Yi Xing, Ph.D. - US grants

Clathrin-coated vesicles (CCV) are intracellular devices for membrane traffic from the cell surface and the trans-Golgi network to endosomes. They are not only important universal vehicles used by all nucleated cells for intracellular membrane traffic, but also have tissue-specific functions, especially in neurons and in the immune system. CCV pathways are involved in metabolic diseases, neurological disorders and viral infections. A critical step in the CCV "life cycle" is clathrin-lattice uncoating, a process mediated by the chaperone Hsc70 and co-chaperone auxilin. I propose to determine the molecular mechanism of the uncoating process, through structural studies and biochemical experiments. I will obtain higher resolution cryoEM structures of clathrin coats of different sizes and designs, to understand the role of invariant interactions in stabilizing the clathrin lattice. I will then determine the positions of Hsc70 and auxilin in these lattices, to visualize how uncoating starts. I will test hypotheses through studies of assembly and uncoating of cages formed from recombinant clathrin bearing specific mutations. I will also attempt to obtain crystals of Hsc70 bound with the auxilin J-domain and of the auxilin PTEN-like domain, to understand the function of auxilin through its x-ray crystal structures. These studies will elucidate the molecular mechanism of Hsc70/auxilin-mediated CCV uncoating and greatly advance understanding of CCV pathways. My work will also provide a general framework for using modern X-ray crystallography and cryoEM techniques to acquire high-resolution information on intricate molecular machines. Clathrin-coated vesicles play essential roles in maintaining the health of neurons and immune system. Defects in CCV pathways correlate with a broad spectrum of human diseases, including familial hypercholesterolemia, leukemias, Alzheimer's disease, influenza and HIV. Understanding steps in CCV life cycle as proposed by this research will assist in the discovery of new therapeutic strategies for diseases involving defects in membrane traffic.

DESCRIPTION (provided by applicant): MicroRNAs (miRs) play important roles in regulating cell function, with more than half of mammalian mRNAs under selective pressure to maintain pairing to miRs. However, with the exception of cancer, little is known regarding miR functions in the lung, and less known about miR expression and functions in airway epithelia. Cystic fibrosis (CF) is caused by mutations in the CFTR gene, a nucleotide regulated anion channel. CFTR is a low abundance mRNA, and loss of function has profound consequences. There is no knowledge of miR function in this disease. We surveyed miR expression in cultured well-differentiated human CF and non-CF airway epithelia and identified four miRs with increased expression in CF cells. Two miR targets of interest that we validated in pilot studies are SIN3A, a transcriptional co-repressor known to interact with regulatory elements in the CFTR promoter, and the CFTR gene product itself. We hypothesize that changes in miR expression resulting from CFTR mutations may contribute to the complex CF lung disease phenotype. There are three Aims: Aim 1. To validate selected mRNA targets of differentially expressed miRs. The transcriptional regulatory gene SIN3A is a predicted miR-138 and -661 target, while CFTR is a target for miRs-494, -509-3p, and -661. We hypothesize that CFTR protein abundance is regulated by miRs at the transcriptional and translational levels. We will confirm the ability of differentially expressed miRs to bind the SIN3A and CFTR 3'UTRs in fusion gene (luciferase) assays. These and additional confirmed targets will be studied in gene specific studies. Aim 2. To examine effects of miRs in the context of airway epithelia. Confirmed targets of miRs-138, -494, -509-3p, -661 will be assessed in gain and loss of function experiments in airway epithelia. For gain-of-function effects on SIN3A and CFTR, we will individually express miRs-138, - 494, -509-3p, -661 in CF and non-CF airway epithelia, and quantitatively assess endogenous SIN3A and CFTR mRNA and protein abundance. For loss of function effects, we will deliver anti-miRs to cells to inhibit interactions of endogenous miRs and their target 3'UTRs, and similarly assess mRNA and protein abundance of targets. For miRs with confirmed CFTR regulation we will further investigate the how miR expression alters cAMP-activated CFTR Cl- channel function in vitro. Similar strategies will be used for other priority targets. Aim 3. To perform expanded expression profiling for a comprehensive discovery of differentially expressed miRs and their target mRNAs in CF and non-CF epithelia. We will use qRT PCR arrays to investigate miR expression patterns in a larger sample of CF (?F508/?F508) and non-CF primary human airway epithelia. MiRs with differential expression in CF will be confirmed by qRT PCR and Northern blot. In addition, CF and non-CF samples used in miR profiling will be profiled by Illumina Solexa mRNA sequencing (mRNA-seq) for their global gene expression patterns. Completion of this aim will provide a robust expression profile for 685 human miRs, expand our genomic efforts and miR target selection through anti-correlations with mRNA expression levels. PUBLIC HEALTH RELEVANCE: MicroRNAs are a class of small non-coding RNAs that play key roles in regulating gene expression. This project will investigate changes in microRNA expression in cystic fibrosis (CF) airway epithelia and how such changes contribute to the complex lung disease phenotype in CF. This will lead to a better understanding of cellular processes that may influence CF disease states, and could provide novel drug targets for therapeutic development.

