1-Benzyl-indole-3-carbinol is a highly potent new small molecule inhibitor of Wnt/β-catenin signaling in melanoma cells that coordinately inhibits cell proliferation and disrupts expression of microphthalmia-associated transcription factor isoform-M

The synthetic indolecarbinol 1-benzyl-I3C inhibits proliferation in melanoma cells by disrupting Wnt signaling and downstream expression of the master regulator MITF-M, independent of oncogenic BRAF. Our preclinical study implicates 1-benzyl-I3C as an attractive candidate molecule for development of novel treatment therapies for human melanoma.

Abstract

1-Benzyl-indole-3-carbinol (1-benzyl-I3C), a synthetic analogue of the crucifer-derived natural phytochemical I3C, displayed significantly wider sensitivity and anti-proliferative potency in melanoma cells than the natural compound. Unlike I3C, which targets mainly oncogenic BRAF-expressing cells, 1-benzyl-I3C effectively inhibited proliferation of melanoma cells with a more extensive range of mutational profiles, including those expressing wild-type BRAF. In both cultured melanoma cell lines and in vivo in melanoma cell-derived tumor xenografts, 1-benzyl-I3C disrupted canonical Wnt/β-catenin signaling that resulted in the downregulation of β-catenin protein levels with a concomitant increase in levels of the β-catenin destruction complex components such as glycogen synthase kinase-3β (GSK-3β) and Axin. Concurrent with the inhibition of Wnt/β-catenin signaling, 1-benzyl-I3C strongly downregulated expression of the melanoma master regulator, microphthalmia-associated transcription factor isoform-M (MITF-M) by inhibiting promoter activity through the consensus lymphoid enhancer factor-1 (LEF-1)/T-cell transcription factor (TCF) DNA-binding site. Chromatin immunoprecipitation revealed that 1-benzyl-I3C downregulated interactions of endogenous LEF-1 with the MITF-M promoter. 1-Benzyl-I3C ablated Wnt-activated LEF-1-dependent reporter gene activity in a TOP FLASH assay that was rescued by expression of a constitutively active form of the Wnt co-receptor low-density lipoprotein receptor-related protein (LRP6), indicating that 1-benzyl-I3C disrupts Wnt/β-catenin signaling at or upstream of LRP6. In oncogenic BRAF-expressing melanoma cells, combinations of 1-benzyl-I3C and Vemurafenib, a clinically employed BRAF inhibitor, showed strong anti-proliferative effects. Taken together, our observations demonstrate that 1-benzyl-I3C represents a new and highly potent indolecarbinol-based small molecule inhibitor of Wnt/β-catenin signaling that has intriguing translational potential, alone or in combination with other anti-cancer agents, to treat human melanoma.

Introduction

Melanomas are the most aggressive form of human malignant skin cancer (1), and the canonical- or β-catenin-dependent Wnt signaling pathway (2, 3) has been implicated to play a critical role in melanoma proliferation, progression, tumor survival, metastasis and chemoresistance (4). In the absence of Wnt, a β-catenin destruction complex is maintained in which Axin and adenomatous polypsosis coli (APC) provide the scaffold to tether active glycogen synthase kinase-3β (GSK-3β), which phosphorylates β-catenin to signal the β-TrCP-mediated ubiquitination and subsequent degradation of β-catenin (5). Wnt binding to its co-receptors, the Frizzled family seven-pass transmembrane receptors along with the one of two members of the low-density lipoprotein receptor-related protein family (LRP5 and LRP6), triggers the phosphorylation and recruitment of disheveled to the co-receptor complex as well as recruits GSK-3β and Axin to LRP5/6 away from the destruction complex (6). As a result, the loss of GSK-3β-dependent phosphorylation of β-catenin allows β-catenin to escape its ubiquitination and degradation. The stabilized β-catenin protein is imported into nucleus where it interacts with the lymphoid enhancer factor/T-cell transcription factor (LEF/TCF) to induce expression of tissue-specific sets of target genes (7, 8). In human cancer cells, expression of β-catenin-regulated gene networks can help drive proliferation and contribute to maintenance of tumorigenic phenotypes (9–11).

Human melanomas can be categorized by distinct mutational profiles that determine the corresponding phenotypes, proliferative capabilities and therapeutic options (12, 13). Several studies implicate an oncogenic role for enhanced Wnt signaling in melanomas that can result from the production and secretion of high levels of Wnt proteins and/or the constitutive or aberrant functioning of downstream components in the Wnt signaling cascade such as β-catenin, Axin and APC (14, 15). For example, differences in expression levels of Wnt2, Wnt5a, Wnt7 and Wnt10b subtypes correlate with the histopathological features of melanoma tumors (16), and many primary melanoma tumors display elevated levels of nuclear β-catenin (17). Constitutive activation of Wnt/β-catenin signaling was shown to enhance the growth of murine melanoma cells in vitro (18), and in a conditional mouse model of melanoma with a melanocyte-specific PTEN loss and expression of oncogenic BRAF-V600E increasing or decreasing β-catenin levels led to enhanced or repressed metastasis, respectively (19).

