|
The Giovanni Armenise-Harvard Foundation Cancer Biology, Genomics, and Post-Genomics Presentations - Day One
Dr. Leder's opening presentation introduced the first major theme of the symposium; namely, that many cancers result from a series of individual genetic mutations that conspire to eventually produce disease. The path to malignancy often begins with a single genetic change, which does not cause runaway cell proliferation until other mutations disrupt or activate specific pathways. Over the years, Dr. Leder's laboratory has created mouse models that demonstrate how a series of mutations can lead to malignant transformation, and which enable researchers to study individual steps along the way. Because some of these steps may represent opportunities for treatment, these models are being put to a new use: helping scientists screen large numbers of drugs with cancer fighting potential. Dr. Leder described a series of experiments using a mouse model for breast cancer. These involved using retroviral vectors to insert selected mutations into transgenic mice that already carried an activated, initiating oncogene. Dr. Leder and his colleagues predicted that some of these insertions would alter additional pathways in ways that would Ü in concert with the transgene carried by the mice Ü accelerate the development of breast malignancies. Using this approach, the researchers identified specific collaborating partnerships between the fibroblast growth factor (FGF) and the Wnt pathways in murine breast cancer, and have shown that Wnt acts on b-catenin and its targets. Building on these findings, Dr. Leder's lab has created a series of transgenic mice that develop malignancy following the introduction of such well-known cancer genes as mutated p53, BRCA-1, or Neu (HER-2). The researchers can now create malignant cells with specific combinations of mutations, which are incorporated into microarrays that can be used to screen literally thousands of small molecules. These microarrays, which are an increasingly important tool for drug development, can be used to identify chemicals that most effectively block proliferation. The hope is that some of these will prove safe and effective enough to treat cancer patients in the future.
Dr. Leder set a theme by establishing that oncogenes act in concert to produce malignant transformation. Dr.Giordano discussed a related kind of malevolent teamwork, in which two different ligand-receptor pairs join forces to promote invasive growth by cancer cells. The first pair is semaphorin 4D and plexin B1. Semaphorins are soluble signals that guide axon growth in neurons, and which bind to plexins, a family of receptors found not only in the nervous system, but also in other types of cells. The second pair consists of a scatter factor, hepatocyte growth factor (HGF), and its receptor Met, a tyrosine kinase from the MET/RON/SEA oncogene family. Although HGF and Met are important during normal embryonic development, they have also been associated with malignancy when mutated in certain ways. Although the biological activities of sema 4D and plexin B1 in nerve cells are well understood, their function in other settings has been unclear. Working with epithelial cells, Dr. Giordano and her colleagues showed that sema 4D could trigger a complex program for invasive growth, including cell-cell dissociation, anchorage-independent growth and branching morphogenesis. The researchers recognized that these same events in epithelial cells can be launched by the binding of scatter factors, such as HGF, to tyrosine kinase receptors such as Met. They also noticed striking structural homology between Met-type receptors and plexins, notably in the extracellular domain. More surprising was the discovery that plexin B1 and Met form a complex, and that when sema 4D binds to its receptor, plexin B1, this stimulates the tyrosine kinase activity of Met, resulting in tyrosine phosphorylation of both receptors. Taken together, these events triggered branching morphogenesis and invasive growth by epithelial cells. Semaphorin had no effect, Dr. Giordano observed, on cells where Met was not expressed Ü indicating that this tyrosine kinase receptor is essential for semaphorin-stimulated invasive growth.
