Abstract
Chemoprevention is considered a valid approach to reduce the incidence of colorectal cancer, one of the most common malignancies worldwide. Here, we investigated the tumor-preventive activity of curaxin CBL0137. This compound represents a new class of nonmutagenic DNA-binding small molecules that alter chromatin stability and inhibit the function of the histone chaperone FACT. Among downstream effects of CBL0137 treatment are activation of p53 and type I interferons and inhibition of NFκB, HSF1, and MYC. In addition, our data show that in both human and mouse colorectal cancer cells in vitro, CBL0137 inhibits the APC/WNT/β-catenin signaling pathway, which plays a key role in colon carcinogenesis. Using quantitative RT-PCR and microarray hybridization, we have demonstrated decreased expression of multiple components and downstream targets of the WNT pathway in colon cancer cells treated with CBL0137. At the same time, CBL0137 induced expression of WNT antagonists. Inhibition of WNT signaling activity by CBL0137 was also confirmed by luciferase reporter assay. Tumor-preventive activity of CBL0137 in vivo was tested in a murine model of colorectal carcinogenesis induced by 1,2-dimethylhydrazine (DMH), which is known to involve WNT pathway dysregulation. After DMH subcutaneous treatment, mice were administered CBL0137 in drinking water. Efficacy of CBL0137 in suppressing development of colorectal cancer in this model was evidenced by reduced incidence of adenocarcinomas and adenomas in both males and females and decrease in tumor multiplicity. These data support the prospective use of CBL0137 in chemoprevention of colorectal cancer as well as of other malignances associated with activated WNT signaling.
Introduction
Colorectal cancer is one of the most common types of malignancies, representing the third most frequently diagnosed cancer (1). Although colorectal cancer treatment options and outcomes have improved, recent advances have been more modest than anticipated. High cost of colorectal cancer treatment is often not justified by the amount of benefit it provides and at the same time, it is out of reach for many patients (2). Globally, there remain 1.2 million cases of colorectal cancer diagnosed each year and colorectal cancer accounts for more than 600,000 deaths annually (8% of all cancer-related deaths). The global burden of colorectal cancer is expected to increase by 60% to more than 2.2 million new cases and 1.1 million cancer deaths by 2030 (1).
Colorectal cancer presents an example of the combined impacts of genetics, epigenetics, and environment/lifestyle. Significant part of individuals diagnosed with colorectal cancer has a family history of cancer (3). Among genetic syndromes predisposed to colorectal cancer are familial adenomatous polyposis (FAP) caused by mutations in the adenomatous polyposis coli (APC) gene and hereditary nonpolyposis colorectal cancer (HNPCC or Lynch syndrome) caused by mutations in MMR genes encoding DNA mismatch repair proteins (4). Colorectal cancer frequently develops on the background of inflammatory bowel disease, and it is the most common cancer among such patients (4). The mechanisms underlying development of sporadic colorectal cancer involve mutations in specific oncogenes (e.g., APC, KRAS) and tumor suppressor genes (e.g., P53) as well as epigenetic changes (5).
Although much research and development effort has focused on colorectal cancer treatment, pharmacologic strategies aimed at prevention of the disease are both appealing and gaining support. Chemopreventive activity against colorectal cancer has been reported for a number of drugs, including NSAIDs, hormone replacement therapy in postmenopausal women, statins, bisphosphonate, and angiotensin inhibitors (6). The most promising data on colorectal cancer chemoprevention were obtained for NSAIDs, which showed tumor preventive effects in many animal studies (7). Epidemiologic studies also indicate that anti-inflammatory drugs have a chemopreventive effect on colon carcinogenesis (8). These data are consistent with the well-established role of inflammatory factors in colorectal cancer development (9). Unfortunately, both aspirin and other NSAIDs are associated with significant adverse effects, including increased risk of upper gastrointestinal and of cardiovascular toxicity, hemorrhagic stroke, and internal bleeding (10). Thus, although chemoprevention is a promising approach to reduce the incidence of colorectal cancer, new cancer-preventive drugs that are both safe and effective are needed.
Recent strategy for colorectal cancer chemopreventive drug development is based on improved understanding of the molecular events involved in colon carcinogenesis (11, 12). Aberrant activation of the APC/WNT/β-catenin signaling pathway (referred to herein as the WNT pathway) plays a major role in both sporadic and hereditary colorectal cancers, and targeting of this pathway for colorectal cancer prevention was recently highlighted as a priority research gap (13, 14). Importantly, NSAIDs, in addition to their anti-inflammatory activity, were shown to suppress a number of COX-independent mechanisms and the WNT signaling pathway, in particular. Chemopreventive activity against colorectal cancer was also described for natural polyphenols (e.g., resveratrol, genistein, and others), which was shown to inhibit WNT pathway in colorectal cancer cell lines and in chemically induced colorectal cancer in mice (15–17).
