Anthony G. Letai - US grants

[unreadable] DESCRIPTION (provided by applicant): Cancer cells frequently and perhaps invariably possess aberrations in the genetic pathway of programmed cell death. The BCL-2 family of proteins plays a critical role in the signaling and execution of death signals. A common mechanism by which cancer cells evade programmed cell death is by altering the ratio of antiapoptotic/ pro-apoptotic BCL-2 members, resulting in a functional excess of BCL-2 antiapoptotic molecules. This proposal describes strategies to test the hypotheses that (a) anti-apoptotic BCL-2 family members can be specifically targeted, and (b) BCL-2 is a valid target for anti-cancer therapy. To address (a), binding interactions between anti-apoptotic BCL-2 family members and BH3 domains from "BH3-only" family members will be examined using fluorescence polarization to determine the specificity of interaction between the two classes of proteins. The relevance of the binding interactions will be validated using functional mitochondrial and cellular assays of apoptosis. From these studies will emerge oligo-peptides which function as prototype inhibitors of anti-apoptotic BCL-2 family members. This approach has already shown promise in the preliminary characterization of a peptide inhibitor of BCL-2 based on the BH3 domain of BAD. To address (b), a mouse model of leukemia which is dependent on BCL-2 for maintenance is described. In this model, mice develop a lymphoid leukemia which remits when expression of BCL-2 is eliminated by treatment with doxycycline. To test the importance of BCL-2 in tumor maintenance in other tissue types, the conditional expression of BCL-2 will be extended to other tissues such as breast and melanocytes. Methods are described for the promotion of cancer in these models. When cancer develops, BCL-2 expression will be controlled by doxycycline to determine if BCL-2 is required for tumor maintenance. Finally, we will use the mitochondrial, cellular and mouse assays developed above to test the efficacy and specificity of BH3 mimetics of both small molecule and peptidomimetic origin. [unreadable] [unreadable]

[unreadable] DESCRIPTION (provided by applicant): It has been hypothesized that cancer cells require a block in apoptosis for survival, due to their numerous phenotype irregularities. One potential mechanism cancer cells might exploit to maintain survival is the expression of antiapoptotic proteins of the BCL-2 family. We have developed a novel method, called BH3 profiling, that detects dependence on antiapoptotic proteins. It is particularly useful as it can be used to study freshly isolated primary cancer tissues without need for further culture ex vivo. We will use this and other more standard techniques to study regulation of apoptosis in several systems. First, we will investigate whether chronic lymphocytic leukemia (CLL) cells are dependent on BCL-2 for survival. We will also examine whether CLL cells are sensitive to treatment with a novel BCL-2 antagonist, ABT-737. We will furthermore determine the molecular events underpinning this sensitivity. Next, we will investigate acquired resistance to BCL-2 antagonism. We have identified a panel of hematopoietic cancer cell lines that are sensitive to ABT-737 treatment. Since acquired resistance to therapy is an important clinical phenomenon, we will study the molecular mechanisms by which cells might become resistant to ABT-737. Identification of these mechanisms will be critical to designing therapies to overcome resistance. We propose a series of mechanism-based combination therapies to be tested for their ability to overcome resistance and facilitate response to ABT-737 in cell lines and CLL cells. Finally, our studies to date suggest that cancer cells are far more likely to be dependent on anti-apoptotic proteins than non-malignant, normal cells. We will test this hypothesis by systematically evaluating the anti-apoptotic requirements of normal blood cells using BH3 profiling. Our goal is to provide a molecular understanding of what may be a dichotomy of vital clinical significance, as it suggests an intriguing therapeutic window between normal and cancer cells. These studies are directed at determining the specific ways cancer cells keep themselves alive, ways that may be used only by cancer cells, but not by normal, healthy cells. Understanding this would allow us to identify targets that would kill cancer cells, but not normal cells. Such a strategy has the promise to improve treatment of cancer by making treatment more selective for cancer cells, and therefore less toxic to normal, healthy tissues. [unreadable] [unreadable] [unreadable]