BRCA1 and 2 have become famous as "breast cancer genes," because germ-line mutations in these genes predispose women to development of early onset breast and/or ovarian cancer. BRCA1 and 2 are highly conserved genes found in a variety of animals as well as humans, and Dr. Livingston and his colleagues have analyzed protein function using a frog embryogenesis model. As a result, his lab has uncovered an unusual and important partnership between the BRCA1 protein and the protein encoded by a similar gene called BARD1. Prior to these experiments, BRCA1 and 2 were known to have tumor suppressing activity and to encode proteins that interact physically with each other and with other proteins as well Ü some known, some not. The full range of function for these pairings was unclear, although there was ample evidence that cells need BRCA 1 and 2 "on patrol," Dr. Livingston said, to protect the integrity of their genome. His lab has been analyzing mechanisms these proteins use to maintain genome integrity, as this is the likely key to their tumor suppressing activity. In frog embryos, the researchers discovered collaboration between the BRCA1 protein and the protein encoded by BARD1, a gene also known to play a role in tumor suppression. It turns out that the two associate to form a heterodimer, and this structure maintains the high levels of the two proteins required for normal development. If the proteins were prevented from associating in this way, and the level of either was depleted during embryogenesis, a range of growth defects and deformities occurred in the frogs. These findings raise provocative questions about what else BRCA1 may be up to. There is evidence that this protein interacts with products of at least seven other genes associated with cancer, some known to cause malignancy and others known to suppress tumor growth. Dr. Livingston speculates that BRCA1 is an "influence peddler" that plays key roles in genome maintenance, proliferation control, and cell differentiation by regulating the activities of other proteins, rather than acting on its own. Future research will explore this idea.
Dr. Minucci's research is part of a long-term exploration of specific molecular events in the pathogenesis of acute promyelocytic leukemia (APL), particularly chromosomal rearrangements of the transcription factor RAR that affect patient survival and response to treatment with retinoic acid. His lab studies the biological activities of a fusion protein that is formed by RAR and the product of a nuclear gene called PML, which prolongs survival and blocks cell differentiation. Now they have found that events in other acute leukemias mimic this joint activity. Dr. Minucci and his colleagues had previously shown that oligomerization of RAR, through a self-association domain present in PML, activates transcriptional co-regulators that in turn recruit histone deacetylases (HDACs). HDACs are required for transcriptional repression of PML-RAR target genes, and for the transforming potential of the fusion protein. More recently, they found evidence that nuclear PML recruits p53, which when modified protects against apoptotic cell death. It now appears that the PML-RAR complex has two mechanisms of action Ü one where p53 activity prolongs survival; the other in which abnormal HDAC recruitment blocks differentiation. The researchers hypothesized that the same mechanisms might also be at work in non-APL forms of acute myelocytic leukemia. Experiments confirmed that oligomerization and altered recruitment of HDACs are also responsible for transformation by the fusion protein AML1-ETO, found in a type of AML. AML1-ETO expression blocked retinoic acid (RA) signaling, suggesting that HDAC recruitment may be a common theme in AMLs Ü and thus an opportunity for treatment. Dr. Minucci's team used a mouse model to show that chemical inhibition of HDAC, in addition to standard retinoic acid treatment, prolonged survival. The HDAC-inhibitor they used was valproic acid, a common anti-seizure medication. The combination of retinoic and valproic acids is currently being tested in a Phase II clinical trial.
Defects in RecQ-like DNA helicases are responsible for genetic instability in Bloom syndrome (BS), Rothmund-Thomson syndrome (RTS) and Werner syndrome (WS). Early in life, people with WS develop not only gray hair and wrinkles, but also typical diseases of old age such as cancer and cardiovascular disease. Dr. Sinclair is a yeast biologist who specializes in the genetic underpinnings of aging, and he is using this simpler organism to shed light on the connection between premature aging and cancer in WS patients. There is evidence that the premature senescence of fibroblasts cultured from WS patients is due to a defect in the telomere, the protective tip that guards chromosomes against erosion. The enzyme telomerase preserves telomere length when cells are in their prime, then declines as they age. In WS fibroblasts, senescence could be averted by introducing telomerase catalytic subunit (hTERT) into the culture. When the telomeres of yeast chromosomes reach a perilously short length, a parallel mechanism saves some of them from premature senescence. Others, however, are rescued by a yeast helicase called Sgs1 that causes the yeast chromosomes to develop abnormally long telomeres. In about 10% of human cancer cells, a poorly understood sequence of events called the "ALT" pathway, for "alternative lengthening of telomeres," is thought to immortalize cells by adding to their telomeres even when no telomerase is present. The Sgs1 gene in yeast is homologous with the WS gene, which encodes a DNA helicase known as WRN. Dr. Sinclair's findings about Sgs1 raise the possibility that the human WRN helicase is an "ALT" factor, and that it could be the missing link between premature aging and early onset of cancer in WS patients. Future studies in animal models may help elucidate the in vivo function of the mammalian RecQ-helicases as well as the mechanisms by which cancer cells proliferate in the absence of telomerase.