Paying attention to the fact that many natural polyphenols demonstrating colorectal cancer chemopreventive activity represent DNA-binding small molecules, we decided to study cancer-preventive effects of a new anticancer drug, small-molecule curaxin CBL0137. Recently, it was shown that this compound binds DNA noncovalently through intercalation and minor groove binding, inhibits topoisomerases (18), and influences epigenetic regulation (19) and DNA packaging (18). Chromatin alterations caused by CBL0137 lead to functional inhibition of the histone chaperone FACT and modulation of the activity of several transcription factors, including P53, NF-B, and HSF-1 (20, 21). Moreover, our recent study revealed CBL0137 influence on type I IFN signaling (19). Using colorectal cancer cell line HCT116, we demonstrated CBL0137 inhibitory effect on COX2 transcription (22). As COX2 catalyzes biosynthesis of prostaglandin E2, which is a known modulator of the WNT pathway, we proposed that CBL0137 might influence WNT signaling pathway (23). Thus, the main goals of this study were to analyze the effect of CBL0137 on the WNT signaling pathway in colorectal cancer cell lines in vitro and to evaluate its tumor-preventive effect in vivo in an animal model. We used DMH-induced colon carcinogenesis in mice (24, 25) as a well-established colorectal cancer model in which tumor development is known to be associated with activation of the WNT signaling pathway (26–28).
Materials and Methods
Chemicals and reagents
DMH (1,2-dimethylhydrazine 2-HC1) was purchased from Sigma-Aldrich Co. LLC; CBL0137 was provided by Incuron, Inc.
The following reagents were used: chloroform (Vekton, JSC, Russian Federation), histomix (BioVitrum LLC., RF), xylol (Merck & Co. Inc.), hematoxylin (Ferak Berlin GmbH, Eosin B (Sigma-Aldrich Co. LLC) and Mayer's Mucicarmine Stain Solution (Sigma-Aldrich), and dimethylsulfoxide (DMSO, Sigma-Aldrich Co. LLC).
Cell lines
Cell lines HT29, HCT116, SW480, and CaCo2 were obtained from ATCC. Cells were used no longer than 20 passages. Mycoplasma cell culture contamination was routinely checked using DAPI staining, followed by fluorescent microscopy. Cells were cultured in DMEM supplemented with 10% FCS and 1% penicillin–streptomycin (Paneco LTD, RF) in a humidified incubator at 37°C, 5% CO2. Mouse colorectal cancer cell line MC-38 (29) was obtained from the Istituto Scientifico Universitario San Raffaele, Milano, Italy.
Cell viability (MTT) assay
HT29, HCT116, SW480, and CaCo2 cells were seeded in 96-well plates (4,000 cells/well). After overnight incubation, CLB0137 was added to the medium to final concentrations of 0.075, 0.15, 0.3, 0.6, and 1.2 μmol/L. After 72 hours, 20 μL of freshly prepared 3-[4,5-dimethylthia-zol2-yl]-2,5-diphenyltetrazolium bromide (MTT, Merck; 5 mg/mL) was added to wells and the plates were incubated for 4 hours. Then, medium was removed and 150 μL of DMSO was added to each well. Relative cell viability was determined, respectively, to the untreated cells (30).
Microarray analyses
HCT116 cells were treated with 1 μmol/L of CBL0137 for 5 or 16 hours. RNA was isolated from cells using TRIZOL reagent (Invitrogen) and processed for hybridization with GeneChip Human Genome U133A 2.0 Array (Applied Biosystems, Thermo Fisher Scientific) in Genomics Facility of Roswell Park Cancer Institute. Two biological replicates of each condition were used for microarray hybridization. Raw data quantitation, background subtraction, and quantile normalization were done according to the manufacturer's instruction. Comparison of gene expression between samples was done using GeneSpring GX software (Agilent).
