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.     

Novel common integration sites targeted by mouse mammary tumor virus insertion in mammary tumors have oncogenic activity

Non-acute transforming retroviruses like mouse mammary tumor virus (MMTV) cause cancer, at least in part, through integration near cellular genes involved in growth control, thereby de-regulating their expression. It is well-established that MMTV commonly integrates near and activates expression of members of the Wnt and Fgf pathways in mammary tumors. However, there are a significant number of tumors for which the proviral integration sites have not been identified. Here, we used high through-put screening to identify common integration sites (CISs) in MMTV-induced tumors from C3H/HeN and BALB/c mice. As expected, members of both the Wnt and Fgf families were identified in this screen. In addition, a number of novel CISs were found, including Tcf7l2, Antxr1/Tem8, and Arhgap18. We show here that expression of these three putative oncogenes in normal murine mammary gland cells altered their growth kinetics and caused their morphological transformation when grown in three dimensional cultures. Additionally, expression of Tcf7l2 and Antxr1/Tem8 sensitized cells to exogenous WNT ligand. As Tcf7l2, Antxr1/Tem8, and Arhgap18 have been associated with human breast and other cancers, these data demonstrate that MMTV-induced insertional mutation remains an important means for identifying genes involved in breast cancer.     

Substitution of feline leukemia virus long terminal repeat sequences into murine leukemia virus alters the pattern of insertional activation and identifies new common insertion sites

The recombinant retrovirus, MoFe2-MuLV (MoFe2), was constructed by replacing the U3 region of Moloney murine leukemia virus (M-MuLV) with homologous sequences from the FeLV-945 LTR. NIH/Swiss mice neonatally inoculated with MoFe2 developed T-cell lymphomas of immature thymocyte surface phenotype. MoFe2 integrated infrequently (0 to 9%) near common insertion sites (CISs) previously identified for either parent virus. Using three different strategies, CISs in MoFe2-induced tumors were identified at six loci, none of which had been previously reported as CISs in tumors induced by either parent virus in wild-type animals. Two of the newly identified CISs had not previously been implicated in lymphoma in any retrovirus model. One of these, designated 3-19, encodes the p101 regulatory subunit of phosphoinositide-3-kinase-gamma. The other, designated Rw1, is predicted to encode a protein that functions in the immune response to virus infection. Thus, substitution of FeLV-945 U3 sequences into the M-MuLV long terminal repeat (LTR) did not alter the target tissue for M-MuLV transformation but significantly altered the pattern of CIS utilization in the induction of T-cell lymphoma. These observations support a growing body of evidence that the distinctive sequence and/or structure of the retroviral LTR determines its pattern of insertional activation. The findings also demonstrate the oligoclonal nature of retrovirus-induced lymphomas by demonstrating proviral insertions at CISs in subdominant populations in the tumor mass. Finally, the findings demonstrate the utility of novel recombinant retroviruses such as MoFe2 to contribute new genes potentially relevant to the induction of lymphoid malignancy.     

Retroviral insertions in the VISION database identify molecular pathways in mouse lymphoid leukemia and lymphoma

AKXD recombinant inbred (RI) strains develop a variety of leukemias and lymphomas due to somatically acquired insertions of retroviral DNA into the genome of hematopoetic cells that can mutate cellular proto-oncogenes and tumor suppressor genes. We generated a new set of tumors from nine AKXD RI strains selected for their propensity to develop B-cell tumors, the most common type of human hematopoietic cancers. We employed a PCR technique called viral insertion site amplification (VISA) to rapidly isolate genomic sequence at the site of provirus insertion. Here we describe 550 VISA sequence tags (VSTs) that identify 74 common insertion sites (CISs), of which 21 have not been identified previously. Several suspected proto-oncogenes and tumor suppressor genes lie near CISs, providing supportive evidence for their roles in cancer. Furthermore, numerous previously uncharacterized genes lie near CISs, providing a pool of candidate disease genes for future research. Pathway analysis of candidate genes identified several signaling pathways as common and powerful routes to blood cancer, including Notch, E-protein, NFκB, and Ras signaling. Misregulation of several Notch signaling genes was confirmed by quantitative RT-PCR. Our data suggest that analyses of insertional mutagenesis on a single genetic background are biased toward the identification of cooperating mutations. This tumor collection represents the most comprehensive study of the genetics of B-cell leukemia and lymphoma development in mice. We have deposited the VST sequences, CISs in a genome viewer, histopathology, and molecular tumor typing data in a public web database called VISION (Viral Insertion Sites Identifying Oncogenes), which is located at . Keith C. Weiser and Bin Liu are authors that contributed equally to this work.     

