Fluorinated N,N'-Diarylureas As Novel Therapeutic Agents Against Cancer Stem Cells

Colorectal cancer is the second leading cause of cancer deaths in the United States (1). This high mortality rate is largely attributed to tumor metastasis and recurrence (2). At initial presentation, more than half of all patients display metastatic disease in regional lymph nodes or systemic organs (e.g., liver). The five-year survival rate for patients with metastatic colorectal cancer is approximately 13%, as compared to 90% for nonmetastatic disease (1). A growing body of evidence suggests that cancer stem cells play a role in the resistance and spread of cancer cells (2–4). Cancer stem cells are a subpopulation of tumor cells that are capable of self-renewal and display resistance to DNA damage-induced cell death, a ruinous property that allows metastatic tumors to survive initial treatments and repopulate. Developing new antineoplastic agents that selectively target cancer stem cells represents a potentially valuable treatment to improve colorectal cancer outcomes (2–4).

Cancer stem cells possess demanding metabolic requirements and survival mechanisms that are potential targets for treatment (5). One such critical pathway involves AMP-activated protein kinase (AMPK), which coordinates growth, autophagy, and metabolism (6). AMPK acts as a metabolic sensor for the ratio of AMP to ATP in cells and has several downstream transcriptional targets (6). Among the antineoplastic effects of AMPK activation are cell-cycle arrest through stabilization of p53 and cyclin-dependent kinase inhibitors p21^waf1 and p27^kip1, inhibition of macromolecular synthesis, inhibition of mTORC1, and inhibition of the Warburg effect (7, 8). In support of the particular importance of AMPK in cancer cells and cancer stem cells, the widely studied AMPK activator, metformin, reduces cancer incidence in diabetic populations for many different cancers, including colorectal cancer (9). Furthermore, in vitro and in vivo studies with metformin specifically show cytotoxicity toward cancer stem cells (10, 11).

We identified the antineoplastic effects of fluorinated N,N'-diarylureas (FND) in a high-throughput screening program (12) and found that several potent FNDs inhibited growth of colorectal cancer cell lines through inhibition of the mTOR pathway (12). Overexpression of mTORC1 and mTORC2 components, Raptor and Rictor, is important to tumorigenesis (13), and the activation of AMPK regulates cell growth by suppressing mTORC1 through direct phosphorylation of the tumor suppressor, TSC2, and Raptor (6). Through this mechanism, we anticipated that AMPK activation would directly inhibit colorectal cancer cell proliferation.

In this study, we investigated the ability of eight FNDs to inhibit growth and induce apoptosis in colorectal cancer metastatic cell lines and stem cells. Activation of AMPK by all FND compounds successfully inhibited cell-cycle progression and subsequent cellular proliferation. These results demonstrate that FNDs exhibit considerable promise in the treatment of metastatic colorectal cancer, predominantly through the inhibition of colorectal cancer stem cells.

FNDs were synthesized as previously described (12). Table 1 shows the FNDs used in this study. Stock solutions (10 mmol/L) in DMSO were stored at −20°C.

Human cell line HT29 was purchased from ATCC. KM20 was provided by Dr. Isaiah J. Fidler (M.D. Anderson Cancer Center, Houston, TX); STR validation analysis (Bio-Synthesis) yielded a full DNA profile and no contamination by another human cell line was detected; there were no matches after searching ATCC, JCRB, or DSMZ databases. HCT116 pik3ca wild-type and mutant cell lines were a gift from Dr. J. Wang (14). Human colorectal cancer stem cell line 1 (no. 36112-39; lot no. 12121800-05) and stem cell line 2 (no. 36112-39; lot no. 1313161-12) were purchased from Celprogen. Cancer stem cells were limited to less than 12 passages.

Cell lines were routinely grown as monolayer cell cultures in 5% CO₂ in air, and 100% relative humidity at 37°C. HT29 and KM20 cell lines were grown in McCoy's 5A medium (Sigma-Aldrich) and supplemented with 10% FBS and 1× antibiotic–antimycotic (Life Technologies). Stem cell lines were grown in Cancer Stem Cell Complete Growth Media with Serum without antibiotic on precoated flasks with Human Colon Cancer Stem Cell Extra-cellular Matrix (both from Celprogen). Cell passages were carried out by detaching adherent cells in a logarithmic growth phase by addition of a mixture of 0.25% trypsin with 0.02% EDTA (Sigma Aldrich) and incubating for 10 to 15 min at 37°C. The number of viable cells was estimated with a cell counter V-CELL XR (Beckman Coulter). Metformin HCl was purchased from Seleckchem.