Quantitative RT-PCR
HT29, HCT116, SW480, and CaCo2 cells were seeded in 6-well plates (250,000 cells/well) and after overnight incubation CLB0137 was added to final concentrations of 0.25 μmol/L, 0.5 μmol/L, or 0.75 μmol/L. After 48 hours of incubation, total RNA was extracted using TRI reagent (Sigma-Aldrich) and purified using RNase-free DNase (Promega). Quantitative RT-PCR was performed using the Bio-Rad CFX Real-Time PCR System. Each sample was tested in triplicate, and results were normalized to the expression of RPL27. PCR primers were designed using BLAST (Supplementary Table S1).
Lentiviral T-cell factor/lymphoid enhancer factor luciferase reporter construct and luciferase assay
Lentiviral stocks were generated containing the construct Cignal Lenti T-cell factor/lymphoid enhancer factor (TCF/LEF) Reporter (Qiagen). HCT116 cells were infected with the lentivirus following the manufacturer's protocol. The infected cells were then treated with CBL0137. Luciferase activity was measured using a commercially available Luciferase Assay (Promega) and Luminometer TD 20/20 (TDI; ref. 30).
Animals
Tumor-preventive effects of CBL0137 were studied in CBA mice obtained from the Stolbovaya Farm of Federal Medical and Biological Agency (http://www.scbmt.ru). Two months old 20- to 22-g mice were used. The protocols for the experiments were approved by the Animal Care and Use Committee of the N.N. Blokhin NRMCO and corresponded to the guidelines for the welfare and use of animals in cancer research adopted by The United Kingdom Coordinating Committee on Cancer Prevention Research (31). All animals were housed 7 to 8 mice per cage with free access to drinking water and a pelleted basal diet.
Testing CBL0137 against DMH-induced colorectal cancer carcinogenesis in mice
Chemically induced colon carcinogenesis was reproduced on CBA mice by subcutaneous injections of 1,2-dimethylhydrazine according to the protocol described in the International Agency for Research on Cancer monograph 71 (24, 32, 33). CBL0137 was administering in drinking water as 0.13 mg/mL solution that corresponds to the maximal tolerated dose of 20 mg/kg/d that was detected and calculating according to the previously described protocol (34). One hundred fifty male and 153 female CBA mice were randomized into the groups of 35 to 44 animals as follows (see Fig. 1): groups 1.1 and 1.2—once weekly subcutaneous injection of DMH (dissolved in sterile water) at a dose of 8 mg/kg for 15 weeks (males) or 20 weeks (females; refs. 32, 33), followed by treatment with CBL0137 in drinking water at a dose of 20 mg/kg/d during the weeks 18 to 40 (males) or 23 to 40 (females); groups 2.1/2.2—DMH treatment only (as described for group 1.1/1.2); groups 3.1/3.2—CBL0137 treatment only (as described for group 1.1/1.2); and groups 4.1/4.2—no treatment. Animals were euthanized after the 40th week. All organs were examined for gross pathology and were also collected for histologic examination.
Scheme of the experimental design.
Histologic analysis
Gross lesions and normal-appearing tissues were fixed in 10% buffered formalin for 3 days. The samples were then dehydrated with alcohol, followed by chloroform and embedded in Histomix. After deparaffinization by m-Xylol and rehydration, sections were stained with hematoxylin and eosin according to the standard procedure (35). Mayer's Mucicarmine staining (according to the protocol of Sigma-Aldrich Co. LLC.) was used to reveal mucin-depleted foci (MDF).
Statistical analysis
Data processing was carried out using Statistics software (StatSoft Inc., 2001, version 6.0) and Origin 8.0 (OriginLab Corporation). The statistical significance of the differences between two samples in RT-PCR analysis and luciferase reporter assay were evaluated by paired samples Student t test.
The statistical significance of the difference between animal groups was calculated with Pearson χ2 test.
Results
CBL0137 shows cytotoxicity to colorectal cancer cell lines
As an initial step toward investigation of the potential utility of CBL0137 as a colorectal cancer preventive agent, we tested the effect of the drug on viability of four human colorectal cancer cell lines (HCT116, HT29, SW480, and CaCo2) and murine adenocarcinoma cell line MC-38 growing in culture. HCT116 harbors three base deletion in the CTNNB1 (β-catenin) gene and in the rest three lines mutations of APC gene were described (36, 37). Cells were treated for 72 hours with doses of CBL0137 ranging from 0.075 μmol/L to 1.2 μmol/L. Dose-dependent toxicity of CBL0137 was observed for all tested colorectal cancer cell lines, with IC50 values of 0.70 μmol/L, 0.67 μmol/L, 0.63 μmol/L, and 0.86 μmol/L in HT29, HCT116, SW480, and CaCo2 cells, respectively (Supplementary Fig. S1).
