Expanded Overview of Our Cancer Research Program
Our development of novel treatments for GBM, which are suitable for rapid transition to clinical analysis, was achieved via the following discoveries. This discovery path demonstrates how each component of our overall strategy is integrated to generate novel medical insights – and also demonstrates the potential importance of beginning the search for new cancer therapies by better understanding the normal progenitor cells from which tumors may arise and understanding the ways in which tumor cells escape existing treatments. Read a detailed review.
Step 1: Intracellular Redox State is Critical in Controlling Progenitor Cell Function and in the Response to Signaling Molecules
Our work begins with understanding fundamental principles in cell regulation, and there is no more fundamental biological principle than electron transfer, for without electron transfer no chemical reactions can occur. To oversimplify a bit, the biology of electron transfer is the basis for what we call the biology of reduction and oxidation (which is referred to as redox biology). This is a topic of broad interest due to the importance of changes in redox regulation in normal development and in pathology. Our own interests in this arena began with the first demonstrations that anti-oxidants can not only provide protection against physiological stressors but also can enable cell survival in the presence of suboptimal concentrations of such trophic factors as nerve growth factor (NGF) and insulin-like growth factor-I (IGF-I) (Mayer and Noble, 1994).
Our studies on glial progenitor cells next provided the first demonstrations that critical precursor cell functions are regulated by intracellular redox state (Smith et al, 2000; Power et al, 2002). We discovered that small changes in redox state modified the balance between division and cell-cycle exit/differentiation, such that more reduced cells undergo more division and more oxidized cells are more likely to cease division and initiate differentiation. We also found that cells isolated from regions of the CNS that are myelinated later in life and are larger (and thus require more oligodendrocytes) are more reduced than those found in regions that are myelinated earlier and require fewer oligodendrocytes.
We further found that signaling molecules, such as thyroid hormone and neurotrophin-3 (NT-3) need to alter redox state in order to modulate the balance between division and cell-cycle exit. This work identified a new essential component of action of well-studied signaling molecules. Present studies address the question of how different aspects of signaling molecule function are integrated in cellular regulation.
Our studies on redox biology triggered a broad and growing interest in metabolic regulation of stem and progenitor cell biology. Multiple independent lines of evidence demonstrate the critical importance of redox state in modulating progenitor cell function.
Step 2: The redox/Fyn/c-Cbl (RFC) Pathway Links Oxidative Changes to Control of Critical Progenitor Cell Functions
Once we knew that redox state modified progenitor cell function, we next set to discover the molecular mechanism(s) responsible for converting redox changes into alterations in cell function.
Our research identified new regulatory pathway that converts small increases in oxidative status into accelerated degradation of multiple proteins that are critical in controlling function of stem cells and progenitor cells (Li et al, 2007). Oxidation of O-2A/OPCs causes activation of Fyn kinase, which then activates the c-Cbl ubiquitin ligase. Activation of c-Cbl enables it to attach ubiquitin to its targets, which include such critically important proteins as the epidermal growth factor receptor (EGFR), the platelet-derived growth factor receptor-α (PDGFRα), the c-Met receptor for hepatocyte growth factor, Notch-1 and beta-catenin. It does not appear to matter whether progenitor cells are oxidized by exposure to signaling molecules, environmental toxicants, chemotherapeutic agents or genetic/epigenetic regulation of redox status – all lead to activation of the RFC pathway.
Step 3: Identifying the Cellular Foundations of Chemobrain
The central dilemma of cancer research is that killing cancer cells is easy, but killing cancer cells without killing normal cells is not. This is a particularly challenge to address, because many of the proteins and molecular mechanisms of central importance in cancer cells are also of central importance in the function of normal cells, and particularly of stem and progenitor cells. Thus, it is not surprising that most current anticancer therapies can cause unacceptable levels of damage in normal tissue.
The CNS provides one of the most important examples of the problems that can be caused by current cancer treatments. Systemic treatment with any of a large variety of chemotherapeutic agents can cause cognitive changes, alterations in motor function and changes in CNS structure in an unacceptably high proportion of patients.
