Skip to content

Advertisement

  • Review
  • Open Access

The utility of transposon mutagenesis for cancer studies in the era of genome editing

Contributed equally
Genome Biology201516:229

https://doi.org/10.1186/s13059-015-0794-y

  • Published:

Abstract

The use of transposons as insertional mutagens to identify cancer genes in mice has generated a wealth of information over the past decade. Here, we discuss recent major advances in transposon-mediated insertional mutagenesis screens and compare this technology with other screening strategies.

Keywords

  • Vemurafenib
  • Insertional Mutagenesis
  • Sleep Beauty
  • Candidate Cancer Gene
  • Common Insertion Site

Introduction

Genome sequencing has revealed a plethora of mutations in cancer, with some tumors carrying tens of thousands of somatic mutations [1]. Importantly, the relevance of these mutations is not always intrinsically clear and as a result must be inferred from the types of mutations observed, their frequency across tumor types, and their predicted effects on protein function. Insertional mutagenesis screens provide a functional readout to complement these sequencing studies, as genes identified by insertional mutagens are likely to represent both functionally important and evolutionarily conserved cancer genes. Insertional mutagenesis studies can also highlight cancer genes or common pathways that are disrupted at low frequency or by processes not immediately obvious from the genome sequence alone.

The first insertional mutagenesis efforts in mice were performed with the murine leukemia virus and the mouse mammary transforming virus to induce lymphoma and mammary tumors [2, 3], respectively, and led to the identification of numerous cancer pathways, including the WNT pathway [4]. However, these viruses were found to be of limited utility for mutagenesis in other tissue types owing to viral tropism and the fact that they only infect replicating cells [5]. Furthermore, as these retroviruses generate insertions that activate gene expression, they almost exclusively tag proto-oncogenes [5], restricting our ability to identify other types of cancer genes such as tumor suppressors.

For these reasons, DNA transposons were developed as insertional mutagens [6]. Transposons are mobile elements that move through the genome by a cut-and-paste process (DNA transposons), or through an RNA intermediate in a copy-and-paste mechanism (retrotransposons) [7]. Endogenous transposons are ubiquitous in vertebrate genomes, comprising approximately 45 % of DNA sequence [8], but are largely silent as a result of inactivating mutations acquired through evolution. The introduction of exogenous DNA transposons allows insertional mutagenesis in a wider spectrum of tissues than the ones that are accessible with retroviruses, and thus the generation of new mouse tumor models [9, 10]. The most commonly used transposon systems are the Sleeping Beauty (SB) and piggyBac (PB) systems [11]. A typical transposon used for in vivo insertional mutagenesis contains splice acceptors (SAs) followed by polyadenylation signals (pA) in both orientations, and a unidirectional promoter upstream of a splice donor (SD). A transposon can either disrupt gene function when it integrates into the body of a gene, thereby intercepting and curtailing transcription through the SA–pA elements, or it can activate expression when inserted upstream of a gene as the promoter–SD module drives expression of downstream sequences (Fig. 1). The pattern and orientation of transposon integration sites therefore often provide a clue as to whether the affected gene encodes a tumor suppressor or an oncogene.
Fig. 1
Fig. 1

Transposons as insertional mutagens. a Sleeping Beauty (SB) and piggyBac (PB) (black rectangles) are mutagenic transposons that can be mobilized from donor loci (left panel) and reintegrated into other loci (right panel). Repeats in the transposon (arrowheads) are recognized by the Sleeping Beauty or piggyBac transposases (ovals), resulting in the transposon being excised from the genome. Reintegration of mobilized SB or PB transposons can occur at TA and TTAA sites, respectively, catalyzed by transposase activity. b Transposon insertion can promote or disrupt gene expression. In the example depicted in this panel, a transposon integrates between exons 3 and 4 (numbered gray boxes) of a gene. This can result in two possible outcomes: (I) the transposon disrupts gene function by hijacking transcription through the splice acceptor-polyadenylation signal (SA-pA) elements, leading to expression of a truncated transcript (exons 1–3); or (II) the transposon drives expression of the downstream gene sequences (exons 4–7) through the promoter-splice donor (SD) elements. Depending on the integration site, transposons can activate or abrogate expression of either the entire mRNA of a gene or only parts of it

Here, we discuss recent advances in cancer gene discovery using transposons and their role in the era of other mutagenesis tools such as clustered regularly interspaced short palindromic repeats/CRISPR-associated protein 9 (CRISPR/Cas9).

Transposon-mediated insertional mutagenesis

In 2005, the groups of David Largaespada, Nancy Jenkins and Neal Copeland reported the use of the Sleeping Beauty transposon system as a tool for the identification of cancer-promoting genes in transgenic mice [12, 13]. Largaespada and colleagues performed whole-body transposon-mediated insertional mutagenesis (TMIM) with the first-generation T2/Onc transposon, accelerating tumorigenesis in mice null for the tumor suppressor p19Arf gene [12]. Using a more active transposon system (T2/Onc2), Dupuy and colleagues induced predominantly hematopoietic tumors following global mutagenesis in wild-type mice [13]. Following these landmark studies, a variety of transgenic mouse strains harboring different versions of transposons and transposases have been generated and utilized for candidate cancer gene discovery. By targeting SB transposase expression to tissues of interest, a variety of cancers have been generated by mutagenesis [1320]. Additionally, several cancer types have been accelerated by TMIM in combination with sensitizing mutations [2127, 29, 30] (Table 1). Collectively, many candidate cancer genes have been identified in the mouse that have subsequently been found to be relevant clinically and prognostically in human malignancies [20, 24] (Table 1). In a similar way, the PB transposon has been used for cancer gene discovery in the hematopoietic system and pancreas [31, 32].
Table 1

