ABSTRACT
Acute myeloid leukemia with KMT2A (MLL) rearrangements is characterized by specific patterns of gene expression and enhancer architecture, implying unique core transcriptional regulatory circuitry. Here, we identified the transcription factors MEF2D and IRF8 as selective transcriptional dependencies of KMT2A-rearranged AML, where MEF2D displays partially redundant functions with its paralog, MEF2C. Rapid transcription factor degradation followed by measurements of genome-wide transcription rates and superresolution microscopy revealed that MEF2D and IRF8 form a distinct core regulatory module with a narrow direct transcriptional program that includes activation of the key oncogenes MYC, HOXA9, and BCL2. Our study illustrates a mechanism of context-specific transcriptional addiction whereby a specific AML subclass depends on a highly specialized core regulatory module to directly enforce expression of common leukemia oncogenes.
Subject(s)
Leukemia, Myeloid, Acute , Myeloid-Lymphoid Leukemia Protein , Gene Rearrangement , Humans , Interferon Regulatory Factors/genetics , Interferon Regulatory Factors/metabolism , Leukemia, Myeloid, Acute/genetics , Myeloid-Lymphoid Leukemia Protein/genetics , Myeloid-Lymphoid Leukemia Protein/metabolism , Oncogenes/geneticsABSTRACT
Epstein-Barr virus (EBV) immortalizes resting B-lymphocytes through a highly orchestrated reprogramming of host chromatin structure, transcription and metabolism. Here, we use a multi-omics-based approach to investigate these underlying mechanisms. ATAC-seq analysis of cellular chromatin showed that EBV alters over a third of accessible chromatin during the infection time course, with many of these sites overlapping transcription factors such as PU.1, Interferon Regulatory Factors (IRFs), and CTCF. Integration of RNA-seq analysis identified a complex transcriptional response and associations with EBV nuclear antigens (EBNAs). Focusing on EBNA1 revealed enhancer-binding activity at gene targets involved in nucleotide metabolism, supported by metabolomic analysis which indicated that adenosine and purine metabolism are significantly altered by EBV immortalization. We further validated that adenosine deaminase (ADA) is a direct and critical target of the EBV-directed immortalization process. These findings reveal that purine metabolism and ADA may be useful therapeutic targets for EBV-driven lymphoid cancers.
Subject(s)
B-Lymphocytes/pathology , Cell Transformation, Viral , Chromatin/genetics , Epstein-Barr Virus Infections/pathology , Epstein-Barr Virus Nuclear Antigens/metabolism , Nucleotides/metabolism , Viral Proteins/metabolism , B-Lymphocytes/metabolism , B-Lymphocytes/virology , Chromatin/metabolism , Epigenesis, Genetic , Epstein-Barr Virus Infections/metabolism , Epstein-Barr Virus Infections/virology , Epstein-Barr Virus Nuclear Antigens/genetics , Herpesvirus 4, Human/physiology , Host-Pathogen Interactions , Humans , Metabolome , Transcriptome , Viral Proteins/geneticsABSTRACT
In sub-Saharan Africa, endemic Kaposi's sarcoma (EnKS) is still prevalent despite high incidence of epidemic Kaposi's sarcoma (EpKS) resulting from the on-going HIV-1 epidemic. While KSHV is clearly the etiologic agent of KS, the mechanisms underlying KS development are not fully understood. For example, HIV-1 co-infection and concomitant immune dysfunction have been associated with EpKS development. However, the direct or indirect role(s) of HIV-1, and therefore of immune suppression, in EpKS remains unclear. How, or whether, EpKS is mechanistically distinct from EnKS is unknown. Thus, the absence of HIV-1 co-infection in EnKS provides a unique control for investigating and deciphering whether HIV-1 plays a direct or indirect role in the EpKS tumor microenvironment. We hypothesized that HIV-1 co-infection would induce transcriptome changes that differentiate EpKS from EnKS, thereby defining the direct intra-tumor role of HIV-1 in KS. Comparison of ART-treated and -naïve patients would further define the impact of ART on the KS transcriptome. We utilized RNA-seq followed by multiparameter bioinformatics analysis to compare transcriptomes from KS lesions to uninvolved control skin. We provide the first transcriptomic comparison of EpKS versus EnKS, ART-treated vs-naïve EpKS and male vs female EpKS to define the roles of HIV-1 co-infection, the impact of ART, and gender on KS gene expression profiles. Our findings suggest that ART-use and gender have minimal impact on transcriptome profiles of KS lesions. Gene expression profiles strongly correlated between EpKS and EnKS patients (Spearman r = 0.83, p<10-10). A subset of genes involved in tumorigenesis and inflammation/immune responses showed higher magnitude, but not unique dysregulation in EnKS compared to EpKS. While gender and ART had no detectable contribution, the trend toward higher magnitude of gene dysregulation in EnKS coupled with the absence of HIV-1 transcripts in EpKS may suggest an indirect or systemic effect of HIV-1 to promote KS tumorigenesis.