Sites of retroviral DNA integration: From basic research to clinical applications

One of the most crucial steps in the life cycle of a retrovirus is the integration of the viral DNA (vDNA) copy of the RNA genome into the genome of an infected host cell. Integration provides for efficient viral gene expression as well as for the segregation of viral genomes to daughter cells upon cell division. Some integrated viruses are not well expressed, and cells latently infected with human immunodeficiency virus type 1 (HIV-1) can resist the action of potent antiretroviral drugs and remain dormant for decades. Intensive research has been dedicated to understanding the catalytic mechanism of integration, as well as the viral and cellular determinants that influence integration site distribution throughout the host genome. In this review, we summarize the evolution of techniques that have been used to recover and map retroviral integration sites, from the early days that first indicated that integration could occur in multiple cellular DNA locations, to current technologies that map upwards of millions of unique integration sites from single in vitro integration reactions or cell culture infections. We further review important insights gained from the use of such mapping techniques, including the monitoring of cell clonal expansion in patients treated with retrovirus-based gene therapy vectors, or patients with acquired immune deficiency syndrome (AIDS) on suppressive antiretroviral therapy (ART). These insights span from integrase (IN) enzyme sequence preferences within target DNA (tDNA) at the sites of integration, to the roles of host cellular proteins in mediating global integration distribution, to the potential relationship between genomic location of vDNA integration site and retroviral latency.     

Gammaretroviral integration into nucleosomal target DNA in vivo

Some of the earliest studies of retroviral integration targeting reported that sites of gammaretroviral DNA integration were positively correlated with DNase I-hypersensitive sites in chromatin. This led to the suggestion that open chromatin was favorable for integration. More recent deep sequencing experiments confirmed that gammaretroviral integration sites and DNase I cleavage sites are associated in genome-wide surveys. Paradoxically, in vitro studies of integration show that nucleosomal DNA is actually favored over naked DNA, raising the question of whether integration target DNA in chromosomes is wrapped in nucleosomes or nucleosome free. In this study we examined gammaretroviral integration by infecting primary human CD4(+) T lymphocytes with a murine leukemia virus (MLV)-based retroviral vector or xenotropic murine leukemia virus-related virus (XMRV), and isolated 32,585 unique integration sites using ligation-mediated PCR and 454 pyrosequencing. CD4(+) T lymphocytes were chosen for study because of the particularly dense genome-wide mapping of chromatin features available for comparison. Analysis relative to predicted nucleosome positions showed that gammaretroviruses direct integration into outward-facing major grooves on nucleosome-wrapped DNA, similar to the integration pattern of HIV. Also, a suite of histone modifications correlated with gene activity are positively associated with integration by both MLV and XMRV. Thus, we conclude that favored integration near DNase I-hypersensitive sites does not imply that integration takes place exclusively in nucleosome-free regions.     

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.     

Organization and integration sites in the human genome of endogenous retroviral sequences belonging to HERV-E family

The present paper reports the characterization of HERV-E endogenous retroviral sequences in the human genome by using three complementary approaches. Firstly, we identified genomic clones containing HERV-E by BLAST screening of human DNA databases by using the entire sequence of a characterized HERV-E clone (M10976) as a query. The genomic structure and integration sites of the HERV-E elements were characterized. Secondly, new integration sites of HERV-E elements were revealed by a retroviral LTR-arbitrary primer-PCR (RELAP-PCR) technique. BLAST analysis of the PCR products identified a subgroup that shows low identity (75%) to the original clone M10976 and slightly higher identity (80%) to a closely related HERV-E (Ac. n. K02166). Finally, we performed FISH mapping, which revealed sites of integration of HERV-E not previously identified at the cytogenetic level. A preliminary analysis of genes mapping in the same bands as HERV-E integration sites was performed: several loci relevant to physiological and/or pathological processes were detected. Our findings may provide clues to identify HERV-E integration sites adjacent to genes with important biological roles.     

The genes encoding the peripheral cannabinoid receptor and alpha-L-fucosidase are located near a newly identified common virus integration site, Evi11.