Capacity of TMIM screens to identify common human cancer genes in three cancer typesa

Cancer type

Genes commonly mutated in human cancer

Recurrent genes identified by TMIM screens

Comments

Novel, functionally validated cancer genes identified in TMIM screens

Colorectal

APC [102]

Yes [19, 22, 23, 51]

 

CNOT1, PDE4DIP, PDCD6IP, ATF2, SFI1 [22]

 

TP53 [102]

Yes (low frequency) [23, 51]

Tp53 has been rarely targeted in any TMIM screen, although upstream regulators such as Cdkn2a are frequently targeted in TMIM screens

ANKRD11, CSNK2A1, MKL2, MYO9A, RNF43, SIN3A, Zfp292 [51]

 

KRAS [102]

Yes (low frequency) [23, 51]

Insertions in Kras have been detected at low frequency in Apc-deficient backgrounds [23], perhaps reflecting the inability of transposons to cause point mutations in target genes through insertion alone. However, upstream and downstream genes within the Kras signaling pathway are targeted in TMIM screens, leading to pathway activation

 

PIK3CA [102]

No

Other phosphoinositide 3-kinase pathway genes have been targeted, for example Pik3r1 and Pten

 

FBXW7 [102]

Yes [23, 51]

 
 

SMAD4 [102]

Yes [23, 51]

 
 

CTNNB1 [102]

Yes [23, 51]

 
 

NRAS [102]

No

Recurrent insertions in Nras have not been found, probably owing to the same reasons applicable to Kras

 

TCF7L2 [102]

Yes [23, 51]

 
 

FAM123B [102]

No

Other components of Wnt–β-catenin pathway are targeted in tumors, for example Apc and Ctnnb1.

Melanoma

BRAF [103, 104]

Yes [49]

Identified as potential mediator of BRAF inhibitor resistance

ERAS [49]

 

NRAS [103, 104]

No

 
 

TP53 [103, 104]

No

 

CEP350 [38]

 

PTEN [103, 104]

Yes [38, 49]

 

MAGI2, PTPRO, Map3k1 [57]

 

CDKN2A [103, 104]

Yes [38, 49, 57]

 
 

NF1 [103, 104]

Yes [38]

Melanoma TMIM screens were performed in a Braf-mutant background. NF1 mutations are typically mutually exclusive from BRAF mutations in human melanoma, potentially accounting for the lack of Nf1 insertions in some screens

 

RAC1 [103, 104]

No

Other genes operating in Rho GTPase pathways have been targeted

 

MAP2K1 [103, 104]

No

Other MAP kinase pathway genes have been targeted in melanoma TMIM screens

 

ARID2 [103, 104]

No

 
 

PPP6C [103, 104]

Yes [38, 49]

Identified as potential mediator of BRAF inhibitor resistance

Pancreatic

KRAS [105, 106]

Yes (low frequency) [24]

Pancreatic cancer TMIM screens were performed in a Kras G12D background

USP9X [24]

 

TP53 [105, 106]

No

Recurrent insertions in Tp53 were not observed despite the high prevalence of TP53 alterations in human pancreatic cancer. Cdkn2a was a recurrent CIS-associated gene, and Usp7, a p53-regulatory deubiquitinase, was targeted at low frequency in one study [24]

Foxp1, Foxp2, Bcl6, Fign [32]

 

CDKN2A [105, 106]

Yes [24, 32]

 
 

SMAD4 [105, 106]

Yes [24, 25]

 
 

PREX2 [106]

No

 
 

TGFBR2 [105, 106]

Yes [24, 32]

 
 

RNF43 [106]

Yes [24, 25]

 
 

KDM6A [106]

Yes [24, 25, 32]

 
 

ARID1A [105, 106]

Yes [24, 32]

 
 

MLL3 [105]

Yes [24, 25]

 

aShown are details of TMIM screens for identifying common human cancer genes in three types of cancer for which more than one screen has been performed. TMIM transposon-mediated insertional mutagenesis

TMIM — technical considerations

Various mouse strains have been generated that express SB or PB transposase in a ubiquitous or conditional manner. With these strains, transposon mobilization can be induced either in the whole animal or in a tissue- or temporal-restricted manner by using an appropriate Cre recombinase allele (Fig. 2). The transposon mice are transgenic strains containing transposon concatemers on a single chromosome. As a consequence, many insertion sites are found locally, and the tendency for local integrations is reported as being higher with SB compared with PB [33]. The number of transposons in the concatemer is also a consideration. Global mobilization of greater than 20–30 transposon copies during embryonic development correlated with increased embryonic lethality [13, 15, 31]. Additionally, increasing transposon numbers amplifies the potential for passenger integrations, which do not contribute to the observed phenotype.
Fig. 2
Fig. 2

Tools for transposon-mediated mutagenesis. a Transposase expression can be either ubiquitous (ub. prom.) or directed to a particular cell or tissue type by using Cre-inducible alleles of the transposase enzyme. In the latter case, a loxP-site-flanked transcriptional stop element (gray triangles and STOP sign, respectively) prevents transcription of the gene encoding transposase. Upon Cre-mediated excision of the loxP-STOP-loxP element, the transposase is expressed in Cre-positive cells. b A variety of transposons have been developed for mutagenesis. SB transposons have been developed that carry either a murine stem cell virus (MSCV) promoter (T2/Onc and T2/Onc2) or the chicken β-actin/CMV enhancer (CAG) promoter (T2/Onc3). To facilitate gene activation, transposons carrying these promoters also contain splice donor (SD) elements, and, for gene disruption, splice acceptor (SA) and polyadenylation (pA) elements (bi-pA bi-directional polyadenylation signal). Versatile SB/PB transposons containing terminal repeats recognized by SB and PB transposases (arrowheads) have also been developed and carry either CAG, MSCV or mouse phosphoglycerate kinase 1 (PGK) promoters (ATP1, ATP2 and ATP3 transposons, respectively). c Using combinations of the aforementioned alleles tabulated here, global or spatiotemporal mutagenesis with co-operating mutations can be performed