Subject(s)
Coinfection/genetics , HIV-1 , Herpesvirus 8, Human , Sarcoma, Kaposi/genetics , Sarcoma, Kaposi/virology , Adult , Female , Gene Expression Profiling , HIV Infections/complications , Humans , Male , Middle Aged , Young AdultABSTRACT
Kaposi's sarcoma-associated herpesvirus (KSHV) is the etiologic agent of Kaposi's sarcoma (KS). It is endemic in a number of sub-Saharan African countries with infection rate of >50%. The high prevalence of HIV-1 coupled with late presentation of advanced cancer staging make KS the leading cancer in the region with poor prognosis and high mortality. Disease markers and cellular functions associated with KS tumorigenesis remain ill-defined. Several studies have attempted to investigate changes of the gene profile with in vitro infection of monoculture models, which are not likely to reflect the cellular complexity of the in vivo lesion environment. Our approach is to characterize and compare the gene expression profile in KS lesions versus non-cancer tissues from the same individual. Such comparisons could identify pathways critical for KS formation and maintenance. This is the first study that utilized high throughput RNA-seq to characterize the viral and cellular transcriptome in tumor and non-cancer biopsies of African epidemic KS patients. These patients were treated anti-retroviral therapy with undetectable HIV-1 plasma viral load. We found remarkable variability in the viral transcriptome among these patients, with viral latency and immune modulation genes most abundantly expressed. The presence of KSHV also significantly affected the cellular transcriptome profile. Specifically, genes involved in lipid and glucose metabolism disorder pathways were substantially affected. Moreover, infiltration of immune cells into the tumor did not prevent KS formation, suggesting some functional deficits of these cells. Lastly, we found only minimal overlaps between our in vivo cellular transcriptome dataset with those from in vitro studies, reflecting the limitation of in vitro models in representing tumor lesions. These findings could lead to the identification of diagnostic and therapeutic markers for KS, and will provide bases for further mechanistic studies on the functions of both viral and cellular genes that are involved.
Subject(s)
Energy Metabolism/genetics , Glucose/metabolism , Lipid Metabolism/genetics , Sarcoma, Kaposi/genetics , Sarcoma, Kaposi/metabolism , AIDS-Related Opportunistic Infections/genetics , AIDS-Related Opportunistic Infections/metabolism , Adult , DNA, Viral/analysis , HIV-1 , Herpesvirus 8, Human/genetics , Herpesvirus 8, Human/physiology , Humans , Male , Middle Aged , Polymerase Chain Reaction , Sarcoma, Kaposi/virology , Sequence Analysis, RNA , Tanzania , Viral Load/genetics , ZambiaABSTRACT
About 15% of human cancer cases are attributed to viral infections. To date, virus expression in tumor tissues has been mostly studied by aligning tumor RNA sequencing reads to databases of known viruses. To allow identification of divergent viruses and rapid characterization of the tumor virome, we develop viRNAtrap, an alignment-free pipeline to identify viral reads and assemble viral contigs. We utilize viRNAtrap, which is based on a deep learning model trained to discriminate viral RNAseq reads, to explore viral expression in cancers and apply it to 14 cancer types from The Cancer Genome Atlas (TCGA). Using viRNAtrap, we uncover expression of unexpected and divergent viruses that have not previously been implicated in cancer and disclose human endogenous viruses whose expression is associated with poor overall survival. The viRNAtrap pipeline provides a way forward to study viral infections associated with different clinical conditions.