For each experiment, cell lines were seeded in two 96-well plates in normal medium (5 × 10³ cells/well, 100 μL). At 24 hours, 100 μL of media with drugs at different concentrations were added to each well. DMSO was used as a treatment control. Cell viability was measured using the Cytoscan-SRB Cell Cytotoxicity Assay (G-Biosciences) according to manufacturer's instructions. Cell viability was plotted as a percentage relative to DMSO treatment alone.

Total protein lysates were resolved on a 4%–12% Bis-Tris gel and transferred to Immobilon PVDF transfer membranes. Membranes were incubated for 1 hour at room temperature in blocking solution (TRIS-buffered saline containing 10% nonfat dried milk and 0.1% Tween 20), followed by an overnight incubation in primary antibodies at 4°C. Membranes were washed three times and incubated with horseradish peroxidase-conjugated secondary antibodies for 1 hour. After three additional washes, the immune complexes on the membranes were visualized using Immobilon Western Chemiluminescent HRP substrate (EMD Millipore) or Amersham ECL (GE Life Sciences).

Antibodies for Western blot analysis included the following: PARP (#9542, 1:1,000), Phospho-AMPKα (#2531, Thr172, 1:1,000), phospho-S6 ribosomal protein (Ser235/236; Cell Signaling); cyclin D1 (Abcam; #AB34175, 1:5,000); β-actin (Sigma Aldrich, #A5441, 1:20,000); anti-rabbit and anti-mouse (Santa Cruz Biotechnology, #SC-2054, #SC-2055, 1:3,000).

The original patient colorectal cancer tumor (F0 generation) was divided and implanted into the flanks of NOD scid gamma mice (The Jackson Laboratory; 005557). When the resulting tumors (F1 generation) grew to 1 cm³, they were resected, divided into 2-mm³ pieces, and implanted into 5 mice (F2 generation). All animal studies were performed in accordance with policies of the Institutional Animal Care and Use Committee and were approved by the Institutional Review Board of the University of Kentucky.

Liberase DH Research Grade (05401054001; Roche Applied Science) was resuspended in sterile water at a 2.5 mg/mL concentration and stored in single-use 100 μL aliquots at −80°C. collagenase/hyaluronidase (07912; StemCell Technologies) was aliquoted into single-use 250 μL aliquots and stored at −80°C. Upon collection, PDX tumors (F2 generation) were placed into DMEM complete cell culture media supplemented with 1× Gibco antibiotic–antimycotic (15240-062; Life Technologies) for transportation. Tumor fragments were minced into 2-mm cubes using scissors and digested in 50 μg/mL Liberase DH (100 μL) and 0.5× collagenase/hyaluronidase (250 μL), diluted in 5 mL of DMEM serum-free media for 4 hours at 37°C with gentle agitation by magnetic stirring bar. No undigested tissue was observed; cells were washed twice with complete cell culture media and transferred into 10% FBS DMEM media supplemented with 1× Gibco antibiotic–antimycotic.

Analysis of variance was used for comparison of cell viability across DMSO and varying dose levels of FND treatment, and adjusted P values using the Holm's procedure were calculated for pairwise comparison of each dose level with DMSO.

We performed a high-throughput screen of N,N'-diarylureas to identify potent AMPK activators and selected eight active agents (Table 1; ref. 12). Next, HT29 cells were treated with the eight agents (FND 4a, 4b, 4c, 4d, 4e, 4f, 4g, 4h) for 48 hours. Decreased cell viability was evident at a drug dose as low as 5 μmol/L, and was below 50% for all eight compounds treated at the 25 μmol/L dose (Table 2). To determine whether the prominent decrease in cell viability was due to cell-cycle inhibition or apoptosis induction, KM20 and HT29 colorectal cancer cells were treated with each of the eight FNDs for 24 hours at 25 or 50 μmol/L. Cyclin D1 expression and PARP cleavage were analyzed by Western blot analysis. Cyclin D1 expression was drastically decreased in both cell lines after treatment with FND 4b, 4c, 4d, 4e, 4f, 4g, and 4h at 25 μmol/L (Fig. 1A). Prominent PARP cleavage was observed after treatment at the 50 μmol/L dose, notably after treatment with the 4b compound (Fig. 1B).

Colorectal cancer stem cell lines were treated with FND 4b, 4d, 4f, 4g, and 4h at 50 μmol/L for 24 hours and analyzed for cyclin D1 expression and PARP cleavage. Cyclin D1 expression was decreased in all treatment groups; prominent PARP cleavage was observed in both stem cell lines only after treatment with FND 4b (Fig. 2A). Activation of apoptosis is especially important for elimination of cancer stem cells, because these cells are progenitors for tumor progression; FND 4b treatment induced apoptosis in both metastatic cancer cells and cancer stem cells. To confirm that AMPK activation was dose responsive in colorectal cancer stem cells, we treated colorectal cancer stem cell line 1 with 5, 25, and 50 μmol/L of FND 4b. Our results showed a dose-dependent AMPK activation after FND 4b treatment (Fig. 2B).

