Phenotypic screening with deep learning identifies HDAC6 inhibitors as cardioprotective in a BAG3 mouse model of dilated cardiomyopathy

Dilated cardiomyopathy (DCM) is characterized by reduced cardiac output, as well as thinning and enlargement of left ventricular chambers. These characteristics eventually lead to heart failure. Current standards of care do not target the underlying molecular mechanisms associated with genetic forms of heart failure, driving a need to develop novel therapeutics for DCM. To identify candidate therapeutics, we developed an in vitro DCM model using induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) deficient in B-cell lymphoma 2 (BCL2)-associated athanogene 3 (BAG3). With these BAG3-deficient iPSC-CMs, we identified cardioprotective drugs using a phenotypic screen and deep learning. From a library of 5500 bioactive compounds and siRNA validation, we found that inhibiting histone deacetylase 6 (HDAC6) was cardioprotective at the sarcomere level. We translated this finding to a BAG3 cardiomyocyte-knockout (BAG3^cKO) mouse model of DCM, showing that inhibiting HDAC6 with two isoform-selective inhibitors (tubastatin A and a novel inhibitor TYA-018) protected heart function. In BAG3^cKO and BAG3^E455K mice, HDAC6 inhibitors improved left ventricular ejection fraction and reduced left ventricular diameter at diastole and systole. In BAG3^cKO mice, TYA-018 protected against sarcomere damage and reduced Nppb expression. Based on integrated transcriptomics and proteomics and mitochondrial function analysis, TYA-018 also enhanced energetics in these mice by increasing expression of targets associated with fatty acid metabolism, protein metabolism, and oxidative phosphorylation. Our results demonstrate the power of combining iPSC-CMs with phenotypic screening and deep learning to accelerate drug discovery, and they support developing novel therapies that address underlying mechanisms associated with heart disease.

Abstract P436: Modeling Genetic Dilated Cardiomyopathy Using Induced Pluripotent Stem Cell-derived Engineered Heart Tissues For Precision Medicine

Recent advancement of human induced pluripotent stem cell-derived cardiomyocyte (hiPSC-CM) technologies has opened the door to next-generation modeling of human cardiac biology and disease. Not only does this alleviate the need for human primary tissue and compensate for the well documented deficiencies of rodent models for cardiovascular disease, but these technologies may lead to the identification of better candidates for clinical development. Unfortunately, most current 2D hiPSC-CM models lack the biochemical, mechanical, and electrical feedback that cardiomyocytes endure in a multi-cellular aligned tissue, which limits their translatability into clinical settings. To address this, we incorporated hiPSC-CMs modeling various genetic dilated cardiomyopathies (DCM) into 3D engineered heart tissues (EHTs). Here, we show these models develop distinguishable contractile defects recapitulating hallmarks of DCM when compared to biologically relevant controls. Moreover, this distinct phenotype is quantitative, reproducible, and demonstrates utility for drug discovery. As this innovative technology continues to develop, EHTs are on the forefront of emerging biomimetic assays that can be used to prevent drug attrition in the late stages of drug development.

Deep learning detects cardiotoxicity in a high-content screen with induced pluripotent stem cell-derived cardiomyocytes

Drug-induced cardiotoxicity and hepatotoxicity are major causes of drug attrition. To decrease late-stage drug attrition, pharmaceutical and biotechnology industries need to establish biologically relevant models that use phenotypic screening to detect drug-induced toxicity in vitro. In this study, we sought to rapidly detect patterns of cardiotoxicity using high-content image analysis with deep learning and induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs). We screened a library of 1280 bioactive compounds and identified those with potential cardiotoxic liabilities in iPSC-CMs using a single-parameter score based on deep learning. Compounds demonstrating cardiotoxicity in iPSC-CMs included DNA intercalators, ion channel blockers, epidermal growth factor receptor, cyclin-dependent kinase, and multi-kinase inhibitors. We also screened a diverse library of molecules with unknown targets and identified chemical frameworks that show cardiotoxic signal in iPSC-CMs. By using this screening approach during target discovery and lead optimization, we can de-risk early-stage drug discovery. We show that the broad applicability of combining deep learning with iPSC technology is an effective way to interrogate cellular phenotypes and identify drugs that may protect against diseased phenotypes and deleterious mutations.

Abstract 5621: Prevalence of GITR expression and pharmacodynamic (PD) biomarkers in syngeneic tumor models treated by a GITR agonist (GITRL-Fc)

GITRL (Glucocorticoid-Induced Tumor Necrosis Factor Receptor Ligand, TNFSF18) is a member of the tumor necrosis factor (TNF) ligand superfamily. GITRL binds and activates the co-stimulatory surface receptor GITR, which promotes proliferation and activation of effector T cells (Teff) and inhibits suppressive activity of regulatory T cells (Treg). It is thus hypothesized that co-stimulation of GITR by agonist agents will promote anti-tumor immunity. We generated a novel single-gene GITRL trimer fused to an immunoglobulin Fc domain (GITRL-Fc) that shows robust single agent antitumor efficacy and immune effects in multiple syngeneic mouse models, suggesting its potential benefit in cancer immunotherapy To investigate the prevalence of GITR expression in human tumors, RNA-Seq data analyses of 33 tumor types in TCGA showed GITR is highly expressed in a subset of solid tumors, including head & neck, lung, breast, esophageal, and bladder cancers. In most solid tumors, GITR expression correlated poorly with T cell markers, implying that GITR may not be exclusive to immune cells and may be expressed in tumor cells as well. Similar findings emerged from RNA-Seq data analysis of patient-derived xenograft (PDX) samples from 24 tumor types. The gene expression data was corroborated by immunohistochemistry (IHC) analysis of GITR expression in 17 tumor types which showed that in addition to immune cells, GITR was expressed on tumor cell membranes. A multi-platform approach was taken to investigate GITRL-Fc pharmacodynamic (PD) biomarkers in tumors and in matched whole blood samples from mice bearing CT26 colon, 4T1 breast, and B16F10 melanoma carcinoma models. Global gene expression levels were profiled by microarray on treated and control tissues. We also monitored the changes of immune cell populations and cytokine secretions by flow cytometry, Luminex and IHC. Immune gene changes were more robust in tumors than in blood samples. In tumor samples, GITRL-Fc increased the gene expression associated with T cells, CD8 T cells, cytotoxicity, Th1 cells, interferon gamma (IFN-γ), natural killer cells, Teff cells, and T cell activation markers. These gene changes were validated by quantitative real-time PCR. Similarly, flow cytometry analysis showed that GITRL-Fc promoted activation of CD4+ effector cells, decreased Treg frequency, and increased the ratio of CD8+ T cell/Treg in the tumor. GITRL-Fc also modulated secretion of cytokines in splenocytes, including an increase in IFN-γ. Taken together, the PD biomarker changes in immune-related gene expression, immune cell populations, and cytokine secretions observed in these preclinical tumor models are consistent with GITRL-Fc mechanism of action and demonstrated target engagement of GITRL-Fc. Additional approaches, including in-silico sorting, to monitor rare immune cell populations in tumor samples will be discussed.

Citation Format: Min Wang, Fiore Cattaruzza, Pete Yeung, Alayne Brunner, Erwan LeScolan, Yuwang Liu, Jennifer Cain, Gilbert OYoung, Earth Light Lowe, Belinda Cancilla, Rose Harris, Tim Hoey, Austin Gurney, John Lewicki, Gretchen Argast, Ann M. Kapoun. Prevalence of GITR expression and pharmacodynamic (PD) biomarkers in syngeneic tumor models treated by a GITR agonist (GITRL-Fc) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 5621. doi:10.1158/1538-7445.AM2017-5621

Paracrine wnt-β-catenin signaling inhibition as a strategy to enhance the efficacy of anti-PD-1 antibody (Ab) therapy in a transgenic model of melanoma

3053

Background: Activation of the Wnt-β-catenin signaling pathway is associated with poor T cell infiltration of tumors. We have previously demonstrated that paracrine Wnt5a-β-catenin signaling is a critical trigger of dendritic cell (DC) tolerization and regulatory T cell (Treg) differentiation in the melanoma microenvironment. In a transgenic BRAFV600EPTEN-/- model, the genetic silencing of melanoma-derived Wnt5a potently enhances infiltrating CD8+T cell effector function and promotes responses to anti-PD-1 Ab therapy. Ipafricept (IPA) is a recombinant Wnt decoy receptor and Vantictumab (VAN) is a Fzd receptor monoclonal Ab. Both molecules inhibit Wnt-β-catenin signaling and have been well-tolerated in ongoing phase I/Ib clinical trials. We explored the ability of IPA/VAN to reverse tumor-mediated immune tolerance and enhance the efficacy of anti-PD-1 Ab immunotherapy in a pre-clinical model that closely recapitulates human melanoma. Methods: Both IPA and VAN were utilized to investigate Wnt-β-catenin inhibition as a strategy for suppressing melanoma-induced DC indoleamine 2,3-dioxygenase (IDO) expression and Treg differentiation in vitro. These agents were further tested for their ability to enhance anti-tumor T cell responses and to augment the efficacy of anti-PD-1 Ab therapy in syngeneic and autochthonous models of BRAFV600EPTEN-/- melanoma. Results: IPA and VAN effectively inhibit Wnt5a and melanoma-induced DC IDO expression and Treg differentiation in vitro. Further studies demonstrate that IPA and VAN significantly augment anti-PD-1 Ab-mediated suppression of primary and metastatic tumor progression in both syngeneic and autochthonous BRAFV600EPTEN-/- melanoma models. These anti-tumor effects correlated with suppressed IDO enzymatic activity, enhanced tumor-infiltrating CD8+T cell/Treg ratios, and increased activation of TRP2 antigen-specific effector T cells. Conclusions: The pharmacological inhibition of paracrine Wnt-β-catenin signaling with IPA and VAN augment the anti-tumor efficacy of anti-PD-1 Ab therapy and represent a promising strategy for further phase I testing in melanoma and other solid tumors.