Engaging the cell death pathway known as apoptosis is critical for the success of many anti-cancer agents, both novel targeted agents and conventional cytotoxic agents. Whether the chemotherapy is conventional cytotoxic chemotherapy, or more modern targeted therapy, it is usually the case that while the initial target is known, the signaling downstream ofthe target that connects it to the intrinsic apoptotic pathway is poorly understood. Therefore, key molecular determinants of sensitivity and resistance to chemotherapy are largely unknown. The consequence, is that we are very bad at predicting what tumors respond to chemotherapy and why. Thus, many patients are exposed to toxic drugs without a good chance at response. Here we propose a systematic approach to connecting the contact of drug to upstream target to the commitment to programmed cell death at the mitochondrion. Since the BCL-2 family of proteins are the key regulators of mitochondrial apoptosis, they will be studied in particular detail. Our goals are to identify complete signaling pathways that are invoked when tumors are killed by anti-mitotic agents or ABT-737. We will use this information to generate predictive models that can predict sensitivity to treatment based solely on certain defined initial conditions. Once these predictive models are validated in in vitro models, we will apply to clinical testing. Furthermore, understanding the signaling pathways that permit successful killing by taxanes may permit the identification of targets in the pathway that can be selectively exploited by less toxic agents. The techniques we will use will include siRNA screening for genes that are essential for priming the apoptotic pathway. In addition, we have developed a new strategy which we call FACS-based BH3 profiling, that allows us to observe, at the single cell level, a cell's death signaling during progression to mitochondrial apoptosis. This strategy will permit a more detailed analysis ofthe role ofthe cell cycle and the role of individual molecules in determining death after treatment with anti-mitotic agents.

DESCRIPTION (provided by applicant): With the growing number of therapies being tested in chronic lymphocytic leukemia (CLL) comes the growing challenge of identifying for each patient which therapies are the best. In this competitive renewal proposal, we propose strategies to identify therapeutic opportunities in individual CLL tumors. A key element of this approach is the ability to predict patient response to individual therapies. We apply many of the basic lessons we have learned in the initial grant period about control of apoptosis, and make use of a tool that we refined in that proposal, BH3 profiling. Building on our prior work in which we found we can correlate in vitro sensitivity to ABT-737 to BH3 profiling results, we propose to test our ability to predict clinical response to the related BCL-2 antagonist, ABT-199 in CLL (Specific Aim #1). In addition, in Specific Aim #2, we propose to identify pathway addiction in individual cases of CLL. In the past, inefficiency of siRNA and shRNA approaches and difficulty of ex vivo culture in CLL have made this issue difficult to address by conventional means. We propose a study using small molecule pathway inhibitors instead of knockdown strategies to evade RNA transfection difficulties. Importantly, we use BH3 profiling to measure early apoptotic signaling within 4-24 of drug treatment, obviating the need for extended ex vivo culture. Our aim is to link early apoptotic signaling in response to small molecule inhibitors to pathway addiction. An important advantage of this approach is that detection of apoptotic signaling in primary CLL in response to drugs provides a rational path to clinical translation. Finally, in Specific Aim #3, we propose to investigate how stromal interactions inhibit apoptotic signaling in CLL cells. Moreover, using in vitro co-culture systems, we will study the efficacy and mechanisms of drugs intended to interrupt CLL interactions with stromal cells. Our goal is to identify which CLL patients will most benefit from such therapies.

DESCRIPTION (provided by applicant): Pancreatic cancer is a deadly disease that is typically refractory to even the most aggressive cytotoxic chemotherapy regimens. Recent advances in chemotherapy for pancreatic cancer have led to some improvement in response rates; however, the majority of patients still do not substantially benefit from therapy. Responses to chemotherapy commonly occur through apoptosis, a form of programmed cell death. Our goal is to identify genes whose loss of function enhances sensitivity of pancreatic cancer cells to apoptosis. BH3 profiling is a well-characterized upstream measurement of the apoptotic sensitivity or priming of cancer cells and has been shown to correlate with response to chemotherapy and disease prognosis in certain cancer types. We will utilize a novel screening approach called RNAi-BH3 Screening (RiB Screening) that combines loss-of-function genetic screening using pooled lentivirally delivered shRNAs with the technology of BH3 profiling. In this approach, we will identify genes whose knockdown results in enhanced priming for apoptosis across several pancreatic cell lines. Based on this loss of function genetic screen, we will attempt to identify known small molecules targeting these genes, and determine whether they alter apoptotic priming and chemosensitivity. Finally, we propose to evaluate potential clinical utility of identified genes or small molecules by determining their effect on BH3 Profiles and chemosensitivity in fresh patient-derived pancreatic cancer samples. The successful identification of such genes and small molecules will not only improve our understanding of chemotherapeutic response, but could also serve as novel biomarkers or therapeutic targets.

