We used two colorectal cell lines; DLD1 was obtained from ATCC (American Type Culture Collection) and LiM6, a derivative of the LS174T cell line, was a kind gift of Dr Robert S. Bresalier (MDACC) (22). All cell lines were grown in RPMI-1640 medium (Invitrogen, Carlsbad, CA, USA) supplemented with 10% FCS (Invitrogen) and 1% penicillin/streptomycin (Invitrogen) at 37°C in a humidified incubator with 5% CO₂.

Aprepitant (AP), an NK1R antagonist and substance P (SP), an NK1R agonist were purchased from Selleck Chemicals and Sigma-Aldrich and were dissolved in DMSO and 0.1 M acetic acid, respectively.

Cell proliferation was assessed using 1-(4,5-dimethylthiazol-2-yl)-3,5-diphenylformazan (MTT) assay. Ten thousand cells/well were seeded into 96-well plates (NUNC, Langenselbold, Germany). After 24 h, cells were treated with increasing doses of AP for 48 h or with DMSO. To assess cell viability, a final concentration of 0.5 mg/ml (MTT formazan powder, Sigma-Aldrich diluted in PBS) was first added to each well followed by 4-h incubation at 37°C. Finally, a lysis solution of 10% SDS and 1 M HCl was added overnight in each well. For the readout, a multi-scanner microplate reader (Tecan GENios Microplate Reader, Männedorf, Switzerland) was used to measure the absorbance at 595 nm. Each experiment was realized three times and each condition was performed in triplicates.

DLD1 cells were seeded onto 6-well plates at a density of 200,000 cells/well. After 24 h, cells were treated for 24 h with 30 μM AP or DMSO. Adherent cells were trypsinized and pooled together with non-adherent cells, rinsed with PBS and fixed with ice cold ethanol 70% for 2 h minimum. Cells were washed with PBS and stained for 30 min with a staining solution: PBS, 0.1% Triton X-100 (Sigma), 0.2 mg/ml RNAse A (Qiagen, Hilden, Germany) and 0.02 mg/ml propidium iodide (Sigma-Aldrich). Cells were analyzed by BD LSRFortessa cell analyzer (BD Biosciences, Heidelberg, Germany).

Sphere formation medium was prepared as follows: DMEM-F12 with 1% penicillin/streptomycin, 10 ng/ml human recombinant βFGF, 20 ng/ml human recombinant EGF, 1% glutamine and B27 serum-free supplement 1X (all from Invitrogen) subsequently filtered with a 0.475-μm ultra Cruz syringe filter (Santa Cruz Biotechnology, Heidelberg, Germany). Parental cells were trypsinized, rinsed twice with PBS and cell counting was performed to seed 500 cells onto a 96-well ultra-low attachment plate (Corning GmbH, Wiesbaden, Germany), each well containing 100 μl of filtered sphere formation medium. Spheres were cultivated for 8 days and 100 μl of media containing the additives (20 μM AP, 100 ng/ml SP or DMSO) was added every 3 days. Sphere formation ability was determined under a microscope.

Cells were first treated with AP (20 or 40 μM), SP (100 ng/ml) or DMSO for 24 h, washed with ice-cold PBS, denatured with a 1% SDS/β-mercaptoethanol lysis buffer solution and stored at −80°C. The Functional Proteomics RPPA Core Facility at MD Anderson Cancer Center analyzed the samples as previously described (23). In brief, probes were spotted on nitrocellulose slides and 172 different proteins were analyzed. Densitometry of these spots was performed and values normalized for protein loading. The values were then transformed to linear values, which were used for our calculations. The quantification was realized by calculating the ratios of the linear protein intensity values of the treated probes with the linear protein intensity values of their respective control probes.