Wnt-driven signaling has also been proposed to play a role in therapeutic escape of melanomas (20). Because approximately 90% of human melanomas express an oncogenic form of BRAF, a key treatment strategy for these patients is the use of BRAF-specific inhibitors such as Vemurafenib (21). Elevated Wnt5A expression was observed in subsets of tumors from patients exhibiting resistance to BRAF inhibitor therapy (14) and was shown to correlate with melanoma progression and poor outcomes with BRAF inhibitor treatment (22). In melanoma cells, the efficacy of Wnt-regulated signaling can be linked to expression of microphthalmia-associated transcription factor isoform-M (MITF-M), the master regulator of melanocyte and melanoma biology (18). MITF-M is a ‘lineage survival’ oncogenic transcription factor that is amplified in approximately 20% melanomas, and its expression levels correlate with decreased overall patient survival (23) and acquired resistance to BRAF inhibitors (24–27). MITF-M has been shown to reprogram multivescicular bodies/late endolysozomes vesicular traffic so that GSK-3β and Axin are sequestered away from the other components of the β-catenin destruction complex, which leads to stabilization of nuclear β-catenin. In a complementary manner, interaction of the LEF/β-catenin transcription complex with MITF-M promoter induces its expression, whereas MITF-M protein can be stabilized by decreased GSK-3β-mediated phosphorylation (28). Therefore, because of the positive feedback loop between Wnt and MITF-M, small molecule inhibitors of the canonical Wnt signaling pathway could potentially inhibit MITF-M expression and melanoma cell proliferation as well as prevent or delay development of acquired resistance to BRAF inhibitors.

We and others have shown that indole-3-carbinol (I3C), a natural phytochemical from cruciferous vegetables such as broccoli, brussels sprouts and cabbage, triggers anti-cancer signaling cascades in a variety of cultured human cancer cells (29–41), including melanoma (42–45), and in vivo in human cancer cell-derived tumor xenografts (46). Furthermore, our previous studies have uncovered the only two known I3C target proteins, elastase and the ubiquitin ligase NEDD4-1, that trigger-specific anti-proliferative cascades in breast cancer and melanoma cells, respectively (42, 47–49). We have identified 1-benzyl-I3C as a highly potent synthetic analogue of I3C that displays significantly enhanced potency compared with the natural phytochemical in human breast cancer cells due to higher efficacy of interactions with its target protein elastase and its increased cellular stability and lipophilic properties (49, 50). In the present study, we demonstrate that 1-benzyl-I3C inhibits proliferation at high potency of a wide range of melanoma cells unrestricted by the presence of oncogenic BRAF. Furthermore, 1-benzyl-I3C strongly disrupts Wnt/β-catenin signaling at or upstream of the Wnt co-repressor LRP6 and downregulates MITF-M expression as part of the anti-proliferative response. Our study demonstrates that 1-benzyl-I3C is a highly potent new Wnt signaling inhibitor that has the potential to be used for targeted therapy of human melanoma.

Materials and methods

The methods used to complete the CCK-8 proliferation assay, flow cytometry, western blot analysis, RT-PCR, tumor xenografts, immunofluorescence, luciferase assays using MITF-M promoter-linked reporter plasmids, transfection of an MITF expression vector and chromatin immunoprecipitations are all standard procedures carried out as described previously (42, 43) and are detailed in the Supplementary Material. The melanoma cell lines G361, SK-MEL-28, SK-MEL-2, SK-MEL-30 and RPMI-7951 and melanocytes were purchased from American type culture collection (ATCC) (Manasas, VA), and were authenticated according to the ATCC guidelines. DM738 melanoma cells were acquired from the tissue culture facility at University of California, Berkeley. The cell culture and treatment conditions are described in the Supplementary Material.

L-cell-conditioned media for TOP FLASH reporter assay

Murine fibroblast L cells stably transfected with a construct expressing high levels of humanWnt3A, and its corresponding empty vector transfected control L cells was obtained as a kind gift from Dr Henk Roelink (Department of Molecular and Cell Biology, University of California at Berkeley). The cells were grown in dulbecco’s modified eagle’s medium (DMEM) containing 4.5 g/L Glucose, 114 mg sodium pyruvate and 2 mM L-glutamine, supplemented with 10% fetal bovine serum (Gemini Bio Products), 2 mM L-glutamine and 2.5 ml of 10000 U/ml penicillin/streptomycin mixture (Gibco, Life Technologies). Conditioned media was harvested in two batches, once during routine media change after 2 days of seeding the cells and once more when the plates were confluent. The cells were trypsinized as described earlier using 0.025% trypsin (Lonza) and pelleting the cells by centrifugation at 4000 rpm for 5 min. The supernatant fractions were filtered through a 0.2 micron filter to obtain the conditioned media. The two batches of conditioned media were mixed in ratio 1:1 and stored in aliquots at −80°C for future use.

TOP FLASH reporter assay

G361 cells were cultured to approximately 80% confluency in a six-well plate (Nunc). The TOP FLASH/FOP FLASH constructs (Millipore) were transfected using superfect transfection reagent (Qiagen) according to the manufactures instructions. Cells in each well were transfected with 2 μg DNA and 4 µl superfect for 2–3 h and then replaced with 2 ml of 1:1 ratio of fresh and conditioned media from cultured L cells secreting Wnt or no Wnt. After 24 h, cells were treated with or without 10 μM 1-benzyl-I3C for another 24 h in conditioned media. Cells were washed with dulbecco’s phosphate-buffered saline (DPBS), resuspended in lysis buffer (Promega), incubated on ice for 15 min, and then centrifuged at 14000 rpm for 1 min at 4°C. Supernatant fractions (20 µl) from each sample was combined with 100 μl luciferase substrate (Promega), and emitted fluorescence was measured using luminometer, Lumat LB 9507 (EG&G Berthold) and the relative light units (RLU) were recorded. The experiment was performed in triplicates and replicated at least 3 times.

