ML264, A Novel Small-Molecule Compound That Potently Inhibits Growth of Colorectal Cancer

This article is featured in Highlights of This Issue, p. 1

Colorectal cancer is the third most prevalent cancer in the United States with approximately 140,000 new cases and 50,000 deaths each year (1). The progression from normal intestinal epithelial tissue to metastatic neoplasm results from the disruption of multiple regulatory mechanisms involving critical signaling pathways that normally regulate proliferation, differentiation, migration, and apoptosis (2, 3). Previous research has coalesced around a model of colorectal cancer development whereby dysregulation of WNT and RAS/MAPK/PI3K signaling pathways is necessary for early development and progression (4). That the sequential impairment of each of these pathways is a requirement for developing intestinal neoplasia has been recently confirmed in an organoid system by Drost and colleagues (5). Despite this knowledge, few therapeutic strategies have been developed to specifically target components of these particular pathways and none have gained FDA approval. Our group has identified Krüppel-like factor 5 (KLF5), a WNT- and RAS/MAPK/PI3K-responsive molecule, as an important regulator of intestinal epithelial cell proliferation that is frequently overexpressed during intestinal tumorigenesis.

KLF5 is a zinc-finger transcription factor highly expressed in the intestinal epithelium crypts (6). We previously demonstrated that transit-amplifying (TA) cells of the intestinal epithelium express high levels of KLF5 that co-localize with the proliferation marker, Ki-67 (6, 7). More recently, we demonstrated that KLF5 is also expressed in the crypt base columnar (CBC) cells of the intestinal crypts (8). CBC cells, which express the leucine-rich-repeat-containing G-protein-coupled receptor 5 (LGR5) gene, proved to be the dividing and long-lived intestinal stem cells (ISC) that give rise to all lineages of the intestinal epithelium (9–12). Importantly, we recently showed that KLF5 is essential for the proliferation and survival of LGR5-expressing ISCs (8). Mechanistically, we and others have shown that KLF5 expression is regulated by both the RAS/MAPK/PI3K and WNT signaling pathways, but via distinct mechanisms. Although the RAS/MAPK/PI3K pathway exerts its action directly through early growth factor 1 (EGR1), a potent transcriptional activator of KLF5 (13, 14), the WNT pathway regulates the rate of the KLF5 protein degradation through GSK3β phosphorylation and a subsequent FBW7α-dependent degradation (15–17). Additionally, we have demonstrated that KLF5 is itself a positive regulator of both the RAS/MAPK and WNT signaling pathways. The RAS/MAPK pathway is affected by KLF5-mediated EGFR activation, whereas WNT signaling is stimulated by the ability of KLF5 to enhance β-catenin stability, nuclear localization, and transcriptional activity (18, 19). Taken together, these observations illustrate the important role that KLF5 plays in the maintenance of homeostasis of intestinal epithelium by governing the activity of stem and TA cells.

Furthermore, we have shown that elevated levels of oncogenic KRAS, which is associated with hyperactive MAPK and PI3K signaling, result in the upregulation of KLF5 expression in both human primary colorectal cancers and transgenic mouse models (14, 20). KLF5 overexpression in turn potentiates the transforming activity of KRAS by increasing the proliferation rate of cells (20). Our group has also demonstrated that reduced Klf5 expression leads to reduced intestinal tumor formation in mice harboring a germline mutation in the tumor suppressor Apc, a crucial component of the WNT pathway, alone or in combination with activating KRAS mutations (18, 20, 21). In addition, it has been recently demonstrated that KLF5 expressed in CBCs facilitates the oncogenic activity of mutated β-catenin promoting development of intestinal adenomas, while Klf5 deletion abrogates this process (22). Moreover, we have evidence that KLF5 expression levels are highest in cancer cells of colorectal cancer origin among the NCI60 panel of cancer cells (23). These lines of evidence suggest that small molecule compounds that decrease KLF5 expression could prove to be an effective therapeutic option for colorectal cancer.

We generated colorectal cancer cell lines stably expressing the luciferase reporter from the human KLF5 promoter and utilized these cells in an ultrahigh-throughput screening (uHTS) approach to identify compounds that modulated KLF5 expression (23, 24). Previously, we demonstrated that this screening method allows for specific identification of compounds that decrease KLF5 expression levels and that inhibit proliferation of colorectal cancer cell lines in in vitro systems (23, 24). Here, we show that ML264, a third-generation small molecule compound that arose from the first-generation of uHTS hits, potently inhibits KLF5 expression, decreases proliferation of colorectal cancer cell lines, and inhibits the growth of xenografts in a mouse model of primary tumor development.