The promoter within the transposon can display tissue-specific activity and thereby influence the phenotype of whole-body insertional mutagenesis screens or the insertion sites that are positively selected for in organ-specific screens. Indeed, the first transposon mouse strains (T2/Onc, T2/Onc2) utilized the murine stem cell virus (MSCV) promoter, which displays a propensity for the development of hematopoietic tumors. However, replacing the MSCV promoter with the chicken β-actin/CMV enhancer (CAG) promoter or the phosphoglycerate kinase 1 (PGK) promoter significantly increased the incidence of solid tumors in both the SB and PB system [14, 31]. Thus, the modularity of transposons and the ability to modify elements such as the promoters they carry can be used to influence the tumor type and incidence.

An important technical consideration in transposon screens is integration bias. SB has been reported to demonstrate a bias towards integration into DNA sequences containing TA nucleotides and appears to preferentially integrate into gene bodies but not into transcriptional start sites (TSSs) [34] (Fig. 3). Conversely, PB, which predominantly integrates into TTAA sequences, displays a preference towards integration into TSSs over gene bodies (Fig. 3). As a consequence, oncogenes are more likely to be identified using PB, whereas transposon integration in tumor suppressors is primarily seen when the SB system is used, but this again is influenced by the promoter elements used in the transposon. Allan Bradley’s group recently reported the development of a conditional PB transposase mouse allele [32], which can direct cell- or tissue-specific expression of PB, and hence directs mutagenesis to a specific cellular compartment. The development of this strain allowed the direct comparison of screening data generated in a mouse model of Kras G12D -driven pancreatic cancer, where a prior screen with the SB transposon system had been performed [24]. The PB screen identified candidate drivers that had also been identified by the pancreatic SB screen as well as novel candidate pancreatic cancer genes, and thus exemplified the complementarity of the SB and PB approaches as in vivo insertional mutagens for cancer gene discovery.
Fig. 3
Fig. 3

Integration biases of SB and PB transposons. The distribution of transposon insertions across genes from 5 kb upstream of the transcription start site (TSS) to 5 kb downstream of the transcription termination site (TTS). Red, transposon insertions in the sense orientation relative to the gene; blue, insertions in antisense direction. Reproduced from [34]

Another consideration that investigators should be mindful of when performing insertional mutagenesis screens is the damage done to the genome by the process of transposition itself as transposons are mobilized from chromosomal integration sites. Excision of PB transposons generally results in no or limited damage to the genome; by contrast, the mobilization of SB transposons leaves behind a two-to-five nucleotide footprint [35]. SB transposon footprints can thus result in frameshift mutations, splicing alterations or promoter disruptions, which in turn could promote tumorigenesis. The mobilization of transposons in cis could also result in chromosomal rearrangements such as deletions or copy-number-neutral changes [36]. Fortunately, these passenger effects appear to be limited [3638], and thus tumor promotion in transposon screens appears to be largely driven by transposon insertion events, but this factor is nevertheless of consideration in the analysis of tumors collected during screening.

TMIM — statistical considerations

Although tumor evolution selects for mutagenic insertions that drive tumorigenesis, each tumor cell will harbor multiple additional inconsequential passenger insertions, as repeated rounds of transposon mobilization and reintegration will result in thousands of integration sites in a polyclonal tumor. Cancer drivers cannot be identified solely by sequencing all of the insertion sites in a given tumor — this merely gives a snapshot of insertion sites at a point in time. Thus, statistical approaches are necessary to reveal regions of the genome that are enriched with insertions more than expected by chance — so-called common insertion sites (CISs). By mapping CISs onto a reference genome, CIS-associated genes can be identified as potential cancer drivers.

A number of statistical approaches have been used to identify CIS-associated genes from transposon screens. Early studies deployed Monte Carlo-based methods and Poisson distributions [39, 40] to define those genomic locations enriched with insertion sites. More recently, Gaussian Kernel Convolution (GKC) approaches [41], gene-centric common insertion site (gCIS) analysis [42] and refined versions of the Poisson approach have been developed [43]. Essentially, all these methods provide a measure of the degree to which insertion sites are enriched at a given locus relative to either a pre-computed background distribution or an insertion dataset derived from tissues in which transposons have been mobilized for a short period of days or weeks, before clonal selection could be operative. The concordance between methods ranges between 60 and 80 %, and thus most investigators use multiple algorithms to identify CISs [23]. Methods such as GKC [41] adjust the significance statistic for a locus (CIS) relative to the frequency of the transposon target site (TA for SB, and TTAA for PB) that can account for some local biases in transposon integration. Both the type and stringency of the CIS-calling methods used to identify insertions affect the classification of co-occurring or mutually exclusive CISs. Reinders and colleagues have developed a two-dimensional GKC method to identify co-operating mutations from virally induced mutagenesis data, a method that has also been applied to TMIM screens [44]. In addition, the Poisson regression insertion model (PRIM) [45] has been used to identify co-occurring gene pairs, and the TAPDANCE algorithm can generate the association of independent CISs by using a Fisher’s exact test [43].

