Pancreatic Cancer Tumors

Until recently, pancreatic cancer was a poorly understood disease. Research in the past decade has shown conclusively, however, that pancreatic cancer is primarily genetic in nature. Inactivation with a variety of tumor-suppressor genes such as p16, DPC4, and p53, coupled with activation of oncogenes such as K-ras, are a few of the mutations that trigger the growth of cancerous cells. Understanding these mutations is critical to a better understanding of familial pancreatic cancer and to the development of gene-based screening tests and therapies, In this article, we review the genetic alterations identified in pancreatic cancer and provide examples of how this information can be applied to patient care.

The last 10 years have witnessed an explosive growth in our understanding of pancreatic cancer, and it is now clear that pancreatic cancer is fundamentally a genetic disease. Simply put, pancreatic cancer is a disease caused by inherited and acquired mutations in cancer-related genes. This insight into the nature of pancreatic cancer is already beginning to have a significant clinical impact. Not only has it improved our understanding of the various pathological manifestations of pancreatic cancers, but it has also improved our ability to diagnose pancreatic cancer in small biopsies, helped us understand why pancreatic cancer aggregates in some families, been the basis for the development of new screening tests for the disease, and provided a rational basis for targeted therapeutics.

We will first review the genetic alterations identified to date in adenocarcinomas of the pancreas. Each of the major classes of cancer-associated genes will be reviewed—tumor-suppressor genes, oncogenes, and DNA mismatch repair genes. We will then provide examples of how an understanding of these genetic alterations can be applied to patient care.
TUMOR-SUPPRESSOR GENES

Tumor-suppressor genes are genes that, when inactivated, convey transforming properties. The tumor suppressor genes most frequently targeted in pancreatic cancer include p16, p53, DPC4, BRCA2, MKK4, STK11, ACVR1B, and the TGF-ß receptor genes (Table 1).[2–10] These genes are inactivated either by homozygous deletion (deletion of both copies of the gene), by loss of one allele (one copy of the gene) coupled with an intragenic mutation in the second allele, or by hypermethylation of the gene's promoter.
p16

The p16 (MTS1/INK4A/CDKN2) tumor suppressor gene on chromosome 9p is inactivated in ~95% of pancreatic adenocarcinomas.[ These inactivations are caused by homozygous deletion in 40% of these carcinomas. A further 40% are caused by loss of one allele coupled by an intragenic mutation in the second, and another 15% are caused by hypermethylation of the promoter of the p16 gene. The protein product of the p16 gene, p16, functions to inhibit the promotion of the cell cycle. p16 directly interferes with ATP-binding by the cyclin-dependent kinases 4 and 6 (CDK4 and CDK6) and causes a change that affects the cyclin D binding site, thereby preventing CDK4 and CDK6 from phosphorylating the Rbl protein. In the absence of p16, phosphorylated Rb1 no longer represses the genes and proteins that promote the G1/S transition. The inactivation of the p16 gene thereby deregulates an important cell cycle checkpoint in almost all pancreatic cancers.
p53

The p53 tumor suppressor gene resides on chromosome 17p and is one of the most frequently inactivated genes in human cancers. The p53 gene is inactivated in 50%–75% of pancreatic cancers. In almost all of these cancers, it is inactivated by loss of one allele coupled with an intragenic mutation in the second. The protein product of the p53 gene, p53, serves to activate the transcription of a large number of genes that initiate apoptosis, cause a G1/S cell cycle arrest, and stabilize a G2/M cell cycle checkpoint. Inactivation of the p53 gene in pancreatic cancer therefore deregulates both the induction of cell death and the regulation of the cell cycle.
DPC4

