whilst inhibitor 4 was much more active with GI$_{50}$ values in the range 32-87 nM. While inhibitor 4 constitutes the most potent compound of the series in all the tested tumor cell lines, compound 2 appears to be the less active in this context. Compound 1 displayed more potent inhibition than the selective inhibitor 2. The selective ICK1$\varepsilon $ 1-2 showed stronger antiproliferative effect on MDA-MB-453 cells. In contrast, the ICK1$\delta/\varepsilon $ 3-4 were more active against T-47D cells. The ratios between GI$_{50}$ values obtained for each compound are shown in Table 1.
| MDA-MB-453 | T-47D | Rb | |
|---|---|---|---|
| GSD0054 (1) | 2.4 士 0.5 | 17 士 1.7 | 7.1 |
| PF-4800567 (2) | 58 士 11 | 79 士 6.4 | 1.4 |
| PF-670462 (3) | 42 士 6.5 | 11 士 2.0 | 3.8 |
| IC 261 (4) | 0.087 士 0.040 | 0.032 士 0.007 | 2.7 |
$^{\rm a}$Values are given in $\mu $M and represent mean values of at least three independent experiments $\pm $ standard deviation.
$^{\rm b}$R is defined as the quotient of high and low GI$_{50}$ for each compound against both cell lines.
To determine whether or not P-gp could affect the activity of GSD0054, we determined the GI$_{50}$ values after 48 h of exposure to ICK1$\varepsilon $ 1-4 in wild type and P-gp- overexpressing SW1573 cells, and in the presence of the P-gp transport inhibitor verapamil (at a fixed concentration of 10 $\mu $M). The standard microtubule-interacting drug paclitaxel was used for comparison purposes. Table 2 shows the GI$_{50}$ values obtained after 48 h of drug exposure. For better comparison of the data, we defined resistance factor (Rf) for a given compound as the ratio of GI$_{50}$ values of the P-gp overexpressing and the wild-type cell line. In the absence of verapamil, we observed that ICK1$\varepsilon $ 1-4 were not affected by P-gp overexpression (Rf $=$ 0.6-3.2). These results indicate that the activity of ICK1$\varepsilon $ 1-4 is not affected by the overexpression of P-gp. In addition, the dose response curves (Fig. fig2) show that only GSD0054 (1) is able to induce net cell killing, which occurs at doses higher than 30 $\mu $M.
| - Verapamil | 十 Verapamil | |||||
|---|---|---|---|---|---|---|
| SW1573 | SW1573/P-gp | Rfb | SW1573 | SW1573/P-gp | Rf | |
| GSD0054(1) | 16 士 2 | 16 士 3 | 1 | 20 士 1 | 17 士 4 | 0.9 |
| PF-4800567(2) | 76 士 26 | 48 士 24 | 0.6 | 85 士 14 | 51 士 25 | 0.6 |
| PF-670462 (3) | 19 士 3 | 19 士 8 | 1 | 21 士 7 | 19 士 6 | 0.9 |
| IC 261 (4) | 0.25 士 0.03 | 0.79 士 0.10 | 3.2 | 0.35 士 0.03 | 0.86 士 0.42 | 2.5 |
| Paclitaxel | 0.0029 士 0.0009 | 4 士 0.5 | 1379 | 0.0030 士 0.0015 | 0.0036 士 0.0015 | 1.2 |
$^{\rm a}$Values are given in $\mu $M and represent mean values of at least three independent experiments $\pm $ standard deviation.
$^{\rm b}$Rf represents the GI$_{50}$ ratio between SW1573/P-gp and SW1573.
As an indirect way to determine CK1$\varepsilon $ inhibition, we evaluated the capacity of compounds 1-4 to reduce $\beta $-catenin expression in T-47D cells. MDA-MB-453 cells were used as a negative control since they barely express detectable nuclear $\beta $-catenin levels. Thus, T-47D cells were treated with compounds 1-4 at three drug concentrations, which were selected from the corresponding GI$_{50}$ values (Table 1). The results are shown in Fig. 2. After 24 h of incubation, the drug GSD0054 (1) was able to reduce significantly the expression of $\beta $-catenin at concentrations of 20 $\mu $M ($\approx $ GI$_{50}$) or higher. Compounds PF-4800567 (2) and PF-670462 (3) did not exert any inhibitory effect on $\beta $-catenin levels. However, treatment of T-47D cells with IC 261 (4) induced a reduction in the expression of $\beta $-catenin after 24 h of exposure at 100 nm ($\approx $ 3$\times$ GI$_{50}$).