Results

Human melanoma cells with different mutational profiles are highly sensitive to the anti-proliferative effects of 1-benzyl-I3C

We initially analysed the potential sensitivity of five different human melanoma cell lines with distinct mutational profiles and normal epidermal melanocytes to 1-benzyl-I3C and I3C. The melanoma cell lines express different combinations of the most common melanoma-associated mutations, namely BRAF, PTEN and N-RAS (Figure 1A table), whereas the melanocytes express wild-type forms of each gene. Based on our previous observations with breast cancer and melanoma cells (43, 46), the melanoma and melanocyte cell lines were treated with or without the optimum concentration of 20 μM 1-benzyl-I3C and 200 μM I3C for 48 h and cell proliferation determined using a CCK-8 assay. Comparison with the percent inhibition of proliferation showed that each of the melanoma cell lines, regardless of their mutational profiles, was highly sensitive to the anti-proliferative effects of 1-benzyl-I3C, unlike the oncogenic BRAF-dependent response of I3C (Figure 1A) (43). The two most sensitive melanoma cell lines (G361 and DM738 cells) to I3C displayed an approximate 50% inhibition of proliferation under conditions in which 1-benzyl-I3C inhibited cell proliferation by almost 90%. Wild-type BRAF-expressing SK-MEL-2 cells were sensitive to anti-proliferative effects of 1-benzyl-I3C, although to a lesser extent than either G361 or DM738 cells. In contrast to both oncogenic BRAF-expressing melanoma cell lines, SK-MEL-2 cells were insensitive to the anti-proliferative effects of I3C (Figure 1A). The proliferation of human normal epidermal melanocytes was unaffected by either 1-benzyl-I3C or I3C, indicating that melanoma cells are significantly more sensitive to the anti-proliferative effect of this synthetic I3C analogue. However, it is worth mentioning that the cultured epidermal melanocytes proliferate much slower than the cultured melanoma cells, which may influence the ability to detect potential indolecarbinol-mediated changes in proliferation.

Figure 1.

Anti-proliferative and cell cycle effects of 1-benzyl-I3C in melanoma cells displaying a range of mutational profiles. (A) Human melanoma cell lines with distinct sets of driver mutations in BRAF, N-RAS and PTEN (see table) as well as normal epidermal melanocytes were treated with 20 µM 1-benzyl-I3C (1B-I3C), 200 µM I3C or with the DMSO vehicle control for 48 h. Cell proliferation was quantified using a CCK-8 assay relative to vehicle control cells. Results are shown as percent inhibition of proliferation. (B) Oncogenic BRAF-V600E-expressing G361 and DM738 melanoma cells as well as wild-type BRAF-expressing SK-MEL-2 melanoma cells were treated with the indicated concentrations of 1-benyzl-I3C (1B-I3C), I3C or the oncogenic BRAF inhibitor Vemurafenib (Vem) for 48 h. Cell proliferation was quantified using a CCK-8 assay. (C) G361, DM738 and SK-MEL-2 melanoma cells were treated with the indicated concentrations of 1-benzyl-I3C (1B-I3C) for 48 h. The DNA content of propidium iodide stained nuclei was quantified by flow cytometry as described in the Materials and method section. The histograms of representative experiments from three independent trials are shown and the percentage of cells in the population displaying G₁ (dark gray), S (light gray) or G₂/M-DNA (moderate gray) content was quantified.

The effects of the clinically used oncogenic BRAF inhibitor Vemurafenib was compared with that of 1-benzyl-I3C and I3C in oncogenic BRAF-V600E-expressing G361 and DM738 cells and in wild-type-expressing SK-MEL-2 cells. Melanoma cells were treated with increasing concentrations of each compound for 48 h and cell proliferation monitored by a CCK-8 assay. As shown in Figure 1B (left and middle panels), in the BRAF-V600E-expressing cells, the 1-benzyl-I3C anti-proliferative responses were at least 10-fold more potent compared with I3C. The concentration of 1-benzyl-I3C needed to observe the maximal proliferative arrest was comparable with the optimally effective concentration of Vemurafenib. Interestingly, 1-benzyl-I3C inhibited proliferation of SK-MEL-2 melanoma cells, which express wild-type BRAF, whereas, these cells remained relatively resistant to the effects of either I3C or Vemurafenib (Figure 1B, right panel). Thus, melanoma cells are sensitive to the anti-proliferative effects of 1-benzyl-I3C independent of the presence of oncogenic BRAF.

To begin to understand the molecular mechanism underlying the anti-proliferative effects of 1-benzyl-I3C, flow cytometry analysis of propidium iodide stained nuclei was performed on G361, DM738 and SK-MEL2 cells treated with the indicated concentrations of 1-benzyl-I3C for 48 h. In all three melanoma cell lines, 1-benzyl-I3C dose dependently induced a G₁ cell cycle arrest, and consistent with the proliferation results, the maximal concentrations of 1-benzyl-I3C caused nearly 90% of the cells in the overall cell populations to arrest in the G₁ phase of the cell cycle (Figure 1C). For the G361 cells, concentrations of 1-benzyl-I3C greater than 10 µM caused a strong apoptotic response, and therefore we could not complete the cell cycle analysis. The synthetic indolecarbinol 1-benzyl-I3C had no effect on cell cycle progression of human melanocytes (see Supplementary Material). Time course studies of G361, DM738 and SK-MEL2 melanoma cells treated with or without 20 µM 1-benzyl-I3C up to 72 h showed that the downregulation of G₁ cell cycle regulator proteins, CDK2, CDK4, CDK6 and Cyclin D1, and upregulation of the CDK inhibitor p21 temporally correlated with the G₁ cell cycle arrest (see Supplementary Material).

