2009 Innovative Research Grant 36-Month Progress Reports
An Emerging Tumor Suppressor Pathway in Human Cancer
Fernando D. Camargo, Ph.D., Children’s Hospital Boston
Hippo is a novel biochemical pathway that can regulate organ growth. It seems to work by preventing further cell division once organs have reached their proper size. Our group believes that Hippo signals participate in a novel and powerful “checkpoint” that restricts cell proliferation and activates cell death. Such a checkpoint would normally enable an organ to “know” its size and stop growing when the appropriate dimensions are achieved. A checkpoint mechanism would also serve to suppress the unchecked cell proliferation that characterizes tumor growth. Our proposal aims to study the existence of this checkpoint and to examine its role in tumor growth. In Aim 1, we proposed to test this hypothesis by using genetically engineered mice in which Hippo signaling can be turned on or off. For Aim 2, we proposed to identify novel proteins and small molecules that can modulate Hippo signaling in cultured cells and that could eventually be used to develop therapies for cancer.
For Specific Aim 1, we have found that inactivation of the Hippo pathway in the skin of mice leads to an expansion of the stem cells that normally form and maintain the skin, and this eventually leads to the development of invasive squamous cell carcinomas. We have also found that a specific sub-class of human squamous cell carcinomas might be driven by alterations in this pathway, indicating that these tumors might benefit directly from therapies targeting components of this pathway. Along these same lines, we have made similar observations in skeletal muscle, where inactivation of the Hippo pathway leads to the amplification of the muscle-specific stem cells, and the subsequent development of invasive rhabdomyosarcomas. We have validated these observations in human cells and tumors suggesting that Hippo might be an important pathway to target in soft-tissue targets. Additional results indicate that the Hippo pathway is an important suppressor of tumor growth in the small intestine and the colon. Our discoveries to date highlight a previously unappreciated global role for Hippo as an important cancer preventing checkpoint in mammals.
We have also demonstrated using mouse models that re-activation of the pathway can lead to an important suppression of growth in a model of basal cell carcinoma and hepatocellular carcinoma. Conversely, in the intestine over-activation of the pathway might not be beneficial and might even be harmful.
For Specific Aim 2, we have finalized a genetic screen to identify kinases or phosphatases (specific subset of proteins in the human genome) that can regulate Hippo signaling. We have identified approximately 40 molecules that can have important effects on Hippo activity. Among these, we have discovered an important connection between Lkb1, a very common gene altered in human cancer, and the Hippo pathway. This means that human cancers with mutations in Lkb1 could be amenable to treatment by manipulating the Hippo pathway. Additionally, another molecule identified in this screen, alpha-catenin, is a known tumor suppressor in the skin and other epithelial tissues. How alpha-catenin prevented tumor growth was unclear, our results suggest that alpha-catenin inhibits growth by directly controlling the activity of the Hippo pathway. Thus, our results would argue that these alpha-catenin mutant tumors would benefit from therapies inhibiting Hippo signaling.
Targeted Inhibition of BCL6 for Leukemia Stem Cell Eradication
Markus Müschen, M.D., Children’s Hospital Los Angeles
Despite significant advances in the treatment of leukemia over the past four decades, the rate of long-term survival has reached a plateau and still large numbers of leukemia patients die, mostly because of relapse and drug-resistance. These two clinical problems were recently attributed to the persistence of leukemia stem cells. If a therapy succeeds in eradicating leukemia stem cells, renewed initiation of the disease (relapse) is no longer possible. Therapeutic progress in recent clinical trials has likely been stalled, partly because current chemotherapy approaches target proliferating bulk leukemia cells rather than non-dividing leukemia stem cells. We now discovered that BCL6, a factor known to play a central role in lymphomas, also plays a key role in the maintenance of leukemia stem cells. Since leukemia stem cells represent the origin of relapse and drug-resistance in leukemia in many cases, the identification of BCL6 as a target for leukemia stem cell eradication holds great promise. BCL6 is a master regulatory factor that controls the production of many different important genes. BCL6 was not previously known to be involved in leukemia. In preliminary studies for this proposal, we have discovered aberrant expression of BCL6 as a central component of a fundamentally novel pathway of leukemia stem cell self-renewal and drug-resistance in a wide array of human leukemias, some of which are still difficult to treat. In these leukemias, drug-treatment results in aberrant production of BCL6 by the leukemia cells, which appears to allow leukemia stem cell to self-renew and become resistant against chemotherapy. Recently a drug has been developed that can attach to BCL6 and block its cancer-causing activities. We found that this BCL6 inhibitor, called RI-BPI, has strong cooperative activity when combined with conventional chemotherapy. This opens up a powerful new therapeutic strategy for leukemia stem cell eradication through targeted inhibition of BCL6. Based on the discovery of BCL6 as a key component of a novel pathway of drug-resistance and stem cell self-renewal in a wide array of leukemias, we propose three Aims to develop these findings towards application in patient care: (1) To test the hypothesis that BCL6 is critical for leukemia-initiation and relapse of leukemia, (2) To determine the frequency and appearance of BCL6-dependent leukemia stem cells in human leukemia samples and (3) To validate the role of the BCL6 inhibitor RI-BPI as a novel therapeutic agent for targeted eradication of leukemia stem cells. Since RI-BPI is currently going through the process of approval for use in clinical trials, we expect to be able to test the power of this approach in clinical trials by the end of the funding period.