CBL0137 modulates expression of WNT signaling pathway components, target genes, and regulators in colorectal cancer cell lines
Because colorectal cancer is known to be associated with overactivity of the WNT signaling pathway, we next tested whether the observed influence of CBL0137 on human colorectal cancer cells might be associated with altered expression of genes involved in the WNT pathway. We compared RNA prepared from HCT116 cells left untreated or treated with 0.5 μmol/L CBL0137 for either 5 or 16 hours using microarray-based global gene expression profiling. We revealed CBL0137-dependent reductions in the expression of WNT signaling pathway receptor genes FZD1, FZD2, and FZD5 and 18 target genes (Fig. 2A; Supplementary Table S2). In contrast, CBL0137 activated expression of two known negative regulators of the WNT signaling pathway, APC2 and WiF1, by up to 2.83- and 37-fold, respectively.
Effect of CBL0137 treatment on WNT signaling in colorectal cancer cell lines. A, Changes in gene expression in HCT116 cells upon 5 or 16 hours of CBL0137 treatment (color scale shows log2 fold change relative to untreated cells). Data are shown for genes coding for WNT pathway component (n = 26), ligand (n = 1), receptors (n = 3), negative regulators (n = 4), and targets (n = 18). B, Relative expression level (fold change in cells treated with CBL0137 during 48 hours compared with untreated cells) of WNT target genes determined by quantitative RT-PCR. C, Relative expression level (fold change in CBL0137-treated cells compared with untreated cells) of WNT-signaling antagonists determined by quantitative RT-PCR. D, Luciferase TCF/LEF reporter analysis. B–D, Influences of CBL0137 on the gene expression levels were significant with P < 0.05.
To extend our findings from the HCT116 microarray hybridization experiment, we used quantitative RT-PCR to analyze the effect of CBL0137 on expression of several target genes of the WNT signaling pathway in all four colorectal cancer cell lines (HT29, HCT116, SW480, and CaCo2). RNA was analyzed from cells treated with 0.25 μmol/L, 0.5 μmol/L, or 0.75 μmol/L CBL0137 for 48 hours. Three tested β-catenin transcriptional targets, proproliferative Cyclin D1, proproliferative and antiapoptotic cMYC, and antiapoptotic Survivin (BIRC5), all showed significant CBL0137 dose-dependent inhibition of expression (Fig. 2B). On the other hand, significant dose-dependent increases in expression were observed for three negative regulators of WNT signaling: DKK3, SFRP1, and WIF-1 (Fig. 2E).
It is well known that TCF/LEF transcription factors are the major endpoint mediators of Wnt pathway throughout metazoans. To determine whether the observed CBL0137-dependent changes in gene expression indeed resulted in altered WNT/β-catenin signaling in treated cell lines, we tested the effect of the compound on expression of a TCF/LEF luciferase reporter construct stably integrated into the genome of HCT116 cells. As shown in Fig. 2D, CBL0137 inhibited expression of the reporter in a dose-dependent manner (up to 3-fold with 0.5 μmol/L CBL0137) after 24-hour treatment. CBL0137 dose-dependent inhibition of reporter expression was also observed when WNT signaling was activated by LiCl (up to 5-fold with 0.5 μmol/L CBL0137), as LiCl enhances WNT pathway via GSK3β inhibition. These data were in accordance with the results obtained by quantitative PCR: the treatment of cells with CBL0137 (0.50 μmol/L) decreased luciferase expression up to 5-fold.
Together, these results show that CBL0137 treatment of human colorectal cancer cell lines causes reduced expression of WNT-pathway components and target genes and yet increased expression of antagonists of the pathway, resulting in suppression of WNT/β-catenin signaling.
Before the analysis of CBL0137 preventive effect in vivo, we checked whether CBL0137 influences Wnt pathway in mouse colorectal cancer cells in vitro using MC-38 adenocarcinoma cells, which was obtained by Corbett and colleagues using DMH-induced chemical carcinogenesis (29). We revealed strong inhibiting effect of CBL0137 on the expression of the main Wnt targets: cMyc, Ccnd1, Birc5, CD44, and Cox2 in MC-38 cells (Fig. 3).
Effect of CBL0137 treatment on WNT signaling in MC-38 adenocarcinoma cells: relative expression level (fold change in cells treated with CBL0137 during 48 hours compared with untreated cells) of WNT target genes determined by quantitative RT-PCR.