Abstract A42: Using a PDX tumor bank to screen for cancer stem cell therapies

OncoMed Pharmaceuticals is focused on discovering novel therapies that target cancer stem cells (CSCs), specifically those which depend on the Notch or Wnt pathways. Since Patient Derived Xenografts (PDXs) recapitulate both tumor cell heterogeneity and maintain the histopathological characteristics of the original tumor, they represent an important preclinical model for CSC drug discovery and can be used to effectively screen for new therapeutic candidates. To this end, OncoMed has established a fully characterized PDX Tumor Bank for its monoclonal antibody discovery and developmental efforts and has used PDX models to advance seven candidates into clinical trials. The goal of this study is to: 1) review the operational steps needed to propagate PDX's; 2) define the quality controls necessary to maintain these tumors; 3) assess how the Tumor Bank has helped in the selection and advancement of these targeted biologics into the clinic; and 4) review the role of the Tumor Bank in defining biomarkers for clinical use.

Tumor specimens were received from surgery and processed into fragments, cell clumps, or dissociated cells. Tumor tissue or cells were either implanted into NOG or Nod/scid mice or frozen for later implantation. Primary tumors that grew were serially transplanted to establish working stocks for drug screening experiments. Every tumor was characterized for a variety of molecular, cellular, and tumor endpoints. All pertinent data and samples were captured into a custom designed database which included lineage diagrams to easily track tumor propagation and characterization endpoints.

The quality control of the established PDXs must be carefully monitored throughout the process. We have used DNA fingerprinting to ensure the identification of each tumor and subsequent passages, but the larger problem of identifying and monitoring the development of spontaneous lymphomas (both human and murine) that can infiltrate and contaminate the PDX required a rigorous monitoring strategy. The infection of mice with Lactate Dehydrogenase Elevating Virus from contaminated reagents can severely affect animal health and, therefore the screening process. Surprisingly, the misdiagnosis of tumors received from surgery was higher than expected and resulted in tumors being correctly reclassified before being used.

The strategy employed to effectively screen selected targets in the Notch and Wnt pathways was to first review the molecular characterization data from our PDX models, and then select appropriate models for in vivo efficacy testing. In order to assess how the Tumor Bank has helped both in the screening process and the identification of biomarkers, we will review both pre-clinical and clinical data for selected OncoMed antibodies. As an example, one anti-CSC agent, OMP-59R5 (Tarextumab), which targets Notch2/3 was tested in ten pancreatic PDX models. Six pancreatic PDX tumors were responders to anti-NOTCH2/3 while four were non-responders. Bioinformatic analysis of the responder/non-responder data sets identified tumors that had high Notch3 gene expression as responders to OMP-59R5 treatment. Based on this preclinical data, Notch3 levels were evaluated in a Phase 1b pancreatic trial as a potential predictive biomarker. In this trial, higher response rate and longer survival was noted in patients with Notch3 high tumors receiving GEM/Nab-P/ Tarextumab (at 5-15mg/kg). These observations are being tested in a placebo-controlled, randomized Ph2 setting.

The creation of a working PDX Tumor Bank across multiply human solid tumor types has allowed us to implement an effective preclinical screening program to select candidate biologics and potential predictive biomarkers for targets in the Notch and Wnt pathways.

Citation Format: James Evans, Chun Zhang, Angie InKyung Park, Alayne Brunner, Min Wang, Cristina Dee-Hoskins, Roger Lopez, Xiaomei Song, Kellie Pickell, Wan-Cheng Yen, Marcus Fischer, Raymond Tam, Gilbert O'Young, Jakob Dupont, Lei Zhou, Austin Gurney, John Lewicki, Tim Hoey, Ann M. Kapoun, Belinda Cancilla. Using a PDX tumor bank to screen for cancer stem cell therapies. [abstract]. In: Proceedings of the AACR Special Conference: Patient-Derived Cancer Models: Present and Future Applications from Basic Science to the Clinic; Feb 11-14, 2016; New Orleans, LA. Philadelphia (PA): AACR; Clin Cancer Res 2016;22(16_Suppl):Abstract nr A42.

Abstract A41: Use of patient-derived tumor xenografts (PDX) for discovery and development of an anti-Notch2/3 monoclonal antibody targeting cancer stem cells

Accumulating evidence has suggested that tumors may arise and grow as a result of the formation of a subset cell population termed cancer stem cells (CSC) or tumor initiating cells. Several research reports have indicated that CSCs are relatively resistant to conventional therapies. Thus, therapeutic strategies that specifically target cancer stem cells could have a major impact on cancer patient survival. Patient-derived tumor xenografts (PDX) have played a major role in the development of new cancer therapies. The advantage of PDX over standard cell-line xenograft models is that PDX retain much of the molecular, genetic, and histological heterogeneity of the original tumor and are minimally passed. In addition, self-renewal and lineage differentiation, the hallmarks of cancer stem cells, can be demonstrated through serial transplantation in immunodeficient mice. Our therapeutic approach in cancer stem cell drug discovery has been to target key developmental pathways that have been strongly implicated in cancer including the Notch pathway.

We carried out tumorigenicity studies in three SCLC PDX models utilizing naïve cells positively enriched for either Notch2 or Notch3 expression. In two of the three models, Notch2+ and/or Notch3+ cell populations resulted in enhanced tumor formation. In a SCLC tumor that expressed low levels of Notch3, the Notch2+ population resulted in 70% tumor formation compared to CD44+ (10%) or CD133+ (45%) enriched populations.

We have developed a monoclonal antibody, tarextumab (OMP-59R5), which selectively inhibits the function of both Notch2 and Notch3. Our preclinical data in PDX models demonstrate that tarextumab was efficacious in inhibiting the growth of various indications with minimal intestinal toxicity. Notably, the sensitivity of tarextumab in combination with gemcitabine or gemcitabine plus nab-paclitaxel in pancreatic tumors was associated with higher levels of Notch3 gene expression. Interference with Notch2/3 signaling by tarextumab delays tumor recurrence, decreases cancer stem cell frequency (as determined by in vivo LDA studies) and modulates the function of tumor vasculature.

Our ALPINE Phase 1b clinical trial indicates that tarextumab is generally well-tolerated and shows signs of anti-tumor efficacy and modulation of Notch pathway signaling in the clinic. Furthermore, we observe a higher response rate and longer survival in patients with Notch3 high tumors receiving gemcitabine/nab-paclitaxel/tarextumab combination therapy. Analysis of pre- and post-treatment tumor biopsies showed an inhibition of Notch pathway and CSC gene signatures. Ongoing Phase 2 clinical trials evaluate first-line treatment with tarextumab in metastatic pancreatic cancer (ALPINE) and small cell lung cancer (PINNACLE). Collectively, these results demonstrate the utility of PDX models for discovery and development of anti-cancer stem cell therapeutics, identifying pharmacodynamic endpoints of drug actions, identifying predictive biomarkers for patient stratification, and translating these preclinical findings into clinical trials.

Citation Format: Marcus M. Fischer, Wan-Ching Yen, Fumiko Axelrod, Christopher Bond, Jennifer Cain, Belinda Cancilla, Randall Henner, Rene Meisner, Aaron Sato, Jalpa Shaw, Tracy Tang, Breanna Wallace, Min Wang, Chun Zhang, Ann Kapoun, Lei Zhou, Jakob Dupont, John Lewicki, Austin Gurney, Tim Hoey. Use of patient-derived tumor xenografts (PDX) for discovery and development of an anti-Notch2/3 monoclonal antibody targeting cancer stem cells. [abstract]. In: Proceedings of the AACR Special Conference: Patient-Derived Cancer Models: Present and Future Applications from Basic Science to the Clinic; Feb 11-14, 2016; New Orleans, LA. Philadelphia (PA): AACR; Clin Cancer Res 2016;22(16_Suppl):Abstract nr A41.

Abstract 2214: GITR ligand fusion protein (GITRL-Fc) induces T cell mediated anti-tumor immune response and can combine with anti-PDL1 to enhance anti-tumor immunity and long-term immune memory

GITRL (Glucocorticoid-Induced Tumor Necrosis Factor Receptor Ligand, TNFSF18) is a member of the TNF superfamily and naturally exists as a membrane-anchored type II protein that self assembles as a trimer. GITRL activates the co-stimulatory receptor GITR. GITR is found primarily on activated T effector (Teff) cells and regulatory T (Treg) cells. Co-stimulation of GITR by agonist agents is hypothesized to promote anti-tumor immunity by enhancing Teff cell activity and inhibiting Treg suppression. We generated a novel single-gene GITRL trimer fused to an immunoglobulin Fc domain (GITRL-Fc). GITRL-Fc activated GITR signaling more effectively than prototype GITR agonist antibody DTA-1. GITRL-Fc promoted a robust anti-tumor immune response in multiple syngeneic mouse tumor models. GITRL-Fc enhanced tumor specific T-cell responses, particularly of the Th1 type, and also led to reduction in Treg-mediated immunesuppressive activity. GITRL-Fc displayed single agent activity in inhibiting tumor growth and promoting complete tumor rejection in the murine CT26 colon carcinoma model and combination activity with anti-PDL1 as compared to anti-PDL1 and control IgG2a alone. Mice “cured” with GITRL or GITRL/anti-PDL1 combination treatments were protected from re-challenge with tumor cells, suggesting the existence of immunologic memory. More mice were protected from tumor re-challenge with the combination of GITRL-Fc and anti-PDL1, as compared to GITRL-Fc alone. Our results demonstrate that agonist GITRL-Fc induces potent T cell responses, overcomes Treg inhibition, and promotes anti-tumor activity in preclinical models as a single agent or in combination with anti PDL1. The mechanism of tumor eradication and induction of long-term immune memory response by the combination is under investigation and will be discussed at the presentation.

