Dissecting the mechanisms of cancer has revealed an ever-increasing number of interconnecting pathways that leaves even those in the field dizzy with the effort to keep up. In theory, it's simple. In an ideal world, cell division, survival, and death are in sync, promoting homeostasis with neither unregulated growth nor inappropriate cell loss. However, the real cellular world is laden with oncogenes and tumor suppressor genes whose products interact in overlapping pathways that, when dysfunctional, can lead to cancer.

Despite tremendous progress in unraveling these networks, researchers have their work cut out for them. Daniel Haber, director of the Massachusetts General Hospital Cancer Center, explains: "We now have tools and approaches that should make progress even faster, but it's hard to be patient." Patience, however, is the prescription to fully understanding the mechanisms behind transformation of normal cells to cancer cells and perversion of normal tissue into malignant tumors. One by one, the genetic pathways are being elucidated.

THE FIRST OF MANY The retinoblastoma (Rb) protein could be called the cell-cycle master switch. Cyclin-dependent kinases (CDKs) phosphorylate Rb, turning on cell proliferation.¹ Otherwise Rb is bound to E2F, a transcription factor, sequestering it from activating genes required for cell proliferation. Like other proteins once thought to have one job, Rb has several; it is implicated in DNA replication, differentiation, DNA repair, cell-cycle checkpoints, and apoptosis. In fact, it may interact with more than 100 different cellular proteins.

Rb was the first tumor suppressor gene to be cloned, although how it acts remains unclear. People with germline mutations develop the retinal tumors that give retinoblastoma (Rb) its name--loss of the second Rb allele is the rate-limiting step. Rb is also mutated in diverse sporadic cancers, including small-cell lung carcinoma and osteosarcoma. Plus, alterations in Rb regulatory proteins, such as p16, cyclin D, and CDK4/6, are common in a spectrum of human tumors. "Rb [is] ... downstream of p16 and cyclinD/CDK4, all of which have been found deregulated in cancer," says Marie Classon, staff scientist at Massachusetts General Hospital's cancer center. Taken together, Rb pathway perturbation is a common and significant event in tumorigenesis.

THE DNA DAMAGE RESPONSE Dysfunction of DNA repair and DNA damage checkpoint pathways is largely responsible for genetic instability, a hallmark of cancer cells. Originally defined as pathways that slow the cell to allow time for repair, DNA damage checkpoints also help prevent chromosomal instability. Recently, mushrooming data expose a nexus of overlapping processes called the DNA damage response pathway (DRP). Rather than merely slowing the cell cycle when DNA is damaged, the DRP is multifunctional, controlling repair, telomere composition and length, transcriptional programs, and in some cases, apoptosis.² "It is now recognized that checkpoints are merely components of a larger DNA damage response pathway that regulates a multifaceted response," says Stephen Elledge, Howard Hughes Medical Institute investigator and Baylor College of Medicine biochemistry professor.

DRP tumor suppressor genes fuel cancer's genetic instability. DRP proteins sense damage, transduce the damage signal, and affect a response. How does the cell sense DNA damage to initiate a response? The recent discovery of ATRIP, which exists in a protein complex with ATR that homes to broken DNA and initiates checkpoint signaling, provides a partial answer. "Several suspected sensors have existed ... but how they work has not been established," says Elledge. ATRIP has yet to be linked to cancer, but it would not be surprising.

Best studied are the damage signal transducers, the protein kinases ATM and ATR. ATM, mutated in patients with ataxia telangiectasia, controls the initial phosphorylation of several key transduction and effector proteins such as p53, and other proteins implicated in DNA repair. In fact, only recently has the role of DRP in controlling DNA repair emerged. For example, DNA damage triggers ATM activation, which in turn phosphorylates a number of substrates including BRCA1 (breast cancer susceptibility protein), Rad55, and Nbs1 (Nijmegen Breakage Syndrome), depending on the DNA lesion. In fact, both ATM and the Nbs1 complex are components of BASC (BRCA1 genome surveillance complex) suggesting that ATM controls some aspects of the repair process as well as its long-held role in DNA damage checkpoints.

The serine/threonine kinases Chk1 and Chk2, both associated with tumorigenesis, are primary downstream targets of ATM. These kinases act like a relay, moving checkpoint signals from ATM and ATR to effector proteins, such as p53. Their roles overlap in regulating the cell's housekeeping functions, such as DNA replication, cell-cycle progression, chromatin remodeling, and apoptosis.³ Germline Chk2 mutations cause some cases of Li-Fraumeni familial cancer syndrome, which produces multiple tumors at a young age, with a predominance of breast cancer and sarcomas.