DESCRIPTION (provided by applicant): The goal of this project is to systematically survey the evolution of pre-mRNA splicing in primates, and elucidate the molecular mechanisms that created species-specific exons and splicing patterns. Alternative splicing in higher eukaryotes generates an enormous regulatory and functional diversity from a limited repertoire of protein-coding genes. It also permits a gene to evolve a new spliced isoform, while still expressing the ancestral spliced isoform. Many genes have species-specific exons and splicing patterns that arose from either small-scale sequence changes that affected essential splicing signals, or large-scale insertions or deletions. However, despite the critical role of splicing during eukaryotic genome evolution, many questions regarding how splicing changes occurred and the evolutionary significance of such changes remain largely unexplored. We propose to combine genomic, computational, and molecular approaches to study splicing changes during primate and human evolution. The specific aims are: 1) To investigate the birth and evolution of new exons in primates, using genome alignments of vertebrate species, extensive exon-level transcriptome profiles of human genes generated by microarray and sequencing-based technologies, and molecular splicing analysis of new exons in humans and nonhuman primates. 2) To globally examine splicing differences between humans and nonhuman primates, by high-density exon junction array and RNA-seq profiling of a large panel of human and primate tissues. 3) To elucidate the mechanisms of splicing evolution in primates, via comparative analysis of splicing regulatory signals and minigene experiments. This project will improve the annotation of human and primate genomes, greatly expand the knowledge of new exons and splicing patterns that are unique to our species, and shed light on how eukaryotic genomes expand their functional repertoire via the evolution of splicing. The results of these studies will elucidate how the evolution of genomic sequences contributed to splicing differences among species. This will provide significant insight into the regulation of splicing, and how genetic variations disrupt splicing in human diseases. PUBLIC HEALTH RELEVANCE: Many human diseases are caused by aberrations in pre-mRNA splicing. This project will systematically survey the evolution of splicing in primates, and elucidate how splicing patterns change as a result of genome sequence evolution. These studies will provide significant insight into how splicing is regulated, and how genetic variations disrupt splicing in human diseases.