Inhibition of MITF-M expression is necessary for the 1-benzyl-I3C anti-proliferative response

Expression of several critical G₁-regulatory genes, such as CDK2 and CDK4, is regulated by MITF-M (51), therefore, MITF-M protein levels and transcript expression were examined in G361, DM738 and SK-MEL-2 melanoma cells over a 72 h time course of cells treated with or without 20 µM 1-benzyl-I3C. MITF-M protein levels were analysed by western blots of cell extracts and transcripts levels determined by RT-PCR of isolated total RNA. As shown in Figure 2A and B, 1-benzyl-I3C treatment caused a rapid and strong downregulation of MITF-M gene products in all three melanoma cell lines relative to the stable production of the control protein HSP90 and control transcript glyceraldehyde 3-phosphate dehydrogenase (GAPDH), respectively. Characteristic of MITF-M, we observed several protein bands in the western blots between 55 and 60 kDa, which is probably due in part to differences in the phosphorylation state (23, 44).

Figure 2.

1-Benzyl-I3C-mediated downregulation of MITF-M gene expression and rescue of the anti-proliferative response by expression of exogenous MITF-M. (A) G361, DM738 and SK-MEL-2 melanoma cells were treated with or without 20 µM 1-benzyl-I3C over a 72 h time course. The levels of MITF-M protein and the control HSP90 protein were determined by western blot analysis of electrophoretically fractionated total cell extracts. (B) G361, DM738 and SK-MEL-2 melanoma cells were treated with or without 20 µM 1-benzyl-I3C for 48 h, and the levels of MITF-M transcripts and control GAPDH transcripts determined by RT-PCR analysis of isolated total RNA. (C) G361 melanoma cells were either transfected with the pCMVMITF expression vector, or the pCMV empty vector control or left untransfected, and each set of cells were then treated with or without 20 μM 1-benyzl-I3C for a submaximal time of 24 h. Cell proliferation was measured using a CCK-8 assay, and results show the mean of three independent experiments ± SEM (*P < 0.01). (D) G361 cells transfected as described in panel C were treated with or without 20 µM 1-benzyl-I3C for 24 h. The protein levels of MITF-M and of two of its downstream target genes, the anti-apoptotic BCL2 and pro-proliferative CDK4, were determined in comparison with the control HSP90 protein by western blot analysis of total cell extracts.

To assess whether the downregulation of MITF-M expression is necessary for the 1-benzyl-I3C anti-proliferative response, exogenous MITF-M protein was expressed in G361 melanoma cells by transient transfection of the pCMVMITF expression vector. Another set of G361 melanoma cells was transfected with the pCMV empty vector, and a third set of cells remained untransfected. Cells were treated with or without 1-benzyl-I3C for 24 h, a time frame shorter than required for complete proliferative arrest but necessary to maintain the cellular levels of the transfected plasmids. Cell proliferation was monitored using a CCK-8 assay, and the protein levels of MITF-M as well as two MITF-M target genes, the cell cycle regulator CDK4 and the anti-apoptotic gene Bcl2, were examined by western blots. Compared with empty vector transfected and untransfected cells, expression of exogenous MITF-M maintained total MITF-M protein at a high level in the presence of 1-benzyl-I3C, strongly attenuated the 1-benzyl-I3C anti-proliferative response (Figure 2C) and rescued the downregulation of CDK4 and Bcl2 (Figure 2D).

Effects of 1-benzyl-I3C on production of Wnt/β-catenin signaling pathway components in cultured melanoma cells and in vivo in melanoma cell-derived tumor xenografts

The effects of 1-benzyl-I3C on expression of specific components of the canonical Wnt cascade were examined in oncogenic BRAF-expressing G361 and DM738 cells as well as in wild-type BRAF-expressing SK-MEL-2 melanoma cells treated over a 72 h time course with or without 20 µM 1-benzyl-I3C. As shown in Figure 3A, western blots of total cell extracts revealed that 1-benzyl-I3C downregulated the protein levels of β-catenin by approximately 75% at the 72 h time point (densitometry data not shown) and upregulated levels of the GSK-3β and Axin components of the β-catenin destruction complex between approximately 3-fold and 8-fold at the 72 h time point depending on the cell line (densitometry data not shown). Although there were cell line variation in the absolute levels, by 72 h treatment with 1-benzyl-I3C, we observed that the Wnt co-receptor LRP5/6 was downregulated between 50 and 90%, and the LEF-1 transcription factor, which interacts with nuclear β-catenin to stimulate expression of LEF-1/β-catenin target genes, was downregulated by approximately 6080% (densitometry data not shown). Thus, 1-benzyl-I3C disrupts the expression of components of the Wnt/β-catenin cascade in a range of human melanoma cells, although one limitation of the interpretation of the synthetic indolecarbinol analogue can selectively target melanoma cells is that we were not able to analyse the Wnt/β-catenin protein components in cultured epidermal melanocytes.

Figure 3.

Effects of 1-benzyl-I3C on levels of the Wnt/β-catenin signaling pathway components in cells and in tumor xenografts. (A) BRAF-V600E-expressing G361 and DM738 melanoma cells as well as wild-type BRAF-expressing SK-MEL-2 melanoma cells were treated with or without 20 μM 1-benzyl-I3C over a 72 time course. The protein levels of the Wnt co-receptor LRP6 and phosphorylated LRP6 (phos-LRP6), the β-catenin destruction complex components Axin and GSK-3β, the transcription factors LEF-1 and β-catenin as well as of the control HSP90 protein were determined by western blot analysis of electrophoretically fractionated total cell extracts. (B) Athymic nude mice with G361 cellderived tumor xenografts were injected subcutaneously with 20 mg/kg body weight 1-benzyl-I3C, 200 mg/kg body weight I3C or with DMSO vehicle control, and the resulting tumor volumes were determined as described in Methods section of the supplemental information. The micrographs show tumors harvested after 18 days of treatment. (C) Ten-micrometer cryostat sections of tumors harvested from 1-benzyl-I3C and DMSO vehicle control-treated animals were analysed for the indicated components of the Wnt signaling pathway by immunofluorescence staining. The results are 63× magnified representative images from three independent experiments.