Like other tyrosine kinase receptors, epidermal growth factor receptor (EGFR) is the entry point into pathways important for signal transduction, endocytosis, and many other key cellular functions. Small GTPases of the Rho family are known to be involved in signal transduction, while members of the Rab family are established players in endocytosis. Dr. Lanzetti and her colleagues set out to investigate how signaling relying on these two classes of GTPases might be integrated. They studied individual intracellular proteins that are substrates for tyrosine kinase activity, and found unexpected collaborations among them. Her laboratory has a long-standing interest in Eps8, a protein that plays a role in endocytosis. When it acts in this capacity, Eps8 collaborates with Rab5, a gatekeeper located just beneath the cell membrane to mediate trafficking. This is part of the Ras pathway that is normally involved in the organization of actin into ruffles when tyrosine kinase receptors are activated. In the present series of experiments, Dr. Lanzetti found that alternative complexes between Eps8 and E3b1 or RN-tre connect signaling pathways for ruffle formation and endocytosis. E3B1 is an adaptor protein that couples Eps8 with the sos1 protein, which in turn activates the small GTPase called Rac. In this configuration, Eps8 contributes to the formation of ruffles. In a second configuration, the SH3 domain connects Eps8 with RN-tre, a GTPase activator protein. Dr. Lanzetti's experiments show that when RN-tre is complexed with Eps8, it inhibits Rab5, an important gatekeeper in the early stages of endocytosis. Knockout experiments show that RN-tre alone does not have this activity. At the same time, RN-tre diverts Eps8 from its Rac-activating function, which inhibits actin reorganization into ruffles. Thus it appears that when RN-tre is complexed with Eps8, signaling through Ras and Rac are affected as well as trafficking through Rab5.
Human papillomavirus is one of the most medically important microbes, due to its causal role in cervical cancer. Of more than 100 strains that have been identified, about two dozen are associated with genital tract lesions and a subset of these put infected women at high risk for cervical cancer. The dangerous strains make oncogenic proteins that transform cells by selectively inactivating tumor suppressor genes, including p53. Extensive research on one of these cancer-causing HPV proteins, E6, has been carried out in Dr. Howley's laboratory over the years. He and his colleagues showed that E6 subverts the normal activity of a cellular protein they call E6 associated protein (E6-AP), causing it to mark p53 protein for proteolytic destruction. Interestingly, this happens only in the presence of the viral protein E6; in uninfected cells, E6-AP is a protein ligase that doesn't interact with p53 at all. This caused the researchers to wonder what else E6-AP might be doing that contributes to oncogenic transformation. The observation that levels of Fyn, a Src-family tyrosine kinase, are high in cells infected with cancer-causing HPV provided an important clue. A long series of experiments in cells and in mice were needed to show that overexpression of E6-AP in HPV-infected cells caused levels of an enzyme Csk to fall. This enzyme turned out to be a player in the ubiquitin pathway, whose role is to flag the oncogenic Fyn protein for destruction. This is a second major role for E6-AP in malignant transformation: it not only inactivates helpful p53, but also elevates potentially dangerous Fyn.
