Chapter 1. p53 Mouse Models of Human Cancer

Introduction

At first glance, mice and humans appear to have little in common. From a genetic, physiological, and anatomical perspective, however, mice are remarkably similar to humans. Indeed, comparisons between the mouse and human genomes have revealed that 75% of mouse genes - a total of 15,187 genes - have a 1:1 relationship with human genes and are predicted to share functions with their human counterparts¹. Although the mouse and human genomes are not identical, they are similar enough that researchers can use genetically engineered mice to study genes linked to human disease. How are mouse models of human disease created? And how do they inform us about human disease?

One way to determine whether a gene contributes to the development of a disease is to silence its expression in a mouse strain. Researchers who do this in the lab are creating what's known as a "knockout" mouse model. Alternatively, researchers can introduce a wild-type or modified version of a gene linked to human disease into a mouse strain to generate a "knock-in" or "transgenic" mouse model. These models, which are discussed in Section 10.6 and in pages 806-807 of your textbook, often exhibit phenotypes that mirror those associated with the corresponding disease in humans. Since the development of genetically engineered mice, countless mouse models of human disease have been generated. The Mouse Genome Informatics (MGI) database (http://www.informatics.jax.org) describes the correlations between mouse models and human disease. This centralized database is invaluable in facilitating the use of mouse models to study human disease. Scientists use these mouse models to carry out preclinical studies aimed at disease prevention, intervention, and treatment. And while none of these models are perfect, they are extremely powerful tools for understanding the basic molecular mechanisms underlying human disease.

Here we'll explore how p53 knockout and knock-in mouse models have expanded our view of cancer-associated genes in humans.

Genetically Engineered Mice and Cancer Research

The use of animal models to study human disease is vital to research because ethical and monetary considerations mean that most of these experiments cannot be performed on human subjects. In studies of cancer genetics, the mouse is the most widely used mammalian model organism. Considerable effort has gone into developing relevant mouse models of various forms of cancer to help pinpoint the genes that are responsible for cancer formation and progression. The knowledge we have gained by using mouse models cannot be understated.

Introduction to p53

Cancer is a class of disease characterized by uncontrolled cell growth or proliferation and, in the case of malignant cancer, invasion of adjacent and distant tissues. Tumor suppressor genes are a category of genes that suppress cancer formation. Some tumor suppressor genes encode proteins that repress genes essential for the progression of the cell cycle. Other tumor suppressor genes couple the cell cycle to DNA damage or initiate programmed cell death. Phenotypes associated with mutations in tumor suppressor genes are most often recessive, meaning that both alleles of the gene must be inactivated before a phenotypic effect is observed. One notable exception to this rule is the p53 gene product.

p53 (also called tumor protein 53, or TP53 and Trp53 in mice) was named "Molecule of the Year" in 1993 by the journal Science². Why is this molecule so important? p53 is a tumor suppressor that has important functions in DNA transcription, cell growth, and cell proliferation. Because of these functions, p53 plays a critical role in cancer prevention. If the p53 gene is damaged or lost, the risk of developing cancer is greatly increased. Interestingly, certain types of mutations in just one p53 allele are sufficient to inactivate the wild-type allele. Over 50% of human tumors contain a mutation or deletion of p53, including breast, cervical, colon, lung, liver, prostate, bladder, and skin cancers. In addition, cancers associated with a p53 mutation tend to be more aggressive, to be more likely to spread to other parts of the body, and to have a worse prognosis.

p53 Knockout Mice

Are mice that lack a functional copy of the p53 gene more prone to cancer? In 1992, Donehower and colleagues achieved the first p53 gene knockout in mice³. This team used homologous recombination in mouse embryonic stem cells to inactivate the p53 gene. Using molecular biology techniques (discussed in Section 10.6 and in pages 806-807 of your textbook), these researchers generated two types of mice: heterozygous mice (p53^+/-) that carried only one functional p53 allele, and homozygous null mice (p53^-/-) that lacked both functional p53 alleles and had no functional p53 protein.

The team monitored the p53^-/- mice, p53^+/- mice, and their wild-type littermates (p53^+/+) for spontaneous tumor development. Not surprisingly, the p53^-/- mice developed tumors at a much earlier age than their heterozygous or wild-type counterparts. The researchers found that 74% of the p53^-/- mice developed tumors by six months of age. This observation was in sharp contrast to the 2% tumor rate observed in p53^+/- mice. When the researchers examined the tissues of p53^-/- mice for tumors, they observed that a remarkable variety of cell types were involved, showing that the loss of p53 was associated with tumors from a wide range of human organs. These null mice have been instrumental for researchers studying p53-associated functions and have prompted numerous studies.

p53 Knock-in Mice

If mice lacking p53 are sensitive to cancer, are mice with extra copies of the p53 tumor suppressor gene more resistant to cancer? To address this question, researchers have also made "super p53" mice⁴. Using molecular biology techniques (discussed in Section 10.6 and in pages 772-773 of your textbook), Garcia-Cao et al. set out to generate mice with an extra copy of p53 that maintained its normal regulation. The result was super p53 mice, which harbor an extra copy of a large DNA segment - called a transgene - that contains the p53 gene to maintain the gene's "natural genomic context."

