Flat adenomas in right colon progressed more rapidly to invasive carcinoma than polypoid adenomas, selleck whereas protruding adenomas in the left colon progressed more rapidly to invasive carcinoma than flat adenomas.25 IELs and apoptotic granules were found in 95% and 98% of the tumors induced by GLU, a mutagen of glutamic acid, but only in 21% of DMH-induced neoplasias. The presence of IELs and apoptotic granules in GLU-induced tumors, and their absence in most of the DMH tumors, is puzzling. However, GLU neoplasias were induced by

daily doses for 24 months, whereas DMH neoplasias by weekly doses, for up to 6 months. It would appear that “slowly growing” colonic GLU STA-9090 neoplasias often attract IELs leading to apoptosis, whereas “rapidly growing” DMH tumors seldom elicit those reactions.26 and 27 AB-stained sections from the colon of rats with simultaneously growing tumors were quantified in an image analyzer. The AB-positive areas occupied only 35% of the mucosa, both in tumor-bearing rats and in nontumor-bearing DMH-treated rats, suggesting that the decrease of mucin production might be related either to the protracted DMH treatment or to a genuine biochemical

premalignant change in the colonic mucosa.28 After DMH injection to 278 rats, 358 neoplasias developed. Of the 60 colonic adenomas, 25% were flat adenomas, and of the 298 colonic carcinomas, 13% were flat carcinomas originating in flat adenomas, 28% protruding carcinomas, and the remaining 30% lymphoid-associated

carcinomas (originating in GALT). After GLU treatment to 112 rats, 52 polypoid adenomas and 11 polypoid carcinomas evolved; flat neoplasias did not develop.27 Taken together, these animal studies suggest that DMH might be the carcinogen of choice to recreate the human model of carcinogenesis, namely nonpolypoid neoplasias, polypoid neoplasias, and GALT carcinomas. The limitation is that DMH elicits in rats a high percentage of GALT carcinomas, a phenotype that infrequently occurs in the human counterpart.29 Of particular interest is the recent finding of Iishi and colleagues30 in rats; using azoxymethane and pravastatin, an inhibitor of ras p21 isoprenylation, an increased number of flat adenomas was achieved. Because few cases of nonpolypoid Branched chain aminotransferase adenomas are being detected and followed in IBD, it remains unknown whether these adenomas develop before, simultaneously, or after the appearance of histologically detected dysplasias or adenomatoid neoplasias. The design of new animal models might be of help in obtaining this basic information. To Drs Tetsuichiro Muto, Teruyuki Hirota, Yo Kato, Tomo Kitagawa, Kyoichi Nakamura, Haruo Sugano, Shozo Takayama, and Takatoshi Ishikawa, Tokyo, Japan. Their guidance, co-operation, and generosity have permitted to compile this article.