Effect of CBL0137 on DMH-induced colon carcinogenesis in mice
To determine whether the effects of CBL0137 on Wnt signaling both in human and mouse colorectal cancer cells observed in vitro would contribute to the tumor-preventive effect in vivo, we used a well-established mouse model of colon carcinogenesis induced by DMH. Mice were given subcutaneous injections of DMH once weekly for 15 weeks (males) or 20 weeks (females) after which CBL0137 was provided via drinking water ad libitum. Four treatment groups were established both for males and females as shown in Figure 1. Group 1 received DMH and CBL0137, group 2 received DMH alone, group 3 received CBL0137 alone, and group 4 did not receive either compound (untreated control group).
Over the course of the experiment, general animal health, as determined by cage-side observations, was similar between all four study groups. As expected, a steady increase in mean animal body weight was observed in all groups during the experiment. However, consistent with the previously published results of Turusov and colleagues (33), groups of animals injected with DMH gained weight more slowly than groups not given DMH. CBL0137 (20 mg/kg in drinking water) did not affect weight gain in either DMH-treated or DMH-untreated mice. For both males and females, there was no significant difference in mean body weight at the end of the experiment between groups that received DMH and CBL0137 versus DMH alone or between groups that received CBL0137 alone versus no treatment (Supplementary Table S3). Thus, CBL0137 did not have any effect on animal body weight by itself and it did not alter the effect of DMH.
DMH-induced colon carcinogenesis was evaluated in all study animals through gross and histopathologic observation following euthanasia at 40th week, with particular focus on the colon as described in the Methods. In collected colons, we analyzed the frequency of adenomas and adenocarcinomas, tumor multiplicity, and the frequency of mucin-depleted/mucin-producing tumors and MDF. The last ones are crypt foci with absent or scant mucous production. They are considered to be preneoplastic lesions, as mucus layer plays an important role in the intestine homeostasis: it provides a physical separation between the epithelial cell monolayer and the luminal contents, a chemical and biochemical barrier that supports the structure of the mucus gel, as well as the ability of the gel to concentrate biological factors secreted by mucosal epithelial cells (38). Recently, it was shown that Muc2 deficiency generates a chronic, low-level inflammatory response and eventual tumor development. Representative images of stained histologic colon sections are shown in Figure 4. There were no signs of colon carcinogenesis in any study animal that did not receive DMH. Thus, no colon adenomas or adenocarcinomas were detected in the untreated group or in the group that received only CBL0137. These animals also lacked MDF in normal-appearing tissue. These findings are presented in Supplements (Supplementary Table S4).
Histologic analysis of colon tissue observed in control and DMH–treated CBA mice. A–D, Microphotographs of H&E-stained sections of mouse colon tissue. A and B, Normal colon epithelium (×40; ×200). C, Tubular adenoma induced by DMH (×100). D, A representative adenocarcinoma induced by DMH (×100). E and F, Mouse colon tissue stained with mucicarmine, which reveals mucus (pink), and hematoxylin-staining basophilic cell structures (blue). E, Mucin-producing normal epithelial tissue (×500). F, Mucin-depleted epithelium of early-stage adenoma (marked area; ×400).
Histologically, the morphology of the colon epithelium of males and females treated with CBL0137 showed no differences compared with corresponding samples from the untreated control group. Normal colon tissue is shown in Figure 4A, B, E.
In contrast, adenomas and adenocarcinomas were observed in the colons of mice treated with DMH (Fig. 4C and D). In addition, some DMH-treated mice had foci within normal-appearing epithelium and/or adenomas characterized by absent or scarce mucous production (Fig. 4F).
Comparisons of overall tumor incidence (adenoma and adenocarcinoma) and tumor multiplicity (number of tumors per animal) clearly demonstrated a tumor-preventive effect of CBL0137 in this study (Fig. 5). Among the 36 males treated with DMH, 28% of animals did not show any colon tumors, 42% developed a single tumor, and 30% had more than two tumors (Fig. 5A; Supplementary Table S4). In contrast, among 44 males treated with DMH and CBL0137, tumors did not appear in 77% of animals, 14% of mice had one tumor, and only four animals (9%) had more than two tumors. A similar tumor-preventive effect of CBL0137 was observed in females. In the group of 39 mice given DMH alone, 28% of mice did not develop any colon tumors, 33% developed a single tumor, and 36% developed more than two tumors. Of 44 females treated with DMH and CBL0137, 69% of animals did not bear any tumors, 21% animals developed one tumor, and only 11% had two or more tumors. Thus, CBL0137 administering after DMH injections both greatly increased the proportion of animals without tumors and decreased tumor multiplicity per animal.
