Retroviral Tagged Cancer Gene Database Research Articles

Abstract 2688: A forward genetics screen of murine brain tumors identifies novel candidate genes involved in gliomagenesis

Abstract Glioma is the most frequent malignant brain tumor in adults. Platelet-derived growth factor (PDGF) signaling is commonly activated in glioma. We have used a retrovirus-driven PDGFB-induced murine glioma model that causes tumors that closely resemble human gliomas of various grades. Knowing that retroviruses have a capacity to induce insertional mutagenesis, we have employed whole genome sequencing to identify potential genes that, together with PDGFB, drive glioma development. Gliomas were induced by RCAS virus injection into the brains of mice expressing the RCAS retroviral receptor from specific promoters. Genomic DNA from tumor cell lines was probed for retroviral tags and sequenced to identify genomic targets of the retrovirus. A streamlined analysis pipeline was developed for retrovirus integration detection and mapping to the reference mouse genome. Integration sites were analyzed and a common integration site (CIS) label was assigned to a gene, given that it was either tagged by a retrovirus more than once within a discovery set or found within the Retroviral Tagged Cancer Gene Database (RTCGD). In a small discovery subset of 15 murine gliomas, we have identified 40 CIS, of which 37 were validated by Sanger sequencing. When compared with previously identified CIS in RTCGD, 5.5% of them were shared with our older screen, where we overexpressed PDGFB from another retrovirus in order to induce glioma. Less CIS genes were shared with other published tumor models induced by viruses driven by other cancer genes/viruses. The majority of genes identified in our screen were tagged twice. However, Nfic, Cuecd1, Thra, Foxj1 and Nrxn1 were tagged three times, Ppfibp1 and Rhbg four times, and Mir29a/29b-1 seven times. As compared to control tumor lines, two top candidate genes, Mir29a and Ppfibp1, demonstrated significantly increased expression in tumor lines in were they were respectively tagged. Mir29a is often found downregulated in human tumors including gliomas, still high levels of Mir29a are sometimes found in certain aggressive cancers and in metastases. Interestingly, we found that specific PDGFR inhibition negatively regulates Mir29a, indicating a possible role for PDGF signaling in Mir29a regulation. Ppfibp1 has not been extensively studied in cancer. However, Ppfibp1 seems to have a subgroup-specific expression in human glioblastoma, making it an interesting candidate for further analysis. Here we present a new screening method that can be employed to identify genes involved in PDGFB-driven gliomagenesis. So far, we have identified 37 candidate genes by whole genome sequencing. Two of the most frequently tagged candidates, Mir29a and Ppfibp1 were upregulated as a consequence of retroviral mutagenesis. Their precise role in driving glioma formation in collaboration with PDGF is currently explored. Citation Format: Matko Cancer, Holger Weishaupt, Gabriela Rosen, Ignas Bunikis, Yiwen Jiang, Smitha Sreedharan, Sara Bolin, Ulf Gyllensten, Oren J. Becher, Lene Uhrbom, Adam Ameur, Fredrik J. Swartling. A forward genetics screen of murine brain tumors identifies novel candidate genes involved in gliomagenesis. [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 2688.

Identification of co-occurring insertions in cancer genomes using association analysis.

Collections of tumor genomes created by insertional mutagenesis experiments, e.g., the Retroviral Tagged Cancer Gene Database (RTCGD), can be analysed to find connections between mutations of specific genes and cancer. Such connections are found by identifying the locations of insertions or groups of insertions that frequently occur in the collection of tumor genomes. Recent work has employed a kernel density approach to find such commonly occurring insertions or co-occurring pairs of insertions. Unfortunately, this approach is extremely compute intensive for pairs of insertions and even more intractable for triples, etc. We present a technique that can efficiently find commonly co-occurring sets of insertions (or other genomic features) of any length by applying Association Analysis (AA) (frequent pattern mining) techniques from data mining. A comparison to the kernel density approach on RTCGD is provided, as well as results of the association approach on two other tumor data sets.