DLD-1 and HCT116 colorectal cancer cell lines were purchased from the ATCC. DLD-1 cells were maintained in RPMI1640 medium supplemented with 10% FBS and 1% penicillin/streptomycin, and HCT116 cells were maintained in McCoy's medium supplemented with 10% FBS and 1% penicillin/streptomycin. We routinely carry out morphology checks on all cell lines and we only passage the cell lines for 3 months. In addition, the cell lines were tested for Mycoplasma contamination. Furthermore, each experiment had appropriate controls to assure the behavior of tested cell lines. The compound ML264 was synthesized at The Scripps Research Institute in the laboratory of Dr. Thomas Bannister (25). The structure of ML264 compound and its synthesis pathway have been previously published (25). For in vitro experiments, ML264 was dissolved in DMSO (Fisher Scientific). For in vivo studies, ML264 was dissolved in the vehicle solution: 80% dH₂O, 10% DMSO, and 10% Tween 80. The antibodies used for this study are listed in Supplementary Table S1.

For cell proliferation experiments, DLD-1 and HCT116 cells were treated with 10 μmol/L ML264 or with vehicle (DMSO). Live cells were collected at 24, 48, and 72 hours after treatment and their numbers were determined by counting using a Coulter counter (Beckman Coulter). Each experiment was done in triplicate. In MTS assay, DLD-1 and HCT116 cells were treated with 10 μmol/L ML264 or with vehicle (DMSO). After 24, 48, and 72 hours of incubations, 20 μL of MTS solution (Promega, Cat. #G3582) was added to each well and an analysis was performed according to the manufacturer's protocol. The measurement of the control (cells with medium and DMSO) was defined as 100% and the results from other measurements were calculated accordingly. Each experiment was done in sextuplicate. A cell-cycle progression assay was performed as described previously (23). Each experiment was done in triplicate. The apoptosis rate was determined using the Alexa Fluor 488 Annexin V/Dead Cell Apoptosis Kit (Life Technologies, Cat. #V13241) according to the manufacturer's instructions with analysis by flow cytometry. Each experiment was done in triplicate.

Total protein was extracted from cells with Laemmli buffer and the analysis was performed as described previously (23).

Total RNA from DLD-1 and HCT116 cells was used for quantitative PCR. RNA was extracted using TRIzol Reagent (Life Technologies, Cat. #15596) according to the manufacturer's instructions. Primers against human KLF5, EGR-1, CTNNB1, CCND1, CCNE1, CCNA2, CCNB1, and GAPDH were purchased from Qiagen. Their respective catalogue numbers are QT00074676, QT00218505, QT00077882, QT00495285, QT00041986, QT00014798, QT00006615, and QT00079247. Quantitative PCR was performed using the QuantiTect SYBR Green RT-PCR Kit (Qiagen, Cat. #204243) as per standard protocols. Observed CT values were then used to calculate fold change using the 2-ΔΔCt method of relative quantification (26). Human GAPDH was used as the housekeeping gene.

Tumors dissected from mice were first fixed in Bouin's fixative (50% ethanol + 5% acetic acid in water) for 1 hour, then fixed overnight in 10% buffered formalin (Fisher Scientific). The tissues were then paraffin-embedded using an automated processor, sectioned at 5 μm, collected onto charged slides, and baked in a 65°C oven overnight, and were subsequently deparaffinized in xylene. Sections were incubated in a 2% hydrogen peroxide in methanol bath to block endogenous tissue peroxidases and were then rehydrated by incubation in a decreasing ethanol bath series (100%, 95%, and 70%) followed by antigen retrieval in citrate buffer solution (10 mmol/L sodium citrate, 0.05% Tween-20, and pH 6.0) at 120°C for 10 minutes using a decloaking chamber (Biocare Medical). Tissue sections were first incubated with blocking buffer (5% BSA in TBS-Tween) for 30 minutes at 37°C and then with primary antibody at 4°C overnight in a humidified chamber with gentle shaking. The list of primary antibodies used is shown in Supplementary Table S1. Sections were washed and incubated with secondary antibodies (horseradish peroxidase–conjugated or fluorescent-tagged) at the appropriate concentration for 30 minutes at 37°C. Betazoid DAB (Biocare Medical) was used to reveal IHC staining in tissues. For fluorescent sections, slides were washed after secondary antibody treatment and then stained with Hoechst (AnaSpec Inc.) and mounted with Prolong gold antifade (Life Technologies). Slides were analyzed under a Nikon Eclipse 90i microscope (Nikon) and representative photomicrographs were taken.