Prevention of DMH-induced colorectal carcinogenesis in mice by CBL0137. A, Colon tumor (adenoma + adenocarcinoma) multiplicity in male and female mice, given DMH injections alone or followed by CBL0137 treatment. B, Adenoma and adenocarcinoma incidence in male and female mice left untreated (control) or treated with DMH alone, DMH + CBL0137, or CBL0137 alone. C, Incidence of MDF in normal-appearing colon tissue, adenomas with mucin-depleted epithelium, and adenomas with mucin-producing epithelium among male and female mice treated with DMH alone or DMH + CBL0137. B and C, Asterisks indicate statistically significant differences between DMH-treated and DMH + CBL0137–treated groups (Pearson's χ2 test, P < 0.01). More detailed information could be found in the Table S4. D, The number of mucin-depleted and mucin-producing adenomas in the DMH-treated and DMH + CBL0137–treated male and female groups. The difference in the proportions of mucin-depleted/mucin-producing adenomas in males is significant, P = 0.06.
Although adenomas were detected much more frequently than adenocarcinomas (Fig. 5B; Supplementary Table S4), incidence of both tumor types was reduced by CBL0137 administering, resulting in increased proportions of tumor-free animals. The difference in the proportion of animals with adenomas (irrespective of tumor multiplicity) between the group of DMH-treated mice and the group of mice treated with DMH and CBL0137 was highly statistically significant (P < 0.01) for both males (54% and 23%, respectively) and females (64% and 25%, respectively; Fig. 5B). CBL0137 similarly reduced the proportions of male and female animals with adenocarcinomas, but the difference was statistically significant only in males (Fig. 5B). Notably, in males, CBL0137 treatment following DMH injection completely eliminated the appearance of adenocarcinomas, as compared with their development in 19% of mice that received DMH alone. In females, the incidence of adenocarcinomas was reduced from 21% to 12%.
The incidence of mucin-depleted adenomas was also significantly lower in both males and females treated with CBL0137 after DMH injection than in those that received DMH alone (24% vs. 55% for males; 19% vs. 68% for females; Fig. 5C). It should be noted that CBL0137 treatment in males led to a substantial reduction in the proportion of mucin-depleted adenomas: from 17/19 in the group of DMH-treated mice to 8/15 in the group of males treated with DMH and then CBL0137 (Fig. 5D).
In the group of females treated with DMH alone, the multiplicity of adenomas was 6.12 (adenomas per mouse) whereas in the group of females treated with DMH, followed by CBL0137, this parameter was 1.64. Therefore, CBL0137 decreased the multiplicity of adenomas in females by 3.73-fold.
For accurate presentation of the in vivo experiment results, we should point out that besides carcinogenesis in the small and large intestines, we observed appearance of the following malignant tumors: anal tumors, renal angiosarcomas in male mice, and uterine sarcomas and ovarian hemangiomas. All the results are presented in the Supplementary Figure S2. We observed statistically significant tumor-preventive influence of CBL0137 on the frequency of DMH-induced renal capsule angiosarcomas in males and uterine sarcomas in females.
Overall, the results of this in vivo experiment demonstrate high chemopreventive efficacy of CBL0137 against colorectal cancer in both male and female mice.
Discussion
WNT and NFκB signaling pathways have been identified as playing key roles in colon carcinogenesis, and thus they are well recognized as promising drug targets for colorectal cancer prevention and/or treatment (4, 12). Inhibition of NFκB by a new class of DNA-binding small molecules (curaxins) represented by CBL0137 has been demonstrated in various colorectal cancer cell lines (20, 22). Here, we report for the first time that CBL0137 also inhibits WNT signaling.
This was demonstrated in colorectal cancer cell lines using three independent approaches: microarray-based gene expression analysis, quantitative RT-PCR, and luciferase reporter assay. Microarray analysis in HCT116 cells revealed a significant inhibitory effect of CBL0137 on expression of many components and downstream targets of the WNT signaling pathway, while expression of WNT antagonists genes WIF1 and APC2 was enhanced. Using quantitative RT-PCR, we demonstrated a dose-dependent inhibitory effect of CBL0137 on expression of several WNT signaling target genes (cMYC, CCND1, and BIRC5) in four different human colorectal cancer cell lines. We also observed dose-dependent activation of WNT signaling antagonists SFRP1, WIF1, and DKK3. It should be noted that naturally occurring inactivation of these WNT signaling antagonists in human colorectal cancer (39) and in DMH/AOM-induced mouse colon tumors occurs mainly via epigenetic silencing (40). CBL0137 is a DNA-binding small molecule and is known to alter chromatin structure through its inhibition of the histone chaperone FACT (18, 21). Our recent demonstration of CBL0137 epigenetic activity using HeLa-TI cells harboring an epigenetically silenced GFP reporter gene (18) suggests that such activity may underlie the observed upregulation of WNT signaling antagonists in colorectal cancer cell lines. Reduced expression of a luciferase reporter gene controlled by the TCF/LEF promoter in colorectal cancer cells treated with CBL0137 provided additional support for our conclusion that CBL0137 is an efficient inhibitor of WNT signaling.