Next, we used compound 4b to evaluate dose-dependent AMPK activation in KM20 and HT29 cells. We observed AMPK activation starting with as little as 10 μmol/L (Fig. 3A). We also examined the dose dependent effects of FND 4b treatment on cell viability (Fig. 3B), cell cycle, and apoptosis induction in KM20 and HT29 cells. KM20 cells were more sensitive to compound 4b treatment and showed marked suppression of cyclin D1 expression at a dose of 10 μmol/L. PARP cleavage was detected in both cell lines starting at a 40 μmol/L dose.

We evaluated whether the pik3ca mutation protects cancer cells against FND 4b treatment. HCT116 pik3ca mutant and wild-type colorectal cancer cell lines were treated with FND 4b at 10, 20, 30, 40, and 50 μmol/L for 24 hours. Both cell lines showed strong cell-cycle inhibition but required higher dosages compared to KM20 and HT29 cell lines. PARP cleavage was noted with FND 4b treatment at 20 μmol/L (Fig. 4A). These results show that the HCT116 pik3ca mutant and wild-type cells have a similar pattern of cell-cycle inhibition and apoptosis in response to FND 4b treatment.

Metformin, which is a known activator of AMPK and is used for diabetes treatment and cancer prevention, activates AMPK at millimolar levels (15). AMPK downregulates TOR/S6K activity via TSC2 in response to stress (16). We treated HT29 and KM20 cells with FND 4b at 10, 20, and 30 μmol/L and with metformin at 5, 10, and 20 mmol/L for 24 hours to compare the effect on cyclin D1 expression and S6 phosphorylation at Ser234/235. Both metformin and FND 4b decreased phosphorylation of the S6 protein, a downstream effector of mTOR that is a crucial signaling pathway for proliferation and survival of cancer cells in vitro as well as tumors in vivo (Fig. 4B and C). Our results demonstrated that FND 4b had a similar effect on cell-cycle inhibition and inhibition of S6 phosphorylation at the 20 μmol/L dose as compared to the 10 mmol/L dose of metformin.

To determine the effect of FND treatment on AMPKα activation, HT29 cells were treated with low dosages (i.e., 5 and 10 μM) of FND 4b for 12, 24, 48, and 72 hours. FND treatment demonstrated a long-lasting effect on AMPK activation at 10 μmol/L (Fig. 5A). Low-dose FND 4b treatment resulted in a strong AMPKα activation, and decreased cyclin D1 expression in colorectal cancer stem cells at 48 hours (Fig. 5B and C). Next, we established a cell line from a patient-derived xenograft (PDX) and treated it with low-dose FND 4b for 48 hours. Cyclin D1 expression was markedly suppressed in the PDX cell line (Fig. 5D). FND treatment at 5, 7.5, and 10 μmol/L significantly decreased stem cells and PDXL proliferation (Fig. 5 B–D, bottom). Low-dose FND treatment did not increase PARP cleavage in all tested cells (data not shown). These results demonstrate a long-lasting effect of low-dose FND 4b treatment on AMPKα activation and cell-cycle inhibition.

The central regulatory role played by AMPK in energy homeostasis makes it an attractive target for treatment of various diseases such as diabetes and cancer (6, 7, 17). Recent efforts have focused on elucidating the complex role of AMPK in cancer cell viability (18). We examined the effects of eight fluorinated N,N'-diarylureas as AMPK activators on colorectal cancer metastatic and stem cell lines to evaluate their potential utility as chemotherapeutic agents for colorectal cancer. The FNDs that were used in this study represent a new class of AMPK activators with no effect on AKT or ERK, unlike other N,N’ diarylureas [12]. We confirmed that FND treatment of colorectal cancer metastatic and stem cell lines resulted in dose-dependent AMPK activation. High doses of traditional AMPK activators, such as AICAR and metformin, were required to achieve anticancer effects (10–12, 19). However, similar to FNDs, more recent AMPK activators showed antitumorigenic effects at lower concentrations. Chen and colleagues (19) showed that plumbagin, an AMPK activator, induced apoptosis and inhibited growth of colorectal cancer cell lines at a treatment concentration of 10 μmol/L. Kaushik and colleagues (20) showed that honokiol, another AMPK activating compound, inhibited melanoma growth and spheroid-forming capacity (i.e., stemness) through AMPK activation at concentrations up to 50 μmol/L. Finally, Valtorta and colleagues (21) showed that one to four diaryl-azetidinone, also a novel AMPK activator, inhibited human colorectal cancer xenografts in mouse models at IC₅₀ in the nmol/L to 1 μmol/L range. Taken together, these studies suggested that targeting cancer cells with novel AMPK activators represent an effective strategy for the treatment of different types of cancers. Thus, the beneficial effects on metastatic cancer cells by AMPK activation provide a rationale for further studies, to determine the potential clinical utility of AMPK activators, like FNDs, on cancer stem cells.

