ERK1/2 regulation of CD44 modulates oral cancer aggressiveness

Abstract

Carcinogen-induced oral cavity squamous cell carcinoma (OSCC) incurs significant morbidity and mortality and constitutes a global health challenge. To gain further insight into this disease, we generated cell line models from 7,12-dimethylbenz(a)anthracene-induced murine primary OSCC capable of tumor formation upon transplantation into immunocompetent wild-type mice. Whereas several cell lines grew rapidly and were capable of metastasis, some grew slowly and did not metastasize. Aggressively growing cell lines displayed ERK1/2 activation, which stimulated expression of CD44, a marker associated with epithelial to mesenchymal transition and putative cancer stem cells. MEK (MAP/ERK kinase) inhibition upstream of ERK1/2 decreased CD44 expression and promoter activity and reduced cell migration and invasion. Conversely, MEK1 activation enhanced CD44 expression and promoter activity, whereas CD44 attenuation reduced in vitro migration and in vivo tumor formation. Extending these findings to freshly resected human OSCC, we confirmed a strict relationship between ERK1/2 phosphorylation and CD44 expression. In summary, our findings identify CD44 as a critical target of ERK1/2 in promoting tumor aggressiveness and offer a preclinical proof-of-concept to target this pathway as a strategy to treat head and neck cancer.

Conflict of interest statement

Disclosure: No conflict of interest

©2011 AACR.

Figures

Figure 1. Variant growth patterns and lymphatic metastasis in a C57BL/6 syngeneic model of OSCC

A. Representative primary OSCC in left floor of mouth/buccal region after 25 weeks of biweekly DMBA treatment (MOC10 parent tumor). B. Representative H&E stained section of primary tumor (MOC10 parent tumor) showing moderately differentiated squamous cell carcinoma (40X magnification). C. Immunofluorescence of MOC10 cell line showing epithelial phenotype with positive cytokeratin staining (green) and DAPI for nuclear staining (40X magnification). D. Representative in vivo growth curves comparing indolent MOC1 and aggressive MOC2 after 1×106 cells were injected into the right flank of C57BL/6 WT mice (**=p<0.01 in all days post day 0). E. Metastatic inguinal draining lymph node (black filled arrow) that is enlarged and discolored compared to contralateral normal appearing lymph node (black open arrow). F. H&E stained metastatic lymph node shows effacement of normal architecture by SCC (LN= lymph node, 20X magnification). G. Summary of LN metastatic capacity of each cell line. (*MOC23 data is shown for RAG2−/− mice, as this line does not form progressive tumors in WT mice)

Figure 2. Increased activation of ERK1/2 is associated with increased OSCC aggressiveness

A. Western blots of MOC1/2/7/10 for phospho-NFκB, NFκB, phospho-AKT, AKT, EGFR, TGFβIIR, and β-actin. STAT3 and phospho-STAT3 were visualized by immunoprecipation and Western blotting. Note in the experiment shown MOC7 had less total STAT3 immunoprecipitated, but in repeat experiments it was found to express similar amounts of STAT3 when compared with the other cell lines (data not shown). B. Western blot of p-ERK1/2 and ERK1/2 for all 6 MOC lines. C. Representative 20× microscopic image of a Transwell migration assay of MOC2 cells treated with vehicle control (DMSO) or U0126 (10μM). D. Quantitation of Transwell migration assay where 4 random sections per filter × 3 filters were counted in a blinded fashion by light microscopy at 20× magnification for both MOC1 and MOC2 cells treated with vehicle or U0126 (***=p<0.001). The percentage decrease relative to vehicle treatment is indicated above the bar graph.