To extend our observations in an in vivo context, G361 melanoma cells were injected into the flanks of immunocompromised athymic nude mice to generate palpable tumor xenografts. The mice were subcutaneously injected in the scruff with either 20 mg/kg body weight 1-benzyl-I3C or with the dimethyl sulfoxide (DMSO) vehicle control everyday through out a 4 week time course and tumor volumes calculated as described in the methods section. The subcutaneous injection route was used to bypass the acid conditions of the stomach and thereby prevent the self-condensation of I3C into its bioactive products. In contrast to the natural indolecarbinol I3C, 1-benzyl-I3C does not self-condense and remains relatively stable (50). As shown in Figure 3B (upper panel), 1-benzyl-I3C strongly attenuated growth of the tumor xenografts compared with the vehicle control that was noticeable within the first week of injections and sustained over the entire time course. The in vivo results with 1-benzyl-I3C were compared with mice injected with 200 mg/kg body weight of I3C, which we have previously reported (43). In mice not injected with any melanoma cells, treatment with either 1-benzyl-I3C or I3C did not cause any detectable weight loss of the mice (data not shown). At termination of the treatments, the size of the residual tumors treated with 1-benzyl-I3C was significantly smaller than the vehicle control and comparable with tumors treated with a 10-fold higher concentration of I3C (Figure 3B, lower panel). The in vivo effects of 1-benzyl-I3C on Wnt signaling pathway components were assessed by immunofluorescence using tumor xenograft tissue sections isolated from mice injected with either 1-benzyl-I3C or with the vehicle control. Consistent with the cell culture results, 1-benzyl-I3C strongly downregulated the tumor levels of expressed LRP6, β-catenin, LEF-1 and MITF-M and upregulated the levels of the β-catenin destruction complex components GSK-3β and Axin (Figure 3C).

1-Benzyl-I3C downregulation of Wnt/β-catenin signaling is linked to the loss of MITF-M expression

The Wnt-induced nuclear localization of β-catenin stimulates MITF-M expression in melanoma cells (17), suggesting that the 1-benzyl-I3C induced downregulation of β-catenin and MITF-M may be linked mechanistically. In a pharmacological approach to test this possibility, oncogenic BRAF-expressing G361 melanoma cells and wild-type BRAF-expressing SK-MEL-2 melanoma cells were pre-treated with Lithium Chloride (LiCl), a non-selective inhibitor of GSK-3β, or with the GSK-3β-specific inhibitor 6-bromoindirubin-3-oxime (BIO), for 3 h prior to treatment with or without 1-benzyl-I3C for 24 h. Because GSK-3β phosphorylation of β-catenin signals β-catenin ubiquitination and degradation, the inhibition of GSK-3β should enhance β-catenin protein levels when Wnt signaling is absent or attenuated. Western blots demonstrated the exposure to either of the GSK-3β inhibitors rescued the 1-benzyl-I3C-mediated downregulation of β-catenin protein with a concomitant rescue of the loss of MITF-M protein and of its downstream cell cycle regulator target gene CDK2 and its anti-apoptotic target gene Bcl2 (Figure 4A). This result implicates a functional connection between the 1-benzyl-I3C disruption of Wnt/β-catenin signaling and inhibition of MITF-M expression.

Figure 4.

Role of β-catenin/LEF-1-mediated regulation of MITF-M expression in the 1-benzyl-I3C mechanism of action. (A). BRAF-V600E-expressing G361 and wild-type BRAF-expressing SK-MEL-2 melanoma cells were pre-treated with either the non-specific GSK-3β inhibitor LiCl (left panel) or the specific GSK-3β inhibitor BIO (right panel) for 3 h. The cells were subsequently treated with or without 20 μM 1-benzyl-I3C for 24 h, and the levels of β-catenin, MITF-M, Bcl2 and CDK2 protein determined by western blot analysis of total cell extracts. The results are representative of three independent experiments. (B) ChIP assay was performed on G361 cells treated with or without 20 μM 1-benzyl-I3C for 48 h using LEF-1 antibodies (IP: LEF-1) or the control IgG with one-percent input as the loading control. The bar graphs quantify the densitometry results from three independent experiments ± SEM (*P < 0.01). (C) G361 cells were transfected with reporter plasmids containing either a wild-type MITF-M promoter (WT), a LEF-1 consensus site mutant (LEF-1 Mut), a BRN2 consensus site mutant (BRN2 Mut) or the control PGL2 empty vector. Luciferase-specific activity was measured in cells treated with or without 20 μM 1-benzyl-I3C for 24 h. The bar graph shows the results of three independent trials in triplicate ± SEM (*P < 0.01).