DESCRIPTION (provided by applicant): The human metabolic syndrome, an archetypical complex disease (involving multiple genes and environmental interactions), is a collection of disorders including obesity, dyslipidemia, hypertension, and insulin resistance, leading to end organ failure and death. While genetic studies have had success in identifying genes related to obesity, dyslipidemia, and hypertension, much of the variation remains unknown. The Lyon Hypertensive (LH) rat has several features common to the human metabolic syndrome - high body weight, cholesterol, and triglycerides, increased insulin and insulin/glucose ratio, and high blood pressure exacerbated by a high salt diet. The Lyon normotensive (LN) control strain is genetically quite similar to the LH, but phenotypically very distinct. Mapping studies in an F2 intercross between the LH and LN identified quantitative trait loci (QTL) contributing to body weight, lipid levels, blood pressure and insulin levels. However, the genes that underlie the substantial phenotypic differences between the genetically similar LH and LN rats are not yet known. Identification of the genes underlying QTL can be facilitated by comparing transcriptomes of the disease and control models and by identifying positional candidate genes and perturbed biological pathways. While gene expression arrays allow for analyses in disease related tissues, they are limited by the features contained on the array as well as the current state of genome annotation. Of increasing importance in the metabolic syndrome and other complex diseases is the occurrence of alternative pre-mRNA splicing. However, conventional gene expression arrays cannot examine alternative splicing patterns. We hypothesize that the metabolic syndrome in the LH rat is due to a complex gene regulatory network which contributes to changes in gene expression and RNA processing. Because of the inherent complexity in common disease, we assert that deep RNA sequencing of transcriptomes from LH and LN rat strains will lead to the identification of gene(s) and mechanisms involved in the metabolic syndrome in the LH rat. We propose to carry out a high throughput RNA sequencing analysis of gene expression and alternative splicing in tissues collected from LH and LN strains. Specifically we will 1) identify genomewide gene expression differences between genetically similar LH and LN strains in disease-associated tissues;and 2) identify alternative splicing differences between LH and LN strains by ultra deep RNA sequencing. Identification of transcriptome signatures associated with obesity and dyslipidemia in animal models will lead to novel disease genes and pathways, and ultimately a better understanding and treatment of the human metabolic syndrome. PUBLIC HEALTH RELEVANCE: The human metabolic syndrome (obesity, dyslipidemia, hypertension, and insulin resistance) and its related end organ failure affects nearly 25% of the US population and has a major impact on health care costs in the US, estimated at over $30 billion annually. This project will examine genomewide patterns of gene expression and RNA processing in disease-associated tissues collected from a rat model of the human metabolic syndrome. Identification of a metabolic syndrome transcriptome signature in animal models will lead to better understanding and treatment of the human disease.

DESCRIPTION (provided by applicant): Huntington's disease (HD) and Huntington's disease-like 2 (HDL2) are remarkably similar autosomal dominant adult onset neurodegenerative disorders, nearly indistinguishable clinically and pathologically. Each disease is ultimately fatal with no effective treatment to stop or slow the relentless progression. HD affects about 30,000 Americans, with a much higher number at risk; HDL2 is rare. The complete explanation for HD and HDL2 pathogenesis remains elusive. A novel strategy for focusing the search for disease mechanisms and therapeutic targets of HD is to determine those points at which the pathogenic pathways of HD and HDL2 converge. A particularly powerful method for implementing this strategy is to compare the transcriptomes of the two diseases. Based on our preliminary evidence, we hypothesize that both abnormal levels of gene expression and abnormal gene splicing will be present in HD and HDL2 and that the sets of these abnormalities will overlap in the two diseases. Here we propose to take advantage of the remarkable similarities of HD and HDL2 to identify convergent pathogenic pathways, via parallel transcriptome characterization of mouse models and human patient samples of HD and HDL2. We propose two specific aims. In aim 1, we will use state of the art exon junction array, RNA sequencing (RNA-Seq), and analytic methods to examine and compare RNA extracted from human HD and HDL2 postmortem brains and mouse models of HD and HDL2 as well as controls. In aim 2, we will experimentally validate expression and splicing abnormalities using high-throughput automated PCR assays and new RNA samples, compare mouse and human data, and use bioinformatics tools to determine common gene sets, pathways, and molecular subnetworks shared by genes showing gene expression or splicing abnormalities in HD and HDL2 brains. We anticipate that the proposed studies will create an extremely valuable resource that will provide a detailed characterization of the HD and HDL2 transcriptomes at an unprecedented resolution and hence fundamentally improve understanding of disease pathophysiology.