Fig.2.Dose-response curves of SW1573 and SW1573/P-gp cells treated with ICK1$\varepsilon $ 1-4 for 48 h.
We used flow cytometry to study the effect of inhibitors 1-4 on the cell cycle. Cells were exposed for 24 h to two drug doses (low and high), chosen based on the GI$_{50}$ values (Table 1). The effect on the cell cycle was dose-dependent and varied for each compound (Fig. 3). Thus, GSD0054 (1) induced a notable accumulation of MDA-MB-453 cells at the G$_{1}$ phase and a slight increase in the sub-G$_{1}$ population was noted that reached 6.3% at the high dose. The effect of GSD0054 (1) on T-47D cells was different from that on $\beta $-catenin-lacking cells. There was a notable increase in the sub-G$_{1}$ population that peaked at about 41% at the high drug dose. In contrast, PF-4800567 (2) showed a quite similar effect on both tumor cell lines, leading to a marked accumulation of cells in the G$_{1}$ phase, concomitant with a decrease in the S and G$_{2}$/M phases. Moving on to PF-670462 (3), we observed a rise in MDA-MB-453 population at the G$_{1}$ phase when exposed to the compound, while the effect on T-47D cells remained unclear due to a possible aneuploid population. Importantly, compound IC 261 (4) exerted a very different effect on the cell cycle leading to an increase in the S and G$_{2}$/M populations in both cell lines.
Fig.3.Expression of $\beta $-catenin in drug-treated T-47D cells. Cells were exposed to compounds 1-4 for 24 h at the indicated doses. C (-): MDA-MB-453, negative control; C ($+$): T-47D, positive control. $\beta $-Actin signals allowed us to normalize the experiment.
In order to gain an in-depth information about redox processes in the antiproliferative activity of ICK1$\varepsilon $ 1-4, the ROS-GloTM H$_{2}$O$_{2}$ assay was used. Importantly, the ROS-Glo$^{\rm TM}$ assay does not use horseradish peroxidase (HRP), which is known to cause a high number of false hits, providing a simple format for both cell-based and enzymatic assays. Since various ROS are converted to H$_{2}$O$_{2}$ in the cell, and H$_{2}$O$_{2}$ is the longest-lived ROS, an increase in H$_{2}$O$_{2}$ can reflect a general increase in the ROS level. In particular, this experiment was applied to breast cancer cell lines MDA-MB-453 and T-47D, in the presence of model compounds at a low and a high dose (based on their GI$_{50}$ values), and an incubation time of 48 h. The results shown in Fig. 5 evidence a dose-dependent increase of ROS levels in the $\beta $-catenin positive tumor cell line (T-47D) for all compounds except for ICK1$\varepsilon $ 2. However, in MDA-MB-453 cells, no relevant increase in ROS production was observed.
Fig.4.Cell cycle histograms of MDA-MB-453 and T-47D untreated cells (control) and cells treated (drug concentration in $\mu $M) for 24 h with compounds 1-4 at the indicated doses.
Fig.5.ROS production increase in MDA-MB-453 and T-47D cells treated with drug (dose in $\mu $M) for 48 h with compounds 1-4 at the indicated concentrations. LD (low dose) of 1, 10, 10, 0.1 and HD (high dose) of 10, 100, 100, 1 $\mu $M for 1-4, respectively.