Figure 3. CD44 is associated with and contributes to increased OSCC aggressiveness

A. FACS analysis of cell surface CD44 expression by pan-CD44 antibody (1M7) and isotype control shown for MOC1, 2 and 10. B. Representative mean fluorescence intensity for cell surface CD44 expression in all 6 cell lines (from one of at least three experiments). C. FACS analysis of cell surface CD44 expression in MOC10 after shRNA knockdown of CD44 with 3 distinct shRNAs (CD44-6, CD44-7, CD44-10) or scramble control shRNA (quantitated by indicated MFI). D. Scratch Test of MOC10 after transduction with indicated CD44 or scramble shRNAs. E. CD44 shRNA knockdown leads to delayed or abrogated growth of MOC10. MOC10 cells (1×104) transduced with scramble, CD44-7, or CD44-10 shRNA were injected into the right flank of C57BL/6 WT mice and monitored. Growth curves only represent average tumor diameter of transplanted tumors that grew out as indicated (***=p<0.001).

Figure 4. ERK1/2 activation regulates CD44 activity in MOC cells

A. Dual immunofluorescence of p-ERK1/2 (green) and CD44 (red) was performed on paraffin-embedded sections of transplanted MOC1 and MOC2 tumors. All cell lines have nuclear staining with DAPI (blue). Images are representative of at least 3 sections with the same exact settings for both tumors (40X magnification). Scale bar in MOC1 represents 100 μM—the same scale applies to all images. B. FACS analysis of cell surface CD44 or CD24 expression on MOC2 after 48 hours of treatment with vehicle control (gray line) or U0126 (10μM, black line). The isotype control for each analysis is indicated by gray shaded curve. MFIs are as indicated. C. FACS analysis of cell surface CD44 expression after MOC1 was transduced with tamoxifen regulated MEK1/R4F and treated with (black line) or without (gray line) 200nM tamoxifen for 48 hours. Isotype control is indicated by the gray shaded curve and MFIs are as indicated. D. MOC1, MOC2, or MOC10 cells were co-transfected with indicated CD44 promoter-luciferase and Renilla plasmids in triplicate. CD44 luciferase plasmids utilized were the basal CD44 promoter (CD44 -97+109) (gray), the full length CD44 promoter (CD44 -1262/+109) (white) or the full length CD44 promoter with a mutated AP-1 site (CD44 -1262/+109 AP-1M) (black). Luciferase activity was normalized to Renilla activity to control for transfection efficiency (***=p<0.001 for basal or mutant AP-1 construct vs. full length CD44 promoter). E. MOC1, 2, and 10 were co-transfected with CD44 full length promoter luciferase (CD44 -1262/+109) and Renilla plasmids. Cells in triplicate were then treated with vehicle (DMSO) or U0126 (10μM) for 24 hours (***=p<0.001 for DMSO vs. U0126 treatment only). F. MOC1 cells were co-transfected with tamoxifen regulated MEK1/R4F, CD44 full length promoter luciferase (CD44 -1262/+109), and Renilla plasmids. Cells in triplicate were treated with or without 200nM tamoxifen (***=p<0.001 for vehicle vs. tamoxifen treatment only). All data are representative of at least 2 independent experiments.

Figure 5. Human OSCC also display a p-ERK1/2 and CD44 relationship

A. Western analysis of pERK1/2 and ERK1/2 of the indicated human OSCC lines. B. FACS analysis of cell surface CD44 expression on human OSCC lines with MFI as indicated. Isotype control (from UPCI: SCC029B) is represented by the gray shaded curve. C. FACS analysis of UPCI:SCC068 cell surface CD44 expression after cells were treated with vehicle control (gray line) or U0126 (10μM, black line) for 48 hours. The gray shaded curve represents isotype control. D. Freshly resected primary human OSCC also display a relationship between p-ERK1/2 and CD44. Primary tumor was digested, hematopoietic cells were excluded, and tumor cells were then stained for CD44 and intracellular p-ERK1/2. Dot plots show control (left) and enhanced p-ERK1/2 staining (right) in the CD44high vs CD44low S-SCC-7133 tumor cells. E. MFI of p-ERK1/2 expression in CD44high and CD44low tumor populations from three separate primary human tumors.