LiM6 and DLD1 cells were first treated with AP (30 μM) or DMSO for 24 h, washed with ice-cold PBS and lysed with a lysis buffer (0.5% Triton X-100, 1 mM sodium orthovanadate, 1 protease inhibitor cocktail tablet in purified water). The protein concentration was determined by using the Bio-Rad protein assay (Bio-Rad Laboratories, Munich, Germany). For each condition, 20 μg of proteins were loaded on 8–12% Tris-glycine gels (Invitrogen), separated by electrophoresis and electroblotted onto nitrocellulose membranes (Amersham Life Sciences, Buckinghamshire, UK). The membranes were then incubated for 2 h in a blocking solution (5% BSA in TBS, 0.1% Tween-20), followed by an overnight incubation with primary antibodies at a 1:1,000 dilution against β-catenin, LRP5 (low-density lipoprotein receptor-related protein 5), MYC, PARP [poly(ADP-ribose) polymerase] and β-actin (all Cell Signaling Technologies, Danvers, MA, USA). Finally, the blots were washed with TBS-0.1% Tween-20 and incubated for 1 h at room temperature with a peroxidase-conjugated goat anti-rabbit IgG antibody (Dako, Hamburg, Germany) at a dilution of 1:2,000. The detection was realized with an enhanced chemiluminescence reaction (ECL Prime Western blotting detection; Amersham Life Sciences) and β-actin served as a loading control.

DLD1 cells were seeded onto 18-mm diameter cover slips (Thermo Scientific, Braunschweig, Germany) in a 12-well plates format (NUNC) at a density of 75,000 cells/well. Cells were treated with 30 μM AP, 70 nM SP or DMSO for 24 h. Cells were then fixed with 4% paraformaldehyde for 15 min, followed by a permeabilization step for 15 min with 0.15% Triton X-100 in PBS and a blocking step for 30 min with 1% BSA in PBS. The cells were then incubated overnight at 4°C with a rabbit primary antibody against β-catenin (Cell Signaling Technologies) diluted in blocking solution at 1:80. After multiple washing steps, cells were incubated in the dark with a goat anti-rabbit secondary antibody conjugated with Alexa Fluor 488 (Invitrogen) at a 1:200 dilution for 1 h. Finally, cover slips were rinsed three times with PBS and mounted on microscope slides with Vectashield (Vector Laboratories, Burlingame, CA, USA) containing 4,6-diamidino-2-pheylindole (DAPI). The images were taken with a Olympus FluoView™ FV1000 confocal microscope.

Total RNA isolation was realized with TRI-reagent (Sigma-Aldrich) and 2 μg were used for cDNA synthesis using random primers (Roche Diagnostics, Penzberg, Germany) and SuperScript II Reverse Transcriptase (Invitrogen). Each PCR reaction was made with 40 ng of cDNA, 500 nM of the primer pair and SoAdvanced™ Universal SYBR Green Supermix (Bio-Rad Laboratories) for a final volume of 20 μl. Thermal cycling was realized by the Mastercycler EP gradient S (Eppendorf, Hamburg, Germany) and consisted of 40 cycles with denaturation at 95°C for 15 sec, annealing at 55°C for 15 sec and elongation at 72°C for 30 sec. All experimental conditions were assessed in duplicate. The primers used were as follows: CTNNB1 (NM_001904.2) forward, 5′-ACGTCCATGGGTGGGACA-3′; reverse, 5′-CTAGGATGTGAAGGGCTCCG-3′. AXIN2 (NM_004655.3) forward, 5′-TATCCAGTGATGCGCTGACG-3′; reverse, 5′-TGTTTCTTACTGCCCACACGAT-3′. MYC (NM_002467.3) forward, 5′-CACCACCAGCAGCGACTCT-3′; reverse, 5′-CAGACTCTGACCTTTTGCCAGG-3′. FOXM1 (NM_202002.2) forward, 5′-CTCCCGCAGCATCAAGCAA-3′; reverse, 5′-GCCAGGACGCTGATGGTCTC-3′.

To assess luciferase activity we used the Dual-Luciferase Reporter Assay system (Promega, Madison, WI, USA). DLD1 and LiM6 cells were seeded in 24-well plates at a density of 35,000 cells/well for 24 h. For the transfection, we used FuGENE HD reagent (Promega) and αMEM medium. We co-transfected inducible Firefly luciferase expressing SuperTOP or SuperFOP vectors (AddGene plasmids nos. 12456 and 12457) and constitutively Renilla luciferase expressing normalization vector pRL-TK (Promega) at a ratio of 50:1. The latter vector was used as a control of transfection efficiency. Simultaneously, the cells were treated with different concentrations of AP as indicated or with DMSO. After 24 h of incubation total cell lysate was extracted using reporter lysis buffer (Promega). Total extract (5 μl), 25 μl of Luciferase substrate LarII solution (Promega) and 25 μl of Stop and Glow 1X solution (Promega) were used to assess Firefly and Renilla luciferase activities, respectively. Finally, Wnt activity was measured by the ratio of SuperTOP/SuperFOP luciferase activities each normalized to the respective Renilla luciferase activities. The results are displayed as relative light unit (RLU) and each experimental condition was assessed in triplicate.

