17, Stage 2 = 0.64, Stage 3 = 0.64, Stage 4 = 0.92; p =0.01, 0.002, and NS, respectively). These data suggest that TLR4 protein expression mirrors what we found in the transcriptome data. Tumor stroma, epithelium, and grade TLR4 staining scores were recorded in the tumor stroma and stratified by tumor grade as follows: well-differentiated = 3.91, moderately-differentiated = 3.02, poorly-differentiated = 3.59, undifferentiated = 3.64 (ANOVA comparing all four categories, p = 0.0005). The TLR4 staining score in the tumor epithelium was classified by tumor grade: well-differentiated = 0.57, moderately-differentiated = 0.84, poorly-differentiated = 0.00, or undifferentiated

= 0.23 (ANOVA comparing Smoothened inhibitor all four categories, p = 9.99 × 10−9). Well-differentiated tumors had a higher stroma:epithelium TLR4 staining ratio than moderately-differentiated tumors (6.86 vs 3.59, respectively). Poor- and un-differentiated tumors had modest stromal staining but little to absent epithelial staining. Survival and recurrence A trend toward statistical significance was observed between increased Bucladesine purchase TLR4 stromal staining and decreased OS (p = 0.16) after correcting for both stage and grade. Marginal significance was observed for the relationship describing increased epithelial TLR4 staining and

decreased OS (p = 0.11). No relation between TLR4 expression and time to tumor recurrence was noted. TLR4 staining in polyps Given the small number of interpretable adenomatous tissue cores on the NCI TMA (n = 15), an additional TMA with adenomas and normal controls was stained. Small sample sizes prevented achievement of significance for all endpoints. Mean TLR4 stromal staining scores were lower in adenomatous polyps (n = 14) than normal tissue (n = 12) controls (adenoma 2.29 versus normal 3.5, W = 95, p = 0.58). Mean TLR4 epithelial staining scores were lower in adenomatous polyps than normal tissue controls

(adenoma 0.57 versus normal 0.67, W = 67, p = 0.30). Mean TLR4 stromal and epithelial staining scores among inflammatory polyps (IP) were higher than normal tissue controls (stroma: IP 5.6 vs normal 3.5, p = 0.22 and epithelium: IP 1.8 versus normal 0.67, p = 0.81). These under-powered observations support the expected finding that inflamed polyps would manifest higher TLR4 Casein kinase 1 levels. Increased TLR4 expression in the epithelium and pericryptal myofibroblasts (PCMs) in CRCs Using cytokeratin staining to identify epithelium, we found that TLR4 is over-expressed in a subset of tumors and that the expression increases from normal to adenoma to cancer. We also observed increased TLR4 staining in the cytokeratin-negative stroma. Given the increased stromal staining of TLR4, we wished to check details clarify which cell types comprise the TLR4-positive stroma in CRCs. Clinical insights from hematoxylin sections suggested fibroblasts as the source for this increased intensity.