Taking Aim at p53: Researchers are targeting the tumor suppressor with vectors, viruses, and small molecules.

POTENT COMBINATION: Joining gene therapy with conventional cancer treatment may more effectively kill recurring tumor cells, thinks Georgetown University Medical Center's Esther Chang. In preclinical experiments, Chang and colleagues administered p53 via a compact, targeting nonviral vector along with DNA-damaging therapy to kill a variety of tumor types. The mice are cancer free 17 months after treatment, Chang reports. Selectivity and ubiquity. Because both properties characterize the tumor

Paul Smaglik
Jan 17, 1999


POTENT COMBINATION: Joining gene therapy with conventional cancer treatment may more effectively kill recurring tumor cells, thinks Georgetown University Medical Center's Esther Chang. In preclinical experiments, Chang and colleagues administered p53 via a compact, targeting nonviral vector along with DNA-damaging therapy to kill a variety of tumor types. The mice are cancer free 17 months after treatment, Chang reports.
Selectivity and ubiquity.

Because both properties characterize the tumor suppressor gene p53, the gene has emerged as one of the top targets in the war against cancer. P53 plays an integral role in activating programmed cell death (apoptosis) and is mutated in about 55 percent of tumor types. Potential relief strategies under development include replacing the defective gene, injecting viruses into tumors, and reactivating the gene's apoptotic function with small molecules.

When Arnold J. Levine, now president of Rockefeller University, and Princeton colleague Daniel I.H. Linzer discovered p53 in 1979, they had no conception that their finding would result in any of these approaches.1 By 1989, they suspected a tumor suppressor role,2 which Bert Vogelstein and colleagues at Johns Hopkins confirmed.3 And in 1994, researchers learned how the tumor suppressor interacted with downstream protein Mdm2 to turn the tumor suppressor function on and off.4 Fifteen years after its characterization, p53 had become a viable target. "It's the single most common mutation in human cancer," Levine remarks in a telephone interview. "If we can learn how to really harness it and work it right, we can try to get back at the cancers and kill them."

Researchers think manipulating the gene could kill only cancerous cells and spare healthy ones--something cancer treatments have yet to do. Existing cancer treatments remain crude, Levine remarks, indiscriminately killing healthy cells along with cancerous ones. Targeting p53 could lead to less toxic, more effective alternatives. "There are reasons to believe it will be selective for the cancer cells because it is activated and initiates apoptosis when an oncogene is active in the cell. Normal cells would be spared apoptosis."

In theory, restoring the p53 function through gene therapy sounds like the most elegant approach. "Using a virus, like adenovirus, you could return p53 to the cell in high amounts so that it's active. The environment of the cell that has the activated oncogene triggers apoptosis, and the cells will die," Levine comments. "Now the downside of that approach is the ability to deliver a virus to every one of the cancer cells. It's not possible. The problem is a delivery problem, not a conceptual problem. If you could deliver [ p53 ] to every cell, you'd have a solution."

Still, some are trying. This fall, a group of researchers at the University of Texas Southwestern Medical Center in Dallas began the first p53 -replacement trial for ovarian cancer. However, other gene therapy proponents are not optimistic about gene therapy alone as a solution for cancer.

"The idea of using a gene itself--a single agent--is not valid anymore," states Esther H. Chang, of Georgetown University Medical Center's Lombardi Cancer Center in a recent telephone interview. Chang notes that, of all the gene therapy-only cancer approaches in clinical trials she's aware of, not one has yet met with real success. Instead, she favors combining p53 replacement with conventional treatment such as chemotherapy or radiation therapy, because even the best viral vectors can't deliver therapeutic agents to every single tumor cell. And without that almost impossibly high level of efficiency, cancer cells not transfected will likely give rise to new cancer cells, yielding an even more chemo- and radiation-resistant tumor. At the National Foundation for Cancer Research's 25th Anniversary Conference at Georgetown University this past fall, Chang presented preclinical findings from experiments that combined p53 delivery via a nonviral vector with conventional treatment. According to those findings, mice with six different implanted tumor types receiving conventional treatment with the nonviral gene therapy had their tumors shrink to nondetectable masses. And 17 months after the treatment, the tumors have not returned, Chang adds.