Retroviral Insertional Mutagenesis In Egr1+/- mice, Haploinsufficient For a Human Del(5q) Myeloid Leukemia Gene, Develop Myeloid Neoplasms With Proviral Insertions In Genes Syntenic To Human 5q

Therapy-related myeloid neoplasm (t-MN) is a distinctive clinical syndrome occurring after treatment with chemotherapy and/or radiotherapy, typically for a primary malignant disease. Loss of the long arm of chromosome 5, del(5q), is the most common recurring cytogenetic abnormality and is observed in 40% of t-MN patients, as well as 10-15% of patients with primary MDS or AML de novo. These deletions typically encompass over 70Mb [spanning 5q14-q33] and numerous genes, making the identification of relevant del(5q) genes very challenging. In previous studies, we identified a commonly deleted region within 5q and identified Egr1 as a del(5q) haploinsufficient myeloid leukemia gene. Using Egr1+/- mice, we previously showed that Egr1 cooperates with mutations, induced by the alkylating agent, ENU, to induce a myeloproliferative disorder with ineffective erythropoiesis (MPD). However, loss of Egr1 on its own was not sufficient for the development of MPD. To identify cooperating mutations in MPDs in Egr1 haploinsufficient mice, we conducted a retroviral insertional mutagenesis (RIM) screen. Egr1 WT (n=61) and Egr1+/- (n=77) neonates were injected with the MOL4070LTR retrovirus. Although the overall survival of MOL4070LTR-treated Egr1+/- and WT controls was similar, Egr1+/- mice developed MPD or AML with a shorter latency and at a higher overall frequency than WT littermate controls. Forty-six percent of WT mice developed myeloid disease versus 61% for Egr1+/- mice, with a median survival of 474 d for Egr1 WT and 389 d for Egr1+/- mice (p=0.03).We mapped the retroviral integration sites in myeloid neoplasms from 29 WT and 46 Egr1+/- mice using barcoded splinkerette PCR and Illumina high-throughput sequencing. To identify and analyze the statistically significant common insertion sites (CISs) we used the TAPDANCE software developed by A. Sarver. In total, 159 CISs were identified in WT mice and 365 CISs were identified in Egr1+/- mice. Several of these CIS-associated genes, such as Sox4, Pim1 and Myb have been previously identified in other genetic screens according to the Retroviral Tagged Cancer Gene Database (RTCGD). As the main goal of this study is to identify mutations that cooperate with Egr1 haploinsufficiency, we were particularly interested in CISs that were identified exclusively or more frequently in Egr1+/- mice with myeloid neoplasms. The TAPDANCE software automatically identifies associations between phenotypes and CISs using Fisher's exact test with multiple testing correction. Using this analysis, we identified six CISs that were statistically associated with myeloid neoplasms in Egr1+/- mice, but not WT mice. The candidate cancer genes proximal to these proviral insertions included Evi1, Gfi1, Evi5, and Cd47. Analysis of transcript levels revealed elevated expression of Evi1, but not Gfi1, Evi5, or Cd47 in myeloid leukemias with proviral insertions proximal to these genes.Of interest was a CIS associated with Egr1+/- mice that mapped to a region of mouse chromosome 18 that is syntenic to human 5q31.2, proximal to the Dnajc18, Ecscr, Tmem173, Cxxc5 and Psd2 genes and adjacent to the commonly deleted region. Moreover, an analysis of co-occurring CISs revealed that this CIS co-occurred with a CIS that mapped to a region of mouse chromosome 13 that is syntenic to human 5q31.1, also deleted in t-MN patients and proximal to the Tifab, and H2afy genes. Of the genes in these two CISs, CXXC5, TMEM173, TIFAB, and H2AFY each show significantly decreased expression in bone marrow cells from t-MN del(5q) patients, consistent with haploinsufficiency. Identification of the relevant del(5q) genes in t-MN continues to be a challenge for developing therapeutic targets. Loss of expression of the tumor suppressor gene, EGR1, which is expressed at haploinsufficient levels in t-MN patients with a del(5q), on its own is not sufficient for the development of myeloid leukemia. Here we performed a forward genetic screen with Egr1+/- mice and have identified several candidate del(5q) genes, including CXXC5, TMEM173, TIFAB, and H2AFY, that should now be evaluated as candidate genes that cooperate with EGR1 haploinsufficiency in the pathogenesis of t-MN. The identification of aberrant pathways resulting from haploinsufficiency of EGR1 in cooperation with these del(5q) genes may potentially lead to the new therapeutic targets for t-MNs with chromosome 5 abnormalities. Disclosures:Largaespada:Discovery Genomics, Inc: Consultancy, Share Holder Other; NeoClone Biotechnology, Inc: Consultancy, Share Holder, Share Holder Other.