The DPC4 (SMAD4/MADH4) tumor suppressor gene resides on chromosome 18q, and its protein product is believed to play a rule in signal transduction from the transforming growth factor-ß (TGF-ß) superfamily of cell surface receptors.[ Smad proteins interact with these receptors, become phosphorylated, and then complex with DPC4 (Smad4).Activated Smad complexes are then translocated to the nucleus, where their binding to DNA stimulates the transcription of nearby genes. The DPC4 gene is inactivated in ~55% of pancreatic carcinomas. In 35%, this occurs by homozygous deletion, and in the remaining 20%, by loss of one allele coupled with an intragenic mutation in the second allele. Inactivation of the DPC4 gene is relatively specific for pancreatic cancer, as the DPC4 gene is less commonly targeted in other tumor types.
BRCA2

The discovery of the second breast cancer gene, BRCA2, was greatly facilitated by the discovery of a small homozygous deletion on chromosome 13q in a pancreatic carcinoma. The protein product of the BRCA2 gene functions in the repair of double-strand DNA breaks. The BRCA2 gene is inactivated in ~7% of pancreatic carcinomas by a germline intragenic mutation coupled with an acquired somatic loss of the remaining wildtype allele.
Other Tumor Suppressor Genes

A number of tumor suppressor genes have been shown to be inactivated in a minority of pancreatic adenocarcinomas. These include the MKK4 gene on chromosome 17p, the STK11 gene (LKB1, the gene responsible for the Peutz-Jeghers Syndrome) on 19p, the ACVR1B (ALK4) gene on chromosome 12q, the p300 histone acetylase gene, the RB1 gene, and the TGF-ß receptor genes TGF-ßR1 and TGF-ßR2. Although each of these genes is inactivated at a relatively low frequency, these genes are important because, in aggregate, they are altered in the majority of pancreatic cancers.
DNA Methylation

As noted earlier in the discussion on the p16 gene, some tumor suppressor genes are inactivated by hypermethylation of CpG islands. Ueki et al recently studied a series of 45 pancreatic carcinomas, in which they detected aberrant DNA methylation of a least one locus in 60% of the carcinomas. The genes targeted included RAR/ß (methylated in 20%), p16 (methylated in 18%), CACNA1G (16%), TIMP-3 (11%), E-cad ( 75), THßS1 (7%), hMLH1 (4%), and DAP kinase (2%). Of note, simultaneous methylation of at least four loci was observed in 14% of the carcinomas. Ueki et al defined this subgroup of cancers with multiple genes methylated as CpG island-methylator-phenotype positive (CIMP +).
ONCOGENES

Oncogenes are genes that, when activated by mutation or overexpression, possess transforming properties. The K-ras gene on chromosome 12 is the most frequently activated oncogene in pancreatic cancers. The K-ras gene encodes for a protein involved in signal transduction, and point mutations in codons 12, 13, or 61 of the K-ras gene activate the gene product. The K-ras gene is activated by point mutation in ~90% of pancreatic carcinomas, the majority of which are in codon 12 of the gene.

Other oncogenes are activated by gene amplification, such as the AKT2 gene on chromosome 19q, which is amplified in ~10% of pancreatic cancers. The c-myb gene on chromosome 6q is also amplified in a similar number.[ 56] The Epstein Barr Virus (EBV) has oncogenic properties, and Wilentz et al have recently detected EBV in a pancreatic cancer.

DNA MISMATCH REPAIR GENES

The DNA mismatch repair genes, as their name suggests, code for proteins that ensure the fidelity of DNA replication. It is therefore not surprising that DNA replication errors (RER +) are the hallmark of inactivation of a DNA mismatch repair gene. These carcinomas also contain associated mutations in the TGF-ßRII gene. Inactivation of the hMLH1 gene by promoter hypermethylation has also been demonstrated in some RER + pancreatic carcinomas. All reported RER+ pancreatic carcinomas have a distinct “medullary” histologic appearance. Inherited mutations in DNA mismatch repair genes are associated with an increased risk of cancer. Therefore, as will be discussed in greater detail in the article on familial pancreatic cancer (Klein et al, this volume), patients with pancreatic cancer with medullary histology may have inherited mutations in a DNA mismatch repair gene. Recognition of this association can lead to appropriate genetic counseling for those patients and their family members.
MITOCHONDRIAL DNA MUTATIONS

