Although nonsteroidal anti-inflammatory drugs (NSAID), including sulindac, have been used extensively as chemopreventive agents for colorectal cancer, results are not consistent. NSAIDs, most reportedly sulindac, often do not cause a complete regression of adenomas and some patients develop resistance to NSAID treatment. In this study, we evaluated the effect of sulindac on colon tumorigenesis in the Apc(Min/+) mouse model. Sulindac (180 ppm) given in drinking water for 9 weeks to Apc(Min/+) mice significantly reduced the size of colon tumors, but actually caused an increase in colon tumor multiplicity relative to untreated controls (average of 5.5 versus 1.6 tumors per mouse, respectively; P < 0.0001). This indicated that the drug could inhibit colon tumor progression but not initiation. As expected, in the small intestine, sulindac significantly reduced tumor size and multiplicity relative to untreated controls (average of 2.3 versus 42.0 tumors per mouse, respectively; P < 0.0001). Generation of a panel of prostanoids was comparably suppressed in the small intestine and colon by sulindac treatment. Sulindac is also known to exert its growth inhibitory effects through regulation of many noncyclooxygenase targets, including p21, beta-catenin, E-cadherin, mitochondrial apoptotic proteins, and peroxisome proliferator-activated receptor-gamma. We found that sulindac treatment protected against E-cadherin loss in colon tumors, with associated inhibition of nuclear beta-catenin accumulation. Importantly, p21(WAF1/cip1) and peroxisome proliferator-activated receptor-gamma expression were absent in colon tumors from sulindac-treated mice, suggesting that loss of these proteins is necessary for drug resistance. Together, these observations may be translatable to designing novel clinical therapies using combinations of agents that target multiple molecular pathways to overcome sulindac resistance.