Limitations of TMIM

TMIM is a powerful tool for in vivo cancer gene discovery, but, as with every technology, there are several limitations. We summarize these limitations here and also allude to them throughout the text. The primary limitation is the inability of the transposons to interrogate the genome in a completely unbiased fashion. Transposons do not integrate into and affect all genes with similar probability owing to factors such as promoter selection within the transposon [31], integration-site preferences [34], local transposon hopping [33], gene size (larger genes are more likely to be affected by transposon integrations) and the relative superior ease of isolating tumor suppressors as the precise transposon integration site and orientation with respect to the target gene are less crucial factors for tumor suppressors compared with those of oncogenes.

Another limitation is that TMIM cannot recapitulate the complete spectrum of mutations that are commonly found in human cancer, such as point mutations. Elevated expression and mutations may not result in identical biological outcomes, and thus transposon-mediated overexpression of proto-oncogenes does not always mimic the effects of somatic, gain-of-function point mutations [46]. Similarly, mutations in tumor suppressors can result in dominant-negative effects that are not recapitulated by transposon-insertion-mediated loss of expression [47]. The insertion spectrum recovered by TMIM screens can also be affected by the sensitizing genetic backgrounds that activate pro-tumorigenic pathways — for example, oncogenic mutants of B-Raf or Kras [24, 32, 38, 48, 49], such that genes that activate the same pathway as the sensitizing mutation are unlikely to be identified in these particular backgrounds. Finally, transposon insertions are unable to recapitulate reciprocal translocations such as BCRABL and other genomic alterations that commonly occur in cancer.

There are also technical and resource limitations to TMIM approaches. For example, investigators might wish to perform drop-out screens designed to identify genes that are detrimental to cells when mutated. Such screens are not feasible with TMIM as such cells are lost during the screening process. Moreover, the generation of mouse cohorts is both time-consuming and costly for in vivo TMIM screens as compound mutant mice carrying three or four transgenic alleles are typically required. Finally, candidate cancer genes identified through TMIM screens in the mouse might not necessarily have equal relevance in human cancer — follow-up validation studies must therefore be performed. Investigators should consider all these limitations when designing transposon screens.

Transposon mutagenesis — beyond the basic screen

Over the past decade, numerous TMIM studies have identified known and novel cancer genes that either promote tumor initiation or co-operate with cancer-sensitizing mutations to drive tumor progression. Recently, novel and elegant ways of employing transposon mutagenesis to query specific cancer processes have been devised. In this section, we summarize recent developments in the TMIM field.

Investigating tumor progression and evolution

TMIM screens have been performed in mice harboring various initiating mutations found in human cancer. Such screens identify drivers of tumor progression and, importantly, might be influenced by the sensitizing mutation. For example, Alexander and colleagues performed TMIM in the hematopoietic system, which resulted in multiple leukemias [50]. A Jak2 V617F -mutant background skewed the disease towards erythroleukemia, and insertions in the ETS transcription factor genes Erg and Ets1 were identified as the most common events. Conversely, when using an activated ERG allele (TLS-ERG) as the sensitizing mutation, the authors identified frequent activating insertions in Jak2, thus validating the co-operation between Jak2 and Erg [50].

In an elegant study, TMIM was utilized to delineate evolutionary events during the progression of colorectal cancer (CRC) [51]. Jenkins, Copeland and colleagues crossed the SB system into different sensitizing backgrounds that carry mutations in genes that act at different stages of CRC: Apc min , Kras G12D , Smad4 +/− or Tp53 R172H (Fig. 4) [51]. Intriguingly, this approach revealed that functional loss of the wild-type Apc allele was the most crucial event for tumor progression in Apc min , Kras G12D and Tp53 R172H tumors, but not in tumors that were initiated by heterozygous loss of Smad4. Instead, those tumors displayed frequent insertions in the wild-type Smad4 allele along with mutually exclusive insertions in Rspo1 and Rspo2 that promoted overexpression of these R-spondins, which are known enhancers of Wnt signaling. In addition, 111 candidate cancer genes were identified that were independent of the initiating mutation.
Fig. 4
Fig. 4

Use of transposon-mediated insertional mutagenesis (TMIM) screening to identify mutations that co-operate with specific genetic lesions associated with different stages of colorectal cancer development. The top panels illustrate a model of colorectal cancer initiation and progression [101], along with genetic alterations associated with these stages. TMIM screens using mouse models carrying mutations in corresponding genes have revealed that Apc was the predominant gene inactivated in tumors from all sensitizing genotypes apart from Smad4 KO/+ cases, where inactivation of the remaining wild-type Smad4 gene is the most frequent insertional event

These studies illustrate how sensitizing mutations can co-operate with transposon-associated lesions and how different pre-existing mutations can sometimes influence the trajectory of subsequent mutation acquisition during tumor development. In the case of human CRC, loss of APC is thought to be the initiating event, whereas mutations in KRAS, TP53 or SMAD4 occur later during tumor progression. Indeed, transposon-insertion-mediated loss of Apc appeared to be a prerequisite for colon tumorigenesis in the Apc min , Kras G12D and Tp53 R172H backgrounds, whereas insertions in Kras and Tp53 are rare in Apc-loss-driven CRC [51] (Table 1; Fig. 4). This finding further supports the notion of APC being the gatekeeper of CRC. Conversely, leukemogenesis is initiated by either mutant Jak2 or Erg and progresses upon transposon insertions in the other gene, suggesting that the temporal sequence of mutation might be irrelevant [50]. Taken together, TMIM is a valuable tool to delineate tumor progression, and future studies that unravel the genetic dependencies of co-operating mutations on different initiating mutations in other cancer types will shed further light on the genetics of tumor progression and might be useful for devising treatment strategies.