? DESCRIPTION (provided by applicant): Inter- and intra-tumoral mitochondrial heterogeneity in response to chemical perturbation in cancer Response to therapy is heterogeneous between tumors of different histology, between different tumors of the same histology, and even among different tumor cells within the same tumor. Much of the cytotoxic response to therapy in cancer is governed by the mitochondrial pathway of apoptosis. Therefore, we hypothesize that heterogeneity in response at all three of these levels is governed by functional and molecular heterogeneity of mitochondria. BH3 profiling provides a functional measure of how close a cell is to the threshold of apoptosis by measuring mitochondrial sensitivity to BH3 peptides, a property we also refer to as apoptotic priming. We have previously shown, in both hematologic and solid tumors, that pretreatment baseline priming of patient tumors predicts clinical response to conventional chemotherapy. Our approach to understanding response to targeted therapies is distinct. In contrast to the ubiquitous nature of the DNA and microtubules that are the targets of conventional chemotherapy, targeted therapies attack vulnerabilities that are selectively present in certain cancer cells. To identify where these vulnerabilities exist, we briefly (less than 24 hours) expose cancer cells to drugs and measure whether the drugs increase apoptotic priming. We have shown that this approach, which we call dynamic BH3 profiling (DBP), can predict death of tumor cells from targeted agents in vitro and in vivo for both hematologic and solid tumors. There are two important advantages of DBP over conventional measures of cytotoxicity. First, DBP can be applied much more efficiently to primary patient samples. Experiments measuring cytotoxicity in cancer cells often require days of culture. Because the long-term culture of primary patient samples is so unreliable, cytotoxic measurements in primary tumors are unreliable. DBP requires no more than a day of ex vivo culture since we measure well before frank cell death occurs, and we have demonstrated its predictive power in primary liquid and solid tumors. Second, there are many useful anti-cancer agents that do not cause frank cell death, but which nonetheless provoke apoptotic signaling that facilitates killing in combination regimens. Classical cytotoxic measurements do not identify these, but DBP does. We propose to study intra- and inter-tumoral heterogeneity based on differential response to compounds that sensitize mitochondria for apoptosis. We will focus on colon cancer tumors, as we have access to primary and PDX specimens of these tumors. Our main goals are to develop a therapeutic toolbox of potentially useful drugs, determine how best to combine these in a personalized way, and also to understand the molecular basis of the heterogeneity of mitochondrial function that underlies differences in response to these drugs.

Project Summary: Despite impressive therapeutic advances in the treatment of CLL over the past decade, relapsed tumors that are often resistant to known therapies remain a problem. Understanding resistance and developing new treatment strategies will be necessary to turn CLL into a curable disease. With few notable exceptions, responses to chemotherapy typically occur through apoptosis, a form of programmed cell death. Our goal is to identify drugs that sensitize CLL cells to apoptosis, and to identify combinations of drugs that would prevent relapse. While chemical screening of conventional cytotoxicity is an attractive strategy to meet this goal, such an approach is impeded by the inability to reliably culture CLL cells for longer than 48 hours ex vivo. This ultimately reflects a technological gap in our ability to chemically perturb CLL ex vivo, and measure functional and clinically relevant phenotypes. Here, we will use a novel chemical screening approach called dynamic BH3 profiling (DBP), which enables functional measurements of chemically induced apoptotic changes, and requires only 6-24 hours of ex vivo culture, thereby maximizing molecular fidelity and tumor cell viability. Using a panel of over 2000 drugs, not only will we identify novel molecules that sensitize CLL cells for apoptosis, but we will also determine chemical apoptotic sensitivities that are lost or gained on relapse for individual patients. Drugs that uniquely sensitize only relapsed tumors present exciting opportunities for combination chemotherapy. Finally to understand mechanisms of apoptotic vulnerabilities in pre- and post-treatment samples, we will correlate our measurements of functional apoptotic responses with the large molecular datasets in Project 1 and 2 for at least 104 CLL patients. In sum, the successful identification of apoptotic sensitizing drugs in CLL, will not only advance our molecular understanding of CLL, but could result in truly novel therapeutic options.

