Concordant loss of MTAP and p16/CDKN2A expression in gastroesophageal carcinogenesis: evidence of homozygous deletion in esophageal noninvasive precursor lesions and therapeutic implications.

The gene that encodes methylthioadenosine phosphorylase (MTAP), an enzyme involved in adenine and methionine salvage pathways, is located on chromosome 9p21 telomeric to the p16INK4A/CDKN2A tumor suppressor gene. Inactivation of the p16INK4A/CDKN2A gene occurs by three different mechanisms: hypermethylation of the gene promoter, intragenic mutation coupled with loss of the second allele, and homozygous deletion. Immunohistochemical labeling for the p16INK4A/CDKN2A gene product parallels gene status but does not elucidate the mechanism of gene inactivation. Since the MTAP gene is often co-deleted with p16INK4A/CDKN2A, concurrent immunolabeling for both proteins can identify cases with homozygous p16INK4A/CDKN2A gene deletion. MTAP loss itself has therapeutic implications since it may confer selective sensitivity to inhibitors of de novo purine biosynthesis, such as L-alanosine. Twelve tissue microarrays were constructed from 92 cases of Barrett-associated adenocarcinomas and precursor lesions and 112 cases of gastric adenocarcinoma and precursor lesions comprising 1161 individual cores. Multiple cores were arrayed from any given case, and when available, included the entire histologic spectrum of intestinal metaplasia-dysplasia-carcinoma. Tissue microarrays were labeled with monoclonal antibodies against MTAP protein (clone 6.9, Salmedix, Inc) and p16 (clone 16P07, Neomarkers). Complete loss of labeling was considered negative, while any labeling (p16: nuclear; MTAP: cytoplasmic and nuclear) was considered positive. Loss of MTAP labeling occurred exclusively in conjunction with loss of p16 labeling, confirming that the previous findings from this group that concurrent loss of MTAP and p16 labeling is a surrogate marker of 9p21 homozygous deletions. Complete loss of MTAP and p16 was seen in 4 of 25 (16%) patients with Barrett's esophagus, 4 of 18 (22%) with low-grade dysplasia, 5 of 39 (13%) with high-grade dysplasia, 17 of 78 (22%) with invasive adenocarcinoma, and 8 of 36 (22%) of metastases. There were 7 cases of esophageal adenocarcinoma with loss of both MTAP and p16 for which precursor lesions were available. In 6 on these 7 cases (85%), the precursor lesion(s) had loss of both MTAP and p16. Lack of MTAP and p16 expression was seen in 11 of 106 (10%) cases of gastric adenocarcinoma. All metaplastic (30 biopsies from 20 cases) and dysplastic (15 biopsies from 13 cases) gastric tissues had both intact MTAP and p16INK4A/CDKN2A gene products. No precursor lesions were available from the gastric cancers that had loss of both MTAP and p16. Two benign gastric hyperplastic polyps also had intact p16 and MTAP. Concurrent MTAP and p16 loss detected by immunohistochemistry can serve as a convenient surrogate for p16INK4A/CDKN2A gene homozygous deletion in archival tissues. Inactivation of p16INK4A/CDKN2A by homozygous deletion appears to be an early event in Barrett carcinogenesis, occurring in noninvasive precursor lesions, including nondysplastic Barrett mucosa, in subsets of cases. In the absence of MTAP, cells depend exclusively on the de novo synthesis pathway for production of adenosine. This loss of MTAP during 9p21 homozygous deletion might be exploited therapeutically using de novo purine synthesis antimetabolites to treat a subset of invasive gastroesophageal adenocarcinomas and esophageal precursor lesions.

The dichotomy in the preinvasive neoplasia to invasive carcinoma sequence in the pancreas: differential expression of MUC1 and MUC2 supports the existence of two separate pathways of carcinogenesis.

Emerging evidence suggests a dichotomy in the dysplasia-CIS-invasive carcinoma sequence in the pancreas. Pancreatic intraepithelial neoplasms (PanINs; small, incidental duct lesions) progress to invasive ductal adenocarcinomas (5-y survival of < 15%), whereas intraductal papillary mucinous neoplasms (large, intraductal tumors with ductal dilatation) are often associated with colloid carcinoma (5-y survival of > 55%). We explored the relationship of these lesions by examining the expression of MUC1 and MUC2, glycoproteins reportedly reflecting "aggressive" and "indolent" phenotypes in pancreas cancer, respectively. Immunohistochemical labeling with MUC1 (clone Ma695) and MUC2 (clone Ccp58) antibodies was performed on PanINs (n = 43), intraductal papillary mucinous neoplasms (n = 74), ductal adenocarcinomas (n = 136), and colloid carcinomas (n = 15). Fifty-four percent of the intraductal papillary mucinous neoplasms expressed MUC2, whereas none of the PanINs did. In contrast, PanINs, especially higher grade lesions, were often positive for MUC1 (61% of PanIN 3), whereas the expression of this glycoprotein was infrequent in intraductal papillary mucinous neoplasms (20%). This dichotomy was further accentuated in the invasive carcinomas with which these two preinvasive pathways are respectively associated: all colloid carcinomas were MUC2+ (100%) and MUC1- (0%), whereas the labeling pattern was the reverse for ductal adenocarcinomas: 63% were MUC1+ and only 1% were MUC2+. These results support a dichotomy in the dysplasia-CIS sequence in the pancreas. Because these two pathways often lead to different types of invasive carcinomas, this is an invaluable model for the study of carcinogenesis. The findings here also support the previous impression that MUC2 (the mucin associated with gel formation) is a marker of the "indolent" pathway (intraductal papillary mucinous neoplasm and colloid carcinoma), whereas MUC1 (the glycoprotein known to have an inhibitory role in cell-cell and cell-stroma interactions as well as in immunoresistance of tumor cells) is a marker of the "aggressive" pathway (PanIN to ductal adenocarcinoma).