"Chk2 mutations predispose to cancer not only in the Li- Fraumeni families, but also in some families with hereditary breast, colon, or prostate cancers," says Jiri Bartek of the Danish Cancer Society's Institute of Cancer Biology. Cancers of the breast, vulva, bladder, colon, and ovary, and osteosarcomas and lymphomas, may contain somatic mutations of Chk2. In contrast, Chk1 defects are found in colon, stomach, lung, and endometrial cancers.

Which tumor suppressor is most important? One stands out: p53. BRCA1, BRCA2, and Chk2 have dominated headlines for their roles in breast cancer. But, says Elledge, "so far, it is p53 by a long shot." It still holds the top spot for being mutated in the highest percentage of tumors relative to any other single tumor suppressor, he adds.

LINKING UP Although the fog is lifting, it will be some time before the milieu of signaling cascades becomes clear. Other pathways that influence the DRP are becoming evident and demonstrating that tumor suppressors need not function individually. Tumor suppressors, growth signaling pathways, and oncoproteins are networked. Lindsey Mayo at Case Western Reserve University's Ireland Comprehensive Cancer Center explains, "Gain in function in certain molecules or loss in the activity in other molecules would have dramatic effects on these pathways." One such pathway is the PTEN-Mdm2-p53 tumor suppressor-oncoprotein network.⁴

Mitogens and cytokines promote cell proliferation by binding to receptors and engaging signaling cascades, many of which produce second messengers such as phosphytidylinositol (3,4,5) triphosphate (PtdIns 3,4,5 P3). This compound then recruits proteins such as the serine/threonine kinase Akt, which through several substrates, promotes cell survival. One such substrate is Mdm2. Akt phosphorylation of Mdm2 allows its entry into the nucleus where it targets p53 for degradation.

In comes PTEN. The PTEN tumor suppressor protein is a dual-specificity phosphatase that normally blocks the Akt cytokine signaling pathway. Therefore, PTEN's inhibition of this signaling blocks Mdm2 from entering the nucleus, which ultimately protects p53, allowing cells to respond to damage by apoptosis. When PTEN is mutated, as in some tumor cells, it cannot fully protect p53. This relationship illustrates the interconnected networks of tumor suppressor proteins, growth signaling pathways, and oncoproteins, such as Mdm2. Their codependence is essential for homeostasis.

What is the significance of these pathways to cancer? Amplification of PtdIns 3-kinase or Akt or activating mutations of this signaling cascade are found in ovarian, colon, and pancreatic cancers. Mdm2 is overexpressed in sarcomas, leukemia, breast and prostate cancers and, predicts high-grade, aggressive, drug-resistant malignancies. Point mutations, deletions, and loss of the p53 gene are prevalent in over 50% of human cancers. PTEN is mutated in all major forms of human cancer, including 40-50% of gliomas. However, in some tumors PTEN is unexpressed without evidence of mutation, says Ramone Parsons, associate professor of pathology and medicine at Columbia University's Institute for Cancer Genetics. "Understanding pathological mechanisms of reduced PTEN expression may reveal novel oncogenic pathways of cancer." It's likely that classical oncogenic pathways synergize with unexplored pathways, he adds. "Facing this complexity is likely to lead to a more accurate and richer description of oncogenesis that will ultimately benefit cancer patients." The entire lot of human cancers may be associated with dysfunctional components of these tumor suppressor-oncogene networks. Now, to untangle them.

THE INK4A/ARF LOCUS-ANOTHER LINK Two additional tumor suppressors, p14ARF (alternative reading frame) and p16INK4a (p16), alternate products of the INK4a locus, also influence p53 and Rb pathways.⁵ ARF increases p53 expression by blocking Mdm2. Then p16 is linked to the Rb pathway by inhibiting CDKs, therefore inactivating Rb, leading to expression of genes whose products are involved in DNA replication. Recent p16 studies indicate that it is a significant player in cancer inhibition.

Homozygous deletion of the INK4A/ARF locus is common in brain, head and neck, and other types of primary tumors. Promoter methylation blocks p16 transcription in many primary tumors of the colon and other organs. And p16 point mutations are found in primary, sporadic esophageal and pancreatic carcinomas. Mutations in p16 segregate with disease in families with melanoma. Cyclin D, the activating unit of Cdk4/6, is also overexpressed in many malignancies, another indication of the significance of the p16/Rb pathways in tumor suppression.