Citation Format: Minu K. Srivastava, Rui Yun, Erin Mayes, Hyun_Bae Jie, Fumiko Axelrod, Jorge Monteon, Ming-Hong Xie, John Lewicki, Tim Hoey, Austin Gurney, Angie Inkyung Park. GITR ligand fusion protein (GITRL-Fc) induces T cell mediated anti-tumor immune response and can combine with anti-PDL1 to enhance anti-tumor immunity and long-term immune memory. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2214.

Abstract 4652: Effects of anti-DLL4 treatment on non-small cell lung cancer (NSCLC) human xenograft tumors

Background: Non-small cell lung cancer (NSCLC) accounts for the vast majority of lung cancers, the leading cause of cancer-related deaths. Notch signaling has been shown to play an important role in lung cancer initiation and progression. Delta-like ligand 4 (DLL4) activates the Notch pathway and is important for cancer stem cell (CSC) survival. Demcizumab (OMP-21M18) is a humanized IgG2 anti-DLL4 antibody currently being tested in a Phase 2 trial in combination with pemetrexed and carboplatin for first-line treatment of patients with NSCLC. Previously, OMP-21M18 in combination with its mouse anti-DLL4 surrogate has been shown to inhibit tumor growth, decrease cancer stem cell frequency, and cause dysfunctional sprouting of new vessels resulting in an anti-angiogenic effect in patient-derived tumor xenograft (PDX) models in breast, colon, ovarian, and pancreatic cancers. Here we show results from NSCLC PDX models.

Methods and Results: Anti-DLL4 treatment was tested in a series of NSCLC PDX models. Because DLL4 inhibition has been shown to have effects on the tumor as well as the vasculature, the combination of OMP-21M18 (targeting human DLL4) and 21R30 (antibody targeting mouse DLL4) treatment in the PDX models was used to model demcizumab treatment in humans. Treatment with anti-DLL4 in combination with chemotherapy inhibited tumor growth in a series of NSCLC PDX models. Additionally, a tumorigenicity assay showed a decrease in the frequency of tumor-initiating cells following treatment with anti-DLL4 and chemotherapy. Gene expression analysis of tumor samples provided insights into the mechanism of action.

Conclusions: Anti-DLL4 treatment in a panel of NSCLC PDX tumor models in vivo showed inhibition of tumor growth and a decrease in the frequency of tumor-initiating cells. Mechanism of action and gene expression analysis of these models treated with anti-DLL4 will be presented. These findings provide additional evidence supporting demcizumab as an effective treatment for NSCLC patients.

Citation Format: Alayne Brunner, Fiore Cattaruzza, Wan-Ching Yen, Pete Yeung, Marcus Fischer, Belinda Cancilla, Gilbert O’Young, Raymond Tam, Yu-Wang Liu, Austin Gurney, John Lewicki, Tim Hoey, Min Wang, Ann M. Kapoun. Effects of anti-DLL4 treatment on non-small cell lung cancer (NSCLC) human xenograft tumors. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4652.

Abstract 404: Development of a RSPO3 CLIA-validated assay as a predictive biomarker for response to anti-RSPO3 antibody treatment in patients with solid tumors

R-Spondin (RSPO) proteins bind to LGR receptors and potentiate Wnt/β-catenin signaling. We have identified a therapeutic anti-RSPO3 antibody targeting the RSPO-LGR pathway. In preclinical studies, RSPO3 gene expression has shown correlation with anti-RSPO3 antibody efficacy in multiple solid tumor types. A qPCR-based RSPO3 assay has been developed as a predictive biomarker for response to the anti-RSPO3 antibody. In addition, RSPO gene fusions may play a role in the activation of Wnt signaling. A gene fusion detection workflow consisting of a RSPO3 CLIA assay, a RSPO3 RUO assay and next generation sequencing (NGS) has also been developed.

We designed 6 qPCR-based assays for the RSPO3 CLIA assay development and 2 assays for the RUO assay. These assays were designed to span exon-exon junctions or target microarray probe set sequences. Amplification sensitivity and specificity were assessed for assay selection. The analytic performance of the candidate RSPO3 CLIA assay and quality control measures were established in a validation study. The validation study included: 1) performance specifications of the RSPO3 assay including analytical sensitivity, linearity, and precision, 2) determination of a reportable range, 3) establishment of a cut-off for the RSPO3 CLIA assay for patient selection, and 4) establishment of quality control procedures. 104 human cancer tissues and 24 independent patient-derived tumor xenografts (PDX) were used in these studies. To evaluate the fusion detection workflow, the RUO assay was performed on samples that tested above the CLIA assay cut-off. The delta Ct difference between the CLIA and RUO assays was calculated to identify potential fusions.

The limit of quantification was established for the RSPO3 CLIA assay. The 95% reference interval was estimated to be (-2.44, 16.02) with 90% confidence interval for the lower bound (-3.45, -2.12) and upper bound (15.26, 16.57). The delta Ct cut-off for the RSPO3 CLIA assay was set based on sensitivity, specificity and prevalence. No statistically significant difference in the total variance across the tested samples was observed. A549 and OV56 were identified to be cell line controls with established acceptable delta Ct limits. Using NGS, RSPO3 fusions were identified in 6 PDX tumors with delta Ct RUO - delta Ct CLIA>7, including a novel fusion. This cut-off was further refined with NGS of 9 clinical samples. Prevalence of the RSPO3 expression and fusions will be presented.

A qPCR based RSPO3 assay was developed and CLIA-validated for use as a potential predictive biomarker for response to anti-RSPO3 therapy. This RSPO3 CLIA assay, together with the fusion detection workflow, will be evaluated in a Phase 1a/b dose escalation study of anti-RSPO3 (OMP-131R10) in advanced solid tumors and in combination with FOLFIRI in metastatic colorectal cancer (NCT02482441).

Citation Format: Chun Zhang, Yuwang Liu, Min Wang, Gilbert OYoung, Joy Kavanagh, Cheryl McFarlane, Fiore Cattaruzza, Pete Yeung, Jennifer Cain, Wan-Ching Yen, Marcus Fischer, Belinda Cancilla, Edwina Dobbin, Michelle McCarthy, Austin Gurney, Leonardo Faoro, John Lewicki, Tim Hoey, Ann M. Kapoun. Development of a RSPO3 CLIA-validated assay as a predictive biomarker for response to anti-RSPO3 antibody treatment in patients with solid tumors. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 404.

Abstract P3-07-57: Development of a 6-gene qPCR RUO-validated assay as a predictive biomarker for response of vantictumab (OMP-18R5; anti-frizzled) in HER2- breast cancer patients

Background: We have developed a monoclonal antibody, vantictumab that blocks canonical WNT/β-catenin signaling through binding of five FZD receptors (1, 2, 5, 7, 8). This antibody inhibits the growth of several tumor types, including breast. Vantictumab reduces tumor-initiating cell frequency and exhibits synergistic activity with standard-of-care (SOC) agents (Gurney et al, 2012). To target breast cancer patients most likely to respond to vantictumab, we undertook a predictive biomarker study.

Methods: We have identified a 6-gene Wnt pathway-related signature, FBXW2, CCND2, RHOU, CTBP2, WIF1, and DKK1, based on microarray gene expression data from 8 breast cancer patient derived xenograft (PDX) models with established in vivo response to vantictumab plus SOC. This signature successfully predicted the response of 8 additional and independent PDX breast tumors. We further developed a qPCR Research Use Only (RUO) assay for the 6 genes to be used on FFPE human breast tumor samples. Multiple assays targeting different regions spanning each mRNA transcript were tested and selected based on PCR efficiency, specificity and sensitivity. We compared assay sensitivity under different cDNA synthesis and pre-amplification conditions: random vs. gene-specific priming, number of pre-amplification cycles, pre-amplification reaction volumes, and cDNA synthesis kits. A repeatability study was performed to test assay performance. The pre-amplification and PCR were repeated over three separate days and across two independent labs.

Results: Our results showed that cDNA synthesis by gene-specific priming followed by 18 cycles of pre-amplification performed the best and the assay is robust with minimal starting FFPE RNA input. The results of the repeatability study were consistent among the different days and the different labs (<5% CV). Using the 6-gene qPCR RUO assay, the signature score from the microarray data was further refined using 12 PDX HER2- breast tumors with known in vivo response to vantictumab with SOC. The prevalence of the 6-gene signature was established using ∼100 HER2- breast cancer samples.

Conclusions: A robust 6-gene RUO-validated assay was developed as a predictive biomarker for vantictumab in HER2- breast cancer. The assay is currently being evaluated in a Phase 1b study of vantictumab with paclitaxel in HER2- breast cancer.

Citation Format: Zhang C, O'Young G, Wikstrom K, Davison T, Yeung P, Cattaruzza F, Yen W-C, Hoey T, Lewicki J, Rachmann R, Kerr P, Hill L, Eason R, McErlean S, Liu Y, Kapoun AM. Development of a 6-gene qPCR RUO-validated assay as a predictive biomarker for response of vantictumab (OMP-18R5; anti-frizzled) in HER2- breast cancer patients. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P3-07-57.

Abstract A30: Predictive and pharmacodynamic biomarkers of vantictumab (OMP-18R5; anti-Frizzled) in non-small cell lung cancer

Background: Vantictumab is a monoclonal antibody that blocks canonical WNT/β-catenin signaling through binding of five FZD receptors (1, 2, 5, 7, 8). This antibody inhibits the growth of several tumor types, reduces tumor-initiating cell frequency (TIC) and exhibits synergistic activity with standard-of-care (SOC) chemotherapeutic agents (Gurney et al., 2012). To target responsive patients and understand the mechanism of action of the drug, we set out to identify predictive and pharmacodynamic (PD) biomarkers of vantictumab in non-small cell lung cancer (NSCLC).