Therapeutically Targeting the Epigenome in Aggressive Pediatric Cancers
Charles Roberts, M.D., Ph.D., Dana-Farber Cancer Institute
Recently there has been a growing realization that some of the critical changes in gene expression required for the development of cancer do not arise via genetic mutations in DNA but rather are ‘epigenetic’ changes that affect gene expression indirectly by affecting DNA packaging. The SWI/SNF complex controls the protein support structure that surrounds key growth genes and is thus at the heart of this epigenetic regulation. Mutations in SNF5, a core subunit of the SWI/SNF complex, are present in the large majority of malignant rhabdoid tumors (MRT), a highly lethal cancer that occurs in kidney, brain, and soft tissues of young children. Inactivating mutations in SNF5 have recently been found to occur in a variety of other cancers as well including epithelioid sarcomas, small cell hepatoblastomas, undifferentiated sarcomas, chondrosarcomas, familial schwannomatosis, and renal medullary carcinomas. Mutation in SNF5 is also the basis of an inherited cancer predisposition syndrome. As the cancers that arise following Snf5 loss appear to be largely driven by the epigenetic consequences, we hypothesize that these cancers will be particularly susceptible to drugs that interfere with epigenetic mechanisms of gene regulation. The experiments in this proposal were designed to reveal the underlying mechanisms by which SNF5 loss affects gene expression and thereby causes cancer with a goal of identifying improved therapies that can be rapidly translated into patients. Of note, during the time of SU2C grant support, cancer genome sequencing studies have revealed remarkable findings: in addition to SNF5, at least 6 other SWI/SNF subunits are frequently and recurrently mutated in a wide variety of cancers, both pediatric and adult, including cancers of lung, breast, stomach, liver, ovary, uterus, kidney, bladder, brain and melanoma. Consequently, the work performed during this IRG now has broad implications for many types of cancer.
Over the course of the IRG support, we exceeded our expectations. While not all of our hypotheses proved correct, we made outstanding progress on others. Indeed, the first clinical trial based upon our work is now open at 6 centers nationwide, and several in Europe. In addition, we anticipate opening of a second clinical trial in 2014 directly based upon work that was accomplished via IRG support. With respect to the specific aims of this proposal, based upon identifying over-expression of miR-21 in SNF5 deficient cancers, we had hypothesized that inactivation of this micro RNA might reverse the growth of Snf5-deficient cancers. This proved incorrect – it had no effect. With respect to changes in histone modifications caused by Snf5 inactivation, we identified the existence of epigenetic antagonism between SNF5 and EZH2, a member of a different chromatin modifying complex. This suggested that inhibiting EZH2 might stop the growth of SNF5 mutant cancers and, indeed, that is precisely what we found in our pre-clinical models, and we’re hopeful that this will be headed into the clinics for patients with rhabdoid tumors in 2014. With respect to or DNA methylation hypothesis, we’re still working on obtaining data to determine the extent to which Snf5 loss affects DNA methylation, and whether this will be a good target. However, we have performed trials of DNA methylation in mice bearing Snf5-deficient tumors. Bottom line: work from my lab has now led to a new, first-time mechanistically targeted clinical trial that is now open, and a new trial directly derived from the SU2C work slated to open in 2014.