Histology of sections was observed upon staining 5 μm sections that were fixed, paraffin-embedded, deparaffinized, and rehydrated as mentioned before. Then they were stained with Hematoxylin Stain Solution, Gill 3 (Ricca Chemical Company), and Eosin Y (Sigma-Aldrich). Sections were dehydrated in an increasing series of ethanol bath (70%, 95%, and 100%), cleared in xylene, and mounted with Cytoseal XYL xylene-based mounting media (Thermo Scientific, Cat. #8312-4).

In in vitro studies, DLD-1 cells were seeded onto slides with medium containing DMSO or 10 μmol/L ML264, fixed and stained with Hoechst (DNA labeling) after 24, 48, and 72 hours of treatment. To quantify the number of mitotic figures five fields, each with 100 cells were counted. For in vivo studies mitotic figures were counted in 15 images of three vehicle- and ML264-treated animals using slides with hematoxylin and eosin (H&E) staining.

Ki-67 was quantified using images from three animals treated with vehicle and with ML264. We analyzed three images per animal per treatment using the ImageJ software (27).

All mice studies were approved by the Stony Brook University Institutional Animal Care and Use Committee. Nude mice were purchased from Jackson Laboratories (Bar Harbor). Animals were housed under specific pathogen-free conditions in ventilated and filtered cages under positive pressure. Xenograft tumors were generated by injecting subcutaneously 5 × 10⁶ DLD-1 human colorectal cells into the right flank of 6- to 7-week-old male nude mice. Tumor volume was determined by caliper measurement and calculated by established methods (28). When tumors reached a volume of about 100 mm³, mice were treated intraperitoneally with varying doses of ML264: 10 mg/kg daily, 10 mg/kg twice per day, and 25 mg/kg twice per day, with each treatment regimen lasting for a duration of 10 days. The vehicle solution was used as the control treatment. Mice were monitored and weighed every 2 days. Experiments were terminated when the tumor's greatest measurement reached 2 cm. Tumors were excised and retained for further analyses.

The analysis of in vitro experiments was performed with Student t test. A value of P < 0.05 was considered statistically significant. This analysis was performed using the GraphPad Prism version 5.00 for Windows (GraphPad Software). A comparison of tumor progression between the different treatment regimens and the control group was performed at 5 and 10 days after treatment by constructing linear mixed models for longitudinal data arising from every experiment. Autoregressive with order 1 structure was used to model the dependence among time points—the best structure compared with compound symmetric and unstructured. Prespecified comparisons between treatment and vehicle group were made. A P value less than 0.05 was considered statistically significant and analysis was performed using SAS 9.3 (SAS Institute Inc.).

ML264 is a third-generation small molecule compound that has been developed with the goal of inhibiting the growth of colorectal cancer cells by reducing the activity of the oncogene KLF5 (25). We designed a series of experiments to test the efficacy of ML264 using DLD-1 and HCT116 colorectal cancer cell lines due to the relatively high levels of expression of KLF5 in these cells. We tested the effects of ML264 on the rate of cell proliferation of colon cancer cells lines DLD-1 (Fig. 1A) and HCT116 (Fig. 1B) over 72 hours. ML264 efficiently inhibited the rate of proliferation of both cell lines. A significant decrease in proliferation was evident within 24 hours of treatment (Fig. 1A and B) and by 72 hours the live cell numbers of ML264- and vehicle-treated cells differed by 15- to 30-fold. Additionally, we performed an MTS assay that allows the quantification of metabolically active cells. As shown in Fig. 1C (DLD-1) and D (HCT116), both cell lines demonstrated a reduced level of viability that is in concordance with the results presented in Fig. 1A and B.