This conclusion is consistent with and further supported by our previously published works showing that CBL0137 has anti-inflammatory effects including COX2 inhibition (20, 22), because one of the bioactive products of COX-2 function is prostaglandin E2 that regulates WNT signaling (23). Moreover, CBL0137 was previously shown to activate NOTCH1 expression in small-cell lung cancer, and an interplay between β-catenin signaling and NOTCH1 effectors has been demonstrated in intestinal tumorigenesis (41). Finally, it is possible that the effect of CBL0137 on cMyc expression that was previously thought to be mediated by CBL0137′s modulation of FACT function (18) could be also due to the impact of the drug on WNT signaling.
The significance of our finding that CBL0137 inhibits WNT signaling is highlighted by the well-appreciated fact that aberrant activation of WNT signaling plays a pivotal role in colorectal carcinogenesis in mammals. Germline mutations in the APC gene result in FAP, the major hereditary predisposition event leading to colorectal cancer development (42).
Having demonstrated that CBL0137 is cytotoxic toward human and mice colorectal cancer cell lines and inhibits the Wnt signaling pathway both in human and murine adenocarcinoma cells, we next tested the drug for chemopreventive activity in vivo using a murine model of colorectal cancer induced by DMH. Rodent models of DMH-induced colorectal cancer were developed after it was recognized in a Guamanian population study that hydrazines from cycad flour induce colon cancer (43). The hydrazine and, in particular, unsymmetric dimethylhydrazine (UDMH), represent rather widespread human chemical carcinogens (24). UDMH is an industrial chemical that enters the environment primarily by emissions from its use in aerospace fuels and from industrial facilities that manufacture, process, or use it (44). DMH is used in experimental study of colon tumors with morphologic and histologic features and pathogenesis similar to those seen in human sporadic colorectal cancer (24). In DMH-induced colon carcinogenesis, similarly to human colorectal cancer, Wnt signaling activation is regarded as critical early events (45, 46). In spite of the fact that the patterns of somatic mutations between mouse and human colorectal cancer are in some way different (47), both of them affect Wnt signaling (28, 44). Notably, aberrant activation of Wnt signaling both in human colorectal cancer and in DMH-induced colon tumors is due to a lack of β-catenin phosphorylation (14, 27, 28). In mammals, Wnt signaling is mainly regulated via proteasomal degradation of β-catenin after its phosphorylation by the β-catenin destruction complex and subsequent ubiquitination. In human colorectal cancer, this process is often disrupted by APC mutations that inactivate the β-catenin destruction complex. In mouse tumors, phosphorylation of β-catenin by the destruction complex is obstructed by point mutations in the GSK-3β phosphorylation consensus motif of β-catenin. Thus, the consequences of β-catenin mutations in mouse DMH-induced tumors and APC and β-catenin mutations in human colorectal cancer are similar. COX2 overexpression is another molecular alteration that is observed both in mouse DMH-induced tumors and in human colorectal cancer (28). Together, these facts indicate that the murine DMH-induced colorectal cancer model is an appropriate animal model for testing the effect of CBL0137 colon carcinogenesis in vivo.