Materials and methods: The response to vanticutmab was established from in vivo efficacy experiments including different treatment groups: control, vantictumab, paclitaxel and vantictumab in combination with paclitaxel. For combination treatment, same day dosing and sequential dosing (paclitaxel dosed 2 days after the antibody) were compared. Samples were collected for PD biomarker analysis. To identify a predictive biomarker for the response to vantictumab in NSCLC patients, gene expression data from 7 NSCLC patient derived xenograft (PDX) models was analyzed. We utilized support vector machine-recursive feature elimination (SVM-RFE, Guyon et al., 2002) to select genes and support vector machine (SVM) for classification.

Results: Vantictumab showed significant tumor growth inhibition as a single agent as well as in combination with paclitaxel. The reduction of TIC and the antitumor efficacy of vantictumab were significantly enhanced with sequential dosing compared with same day dosing. These findings suggested that optimal synergy occurs using sequential dosing, likely due to enhanced blockade of cell cycle progression at mitosis. PD biomarker analysis confirmed inhibition of genes in Wnt, Notch, and stem cell pathways by vantictumab both as a single agent and also in combination with paclitaxel. Wnt pathway targets including AXIN2 and LEF1 were down-regulated significantly by vantictumab in both sequential dosing and same day dosing confirming the mechanism of action. From a series of 7 in vivo efficacy PDX experiments, LEF1 was identified as a predictive biomarker of vantictumab response and achieved the best performance with cross-validated positive predictive value (PPV) = negative predictive value (NPV) = sensitivity = specificity = 100%. Strong correlation was also observed between LEF1 gene expression and the ratio of tumor volume. Furthermore, LEF1 was able to successfully predict the response to vantictumab in 2 independent NSCLC PDX models. Prevalence estimation for LEF1 ranged from 35% to 50% based on public microarray datasets. LEF1 was also found to be significantly correlated with the response to vantictumab in combination with paclitaxel in 12 NSCLC PDX models (p = 0.0162), indicating LEF1 as a potential predictive biomarker of the response vantictumab as a single agent and in combination with SOC in NSCLC.

Conclusions: A biomarker study for the pharmacodynamics and response to vantictumab was performed using a series of PDX NSCLC models. PD biomarkers were identified which confirmed the mechanism of action of vantictumab. LEF1 was identified as a predictive biomarker and is being evaluated in the Phase 1b study of vantictumab in combination with SOC in previously treated NSCLC: NCT01957007. Comprehensive PD and predictive biomarker data will be presented.

Citation Format: CHUN ZHANG, Fiore Cattaruzza, Pete Yeung, Wan-Ching Yen, Marcus Fischer, Alayne Brunner, Min Wang, Belinda Cancilla, Rainer Brachmann, Tim Hoey, John Lewicki, Ann M. Kapoun. Predictive and pharmacodynamic biomarkers of vantictumab (OMP-18R5; anti-Frizzled) in non-small cell lung cancer. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr A30.

Abstract 2322: The Hippo signaling pathway mediates BMP inhibition of cancer stem cells

Multiple developmental signaling pathways cooperatively regulate cancer stem cell (CSC) function. Among them, the BMP pathway has been implicated in brain CSC tumorigenicity. To examine whether BMP is a general regulator of CSCs across diverse organ systems we utilized lentivirus-mediated BMP4 expression in patient-derived xenograft models in NOD/SCID mice. The prevention of tumor growth was observed in breast, colon, ovarian, and melanoma models. Similarly, adenovirus-mediated systemic delivery of BMP4 inhibited tumor growth, dramatically affected CSC-specific cell surface markers and decreased CSC frequency. Microarray analysis of BMP4-treated tumors showed that a set of genes encoding for components of the actin cytoskeleton was significantly enriched amongst regulated genes and identified Hippo pathway genes as potential mediators of the BMP4 anti-tumor response with up-regulation of SAV1 and LATS2 and down-regulation of BIRC5. Hippo signaling regulates organ size, stem cell biology and cancer through a protein kinase complex including MST1/2, SAV1, and LATS1/2, which phosphorylates the downstream nuclear effector YAP/TAZ. Phosphorylated YAP/TAZ is unable to induce the expression of cell proliferation and anti-apoptotic genes through interaction with transcriptional co-activators such as TEAD. The Hippo pathway, known to cross-talk with BMP signaling at multiple levels, recently emerged as a cellular sensor of the physical environment through the actin cytoskeleton. We thus asked whether modulating Hippo signaling affected tumor growth. YAP1 overexpression resulted in tumor growth decrease. A deletion variant of YAP, YAP1-291, possessing a deletion of the transactivation domain has been shown to act as a dominant-negative through binding to TEAD. Intriguingly YAP1-291 was also sufficient to strongly attenuate growth of diverse BMP4-sensitive and BMP4-insensitive tumor models suggesting that transactivation through TEAD is not the only route by which YAP modulates tumor growth. YAP anti-proliferative properties were reported before. In order to reconcile the dual function of YAP in cancer, we propose that the transcriptional co-activator's TEAD-mediated oncogenic function may be counter-balanced by its potentiation of SMAD signaling in response to BMP.

Citation Format: Cecile Chartier, Shirley Ma, Dragana Antic, Elsa Quintana, Janak Raval, Min Wang, Cristina Dee-Hoskins, Ann Kapoun, Tim Hoey, John Lewicki, Austin Gurney. The Hippo signaling pathway mediates BMP inhibition of cancer stem cells. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2322. doi:10.1158/1538-7445.AM2015-2322

Abstract 255: Dual targeting of Delta-like ligand 4 (DLL4) and programmed death 1(PD1) inhibits tumor growth and generates enhanced long-term immunological memory

Blocking DLL4, a Notch ligand, effectively inhibits tumor growth by increasing non-functional angiogenesis and decreasing the cancer stem cells (CSC) population. We are currently testing an anti-DLL4 antibody, demcizumab, in Phase1B trials in NSCLC, pancreatic, and ovarian cancer. DLL4 is also known to modulate immune responses. In the current study we examine the impact of anti-DLL4 on anti-tumor immune responses as a single agent and in combination with the key immune checkpoint inhibitor Programmed Cell Death Protein 1 (PD1). While the recent clinical success of PD1 inhibitors represents a new and promising cancer immunotherapeutic approach, high initial response rates are often associated by a lack of long-term, durable effects in a significant number of patients. Therefore, we hypothesized that dual blockade of DLL4 and PD1 might further impact tumor growth by further enhancing anti-tumor immune immunity. Our data demonstrates that dual blockade of DLL4 and PD1 using antibodies not only reduces tumor growth, but also led to tumor rejection in ∼50% in CT26WT tumor-bearing mice, similar to those treated with anti-PD1 alone (no tumor rejection was observed with anti-DLL4 alone). Anti-PD1 increased specific CD8+ T cell-mediated IFN-γ production while decreasing IL6. Anti-DLL4 treatment reduced IL17 production. Interestingly, only the dual blockage led to increased production of IL2 by splenocytes. Since IL2 is required for secondary population expansion of CD8+ memory T cells, increased IL2 in the combination group suggests potential for increased T cell activation, maintenance and memory T cell function, as compared to single agent anti-DLL4 and anti-PD1. While anti-PD1 reduced inhibition of CD4+ T cell proliferation by Tregs, the dual blockade significantly reduced Treg-mediated CD8+ T cell suppression. Furthermore, both effector and memory CD8+ T cell frequencies were increased within the total CD8+ T cell population. Interestingly, anti-PD1 decreased granulocytic MDSCs, while anti-DLL4 reduced monocytic MDSCs. Mice cured with single-agent anti-PD1 and anti-DLL4/anti-PD1 combination treatments were protected from series of re-challenge with tumor cells, suggesting the existence of immunologic memory. Interestingly, more mice were protected from tumor re-challenge when both DLL4 and PD1 were blocked, as compared to PD1 alone. Surprisingly, mice previously treated with the anti-DLL4/anti-PD1 combination produced more IL2, clearly indicating the role of DLL4 blockade in enhancing anti-tumor immunity. Therefore, these results show that dual targeting of DLL4 and PD1 may be an effective and durable cancer therapy by increasing anti-tumor immune response and promoting long-term immunological memory.

Citation Format: Minu Srivastava, Christopher L. Murriel, Julie Roda, Hyun-Bae Jie, Fumiko Axelrod, Ming-Hong Xie, Rui Yun, Erin Mayes, Trevor Bentley, Belinda Cancilla, Raymond Tam, Tracy Tang, Ann Kapoun, John Lewicki, Tim Hoey, Austin Gurney, Angie Inkyung Park. Dual targeting of Delta-like ligand 4 (DLL4) and programmed death 1(PD1) inhibits tumor growth and generates enhanced long-term immunological memory. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 255. doi:10.1158/1538-7445.AM2015-255

Abstract 1549: Development and validation of a biomarker for prospective selection of Notch1 activation in patients with certain advanced solid tumors in a first-in-human phase1 study of the cancer stem cell targeting antibody OMP-52M51 (anti-Notch1)

The Notch pathway plays a key role in embryonic development, the regulation of stem and progenitor cells, and is implicated centrally in many forms of human cancer. Notch1 is known to be frequently activated in certain solid tumor types. OMP-52M51 is a humanized IgG2 antibody that inhibits the signaling function of the Notch1 receptor. Mouse xenograft studies using minimally-passaged, patient-derived xenografts have shown that OMP-52M51 impedes tumor growth and selectively eliminates CSCs in a range of tumor types particularly in tumors with activated Notch1 signaling.