We performed docking calculations against the binding site of CK1$\varepsilon $. The results for compounds 1-4 expressed as docking score (representing the variation in free energy of binding) are given in Table 3. Glide software provided energy scores of each compound; hence this value was a modality to predict how favorable the interaction between the protein and the ligand is. Thus, lower docking scores (more negative energy) indicate better interaction. As seen in Table 3, ICK1$\varepsilon $ 2 displays the most favorable interaction with CK1$\varepsilon $, with a docking score of $-$10.8 kcal/mol. However, the remaining inhibitors showed docking scores in the range $-$5.8 to $-$6.5 kcal/mol. We also docked compound IC 261 (4) against $\beta $-tubulin either at the colchicine or the GTP site. The docking results indicate that IC 261 (4) has more favorable interactions with $\beta $-tubulin at the colchicine binding site ($-$6.2 kcal/mol) than at the GTP site ($-$4.5 kcal/mol). These results are in agreement with the reported experimental findings [7].
| Compound | Score (kcal/mol) |
|---|---|
| GSD0054(1) | -5.8 |
| PF-4800567 (2) | -10.8 |
| PF-670462 (3) | -6.5 |
| IC 261 (4) | -5.9 |
In our group, we seek rational approaches based on phenotypic changes induced in cells by bioactive compounds that may allow the identification of possible biological targets for any given small molecule. With this strategy, we have discovered the novel CK1$\varepsilon $ selective inhibitor GSD0054 (1) [8]. CK1$\varepsilon $ is a modulator of the Wnt/$\beta $-catenin signaling pathway and represents a potential target for developing anticancer drugs with high therapeutic index [7]. To the best of our knowledge, selective CK1$\varepsilon $ inhibitors are scarce. One of the main drawbacks in the search for ICK1$\varepsilon $ is that CK1$\delta $ holds $>$ 98% sequence identity with CK1$\varepsilon $. Thus, inhibitors in public databases in general are reported as dual ICK1$\delta/\varepsilon $ [11]. Considering the differences in biological involvement of both isoforms, the reports in the literature seem confusing and produce apparently contradictory results. In the current study, we performed some phenotypic studies on our ICK1$\varepsilon $ GSD0054 (1) and compared the results with the established compounds ICK1$\varepsilon $ PF-4800567 (2), and the ICK1$\delta/\varepsilon $ PF-670462 (3) IC 261 (4).
Initially, we found that CK1$\varepsilon $ inhibitors 1-4 exerted anti-prolifer- ative effects on MDA-MB-453 and T-47D cell lines (Table 1). Thus, selective CK1$\varepsilon $ inhibitors (1-2) were more active against MDA-MB-453 ($\beta $-catenin negative) cells, whilst T-47D ($\beta $-catenin positive) cells were more sensitive to the dual CK1$\delta/\varepsilon $ inhibitors (3-4). The discouraging outcome of this result is that all compounds were able to induce growth inhibition independently of the Wnt/$\beta $-catenin signaling pathway. In this particular context, it was found that compound 2 did not exert anti-proliferative effects on human cell lines HEK293 and HT1080 after 7 days of exposure at 1 $\mu $M [7]. Our results demonstrate that larger doses of 2 are necessary to observe growth inhibition. At this point, we wondered if the differences in potency observed for compounds 1-4 could be related to other factors. We next explored the ability of P-gp to reduce the antiproliferative activity of compounds 1-4. This transmembrane efflux transporter was initially characterized as overexpressed in tumor cells exhibiting multiple drug resistance to a wide spectrum of structurally-unrelated anticancer drugs [12]. None of the tested compounds was affected by the overexpression of P-gp (Table 3).
When the expression of $\beta $-catenin in T-47D cells was studied, only compounds 1 and 4 induced a downregulation of this protein after 24 hours of treatment, in a dose-dependent manner (Fig. 3). Our findings do not agree with Cheong et al. [7], who, based on the data obtained from the comparison between compounds 2 and 4 proposed that the combined inhibition of CK1$\delta/\varepsilon $ was more effective in promoting inhibition of Wnt/$\beta $-catenin signaling and cell proliferation. Our data suggest that either compounds 2 and 3 do not inhibit CK1$\varepsilon$ and subsequently $\beta $-catenin activity in these cell lines, or that CK1$\varepsilon$ inhibition might have other consequences distinct of Wnt/$\beta $-catenin inhibition that are also able to inhibit growth. Similar results were found by Rodriguez et al. [13] when they observed that suppression of CK1$\varepsilon$ expression using shRNA constructs led to an antiproliferative effect in ovarian cancer cells, which was independent of Wnt/$\beta $-catenin activity. Therefore, it is possible that inhibition of CK1$\varepsilon$ interferes with cell growth without affecting $\beta $-catenin expression, partially explaining the profiles observed for compounds 2 and 3.