In our study, we evaluated the effect of FNDs on colorectal cancer stem cells. Many cancer stem cell models have been described, but a unifying theme in tumorigenesis is that tumors are not homogeneous. A hierarchical structure within tumors includes a small subset of cells with long-term self-renewal and differentiation capacity (3). Colorectal cancer, in particular, is known to possess cancer stem cells (22). To survive, repopulate, and differentiate, cancer stem cells have developed several mechanisms that provide resistance to the effects of antineoplastic agents. These mechanisms include resistance to DNA damage, drug penetration, apoptosis, and reactive oxygen species (ROS; refs. 3–5). Upregulation of the DNA repair machinery and downregulation of telomerase allow cancer stem cells to maintain their genetic integrity. Cancer stem cells also upregulate efflux transporters from the ATP-binding cassette (ABC) gene family to efficiently extrude chemotherapy agents. Resistance to apoptosis occurs, in part, through Akt pathway activation and increased production of apoptosis inhibitor proteins. Finally, cancer stem cells increase their free radical scavenging machinery to protect themselves from ROS damage (4). Several studies have investigated the role of stem cells in treatment resistance (2, 11). For example, Zhou and colleagues (2) created cisplatin-resistant colorectal cancer cells and found that these cells had stem cell-like markers. Treatment with salinomycin caused an accumulation of ROS and down-regulation of antiapoptotic molecules, which suggested that salinomycin was able to overcome the cisplatin resistance of colorectal cancer cells. Because cancer stem cells possess these survival mechanisms, cancer treatment plans should include therapy to specifically target cancer stem cells in order to reduce future recurrence and metastasis.

Although the specific role of AMPK in cancer stem cells compared with the non-stem cell population has not yet been fully characterized, several studies have shown that metformin, an AMPK activator, is preferentially cytotoxic to cancer stem cells in pancreatic and breast cancers (10, 11, 23). Similarly, in our present study, we showed that FND treatment resulted in cell-cycle arrest in colorectal cancer stem cells. AMPK activation led to inhibition of macromolecular synthesis, inhibition of mTORC1, and onset of the “anti-Warburg effect” (7, 24, 25), all of which led to the inhibition of cancer cell proliferation. Hardie and Alessi (7) described several AMPK tumor suppressor functions that included cell-cycle arrest through stabilization of p53, p21, and p27. Particularly in cancer stem cells, it is likely that the anti-Warburg effect targets the utilization of anaerobic glycolysis by cancer stem cells, and decreases their ability to use AKT-mTOR. AMPK activation of apoptosis resulted from p53 phosphorylation; phosphorylated AMPK subsequently accumulated in the mitochondria and interacted with Bak to induce apoptosis (8). Our results demonstrated that AMPK activation reduced the ability of cancer stem cells to hijack energy metabolism for growth and proliferation and triggered an apoptotic cascade. In summary, we demonstrated that FNDs activated AMPK at micromolar concentrations, which resulted in consistent cell-cycle arrest and apoptosis in all cancer cell lines that were studied. These findings provide additional information regarding the role of AMPK activation in cancer stem cells and demonstrate a potential role of FNDs in the treatment of metastatic colorectal cancer.

C. Liu has ownership interest (including patents) in Liu-Watt, LLC. No potential conflicts of interest were disclosed by the other authors.

Conception and design: P. Rychahou, C. Liu, D.S. Watt

Development of methodology: P. Rychahou, D.S. Watt

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): D.E. Kenlan, V.M. Sviripa, D.S. Watt

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D.E. Kenlan, V.M. Sviripa, H.L. Weiss, D.S. Watt

Writing, review, and/or revision of the manuscript: D.E. Kenlan, P. Rychahou, V.M. Sviripa, C. Liu, D.S. Watt, B.M. Evers

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D.E. Kenlan, D.S. Watt

Study supervision: P. Rychahou, D.S. Watt, B.M. Evers

The authors thank the Markey Cancer Center's Research Communications Office for help with manuscript preparation, and the Biostatistics and Bioinformatics Shared Resource Facility for statistical analysis. The contents of this manuscript are solely the responsibility of the authors and do not necessarily represent the official views of the grant funding agencies.

NCI P30 CA177558 (to B.M. Evers), NCATS UL1TR000117 (to B.M. Evers), R01 DK048498 (to B.M. Evers), Markey Cancer Foundation (to B.M. Evers), NIH R21 CA139359 (to C. Liu and D.S. Watt), NIH R01 CA172379 (to C. Liu and D.S. Watt), Office of the Dean of the College of Medicine (to D.S. Watt), NIH P20 RR020171 (to D.S. Watt).

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.