We previously reported the frequency of Notch pathway activation across a large panel of human tumors (n>600) by an Immunohistochemistry (IHC) assay that detects the activated form of Notch1 using an antibody that specifically recognizes the Notch1 intracellular domain (ICD). Using this test and a rigorous H-score cut-off, we found elevated Notch1 ICD in 7-53% of the following cancers: chemo-resistant breast (29%), gastric (13%), cholangiocarcinoma (20%), esophageal (27%), hepatacellular carcinoma (HCC, 7%), small cell lung cancer (SCLC, 12%), pancreatic (12%) and colorectal cancer (53%).

Here we developed a specific CLIA-validated IHC assay capable of identifying patients with Notch1 activation. The assay has a rapid turn-around time and shows robust precision, reproducibility and accuracy. The sensitivity of the method was confirmed by comparing results obtained on xenograft tumor tissue samples with known levels of Notch1 expression and response to anti-Notch1 therapy. From these results, the cut-off of the assay was set. This predictive biomarker was specifically designed to screen patients for prospective selection in the first-in-human Phase1 study of OMP-52M51 (anti-Notch1) in certain advanced solid tumors (NCT01778439). Detailed assay validation data and its application to the analysis of clinical trial samples will be highlighted.

Citation Format: Belinda Cancilla, Raymond Tam, Chun Zhang, Steve Anderson, John Lewicki, Tim Hoey, Bryan McCune, Lori Johnson, Esohe Idusogie, Ann M. Kapoun. Development and validation of a biomarker for prospective selection of Notch1 activation in patients with certain advanced solid tumors in a first-in-human phase1 study of the cancer stem cell targeting antibody OMP-52M51 (anti-Notch1). [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 1549. doi:10.1158/1538-7445.AM2015-1549

Abstract A33: Combined therapy of a novel anti-NOTCH2/3 antibody with paclitaxel inhibits tumor growth in a patient-derived breast tumor xenograft

Cancer stem cells (CSCs) have been identified as the small subpopulation of cells in many cancers and are thought to be responsible for cancer initiation, progression, metastasis, recurrence and drug resistance. The NOTCH pathway, comprised of ligands JAG1, JAG2, DLL1, DLL3 and DLL4 and receptors NOTCH1, NOTCH2, NOTCH3 and NOTCH4, plays a critical role in stem cell signaling, cell proliferation, survival, apoptosis, and differentiation. Because of the existing body of data that implicates aberrant NOTCH pathway activation in resistance to chemotherapeutic agents in epithelial and hematologic tumors, we hypothesized that NOTCH inhibition would enhance the anti-tumor activity of chemotherapeutic agents.

We have identified a NOTCH3 activating mutation in a patient breast cancer and developed a xenograft tumor model from the primary patient sample, OMP-B37. The combination of paclitaxel and OncoMed's anti-Notch2/3 antibody, OMP-59R5, inhibited the tumor growth of OMP-B37 xenograft tumors. In addition, we found that treatment with OMP-59R5 drove changes in tumor heterogeneity as indicated by the increased expression of markers associated with a mature cellular fate. The down-regulation of tumor growth promoting genes and NOTCH pathway markers was observed with OMP-59R5 treatment. The enhanced activity of combined OMP-59R5 and paclitaxel therapy in patient-derived tumor xenografts suggests that there is therapeutic potential in combining targeted NOTCH pathway inhibition with chemotherapeutic agents.

Citation Format: Jie Wei, Jennifer Cain, Jalpa Shah, Breanna Wallace, Tracy Tang, Min Wang, Christopher Murriel, Cancilla Belinda, Austin Gurney, Aaron Sato, John Lewicki, Ann Kapoun, Tim Hoey. Combined therapy of a novel anti-NOTCH2/3 antibody with paclitaxel inhibits tumor growth in a patient-derived breast tumor xenograft. [abstract]. In: Proceedings of the AACR Precision Medicine Series: Drug Sensitivity and Resistance: Improving Cancer Therapy; Jun 18-21, 2014; Orlando, FL. Philadelphia (PA): AACR; Clin Cancer Res 2015;21(4 Suppl): Abstract nr A33.

Abstract 213: Novel NOTCH3 activating mutations identified in tumors sensitive to OMP-59R5, a monoclonal antibody targeting the Notch2 and Notch3 receptors

The Notch signaling pathway, driven by four Notch family receptors (NOTCH1-4) and five canonical Notch ligands (DLL1, DLL3, DLL4, JAG1, and JAG2) in humans, critically regulates key functions during embryonic development as well as stem cell maintenance and differentiation in adult tissues. Altered Notch signaling activity has been documented in many types of cancers; depending on the tumor type, Notch signaling can be oncogenic or tumor suppressing. Activating mutations in NOTCH1 have been identified in hematopoietic cancers including T-cell acute lymphoblastic leukemia and chronic lymphoblastic leukemia. With the advance in high-throughput genomic sequencing platforms, mutations in the NOTCH genes have also been identified in solid tumors, albeit at much lower frequencies. Here we report for the first time the identification of two activating mutations in NOTCH3. One is a frameshift mutation in the PEST domain, and the other a frameshift mutation in the ankyrin (ANK) domain. The PEST frameshift mutation leads to the production of C-terminally truncated NOTCH3 protein, and by Western Blotting of nuclear fractions of tumor tissues and Immunohistochemistry (IHC) we show that this mutant NOTCH3 accumulates stably in the nuclei of the breast tumor in which it was identified. Breast xenograft tumors carrying this mutation are highly sensitive to OMP-59R5, a ligand-blocking antibody targeting the Notch2 and Notch3 receptors, suggesting that the mutation, although activating, is still ligand dependent. Consistent with this finding, expression of this mutant NOTCH3 constructed by site-directed mutagenesis exhibits higher baseline activity compared to wild type NOTCH3 and is equally responsive to ligands in in vitro NOTCH reporter assays. We are not able to observe increased nuclear accumulation of the ANK frameshift mutant NOTCH3 in the colon tumor in which it was identified. However, in vitro the ANK mutant NOTCH3 has higher baseline activity than wild-type NOTCH3 in the absence of ligands and remains highly responsive to ligands. The colon xenograft tumors carrying this NOTCH3 ANK mutation are also sensitive to OMP-59R5 in in vivo efficacy studies. Taken together our data suggest that NOTCH3 is oncogenic in certain subtypes of tumors, and tumors with these activating mutations may predictive sensitivity to therapeutic antibodies targeting the receptor.

Citation Format: Breanna Wallace, Min Wang, Chris Muriel, Jennifer Cain, Belinda Cancilla, Jalpa Shah, Jie Wei, Austin Gurney, John Lewicki, Aaron Sato, Tim Hoey, Tracy Tang, Ann M. Kapoun. Novel NOTCH3 activating mutations identified in tumors sensitive to OMP-59R5, a monoclonal antibody targeting the Notch2 and Notch3 receptors. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 213. doi:10.1158/1538-7445.AM2013-213

Abstract 3728: Anti-Notch1 antibody (OMP-52M51) impedes tumor growth and cancer stem cell frequency (CSC) in a chemo-refractory breast cancer xenograft model with an activating Notch1 mutation and screening for activated Notch1 across multiple solid tumor types

The Notch pathway plays a central role in embryonic development, the regulation of stem and progenitor cells, and is implicated in many human cancers. Notch1 is also known to be activated by oncogenic mutations in tumors including T cell leukemia, chronic lymphocytic leukemia and non-small cell lung cancer. Here we report the discovery of similar mutations in breast cancer. We have developed a primary human xenograft model from a patient bearing an activating Notch1 mutation and have shown that blocking Notch signaling in this model with a novel anti-Notch1 antibody, OMP-52M51, impeded tumor growth and reduced cancer stem cell (CSC) frequency. OMP-52M51 is a humanized antibody that binds with high affinity and selectivity to human Notch1 and antagonizes Notch signaling.

Additionally, we sought to determine the frequency of Notch pathway activation across a large panel of human tumors (n>600) by an immunohistochemistry assay that detects the activated form of Notch1 using a polyclonal antibody that specifically recognizes the Notch1 intracellular domain (ICD). Using this test and a rigorous H-score cut-off of 30, we found elevated Notch1 in 7-29% of the following cancers: chemo-resistant breast, gastric, cholangiocarcinoma, esophageal, hepatacellular carcinoma (HCC), and small cell lung cancer (SCLC).

The class of patients that showed the highest frequency of elevated Notch1.ICD was chemo-resistant breast cancer (∼30%). This frequency was significantly higher than in unselected breast cancer patients, suggesting that Notch1 signaling plays a significant role in breast cancer chemoresistance. Interestingly, the human xenograft model which showed sensitivity to OMP-52M51 was derived from a patient that failed to respond to pre-operative chemotherapy and developed metastatic disease following surgery. Thus, chemo-resistant breast cancer as well as the other tumors with high Notch1.ICD may represent important classes of patients who could benefit from anti-Notch1 therapy in the clinic.