Determining the evolutionary history of mutations within tumors can inform our understanding of the mutational forces that shape cancer development. To assess tumor clonality in a more quantitative fashion, new methods to estimate the frequency of transposon insertions in tumors have been devised. Historical methods to retrieve insertion sites have been based primarily on PCR amplification of restriction-endonuclease-digested, adaptor-ligated tumor DNA, followed by high-throughput sequencing. However, sequence coverage cannot be used to infer tumor clonality accurately owing to PCR biases as a result of the variable distribution of restriction enzyme sites in the genome. An alternative approach, called shear-splink, was developed by Jonkers and colleagues that fragments DNA by acoustic shearing, mitigating this bias [52]. In addition, as DNA is fragmented at random, each fragment harbors a potentially unique stretch of DNA that can serve as a molecular barcode. Quantification of these barcodes permits estimation of transposon clonality within a heterogeneous sample. Rad and colleagues used a similar approach, termed quantitative insertion site sequencing (QIseq), to illustrate the marked genetic complexity of pancreatic tumors [32]. Although these approaches can estimate transposon clonality, they cannot distinguish between transposon heterogeneity arising during tumor evolution in a monoclonal sample and multiple distinct insertions in a polyclonal tumor population.

Identifying genes involved in metastasis

In addition to identifying genes involved in tumor initiation and progression, TMIM has been performed to discover genes that promote tumor dissemination. Largaespada and colleagues expressed the SB system in p53-deficient mouse osteoblasts and identified candidate genes involved in metastasis by comparing transposon insertions from osteosarcoma metastases with those found in primary tumors [53]. Approximately one-third of CIS-associated genes found in metastases were evident in primary tumors. Furthermore, from this analysis, five candidate oncogenes and 38 tumor suppressors were identified, including nine genes that have been implicated previously in cancer metastasis. To study further the evolutionary relationships between metastases and parental ancestors, the authors conducted parsimony analysis of tumors using transposon integration sites as molecular footprints. Osteosarcoma metastases were found to be highly clonal but appeared to show different patterns of evolution from the primary tumor.

Taylor and colleagues performed a TMIM screen aimed at identifying genes affecting dissemination of medulloblastoma in Ptch1 +/− heterozygous null or mutant Tp53 mouse backgrounds [54]. Interestingly, the authors found that both transposon-driven mouse and human metastatic medulloblastoma are clonal but divergent from the primary tumor, suggesting that only a rare subclone in the primary tumor is able to metastasize. Four of the identified candidate genes were validated as drivers of medulloblastoma dissemination by retroviral delivery of these candidates to the cerebellum in combination with overexpression of the Ptch1 ligand sonic hedgehog (Shh) [55]. These studies demonstrated the utility of TMIM screens to discover drivers of metastatic spread, and further studies will identify candidate metastasis genes in certain genetic backgrounds and tumor types. Some mouse cancer models might not be suitable for identification of metastasis genes by TMIM because the mice have to be sacrificed before the formation of macroscopic metastases owing to the primary tumor size. However, surgical removal of the primary tumor to allow more time for metastasis growth or transplantation of primary tumor cells into syngeneic wild-type mice could circumvent this issue. Nonetheless, these reports illustrate how TMIM can be employed to query the clonal relationship of a primary tumor and its metastases, complementing the use of transposons to identify genes involved in tumor progression.

Identifying alterations in cancer pathways

Apart from identifying genes promoting tumor progression, TMIM screens have been used to define the most prominent signaling pathways deregulated in tumors. Using the TAPDANCE tool, Largaespada and colleagues performed a pathway-centric analysis of alterations in Tp53-mutant, EGFR-driven peripheral nerve sheath tumors to identify roles for the phosphoinositide 3-kinase (PI3K)-AKT-mTOR, mitogen-activated protein kinase (MAPK) and Wnt/β-catenin pathways in the development of this tumor type [56]. Novel pathways have also been revealed in melanoma driven by oncogenic B-Raf V600E . Xu and colleagues identified a network involving Magi2 with a PB screen at low transposon copy number and also found insertions in Map3k1 and Map3k2 that resulted in ERK activation [57]. However, these insertions occurred in melanomas that had not recombined the conditional oncogenic B-Raf V600E allele. Although not examined, this suggests that aberrant MAP3K1/MAP3K2 activation could represent another means to activate the MAPK pathway in human melanoma besides the common BRAF and NRAS mutations. The melanoma SB screen performed by Jenkins, Copeland and colleagues identified numerous candidate cancer genes, and pathway analysis found significant enrichment of CIS-associated genes in many cancer-related signaling pathways, including Wnt/β-catenin, TGF-β, PI3K and MAPK signaling, as well as in many biological processes [38]. Recently, it was shown that, by integrating SB TMIM in mice and mutation analysis of human cancer genomes, loss of function of the transcription factor CUX1 drives myeloid malignancy and other cancer types [20]. It was demonstrated that CUX1 antagonizes the PI3K–AKT signaling pathway by regulating transcription of the PI3K inhibitor PIK3IP1. Finally, a SB medulloblastoma screen in Ptch1 +/− mice identified candidate cancer genes and associated protein networks capable of distinguishing the molecular subgroups of human medulloblastoma, demonstrating the power of transposon screens to recapitulate the genetic changes in human cancer [58].