Part of the reason for the efficacy lies in the design of the vector, Chang suspects. "The problem with the viral vectors--whether they're adeno-, retro-, AAV (adeno-associated), herpes, you name it--[is that] at least at this time, none are tumor-targeting." Instead, Chang and colleagues used a cationic lipid because they could engineer it to bind to ligands present in tumor cells but not in healthy cells. "We target it in such a way that the liposome encapsulating the therapeutic gene will home to the tumor--not only the primary tumor, but the metastatic tumors." That ligand-liposome complex allows researchers to inject the vector into the mice tail veins. The vector then travels through the bloodstream and sticks only to cancerous cells, Chang hypothesizes.

Chang also thinks that the small size of their vector allows a higher transfection rate; one of the drawbacks to nonviral vectors has been their larger sizes.5 "We make the liposomes so compact that they are able to leak through the tumor microvessels." That compactness allows systemic injection, which Chang favors over injections of vectors directly into a tumor. Next, Chang hopes to find a pharmaceutical partner and begin clinical trials. If human trials succeed, the next step will be to convince clinicians to reduce the doses of chemotherapy and radiation therapy administered with the gene therapy. "I have to demonstrate to them--in the clinic, not in mice--that in patients we are able to see total elimination of recurrence .... Then they will be willing to start, very cautiously, dropping some doses, which will come later. That's really my ultimate dream."


Frank McCormick
Onyx Pharmaceuticals Inc.'s approach could perhaps be described as gene therapy without the gene. The Richmond, Calif., company's strategy involves injecting an attenuated adenovirus--often used as a gene therapy vector--into p53 -deficient head and neck tumors. "We're not delivering a gene," Frank McCormick, Onyx founder, explained in an interview shortly after he presented preliminary Phase II findings at a New York Academy of Sciences cancer conference this past November. "We're actually using the lack of p53 to allow a virus to grow and kill cells." Adenoviruses require p53 to replicate in normal cells. When a virus infects a normal cell, p53 is activated in the cell. In healthy cells, proteins produced by p53 fight viral replication. However, some viruses have developed a strategy to fight back. Unattenuated adenoviruses produce a protein called E1B55K, which binds to p53 directly and neutralizes it. "In cancer cells, p53 is already missing, so you don't need E1B anymore," McCormick explains. "The cancer cells make up for the defect in the virus."

At the conference, McCormick presented slides of patients with fist-size tumors that shrank to nonvisible masses less than four weeks after treatment. Matching Chang's hypothesis, the viral treatment proved more effective when accompanied with chemotherapy and radiotherapy. Earlier in the trial, McCormick and Onyx colleagues thought the virus fought tumors by aping the apoptotic function of p53.6 Now they suspect other mechanisms may be at work as well. "The body's basically destroying such an enormous mass of tumor so quickly after an infection with a virus ... suggests that immune effectors--macrophages and so on--recognize infected cells in the tumor and home into the tumor; then they can produce TNF-alpha and other proteins that are toxic and basically kill uninfected cells."

Levine, who calls the Onyx trial "promising," agrees with McCormick's new hypothesis. "Setting up a viral infection in a tumor probably brings in a strong immune response," Levine comments. "That may even be part of the 'bystander effect.'" Understanding the controversial bystander effect may be important in developing future therapies. Right now researchers don't know how much of Onyx's success can be attributed to their p53-dependent mechanism and how much can be linked to an immune response. "Just letting any virus replicate in the tumor cells will bring in immune cells in response," he continues. "That would happen both in Onyx's approach and ... in the delivering of p53 by adenovirus."