Citation Format: Belinda Cancilla, Jennifer Cain, Min Wang, Lucia Beviglia, Jalpa Shah, Austin Gurney, John Lewicki, Laura Esserman, Tim Hoey, Ann M. Kapoun. Anti-Notch1 antibody (OMP-52M51) impedes tumor growth and cancer stem cell frequency (CSC) in a chemo-refractory breast cancer xenograft model with an activating Notch1 mutation and screening for activated Notch1 across multiple solid tumor types. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3728. doi:10.1158/1538-7445.AM2013-3728

Abstract 2698: In vivo imaging of human pancreatic tumor and melanoma models in NOD/SCID mice: Preclinical evaluation of metastases and inhibition by anti-DLL4 treatment

Pre-existing metastases at the time of diagnosis are often the cause of death in patients suffering from a number of malignancies, including pancreatic cancer and melanoma. Successful development of effective anti-metastatic cancer therapies requires preclinical metastasis models that mimic the clinical condition of human cancer, sensitive methods of metastasis detection in mice bearing orthotopic tumors, and real-time evaluation of anti-metastatic activity of novel treatments. We have developed orthotopic tumor models in NOD/SCID mice utilizing pancreatic and melanoma tumor cells that are derived directly from human cancer patient specimens. These tumors retain much of the original human cancer heterogeneity, phenotype, and the ability to form metastases in the lungs, liver, intestines, and brain _ recapitulating the metastatic invasiveness seen in the clinic. Since the ability of optical imaging methods to detect tumors and metastases in deep tissues is limited by tumor size and light penetration depth, we employed Zymera's Bioluminescence Resonance Energy Transfer-Quantum Dot (BRET-Qdot®) nanoprobes conjugated with streptavidin ([BRET-Qdot]-SA) to improve sensitivity for detection of both early stage primary tumors and small metastases. The cell surface proteins, MCAM in melanoma and EpCAM in pancreatic tumors, were targeted with human specific biotinylated anti-MCAM or -EpCAM antibodies, (respectively) delivered intravenously or intraperitoneally into NOD/SCID mice and detected with the [BRET-Qdot®]-SA nanoprobe. Light emission from the tumor-bound [BRET-Qdot®]-SA was activated by the catalytic conversion of the luciferase substrate coelenterazine. The novel and highly sensitive BRET-Qdot® technology emits 655 nm bioluminescence light for in vivo imaging of deep tissues. This technology allowed the growth of primary tumors and metastases to be measured overtime in the same mice. Moreover, in efficacy studies we have treated the mice with antibodies developed against the human and the murine DLL4 to target both tumor cells and vasculature that cooperate for the formation and growth of primary tumor and metastases. Treatment with anti-DLL4 inhibits metastases in the lungs, liver, intestine and brain. Finally, we have developed a more clinically relevant melanoma metastasis model consisting of the surgical excision of the primary tumor. This model, characterized by an enrichment of aggressive growth of metastases in the lungs, has been also utilized to evaluate the inhibitory activity of anti-DLL4 antibodies.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 2698. doi:10.1158/1538-7445.AM2011-2698

Abstract P6-15-10: Targeting Delta-Like 4 Ligand (DLL4)/Notch Signaling by a Novel Anti-DLL4 Antibody Inhibits Tumor Growth through Altering Cancer Stem Cell and Epithelial-to-Mesenchymal Transition-Associated Genes in Triple Negative Breast Cancer

Triple negative breast cancer (TNBC, ER-, PR-Her2-) accounts for 10% of all types of breast cancer and is known to be particularly aggressive and refractory to current therapies. Chemotherapeutic agents, such as taxanes, are currently the only treatment option for this type of breast cancer. However, these treatments often result in local and systemic relapse. While several promising targeted agents have improved the relapse rate in the combination with chemotherapy in patients with pathologic complete response, those who do not achieve pathologic complete response still have a poor prognosis. We previously demonstrated that interfering with Deltalike 4 ligand (DLL4)/Notch signaling by an anti-DLL4 antibody decreased tumor growth and delayed tumor recurrence following paclitaxel treatment through reducing breast cancer stem cell fraction and tumorigenicity (Proc Am Asso Cancer Res. 51:4, 2010). CD44+/CD24low cells have been demonstrated to be highly tumorigenic in breast tumors and are associated with treatment resistance (PNAS U S A, 106: 13820-13825, 2009). To gain insights into the mechanism of action of anti-DLL4-mediated TNBC growth inhibition, we evaluated the effect of anti-DLL4 on breast cancer CD44+/CD24low the expression patterns of a list of genes involved in sternness, differentiation, and key stem cell signaling pathways in TNBC xenograft tumors following paclitaxel with and without antibody treatment in vivo. Our data showed that paclitaxel treatment resulted in significant changes of gene expression patterns in the CD44+/CD24low cells in the residual tumors. We further analyzed these genes and identified several genes related to Notch signaling, stem cell maintenance and epithelial-to-mesenchymal transition that were modulated differentially by treatment with paclitaxel and/or anti-DLL4. In a metastatic TNBC orthotopic tumor model, the anti-DLL4 treatment resulted in tumor growth inhibition at the primary site and distant organs. Taken together, our findings suggest that targeting DLL4 may improve the efficacy of current treatments, delay tumor recurrence, and reduce metastatic growth in TNBC.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P6-15-10.

Abstract A29: Discovery and preclinical to clinical evaluation of biomarkers for Anti-DLL4, a monoclonal antibody targeting cancer stem cells

Blockade of DLL4 signaling reduces tumor growth in preclinical human xenograft models by disrupting productive angiogenesis and reducing proliferation of tumor cells. Importantly in these models, blocking DLL4 also reduces cancer stem cell (CSC) frequency in multiple solid tumor types. Currently, there exists an important need to identify and evaluate pharmacodynamic biomarkers that indicate the biological activity of agents, such as anti-DLL4, that are in clinical studies. Towards this aim, we used numerous human xenograft models derived from primary human tumors to identify and characterize biomarkers for anti-DLL4. Here, we present data from these studies and describe anti-DLL4 biomarkers in tumors, in hair follicles, and in blood. We show that several vascular-related genes are significantly upregulated by anti-DLL4 in tumors including, Egfl7, Apln, Cldn5, and Notch4. We also demonstrate significant downregulation of CSC gene signatures and modulation of differentiation genesets. A subset of the identified markers were validated at the protein level and also tested in a human skin graft model, where the human tumor is engrafted in a human microenvironment/vasculature. Finally, a DLL4-Notch gene signature was identified that showed a significant association with breast cancer survival using publicly available cancer gene sets. In surrogate tissues, we identified anti-DLL4 biomarkers in blood and in hair follicles. These markers from surrogate tissues were translated into a phase I trial of Anti-DLL4 antibody (OMP-21M18), and an interim analysis of this data demonstrates that OMP-21M18 affects the Notch pathway in patient samples.

Citation Information: Clin Cancer Res 2010;16(14 Suppl):A29.

Targeting Cancer Stem Cells and Vasculature by a Novel Anti-Delta-Like 4 Ligand (DLL4) Antibody for Treatment of Triple Negative Breast Cancer

Triple negative (ER-, PR- Her2-) breast cancer accounts for 10% of all types of breast cancer and is known to particularly aggressive and refractory to current therapies. Chemotherapeutic agents, such as taxanes, are currently the only treatment option for this type of breast cancer, however, these treatments often result in high rates of local and systemic relapse. It has been suggested that cancer stem cells drive tumor growth and progression and are preferentially resistant to many current therapies. Delta-like 4 ligand (DLL4) is an important component of the Notch signaling pathway which is known to mediate stem cell self-renewal and vascular development. We hypothesized that targeting cancer stem cells and the tumor vasculature by interfering with the DLL4/Notch pathway will improve treatment outcome. We developed anti-DLL4 antibodies that recognize either the human or murine protein and have potent binding and antagonist activities. We have used these antibodies to investigate the role of DLL4 in triple negative breast cancer and to probe the mechanism in tumor and stromal cells in xenograft models derived from primary breast tumors. These studies showed that anti-DLL4 was efficacious as a single agent and in combination with paclitaxel (Taxol) against triple negative breast tumors including a tumor that developed paclitaxel resistance. Gene expression analysis showed that anti-DLL4 affected vascular-related genes in the stroma and Notch target genes in the tumor and stroma. Furthermore, inclusion of anti-DLL4 delayed breast tumor recurrence following termination of paclitaxel treatment. A tumorigenicity study indicated that treatment with anti-DLL4 decreased cancer stem cell frequency as measured by in vivo limiting dilution assay, whereas paclitaxel alone was ineffective. The combination of anti-DLL4 and paclitaxel further decreased cancer stem cell frequency compared to single agents. Our findings provide a rationale for targeting cancer stem cells and tumor vasculature through inhibition of the DLL4/Notch signaling for breast cancer treatment.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 5071.

Colorectal cancer stem cells are enriched in xenogeneic tumors following chemotherapy

Background

Patients generally die of cancer after the failure of current therapies to eliminate residual disease. A subpopulation of tumor cells, termed cancer stem cells (CSC), appears uniquely able to fuel the growth of phenotypically and histologically diverse tumors. It has been proposed, therefore, that failure to effectively treat cancer may in part be due to preferential resistance of these CSC to chemotherapeutic agents. The subpopulation of human colorectal tumor cells with an ESA⁺CD44⁺ phenotype are uniquely responsible for tumorigenesis and have the capacity to generate heterogeneous tumors in a xenograft setting (i.e. CoCSC). We hypothesized that if non-tumorigenic cells are more susceptible to chemotherapeutic agents, then residual tumors might be expected to contain a higher frequency of CoCSC.

Methods and Findings

Xenogeneic tumors initiated with CoCSC were allowed to reach ∼400 mm³, at which point mice were randomized and chemotherapeutic regimens involving cyclophosphamide or Irinotecan were initiated. Data from individual tumor phenotypic analysis and serial transplants performed in limiting dilution show that residual tumors are enriched for cells with the CoCSC phenotype and have increased tumorigenic cell frequency. Moreover, the inherent ability of residual CoCSC to generate tumors appears preserved. Aldehyde dehydrogenase 1 gene expression and enzymatic activity are elevated in CoCSC and using an in vitro culture system that maintains CoCSC as demonstrated by serial transplants and lentiviral marking of single cell-derived clones, we further show that ALDH1 enzymatic activity is a major mediator of resistance to cyclophosphamide: a classical chemotherapeutic agent.