These studies suggest that pathway and network analyses can provide insight into mechanisms of human disease and might predict survival and treatment outcomes. Thus, TMIM is a powerful approach to unravel the functional association of altered signaling pathways or cell-biological processes with cancer development. Conventional sequencing efforts can fail to identify such associations because the mutation rate of individual genes regulating these pathways or processes is not above the background mutation rate. Moreover, although TMIM cannot recapitulate activating mutations of proto-oncogenes, pathway analyses of TMIM datasets can reveal the crucial functions downstream of oncogenes that are commonly mutated in human cancer.

Identification of novel mechanisms of gene deregulation

In cancer cells, loss of mRNA and protein expression can occur without any obvious genetic alteration in corresponding protein-coding regions. Notably, recent TMIM studies have identified novel non-coding regulatory regions and other mechanisms of gene deregulation that promote tumorigenesis. For example, a PB screen identified recurrent transposon insertions in a 200-kb noncoding region (Ncruc) upstream of the Cdkn2a gene [32], which encodes the tumor suppressors p16Ink4a and p19Arf and is frequently inactivated by prototypic gene-body insertions in both SB and PB pancreatic cancer screens [24, 32]. Transposon insertions in or genomic loss of the Ncruc region were associated with reduced expression levels of Cdkn2a in cis, demonstrating the power of PB insertional mutagenesis screens to identify non-coding DNA regions or genes with crucial roles in tumorigenesis.

Although target-site preferences suggest that PB-based TMIM screens might be more useful to identify regulatory elements compared with SB transposons (Fig. 3), SB-mediated screens have also been fruitful in identifying atypical mechanisms of gene deregulation in cancer. For example, Dupuy and colleagues performed a SB-mediated hepatocellular carcinoma (HCC) screen and found frequent insertions in the complex imprinted Dlk1-Dio3 locus. A domesticated retrotransposon, Rtl1, located in this locus was shown to be overexpressed in all tumors with Dlk1-Dio3 insertions [59]. Furthermore, ectopic overexpression of Rtl1 in mouse livers induced HCC, validating Rtl1 as a novel cancer driver. Examination of human liver tissue showed that Rtl1 is transcriptionally inactive in normal liver but can be reactivated in human HCC, supporting a role for Rtl1 in human HCC development.

In a SB-mediated TMIM screen aimed at identifying genes that co-operate with oncogenic B-Raf in melanoma development, a significant enrichment of genes was discovered among the CISs that encode mRNAs with the ability to regulate the expression of the tumor suppressor Pten [48]. These so-called competitive endogenous RNAs control Pten levels as microRNA decoys, in a protein-coding-independent fashion. While these CIS-associated genes are classical protein-coding genes, our analysis highlighted a non-coding function of their mRNAs. Only 2 % of the mammalian genome encodes protein-coding genes; however, the non-coding portion of the genome, both transcribed (e.g., microRNAs, long non-coding RNAs) and non-transcribed (e.g., enhancers), plays crucial roles in physiology and pathology. TMIM screens have barely scratched the surface of the non-coding space, and re-analyzing existing SB and PB mutagenesis data might reveal additional non-coding insertion hotspots.

Identifying mechanisms of resistance to therapy

TMIM has been useful in identifying genes that mediate therapeutic drug resistance both in vitro and in vivo. Schmidt and colleagues conducted a PB screen in four different human cell lines derived from neuroblastoma, breast and cervical cancer to identify genes whose overexpression mediates resistance to paclitaxel [60]. Interestingly, while the authors identified multiple CISs in the four cell lines, the only CIS that was common to all four cell lines was the ABCB1 gene [60], which encodes an ABC-transporter associated with multi-drug resistance [61]. This suggests the existence of both cancer-type-specific and common mechanisms of drug resistance. In addition, Xu and colleagues performed a PB screen in melanoma cells and identified BRAF and CRAF as mediators of resistance to the BRAF inhibitor vemurafenib [62], recapitulating previous observations in human melanoma patients and cell lines treated with vemurafenib [6365].

In diploid cells, biallelic inactivating transposon insertions that completely abrogate gene expression are rare compared with monoallelic events, thus hampering the identification of genes that promote drug resistance only upon complete loss of expression. To tackle this issue, Ashworth and colleagues [66] took advantage of a haploid mouse embryonic stem (ES) cell system to screen for mediators of olaparib toxicity, in which inactivating transposon insertion can result in complete loss of gene expression. The authors identified the poly [ADP-ribose] polymerase 1 gene Parp1 as a mediator of olaparib toxicity, and their results suggested that loss of Parp1 could result in olaparib resistance in patients [66]. In another mouse ES cell screen, Jonkers and colleagues identified loss of the gene 53bp1 as a mediator of survival and DNA-damage responses in Brca1-null cells [67]. Reduced 53BP1 expression was associated with basal-like, triple-negative, and BRCA1/2-mutant breast cancer in humans, suggesting that downregulation of 53BP1 might be an important survival factor in such tumors, particularly during chemotherapy-induced DNA damage. These studies demonstrate the utility of TMIM to identify mediators of resistance in human cancer cell lines as well as ES cells.

