Search

The CD133 antigen AC133/1 is a putative marker for enriching certain stem cells and cancer progenitor cells. Despite its wide application, the function and regulation of its expression remains unclear. Given the presence of its transcript in more differentiated cells, epigenetic regulation of CD133 transcripts alone cannot vouch for its utility as a marker. Therefore, understanding the regulation of CD133 protein expression is important for further defining the cellular state of tumor-initiating cells. We used the MAPLE system (Mak et al, Mol Cell Proteomics, 2010) to identify bona fide protein interaction partners of CD133. We present the microtubule-associated histone deacetylase 6 (HDAC6) as a physical interaction partner of CD133. This interaction was validated by co-immunoprecipitation/Western Blot analyses and by a membrane yeast two-hybrid (MYTH) assay. In order to understand the functional consequence of the HDAC6 and CD133 interaction, we used lentiviral short-hairpin RNAs to knockdown HDAC6 in multiple cell types, including the human epithelial colorectal adenocarcinoma lines Caco- 2 and HT-29 and in the retinoblastoma cell line Weri-Rb-1. Interestingly, we observed a significant reduction of CD133 protein expression. This reduction was shown to be dependent on HDAC6 deacetylase activity using a deacetlyase deficient HDAC6 mutant as well as when using the small molecule inhibitor of HDAC6, tubacin. HDAC6 promotes de-acetylation of alpha-tubulin to influence re-cycling of cell surface proteins. Furthermore, when we performed a CD133 MYTH screen against a cDNA library isolated from an adult human brain, we identified the late endosomal- and lysosomal-associated marker CD63 as an interaction partner of CD133. In fact, treatment of Caco-2 cells with tubacin resulted in trafficking of CD133 into endosomes for lysosomal degradation as determined by increased co-localization with CD63. In an attempt to study the affect of HDAC6 knockdown on cell viability, we noticed that CD133 knockdown significantly impacted cell proliferation and clonogenicity of Caco-2 cells. We determined that the cause of this is likely due to increased differentiation as determined by the level of alkaline phosphatase activity and by quantitative realtime PCR for markers of differentiated colon cells. To assess the tumorigenic potential of Caco-2 cells with reduced CD133 expression in vitro we performed soft agar colony-forming assay, which resulted in the loss of anchorage independent growth, a well-known hallmark of cancer. Together, these results suggest that CD133 can exercise a function in establishing and maintaining a primitive cancer cell state and loss or reduction of its expression marks more differentiated cancer cells that may be nontumorigenic. Citation Information: Clin Cancer Res 2010;16(14 Suppl):A11.