Conclusions

CoCSC are enriched in colon tumors following chemotherapy and remain capable of rapidly regenerating tumors from which they originated. By focusing on the biology of CoCSC, major resistance mechanisms to specific chemotherapeutic agents can be attributed to specific genes, thereby suggesting avenues for improving cancer therapy.

Expression and purification of lacZ and trpE fusion proteins

Fusion proteins are commonly used as a source of antigen for producing antibodies and in many cases can be useful for biochemical analyses. This unit describes two widely used expression systems for producing large amounts of proteins in E. coli. One system expresses lacZ fusions using the pUR series of vectors and the other expresses trpE fusions using the pATH vectors. The gene of interest is first subcloned into either a pUR or pATH vector in the correct reading frame. The correct transformant is selected, grown, and then induced with either IPTG or IAA. Sonication of cells in the presence of protease inhibitors is used to prepare extracts containing both types of fusion proteins, as well as other types of proteins overexpressed in E. coli. The extracts are checked for the presence of fusion protein on an SDS-polyacrylamide gel.

Multiclass cancer classification and biomarker discovery using GA-based algorithms

MOTIVATION

The development of microarray-based high-throughput gene profiling has led to the hope that this technology could provide an efficient and accurate means of diagnosing and classifying tumors, as well as predicting prognoses and effective treatments. However, the large amount of data generated by microarrays requires effective reduction of discriminant gene features into reliable sets of tumor biomarkers for such multiclass tumor discrimination. The availability of reliable sets of biomarkers, especially serum biomarkers, should have a major impact on our understanding and treatment of cancer.

RESULTS

We have combined genetic algorithm (GA) and all paired (AP) support vector machine (SVM) methods for multiclass cancer categorization. Predictive features can be automatically determined through iterative GA/SVM, leading to very compact sets of non-redundant cancer-relevant genes with the best classification performance reported to date. Interestingly, these different classifier sets harbor only modest overlapping gene features but have similar levels of accuracy in leave-one-out cross-validations (LOOCV). Further characterization of these optimal tumor discriminant features, including the use of nearest shrunken centroids (NSC), analysis of annotations and literature text mining, reveals previously unappreciated tumor subclasses and a series of genes that could be used as cancer biomarkers. With this approach, we believe that microarray-based multiclass molecular analysis can be an effective tool for cancer biomarker discovery and subsequent molecular cancer diagnosis.

G protein-coupled receptors GPR4 and TDAG8 are oncogenic and overexpressed in human cancers

The GPR4 subfamily consists of four G protein-coupled receptors that share significant sequence homology. In addition to GPR4, this subfamily includes OGR1, TDAG8 and G2A. G2A has previously been shown to be a potent transforming oncogene for murine 3T3 cells. Here we show that GPR4 also malignantly transforms NIH3T3 cells and that TDAG8 malignantly transforms the normal mammary epithelial cell line NMuMG. Overexpression of GPR4 or TDAG8 in HEK293 cells led to transcriptional activation from SRE- and CRE-driven promoters, independent of exogenously added ligand. TDAG8 and GPR4 are also overexpressed in a range of human cancer tissues. Our results suggest that GPR4 and TDAG8 overexpression in human tumors plays a role in driving or maintaining tumor formation.

Stat4 regulates multiple components of IFN-gamma-inducing signaling pathways

Stat4 is activated in response to IL-12. Most functions of IL-12, including the induction of IFN-gamma, are compromised in the absence of Stat4. Since the precise role of Stat4 in IFN-gamma induction has not been established, experiments were conducted to examine Stat4 activation of IFN-gamma and other genes required for cytokine-induced expression of IFN-gamma. We first examined IL-12 signaling components. Basal expression of IL-12Rss1 and IL-12Rss2 is decreased in Stat4-deficient cells compared with that in control cells. However, IL-12 was still capable of inducing equivalent phosphorylation of Jak2 and Tyk2 in wild-type and Stat4-deficient activated T cells. We have further determined that other cytokine signaling pathways that induce IFN-gamma production are defective in the absence of Stat4. IL-18 induces minimal IFN-gamma production from Stat4-deficient activated T cells compared with control cells. This is due to defective IL-18 signaling, which results from the lack of IL-12-induced, and Stat4-dependent, expression of the IL-18R. Following IL-12 pretreatment to induce IL-18R, wild-type, but not Stat4-deficient, activated T cells demonstrated IL-18-induced NF-kappaB DNA-binding activity. In addition, IL-12-pretreated Stat4-deficient activated T cells have minimal IFN-gamma production followed by stimulation with IL-18 alone or in combination with IL-12 compared with control cells. Thus, Stat4 activation by IL-12 is required for the function of multiple cytokine pathways that result in induction of IFN-gamma.

Identification of a STAT4 binding site in the interleukin-12 receptor required for signaling

The specificity of the various STAT SH2 domains for different tyrosine-containing peptides enables cytokines to activate different signaling pathways and to induce distinct patterns of gene expression. We show that STAT4 has a unique peptide specificity and binds to the peptide sequence pYLPSNID (where pY represents phosphotyrosine). This motif is found at tyrosine residue 800 in the beta2 subunit of the interleukin-12 receptor and is required for DNA binding and transcriptional activity of STAT4. Our data demonstrate that transfection of interleukin-12 receptor beta1 and beta2 subunits is sufficient for STAT4 activation but not for STAT1 or STAT3 activation.

The roles of nuclear factor of activated T cells and ying-yang 1 in activation-induced expression of the interferon-gamma promoter in T cells

Nuclear factor of activated T cells (NFAT) plays an important role in expression of many cytokine genes including interleukin-2 and interleukin-4. However, its role in interferon-gamma (IFN-gamma) expression is not well understood. In the current studies, two strong NFAT-binding sites in the IFN-gamma promoter were identified by DNase I footprint analysis at positions -280 to -270 and -163 to -155. NFATp bound independently to both sites and was required for the formation of a composite element with AP-1 spanning position -163 to -147. In Jurkat T cells and primary lymphocytes, activation-induced expression of IFN-gamma reporter constructs containing point mutations in either NFAT site or the AP-1 component of the composite site was decreased by approximately 40-65%. Despite elimination of both strong NFAT-binding sites, the IFN-gamma promoter remained completely sensitive to inhibition by cyclosporin. This suggests that other elements in the IFN-gamma promoter, such as the IFN-gamma proximal element, are sufficient for cyclosporin sensitivity of this gene. Ying-Yang 1 (YY1), a potential inhibitor of IFN-gamma expression, binds to sites located between the two NFAT sites. Mutation of the YY1 sites alone had little effect on IFN-gamma promoter activity. However, mutation of both the NFAT and YY1-binding sites abolished activation-induced expression in primary murine splenocytes but not in Jurkat T cells. This suggests that under some conditions, YY1 may play a positive role in activation-induced transcription of IFN-gamma.

STAT structure and function in signaling

The phenotypes of various STAT knockout mice reveal an unexpected specificity in the biological roles of these molecules. The mechanisms involved in generating selectivity and modulating STAT activity have been the focus of intense studies. This work has led to the discovery of novel families of proteins that regulate Jak-STAT signaling. Recently, the structures of a STAT dimer/DNA complex and of the amino-terminal domain have been solved, providing new insights into the function of these versatile proteins.

The transcription factor NFAT4 is involved in the generation and survival of T cells

Nuclear factor of activated T cells (NFAT) is a family of four related transcription factors implicated in cytokine and early response gene expression in activated lymphocytes. Here we report that NFAT4, in contrast to NFATp and NFATc, is preferentially expressed in DP thymocytes. Mice lacking NFAT4 have impaired development of CD4 and CD8 SP thymocytes and peripheral T cells as well as hyperactivation of peripheral T cells. The thymic defect is characterized by increased apoptosis of DP thymocytes. The increased apoptosis and hyperactivation may reflect heightened sensitivity to TcR-mediated signaling. Further, mice lacking NFAT4 have impaired production of Bcl-2 mRNA and protein. NFAT4 thus plays an important role in the successful generation and survival of T cells.

The transcription factor NF-ATc is essential for cardiac valve formation

Nuclear factor of activated T cells (NF-AT) is the name of a family of four related transcription factors that may be needed for cytokine gene expression in activated lymphocytes. Here we report that mice with a targeted disruption of the NF-ATc gene show an unexpected and dramatic defect in cardiac morphogenesis, with selective absence of the aortic and pulmonary valves, leading to death in utero from congestive heart failure at days 13.5-17.5 of gestation. In contrast, tricuspid and mitral valve morphogenesis is normal. NF-ATc is the first transcription factor known to be expressed only in the endothelial cells of the heart. As in T cells, nuclear translocation of NF-ATc in cardiac endothelial cells is controlled by the calcium-regulated phosphatase calcineurin: NF-ATc remains cytoplasmic in normal embryos cultured with cyclosporin A, an inhibitor of calcineurin. Abnormal development of the cardiac valves and septae is the most frequent form of birth defect, yet few molecular regulators of valve formation are known. Our results indicate that NF-ATc may play a critical role in signal-transduction processes required for normal cardiac valve formation.

Delayed lymphoid repopulation with defects in IL-4-driven responses produced by inactivation of NF-ATc

The NF-AT family of transcription factors activates early immune response genes such as cytokines. In the adult, NF-ATc is expressed exclusively in the lymphoid system and is induced upon lymphocyte activation. NF-ATc null mutant mice die in utero of cardiac failure, precluding analysis of the role of NF-ATc in lymphocyte activation. By using RAG-2-deficient blastocyst complementation, we now demonstrate that young, highly chimeric mice lacking NF-ATc have impaired repopulation of both thymus and peripheral lymphoid organs. Furthermore, NF-ATc deficiency impaired T lymphocyte activation and secretion of IL-4. B lymphocytes displayed reduced proliferation and a selective loss of IL-4-driven immunoglobulin isotypes both in vivo and in vitro. Our data demonstrate that NF-ATc is essential for the optimal generation and function of mature T and B lineage cells, with an especially profound effect on IL-4-driven responses.