Although most classical cancer-associated genes are coded for by nuclear DNA, Jones et al recently demonstrated that mitochondrial DNA mutations are common in pancreatic carcinomas. They sequenced the complete 16.5 kb mitochondrial genome in 15 pancreatic cancer cell lines and xenografts and demonstrated homoplasmic mitochondrial DNA somatic mutations in nearly all of the cancers. The intracellular mass of mitochondrial DNA in these samples was also increased in the pancreatic cancers relative to normal cells. Mathematical modeling of the evolution of these mutations suggested that many of the mutations might represent the random genetic drift of natural variants towards homoplasmy. Nonetheless, these mitochondrial mutations may be useful targets for screening tests because of the high levels of mitochondrial DNA in tumors.
GENE EXPRESSION

An exciting extension of our understanding of the DNA alterations in invasive pancreatic cancers is the development of new technologies to evaluate gene expression in tissues. Analyses of the patterns of gene expression in pancreatic cancer using these technologies has led to the identification of genes differentially expressed in pancreatic cancer. Finding differentially expressed genes can, in turn, form the basis of a better understanding of the biology of pancreatic cancer, aid in the development of new screening tests for pancreatic cancer, and result in the identification of new targets for novel therapies.
APPLICATIONS

Our improved understanding of the fundamental genetic basis for the development of pancreatic cancer has a number of applications to patient care. These include a more complete understanding of the pathology of the disease, improved diagnosis, a better understanding of why pancreatic cancer aggregates in some families, new screening tests, and a rational basis for selecting therapy.
Molecular Pathology

Molecular genetic analyses have provided new insights into the pathological manifestations of a number of pancreatic neoplasms. One example of the association between microsatellite instability, medullary histology, and familial pancreatic cancer was noted above. Other examples include the “undifferentiated carcinomas with osteoclast-like giant cells.” The prominent osteoclast-like giant cells in these neoplasms originally led pathologists to classify these neoplasms as mesenchymal. Molecular analyses, however, revealed the true epithelial nature of these neoplasms, and the name given to the entity was therefore changed from “osteoclast-like giant cell tumor” to “undifferentiated carcinoma with osteoclastolike giant cells.” Similarly, molecular analyses of high-grade spindle cell neoplasms arising in association with cystic neoplasms helped establish the origin of the high-grade spindle cells from the epithelial cells of the cystic neoplasm. Molecular genetics has therefore helped establish the fundamental direction of differentiation of a number of pancreatic neoplasms.

Perhaps the greatest impact of molecular genetics is on our understanding of the precursor lesions that lead to invasive pancreatic cancer. For decades, microscopic intraductal proliferations were recognized adjacent to some infiltrating pancreatic carcinomas. A number of investigators hypothesized that these intraductal proliferations were noninvasive precursors to invasive cancer, but this hypothesis was difficult to prove in static resection specimens. The genetic alterations in these lesions have been characterized using genetic analysis of microdissected lesions and immunohistochemical labeling of in situ lesions. Not surprisingly, these intraductal proliferations harbor clonal genetic alterations in the same cancer associated genes that are targeted in infiltrating carcinomas of the pancreas. In addition, the frequency of these alterations increases with increasing degrees of histologic atypia in the duct lesions. From these alterations, we can conclude that the ductal proliferations identified adjacent to infiltrating carcinomas are neoplastic (they harbor clonal alterations in cancer-associated genes) and that they are the precursors to invasive pancreatic cancer. Just as there is a progression in the colorectum from adenoma to infiltrating carcinoma, so too is there a progression in the pancreas from duct lesions (most appropriately called “Pancreatic Intraepithelial Neoplasias” (PanIN)) to infiltrating adenocarcinoma.This conclusion will have a significant impact on our approach to pancreatic cancer. It suggests that advanced forms of these early non-invasive pancreatic neoplasms can be detected and removed before they progress to infiltrating cancer.
Diagnosis