A CAG repeat expansion in exon 1 of the HD gene product, huntingtin, causes Huntington's disease (HD), a fatal neurodegenerative disease for which there is no cure or neuroprotective treatment. Dysregulation of transcription is a major feature of HD pathogenesis, as indicated by a large body of work using RNA array techniques, and work on specific transcription factors and their targets. More recent studies have also suggested a role for huntingtin in RNA processing. Prior work on gene expression alterations in HD brain tissues used 3' biased gene expression arrays. Of increasing importance in many human diseases, particularly neurodegenerative diseases, is the occurrence of aberrant alternative pre-mRNA splicing. However, conventional gene expression techniques are not well suited to quantitative analysis of alternative splicing patterns, and do not sample rare transcripts well. Several lines of evidence from our preliminary work suggest global splicing abnormalities in HD. For example, we reported earlier that microRNA miR-124 was significantly reduced in HD brains. Work by Maniatis and colleagues showed that miR-124 promotes neuronal-specific alternative splicing events by down- regulating an important tissue-specific splicing regulator, polypyrimidine tract-binding protein (PTBP1). Consistent with the decrease in miR-124, we have preliminary evidence for significantly increased PTBP1 mRNA levels in HD patient samples. Moreover, preliminary data suggest that several exons in genes regulated by PTBP1 show corresponding changes in exon inclusion/exclusion in HD brain. Inclusion or exclusion of non-constitutive exons can have dramatic effects on transcript stability and protein activity. Thus transcriptome alterations in HD may extend beyond up- and down-regulated genes to include changes in gene and protein isoforms. Assessing these events on a global scale for HD will aid efforts to unravel disease pathophysiology, and may identify new drug targets for therapy. In our work, which encompasses 3 aims, we will move from identification of the altered HD transcriptome, to validation, to in vitro and in vivo studies to test their relevance on HD phenotypes. These studies combine the genomics and bioinformatics expertise of the Xing lab and the HD expertise of the Ross and Davidson labs. The functional relevance of those changes will be elucidated using gain and loss of function studies in the Davidson and Ross labs, where both groups have substantial experience with HD models.

DESCRIPTION (provided by applicant): The central objective of this project is to elucidate the roles of epigenetics and chromatin states in RNA splicing regulation. Eukaryotic cells generate astonishing regulatory diversity and as a consequence exceedingly complex phenotypes from a finite set of genes. Alternative pre-mRNA splicing plays an essential role in creating this regulatory diversity by generating multiple RNA isoforms from a single gene. Traditionally, splicing was considered as a post-transcriptional process, and studies of splicing regulation have largely focused on the roles of cis splicing regulatory elements and their interactions with canonical RNA-binding splicing factors. However, recent studies of eukaryotic epigenomes and transcriptomes have revealed a surprisingly complex picture of splicing regulation shaped by chromatin states and epigenetic marks. Exons are characterized by increased levels of nucleosome positioning, DNA methylation, and certain histone modifications. Many introns are spliced co-transcriptionally when the nascent RNAs are tethered to the chromatin, and changes in the transcription elongation rate or epigenetic marks can influence exon splicing patterns. Despite these exciting findings, many questions about epigenetic regulation of splicing remain unresolved. We propose to systematically investigate chromatin and epigenetic regulation of RNA splicing, by taking advantage of the broad and deep epigenome and transcriptome data generated by the Epigenome Roadmap project. By correlating transcriptome profiles to epigenome profiles across diverse cell types, we aim to address a series of important questions regarding epigenetic regulation of co-transcriptional and post-transcriptional RNA splicing. In three aims, we will investigate epigenome-splicing correlation in diverse tissues and cell types (Aim 1), identify combinatorial chromatin states and long-range interactions associated with splicing (Aim 2), and elucidate how epigenetic determinants affect chromatin-associated splicing (Aim 3). The proposed studies will significantly advance our understanding of splicing regulation, and how epigenetic signals influence alternative splicing in normal and diseased cells. In addition, through the proposed work we will develop novel computational methods for linking epigenome signatures to RNA splicing patterns. We anticipate that these tools will be of broad interest and utility to researchers studying epigenome and transcriptome regulation in diverse biological systems.