New methods for finding common insertion sites and co-occurring common insertion sites in transposon- and virus-based genetic screens

Insertional mutagenesis screens in mice are used to identify individual genes that drive tumor formation. In these screens, candidate cancer genes are identified if their genomic location is proximal to a common insertion site (CIS) defined by high rates of transposon or retroviral insertions in a given genomic window. In this article, we describe a new method for defining CISs based on a Poisson distribution, the Poisson Regression Insertion Model, and show that this new method is an improvement over previously described methods. We also describe a modification of the method that can identify pairs and higher orders of co-occurring common insertion sites. We apply these methods to two data sets, one generated in a transposon-based screen for gastrointestinal tract cancer genes and another based on the set of retroviral insertions in the Retroviral Tagged Cancer Gene Database. We show that the new methods identify more relevant candidate genes and candidate gene pairs than found using previous methods. Identification of the biologically relevant set of mutations that occur in a single cell and cause tumor progression will aid in the rational design of single and combinatorial therapies in the upcoming age of personalized cancer therapy.

Sox4 and CREB Cooperate in Myeloid Transformation,

Abstract 3452The cAMP Response Element Binding Protein, CREB, is a transcription factor that is critical for cell proliferation and survival in neuronal and hematopoietic cells. We previously reported that CREB is overexpressed in leukemic blasts from patients with AML. CREB overexpression is also associated with an increased risk of relapse and decreased event-free survival in AML patients. We generated transgenic mice in which CREB is expressed under the control of the myeloid specific promoter, hMRP8. In vitro, bone marrow progenitors from CREB transgenic mice have increased proliferative potential and replating ability and these mice develop myeloproliferative disease, but not AML. Therefore, CREB is not sufficient to fully transform hematopoietic cells. To identify genes that cooperate with CREB, we performed retroviral insertional mutagenesis with CREB or wildtype C57/Blk6 mice. Newborn mice were infected with the MOL4070LTR retrovirus. Leukemia was observed in both CREB transgenic and wild type mice, however, latency for disease was significantly shortened in CREB transgenic mice at 9 months vs. 14 months for wild type mice (p<0.001). Infected mice developed hepatosplenomegaly (weighing up to 10-fold more than spleens from uninfected wild type mice). The leukemic phenotype by FACs analysis showed that 75% (24/32 mice analyzed) had AML with blast cells expressing Gr-1/Mac-1. Therefore, wild type MOL4070LTR infected mice had decreased penetrance for myeloid disease (45% or 5/11 mice analyzed). To identify potential cooperating oncogenes, genomic DNA was subjected to long-mediated PCR analysis to amplify genomic sequences upstream of the virus. Viral integration sites were identified by Southern blot analysis with genomic DNA from spleens from mice with leukemia. These sequences were mapped to the mouse genome to identify the chromosomal location. Blast analysis of retroviral integration sites (RIS) was performed on the publicly available Ensembl database and compared to known common integration sites on the Retroviral Tagged Cancer Gene Database (RTCGD http://rtcgd.ncifcrf.gov/). Sequencing of integration sites identified previously known oncogenes, such as Evi1/Mds, Evi5/Gfi1, Myb/Ahi, and Ras. Common integration sites identified in multiple mice include, sox4, Evi5/Gfi1, Myb/Ahi, Cbfa2t3. The highest incidence of integration in the mutagenesis screen was in the sox4 gene. Sox4 is a transcription factor that regulates embryonic development and cell fate. Transduction of CREB transgenic mouse bone marrow progenitor cells with a sox4 retrovirus increased survival and self-renewal by 2-fold (p<0.01) in serial replating and methylcellulose colony assays compared to wild type bone marrow progenitors. CREB overexpressing bone marrow progenitor cells stained with Brdu had a 10% increase in cycling cells compared to wild type bone marrow progenitor cells transduced with sox4 retrovirus (p<0.0001). Consistent with this observation is that sox4 retroviral transduction of CREB transgenic bone marrow cells led to increased CREB and cyclin D expression (p<0.01 and p<0.001, respectively). CREB transgenic mouse bone marrow transduced with sox4 retrovirus and stained with Annexin/7AAD also showed decreased apoptosis when cultured in low serum and absence of growth factor (p<0.05). Furthermore, the expression of the sox4 gene was increased in bone marrow cells from AML patients that also overexpressed CREB. Our results indicate that sox4 and CREB cooperate and contribute to transformation of hematopoietic progenitor cells. Disclosures:No relevant conflicts of interest to declare.