Cooperative DNA binding and sequence-selective recognition conferred by the STAT amino-terminal domain

STAT proteins (signal transducers and activators of transcription) activate distinct target genes despite having similar DNA binding preferences. The transcriptional specificity of STAT proteins was investigated on natural STAT binding sites near the interferon-gamma gene. These sites are arranged in multiple copies and required cooperative interactions for STAT binding. The conserved amino-terminal domain of STAT proteins was required for cooperative DNA binding, although this domain was not essential for dimerization or binding to a single site. Cooperative binding interactions enabled the STAT proteins to recognize variations of the consensus site. These sites can be specific for the different STAT proteins and may function to direct selective transcriptional activation.

Impaired IL-12 responses and enhanced development of Th2 cells in Stat4-deficient mice

Interactions between cytokine and receptor lead to the activation of multiple signalling molecules, including the family of signal transducer and activator of transcription (STAT) proteins. STAT4 is one member of this family, and is activated only in response to the cytokine interleukin (IL)-12 (refs 5, 6). By gene targeting, we have generated mice deficient in STAT4 to determine whether the function of this transcription factor is redundant with other signalling molecules activated by IL-12. IL-12-induced increases in the production of interferon (IFN)-gamma cellular proliferation and natural killer (NK) cell cytotoxicity are abrogated in lymphocytes from STAT4-deficient mice. The development of Th1 cells in response to either IL-12 of Listeria monocytogenes is also impaired in the absence of Stat4. Furthermore, Stat4-deficient lymphocytes demonstrate a propensity towards the development of Th2 cells. These results demonstrate that Stat4 is essential for mediating responses to IL-12 in lymphocytes, and regulating the differentiation of both Th1 and Th2 cells.

Differentiation of T-helper lymphocytes: selective regulation by members of the STAT family of transcription factors

Interleukin-4 (IL-4) and interleukin-12 (IL-12) control the differentiation of T-helper cells. Here we summarize studies which investigate the mechanism by which these cytokines selectively reprogramme gene expression in T-lymphocytes. Cytokine stimulation leads to the phosphorylation of specific tyrosine residues within the intracellular domain of the corresponding cytokine receptor. These phosphotyrosines serve as docking sites for latent, cytoplasmic transcription factors known as signal transducers and activators of transcription (Stat) proteins. Receptor/Stat interaction is mediated by the src homology 2 (SH2) domain of the corresponding Stat protein. Although Stat binding to the intracellular domain of the cytokine receptor strongly depends on the phosphotyrosine residue, the recruitment of a specific Stat protein is dictated by amino acid residues C-terminal to the phosphotyrosine. Specific docking sites within individual cytokine receptors have been identified for almost all Stat proteins. The direct coupling between cytokine receptor and transcription factor helps to explain how different cytokines elicit distinct patterns of gene expression.

Hyperproliferation and dysregulation of IL-4 expression in NF-ATp-deficient mice

NF-ATp is a member of a family of genes that encodes the cytoplasmic component of the nuclear factor of activated T cells (NF-AT). In this study, we show that mice with a null mutation in the NF-ATp gene have splenomegaly with hyperproliferation of both B and T cells. They also display early defects in the transcription of multiple genes encoding cytokines and cell surface receptors, including CD40L and FasL. A striking defect in early IL-4 production was observed after ligation of the TCR complex by treatment with anti-CD3 in vivo. The transcription of other cytokines including IL-13, GM-CSF, and TNF alpha was also affected, though to a lesser degree. Interestingly, the cytokines IL-2 and IFN gamma were minimally affected. Despite this early defect in IL-4 transcription, Th2 development was actually enhanced at later timepoints as evidenced by increased IL-4 production and IgE levels in situations that favor the formation of Th2 cells both in vitro and in vivo. These data suggest that NF-ATp may be involved in cell growth, and that it is important for the balanced transcription of the IL-4 gene during the course of an immune response.

Novel NFAT sites that mediate activation of the interleukin-2 promoter in response to T-cell receptor stimulation

The transcription factors NFAT and AP-1 have been shown to be essential for inducible interleukin-2 (IL-2) expression in activated T cells. NFAT has been previously reported to bind to two sites in the IL-2 promoter: in association with AP-1 at the distal antigen response element at -280 and at -135. On the basis of DNase I footprinting with recombinant NFAT and AP-1 proteins, gel shift assays, and transfection experiments, we have identified three additional NFAT sites in the IL-2 promoter. Strikingly, all five NFAT sites are essential for the full induction of promoter activity in response to T-cell receptor stimulation. Four of the five NFAT sites are part of composite elements able to bind AP-1 in association with NFAT. These sites display a diverse range of cooperativity and interdependency on NFAT and AP-1 proteins for binding. One of the NFAT sites directly overlaps the CD28-responsive element. We present evidence that CD28 inducibility is conferred by the AP-1 component in NFAT-AP-1 composite elements. These findings provide further insight into the mechanisms involved in the regulation of the IL-2 promoter.

Isolation of two new members of the NF-AT gene family and functional characterization of the NF-AT proteins

The activation of cytokine genes in response to antigenic stimulation of T cells is mediated by NF-AT proteins. Previous studies have identified two NF-AT proteins, NF-ATp and NF-ATc, that are homologous within a 290 aa domain distantly related to the Rel domain. We have isolated two additional members of this gene family, NF-AT3 and NF-AT4, which encode proteins 65% identical to the other NF-AT proteins within the Rel domain. The four NF-AT genes are transcribed in different sets of tissues that included many sites of expression outside the immune system. The Rel homology domain is sufficient for DNA recognition and cooperative binding interactions with AP-1. Although other members of the Rel family bind DNA as dimers, NF-AT proteins are monomers in solution or bound to DNA. Transfection assays indicate that each of the four NF-AT proteins can activate the IL-2 promoter in T cells.

Coordinate and cooperative roles for NF-AT and AP-1 in the regulation of the murine IL-4 gene

The transcription factor NF-AT plays an essential role in the inducible transcription of several cytokine genes during T cell activation. The distal NF-AT site of the murine IL-2 promoter binds both NF-AT and AP-1 proteins, and thus represents a composite regulatory site that integrates Ca(2+)- and PKC-dependent signaling pathways in T cell activation. However, the individual contributions of the NF-AT and AP-1 components to promoter activity via this composite site have not been resolved, owing to the absence of a clearly defined AP-1 binding site, which, when mutated abolishes AP-1 binding. We describe here an apparently analogous NF-AT/AP-1 composite site in the murine IL-4 promoter, which can be mutated to selectively block the recruitment of each component. We show that the cooperative and coordinate involvement of both NF-AT and AP-1 is necessary for full activity of the NF-AT/AP-1 composite site, and, ultimately, the entire IL-4 promoter.

Drosophila TAFII40 interacts with both a VP16 activation domain and the basal transcription factor TFIIB

Enhancement of RNA polymerase II transcription by the viral transactivator VP16 requires the TFIID complex, which consists of the TATA-binding protein (TBP) and TBP-associated factors (TAFs). Here we report the molecular cloning, expression, and biochemical characterization of Drosophila TAFII40 (dTAFII40), a subunit of TFIID. In vitro protein-protein interaction assays revealed direct binding between dTAFII40 and a 39 amino acid VP16 activation domain. In addition, affinity chromatography indicated a direct binding of the basal factor TFIIB to immobilized dTAFII40. Since VP16 also binds TFIIB, our results suggest a ternary interaction among an activator, a coactivator, and a basal transcription factor. Antibodies directed against dTAFII40 inhibited activation by GAL4-VP16 without affecting basal transcription. These results, taken together with previous studies of Sp1 and dTAFII110, establish that different activators interact with distinct TAFs in the TFIID complex and that TAFs can contact both activators and basal factors.

A new factor related to TATA-binding protein has highly restricted expression patterns in Drosophila

The TATA-binding protein TBP is necessary for the transcription of eukaryotic genes. Multi-protein complexes formed by TBP and different TBP-associated factors are involved in the initiation of transcription by polymerases I and II, and probably III as well. During the formation of an active initiation complex, TBP makes specific contacts with other proteins, for example TFIIB and RNA polymerase II (refs 2-4). Here we describe the cloning and characterization of a Drosophila gene product with considerable sequence similarity to TBP and a highly restricted expression pattern in the embryo. This TBP-related factor is a DNA-binding protein but is not likely to be a basal transcription factor. Our results suggest that TBP-related factor is a sequence-specific transcription factor that shares the DNA-binding properties of TBP.

Molecular cloning and functional analysis of Drosophila TAF110 reveal properties expected of coactivators

The general transcription factor TFIID is a multiprotein complex containing the TATA-binding protein and several associated factors (TAFs), some of which may function as coactivators that are essential for activated, but not basal, transcription. Here we describe the isolation and characterization of the first gene encoding a TAF protein. The deduced amino acid sequence of TAF110 revealed the presence of several glutamine- and serine/threonine-rich regions reminiscent of the protein-protein interaction domains of the regulatory transcription factor Sp1 that are involved in transcription activation and multimerization. In both Drosophila cells and yeast, TAF110 specifically interacts with the glutamine-rich activation domains of Sp1. Moreover, purified Sp1 selectively binds recombinant TAF110 in vitro. These findings taken together suggest that TAF110 may function as a coactivator by serving as a site of protein-protein contact between activators like Sp1 and the TFIID complex.