PROJECT SUMMARY Women and men are well known to have different propensities to neuropsychiatric illness, but the source of these differences is not understood. In particular, the molecular events that determine functional dimorphism between the female and male brain need to be defined. We will examine how the female-specific long noncoding RNA Xist and its newly identified interaction with the pre-mRNA splicing regulators PTBP1 and 2 affect gene expression and alternative splicing in the female brain. The project will use expertise and tools developed in three labs for the study of Xist RNA, of neuronal splicing regulation by the PTB proteins, and of gene expression and alternative splicing using computational methods. RNA-seq data from defined regions of both human and mouse brain will be analyzed using the new rMATS analysis tool to create a large statistically robust database of differential gene expression and alternative pre-mRNA splicing between males and females. Expression and splicing changes will be correlated with changes in PTBP1/2 mRNA and Xist across the same datasets to define genes potentially regulated by these molecules at the transcriptional and post- transcriptional levels. Female specific patterns of expression and splicing caused by the XX genotype will be distinguished from events driven by female hormones using four core genotype mice. Xist targeting will be confirmed using conditional Xist alleles that allow either removal or activation of Xist during brain development and measurement of the resulting changes in splicing. The expression of Xist relative to PTBP2 will be quantified over neuronal differentiation in culture. The PTBP targeting of Xist-dependent changes in splicing will be confirmed in PTBP2 knockout and PTBP1 transgenic mice, and by transcriptome-wide binding analyses by iCLIP. Alternative splicing is a widespread mechanism of gene regulation, but has been only minimally examined in relation to the XX genotype of female cells. Using sophisticated new genome-wide methods and molecular tools, these studies promise to identify new genetic determinants of sexual dimorphism in the mammalian brain and to elucidate their underlying molecular mechanisms. In the longer term, the identified molecular changes driven by Xist and PTBP will provide entrée to the examination of the functional consequences of these dimorphisms.

PROJECT SUMMARY In eukaryotes, one gene can give rise to multiple protein isoforms through various types of alternative pre- mRNA processing (e.g., alternative splicing), contributing significantly to proteome complexity. Differential isoform expression manifests in pathogenesis of diseases from heart failure to neurodegeneration, as well as cellular responses to environmental stress including alcohol and oxidative damage. Advances in RNA-seq technology have led to the discovery of many novel alternative isoforms, but their biological impact is often unclear in the absence of protein information. Conversely, shotgun proteomics technology enables large-scale characterization of proteins, but the limitations of ?one-gene, one-product? databases prohibit their utility in protein isoform identification. Deeper insights into the biology of alternative isoforms require combining the complementary strengths of transcriptomics and proteomics. Accordingly, the integration of technical platforms from mRNA to protein has become an indispensable step in advancing a holistic portrait on gene products. Among the key challenges is the segregation of proteomics and transcriptomics repositories, as well as the disconnect of respective data analysis pipelines and expertise. Despite recent progress, there is an urgent and unmet need for well-integrated and user-friendly computational platforms that can support everyday biomedical researchers in harnessing diverse data types for multi-omics studies. The central goal of this project is to create a unified platform to decode alternative isoforms from RNA- seq/Ribo-seq data, and to guide shotgun proteomics characterization of protein isoforms. Our approach capitalizes on the rapid revolution of Big Data sciences in recent times, where new frontiers in multi-omics integration now make it possible to traverse heterogeneous computational resources and data types seamlessly. We will design, construct, and implement an integrative proteotranscriptomics framework (ProteoSeq), which will combine novel analytical models and custom proteomics workflows to coalesce transcriptomics and proteomics data for large-scale characterizations of alternative protein isoforms. Our proposal details three data science aims, which will (i) develop methods to infer full-length mRNA and protein isoforms from hybrid (short-read/long-read) RNA-seq and Ribo-seq data; (ii) engineer an integrative platform for users to analyze protein isoforms from proteotranscriptomics data on the cloud; and (iii) validate and accrue protein evidence for alternative isoforms in diverse high-value datasets. Our efforts aim to synergize two currently fragmentary omics fields and thereby empower inquiries on the regulations of alternative isoforms in health and disease. We envision the proposed computational tools will be generalizable to multiple biomedical disciplines, and will serve the broad scientific community for routine multi-omics investigations in translational medicine.

