Colorectal tumours have proven to be an excellent system in which to identify and study the genetic alterations involved in the development of a common human neoplasm. A prevalent view is that colorectal tumours appear to arise as the result of multiple genetic alterations in the alleles of both oncogenes and tumour suppressor genes. The accumulation of genetic alterations appears to accompany the clinical and biological progression of the tumours and may determine the phenotype of the tumour cells. In addition to the many somatic alterations identified at various stages of colorectal tumour development, recent studies have led to the identification of the adenomatous polyposis coli (APC) gene, which, when mutated in the germline, predisposes to the development of colorectal tumours. On the basis of studies of inherited and somatic mutations in colorectal tumours, a genetic model for colorectal cancer development has been proposed. Although the model is undoubtedly incomplete, it nevertheless provides a useful framework for further studies of the multiple events that underlie human tumour initiation and progression. Numerous questions remain to be answered, including identification of the normal function of the genes implicated in tumorigenesis, how mutations in these genes arise and are selected for and what the relative contribution of the altered genes is to various stages of the neoplastic process. Nevertheless, an optimistic outlook is that fundamental insights into the pathogenesis of human cancer are within our reach.

The adenovirus early region 1A (E1A) oncogene interferes with the expression level and activity of the AP-1 transcription factor family. E1A abolished the transactivating function of AP-1 (Jun/Fos), which binds to the 12-O-tetradecanoylphorbol-13-acetate-responsive element of the collagenase gene (collTRE). In contrast, the activity of another member of the AP-1 family that binds to the c-junTRE was not repressed. The mRNA expression of the c-jun gene was, in fact, strongly elevated in various cell types expressing the E1A gene of either adenovirus type 5 (Ad5) or Ad12. The regulation of the junB gene by adenovirus E1A, on the other hand, depended both on the cell type and on the transforming adenovirus serotype. The fact that E1A-induced alterations in the repertoire of AP-1 transcription factors depend on its transforming domain in conserved region 1 suggests that the effects are relevant for the transformation process.

The growth factor-inducible cellular genes JE, c-myc and stromelysin (sml) are strongly repressed upon transformation by adenovirus E1A. As E1A proteins are multifunctional and apparently contain distinct domains (conserved regions 1, 2 and 3), each with a specific effect on gene regulation and cell-transformation, we have investigated which of the three conserved regions are responsible for the reduced expression of these genes. To this end, we monitored the expression of the JE, sml and c-myc genes in a panel of normal rat kidney (NRK) cells expressing different mutant E1A genes. Only CR1, and not CR2 or CR3 were found to be essential for the repression of the genes, indicating that CR1, one of the regions essential for cell transformation, represents an autonomous gene regulatory function that can operate in the absence of CR2. We also show that the association of E1A proteins to a 300 kD cellular protein in NRK cells coincides with the ability to repress these genes.

The stromelysin (sml) gene encodes a secreted protease which degrades components of the extracellular matrix. Transformation of NRK49F cells by the E1A region of adenovirus (Ad) type 5 or 12 reduces sml RNA levels, whereas various growth factors or EJras-mediated transformation stimulate sml gene expression in these cells. Nuclear run-on experiments show that AdE1A, growth factors and EJras act on sml gene expression at the level of transcription. Although the sml gene is strongly suppressed in AdE1A-transformed cells, treatment with growth factors or transfection of EJras still causes a raise in sml mRNA levels, indicating that E1A does not block the induction mechanism itself. The effect of glucocorticoid hormones on sml gene expression is very similar to that of AdE1A, in that mRNA levels are lowered without affecting the induction phenomenon. This similarity may provide a clue to the mechanism by which AdE1A represses cellular gene activity.

DNAs from human pancreatic adenocarcinomas were analyzed for the presence of mutations in codons 12, 13 and 61 of the NRAS, KRAS and HRAS gene. Formalin-fixed and paraffin-embedded tissue was used directly in an in vitro amplification reaction to expand the relevant RAS sequences. The mutations were detected by selective hybridization using mutation-specific synthetic oligonucleotides. In 28 of the 30 patients we found a mutation in codon 12 of the KRAS gene. This result confirms the findings of Almoguera et al. [Cell 53 (1988) 549-554] that KRAS mutations occur frequently in adenocarcinomas of the exocrine pancreas. The mutations are predominantly G-T transversions, in contrast to the KRAS mutations in colon tumors which are mainly G-A transitions. Furthermore, in a portion of the tumors the mutation appears to be homozygous.

Primary esophageal squamous cell carcinomas from 41 patients were analyzed for the presence of proto-oncogene alterations associated with this malignancy. The occurrence of activating ras gene mutations in 25 tumors was determined using oligomer hybridization of target sequences amplified by polymerase chain reaction. We found no evidence for mutations in codons 12 and 61 of the H-ras, K-ras, and N-ras genes, nor in codon 13 of the K-ras and N-ras loci in any of these tumors. The apparent absence of activated ras oncogene in esophageal cancers represents a possible exception to the presence of these mutations found consistently in numerous other types of human malignancies, and is in striking contrast to the 40% prevalence of ras mutations in human colorectal cancers. Southern blot hybridization experiments with DNAs from tumors demonstrated amplification of the epidermal growth factor receptor gene (c-erbB) in two of 25 carcinomas. No amplification of the structurally related c-erbB2 (neu) gene was detected. In three out of 12 carcinomas, the level of epidermal growth factor receptor RNA was significantly higher than in normal esophageal mucosal tissue. Our results suggest that enhanced transcription of the epidermal growth factor receptor gene is associated with the development of some esophageal cancers.