Our improved understanding of the fundamental genetic alterations in pancreatic carcinomas has also improved our ability to diagnose pancreatic carcinoma. A variety of conditions including reactive epithelial changes in chronic pancreatitis, can histologically mimic well-differentiated carcinoma. PCR-based gene analyses and in situ colorimetric detection systems for the proteins have been developed, which can now be applied to difficult biopsies. For example, as noted earlier, the DPC4 tumor suppressor gene on chromosome 18q is inactivated in ~55% of pancreatic carcinoma. Wilentz et al have demonstrated that immunolabeling for the DPC4 gene product mirrors DPC4 gene status. That is, pancreatic carcinomas that have inactivated DPC4 genes do not label with antibodies to the DPC4 protein, while normal and reactive tissues, which have intact DPC4, strongly label. The high degree of correlation between gene status and protein level is probably due to a combination of frequent homozygous deletions (that would completely abrogate protein production) and the targeting of mutant DPC4 protein for degradation by the ubiquitin-proteasome system, immunolabeling for DPC4 can be a helpful adjunct to establish a diagnosis of pancreatic carcinoma in that a loss of expression would support the diagnosis of carcinoma.
Families

As will be discussed in greater detail in the review by Klein et al (this volume), our improved understanding of the fundamental genetic basis for the development of pancreatic cancer has also had a significant impact on familial pancreatic cancer. Selected individuals from kindreds in which there has been an aggregation of pancreatic cancer can now be tested for inherited germline mutations in genes known to predispose carriers to pancreatic cancer. Individuals found not to carry a mutation will be relieved of their anxiety, while those found to carry a mutation associated with an increased risk of developing pancreatic cancer may choose careful screening or even prophylactic surgery.
Screening

Techniques have recently been developed to detect rare mutant genes even when these genes are admixed with thousands of normal copies of the gene. The acquired genetic alterations present in a cancer therefore provide new targets for the development of screening tests for that cancer. For example, mutant p53 can be detected in the urine of patients with bladder cancer, and K-ras mutations can be detected in stool samples obtained for patients with colorectal cancer. Similarly, a better understanding of the genetic alterations characteristic of invasive pancreatic cancer may lead to the development of novel screening tests for invasive pancreatic cancer. Extending this concept further, we can predict that a better understanding of the genetic alterations found in PanINs may lead to the development of new tests for preinvasive, and therefore curable, pancreatic neoplasia. For example, Wilentz et al have demonstrated that activating point mutations in codon 12 of the K-ras oncogene can be detected in the duodenal fluid of patients with pancreatic cancer, and Caldas et al have detected mutant K-ras genes shed from invasive pancreatic cancers and from PanIN lesions in the stool.

Treatment

A better understanding of the functions of cancer-associated genes may also lead to novel therapies and to therapies tailored to specific cancers based on the genetic make-up of the cancer. For example, one of the functions of BRCA2, the protein product of the BRCA2 gene, is to cooperate in the repair of double-stranded DNA breaks. Abbott et al have demonstrated that pancreatic carcinomas with BRCA2 gene mutations are more sensitive to radiation and to the chemotherapeutic agents that induce double-stranded DNA breaks. In addition to improving the rational application of existing drugs to the treatment of pancreatic cancer, an improved understanding of the genetics of pancreatic cancer may, one day, lead to the development of new agents specifically designed to replace a cellular function abrogated by a gene mutation.

The last decade has witnessed a dramatic growth in our understanding of pancreatic cancer. It is now clear that pancreatic cancer is a genetic disease. The genes targeted in pancreatic cancer include K-ras, p16, p53, DPC4, BRCA2, and the DNA mismatch repair genes hMLH1 and hMSH2. This knowledge has already led to improvements in patient care.

Last updated Jan 2/07