To determine the mechanism by which ML264 inhibits cancer cell proliferation, we first investigated the effect of ML264 on the number of mitotic figures in DLD-1 cells, cultured in the presence of ML264 or vehicle, over a 3-day period. As seen in Fig. 1E and F, ML264 treatment significantly reduced the number of cells undergoing mitosis in DLD-1 cells at 24, 48, and 72 hours. This result suggests that ML264 either prevents cells from entering into or progressing through mitosis. Next, we treated DLD-1 and HCT116 cells with ML264 or with vehicle for up to 72 hours and assessed the cell-cycle profile or rate of apoptosis by staining with propidium iodide (PI) alone (Fig. 2A and B) or in combination with Annexin V (Supplementary Fig. S1A and S1B), respectively. ML264 treatment led to a significant decrease in the population of cells in G₀–G₁ phase at 48 and 72 hours for both cell lines (Fig. 2A and B, top). In ML264-treated DLD-1 cells, there was a significant increase in the S-phase population at 24 hours, which persisted until 72 hours (Fig. 2A, middle). In ML264-treated HCT116 cells, an increase in S-phase cells was seen at 48 hours that persisted through 72 hours (Fig. 2B, middle). Additionally, we noted changes in the number of ML264-treated cells in the G₂–M phase. In DLD-1 cells, an initial decrease in G₂–M cells was seen at 24 hours and this was followed by an increase in this population at 48 and 72 hours (Fig. 2A, bottom). In contrast, ML264-treated HCT116 cells showed a decrease in G₂–M population at 24 and 48 hours (Fig. 2B, bottom). Furthermore, Annexin V/PI staining of these cells demonstrated only a modest, though significant, increase in apoptotic cells following ML264 treatment (Supplementary Fig. S1A and S1B; Annexin V and PI double-positive fraction). Early apoptotic cells (Annexin V single-positive) increased in ML264-treated DLD-1 cells at 48 hours and in HCT116 cells at 48 and 72 hours. In both DLD-1 and HCT116 cells, late apoptotic cells (Annexin V and PI double-positive fraction) increased. In DLD-1 cells the increase was noted at 48 hours and in HCT116 at 48 to 72 hours after treatment with ML264. It is unlikely that these modest changes in apoptotic cells explain the dramatic difference in cell number that was initially observed. Our data indicate that inability of DLD-1 and HCT116 cells to enter mitosis could be a consequence of S-phase arrest upon ML264 treatment. Thus, these results suggest that ML264 inhibits cell proliferation mainly by modifying cell-cycle progression, though the induction of apoptosis may also contribute to this phenotype.

ML264 and related analogues emerged from a screen that was designed to identify small molecule that inhibit the expression of oncogenic KLF5 and that therefore would inhibit the growth of colorectal cancer cells (23). We investigated how ML264 affects the expression levels of proteins involved in several of the signaling pathways (MAPK, WNT, and PI3K) that regulate either KLF5 expression and activity or the progression through the cell cycle (14, 24, 29, 30). We collected protein and RNA samples over a period of 3 days from DLD-1 and HCT116 colorectal cancer cells that had been treated with vehicle or with ML264 and analyzed them by Western blot analysis and by qRT-PCR. As seen in Fig. 3A and B, the protein levels of p-EGFR and EGFR are differentially modulated by ML264 in a cell line–dependent manner. In DLD-1 cells, the basal levels of EGFR protein are not significantly changed, yet there is an increase in the phosphorylation status of EGFR at 48 and 72 hours after treatment with ML264 as compared with control (Fig. 3A). On the other hand, we observed a slight decrease in the basal levels of EGFR in HCT116 cells treated with ML264 over the 3-day period and a decrease of its phosphorylation status at 48 and 72 hours (Fig. 3B). The basal levels of ERK decreased and this decline was accompanied by an upregulation of its phosphorylated form, in both DLD-1 and HCT116 cells. However, both ML264-treated cell lines showed a decrease in the protein levels of KLF5 that parallels a reduction in the levels of the transcription factor early growth response 1 (EGR1), a direct activator of KLF5 expression (Fig. 3A and B; ref. 13).