An Integrated Genetic-Genomic Approach for the Identification of Novel Cancer Loci in Mice Sensitized to c-Myc-Induced Apoptosis

Deregulated c-Myc is associated with a wide range of human cancers. In many cell types, overexpression of c-Myc potently promotes cell growth and proliferation concomitant with the induction of apoptosis. Secondary genetic events that shift this balance either by increasing growth and proliferation or limiting apoptosis are likely to cooperate with c-Myc in tumorigenesis. Here, the authors have performed large-scale insertional mutagenesis in Eμ-c-myc mice that, through mdm2 loss of function mutations, are sensitized to apoptosis. The authors chose to use this genetic background based on the hypothesis that the high level of apoptosis induced by c-Myc overexpression in MDM2-deficient mice would act as a rate-limiting barrier for lymphoma development. As a result, it was predicted that the spectrum of retroviral insertions would be shifted toward loci that harbor antiapoptotic genes. Nine novel common insertion sites (CISs) specific to mice with this sensitized genetic background were identified, suggesting the presence of novel antiapoptotic cancer genes. Moreover, cross-comparing the data to the Retroviral Tagged Cancer Gene Database, the authors identified an additional 23 novel CISs. Here, evidence is presented that 2 genes, ppp1r16b and hdac6, identified at CISs, are bona fide cellular oncogenes. This study highlights the power of combining unique sensitized genetic backgrounds with large-scale mutagenesis as an approach for identifying novel cancer genes.

The Mir-17-19b Microrna Cluster Quantitatively Regulates the Self- Renewing Leukemia Stem Cell Compartment in MLL Leukemia

The critical molecular determinants for leukemia stem cell (LSC) generation and maintenance are largely unknown. While the majority of studies focus on investigating key transcription factors and cytokines involved in hematological malignancies, very little is known about the roles of recently discovered microRNAs in leukemogenesis. Bioinformatic analysis of the Mouse Retroviral Tagged Cancer Gene Database (Neal Copeland et al.) revealed the 5′ regulatory region of the miR-142 gene as a recurrent target for retroviral integrations in a variety of mouse leukemia and lymphoma cancer models, including MLL-induced leukemia, thereby implicating miR-142 as a potential collaborating oncomir in hematologic malignancies. Subsequent expression analyses of differentially expressed microRNAs in LSC-enriched cell fractions (c-kit+) from MLL leukemias in comparison with various stages of normal hematopoietic progenitors indicated that the miR-17-5p family of microRNAs is highly expressed in LSC-enriched fractions and their possible normal counterparts, granulocyte-macrophage progenitors. Moreover, the expression of miR-142 and miR-17-5p family microRNAs was substantially reduced upon differentiation and loss of self-renewal in cells transformed by the conditional MLL-ENL-ER oncogene following withdrawal of tamoxifen. Forced expression of the miR-17-19b polycistron or miR-142 substantially shortened the latency for MLL leukemia development. In particular, leukemias expressing high levels of the miR-17-19b polycistron displayed a higher frequency of LSCs due to an efficient block of differentiation and enhanced proliferation associated with reduced expression of p21, a cyclin dependent kinase inhibitor previously implicated as a direct target of miR-17-19b microRNAs. Knock down of p21 in MLL oncogene transduced cells recapitulates the MLL-leukemic phenotype induced by over-expression of the miR-17-19b cluster including a significant decrease in MLL leukemia latency. This observation validates p21 as a biologically relevant in vivo target of the miR-17-19b polycistron in MLL leukemia. Thus, microRNAs are pathogenically important for quantitatively modulating the self-renewing LSC compartment in a murine model of MLL leukemia. Potential requirements of the miR-17-19b microRNA cluster for induction and/or maintenance of MLL leukemia are currently under study.

