c-Cbl mediates the degradation of tumorigenic nuclear β-catenin contributing to the heterogeneity in Wnt activity in colorectal tumors

Despite the loss of Adenomatous Polyposis Coli (APC) in a majority of colorectal cancers (CRC), not all CRCs bear hallmarks of Wnt activation, such as nuclear β-catenin. This underscores the presence of other Wnt regulators that are important to define, given the pathogenic and prognostic roles of nuclear β-catenin in human CRC. Herein, we investigated the effect of Casitas B-lineage lymphoma (c-Cbl) on nuclear β-catenin, which is an oncoprotein upregulated in CRC due to loss-of-function APC or gain-of-function CTNNB1 mutations. Despite mechanistic rationale and recent discoveries of c-Cbl's mutations in solid tumors, little is known about its functional importance in CRC. Our study in a cohort of human CRC patients demonstrated an inverse correlation between nuclear β-catenin and c-Cbl. Further investigation showed that the loss of c-Cbl activity significantly enhanced nuclear β-catenin and CRC tumor growth in cell culture and a mouse xenograft model. c-Cbl interacted with and downregulated β-catenin in a manner that was independent of CTNNB1 or APC mutation status. This study demonstrates a previously unrecognized function of c-Cbl as a negative regulator of CRC.

by centroids [47][48][49]. This algorithm treats each object as having a location in space and uses a heuristic to find centroid seeds for clustering. It finds partitions such that objects within each cluster are as close to each other as possible, and as far from objects in other clusters as possible. The approach requires that one specify the number of clusters to be partitioned and a distance metric to quantify how close two objects are to each other. For our case, we used a squared-Euclidean distance metric defined as where x is the pixel of interest and c is the centroid, and d is the computed distance between the pixel and the centroid. Note that each centroid is the mean of the points in that cluster. For each pixel in the image, the clustering approach returns an index corresponding to a cluster. Using this index, one can separate objects by color. In order to ensure robustness, the clustering procedure using new initial cluster centroid positions was repeated ten times, and the solution with the lowest cluster sums of point-to-centroid distances was selected. Note that since the color information exists in the a*b* space, the objects derived using the clustering approach are pixels with a* and b* values.
Based on general observation of all the original images, the expert identified three basic subregions within each image: nuclei and its neighborhood, luminal area, and the interstitial space.
This identification served as the basis for us to sub-divide each image into three clusters as the first step of color-based segmentation.
For the case of c-Cbl, the k-means algorithm (k=3) was used on each transformed color image in a*b* space and segmented into three clusters. From the three segmented clusters, the expert then identified the cluster that encapsulated the cytoplasmic area within the entire image. A size-based filtering operation was then performed on the identified cluster to eliminate all connected components that have fewer than a threshold level of pixel area. Several threshold values for the pixel areas were explored. After careful consideration, all connected components that have fewer than 5000 pixels were removed from the identified cytoplasmic cluster. The resulting image was verified by the expert as the one comprising a filtered sub-region of c-Cbl content as exemplified by the colored staining within the cytosolic regions of each cell within the tissue.
For the case of β-catenin, the same clustering algorithm was used to first divide the transformed color image in a*b* space into three clusters. As the goal here was to estimate the amount of nuclear β-catenin, separating the nuclei that contained β-catenin from the other nuclei within the interstitial and other tissue regions were needed. Therefore, each of the three clusters derived from the color image in a*b* space was divided into two sub-groups using the same k-means (k=2) clustering approach. The final outcome of this 2-tier clustering approach was 6 non-overlapping images. The expert then identified the images with nuclear β-catenin.
For many cases, one out of six images was identified to contain exclusive nuclear β-catenin but for some cases, there were at least 2 images that were identified to contain nuclear β-Catenin.