Drug resistance in patients develops in the context of a supporting microenvironment and, thus, in vitro approaches might be limited in their ability to identify resistance genes. To avoid this shortcoming of in vitro drug-resistance screens, a SB screen in a B-Raf V600E-driven mouse model of melanoma was performed. This identified transposon insertion sites in treatment-naïve tumors as well as melanomas treated with the vemurafenib progenitor compound PLX4720 [49]. Insertions in several known mediators of resistance were enriched in the PLX4720-treated tumors, validating this approach for resistance gene discovery. An ERAS-AKT-BAD signaling axis was validated as a mediator of drug resistance, which mimics the paracrine mechanism of stromal hepatocyte growth factor-mediated resistance [68, 69]. Curiously, many of the genes that have been previously identified in cell lines as promoters of resistance through reactivation of MAPK signaling were not identified in this in vivo study. A possible explanation is that such mutations are preexisting in patients only in a minor tumor subclone that no longer relies on oncogenic BRAF signaling. Conversely, transposon mobilization was induced concomitantly with the initiating B-Raf mutation in the resistance TMIM. In these tumor cells, transposon insertions that would otherwise result in MAPK activation might be negatively selected owing to functional redundancy with oncogenic B-Raf. Thus, additional insight might be gained from studies in which transposon mobilization is induced at the time of drug treatment.

Novel approaches of employing transposon mutagenesis

In vivo transposon mutagenesis requires up to four transgenic alleles to accelerate tumorigenesis in a tissue-specific manner in a sensitizing background. Generating and maintaining compound mutant mouse strains is time consuming and costly, prompting alternative ways of utilizing the transposon systems. Molyneux and colleagues transduced immortalized primary human bone mesenchymal cells with SB and a lentivirus harboring the elements of a SB transposon, and, when injected into mice, the transplanted cells produced myxofibrosarcomas [70]. For human candidate cancer gene discovery, both the insertions of the parental lentivirus as well as the remobilized transposons were mapped. In another study, neural stem cells were derived from transgenic mice harboring the SB system and a Nestin-Cre allele [71]. Following in vitro differentiation, the neural stem cells were immortalized through SB mutagenesis and the resulting immortalized astroglial-like cells were injected into SCID mice to identify genes that drive glioblastoma formation. CIS mapping of immortalized cell lines and tumors identified partially overlapping CISs, suggesting differential roles of the identified genes during immortalization and tumorigenesis. In vitro delivery of the transposon system components followed by orthotopic or subcutaneous transplantation thus represents another means for in vivo selection and identification of candidate cancer genes.

The SB transposon system has also been used as a reverse-genetics tool to validate candidate cancer genes. Futreal and colleagues created transposons with both SB and PB terminal repeats that also harbored IRES-cDNA cassettes [72], such that the cDNA cargo was expressed only when transposon insertion occurred in transcribed genes. Using these transposons, the authors tested kinases with point mutations encoding putative gain-of-function oncogenic alleles. Mice were generated carrying multiple transposons with different cDNA cargos and crossed to SB transgenic mice, leading to tumorigenesis by in vivo selection of the kinase mutants with the highest oncogenic potential in somatic cells. This report elegantly displays how the transposon system can be utilized to discern the relative oncogenic properties of several candidate genes simultaneously in all or selected organs.

To extend the utility of TMIM to another model system, transgenic rats carrying the components of the SB or PB system have been created [73]. The transposons carried both SB and PB terminal repeats as well as a tyrosinase expression cassette, permitting coat-color-based phenotyping for transposon zygosity and genomic position effects on tyrosinase expression in albino rat backgrounds. In the future, it will be interesting to determine the overlap in cancer genes identified by TMIM screens in mouse and rat and their relevance to human cancer.

Comparison with other technologies

Other methods of forward-genetic screens for the promotion of tumorigenesis and related phenotypes in vivo include the use of cDNA or short hairpin RNA (shRNA) libraries for gain-of-function or loss-of-function screens, respectively. In addition, the CRISPR/Cas9 system, a novel powerful tool for genome editing [74, 75], can be employed for gain-of-function and loss-of-function screens. The conventional CRISPR/Cas9 system uses a short guide RNA (sgRNA) to direct the Cas9 DNA endonuclease to a complementary DNA target, resulting in double-strand DNA cleavage, which can result in loss-of-function frameshift indels within exons when DNA breaks are repaired by error-prone non-homologous end-joining mechanisms. Alternative Cas9 enzymes, lacking endonuclease activity, have been engineered that promote transcriptional repression [76, 77] or activation [7880] of target genes when coexpressed with targeting sgRNAs. These approaches have several advantages and disadvantages compared with TMIM, and the different approaches thus provide complementary technologies for cancer gene discovery (Table 2).
Table 2

Comparison of genome-wide TMIM, CRISPR/Cas9 and shRNA/cDNA expression technologies

Feature

TMIM

CRISPR/Cas9

shRNA and cDNA libraries

Screen set-up

Two-component system (transposase and transposon)

Comprehensive delivery of libraries can be technically challenging

Comprehensive delivery of libraries can be technically challenging

Possible types of mutation

Activating and disruptive mutations possible owing to transposon insertion or remobilization

Disruptive mutations (knockout libraries)

Knockdown or overexpression — potentially reversible

Chromosomal deletions and translocations are possible (knockout libraries)

Chromosomal deletions and rearrangements are rare

Mutations either can (transcription repression [76, 77] and activation [79, 80] libraries) or cannot be reversed (knockout libraries)

Mutations can be reversed following transposon remobilization

Mutagenesis efficiency

Biallelic gene inactivation rare in diploid cells

Biallelic mutation achievable with knockout libraries [88] and 90–99 % knockdown efficiency achievable with repression libraries [89]

≥70 % gene knockdown with validated shRNA clones [8284]

>2 standard deviation overexpression by 90 % of cDNA expression vectors [90]

Undesired and off-target effects

Local hopping effects [31], passenger insertions

Minimal off-target effects [78, 81, 88, 89, 95, 96]

Off-target effects can be significant [9799]

Viral-associated insertional mutagenesis possible [93, 94]

Viral-associated insertional mutagenesis possible [93, 94]