Subject(s)
Deep Learning , Neoplasms , Viruses , Humans , Neoplasms/genetics , Viruses/genetics , Genome, Viral , High-Throughput Nucleotide SequencingABSTRACT
Lineage-defining transcription factors form densely interconnected circuits in chromatin occupancy assays, but the functional significance of these networks remains underexplored. We reconstructed the functional topology of a leukemia cell transcription network from the direct gene-regulatory programs of eight core transcriptional regulators established in pre-steady state assays coupling targeted protein degradation with nascent transcriptomics. The core regulators displayed narrow, largely non-overlapping direct transcriptional programs, forming a sparsely interconnected functional hierarchy stabilized by incoherent feed-forward loops. BET bromodomain and CDK7 inhibitors disrupted the core regulators' direct programs, acting as mixed agonists/antagonists. The network is predictive of dynamic gene expression behaviors in time-resolved assays and clinically relevant pathway activity in patient populations.
ABSTRACT
DAXX and ATRX are tumor suppressor proteins that form a histone H3.3 chaperone complex and are frequently mutated in cancers with the alternative lengthening of telomeres (ALT). Here, we show that DAXX and ATRX knock-out (KO) U87-T cells that have acquired ALT-like features have defects in p53 chromatin binding and DNA damage response. RNA-seq analysis revealed that p53 pathway is among the most perturbed. ChIP-seq and ATAC-seq revealed a genome-wide reduction in p53 DNA-binding and corresponding loss of chromatin accessibility at many p53 response elements across the genome. Both DAXX and ATRX null cells showed a depletion of histone H3.3 and accumulation of γH2AX at many p53 sites, including subtelomeres. These findings indicate that loss of DAXX or ATRX can compromise p53 chromatin binding and p53 DNA damage response in ALT-like cells, providing a link between histone composition, chromatin accessibility and tumor suppressor function of p53.
Subject(s)
Chromatin , Histones , Co-Repressor Proteins , DNA Damage , DNA Helicases , Genes, Tumor Suppressor , Molecular Chaperones , Nuclear Proteins , Tumor Suppressor Protein p53 , X-linked Nuclear ProteinABSTRACT
Patient-derived xenografts (PDXs) are resected human tumors engrafted into mice for preclinical studies and therapeutic testing. It has been proposed that the mouse host affects tumor evolution during PDX engraftment and propagation, affecting the accuracy of PDX modeling of human cancer. Here, we exhaustively analyze copy number alterations (CNAs) in 1,451 PDX and matched patient tumor (PT) samples from 509 PDX models. CNA inferences based on DNA sequencing and microarray data displayed substantially higher resolution and dynamic range than gene expression-based inferences, and they also showed strong CNA conservation from PTs through late-passage PDXs. CNA recurrence analysis of 130 colorectal and breast PT/PDX-early/PDX-late trios confirmed high-resolution CNA retention. We observed no significant enrichment of cancer-related genes in PDX-specific CNAs across models. Moreover, CNA differences between patient and PDX tumors were comparable to variations in multiregion samples within patients. Our study demonstrates the lack of systematic copy number evolution driven by the PDX mouse host.
Subject(s)
DNA Copy Number Variations/genetics , Xenograft Model Antitumor Assays , Animals , Databases, Genetic , Gene Expression Regulation, Neoplastic , Humans , Mice , Neoplasm Metastasis , Polymorphism, Single Nucleotide/genetics , Exome SequencingABSTRACT
Anti-PD-1 therapy is used as a front-line treatment for many cancers, but mechanistic insight into this therapy resistance is still lacking. Here we generate a humanized (Hu)-mouse melanoma model by injecting fetal liver-derived CD34+ cells and implanting autologous thymus in immune-deficient NOD-scid IL2Rγnull (NSG) mice. Reconstituted Hu-mice are challenged with HLA-matched melanomas and treated with anti-PD-1, which results in restricted tumor growth but not complete regression. Tumor RNA-seq, multiplexed imaging and immunohistology staining show high expression of chemokines, as well as recruitment of FOXP3+ Treg and mast cells, in selective tumor regions. Reduced HLA-class I expression and CD8+/Granz B+ T cells homeostasis are observed in tumor regions where FOXP3+ Treg and mast cells co-localize, with such features associated with resistance to anti-PD-1 treatment. Combining anti-PD-1 with sunitinib or imatinib results in the depletion of mast cells and complete regression of tumors. Our results thus implicate mast cell depletion for improving the efficacy of anti-PD-1 therapy.