To further understand the similarities and differences in the mechanism of action of the four inhibitors, we proceeded with cell cycle analysis by flow cytometry. Interestingly, we observed that only compound 1 was able to produce cell death in T-47D cells as deduced from the increase in the sub-G$_{1}$ population (Fig. 4). This effect was not observed for MDA-MB-453 cells, marking a difference in the response between these cell lines after exposure to compound 1. However, the effect of inhibitors 2-4 only induced an accumulation of cells at different cell cycle phases. Our data partially support the findings of Cheong et al. [7], who described that compounds 2 and 3 were unable to induce cancer cell death. In addition, compound 4 was found to bind to $\beta $-tubulin with an affinity similar to colchicine that is a potent inhibitor of microtubule polymerization [7]. This is consistent with our cell cycle results, where inhibitor 4 induced accumulation of cells at G$_{2}$/M phase. Similarly, treatment of primary mouse embryo fibroblasts with compound 4 caused an accumulation of cells in G$_{2}$/M phases of the cell cycle, but it was not able to produce cancer cell killing [14].
As aforesaid, compound 1 was able to produce selective cell death in T47D cells which express $\beta $-catenin, where the sub-G$_{1}$ population accounted for 41% after exposure to a 60 $\mu $M concentration ($\approx $ 3$\times$ GI$_{50}$). In contrast, the sub-G$_{1}$ population was only 6% in cells lacking active $\beta $-catenin (MDA-MB-453) after exposure to 15 $\mu $M of compound 1 ($\approx $ 5$\times$ GI$_{50}$). This result is consistent with previous findings that selective cell killing may be occurring when Wnt signaling is known to be aberrantly activated [15]. When considering ROS production (Fig. 5), MDA-MB-453 cells were not affected by compound treatment. This is consistent with the cell cycle experiments. However, the effect on T-47D cells cannot be correlated with the cell cycle experiments, except for our ICK1$\varepsilon $ GSD0054 (1).
The docking results showed discrepancies with the antiproliferative data. Thus, the best docking score does not assure an actual superior cell growth inhibitory activity. In this context, although compound 2 displayed the most favourable interaction with CK1$\varepsilon $, it turns out to have the least potent antiproliferative activity, a null effect on $\beta $-catenin expression, no cell killing and lack of ROS production. One could argue that the docking results for this compound are biased towards more favourable outcome. This compound was co-crystallized in silico with the protein structure that we used in our docking calculations and therefore the protein is in a perfect conformation to accommodate it, producing better docking scores. However, our lead compound 1 having a docking score of $-$5.8 kcal/mol was able to: (a) inhibit cell proliferation, (b) inhibit $\beta $-catenin expression in T-47D cells, (c) induce ROS production and also (d) induce cell killing as observed with the accumulation of cells in the sub-G$_{1} $ phase.
In summary, we have performed a preliminary study of the mechanism of action of GDS0054 (1) in MDA-MB-453 and T-47D cells. We observed that our lead compound exhibits antiproliferative activity against MDA-MB-453 and T-47D cells despite the fact that MDA-MB-453 cells do not have active $\beta $-catenin. Cell cycle studies indicated that selective cell killing occurred in the more resistant, but $\beta $-catenin active, tumor cell line T-47D. Furthermore, GSD0054 (1) was not a transport substrate of P-gp. The comparative study with other three known CK1$\varepsilon $ inhibitors revealed notable differences among the mechanism of action of all compounds.
Taken collectively, the results indicate that the mechanism of action might differ depending on the tumor cell line; it is possible that it affects different intracellular pathways or that the anti-proliferative effect is independent of $\beta $-catenin function. Further mechanistic studies are required to understand the exact role of GSD0054 (1) and its potential use as a therapeutic agent against cancer.
I.L. thanks CONACYT (Mexico) for a postdoctoral grant. G.S.-D thanks the EU Social Found (FSE) and the Canary Islands ACIISI for a pre-doctoral grant.
The authors declare no conflict of interest.