β-Catenin binds with LEF-1/TCF to form a transcription factor complex, including specific co-activators, that can bind to the CTTTGAT consensus site between −199 and −193 bp within the MITF-M promoter. Chromatin immunoprecipitation was used to test whether the 1-benzyl-I3C downregulation of β-catenin and LEF-1 results in the loss of endogenous LEF-1 binding to the MITF-M promoter. G361 melanoma cells were treated with or without 20 µM 1-benzyl-I3C for 48 h and then the sheared and cross-linked chromatin DNA-protein complexes were immunoprecipitated with either anti-LEF-1 or with an IgG control antibodies. PCR analysis using primers specific to the LEF-1-binding site in the MITF-M promoter revealed that treatment with 1-benzyl-I3C significantly disrupted nuclear LEF-1 interactions with the MITF-M promoter (Figure 4B). One-percent input was used as a loading control.

To assess whether 1-benzyl-I3C treatment downregulates MITF-M promoter activity in an LEF-1/TCF-binding site-dependent manner, G361 melanoma cells were first transiently transfected with a wild-type -333/120 MITF-M promoter luciferase reporter plasmid (WT), and luciferase activity was determined in cells treated with or without 20 µM 1-benzyl-I3C for 24 h. As shown in Figure 4C, MITF-M promoter-driven luciferase activity was significantly abrogated in the 1-benzyl-I3C-treated cells compared with the DMSO vehicle-treated control. Mutation of the LEF-1/TCF consensus site (LEF-1 Mut) prevented the 1-benzyl-I3C downregulation of MITF-M promoter activity, which implicates the direct involvement of LEF-1/TCF in the inhibition of MITF-M promoter activity by 1-benzyl-I3C (Figure 4C). In contrast, mutation of the BRN2-binding site, which is a downstream component of oncogenic BRAF signaling (44), had no significant effects on the 1-benzyl-I3C downregulation of MITF-M promoter activity (Figure 4C).

Disruption of Wnt-regulated LEF-1/TCF transcriptional activity by 1-benzyl-I3C

The TOP FLASH assay was used to functionally assess whether 1-benzyl-I3C acts through the disruption of the Wnt/β-catenin signaling cascade. G361 melanoma cells were cultured in conditioned media harvested from either mouse fibroblast L cells stably transfected with either an empty CMV vector (L-cell-conditioned medium) or from L cells transfected with pCMV-Wnt, which secrete exogenous Wnt (L-cell Wnt-conditioned medium). The G361 cells were transiently transfected with either a Wnt inducible TOP-luciferase reporter construct driven by a minimal promoter and an enhancer region containing multiple tandem repeats of the wild-type TCF/LEF-1 response element or transfected with a negative control FOP construct containing non-inducible mutated TCF/LEF-binding sites. The resulting luciferase activity was determined from cell extracts post-treatment with or without 20 µM 1-benzyl-I3C for 24 h. As shown in Figure 5A, in absence of 1-benzyl-I3C, G361 cells cultured in Wnt-containing L-cell-conditioned media stimulated the LEF-1/TCF-driven transcriptional activity of the TOP reporter plasmid by ~1000-fold compared with cells cultured in control L-cell-conditioned media. Relatively, little luciferase activity was detected in cells transfected with the negative control FOP reporter plasmid, demonstrating the LEF-1/TCF-dependent specificity of the luciferase transcriptional activity in this assay. Treatment with 1-benzyl-I3C strongly downregulated the Wnt-conditioned medium activated luciferase activity generated from the TOP plasmid to levels nearly equivalent to that observed with the FOP negative control plasmids (Figure 5A). These results functionally demonstrate that 1-benzyl-I3C disrupts Wnt-activated LEF-1/TCF transcriptional activity.

Figure 5.

Involvement of canonical Wnt signaling in 1-benzyl-I3C responsiveness of melanoma cells. (A) G361 melanoma cells were transfected with either a LEF-1/β-catenin inducible luciferase reporter construct driven by tandem repeats LEF-1/TCF response elements (TOP construct) or with a luciferase reporter plasmid containing mutated non-inducible LEF-1/TCF response elements (FOP construct) to serve as negative control. For each condition, the cells were cultured in either L-cell Wnt-conditioned media (right panel) or L-cell-conditioned media with no secreted Wnt (left panel). Subsequently, the cells were treated with or without 20μM 1-benzyl-I3C for 24 h, and luciferase activity (RLU) was measured. The bar graph shows the results of three independent trials in triplicate ± SEM (*P < 0.01), and the table inserts show the obtained numerical values. (B) G361 cells were co-transfected with the TOP or FOP reporter plasmids in the presence of either a constitutively active LRP6 expression plasmid (right panel) or with a wild-type LRP6 expression plasmid. The cells were cultured in L-cell Wnt-containing medium and then treated with or without 20 μM 1-benzyl-I3C for 24 h. The luciferase reporter plasmid activity was measured (RLU), and the bar graphs show the results of three independent trials in triplicate ± SEM (*P < 0.01). The table inserts show the obtained numerical values.

A critical component of the canonical Wnt cascade is LRP5/6, the Wnt co-receptor with Frizzled that directs downstream signaling leading to the Wnt-dependent activation of the LEF-1/TCF transcriptional activity. To determine whether 1-benzyl-I3C disrupts signaling relatively upstream in the Wnt pathway, G361 melanoma cells were co-transfected with either a constitutively active LRP6 expression vector or a wild-type LRP6 expression vector, along with the TOP-luciferase reporter plasmid or the negative control FOP-luciferase reporter plasmid. Luciferase activity was determined in cell extracts isolated from cells treated with or without 20 μM 1-benzyl-I3C for 24 h. As shown in figure 5B (left panel), in cells transfected with the wild-type LRP6 vector, 1-benzyl-I3C significantly inhibited LEF-1/TCF transcriptional activity compared with vehicle control-treated (or untreated) cells. However, transfection with the constitutively active LRP6 expression vector prevented the 1-benzyl-I3C inhibition of LEF-1/TCF-driven TOP-luciferase activity (Figure 5B, right panel). The near baseline luciferase values observed with the FOP-luciferase plasmid confirm that the detected luciferase activity is specific for LEF-1/TCF transcriptional activity. We therefore propose that 1-benzyl-I3C disrupts Wnt signaling at the level of Wnt co-receptors or perhaps Wnt itself.