Valid area estimation of sub-regions per image:
For both c-Cbl and β-catenin, estimation of the size(s) of the resulting sub-region(s) was performed by computing the fraction of the non-zero pixels within the entire image(s). This resulted in a measure of cytoplasmic c-Cbl and nuclear β-catenin contents, respectively. When more than one image was identified (for nuclear β-catenin), a sum of the computed fractions of the images was considered as the valid area. Germany) were grown with doxycycline 100 mg/ml for 5 days to induce the expression of dnTCF4, as described previously [30].
Chemicals. Sorafenib, Foretinib and Gefitinb all dissolved in DMSO were obtained from LC laboratories. Emetine was obtained from Sigma and MG132 from Calbiochem, EMD Millipore.
Immunoblotting and Immunoprecipitation. Cells were lysed in 50 mM Tris-HCl, pH 7.6, 150 mM NaCl, 30 mM EDTA, 0.5% Triton X-100 with complete protease inhibitor (Roche Applied Science). Immunoblotting and immunoprecipitation were performed, as described previously (5) and all the antibodies were from Cell Signaling (MA, USA), unless specified otherwise.
Immunofluorescence. Cells were grown in chamber slides (Lab-Tek) and fixed and processed as described previously. Alexa 488 goat anti-rabbit and Alexa 647 goat anti-mouse (Molecular Probes, Life Technologies) were used as secondary antibodies. ImageJ was used to generate the profile and the scatter plots, as described previously [16,29]. Cellular Fractionation. Subcellular fractionation was performed using Dounce homogenization, as described previously [29].
Spheroid formation assay. 10,000 CRC cell were seeded in low adhesion plates (Corning ®) for 48 hours in complete growth medium and the spheroid colonies were counted in a blinded manner.

TCF/β-Catenin-responsive Luciferase Reporter Assay. The cells seeded in 6-well plates
stably expressing c-Cbl or c-Cbl-70Z were cotransduced with lentiviral particles of pBARLS or pfuBARLS. After 48 h of transfection, luciferase assays were performed using the Dual-Luciferase Kit (Promega) and normalized using protein content determined by the Bradford assay (Bio-Rad).
Generation of Viral Particles. Retroviral constructs with c-Cbl overexpressing or c-Cbl or β-TrCP shRNA constructs, as described previously [16,17], were cotransfected into HEK293T cells along with packaging, envelope, and reverse transcriptase vectors using Lipofectamine 2000 (Life Technologies) per the manufacturer's instructions. Medium containing active viral particles collected after 48 h was centrifuged and stored at 80 °C. Lentiviral particles of TOPand FOP-Flash were generated similarly by cotransfecting the lentiviral constructs with packaging, envelope, and reverse transcriptase vectors using Lipofectamine 2000 per the manufacturer's instructions. For viral transduction, the cells were seeded at 50-60% confluence.
The cells were treated overnight with the medium containing active viral particles along with hexadimethrine bromide (Sigma), a cationic polymer, to increase the efficiency of infection.
Puromycin (Sigma) selection was initiated after 24 h. The cells were harvested after four days to examine the effect on protein levels.

3
[H] Thymidine incorporation assay. 5,000 CRC cells seeded in 96 well plate were serum starved overnight. The cells were stimulated using DMEM medium containing 5% serum for 24 hours after which they were subjected to 1μCi of 3 [H] Thymidine overnight. The lyzed cells were counted for the radioactivity using the LabLogic 300SL Liquid Scintillation Counter.  Figure 1C. For example, average normalized nuclear β-catenin less than 0.38 was considered low.
#= Average normalized c-Cbl cut-off is based on mean c-Cbl (0.74) shown by the vertical line in Figure 1C. For example, normalized average c-Cbl less than 0.74 was considered low.  for 16 hours were lyzed and immunoprecipitated using β-catenin antibodies and then probed with ubiquitin antibody. Five percent of lysates were probed as inputs. A representative of two independent experiments is shown.

Supplementary Figure 3. c-Cbl regulates CRC proliferation and spheroid formation
A. c-Cbl-70Z, an E3 ligase-deficient and a dominant negative form of c-Cbl increases CRC cell proliferation harboring wild-type β-catenin. RKO stably expressing control or c-Cbl-70Z constructs were serum starved for 24 hours and stimulated with 5% FBS. 3 [H] incorporation assay was performed after 24 h. An average of 6 samples done in duplicates is shown. A Student's t-test was performed. Error bars = SEM.

Supplementary Figure 4. Direct and RTK-mediated regulation of Wnt/ β-catenin by c-Cbl
IC50 of Gefitinib was determined as above and was found to be 6 μM for HCT116 cells.