Previously, we and others have shown that KLF5 plays an important role in regulating the activity of β-catenin, although KLF5 in turn is regulated by the WNT and PI3K signaling pathways (18, 22, 24). Our Western blot analyses show a downregulation of the basal levels of AKT and GSK3β in both cell lines following ML264 treatment, a decrease that was accompanied by a similar decline in pAKT and pGSK3β levels (Fig. 3C and D). We also observed a decrease in the levels of the active, nuclear form of β-catenin, which is phosphorylated at serine 552. This reduction in β-catenin phosphorylated at serine 552 is likely the result of the decrease in AKT levels, as this kinase is known to regulate the phosphorylation status of serine 552 (31, 32). Additionally, we have noticed that in DLD-1 cells treated with DMSO β-catenin phosphorylated at serine 552 predominantly localizes to the cells that are undergoing mitosis and that it is absent in the DLD-1 cells treated with ML264 for the 3-day period (data not shown). Furthermore, we have analyzed the levels of β-catenin that is phosphorylated at threonine 41/serine 45, which is a pool of β-catenin destine for degradation as well as the levels of total β-catenin (33). No significant difference in total β-catenin and β-catenin phosphorylated at threonine 41/serine 45 were observed in ML264-treated DLD-1 cells. That observation, combined with the fact that levels of GSK3β were reduced, suggests that the degradation pathway of β-catenin is not upregulated upon ML264 treatment in DLD-1 cells. In contrast to DLD-1, HCT116 cells show reduced levels of both total β-catenin and the levels of β-catenin phosphorylated at threonine 41/serine 45 following ML264 treatment.

Based on our results, ML264 affects the RAS/MAPK signaling pathway primarily by decreasing the levels of EGR1 and KLF5. A preliminary qPCR analysis showed that there is a decrease in the mRNA levels of both genes already after 8 hours of treatment with ML264 relative to control (data not shown). This implies that ML264 may directly impact the transcription of EGR1 and KLF5 by targeting a common regulator or alternatively that ML264 may inhibit the expression of EGR1, which in turns diminishes the expression level of KLF5. As shown, ML264 modifies cycle-cell progression by inhibiting the progression through mitosis and by S-phase block (Fig. 2). Consequently, the changes in the PI3K/AKT pathway activity can have a dual origin. It has been previously shown that PI3K/AKT pathway can regulate the progression of cells through G₁–S and G₂–M phases. On the other hand, the modification of cell-cycle progression may in turn influence expression levels of AKT. Our data demonstrate that ML264 in DLD-1 and HCT116 cells decreased the levels of β-catenin phosphorylated at serine 552 and additionally in HCT116 ML264 treatment reduced the levels of total β-catenin (Fig. 3C and D). The reduced levels of β-catenin phosphorylated at serine 552 could be a result of downregulation of the expression level of AKT, and its reduced activation. The exact mechanism by which ML264 regulates the RAS/MAPK/PI3K and WNT signaling pathways is still unknown. Ongoing studies that pertain to the analysis of the protein and RNA levels of the components of the RAS/MAPK/PI3K and WNT pathways at much shorter intervals (e.g., hourly) are currently under way.

Altogether, these results suggest that ML264 inhibits the MAPK pathway by reducing EGR1 and KLF5 levels and further that this causes reduced levels of active β-catenin (β-catenin Ser552), which is associated with the downregulation of AKT.

As shown earlier, treatment with ML264 altered the cell-cycle progression of DLD-1 and HCT116 cells as compared with control (Fig. 2). It has been previously shown that KLF5 is a pro-proliferative factor in colorectal cancer cells and that it regulates the expression levels of cyclins D1 and B (7, 14, 34–38). We therefore analyzed the expression levels of cyclins D1, E, A2, and B1 in DLD-1 and HCT116 cells that were treated with ML264 or with vehicle in a manner corresponding to that shown in Fig. 2. As shown in Fig. 4A and B, cyclins E, A2, and B1 are downregulated in both cell lines with changes evident even in the first day of treatment with ML264. However, cyclin D1 levels are differentially regulated in these cell lines. In HCT116 cells, we observed a decrease in cyclin D1 levels at 24 and 48 hours after ML264 treatment (Fig. 4B). However, this trend was not apparent in ML264-treated DLD-1 cells (Fig. 4A). We next investigated whether the changes in protein levels were the result of decreased transcription of the corresponding cyclin genes. As shown in Fig. 4C and D, we failed to see a significant change in the cyclin D1 mRNA levels for either cell line that was treated with ML264 (with exception of DLD-1 cells after 72 hours treatment). However, mRNA levels for cyclins E, A2, and B1 in both cell lines were downregulated during the 3-day treatment with ML264. Here, we showed that ML264 affects cell-cycle progression by increasing the population of the cells in the S-phase and by inhibiting cells from entering mitosis or from transitioning through mitosis (Fig. 2). The rate of cell proliferation can be affected in many ways, for example, by perturbing the cell cycle, by arrest, and/or by promoting apoptosis (39, 40), and is associated with cell-cycle progression and is regulated by cyclins (e.g., A2, B1, and E1). Our results (Fig. 4A–D) show that ML264 affects the expression levels of RNA and protein of cyclin B1 (which is responsible for the transition through mitosis), cyclin A2 (which regulates G₁–S and G₂–M transition), and cyclin E1 (which plays a role in the G₁–S transition; ref. 41). Previously, we demonstrated that KLF5 is a positive transcriptional regulator of cyclin D1 and cyclin B1 (37). Current data suggest that upon treatment with ML264, cyclin B1 expression levels are decreased, which may be a direct effect of KLF5 downregulation. Additionally, it has been shown that KLF5 regulates the promoter activity of cyclin E, although promoting proliferation of vascular smooth muscle (42). The relationship between KLF5 and cyclin A2 has not been studied until now. Taken together, these results show that ML264 decreases the levels of EGR1 and KLF5 and modulates the cell-cycle progression by perturbing the levels of cyclins.