Endogenous Small Molecules that Regulate Signaling Pathways in Cancer Cells
Rajat Rohatgi, M.D., Ph.D., Stanford University School of Medicine
A major goal in cancer biology is the comprehensive understanding of signals that drive the growth and spread of cancer cells. My long-term goal is to develop methods to isolate new small molecules that play a role in cancer signaling and then to identify the proteins that interact with these small molecules. Such small molecule-protein pairs are likely to be particularly good drug targets in oncology. To develop tools for this endeavor, we are focusing on the identification of small molecules that regulate the “Hedgehog” signaling circuit. Damage to this circuit has been shown to drive the development of a large number of adult and childhood cancers. Major progress over the course of this grant, now published in Nature Chemical Biology, has been the unexpected discovery that cholesterol-like small molecules called oxysterols can directly influence the protein Smoothened, a cancer-driving protein that is the major drug target in this pathway. This discovery of the mechanism by which these cholesterol-like molecules influence a cancersignaling pathway was a major goal in our initial proposal. Importantly, we have shown that this mechanism can be exploited to develop a new class of anti-Hedgehog drugs based on a cholesterol scaffold. We believe that this discovery supports our initial IRG hypothesis that endogenous small molecules can have dramatic effects in the activities of cancer-relevant proteins, and we are now actively engaged in discovering such regulatory interactions and establishing their therapeutic relevance.
Genetic Approaches for the Next Generation of Breast Cancer Tailored Therapies
José M. Silva, Ph.D., Columbia University Medical Center
Cancer therapy has radically changed during the last decade. Novel therapies based on the specific molecular changes that drive tumorigenesis in every patient are emerging as low toxicity and more efficient alternatives to classical treatments. An alternative promising approach for the design of these personalized therapies is the use of genetic synthetic lethal interactions. These occur when two genetic alterations that are individually innocuous appear in the same cell, causing growth inhibition. This concept can be exploited to identify genes that, when inhibited, exclusively reduce the viability of tumor cells that carry a preexisting genetic lesion.
Recently, RNA interference (RNAi) technology has emerged as a very powerful approach to attenuate the expression of any chosen gene. Thus, we envision using RNAi to identify genes that, when attenuated, exclusively reduce the viability of tumor cells carrying specific genetic lesions without affecting normal cells. During the last years, my group has pioneered the development of RNAi based genetic tools for studies in mammalian cells. This technology represents a unique opportunity to identify synthetic lethal effects with major cancer alterations. In this project we proposed to apply our state-of-the-art technology to uncover synthetic lethal interactions with the major breast cancer genes.
Our proposal is divided into three specific aims that represent the transition from target discovery and validation to mechanistic characterization.
-Specific Aim 1. Identify genes that interact with the major breast cancer alterations to produce synthetic lethality in vitro (1st year): During the first year of this project, we proposed to complete four genome-wide RNAi screens in vitro to identify target genes that, upon inhibition, reduce the cell viability in breast cancer cells with any of the major breast cancer alterations; ErbB2, c-Myc, Cyclin-D or RB. In our pre-defined milestones we estimated that two of the four RNAi screens would be completed during the first six months and the rest during the second semester of the first year. At this point, the four genome-wide RNAi screens have been completed and analyzed.
-Specific Aim 2. Model selected lethal interactions in vivo (1st and 2nd year): Completion of the above mentioned screens has provided us with a list of candidates that were further validated in vivo by candidate driven RNAi screens in mouse models during the second year of the award.
-Specific Aim 3.Initial characterization of the molecular mechanism of the genetic lethality (2nd and 3nd year): Upon completion of Aim 2, we selected the most promising (2-3) target (inhibiton of STAT3 in ErbB2+ tumors) to investigate the biology of the lethal phenotype in more detail.
Modulating Transcription Factor Abnormalities in Pediatric Cancer
Kimberly Stegmaier, M.D., Dana-Farber Cancer Institute, Children’s Hospital Boston, and the Broad Institute of Harvard and MIT
There is an urgent unmet need for more efficacious therapies for childhood cancers. Many of the cancerpromoting, tumor-specific proteins in these pediatric cancers, however, have been considered “undruggable” by traditional drug discovery approaches. One class of challenging cancer-promoting proteins is proteins that bind to DNA called transcription factors. Our laboratory developed new chemical genomic approaches to target these elusive proteins. We applied these approaches to two pediatric malignancies: Ewing sarcoma, the second most common cause of bone cancer in children, and neuroblastoma, the most common extracranial solid tumor of childhood, both diseases characterized by the expression of cancer-promoting transcription factors, and diseases where treatment for patients with high-risk tumors remains poor.