Figure 1: Validation of p53 function in "super p53" transgenic mice.
Figure 1 Source: Figure 3 (Panel A only) from Garcia-Cao, I. et al. Super p53 mice exhibit enhanced DNA damage response, are tumor resistant and age normally. Embo J 21, 6225-35 (2002), doi: 10.1093/emboj/cdf595, http://www.nature.com/emboj/journal/v21/n22/fig_tab/7594814a_F3.html

As an important control experiment, the research group first tested the "functionality" of the p53 transgenic allele by mating transgenic mice (p53^tg) with p53 null mice (p53^-/-) that lack both copies of the p53 allele and therefore have no functional p53 protein. They were able to show that these p53^-/-;tg mice responded to DNA damage in a manner similar to p53^+/- mice, suggesting that the p53 protein produced by the transgene is functional and subject to the same regulatory mechanisms as the endogenous p53 protein (Figure 1)⁴.

Figure 2: Super p53 mice are tumor resistant.
Figure 2 Source: Figure 5 (Panel A only) from Garcia-Cao, I. et al. Super p53 mice exhibit enhanced DNA damage response, are tumor resistant and age normally. Embo J 21, 6225-35 (2002), doi: 10.1093/emboj/cdf595, http://www.nature.com/emboj/journal/v21/n22/fig_tab/7594814a_F5.html

Having established the functionality of the p53 transgenic allele, the researchers tested the tumor resistance of the super p53 mice. Exposing these animals to chemicals that cause cancer revealed that whereas in wild-type mice (p53^+/+) the rate of tumor development was 92%, in super p53 mice (p53^+/+;tg) it was only 33% (Figure 2)⁴. In addition, the team showed that the rate of "spontaneous" tumor formation was much lower in the super p53 mice (only one out of six mice, approximately 17%) compared to the wild-type mice, which had a 47% incidence of spontaneous tumor formation (Figure 3)⁴. The investigators observed a linear correlation between the p53 gene dosage (one copy versus two copies versus three copies) and the enhanced response to DNA damage and tumor resistance. This correlation was somewhat of a surprise (even to the investigators). The researchers proposed that the presence of the extra copy of p53 may make the mice more tolerant to p53 inactivation by stochastic mutations. In other words, the extra p53 transgene may serve as a "spare tire" for the mouse.

Figure 3: Super p53 mice develop fewer spontaneous tumors.
Figure 3 Source: Figure 6 (Panel B only) from Garcia-Cao, I. et al. Super p53 mice exhibit enhanced DNA damage response, are tumor resistant and age normally. Embo J 21, 6225-35 (2002), doi: 10.1093/emboj/cdf595, http://www.nature.com/emboj/journal/v21/n22/fig_tab/7594814a_F6.html

These landmark mouse models were invaluable in helping researchers understand the complex functions and interactions of p53. Similar models have been created for myriad other cancer-related genes and have been very helpful for researchers studying human cancer biology.

What Limitations Should We Consider?

Academic researchers, clinicians, and pharmaceutical researchers use mouse models extensively, from basic research on human disease to preclinical studies. But while mouse models of human diseases represent an invaluable tool for scientists, there are a number of limitations we need to consider so that we may improve on existing models. For example, most human disorders are polygenic, meaning that a combination of alleles or genes is involved the disease. In contrast, many of the genetically engineered mice target a single gene (that is, they are monogenic). In addition, humans display far greater genetic diversity than do the inbred lines of mice used in the laboratory. Finally, the contribution of external (environmental) factors is rarely considered when evaluating the phenotypes of genetically engineered mice. Some of these environmental factors include diet, exercise, exposure to pollutants, and stress - all of which are tightly controlled and normalized in laboratory settings (Figure 4)⁵.

Figure 4: Schematic representation of five key environmental effects on mouse model systems.
Figure 4 Source: Figure 4 from Beckers, J., Wurst, W. & de Angelis, M. H. Towards better mouse models: enhanced genotypes, systemic phenotyping and envirotype modelling. Nat Rev Genet 10, 371-80 (2009), doi:10.1038/nrg2578, http://www.nature.com/nrg/journal/v10/n6/fig_tab/nrg2578_F4.html

Summary

Despite their limitations, mouse models of human disease represent the best mammalian model systems researchers have to work with. To date, mouse models have contributed an enormous body of knowledge to the scientific and medical communities. In particular, studies of the p53 knockout and super p53 mice have provided exciting new insights into the underlying causes and treatment of human cancer.

References

Quiz