Combined anti-proliferative effects of the synthetic indolecarbinol 1-benzyl-I3C and the BRAF inhibitor Vemurafenib

Vemurafenib selectively targets the oncogenic BRAF pathway, and our current results show that 1-benzyl-I3C targets the Wnt/β-catenin signaling pathway. We therefore hypothesized that a combination of the two compounds might be highly effective in the ability to inhibit the proliferation of human melanoma cells. To test this possibility, oncogenic BRAF-expressing G361 and DM738 melanoma cells, as well as wild-type BRAF-expressing SK-MEL-2 melanoma cells, were treated with various combinations of different concentrations of 1-benzyl-I3C and/or Vemurafenib, and a CCK-8 assay used to monitor the effects on cell proliferation. As shown in Figure 6A, treatment of either oncogenic BRAF-expressing cell lines with combinations of 1-benzyl-I3C and Vemurafenib induced an anti-proliferative response that resulted from the simultaneous disruption of both the Wnt/β-catenin and the oncogenic BRAF signaling pathways. When wild-type BRAF-expressing SK-MEL-2 cells are similarly treated, the effects of a combination of 1-benzyl-I3C and Vemurafenib reflected that of 1-benzyl-I3C alone because of the ineffectiveness of BRAF inhibitors in this melanoma cell genotype. In these experiments, the cells were treated for 24 h because treatment with combinations of 1-benzyl-I3C and Vemurafenib for 48 h or longer caused significant apoptosis.

Figure 6.

Effects of combinations of 1-benzyl-I3C and Vemurafenib on melanoma cell proliferation, MITF-M protein levels and on the levels of components of the Wnt and oncogenic BRAF signaling pathways. (A) Oncogenic BRAF-expressing G361 and DM738 cells as well as wild-type BRAF-expressing SK-MEL-2 cells were treated with the indicated concentration combinations of 1-benzyl-I3C (1B) and Vemurafenib (V) for a suboptimal time of 24 h, and inhibition of proliferation was monitored using a CCK-8 assay. The results represent the average of three independent experiments with a mean ± SEM (P < 0.01) shown in the bar graphs. (B) Melanoma cells were treated with 20 μM I3C and/or 10 μM Vemurafenib for 24 h. The protein levels of MITF-M, the Wnt signaling components β-catenin and GSK-3β, the oncogenic BRAF signaling components MEK-p and MEK1/2, as well as the control HSP90 were determined by western blots of total cell extracts.

The levels of Wnt signaling components, phosphorylated MEK (MEK-p), which is downstream of oncogenic BRAF signaling, and MITF-M were examined in oncogenic BRAF-expressing G361 and DM738 cells and in wild-type BRAF-expressing SK-MEL-2 cells treated for 24 h with or without 20 µM 1-benzyl-I3C in the presence or absence of 10 µM Vemurafenib. The protein levels of MITF-M, MEK-p, β-catenin and GSK-3β were analysed by western blot analysis. In both G361 and DM738 cells, the combination of 1-benzyl-I3C and Vemurafenib downregulated MITF-M protein levels by greater than 90% (densitometry data not shown) in a manner that reflected the loss of both the Wnt/β-catenin and the oncogenic BRAF signaling pathways (Fig. 6B). 1-Benzyl-I3C, but not Vemurafenib, upregulated GSK-3β levels and downregulated β-catenin levels, and in a complementary manner, Vemurafenib, but not 1-benzyl-I3C, downregulated phospho-MEK levels. In SK-MEL-2 cells treated with both 1-benzyl-I3C and Vemurafenib, the downregulation of MITF-M protein was due to the disruption of Wnt signaling by 1-benzyl-I3C with the cells remaining insensitive to Vemurafenib. Overall, the combinational anti-proliferative effects of 1-benzyl-I3C and Vemurafenib mirrored the downregulation of MITF-M levels in each of the tested melanoma cell lines.

Discussion

Only a limited number of effective small molecule inhibitors of Wnt/β-catenin signaling has been developed to date (52). Our study establishes that 1-benzyl-I3C is a new indole-based inhibitor of Wnt/β-catenin signaling in human melanoma cells that strongly downregulates expression of the master melanoma regulator MITF-M. In both cultured cells and in vivo in melanoma cell-derived tumor xenografts, 1-benzyl-I3C disrupted canonical Wnt signaling that resulted in the downregulation of β-catenin and LEF-1 protein levels and increased levels of the GSK-3β and Axin, which function in the β-catenin destruction complex. In contrast to I3C, in which anti-proliferative effectiveness was restricted to oncogenic BRAF-expressing melanoma cells, 1-benzyl-I3C displayed a significantly more versatile anti-proliferative response profile encompassing cells lines expressing an oncogenic or wild-type BRAF. Thus, compared with the natural indolecarbinol I3C, the synthetic analogue 1-benzyl-I3C displays an increased potency as well as an expanded set of melanoma cell genotypes that are sensitive to the anti-proliferative effects of the synthetic compound. Our results suggest that 1-benzyl-I3C has more versatile anti-cancer properties than I3C because of its ability to disrupt Wnt/β-catenin signaling, a pathway that is active in human melanomas (2–6). Using in vitro biochemical, in silico structural and cellular strategies that we employed to establish that 1-benzyl-I3C directly binds to and strongly inhibits the enzymatic activity of the ubiquitin ligase NEDD4-1 (53), we are attempting to identify the precise 1-benzyl-I3C target protein within the Wnt/β-catenin cascade that triggers the loss of downstream signaling. In this regard, expression of a constitutively active form of the LRP6 Wnt co-receptor rescued the 1-benzyl-I3C downregulation of LEF-1 transcriptional activity in melanoma cells, suggesting that the direct target of 1-benzyl-I3C is at or upstream of LRP6 function, including potentially Wnt itself.

Wnt/β-catenin signaling is aberrantly activated in one-third of melanomas with this subset showing poor prognosis (16, 51). Signaling through this pathway has been shown to maintain high levels of MITF-M expression, which in turn plays a central role in melanoma cell cycle progression and the overall proliferative state (23). Cellular variations of MITF-M levels have been linked to the efficiency of and resistance to melanoma therapeutic regimes (24, 26, 54), suggesting that increases in MITF-M levels may be mediating the Wnt-dependent therapeutic escape in melanoma (14, 15, 55). This hypothesis is underscored by our observation that the 1-benzyl-I3C-induced anti-proliferative effects and disruption of Wnt signaling in melanoma cells results in the loss of endogenous LEF-1 binding to the MITF-M promoter. As a result, MITF-M promoter activity is downregulated, accounting for reduced MITF-M transcript and protein levels as well as subsequent loss of expression of MITF-M target genes such as Bcl2 and CDK4.

We also observed that 1-benzyl-I3C inhibited proliferation of melanoma cells independent of the expression of oncogenic BRAF, most probably due to functioning of the Wnt/β-catenin signaling pathway in a wider range of melanoma cell lines with different mutational profiles. In oncogenic BRAF-expressing melanoma cell lines, combinations of 1-benzyl-I3C and Vemurafenib displayed a strong anti-proliferative response. The combined anti-proliferative effects of 1-benzyl-I3C and Vemurafenib converged at and mirrored the downregulation of MITF-M levels, and optimal combinations of each compound nearly ablated expression of MITF-M. Prolonged treatment with BRAF pathway inhibitors can lead to the emergence of acquired drug resistance, and among many reported mechanisms of post-treatment disease progression is an overexpression of MITF-M (24). In another study, MITF-M expression was found to be sufficient to render melanoma cells resistant to MEK or extracellular signal-regulated kinase (ERK) inhibitors (25, 26). Hence, the combination of a BRAF inhibitor and a Wnt signaling inhibitor could prove to be highly effective strategy for anti-melanoma combinational therapy that can potentially delay the emergence of acquired drug resistance because of the simultaneous disruption of two critical proliferative cascades.

We observed that 1-benzyl-I3C displays a highly potent anti-proliferative effect that disrupts Wnt/β-catenin signaling in a wide range of melanoma cell genotypes. This synthetic indolecarbinol compound does not alter the proliferation of normal epidermal melanocytes, and does not cause any detectable weight loss of injected athymic mice, although an important caveat is that a more detailed analysis of the in vivo effects 1-benzyl-I3C needs to be undertaken in order to establish the in vivo tolerance of mice to treatment with 1-benzyl-I3C. In wild-type BRAF-expressing SK-MEL-2 melanoma cells, 1-benzyl-I3C strongly inhibited proliferation and downregulated MITF-M. Current therapeutic options for wild-type BRAF-expressing melanoma patients are limited by the heterogeneity of the disease and the lack of identified driver mutations (13). Inhibition of MEK alone has been reported to be insufficient to treat wild-type BRAF-expressing melanomas (55), although one study in a mouse melanoma model showed that treatment with MEK inhibitors induced apoptosis but did not trigger a cell cycle arrest (56). Wild-type BRAF-expressing melanoma can display hyperactivated Wnt signaling along with upregulated β-catenin expression, suggesting the potential use of Wnt signaling inhibitors as treatment options for this type of melanoma. Wnt signaling is also aberrantly activated in many other types of human malignancies as well as in cancer stem cells (57–59), which has spurred clinical trials testing the effectiveness of Wnt signaling inhibitors, although not currently in melanomas (60). In this regard, because 1-benzyl-I3C disrupts Wnt/β-catenin signaling, conceivably this indolecarbinol analogue could potentially alter a variety of Wnt/β-catenin-dependent processes, and an important future direction will complete a detailed pharmacokinetic study prior to any clinical studies. We propose that 1-benzyl-I3C is an intriguing new and highly potent Wnt signaling inhibitor that can conceivably be developed as a mono-therapeutic molecule or potentially combined with other oncogenic pathway inhibitors to bring about a more robust anti-proliferative response in melanoma cells as well as other human cancer types.

Supplementary material

Supplementary Figures 1 and 2 can be found at Carcinogenesis online.

Funding

National Institutes of Health Public Service grant CA164095 awarded from the National Cancer Institute.

Abbreviations

BRN2

brain 2

CDK

Cyclin Dependent Kinase

CMV

cytomegalovirus

GSK-3β

glycogen synthase kinase-3β

I3C

indole-3-carbinol

IgG

Immunoglobulin G

LEF/TCF

lymphoid enhancer factor/T-cell transcription factor

LRP6

low-density lipoprotein receptor-related protein

MEK

Mitogen activated kinase/Extracellular regulated kinase Kinase

MITF-M

microphthalmia-associated transcription factor isoform-M

PTEN

Phosphatase, tensin homolog deleted on chromosome ten

TrCP

Transducin repeat-containing protein