In Ewing sarcoma, the majority of tumors express the cancer-promoting transcription factor, EWS/FLI. We developed an alternative to traditional drug discovery using DNA microarrays (“gene chips”) to characterize the genes that are turned on or off in the presence or the absence of the Ewing sarcoma protein (Gene Expressionbed High-throughput Screening (GE-HTS)). We completed a screen of over 10,000 chemicals, prioritized 160 top scorers, and focused our attention on several molecules already FDA-approved in humans or in clinical development. These compounds now serve as tools to further our understanding of Ewing sarcoma development and as leads toward clinical trial development. Moreover, we have “platformized” GE-HTS at the Broad Institute to enable access to many investigators. In parallel to these efforts, we also conducted a screen to identify more druggable targets in this disease. We identified the protein focal adhesion kinase (FAK) as highly activated in Ewing sarcoma tumors. Treatment of Ewing sarcoma cells with a FAK-inhibitory drug, PF-562271, impaired viability and induced cell death. Additionally, PF-562271 attenuated Ewing sarcoma growth in mouse xenograft models. With FAK inhibitors currently in clinical trials for adult malignancies, these findings may bear immediate relevance to patients with Ewing sarcoma.
A second approach taken by my laboratory is large-scale screening of genetically-defined cancer cell lines for response to specific small molecules of interest. We collaboratively screened a panel of over 600 geneticallycharacterized cancer cell lines for response to a new class of epigenetic-modifying drugs called BET bromodomain inhibitors. BET bromodomain inhibitors have been studied in a handful of discrete malignancies, but genomic biomarkers to direct clinical translation have been lacking. In our study, integration of genetic features with chemosensitivity data demonstrated a robust correlation between amplification of MYCN, a gene encoding for a transcription factor, and sensitivity to bromodomain inhibition. We have characterized the mechanistic and translational significance of this finding in neuroblastoma, a childhood cancer with frequent MYCN amplification. Genome-wide expression analysis demonstrated downregulation of the MYCN transcriptional program and the expression of the MYCN transcript. Functionally, bromodomain-mediated inhibition of MYCN induced cell death in neuroblastoma cells. BET inhibition conferred a significant survival advantage in three mouse models of neuroblastoma. A clinical trial to test BET bromodomain inhibitors in patients with neuroblastoma is now in development.
Noninvasive Molecular Profiling of Cancer via Tumor-Derived Microparticles
Muneesh Tewari, M.D., Ph.D., Fred Hutchinson Cancer Research Center
Obtaining molecular information about an individual cancer patient’s tumor would allow for the development of patient-specific treatment plans. However, obtaining tumor biopsy samples in many cases requires invasive surgery which can be painful, disfiguring and potentially dangerous. Consequently, alternative non-invasive sampling methods are needed which could serve as a reliable surrogate for actual tumor tissue. The approach we are pursuing takes advantage of the fact that cancer cells release information into the bloodstream. This information is packaged into what are called tumor-derived microparticles; essentially small parcels derived from the contents of the tumor cells. The goal of this project is to develop methods to efficiently capture these particles from patient blood samples and decode the information within them in order to gain molecular information about the cancer cells from which they originated. Over the course of this project we developed methods for the capture and evaluation of a number of different particles types. In doing so, we were able to focus our efforts on a specific type of microparticle known as as an exosome. We have developed methods to specifically and efficiently capture an purify exosomes from clinical samples. Molecular characterization of the contents of these capture particles is in progress.
Furthermore, we also recently discovered that although exosomes do contain some of the information we seek, much of this information is in fact present is a larger, mid-sized class of particles that we are currently characterizing. Consequently, we are now engaged in investigating the nature of these particles using different techniques that will allow us to determine their size, as well as to identify molecules on their surface that will permit us to capture them from blood samples. Taken together, we believe that these various particle types will provide molecular information derived from the tumor and serve as a tumor surrogate. Our ultimate goal is to obtain key information using a blood sample that can inform critical clinical decisions such as choosing the most effective drug for an individual patient.