Genome coverage

Whole genome in principle, but affected by integration-site preferences, local hopping and chromatin accessibility

Dictated by library design

Dictated by library design

Knockout libraries

~8000 human, ~15,000 mouse genes (NKI shRNA library) [85]; >20,000 human and mouse genes (TRC shRNA library) [82, 83]; ~60,000 human and mouse genes (Hannon–Elledge shRNA library) [84]; >17,000 human genes (cDNA expression library) [90]

GeCKOv2 [86]: ~20,000 genes, 1000–2000 miRNAs for human and mouse. Koike-Yusa et al . [87]: ~20,000 mouse genes. Wang et al . [88]: ~7000 human genes

CRISPR-activation libraries

CRISPRa [89]: ~16,000 human genes. SAM [78]: all human RefSeq genes

CRISPR-inhibition library [89]

~16,000 human genes

NKI Netherlands Cancer Institute, shRNA short hairpin RNA, TMIM transposon-mediated insertional mutagenesis, TRC The RNAi Consortium

One major pitfall of shRNA, cDNA and CRISPR/Cas9 screens is that these approaches allow for identification of either tumor suppressors or oncogenes, but not both at the same time [78, 81]. By contrast, TMIM has the ability to detect both tumor suppressors and oncogenes simultaneously owing to the genetic elements within the transposons that intercept and promote transcription (see discussion above). Comprehensive shRNA [8285], sgRNA [78, 8689] and cDNA [90] libraries have been created for forward-genetic screens. However, the task of delivery of these libraries to the cell type of interest for in vivo screens is not trivial. Usually, libraries are delivered in vitro, followed by orthotopic or subcutaneous transplantation of the library-infected cells [91]. While this can be a viable approach in many cases, it might not always accurately recapitulate tumor progression in its natural environment [92] and might therefore select for false-positive candidate cancer genes. In addition, delivery of libraries with lentiviruses can cause tumor-promoting insertional mutagenesis [93, 94] that remains undetected unless these insertion sites are mapped in conjunction with shRNA/sgRNA/cDNA identification. TMIM does not face the issue of library delivery as the transposons are already included in the genome of transgenic mouse strains, and transposon mobilization is readily achieved in virtually any cell type. However, owing to the local hopping effect [34] observed in TMIM, the donor chromosome containing the parental transposon concatemer has to be excluded from the analysis. Thus, to probe all chromosomes by TMIM, more than one transposon mouse strain has to be used [31].

Another bias of shRNA and CRISPR/Cas9 screens is that shRNAs and sgRNAs are designed to target specific sequences. Thus, these screens are inherently biased, although whether this impacts candidate cancer gene discovery remains to be determined. Moreover, while shRNA and sgRNA design algorithms generate sequences with minimal predicted off-target effects, such effects cannot be excluded experimentally [78, 81, 88, 89, 9599]. To control for off-target effects by shRNA and sgRNAs, bona fide hits need to be identified by more than one shRNA or sgRNA. In TMIM, the number of transposon insertion sites in a predefined genomic window determines the statistical significance of a CIS [3941, 43]. However, owing to the continued hopping of unselected transposons and the consequential heterogeneity of tumors with hundreds of passenger insertions, accurate CIS calling remains challenging. Not only are bona fide candidate cancer genes excluded and false positives included following the statistical analysis, CISs might also affect more than one gene. Thus, proper functional validation of any candidates identified by these screening methods is an absolute requirement.

Current sgRNA, shRNA and cDNA libraries are fairly comprehensive, but they do not yet match the ability of TMIM to query virtually the entire genome. However, it is difficult to identify small genetic entities such as microRNAs and enhancers because the likelihood of transposon insertions in the precise locations that would affect their expression or activity is lower. With the CRISPR/Cas9 system, these genes and genetic elements can be targeted and inactivated directly. Indeed, commercially available CRISPR/Cas9 libraries already contain sgRNAs targeting microRNAs [86], and libraries targeting other genetic elements will surely be developed in the near future. Another consideration is that complete target repression is not achieved by either shRNA or TMIM. shRNAs vary drastically in their ability to repress target mRNAs, and transposon insertion is typically observed in only one allele. These technologies are thus biased towards the identification of candidate cancer genes whose incomplete repression promotes tumorigenesis, such as haploinsufficient tumor suppressors or tumor suppressors that readily undergo loss-of-heterozygosity. Conversely, the CRISPR/Cas9 system readily generates biallelic deletions [88, 100] and is therefore able to discover genes that will yield phenotypes only after homozygous loss. Thus, genome coverage and gene dosage are important considerations when choosing a screening system.

Finally, insertions and deletions introduced by the CRISPR/Cas9 system occur through error-prone non-homologous end joining [75]. It is therefore possible that in-frame indels are generated that do not abrogate protein expression [87] but alter proper biological function. This, in turn, could yield different phenotypes compared with those arising from the absence of the protein and could affect the outcome and/or interpretation of the screen. In-frame indels will be selected for if they provide a biological advantage, and are therefore distinguishable from indels that result in frameshifts. Such in-frame indels might reveal interesting aspects about the biology of certain proteins; however, their relevance to human disease will have to be determined on a case-by-case basis. In summary, the different technologies for forward-genetic screening have various pros and cons that need to be considered when designing a screen experiment.

Concluding remarks

The past few years have brought remarkable advances in the field of transposon-mediated insertional mutagenesis. First, technological developments have enabled investigators to identify cancer genes in an ever-expanding array of cell types and at various stages of tumor evolution. Second, improvements to bioinformatics and statistical methods ensure the identification of crucial cancer genes and pathways and the exclusion of false-positive hits. