Characterization of signal transduction pathways that activate tumor suppressor genes.
p53 is part of a stress response pathway.
To survive an organism requires that its cells be able to properly respond to environmental insults. Failure to do so can result in a variety of diseases including cancer. Beginning in 1991, evidence has mounted showing that a major sensor to environmental insults is the p53 tumor suppressor gene. At the cell biology level the response to DNA damage results in an increase in p53 protein and its nuclear accumulation. Upon activation p53 directly transactivates genes such as GADD45 and p21waf1/cip1 which mediate some aspects of DNA repair and cell cycle inhibition respectively. Although not as well described, p53 has also been shown to repress transcription, mediate DNA mismatch repair, inhibit homologous recombination and increase the activities of topoisomerase I. At the cellular level p53 can initiate cell cycle inhibition, apoptosis, or differentiation. At the organismal level p53 is required for preventing cancer formation. After eight years of data compilation it is now clear that p53 is inactivated in 40-50% of human cancers by either direct mutation, or by inactivation by viral and cellular oncogenes. While a number of well established downstream p53 functions have been characterized what remains unclear, and is the subject of intense research, is how p53 receives stress signals. The goal of my research program is to define the molecular steps that lead to p53 activation.
Classification of reagents that activate the p53 signal transduction pathway.
Upon stress the p53 protein level increases and accumulates in the nucleus. Reagents that activate p53 can be broadly classified into two categories: 1. Those that directly damage DNA; 2. Those that act on some other uncharacterized molecule within the cell. Examples of reagents that can directly damage DNA include ionizing radiation, UV light, and Actinomycin D. Conditions that do not directly damage DNA but still activate p53 are mild heat shock, activated H-ras protein, hypoxia and lowering of ribonucleotide pools. As a matter of fact, as far as I know, p53 responds to a broader range of reagents that any other environmental sensor. It is probably the universal requirement for the p53 response that accounts for its functional absence in many different types of tumors. If one presumes that each type of tissue can be exposed to unique toxins, whether external or internal, it is reasonable for p53 to respond to such a broad range of toxins.
There appears to be multiple pathways that lead to p53 activation. Evidence for this hypothesis is derived from the fact that p53 responds to different stress reagents with different kinetics. For example, in breast cancer epithelial cells, we have observed that the p53 level quickly increases in response to ionizing radiation resulting in cell cycle arrest. On the other hand, when breast epithelial cells are treated with Actinomycin D, the p53 level increases much more slowly. We have also observed that Actinomycin D treatment results in p53 nuclear accumulation and strong induction of the downstream effector gene MDM2. This stands in contrast to the fact that ionizing radiation fails to result in nuclear accumulation with little induction of MDM2. Furthermore, simultaneous treatment with both stressors shows that the ionizing radiation response is dominant (i.e. cytoplasmic retention of p53). My interpretation of this data is that once p53 is committed to the ionizing radiation pathway it is desensitized to Actinomycin D. Why p53 responds differently to two DNA damaging agents is unclear. It is possible that p53 is retained in the cytoplasm to modulate the translation of certain transcripts such as CDK4 (a cyclin-dependent kinase). p53 has been shown to decrease the translation of CDK4 transcript in other systems. We would like to uncover the chemical mechanisms governing these varied stress responses.
What will be gained by uncovering this pathway?
The fact that approximately 50% of cancers contain wild-type p53 suggests one of two possibilities: 1. p53 retains a normal function; 2. p53 function is abrogated by a second-site mutation. One particular example of the latter scenario is patients with a rare inherited disorder known as ataxia telengiectasia (AT). Patients with AT suffer from severe muscle atrophy and extreme radiosensitivity. They usually die in their late teens. Recently, it was observed that some of these patients lack a gene called ATM, which was recently been cloned. Experiments with cells from mice that are null for ATM have clearly shown that the p53 response to ionizing radiation is lacking. Uncovering the p53 response pathway may indicate other genes responsible for cancer or extreme sensitivity to environmental toxins. Interestingly, other types of stress reagents such as UV do not require ATM to signal p53 again demonstrating that multiple pathways activate p53. Thus, discovery of upstream molecules in the p53 signaling pathway could define other genes that are targeted in cancers and, perhaps, other diseases.
One can imagine that cancers may initiate because of a failure to elicit a proper p53 response to stress reagents. Immunohistochemical analysis of cancer tissue expressing mutant p53 protein generally indicate that p53 levels are elevated in the nucleus. This indicates that the p53 pathway has been activated. From this perspective, it is easy to see why most cancers express mutant p53 at high levels in the nucleus. I interpret this as evidence that the p53 pathway has been activated but is incapable of performing its function. Understanding the players immediately upstream of p53 will offer the chance to design very specific prodrugs that can be activated by the same molecules that activate p53 within cancer cells.
Another reason why it is important to understand the pathway leading to p53 activation is to design and screen potential chemopreventive reagents. Chemopreventive reagents would be any compound that prevents cancer. An important test of such a drug would be to protect cells from DNA damaging agents which often damage DNA through oxidative reactions. Thus, antioxidants are good candidates for chemopreventive agents. However, it is important to be sure that chemopreventive agents do not merely mask the DNA damage signal from reaching p53. If that should happen p53 would be rendered ineffective in cases where DNA damage has occurred. Thus, the most effective chemopreventive reagents are those that either prevent DNA damage or those that ensure that damaged cells have properly activated p53. Activated p53 then, will either lead to DNA repair or apoptosis. Cells that have properly activated p53 are probably assured not to contribute to tumorigenesis.
Redox regulation of p53
Using our model system, we have investigated a variety of compounds as potential inhibitors of p53 activation. We observed that pyrrolidine dithiocarbamate (PDTC), a metal chelating agent, that inhibits p53-mediated activation of downstream targets. PDTC also prevents p53 from being degraded by E6 protein, a Human Papilloma Virus protein that normally leads to rapid p53 degradation. This set of studies suggested that PDTC may act on p53 itself. Based on this data, we designed a novel assay to determine if p53 cysteine residues were oxidized by PDTC in vivo. Indeed, we found that p53 was oxidized by PDTC, which correlated with inhibition of p53 activities. To further test this possibility, we obtain a series of p53 cysteine substitution mutants from Dr. Kristine Mann (University of Alaska). We will express these mutants in p53 null cells and map the cysteine residue that is oxidized. We also intend to test if other oxidizing agents such as hydrogen peroxide and nitric oxide oxidize p53 in vivo.
MDM2 Amplification Database
A continuing interest of mine is MDM2, an oncoprotein that regulates p53 activity by directly binding to p53 and mediating its destruction through p53 ubiquitination. This natural inhibitor of p53 actually appears to be part of a negative feedback loop. After p53 is activated by DNA damage p53 transactivates the MDM2 gene. The MDM2 protein product then binds p53 and prevents p53 transactivation. Several groups found that MDM2 is amplified in a variety of human tumors. As part of two reviews recently published (Momand and Zambetti, 1997; Momand et al., 1998) I compiled MDM2 amplification frequency data on 28 different types of malignancies from peer-reviewed journals. Over 3,900 samples were analyzed. I have calculated the frequency of MDM2 amplification in each tumor type and have just constructed a Web site in collaboration with other scientists at City of Hope to make this data available to the scientific community. The address is http://www.infosci.coh.org/mdm2. The overall MDM2 amplification frequency is 7% and it is represented in 14 different cancers. The highest frequency of MDM2 amplification occurs in soft tissue tumors (20%) and osteosarcomas (16%).