Subject(s)
Drug Resistance, Neoplasm , Lymphocytes, Tumor-Infiltrating/immunology , Mast Cells/immunology , Programmed Cell Death 1 Receptor/antagonists & inhibitors , Animals , B-Lymphocytes/drug effects , B-Lymphocytes/immunology , Drug Resistance, Neoplasm/drug effects , Humans , Immune Checkpoint Inhibitors/pharmacology , Lymphocytes, Tumor-Infiltrating/drug effects , Mast Cells/drug effects , Melanoma/immunology , Melanoma/pathology , Melanoma/therapy , Mice, Transgenic , Programmed Cell Death 1 Receptor/metabolism , Sunitinib/pharmacology , Sunitinib/therapeutic use , T-Lymphocytes/drug effects , T-Lymphocytes/immunologyABSTRACT
Low-dose CT (LDCT) is widely accepted as the preferred method for detecting pulmonary nodules. However, the determination of whether a nodule is benign or malignant involves either repeated scans or invasive procedures that sample the lung tissue. Noninvasive methods to assess these nodules are needed to reduce unnecessary invasive tests. In this study, we have developed a pulmonary nodule classifier (PNC) using RNA from whole blood collected in RNA-stabilizing PAXgene tubes that addresses this need. Samples were prospectively collected from high-risk and incidental subjects with a positive lung CT scan. A total of 821 samples from 5 clinical sites were analyzed. Malignant samples were predominantly stage 1 by pathologic diagnosis and 97% of the benign samples were confirmed by 4 years of follow-up. A panel of diagnostic biomarkers was selected from a subset of the samples assayed on Illumina microarrays that achieved a ROC-AUC of 0.847 on independent validation. The microarray data were then used to design a biomarker panel of 559 gene probes to be validated on the clinically tested NanoString nCounter platform. RNA from 583 patients was used to assess and refine the NanoString PNC (nPNC), which was then validated on 158 independent samples (ROC-AUC = 0.825). The nPNC outperformed three clinical algorithms in discriminating malignant from benign pulmonary nodules ranging from 6-20 mm using just 41 diagnostic biomarkers. Overall, this platform provides an accurate, noninvasive method for the diagnosis of pulmonary nodules in patients with non-small cell lung cancer. SIGNIFICANCE: These findings describe a minimally invasive and clinically practical pulmonary nodule classifier that has good diagnostic ability at distinguishing benign from malignant pulmonary nodules.
Subject(s)
Biomarkers, Tumor/genetics , Carcinoma, Non-Small-Cell Lung/diagnosis , Gene Expression Profiling , Lung Neoplasms/diagnosis , Multiple Pulmonary Nodules/diagnosis , Tomography, X-Ray Computed/methods , Aged , Algorithms , Biomarkers, Tumor/blood , Carcinoma, Non-Small-Cell Lung/blood , Carcinoma, Non-Small-Cell Lung/diagnostic imaging , Carcinoma, Non-Small-Cell Lung/genetics , Diagnosis, Differential , Female , Gene Expression Regulation, Neoplastic , Humans , Lung Neoplasms/blood , Lung Neoplasms/diagnostic imaging , Lung Neoplasms/genetics , Male , Middle Aged , Multiple Pulmonary Nodules/blood , Multiple Pulmonary Nodules/diagnostic imaging , Multiple Pulmonary Nodules/genetics , Prospective StudiesABSTRACT
Due to their cytotoxic activities, many anticancer drugs cause extensive damage to the intestinal mucosa and have antibiotic activities. Here, we show that cisplatin induces significant changes in the repertoire of intestinal commensal bacteria that exacerbate mucosal damage. Restoration of the microbiota through fecal-pellet gavage drives healing of cisplatin-induced intestinal damage. Bacterial translocation to the blood stream is correspondingly abrogated, resulting in a significant reduction in systemic inflammation, as evidenced by decreased serum IL-6 and reduced mobilization of granulocytes. Mechanistically, reversal of dysbiosis in response to fecal gavage results in the production of protective mucins and mobilization of CD11b+ myeloid cells to the intestinal mucosa, which promotes angiogenesis. Administration of Ruminococcus gnavus, a bacterial strain selectively depleted by cisplatin treatment, could only partially restore the integrity of the intestinal mucosa and reduce systemic inflammation, without measurable increases in the accumulation of mucin proteins. Together, our results indicate that reconstitution of the full repertoire of intestinal bacteria altered by cisplatin treatment accelerates healing of the intestinal epithelium and ameliorates systemic inflammation. Therefore, fecal microbiota transplant could paradoxically prevent life-threatening bacteremia in cancer patients treated with chemotherapy.