Subsequent to find that ML264 inhibits proliferation of colorectal cancer cell lines in vitro, we evaluated its effectiveness in inhibiting growth of tumor xenografts in nude mice. In these experiments, DLD-1 cells were subcutaneously injected into nude mice until a tumor volume of approximately 100 mm³ was achieved. The mice were then injected intraperitoneally for 10 days with ML264 according to the following regimens: 10 mg/kg (once per day), 10 mg/kg (twice per day), and 25 mg/kg (twice per day). In all cases controls were similarly maintained, DLD-1 cells were subcutaneously injected into nude mice until a tumor volume of approximately 100 mm³ was achieved, and mice were then injected intraperitoneally for 10 days with vehicle, as described in the Materials and Methods section. As shown in Fig. 5A, single daily injections of ML264 at 10 mg/kg did not significantly affect tumor growth. However, twice daily injections of ML264 at 10 or 25 mg/kg resulted in significant reductions in tumor growth (Fig. 5B and C), and this effect could be detected as early as 2 days after the first injection. The data also show that there is a concentration-dependent effect of ML264 on the tumor volume. Statistical analysis of tumor growth revealed significant tumor size reduction in mice treated twice daily with ML264 compared with those receiving only vehicle at day 5 and 10. It is noteworthy that none of the treatment regimens affected the weight of the mice (Fig. 5D–F). At the end of 10 days of treatment we harvested tumors from the treated and control mice, photographed the tumors, and sectioned them for further analysis. The photographs shown in Fig. 5G clearly demonstrate the dose-dependent differences in tumor size resulting from ML264 treatment as compared with controls.

Subsequently, we performed H&E staining of the tumors (Fig. 6A). As shown in Fig. 6B and C, there is a significant reduction in the number of mitotic figures in xenografts from mice treated twice daily with ML264 at 25 mg/kg. This explains the lack of xenograft growth in ML264-treated mice (Fig. 5C). The initial histologic analysis indicated that there is an increase in the level of fibrosis and inflammation in the xenografts treated with ML264 in comparison to vehicle treated. Thus, we stained the tumors for vimentin to detect fibroblasts associated with fibrosis. Here, we observed a significantly increased density of fibroblasts in the ML264-treated xenografts (Fig. 6D). Consequently, we stained the tissues for the presence of inflammatory cells. As shown in Fig. 6E, xenografts treated with ML264 had increased levels of mononuclear phagocytes staining for Mac-3. As KLF5 is pro-proliferative marker and a potential target for ML264's activity, we next investigated the impact of ML264 on KLF5 levels in these xenografts. As seen using IHC (Fig. 7A and Supplementary Fig. S2), there is a reduction in KLF5 levels and an absence of KLF5 from the nuclei in the xenografts from mice treated with ML264. This result was corroborated by a Western blot analysis for KLF5, shown in Fig. 7B. We also examined these xenografts for EGR1, a major transcriptional regulator of KLF5, and saw a significant reduction of EGR1 in xenografts from mice that had been treated with ML264 (Fig. 7C and Supplementary Fig. S3). Furthermore, we performed staining for Ki-67, a well-characterized marker of cell proliferation (43), to assess the proliferation status of the xenografts with or without treatment. As shown in Fig. 7D and E, the levels of Ki-67 staining are significantly reduced in xenografts from mice that were treated with ML264 as compared with those from control mice.