Insertional Gene Activation by Lentiviral and Gammaretroviral Vectors

Gammaretroviral and lentiviral vectors are promising tools for gene therapy, but they can be oncogenic. The development of safer vectors depends on a quantitative assay for insertional mutagenesis. Here we report a rapid, inexpensive, and reproducible assay which uses a murine cell line to measure the frequency of interleukin-3 (IL-3)-independent mutants. Lentiviral and gammaretroviral vectors cause insertional mutagenesis at similar frequencies; however, they use different mechanisms. Human immunodeficiency virus (HIV)-based vectors generate mutants by insertion only into the growth hormone receptor (Ghr) locus. The HIV enhancer/promoter is active in the absence of the HIV Tat protein in this locus, and an HIV/Ghr spliced transcript expresses GHR and cells respond to GH. Deletion of the enhancer/promoter in a self-inactivating HIV-based vector prevents this mechanism of insertional mutagenesis. In contrast, gammaretroviral vectors insert into other loci, including IL-3 and genes identified as common insertion sites in the Retroviral Tagged Cancer Gene Database (RTCGD).

Co-occurrence analysis of insertional mutagenesis data reveals cooperating oncogenes

Cancers are caused by an accumulation of multiple independent mutations that collectively deregulate cellular pathways, e.g. such as those regulating cell division and cell-death. The publicly available Retroviral Tagged Cancer Gene Database (RTCGD) contains the data of many insertional mutagenesis screens, in which the virally induced mutations result in tumor formation in mice. The insertion loci therefore indicate the location of putative cancer genes. Additionally, the presence of multiple independent insertions within one tumor hints towards a cooperation between the insertionally mutated genes. In this study we focus on the detection of statistically significant co-mutations. We propose a two-dimensional Gaussian Kernel Convolution method (2DGKC), a computational technique that identifies the cooperating mutations in insertional mutagenesis data. We define the Common Co-occurrence of Insertions (CCI), signifying the co-mutations that are statistically significant across all different screens in the RTCGD. Significance estimates are made on multiple scales, and the results visualized in a scale space, thereby providing valuable extra information on the putative cooperation. The multidimensional analysis of the insertion data results in the discovery of 86 statistically significant co-mutations, indicating the presence of cooperating oncogenes that play a role in tumor development. Since oncogenes may cooperate with several members of a parallel pathway, we combined the co-occurrence data with gene family information to find significant cooperations between oncogenes and families of genes. We show, for instance, the interchangeable cooperation of Myc insertions with insertions in the Pim family. A list of the resulting CCIs is available at: http://ict.ewi.tudelft.nl/~jeroen/CCI/CCI_list.txt.

Detecting Statistically Significant Common Insertion Sites in Retroviral Insertional Mutagenesis Screens