Remarkably, the damage done to the CNS by cancer treatments had been virtually ignored for many years by neuroscientists, stem cell scientists and cancer researchers. At the time we become interested in this problem (in 2005) it was possible to find more laboratories studying a genetic disease that affects only a few thousand people world wide than were studying the problems caused by treatments that are applied to millions of cancer patients each year.
We therefore targeted our attention on understanding the biological foundations underlying the adverse neurological effects of treatment with systemic chemotherapy. We found that the progenitor cells and oligodendrocytes of the CNS were more vulnerable to the effects of multiple chemotherapeutic agents than were human cancer cells (Dietrich et al, 2006; Han et al, 2008). Such toxicities occurred rapidly after treatment and often were durable in their effects. Moreover, when we studied the delayed effects of systemic treatment with chemotherapy, we found that the damage seen weeks or months after the cessation of treatment was even worse than that seen while treatment was ongoing (Han et al, 2008).
Happily, our work had profound effects on the research community. This research was reported in newspapers and in scientific and medical websites around the world, spurring great interest in this problem. Multiple laboratories are now engaged in attempts to understand why adverse neurological effects occur and how they can be treated – or how they can be prevented (a problem that we also have addressed using our strategies of discovering new uses of existing compounds).
Step 4: Selectively Attacking GBMs, Eliminating Cancer Stem Cells and Simultaneously Attacking Multiple Critical Cancer Control Nodes by Restoring Normal Function of the RFC Pathway in Cancer Cells
If normal progenitors are vulnerable to existing anticancer agents but cancer cells are not, then it is essential to understand why this is so and to develop treatments that kill cancer cells more effectively than normal cells.
We were particularly interested in using activation of the RFC pathway to attack cancer cells (due to the multiple proteins critical in cancer stem cell function that are controlled by c-Cbl), but first had to confront the problem that GBM cells are not impaired by oxidation. For example, one of the central regulators of redox balance is the ratio between reduced glutathione (GSH) and oxidized glutathione (GSSG). In primary cells, this ratio is normally at least 50: 1 in favor or GSH – but we found that in GBM cells even increasing GSSG content to where the ratio is <3:1 (an oxidation status that normal cells cannot survive) had no effect on the GBM cells.
We found that GBM cells escape the consequences of oxidation by inhibiting RFC pathway function and c-Cbl activation with a protein (called Cool-1/ß-pix) (Stevens et al, 2014). In GBM cell lines and tumor biopsies – but not in normal brain or normal glial progenitor cells – c-Cbl is found in a complex with Cool-1/ß-pix. Disruption of this inhibitory complex by Cool-1 knockdown enhanced sensitivity to chemotherapeutic agents and decreased division.
Perhaps most importantly, restoration of normal c-Cbl function by Cool-1 knockdown eliminated GBM stem cells, such that the resultant cells could not make a tumor unless they re-expressed Cool-1. Moreover, restoration of normal c-Cbl function enabled us to simultaneously decrease levels of multiple proteins considered by others to be potential therapeutic targets, including EGFR, Notch-1, ß-catenin, CD133 and Sox-2. Restoration of c-Cbl function enables us to attack all of these targets with a single intervention.
Step 5: Determining Whether Pharmacological Restoration of c-Cbl Function is a Viable Therapeutic Approach
Restoration of normal c-Cbl function is an attractive therapeutic target for multiple reasons. The ability to eliminate GBM stem cells, to simultaneously attack multiple proteins critical in GBM cell function, and to selectively increase the toxicity of therapeutic agents for GBM cells but not for normal CNS progenitor cells are already sufficient reasons to be interested in this approach. But there are also multiple other reasons to be interested in this pathway:
Application to other cancers: To determine whether other cancers show a similar dependence on inhibition of c-Cbl function by production of inhibitory proteins, we next examined basal-like breast cancers, the most difficult breast cancers to treat. The category of basal-like breast cancers contains multiple sub-types, based on their molecular properties, but all share the property of being resistant to the major therapies used in successfully treating breast cancer.
We found that basal-like breast cancer cells also inhibit c-Cbl function, and that restoring c-Cbl function similarly provides a means of attacking the cancer stem cells that also are important in these cancers.
Suitability for pharmacological targeting: It has long been known that some lymphoid tumors can harbor c-Cbl mutations that convert this protein from a tumor suppressor gene to an oncogene. Restoring c-Cbl function in a mutated protein is a difficult problem – but if one can restore c-Cbl function by attacking an inhibitory protein (particularly an inhibitory protein that is not mutated) this becomes very attractive for pharmacological interventions.
Our work on basal-like breast cancers provided the first demonstration of the potential potency of pharmacological strategies for restoring c-Cbl function. We used pharmacological inhibition of Cdc42, which plays a key role in c-Cbl inhibition in basal-like breast cancer cells to examine this theoretically attractive therapeutic approach. Treatment of cells with this inhibitor also eliminated cancer stem cells. Moreover, even though this inhibitor was not optimized for in vivo use, it had sufficient potency to demonstrate that initiation of treatment after tumors were established offers a promising therapeutic approach (Chen et al, 2013).
Overcoming the challenge of tumor heterogeneity: One of the attractive features of restoring RFC pathway regulation of c-Cbl is that it may enable us overcome one of the most difficult obstacles to cancer treatment, that of tumor micro-heterogeneity. It is not just that tumors in different patients differ greatly in terms of the specific mutations acquired and the genetic background in which they operate. More problematic, as tumors become more malignant their internal heterogeneity (generally referred to as micro-heterogeneity) increases. By the time a tumor is malignant, cancer patients do not harbor a single disease entity. Instead, they harbor dozens or more (and perhaps many more) molecularly different cancers.
Micro-heterogeneity is one of several difficult challenges to the hopes that personalized medicine will revolutionize cancer treatment. More accurate administration of treatments, in a manner that mirrors the individual features of injury or disease in each individual, has long been a goal of medical practice. This approach enables existing therapies to be employed with greater precision. Whether it is also a strategy for discovering novel therapies, however, is less clear. Aside from the potentially prohibitive costs, too much of the current thinking about personalized medicine is dependent on various types of genetic analysis, and not sufficiently on analysis of protein function and metabolic regulation – which means that we are not spending enough time on the aspects of cells that actually make them work.
As an alternative to changes that may be present only in subpopulations of tumor cells, we found c-Cbl inhibition in multiple independent GBM populations, in every GBM biopsy specimen we examined and in multiple independent basal-like breast cancer populations (Chen et al, 2013; Stevens et al, 2014). These tumor populations come from different people, with different genetic backgrounds, and each would have had their own unique history of mutational changes before they became cancer cells. Yet they all share a dependence on inhibiting c-Cbl. While there may be differences in the exact means by which c-Cbl is inhibited in different cancers, there appear to be only a small number of means of achieving this outcome. Thus, rather than the hyper-personalized cancer treatments being pursued by others, our strategy appears likely to have broad applicability to multiple different people.
Rationally enhancing the utility of other treatments: The complexity of cancer means that the odds of a single agent treating cancer are infinitesimally small. Therefore what is essential is the development of rational combinatorial strategies. As co-application of two experimental therapies in a clinical trial setting is not attractive to regulatory agencies, and has other problems, it is critical to determine optimal combinations between new experimental treatment and the multiple existing therapeutic agents.
In our studies on basal-like breast cancer we demonstrated that we could harness pro-oxidative and estrogen receptor-a independent activities of tamoxifen as a means of increasing c-Cbl activation via the RFC pathway. While tamoxifen or the Cdc42 inhibitor we used provided no benefits in respect to controlling tumor growth in vivo when applied individually, the combination of the two provided the ability to control growth basal-like breast cancer cells in vivo with tamoxifen.
Step 6: Drug Discovery
To turn our discoveries into therapies we need to identify pharmacological agents that can re-establish c-Cbl function. Moreover, as the average lifespan for a GBM patient from time of diagnosis is about 15 months, it is necessary to discover drugs that can move quickly into clinical analysis.
We have discovered multiple FDA-approved drugs have the unexpected activity of restoring normal c-Cbl function in GBMs and also in other types of cancer. These drugs, used in rational combinations with existing therapies, appear to allow virtually complete control of GBM growth in mouse models of aggressive human tumors. Moreover, these approaches are effective at eliminating cancer stem cells. Details of our studies on drug discovery will be added at a later time.