A Transformative Technology to Capture and Drug New Cancer Targets
Loren D. Walensky, M.D., Ph.D., Dana-Farber Cancer Institute
The goal of my SU2C project was to create a powerful, new, and versatile approach to identifying and drugging new cancer targets. To that end, we combined a chemical technology termed “hydrocarbon stapling” that restores natural shape to bioactive peptide alpha-helices with a protein capture technology in order to trap and characterize critical targets involved in defective signaling in cancer. With the first year of SU2C funding, we achieved our goal of chemically synthesizing a pilot panel of “photoreactive stabilized alpha-helices” or pSAHs, which are the chemical tools designed to capture and characterize new cancer targets. We examined and optimized the sensitivity and specificity of these new tool compounds for crosslinking to discrete physiologic targets of the BCL-2 pathway, a key signaling network implicated in cancer pathogenesis and chemoresistance. This SU2C sponsored proof-of-concept study was published as a cover article in Cell’s Chemistry and Biology in December of 2010. In this publication, we demonstrated our capacity to successfully and reproducibly generate pSAHs that recapitulate the structure of distinct bioactive domains and deploy them to trap, purify, and identify their natural cellular targets with high fidelity. In addition, we reported a rapid and reliable method for inputing our crosslinking data into a binding site algorithm that employs mass spectrometry and computational docking analysis to calculate model structures of the key-in-lock binding interfaces we discovered. This critical information provided the basis for validating new proteinprotein interactions for drug development and therapeutic targeting. Having defined and published the “rules” for successful production and application of pSAHs by the end of year 1, we dedicated year 2 support to expanding our arsenal of pSAH constructs in order to home in on key protein interactions that drive cancer. In doing so, we ultimately synthesized a library of over 80 photoreactive helices spanning the death domains of 10 seminal BCL-2 family apoptosis proteins implicated in oncogenesis and the response to chemotherapy. In applying these unique reagents during the final year of the grant, we characterized novel interaction sites at the atomic level, identified unanticipated interactors, and developed a versatile chemical toolbox for validating and modulating these critical protein interactions in cancer cells. In addition to the development and deployment of the technology itself, our flagship achievement from the grant period was the discovery and characterization of novel interaction surfaces to “inhibit the inhibitors” and “activate the activators” of cell death in resistant human cancers. These blueprints provided fundamental new directions for targeted drug development in cancer, and have already catalyzed the advancement of novel stapled peptide and small molecule modulators that reactivate cancer cell death.
Functional Oncogene Identification
David M. Weinstock, M.D., Dana-Farber Cancer Institute
Despite advances in diagnosis and treatment, more than one-half of adults with cancers of the blood (i.e., leukemia, lymphoma and multiple myeloma) will die from their disease. One of the limitations in our current approach is that most cancer chemotherapy does not target abnormalities unique to the tumor cells, but instead kills all growing cells. Thus, the identification of specific cancer-associated abnormalities is an essential first step toward newer and more effective therapies. We developed a system to identify new targets for therapy directly from leukemia and lymphoma samples. Briefly, we isolate the many millions of pieces of genetic material from a tumor sample and then individually insert each into cells that can only grow in a special kind of culture. If one of the pieces of genetic material has a cancer-promoting effect, it allows the cells to grow in normal culture. Thus, any cell that survives in the normal culture must contain a piece of genetic material from the tumor that has a cancer-promoting effect. We can easily identify that piece of genetic material and then confirm that it is important for the tumor’s growth. The system we developed is efficient and can be scaled up to analyze a large number of individual specimens. Using this approach, we have already discovered a new cancer protein called CRLF2 in some cases of acute lymphocytic leukemia. The overall goal of our Stand Up To Cancer Innovative Research Grant proposal is to identify important alterations that promote the growth of other types of blood cancer. During the funding period, we utilized our approach to screen 16 types of human leukemia and lymphoma samples for new mutations. From this screen and from sequencing large amounts of DNA in the specimens, we identified multiple mutations that have never been described from tumor specimens. Of particular interest, we identified mutated versions of proteins that can be targeted with available drugs. We are in the process of confirming that the mutations we identified contribute to tumor growth. We are also defining the frequency of these mutations in other leukemia and lymphoma specimens. These studies have led directly to additional funding from the Leukemia and Lymphoma Society, Claudia Adams Barr Program in Cancer Research and Dana-Farber/Novartis Drug Discovery Program, with the goal of identifying new treatment approaches that target these alterations.