Subject(s)
Cisplatin/pharmacology , Dysbiosis/therapy , Fecal Microbiota Transplantation , Intestines/microbiology , Ovarian Neoplasms/drug therapy , Peritoneal Neoplasms/drug therapy , Animals , Antineoplastic Agents/pharmacology , Dysbiosis/mortality , Dysbiosis/pathology , Female , Gastrointestinal Microbiome , Intestines/drug effects , Intestines/pathology , Ovarian Neoplasms/microbiology , Ovarian Neoplasms/pathology , Peritoneal Neoplasms/microbiology , Peritoneal Neoplasms/pathology , Tumor Cells, CulturedABSTRACT
Both p150 and p110 isoforms of ADAR1 convert adenosine to inosine in double-stranded RNA (dsRNA). ADAR1p150 suppresses the dsRNA-sensing mechanism that activates MDA5-MAVS-IFN signaling in the cytoplasm. In contrast, the biological function of the ADAR1p110 isoform, which is usually located in the nucleus, is largely unknown. Here, we show that stress-activated phosphorylation of ADAR1p110 by MKK6-p38-MSK MAP kinases promotes its binding to Exportin-5 and its export from the nucleus. After translocating to the cytoplasm, ADAR1p110 suppresses apoptosis in stressed cells by protecting many antiapoptotic gene transcripts that contain 3'-untranslated-region dsRNA structures primarily comprising inverted Alu repeats. ADAR1p110 competitively inhibits binding of Staufen1 to the 3'-untranslated-region dsRNAs and antagonizes Staufen1-mediated mRNA decay. Our study reveals a new stress-response mechanism in which human ADAR1p110 and Staufen1 regulate surveillance of a set of mRNAs required for survival of stressed cells.
Subject(s)
Adenosine Deaminase/metabolism , Apoptosis , Cytoskeletal Proteins/antagonists & inhibitors , RNA Stability , RNA-Binding Proteins/antagonists & inhibitors , RNA-Binding Proteins/metabolism , Humans , MAP Kinase Signaling System , Phosphorylation , Protein Processing, Post-Translational , Stress, PhysiologicalABSTRACT
PURPOSE: To define the safety and effectiveness of T cells redirected against follicle-stimulating hormone receptor (FSHR)-expressing ovarian cancer cells. EXPERIMENTAL DESIGN: FSHR expression was determined by Western blotting, immunohistochemistry, and qPCR in 77 human ovarian cancer specimens from 6 different histologic subtypes and 20 human healthy tissues. The effectiveness of human T cells targeted with full-length FSH in vivo was determined against a panel of patient-derived xenografts. Safety and effectiveness were confirmed in immunocompetent tumor-bearing mice, using constructs targeting murine FSHR and syngeneic T cells. RESULTS: FSHR is expressed in gynecologic malignancies of different histologic types but not in nonovarian healthy tissues. Accordingly, T cells expressing full-length FSHR-redirected chimeric receptors mediate significant therapeutic effects (including tumor rejection) against a panel of patient-derived tumors in vivo In immunocompetent mice growing syngeneic, orthotopic, and aggressive ovarian tumors, fully murine FSHR-targeted T cells also increased survival without any measurable toxicity. Notably, chimeric receptors enhanced the ability of endogenous tumor-reactive T cells to abrogate malignant progression upon adoptive transfer into naïve recipients subsequently challenged with the same tumor. Interestingly, FSHR-targeted T cells persisted as memory lymphocytes without noticeable PD-1-dependent exhaustion during end-stage disease, in the absence of tumor cell immunoediting. However, exosomes in advanced tumor ascites diverted the effector activity of this and other chimeric receptor-transduced T cells away from targeted tumor cells. CONCLUSIONS: T cells redirected against FSHR+ tumor cells with full-length FSH represent a promising therapeutic alternative against a broad range of ovarian malignancies, with negligible toxicity even in the presence of cognate targets in tumor-free ovaries. Clin Cancer Res; 23(2); 441-53. ©2016 AACR.