Results are expressed as the mean ± standard error of the mean (SEM). All statistical comparisons were made with a standard t-test and Mann-Whitney U test using biostatistics software from GraphPad Prism^® (La Jolla, CA, USA). The criterion for significance was p<0.05 and p<0.01 for all comparisons.

We have recently described that upon NK1R blockage, growth of hepatoblastoma cells could be inhibited in vitro and in vivo (19). Here, we treated colon cancer cell lines LiM6 (β-catenin mutation) (24) and DLD1 (APC mutation) (25) with increasing doses of AP for 48 h and then performed standard MTT assays. We observed a dose-dependent inhibition of cell growth in both cell lines (Fig. 1A).

In order to have a better understanding of the downstream molecular mechanisms responsible for these observed effects, we used reverse phase protein array (RPPA) analysis. We treated human CRC cells LiM6 and DLD1 with 20 or 40 μM AP for 24 h and screened for changes in protein expression level of 172 proteins involved in multiple signaling pathways that are typically altered in cancer (23). Values from 20 and 40 μM were put into relation to the control value of each cell line and the pathway molecule (Fig. 1B).

To scrutinize the regulation pattern in the two cell lines, we calculated the mean of the values relative to the control of 20 plus 40 μM conditions for each molecule and displayed the results in Fig. 1C. The cut off values are specified in the figure legend. Only a small number of proteins exhibited an upregulation (red) or downregulation (green) common in the two cell lines. Interestingly, we found that important proteins associated with Wnt signaling were downregulated, including the Wnt target MYC and the β-catenin interacting factor FOXM1 (Fig. 1D). Taken together, these findings suggest inhibition of Wnt signaling through inhibition of NK1R. Of note, in these early experiments β-catenin, one key component of the Wnt signaling pathway, was not downregulated in DLD1 (Fig. 1C and D).

Apart from Wnt signaling, we also found the AKT/mTOR pathway to be inhibited (Fig. 2, right panels) at the level of its downstream proteins phospho-4EBP1/2 and phospho-p70S6K (grey bars) compared to their total forms (black bars). Total AKT was also dose-dependently downregulated upon AP treatment in the two cell lines, whereas no clear trend could be extracted from its phospho-form at Thr308. In contrast, phosphorylation of AKT at Ser473 was upregulated (Fig. 2, left panels) along with its substrate PRAS40 at Thr246, a member and repressor of the TORC1 complex (26). These results indicate that apart from decreased Wnt signaling, inhibition of NK1R might lead to additional downregulation of the AKT/mTOR signaling pathway.

It is known that NK1R antagonists induce apoptosis in human hepatoblastoma cancer cells and other cancers (19,20). Therefore, as a next step, we investigated whether this held true in our setup. First, we screened our data obtained by RPPA for alterations in apoptosis markers. We found that cleaved caspase 7 and phospho-c-Jun (Ser73) were upregulated upon AP treatment (Fig. 3A), two well known mediators of apoptosis (27,28).

In order to characterize apoptosis in more detail, we treated DLD1 cells for 24 h with increasing concentration of AP and performed western blot analysis for PARP which executes the apoptotic process (29). We found a dose-dependent increase of cleaved PARP indicating an activation of apoptosis, whereas SP had no effect (Fig. 3B). Further, Her3 was found down-regulated (Fig. 3A, right panels) and targeting this receptor in colorectal cancer cells has been described to induce a G2 arrest as well as apoptosis (30). We then analyzed the effect of AP on the expression of cell cycle regulators. In our RPPA data, we found cyclin B1 and CHK1 to be downregulated, whereas p53 was slightly upregulated (Fig. 3C). These proteins are known to regulate the G2-M transition (31). To investigate more in depth the growth inhibition observed in DLD1 and to confirm our RPPA data, we treated the cells for 24 h with 30 μM AP and analyzed the cell cycle profile by propidium iodide staining and FACS analysis. As shown in Fig. 3D, AP induced a G2 arrest (33% compared with 26% for DMSO-treated cells, white bar) and an increase of cells in the subG1 (12.7% compared with 1.63% for DMSO-treated cells, black bar) phase indicating either late apoptosis or necrosis.