Rather than depending on little-understood immune responses of injected viruses, or hoping for higher rates of transfection through gene therapy, Levine favors a small-molecule approach. He suspects that small molecules acting on the p53 pathway will prove a more reliable treatment for some forms of cancers. McCormick disagrees, noting that he first looked downstream of p53 before taking up Onyx's viral approach. "It's still very difficult to imagine how a small molecule can deal with the effects of losing functional p53, " McCormick notes. "Lack of a function doesn't really give you much to target in a classical medicinal chemistry approach."

However, Levine notes that 30 percent of tumor types with wildtype p53 intact also have high levels of the negative activator protein Mdm2 available. This protein, produced by p53, binds to the tumor suppressor protein, and, in effect, turns it off, allowing tumor cells to divide out of control. Finding a way to break that complex with a small molecule that binds to Mdm2 would free p53. "When p53 is free, then it is active to kill."

David P. Lane, professor of biochemistry at the University of Dundee, and colleagues also suspect this approach will be effective, Lane reported at the New York cancer meeting. They localized the binding site of Mdm2 on p53 and found that the downstream protein can block activation of p53 or transcription. In addition, they disrupted the formation of the p53 -Mdm2 complex with proteins and peptides, providing a proof of principle. The next step remains finding a small molecule with similar activity. "There is a strong need to develop therapies that achieve the same therapeutic effects, but are not genotoxic," Lane commented.

Levine suspects that such molecules are being pursued via high-throughput screens at pharmaceutical companies around the world. "For '99, I'm hoping we have a good drug," comments Levine, who laughed when asked if he was aware of any candidates in the pipeline. "That's a dream. That's not a fact."


Arnold J. Levine
P53 Status:
New Indicator Could Improve Old Treatment

While researchers are justifiably excited about the tumor suppressor gene p53 as a target for new anticancer drugs, one of the gene's founders suspects that knowing a tumor's p53 status will help clinicians better choose from an array of existing treatments. Until recently, clinicians had little scientific basis to pick one DNA-damaging treatment over another, notes Arnold J. Levine, president of Rockefeller University. "We've used standard therapy, but we never knew why it worked." Knowing a tumor's p53 status could change that, Levine adds.

Levine notes that, for lymphomas, p53 status serves as a good predictor of standard chemotherapy efficacy. For example, p53 status probably explains why chemotherapy has been fairly successful in children with acute lymphoblastic leukemia (ALL). "If they have wildtype p53 and they are treated with DNA-damaging agents, the DNA damage induces p53, the [cancer] cells undergo apoptosis, and the child has a very high probability of being cured." In contrast, a child with mutant p53 will not respond. So, a clinician might be best served by attacking a different target and perhaps avoiding unnecessary treatments with myriad side effects. "Why treat someone who you know is not going to respond?" However, answering that question for nonlymphoblastic tumor types will require more research, Levine concludes. --P.S.

  • D.I.H. Linzer and A.J. Levine, "Characterization of a 54K Dalton cellular SV40 tumor antigen present in SV40-transformed cells and uninfected embryonal carcinoma cells," Cell, 17:43-52, 1979.

  • C.A. Finlay et al., "The p53 proto-oncogene can act as a suppressor of transformation," Cell, 57:1083-93, 1989.

  • P.W. Hinds et al., "Mutant p53 DNA clones from human colon carcinomas cooperate with ras in transforming primary rat cells--a comparison of the hot-spot mutant phenotypes," Cell Growth and Differentiation, 1:571-80, 1990.

  • C.Y. Chen et al., "Interaction between p53 and Mdm2 in a mammalian-cell cycle checkpoint pathway," Proceedings of the National Academy of Sciences, 91:2684-8, 1994.

  • P. Smaglik, "Viral vs. Nonviral in Gene Therapy: Which Vector Will Prevail?" The Scientist, 12[13]:1, June 22, 1998.

  • P. Smaglik, "Biotech Firms On Quest For Apoptotic Therapies," The Scientist, 12[6]:1, March 16, 1998.