PROJECT SUMMARY The central objective of this renewal R01 project is to elucidate the variation and regulation of alternative splicing (AS) in human transcriptomes. Eukaryotic cells generate astonishing regulatory diversity and complex phenotypes from a finite set of genes. AS of precursor mRNA is a mechanism essential for generating this regulatory diversity. Almost all multi-exon human genes are alternatively spliced. Widespread changes in AS occur between species, within human populations, in response to developmental signals and environmental perturbations, and in disease pathogenesis. Despite the importance of AS in gene regulation and disease, as well as extensive interest and research activities in this field, there remain many open questions and significant knowledge gaps regarding the landscape, regulation, and functional consequence of AS variation in human transcriptomes. Powerful sequencing technologies for characterizing transcriptome complexity (RNA-seq) and protein-RNA interaction (CLIP-seq), as well as the vast amounts of data continuously deposited into the public domain, create exciting and unprecedented opportunities for studies of AS. This proposal integrates data- driven research leveraging big transcriptome data with hypothesis-driven research using molecular and genomic tools. In three aims, we will investigate the evolution of AS in primates (Aim 1), the genetic variation and phenotypic association of AS in human populations (Aim 2), and epigenetic regulation of AS across human tissues and cell states (Aim 3). We will also develop and disseminate innovative computational and statistical methods for analyzing AS using large, heterogeneous sequencing datasets. As in our previous funding cycle, we will tap into our extensive network of expert collaborators to amplify the impact of our work and pursue new research opportunities within and beyond the scope of the proposed project. Collectively, our research will generate significant novel insights into the variation, regulation, and function of AS, and create broadly applicable computational tools for studying AS variation and mRNA isoform complexity in diverse biomedical disciplines.

ABSTRACT We present a discovery and therapeutics development program to establish alternative splicing as a therapeutic target of Myc-driven prostate cancer. The Myc-family of proto- oncogenes, including c-Myc, N-Myc, and L-Myc, play a central role in the pathogenesis of this and many other cancers. Yet therapies inhibiting Myc action have yet to reach the clinic. As transcription factors, Myc proteins are difficult to target with either small molecules or immunotherapeutics. Indirect strategies that target downstream effectors of the Myc expression program may prove more successful. Our strategy is to target the splicing factors and alternatively spliced isoforms that the multiple Myc paralogs rely on to drive prostate cancer. Myc has been recently shown to control alternative splicing patterns that are crucial to Myc-driven tumor growth. We hypothesize that these tumors rely on specific splicing regulatory proteins that can be identified and potentially targeted with small molecules. In parallel, we hypothesize that Myc-driven alternative splicing will create cancer-specific protein isoforms suitable for immunotherapeutics development. Given the ubiquity of Myc deregulation in human cancer, we anticipate that our results will be of broad relevance to the cancer research community. We have assembled a team of investigators at UCLA with extensive experience in bioinformatics (Yi Xing), alternative splicing (Douglas Black), and cancer cell biology and immunology (Owen Witte). Our proposal integrates the analyses of cancer genomic data with focused experimental research using unique Myc-transformed human prostate materials. We will gather data on Myc-dependent alternative splicing in normal prostate tissues and primary cancers from large datasets (TCGA, GTEx) as well as datasets representing advanced disease states. Total proteomics analysis will be conducted on our transformed materials to confirm protein expression of candidate isoforms. We will employ a high-throughput screening platform developed in one of our labs to identify genetic or chemical modulators of Myc-dependent splicing events. Candidate cell surface isoforms selected for immunotherapeutics development will carry cancer- specific exon-exon junctions suitable for antibody development with phage display libraries. High-affinity, high- specificity antibodies will be built into CAR T-cells for further development