A new common region of virus integration, Evi11, has been identified in two retrovirally induced murine myeloid leukemia cell lines, NFS107 and NFS78. By interspecific backcross analysis, it was shown that Evi11 is located at the distal end of mouse chromosome 4, in a region that shows homology with human 1p36. The genes encoding the peripheral cannabinoid receptor (Cnr2) and alpha-L-fucosidase (Fuca1) were identified near the integration site by using a novel exon trapping system. Cnr2 is suggested to be the target gene for viral interference in Evi11, since proviruses are integrated in the first intron of Cnr2 and retroviral integrations alter mRNA expression of Cnr2 in NFS107 and NFS78. In addition, proviral integrations were demonstrated within the 3' untranslated region of Cnr2 in five independent newly derived CasBrM-MuLV (mouse murine leukemia virus) tumors, CSL13, CSL14, CSL16, CSL27, and CSL97. The Cnr2 gene encodes a seven-transmembrane G-protein-coupled receptor which is normally expressed in hematopoietic tissues. Our data suggest that the peripheral cannabinoid receptor gene might be involved in leukemogenesis as a result of aberrant expression of Cnr2 due to retroviral integration in Evi11.     

Integration of proviral sequences, but not at the common integration sites of the FGF8 locus, in an androgen-dependent mouse mammary Shionogi carcinoma.

A retroviral insertional mutation, especially by mouse mammary tumor virus (MMTV), is a major cause of murine mammary tumorigenesis. Prompted by our previous finding that FGF8, an insertionally activated cellular oncogene, is highly expressed in androgen-dependent mouse mammary Shionogi carcinoma cells, we here investigated retroviral integration adjacent to the fgf8 locus in Shionogi carcinoma. In the genomic Southern blots for fgf8 and its 5'-upstream gene npm3, the hybridized fragments were identical to the host DD/Sio mice, the original Shionogi carcinoma 115 tumor, and a pair of cultured Shionogi carcinoma cell lines of SC-3 and SC-4, suggesting that no retroviral integration occurred around either loci. The genomic cloning for the fgf8 locus from SC-3 cells also confirmed no MMTV integration. In addition, npm3, which is usually coactivated with fgf8 by MMTV insertion,was not up-regulated by androgens in SC-3 cells. All these findings led us to conclude that no retroviral insertion was present at the common integration sites adjacent to the fgf8 locus in Shionogi carcinoma although we demonstrated in this study that multiple proviral sequences of MMTV, Moloney murine sarcoma virus and FBJ-murine sarcoma virus are integrated into SC-3 cells in association with their distinct promoter activity in SC-3 cells.     

Large-scale identification of disease genes involved in acute myeloid leukemia

Acute myeloid leukemia (AML) is a heterogeneous group of diseases in which chromosomal aberrations, small insertions or deletions, or point mutations in certain genes have profound consequences for prognosis. However, the majority of AML patients present without currently known genetic defects. Retroviral insertion mutagenesis in mice has become a powerful tool for identifying new disease genes involved in the pathogenesis of leukemia and lymphoma. Here we have used the Graffi-1.4 strain of murine leukemia virus, which causes predominantly AML, in a screen to identify novel genes involved in the pathogenesis of this disease. We report 79 candidate disease genes in common integration sites (CISs) and 15 genes whose family members previously were found to be affected in other studies. The majority of the identified sequences (60%) were not found in lymphomas and monocytic leukemias in previous screens, suggesting a specific involvement in AML. Although most of the virus integrations occurred in or near the 5' or 3' ends of the genes, suggesting deregulation of gene expression as a consequence of virus integration, 18 CISs were located exclusively within the genes, conceivably causing gene disruption.     

Methods for integration site distribution analyses in animal cell genomes

The question of where retroviral DNA becomes integrated in chromosomes is important for understanding (i) the mechanisms of viral growth, (ii) devising new anti-retroviral therapy, (iii) understanding how genomes evolve, and (iv) developing safer methods for gene therapy. With the completion of genome sequences for many organisms, it has become possible to study integration targeting by cloning and sequencing large numbers of host–virus DNA junctions, then mapping the host DNA segments back onto the genomic sequence. This allows statistical analysis of the distribution of integration sites relative to the myriad types of genomic features that are also being mapped onto the sequence scaffold. Here we present methods for recovering and analyzing integration site sequences.     