Spinocerebellar ataxia type 1 (SCA1) is an inherited neurodegenerative disease caused by excess glutamine repeats in the gene that codes for ataxin-1. Healthy genes have less than 41 glutamine repeats; mutations result in as many as 83 and there is a direct correlation between the number of repeats and disease severity. In the brain, the disease destroys Purkinje cells and pontine and olivary nuclei. SCA1 resembles Huntington's disease: onset is late although the genetic defect is present from birth, and once the disease starts it is progressive and devastating. Dr. Servadio set out to find whether it was possible to reverse pathogenic changes associated with abnormal ataxin-1. His laboratory approached this question by creating a mouse model in which the timing and severity of disease progression could be manipulated. Such a model not only could contribute to the understanding of disease pathogenesis, but also could be useful in evaluating the possible efficacy of novel therapies. Dr. Servadio's team used the "Tet-off" system to create an inducible model of SCA1. They put expression of recombinant ataxin-1 under control of the synthetic tTA transactivator. In this system, giving tetracycline to the mice would inactivate the pathogenic ataxia-1 gene, while withdrawing the antibiotic would allow disease to progress. The researchers also refined their model by breeding transgenic mouse strains that enabled them to regulate expression of mutated ataxin-1 in specific brain regions. Unfettered ataxin-1 expression resulted in smaller litters, reduced body size, cranio-facial dysmorphology, and performance deficits in newborn mice. None of these defects were observed when ataxin-1 transgene was repressed by continuous tetracycline administration. On rare occasions, some birth abnormalities could be reversed by administering tetracycline and shutting down ataxin-1. In mice that were several months old, SCA1 could also be induced by withdrawing tetracycline and allowing ataxin-1 to flourish. These tools, which enable scientists to manipulate the onset, severity, and progression of symptoms, are expected to be widely used in the study of this disease. Dr. Servadio's lab created these models with support from Telethon (Italy) and from the National Ataxia Foundation (Minneapolis, MN).
One of Darwin's greatest insights was that each individual cell, and each living organism, constantly strives to leave behind the maximum number of progeny. Cancer cells, unfortunately for patients, proliferate better than most. And as Dr. Leder suggested in the symposium's opening talk, many different genes play a role in runaway cell growth. New technology such as RNA and DNA microarrays now enable scientists to "fingerprint" pathological states such as cancer, and to determine the full spectrum of genes that are involved in cell proliferation and its regulation. Dr. Church and his colleagues at the Lipper Center for Computational Genetics are also developing functional genomics methods that enable them to tell not only which genes are expressed, but also to quantify expression and determine when it occurs in the cell cycle. When Dr. Church's group used DNA microarrays to study gene expression in cancer cells, they found that the cells have only small growth advantages at first, but over time these are exponentially amplified to optimize their growth. Because DNA must make RNA in order to generate proteins, they used RNA microarrays to determine when specific genes were transcribed and in what quantity. Whether these transcription patterns can be observed in vivo is an important question, of course, and the researchers are using animal models to find out. They're using mass spectrometry to track in vivo protein expression. In the post-genome era, Dr. Church believes that genetic fingerprints of different tumors will lead to more precise diagnosis and treatment. Treating all cancers the same, as clinicians know, results in toxicity and failure to respond. The ability to profile individual malignancies could change that.
Dr. Alcalay's laboratory is one of several at her institution with a persistent interest in chromosomal translocations in acute myeloid leukemia (AML). Recently, her group has been experimenting with various high-throughput DNA screens as a means for identifying the full spectrum of biological activities associated with such translocations. In particular, the researchers have looked for overlapping gene expression patterns for the AML1/ETO fusion protein of M2-AML and the RARa-fusion proteins associated with acute promyelocytic leukemia (APL). At the outset, several similarities were known:
Dr. Alcalay and her colleagues used nylon filter arrays containing 18,376 human cDNAs for a preliminary analysis of gene expression patterns of U937 hematopoietic precursor cells. These cells expressed either the PML/RAR or the AML1/ETO fusion proteins, and the researchers exposed them to substances that stimulate differentiation. This filtering method yielded 1,294 putative target genes. This technique was not very discriminating, however, and additional tests showed that most were not true targets. More exacting Northern blot studies winnowed the candidates down to 12 non-regulated and 64 regulated target genes. Of the 64 regulated genes, half are common targets for PML/RAR and AML1/ETO and half of those are repressed by both fusion proteins. The repressed genes are induced during differentiation. A separate series of experiments with Affymetrix DNA chips identified several hundred genes that are repressed by the products of two or more fusion proteins, pointing to numerous common pathways in acute myeloid leukemogenesis. In AML mice, retinoic acid therapy induced expression of some of these same genes, which Dr. Alcalay said confirms their importance. Symposium Pages
|
Cancer: An Unfortunate Genetic Collaboration