ARF's role as a tumor suppressor stems mostly from work in INK4a/ARF knockout mice. Surprisingly, these mice express p16, but not ARF, and they succumb to sarcomas and lymphomas in their first year. Researchers have found more p16 mutations in tumors than ARF mutations; however, methylation of the ARF promoter may be another mechanism of blocking ARF. ARF promoter methylation is found in some human tumors independent of p16 promoter methylation. Perhaps ARF is the dominant tumor suppressor in the mouse, whereas p16 plays a major tumor suppressor role in people. "The p16 selective knockout mice ... will be key reagents," says Greg Enders of the University of Pennsylvania. "The limitations will be the extent to which mouse cancer models can faithfully reproduce human cancer types and whether the signaling to p16 and ARF in the mouse is similar to that in human cells."

It will be important to identify the steps in tumorigenesis at which p16 and ARF intervene, to understand p16 regulation, and to elucidate the impact of these proteins on tumor development, says Enders. "The future lies, in part, in answering these questions, using mouse tumor models and careful comparisons to human tumors and cell lines."

The progress made in uncovering genetic lesions that trigger human cancer has become practical: For example, specific kinase inhibitor therapy for chronic myeloid leukemia is now available, as is expression profiling to highlight tumor-specific genetic make-ups that are histologically indistinguishable. Insights will continue to emerge from concerted high-throughput approaches, from discoveries in individual laboratories, and from unexpected results in completely divergent fields of research. Predictions are difficult to make when the field of clinical oncology will be transformed by basic science research.

However, Haber is hopeful. "There is every reason to be optimistic that the pace will continue to accelerate," despite the twists, turns, and length of the road.

Linda B. Schultz (LBSchultz@Sciwriting.com) is a freelance science writer in Suwanee, Ga.

	p53: A FAMILY'S VIEW To a family with Li-Fraumeni cancer syndrome, p53 is more than a transcription factor that lies at the crossroads of the cell cycle. A germline mutation sets the stage for multiple, early cancers. Patricia Holm's family is one of only about 100 with the condition. Her story: At 25 years old my husband Timothy Whittaker developed liposarcoma, a rare cancer of the fat cells. He suffered greatly and weighed 55 pounds when he died--eight months after diagnosis. The younger of our girls, Jennifer, was born in 1973. She was perfectly healthy throughout her childhood, but as her schooling began she had trouble in her studies. Between her junior and senior high school years, Jennifer developed a headache that sent us to the emergency room three times in four days. A computed tomography (CT) scan taken on the third visit revealed a lemon-sized tumor in her left parietal lobe. The first resection pathology revealed a pleomorphic xanthoastrocytoma. Four months later pathology revealed anaplastic astrocytoma. After nine months of chemotherapy and radiation, the tumor was back; the pathology report: glioblastoma multiforme--a death sentence. In the 27 months from diagnosis to death every approved and unapproved method of treatment was used. We went to four states and two countries seeking a cure. There wasn't one. Exactly 19 years after the death of Timothy, Jennifer died at the age of 20. Doctors assured me that liposarcoma and brain tumors have absolutely no connection. Bad luck? A witch's curse? What was going on? The answer would come much later. In August 1998 our other daughter Kimberly gave birth to Grace. Shortly thereafter Kim felt a thickening in her left breast. Repeated attempts to alarm her ob-gyn failed, as he saw this as normal breast change. In May 2000, Kim felt a lump under her arm that in three weeks became the size of a golf ball. She then had her general medical doctor take a look; he sent her for a mammogram the same day. While no lump was detected, the image of her breast showed many microcalcifications and looked like a starry night! An ultrasound was done, and there on the screen was an enormous black mass. "Oh my God, not again," I thought. The next 18 months consisted of chemotherapy, radiation, double mastectomies, and 40 weekly treatments of Herceptin. The pathology was intraductal carcinoma. After all her treatments, the pathology on the removed breasts was ductal carcinoma in situ only; she had a complete response to chemotherapy. Finally a victory over cancer! Today Kim is doing very well. Shortly after diagnosis, a genetic work-up revealed that Kim had Li-Fraumeni syndrome. Kim's daughter, Grace, who has a 50% chance of inheriting the mutation and is now 4 years old, has not been tested. But as you can imagine, for neither Grace nor her mother there is no such thing as a simple headache, no harmless sore throat, no lymph node that loving fingers don't glide over in hopes that there will be no knot under the skin. It can be a daily struggle to live with the knowledge that there is a bomb ticking in every cell of your body that is lying in wait for the right stimulus to set it off. We have found a way not to give in to our fear. We live our lives, find love and beauty in every day, help others dealing with cancer and genetic mutations, and remain ever vigilant in our early detection measures. --Patricia Holm