The results of our in vivo study demonstrated a clear, statistically significant chemopreventive effect of CBL0137 against colorectal cancer in both male and female mice. At the same time, chronic administering of CBL0137 via drinking water did not cause any detectable changes in normal organs and tissues of mice of either gender. The latter finding is consistent with the results of previous studies demonstrating that CBL0137 is generally safe and well tolerated in mammals (48). It is well known that colon adenomas and carcinomas appear in men at an earlier age and at a higher rate than in women. This was also observed in rodents, particularly in mice treated with azoxymethane (48). These gender-based differences in colon carcinogenesis are likely due to the tumor-promoting effects of testosterone. Taking this into consideration, different protocols of carcinogen treatment have been developed for males and females (32, 33). Applying this to our study, we administered DMH to males for 15 weeks and to females for 20 weeks. Among DMH-treated males, CBL0137 administering reduced adenoma incidence by approximately 2.5-fold and completely prevented appearance of adenocarcinomas. The proportion of males that remained tumor-free till the end of the study increased from 24% to 73% with CBL0137 treatment. Moreover, although the proportion of mucin-producing adenomas in the group of DMH-treated males was 2/17, in the group of males given CBL0137 after DMH, the corresponding proportion changed to 5/8. As Blache and colleagues (49) showed that MUC-2 gene, encoding the main intestinal mucin, is suppressed by SOX9, which is one of the main targets of Wnt signaling pathway, statistically significant difference (P value = 0.06) in the proportions of mucin-producing and mucin-depleted adenomas between the analyzed groups may be considered as an indirect confirmation of CBL0137 influence on Wnt signaling pathway in DMH-transformed cells in vivo (Fig. 4C; Supplementary Table S4).
In DMH-treated females, CBL0137 inhibited adenoma and adenocarcinoma incidence up to 2.6- and 1.8-fold, respectively. The proportion of females without colon tumors in the group treated by DMN and CBL0137 was higher (68%) than in the group of females treated by DMH only (28%).
Consistent with our findings in the mouse DMH-induced colorectal cancer model, chemopreventive activity of CBL0137 was previously demonstrated in two other preclinical cancer models. First, chronic administering of CBL0137 was shown to inhibit tumor onset in MMTV-neu mice without causing any detectable changes in normal organs and tissues (34). The preventive effect of CBL0137 was evidenced by reduced incidence of spontaneously developing mammary carcinomas, delayed tumor progression, and prolonged animal survival. Second, in a murine model of MYCN-driven neuroblastoma, low-dose CBL0137 administering on days 6 to 28 after birth significantly delayed tumor growth in TH-MYCN+/+ mice (50). Thus, there is a growing body of evidence supporting the potential clinical usefulness of CBL0137 as a chemopreventive agent.
In summary, our results indicate that CBL0137 is an efficient inhibitor of WNT signaling in colorectal cancer cell lines. This reveals an additional mechanism of action for multitargeted anticancer agents, the curaxins. Together with the previously demonstrated NFκB-inhibitory and anti-inflammatory activities of CBL0137, suppression of WNT signaling provides a mechanistic explanation for the significant chemopreventive influence of CBL0137 on DMH-induced murine colon carcinogenesis. These data highlight the strong potential of CBL0137 as an agent with broad applicability against many types of tumors associated with aberrant activation of WNT signaling.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Authors' Contributions
Conception and design: K. Kirsanov, E. Shipaeva, G.A. Belitsky, A.V. Gudkov, K.V. Gurova, M.G. Yakubovskaya
Development of methodology: K. Kirsanov, T. Fetisov, V. Maksimova, L. Trukhanova, E. Antoshina, T. Gor'kova, O. Morozova, A. Safina, A.A. Purmal, A.V. Gudkov, M.G. Yakubovskaya
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K. Kirsanov, T. Fetisov, E.A. Lesovaya, V. Maksimova, L. Trukhanova, E. Antoshina, T. Gor'kova, O. Morozova, A. Safina, M.G. Yakubovskaya
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K. Kirsanov, R. Salimov, G.A. Belitsky, M.G. Yakubovskaya
Writing, review, and/or revision of the manuscript: K. Kirsanov, E.A. Lesovaya, E. Shipaeva, A.A. Purmal, G.A. Belitsky, K.V. Gurova, M.G. Yakubovskaya
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D. Fleyshman, R. Ivanov, M.G. Yakubovskaya
Study supervision: R. Salimov, E. Shipaeva, A. Leonov, M.G. Yakubovskaya
Acknowledgments
We thank Angelo Manfredi and Annalisa Capobianco (Istituto Scientifico Universitario San Raffaele, Milano) for kindly providing us with the mice colon cancer MC-38 cell line.
This work was supported by Incuron LLC, and by Russian Science Foundation (to Marianna G. Yakubovskaya, 17-15-01526).
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.
Footnotes
Note: Supplementary data for this article are available at Cancer Prevention Research Online (http://cancerprevres.aacrjournals.org/).
Cancer Prev Res 2020;13:53–64
- Received April 15, 2019.
- Revision received August 13, 2019.
- Accepted October 17, 2019.
- Published first October 25, 2019.
- ©2019 American Association for Cancer Research.