Subject(s)
Immunotherapy , Ovarian Neoplasms/therapy , Receptors, FSH/immunology , T-Lymphocytes/immunology , Animals , Ascites/immunology , Ascites/pathology , Exosomes/immunology , Exosomes/pathology , Female , Gene Expression Regulation, Neoplastic/immunology , Humans , Immunohistochemistry , Mice , Ovarian Neoplasms/genetics , Ovarian Neoplasms/immunology , Ovarian Neoplasms/pathology , Receptors, Antigen, T-Cell/immunology , Receptors, FSH/genetics , Xenograft Model Antitumor AssaysABSTRACT
Special AT-rich sequence-binding protein 1 (Satb1) governs genome-wide transcriptional programs. Using a conditional knockout mouse, we find that Satb1 is required for normal differentiation of conventional dendritic cells (DCs). Furthermore, Satb1 governs the differentiation of inflammatory DCs by regulating major histocompatibility complex class II (MHC II) expression through Notch1 signaling. Mechanistically, Satb1 binds to the Notch1 promoter, activating Notch expression and driving RBPJ occupancy of the H2-Ab1 promoter, which activates MHC II transcription. However, tumor-driven, unremitting expression of Satb1 in activated Zbtb46(+) inflammatory DCs that infiltrate ovarian tumors results in an immunosuppressive phenotype characterized by increased secretion of tumor-promoting Galectin-1 and IL-6. In vivo silencing of Satb1 in tumor-associated DCs reverses their tumorigenic activity and boosts protective immunity. Therefore, dynamic fluctuations in Satb1 expression govern the generation and immunostimulatory activity of steady-state and inflammatory DCs, but continuous Satb1 overexpression in differentiated DCs converts them into tolerogenic/pro-inflammatory cells that contribute to malignant progression.
Subject(s)
Dendritic Cells/immunology , Gene Expression Regulation, Neoplastic , Histocompatibility Antigens Class II/immunology , Matrix Attachment Region Binding Proteins/immunology , Ovarian Neoplasms/immunology , Animals , Cell Differentiation , Cell Proliferation , Cell Transformation, Neoplastic , Dendritic Cells/pathology , Female , Galectin 1/genetics , Galectin 1/immunology , Histocompatibility Antigens Class II/genetics , Histones/genetics , Histones/immunology , Humans , Immune Tolerance , Immunoglobulin J Recombination Signal Sequence-Binding Protein/genetics , Immunoglobulin J Recombination Signal Sequence-Binding Protein/immunology , Interleukin-6/genetics , Interleukin-6/immunology , Matrix Attachment Region Binding Proteins/antagonists & inhibitors , Matrix Attachment Region Binding Proteins/genetics , Mice , Mice, Knockout , Neoplasm Transplantation , Ovarian Neoplasms/genetics , Ovarian Neoplasms/pathology , Promoter Regions, Genetic , Protein Binding , RNA, Small Interfering/genetics , RNA, Small Interfering/immunology , Receptor, Notch1/genetics , Receptor, Notch1/immunology , Signal Transduction , Transcription Factors/genetics , Transcription Factors/immunologyABSTRACT
The emergence of tumor cells with certain stem-like characteristics, such as high aldehyde dehydrogenase (ALDH) activity due to ALDH1A1 expression, contributes to chemotherapy resistance and tumor relapse. However, clinically applicable inhibitors of ALDH activity have not been reported. There is evidence to suggest that epigenetic regulation of stem-related genes contributes to chemotherapy efficacy. Here, we show that bromodomain and extraterminal (BET) inhibitors suppress ALDH activity by abrogating BRD4-mediated ALDH1A1 expression through a super-enhancer element and its associated enhancer RNA. The clinically applicable small-molecule BET inhibitor JQ1 suppressed the outgrowth of cisplatin-treated ovarian cancer cells both in vitro and in vivo Combination of JQ1 and cisplatin improved the survival of ovarian cancer-bearing mice in an orthotopic model. These phenotypes correlate with inhibition of ALDH1A1 expression through a super-enhancer element and other stem-related genes in promoter regions bound by BRD4. Thus, targeting the BET protein BRD4 using clinically applicable small-molecule inhibitors, such as JQ1, is a promising strategy for targeting ALDH activity in epithelial ovarian cancer. Cancer Res; 76(21); 6320-30. ©2016 AACR.