Stopping Cancer Cell Invasion
By: Carol A. Rouzer, VICB Communications
Published: February 25, 2011
Cancer cells often lack E-cadherin, a protein that “glues” normal cells together. The discovery of new small molecules that restore E-cadherin expression reveals a relationship between E-cadherin levels and cancer cell invasiveness.
A hallmark of most cancer cells is the loss of normal cell-to-cell interactions. In contrast to the well-ordered arrangement of cells in healthy tissues, cancer cells fail to make contact with neighboring cells, and they grow in a seemingly haphazard fashion. This absence of systematic intercellular relationships contributes to a cancer cell’s ability to invade normal tissues and metastasize to distant parts of the body. Thus, Vanderbilt Institute of Chemical Biology (VICB) investigators Craig Lindsley, Dave Weaver, and Alex Waterson joined with Dan Beauchamp’s lab in the Vanderbilt Ingram Cancer Center (VICC) to search for small molecules that restore normal cell-to-cell adhesion between cancer cells (Stoops et al.  ACS Chem. Biol. published online Feb. 9, DOI: 10.1021/cb100305h).
Figure 1. Fluorescence microscopy of human endothelium highlighting cadherin (green) between cells. Image courtesy of Wikimedia Commons under the GNU Free Documentation License.
For many types of cells, a key component of the “glue” that holds cells together is the class of proteins called cadherins (Figure 1). In epithelial cells, E-cadherin exists as a dimer of two long identical protein chains that extend across the cell membrane. On the exterior surface of the membrane, the proteins form an interdigitating network between cells, adhering them together. Inside the cell, E-cadherin binds to other proteins, including p120, α-catenin, β-catenin, and actin filaments to form an adherence complex. Many cancers exhibit abnormally low expression of E-cadherin. Not only does this reduce the interactions between the cells, but it also results in a failure to form normal intracellular adherence complexes. In the absence of E-cadherin, β-catenin can be transported to the nucleus and serve as a transcription factor to promote cell growth, contributing further to the malignant behavior of the cell (Figure 2).
Figure 2. Figure 2. Diagram of E-cadherin complex. Ecadherin is a dimer of two identical protein chains that pass through the cell membrane. On the outside of the cell, E-cadherin molecules interact with those on adjacent cells, forming an adhesive bond between them. Inside the cell, E-cadherin binds to additional proteins, p120, α-catenin, β-catenin, and actin polymers. In addition to its role in the cell adhesion complex with E-cadherin, β-catenin plays a role in signal transduction pathways that are involved in cell growth and differentiation. Reproduced from Stoops et al.  ACS Chem. Biol. published online Feb. 9, DOI: 10.1021/cb100305h. Copyright American Chemical Society.
Because of its key role in cell-to-cell interactions, the VICB/VICC team of investigators focused on identifying small molecules that normalize E-cadherin expression in tumor cells. Prior work had shown that low levels of E-cadherin expression in many tumors result from excessive methylation of the DNA in the E-cadherin gene or deacetylation of histones in the region of the E-cadherin gene. Molecules that block DNA methylation or histone deacetylation restore E-cadherin expression in these tumors, but this approach is nonspecific and can affect the expression of many genes in addition to the gene for E-cadherin. To identify molecules that act more specifically, the team first established a high-throughput screen (HTS) for E-cadherin expression. This fully automated assay was based on the fluorescence signal generated by Alexa fluor 488-tagged antibody bound to E-cadherin on the surface of human colorectal carcinoma (SW620) cells. For assay development, the investigators fully exploited the resources of the VICB HTS Core Facility, including the F3 robotic arm (Thermo Fisher), which moved 384 well plates from a Cytomat II Incubator (Thermo Fisher) to an ELx405 plate washer (BioTek) and a Multidrop dispenser (Thermo Fisher) (Figure 3), and the Isocyte laser-scanning fluorometer (Blueshift Biotechnologies), which measured the fluorescence signal. The assay allowed the efficient screening of 83,200 compounds from the VICB library, resulting in 30 confirmed hits
Figure 3. The F3 robotic arm in the VICB HTS core laboratory.
Follow-up studies on the initial hits led the team to select compound 1 (Figure 4) for optimization efforts. A parallel synthesis approach provided a library of compounds with varying substituents on the amino side of the amide bond. Then, based on structure-activity relationship data gleaned from these compounds, a matrix library was generated that systematically varied substituents on both sides of the molecule. In all, team chemists synthesized over 200 new compounds, many of which displayed equal or better activity than compound 1. The investigators selected four of these active compounds and one inactive one (to serve as a negative control) for further evaluation (Figure 4).
Figure 4. Structures of compounds that increase E-cadherin expression in SW620 cells (EC50 values provided), and an inactive compound used as a negative control.
An “In Cell Western” assay used the HTS Core Facility’s Odessey IR imaging system (LI-COR Biosciences) to obtain concentration-dependence data on the ability of the selected compounds to increase E-cadherin expression in SW620 cells. Further experiments indicated that the compounds were also active in human squamous cell lung carcinoma (H520)cells, with a similar range of potencies. Fluorescence microscopy of SW620 cells treated with the active compounds revealed that newly expressed E-cadherin was localized to the plasma membrane. The cells also exhibited increased levels of β-catenin, that co-localized with E-cadherin, suggesting that complete adherence complexes had been formed. Polymerase chain reaction and more extensive RNA-Seq analysis confirmed, for compounds 1 and 8j, that increased E-cadherin protein levels were the result of increased gene transcription in treated SW620 cells. The RNA-Seq studies also revealed modulation of a variety of additional signaling pathways, many of which are important to the cancer phenotype.
The compounds varied in their ability to suppress cell proliferation, from no effect for compounds 1 and 8g, to a mild inhibition for compounds 14j and 10k, to approximately 50% inhibition for compound 8j. However, all compounds suppressed the invasion of SW620 cells in a matrigel chamber, supporting the hypothesis that E-cadherin expression is inversely correlated with tumor cell invasiveness. Initial attempts to identify the exact mechanism of action of the active molecules, however, were not successful. Screening of a wide range of ion channels, receptors, and proteins known to modulate tumor invasiveness failed to reveal any effect of the compounds at these targets. Treated cells did reveal increased histone acetylation, but a representative test compound failed to directly inhibit any of the eleven histone deacetylase isoforms.
The results indicate the potential value of screening for E-cadherin expression as a tool to identify inhibitors of cancer cell invasiveness. The active compounds provide a starting point for the potential development of new anti-tumor agents with a unique mechanism of action. However, further work will clearly be required to determine the exact molecular target, and to further refine this exciting new class of small molecule probes.