SUMMARY/ABSTRACT The lack of effective glioblastoma treatments poses a significant health problem and highlights the need for novel and innovative approaches. Immunotherapy is an appealing strategy because of the potential ability for immune cells to traffic to and destroy infiltrating tumor cells in the brain. New information suggests that patients mounting immune responses after immunotherapy preferentially recognize novel neoantigens created by tumor-specific mutations. Our data, and that from other immunotherapeutic strategies for patients with cancer, suggest that the vast majority of tumor-specific T cells induced by such personalized, patient-specific immunotherapies do NOT recognize well-characterized, known antigens. Such information is consistent with recent data from other immune-responsive cancers, such as melanoma, in which the percentage of tumor-specific T cells recognizing known antigens was less than 1%. In order to design the most effective immunotherapeutic strategies for glioblastoma, we believe that it is critical to understand which antigens tumor-specific T cells recognize in this disease. Our hypothesis is that glioblastoma patients treated with immunotherapy will mount anti-tumor immune responses against specific mutations and splice variants in their individual tumors. Similarly, our other recent findings strongly suggest that the addition of PD-1 antibody (mAb) blockade to DCVax enhances both the intra-tumoral CD8+ T cell response and clinical benefit in pre-clinical studies. Furthermore, the timing of PD-1 mAb blockade is immunologically relevant; our unpublished, recent clinical trial results highlight how the neoadjuvant (prior to surgery) treatment with PD-1 mAb blockade induces enhanced anti-tumor immune responses and clinical benefit. We hypothesize that the addition of PD-1 mAb blockade should amplify the neoantigen-specific T cell response induced by DC vaccination, both in the blood and the tumor. To test these important questions, In Aim 1, we will develop a new bioinformatics pipeline to predict neoantigens that arise specifically from the types of genetic alterations that occur in GBM. In Aim 2, will create immunocompetent murine glioma models to test the importance of neoantigens. Finally, in Aim 3, we will identify neoantigen-specific T cells from both the TIL population and peripheral blood of GBM patients treated with immunotherapy. These studies span the continuum of translational research in brain tumor immunotherapy and will likely provide informative new insights for the development of new, rational immune-based strategies for brain tumor patients.

ABSTRACT We present a collaborative immunotherapeutics discovery program that exploits alternative pre-mRNA splicing as a source of cancer-specific epitopes for T-cell receptor (TCR) therapy of small cell carcinomas of the prostate and lung. Small cell carcinomas arise from many different epithelial tissues but are generally aggressive, have no curative treatment, and carry a dire prognosis. Small cell lung cancer (SCLC) is the most common subtype. Small cell prostate cancer (SCPC) is rare as a primary disease but is becoming increasingly common as a late-stage phenotypic transition in response to hormone-deprivation therapy. Emerging research indicates that despite their disparate tissues of origin, SCPC and SCLC are highly similar in behavior and molecular phenotype. This suggests effective targeted therapies could address both malignancies. Our strategy is to define cancer-specific epitopes created by alternative pre-mRNA splicing in small cell carcinomas and then use these targets to develop TCR-based therapeutics. Chimeric antigen receptor T-cell (CAR-T) therapies targeting cell surface proteins have been developed for some hematological malignancies, but this strategy has been unsuccessful for epithelial tumors. The limited cancer specificity of the target epitope has led to significant on-target, off-tumor toxicities in human trials. We have chosen to pursue TCRs to expand the pool of available targets beyond the cell surface. We hypothesize that tapping into the additional proteomic diversity revealed by a detailed analysis of alternatively spliced exons will provide better targets. Our team of principal investigators includes experts in the computational biology of alternative splicing (Yi Xing), cancer cell biology and immunology (Owen Witte), and hematopoietic cell development and immunology (Gay Crooks). We are compiling RNA-Seq data on small cell cancers and normal tissues from public datasets and new human cell line models of SCPC & SCLC derived from benign cells by lentiviral transduction. This combined dataset serves as the foundation for our discovery effort. We plan to pair this with total proteomics analysis to identify spliced isoforms that affect protein composition. This data will be further integrated with immunopeptidomics assays that define the pool of peptides presented to the immune system by the target cancer cells. Epitopes derived from alternative splicing events that show high cancer specificity, protein expression, and predicted or observed epitope presentation will be prioritized for TCR development. We will use these epitopes to select TCRs from naïve human T-cell populations using a highly organotypic in vitro artificial human thymic culture system developed in the Crooks laboratory.