Tnfrsf13c (Baffr) is mis-expressed in tumors with murine leukemia virus insertions at Lvis22

In susceptible strains of mice, leukemia is caused by the somatic integration of murine leukemia retroviruses into the host genome. Integration sites that are common to several tumors are likely to affect genes that are important in oncogenesis. Here we present the analysis of a common site of retroviral integration on mouse chromosome 15, which includes the genomic structure of three genes near the integration site. One of the genes misexpressed at the insertion site has recently been characterized as a B-cell receptor, Tnfrsf13c (formerly Baffr), indicating that this approach is useful in defining genes that function in lymphocyte development and tumor progression. Current genome databases provide powerful resources for the rapid identification of genes at common proviral insertion sites. The characterization of these genes in tumor samples will allow a function to be assigned to many novel loci identified by the genome sequencing projects.     

Insertional Mutagenesis and Deep Profiling Reveals Gene Hierarchies and a Myc/p53-Dependent Bottleneck in Lymphomagenesis

Retroviral insertional mutagenesis (RIM) is a powerful tool for cancer genomics that was combined in this study with deep sequencing (RIM/DS) to facilitate a comprehensive analysis of lymphoma progression. Transgenic mice expressing two potent collaborating oncogenes in the germ line (CD2-MYC, -Runx2) develop rapid onset tumours that can be accelerated and rendered polyclonal by neonatal Moloney murine leukaemia virus (MoMLV) infection. RIM/DS analysis of 28 polyclonal lymphomas identified 771 common insertion sites (CISs) defining a ‘progression network’ that encompassed a remarkably large fraction of known MoMLV target genes, with further strong indications of oncogenic selection above the background of MoMLV integration preference. Progression driven by RIM was characterised as a Darwinian process of clonal competition engaging proliferation control networks downstream of cytokine and T-cell receptor signalling. Enhancer mode activation accounted for the most efficiently selected CIS target genes, including Ccr7 as the most prominent of a set of chemokine receptors driving paracrine growth stimulation and lymphoma dissemination. Another large target gene subset including candidate tumour suppressors was disrupted by intragenic insertions. A second RIM/DS screen comparing lymphomas of wild-type and parental transgenics showed that CD2-MYC tumours are virtually dependent on activation of Runx family genes in strong preference to other potent Myc collaborating genes (Gfi1, Notch1). Ikzf1 was identified as a novel collaborating gene for Runx2 and illustrated the interface between integration preference and oncogenic selection. Lymphoma target genes for MoMLV can be classified into (a) a small set of master regulators that confer self-renewal; overcoming p53 and other failsafe pathways and (b) a large group of progression genes that control autonomous proliferation in transformed cells. These findings provide insights into retroviral biology, human cancer genetics and the safety of vector-mediated gene therapy.     

Radiation leukemia virus common integration at the Kis2 locus: simultaneous overexpression of a novel noncoding RNA and of the proximal Phf6 gene

Retroviral tagging has been used extensively and successfully to identify genes implicated in cancer pathways. In order to find oncogenes implicated in T-cell leukemia, we used the highly leukemogenic radiation leukemia retrovirus VL3 (RadLV/VL3). We applied the inverted PCR technique to isolate and analyze sequences flanking proviral integrations in RadLV/VL3-induced T lymphomas. We found retroviral integrations in c-myc and Pim1 as already reported but we also identified for the first time Notch1 as a RadLV common integration site. More interestingly, we found a new RadLV common integration site that is situated on mouse chromosome X (XA4 region, bp 45091000). This site has also been reported as an SL3-3 and Moloney murine leukemia virus integration site, which strengthens its implication in murine leukemia virus-induced T lymphomas. This locus, named Kis2 (Kaplan Integration Site 2), was found rearranged in 11% of the tumors analyzed. In this article, we report not only the alteration of the Kis2 gene located nearby in response to RadLV integration but also the induction of the expression of Phf6, situated about 250 kbp from the integration site. The Kis2 gene encodes five different alternatively spliced noncoding RNAs and the Phf6 gene codes for a 365-amino-acid protein which contains two plant homology domain fingers, recently implicated in the B?rjeson-Forssman-Lehmann syndrome in humans. With the recent release of the mouse genome sequence, high-throughput retroviral tagging emerges as a powerful tool in the quest for oncogenes. It also allows the analysis of large DNA regions surrounding the integration locus.