Retroviral insertional mutagenesis screens, which identify genes involved in tumor development in mice, have yielded a substantial number of retroviral integration sites, and this number is expected to grow substantially due to the introduction of high-throughput screening techniques. The data of various retroviral insertional mutagenesis screens are compiled in the publicly available Retroviral Tagged Cancer Gene Database (RTCGD). Integrally analyzing these screens for the presence of common insertion sites (CISs, i.e., regions in the genome that have been hit by viral insertions in multiple independent tumors significantly more than expected by chance) requires an approach that corrects for the increased probability of finding false CISs as the amount of available data increases. Moreover, significance estimates of CISs should be established taking into account both the noise, arising from the random nature of the insertion process, as well as the bias, stemming from preferential insertion sites present in the genome and the data retrieval methodology. We introduce a framework, the kernel convolution (KC) framework, to find CISs in a noisy and biased environment using a predefined significance level while controlling the family-wise error (FWE) (the probability of detecting false CISs). Where previous methods use one, two, or three predetermined fixed scales, our method is capable of operating at any biologically relevant scale. This creates the possibility to analyze the CISs in a scale space by varying the width of the CISs, providing new insights in the behavior of CISs across multiple scales. Our method also features the possibility of including models for background bias. Using simulated data, we evaluate the KC framework using three kernel functions, the Gaussian, triangular, and rectangular kernel function. We applied the Gaussian KC to the data from the combined set of screens in the RTCGD and found that 53% of the CISs do not reach the significance threshold in this combined setting. Still, with the FWE under control, application of our method resulted in the discovery of eight novel CISs, which each have a probability less than 5% of being false detections.

Cooperative Pathways to Acute Myeloid Leukemia Include the Combining of Transcription Factor Alterations: PML-RARα Cooperates with SOX4.

One paradigm for transformation of normal myeloid cells into acute myeloid leukemia (AML) is the combining of transcription factor alteration with activation of signaling pathways, including cytokine receptor activation. To identify possible alternative pathways to leukemogenesis we crossed MRP8-PML/RARA transgenic mice with BXH-2 mice, which harbor an endogenous retrovirus that causes AML. Approximately half of the leukemias that arose in this cross showed features of acute promyelocytic leukemia (APL). We identified 22 insertions sites in 8 APL-like leukemias. Of these, 7 represented common insertion sites in the Mouse Retroviral Tagged Cancer Gene Database (RTCGD, http://rtcgd.ncifcrf.gov/mm4/index.html). We introduced into a retroviral vector cDNAs encoding 6 genes located at retroviral insertion sites identified in PML/RARA leukemias (Sox4, Meis1, Lck, Sfpi1, Nfil3, Cblb). PML/RARA transgenic bone marrow was transduced with these retroviruses and the marrow was used to reconstitute lethally irradiated recipient animals. Recipients of PML/RARα transgenic marrow transduced with a SOX4 retrovirus developed APL-like leukemias in 3 months. SOX4 is an HMG box transcription factor that exhibits sequence specific DNA binding and transactivation. It is the most common insertion site in RTCGD, and has been found to be overexpressed in small cell lung cancer and medulloblastoma, as well as other tumors. In light of other data indicating that combining transcription factor abnormalities can rapidly induce leukemia (e.g. Hoxa9 + Meis1), our finding that SOX4 + PML/RARα is potently leukemogenic supports the hypothesis that such cooperativity represents another important paradigm in myeloid leukemogenesis.

RTCGD: retroviral tagged cancer gene database.

Retroviral insertional mutagenesis in mouse hematopoietic tumors provides a potent cancer gene discovery tool in the post-genome-sequence era. To manage multiple high-throughput insertional mutagenesis screening projects, we developed the Retroviral Tagged Cancer Gene Database (RTCGD; http://RTCGD.ncifcrf.gov). A sequence analysis pipeline determines the genomic position of each retroviral integration site cloned from a mouse tumor, the distance between it and the nearest candidate disease gene(s) and its orientation with respect to the candidate gene(s). The pipeline also identifies genomic regions that are targets of retroviral integration in more than one tumor (common integration sites, CISs) and are thus likely to encode a disease gene. Users can search the database using a specified gene symbol, chromosome number or tumor model to identify both CIS genes and unique viral integration sites or compare the integration sites cloned by different laboratories using different models. As a default setting, users first review the CIS Lists and then Clone Lists. CIS Lists describe CISs and their candidate disease genes along with links to other public databases and clone lists. Clone Lists describe the viral integration site clones along with the tumor model and tumor type from which they were cloned, candidate disease gene(s), genomic position and orientation of the integrated provirus with respect to the candidate gene(s). It also provides a pictorial view of the genomic location of each integration site relative to neighboring genes and markers. Researchers can identify integrations of interest and compare their results with those for multiple tumor models and tumor types using RTCGD.