In the present study, we have shown that ML264 efficiently inhibits the growth of colorectal cancer cells in vitro (Fig. 1) and in vivo (Figs. 5–7). We were able to show that in an in vitro system ML264 regulates the cell-cycle progression while inducing apoptosis minimally (Fig. 2 and Supplementary Fig. S1). Interestingly, we have noticed that the arrest of cells in S-phase and the induction of apoptosis (Fig. 2 and Supplementary Fig. S1) are accompanied by increased levels of phosphorylated ERK and by decreased levels of phosphorylated AKT (Fig. 3). This phenomenon has been reported previously and has an important implications for the approach, particularly in the development of combinatorial drug treatments that target the MAPK signaling pathway (44–46). Our results demonstrate that ML264 significantly inhibits cellular proliferation, and mitosis in particular, in both in vitro and in vivo systems, as shown by analyses of (i) mitotic figures in DLD-1 cells treated in culture (Fig. 1E and F) and (ii) mitotic figures and Ki-67 staining of DLD-1–derived xenografts (Figs. 6B and C and 7D and E). Our studies showed that ML264 efficiently modulates cell-cycle progression and affects the expression pattern of cyclin B1, A2, and E (Figs. 1, 2, and 4). It has been previously shown that EGR1 and KLF5 regulate progression through the cell cycle by modifying the levels of cyclins D1 and B1, also that KLF5 affects cyclin E (37, 42). It is feasible that ML264 acts by decreasing the levels of EGR1 and KLF5, inhibiting the expression of the mentioned cyclins, and interrupting the cell-cycle progression.

ML264 originated from an ultrahigh-throughput screen to identify compounds that inhibit the proliferation of colorectal cancer cells by targeting the expression of KLF5 (25). In these experiments, we have demonstrated that ML264 is capable of reducing the expression levels of KLF5 in cells treated in culture (Fig. 3A and B) and in a tumor xenograft model of tumorigenesis (Fig. 7A and B and Supplementary Fig. S2). Moreover, we have found that this compound has a profound effect on the expression levels of EGR1, a direct transcriptional regulator of KLF5 (Figs. 3A and B and 7C, and Supplementary Fig. S3). At present, the exact mechanism by which ML264 inhibits KLF5 is unknown. Studies are currently under way to assess the mechanism of action and define if it has either a direct or indirect effect on KLF5 expression.

Our in vitro Drug Metabolism and Pharmacokinetics (DMPK) studies suggested that ML264 would have high stability to first pass metabolism due to its high stability upon exposure to hepatic microsomes. ML264 also does not inhibit the activity of cytochrome P450 isoenzymes, indicative of its potential for safety with respect to drug–drug interactions. Importantly, it was inactive against a selected panel of 47 kinases and 67 protein targets of therapeutic and/or toxicologic interest (25). Our in vivo DMPK studies demonstrated that ML264 has 2-hour half-life in mice. Interestingly, ML264 also displays 47% oral bioavailability in mice, a very promising feature for compound development with potential therapeutic advantages. Ongoing studies in our laboratory aim to identify the direct target of ML264 and to test its role in modulating the early steps of development of intestinal adenomas using Apc^Min/+/KRAS^V12G mice models.

No potential conflicts of interest were disclosed.

Conception and design: A. Ruiz de Sabando, C. Wang, T.D. Bannister, V.W. Yang, A.B. Bialkowska

Development of methodology: A. Ruiz de Sabando, C. Wang, Y. He, M. García-Barros, K.R. Shroyer, T.D. Bannister, V.W. Yang, A.B. Bialkowska

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Ruiz de Sabando, C. Wang, Y. He, J. Kim, T.D. Bannister, A.B. Bialkowska

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A. Ruiz de Sabando, M. García-Barros, K.R. Shroyer, T.D. Bannister, V.W. Yang, A.B. Bialkowska

Writing, review, and/or revision of the manuscript: A. Ruiz de Sabando, M. García-Barros, K.R. Shroyer, T.D. Bannister, V.W. Yang, A.B. Bialkowska

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): V.W. Yang

Study supervision: V.W. Yang, A.B. Bialkowska

The authors thank Dr. Jie Yang from the Department of Preventive Medicine, Stony Brook University, for assistance regarding biostatistical analysis of the tumor growth. They also thank Research Flow Cytometry Core in the Department of Pathology, Stony Brook University, for assistance with data analysis.

This work was supported by grants from the NIH (DK052230, CA084197, and CA172113 to V.W. Yang; CA172113 to A.B. Bialkowska; CA172517 to M. García-Barros; and CA172113 and MH084512 to T.D. Bannister).

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.