After analyzing the RPPA results above, we found strong evidence that AP might induce a reduction of Wnt activity. To validate this hypothesis, we then carried out Super TOP/FOP (STF) assays in non-treated cells to assess baseline Wnt activity in DLD1 and LiM6 cells and compared their activity to the pancreatic cancer cell line L3.6pl, a cell line known to express little Wnt. As expected, we found that both colon cancer cell lines displayed high levels of Wnt activity with a ~500-fold increased activity compared to L3.6pl (Fig. 4A).

Next, we treated the colon cancer cells with increasing doses of AP. After a one-time treatment, we discovered a robust inhibition of Wnt activity in both colorectal cell lines (Fig. 4B). These results were corroborated by qRT-PCR data in DLD1. The Wnt target gene MYC was dose-dependently downregulated upon AP treatment along with AXIN2 and FOXM1 (Fig. 4C). Intriguingly, in contrast to the RPPA data, CTNNB1 was also downregulated, although the trend was not as convincing as observed in the other genes. Thus, instead of targeting β-catenin for degradation, AP might induce a decrease of Wnt activity by sequestrating it away from the nucleus resulting in decreased TCF/β-catenin activity, ultimately leading to a lower Wnt target gene expression.

To further investigate this hypothesis and the down-regulation of the β-catenin/Wnt signaling pathway following AP treatment, we cultured cells with 30 μM AP or 70 nM SP for 24 h, stained them for β-catenin and analyzed them with confocal microscopy. At the membrane, β-catenin is required for cell adhesion where it complexes with E-cadherin. Consequently, despite being protected from degradation, β-catenin is not available for transducing a signal to the nucleus in this state (32). We observed a strong accumulation of membrane-bound β-catenin upon AP treatment (Fig. 4D). However, SP stimulation did not seem to affect β-catenin/Wnt signaling when compared to the control.

Finally, we treated the cells with 30 μM AP and performed western blot analysis. We found that the Wnt co-receptor LRP5 as well as the Wnt target MYC had markedly lower expression upon AP treatment compared to the respective control. In the same experiment, we analyzed the expression of β-catenin following AP treatment. Similar to the RPPA data, we did not observe any striking changes of total β-catenin in whole cell lysates of LiM6 or DLD1 (Fig. 4E). Taken together, these data demonstrate that targeting the NK1R in colon cancer cell lines markedly reduces Wnt-signaling.

In a next step, we investigated whether this held true not only in differentiated colon cancer cell lines, but also in colon cancer stem cells (CSCs). We grew colorectal CSC-like cells as described (33) and assessed sphere formation ability (SFA) in both DLD1 and LiM6 (Fig. 4F). Upon AP treatment, we discovered a striking decrease, both in sphere number and size. Interestingly, activation of the SP/NK1R system with recombinant SP significantly increased the SFA of DLD1 CSCs in number and size, whereas in LiM6, we were not able to find differences between SP-stimulated cells and control (Fig. 4F).

Taken together, these experiments show that inhibition of the SP/NK1R system with AP decreased the canonical Wnt signaling in colorectal cancer cells, likely by arresting β-catenin in its membrane-bound localization. Also, inhibition of the SP/NK1R system inhibited the growth of anoikis-resistant CSC-like colorectal spheres.

In order to better understand the inhibitory effect on Wnt signaling, we sought to investigate whether differences exist regarding the response rate to AP treatment within cell populations that are constitutively active regarding the β-catenin/Wnt pathway. Such cells with high constitutive Wnt activation have been described to have higher stemness properties (6). To make our findings comparable to those from Vermeulen et al, we transduced the colorectal cell lines LiM6 and DLD1 with a 7xTCF-eGFP/mCherry (7TGC) lentiviral construct (34) allowing us to separate cells with different Wnt activities by single cell sorting via FACS (Fig. 5A). Wnt activity in cells with the same level of mCherry intensity then correlates with their GFP intensity. After separation, we independently cultured Wnt^high and Wnt^low expressing cells and subsequently treated these cells with increasing doses of AP for 48 h (Fig. 5B). Interestingly, we did not observe a difference in cell survival for any of the doses tested between cells with high or low Wnt activity. These findings suggest that the apoptosis-inducing property of NK1R targeting is maintained even in cells with supposedly increased cancer stemness potential (Wnt^high). Furthermore, our data indicate that in a particular cell line or population, the observed inhibitory effects caused by NK1R antagonism are independent of the initial Wnt baseline activation.