The etiology of cancer is complex and involves multiple events. However, among these changes must be disruption of normal control of cell proliferation and genomic stability. In mammals, cell proliferation is primarily controlled at the point of cell cycle entry, the G1- to S-phase transition. A large number of genes exist that can influence cell proliferation. They normally function to control cell division spatially and temporally during embryonic development. Many of these genes, when activated in an inappropriate tissue or at an abnormal time, can cause cells that are not dividing to enter the cell cycle and begin to proliferate. They do so by interfacing with the basic cell cycle–regulatory machinery to activate cell cycle entry. As tumor cells evolve, they lose their ability to prevent mutations and maintain control of genomic stability. This allows mutations in growth regulatory genes to accumulate and tumors to form. We hope that by understanding the basic regulatory machinery responsible for catalyzing cell cycle entry and understanding how cells maintain genomic stability, we will uncover mechanisms that may be points through which general growth-inhibitory therapies may be targeted to attack cancer.
Cellular Responses to DNA Damage
Cancer occurs by the accumulation of mutations in growth-controlling genes. Conditions under which increased frequencies of mutations occur often lead to cancer, and there are several examples of human diseases that cause higher mutation rates and result in a predisposition to cancer. Enhanced mutation frequencies can result from environmental factors, such as increased exposure to DNA-damaging agents, or genetic factors that decrease the efficiency of normal DNA repair processes. Organisms have evolved elaborate sensory networks to detect and repair DNA damage to prevent alterations in their genetic material. Cellular responses to DNA damage and to blocks in DNA replication are similar: cells arrest progression through the cell cycle at distinct transitions and induce the transcription of genes whose products facilitate DNA repair.
The capacity to detect and respond to DNA damage is central to an organism's ability to avoid damage to its genetic material. To investigate how eukaryotic cells accomplish this task, we have undertaken genetic approaches to identify genes responsible for this in yeast and humans. We have uncovered a signal transduction pathway that senses and responds to DNA damage. Near the top of this pathway is a pair of PIK-related protein kinases, ATM and ATR. Of these, ATR is the most important, as it, together with its regulatory partner, ATRIP (ATR-interacting protein), responds to problems in DNA replication, the major source of genomic instability in cells. We recently discovered that the ATR-ATRIP complex senses DNA damage by binding to single-stranded DNA (ssDNA) coated with the binding protein RPA (replication protein A). This RPA-ssDNA complex, which is the signal for DNA damage and stalled replication forks, initiates the checkpoint response.
A second sensor of DNA damage is the Rad17/Rfc2–5 complex. We discovered that this complex also recognizes RPA-ssDNA complexes and when it does, it loads a PCNA (proliferating cell nuclear antigen)-related trimeric complex called 9-1-1 onto sites of DNA damage. Rad17 and 9-1-1 interact with the ATR-ATRIP complex on DNA to initiate a solid-state signaling apparatus at the site of damage and replication stress.
A key aspect of this pathway is the identification of the substrates for the ATM and ATR kinases in response to DNA damage. Through proteomic analysis, we have identified more than 700 proteins whose phosphorylation increases in response to DNA damage in vivo. These substrates implicate a large number of novel regulatory pathways in the DNA damage response and establish the framework for understanding this response. It is clear that the presence of DNA damage profoundly alters cellular physiology. Using this substrate list, we have identified several new proteins involved in the DNA damage response, including two new human disease genes, the Fanconi anemia I (FANCI) gene required for cross-link repair and the SMARCAL1 protein defective in Schimke immunoosseous dysplasia (SIOD). We have also discovered a new Brca1 complex, the A complex, which includes Brca1, Bard1, and three new proteins, Abraxas, Nba1, and Rap80.
Genetic Screens in Mammals
The tremendous power of genetics to unravel the mechanisms underlying multiple pathways in yeast, Drosophila, and Caenorhabditis elegans has transformed our understanding of cell biology, neurobiology, and development. Until recently, large-scale genetic screens have not been possible in mammals. Recently, however, sophisticated tools have emerged that have allowed us to consider performing both gain-of-function and loss-of-function genetic screens in mammals. Using the power of small interfering RNAs (siRNAs) to silence gene expression, we are now undertaking loss-of-function screens in mammalian cells. In collaboration with Gregory Hannon (HHMI, Cold Spring Harbor Laboratory), we have developed a sequence-verified genomic library of short hairpin RNAs in retroviral vectors that cover all known human and mouse genes. We are using our bar code–microarray technology to apply these libraries to problems involved in cell cycle control, checkpoint signaling, tumorigenesis, and transcriptional control of important regulators of cell proliferation. We discovered a collection of human genes whose overexpression or loss confers cancer-like phenotypes on cells. Many of these are altered in human cancer.
We now have the ability to identify genes upon which cancer cells depend for survival. We have identified a number of genes whose depletion is lethal to cells bearing Kras mutations, and we identified a large mitotic network of genes Kras depends on for survival.
We have also developed new genetic technologies that allow identification of substrates for ubiquitin ligases. This methodology, called Global Protein Stability (GPS) profiling, allows us to measure the half-lives of thousands of proteins simultaneously. There are more than 700 ubiquitin ligases that control the abundance of cellular proteins, and this level of regulation is likely to be as profound as transcriptional control. Using GPS, we have identified several hundred substrates of the Cul1–5 ubiquitin ligase family. We plan to use these proteins to understand how the proteome is remodeled after translation in the same way that microRNAs remodel the transcriptome after transcription and how this changes in response to stimuli. Thus we are taking a systems biology approach to investigate the plasticity of the proteome.