Summary I am an oncologist and cancer biologist supervising a laboratory focused on identifying therapies that selectively induce apoptosis in cancer cells. My initial contributions to the apoptosis field came with separating certain pro-death BCL-2 family BH3-only proteins into ?sensitizers? and ?activators? based on pro-apoptotic function. This finding drove my interest in the possibilities of inhibiting BCL-2 function with drugs that mimicked the BH3 domain of pro-apoptotic proteins. I designed the first mouse model that demonstrated that loss of BCL-2 function by itself could be sufficient to drive a cancer into remission. Following this, I designed a tool called BH3 profiling ? exposing mitochondria to synthetic oligo-peptides based on the amphipathic alpha- helical BH3 domains of proapoptotic proteins and measuring mitochondrial outer membrane permeabilization (MOMP). By using certain selectively-interacting BH3 peptides, I could use BH3 profiling to identify cells that were especially sensitive to BH3 inhibition. I used BH3 profiling to help launch clinical trial programs of the BCL-2 inhibitor venetoclax in several diseases. Most successful among these so far have been programs in chronic lymphocytic leukemia and acute myelogenous leukemia, the former of which has already yielded FDA approvals. Using different BH3 peptides, BH3 profiling can measure overall apoptotic priming, or proximity to the threshold of apoptosis. We used this aspect to demonstrate that differential apoptotic priming is perhaps the most significant determinant of successful chemotherapy treatment. Moreover, differential apoptotic priming is the main reason that there is a therapeutic index for conventional chemotherapy ? most non-malignant somatic cells are far less primed for apoptosis than chemosensitive cancer cells. Building on this finding, we asked whether we could identify drugs that could enhance apoptotic priming selectively in cancer cells. We found that we could measure increased apoptotic priming within hours of exposing cancer cells to effective drugs using dynamic BH3 profiling (DBP). Increased priming is measured as increased sensitivity of mitochondria in treated cells to BH3 peptides compared to untreated controls. Over the past few years, we have found that an increased priming by a drug in DBP is an excellent predictor of in vivo activity in human and mouse models, in solid and liquid tumors. An important advantage of DBP over most other ex vivo drug sensitivity strategies is that DBP requires no more than 24 hours of ex vivo culture. This overcomes the major obstacle to the general application of such strategies, since many cancers cannot adapt to long-term ex vivo culture, and if they do, they are phenotypically altered so as to degrade the information they can provide. We are exploring DBP as a discovery tool and predictive biomarker in many liquid and solid tumors. Moreover, we are using it as a tool to identify drugs that can make target tumor cells more sensitive to immuno-oncology therapies.

Abstract The majority of people diagnosed with acute myelogenous leukemia (AML) die of their disease, despite availability of active standard chemotherapy regimens, targeted therapies, and allogeneic transplantation. We have found that we can probe mitochondrial apoptotic signaling of patient myeloblasts to identify active therapies in AML using BH3 profiling. Prior exploitation of BH3 profiling led to the FDA approval of venetoclax in combination with hypomethylating agents or low-dose cytarabine. Here we present an innovation, Dynamic BH3 Profiling, which we propose to use to identify active single agents as well as combinations in AML. High priority will be given to combinations with BH3 mimetic drugs. We will also include in our studies candidate small molecules that emerge from the other Projects. Promising candidates will advance to clinical trials in collaboration with Clinical Core 3. For the first time, we propose to accelerate novel clinical trials in AML by using dynamic BH3 profiling to prioritize combinations that drive high apoptotic signaling in patient myeloblasts. In prior work, we have found that BH3 profiling can predict response to induction regimens in AML. We have installed BH3 profiling in a clinical laboratory where we can now test over a hundred AML samples per year. Building on our prior work, and working with Biostatistics Core 2, we will construct predictive biomarkers based on BH3 profiling for response to standard 7+3 induction, as well as to the newer venetoclax plus azacytidine. We will test inclusion into our predictive tool information like ELN criteria, age, and pathology. The expected output is a set of predictive tools that will offer likelihood of CR/CRi for individual patients for both induction regimens. It is hoped that this tool will help guide patients to the induction therapy that is best for them. While we exploit the concept of apoptotic priming to predict clinical response and identify active regimens in AML, we know little about the molecular determinants of the apoptotic priming phenotype. We will utilize genomic, transcriptomic, and proteomic tools to investigate the differences between myeloblasts in different states of apoptotic priming. While this work will focus on AML, it is expected that it will reveal principles of molecule determination of priming that will be applicable to other cell types.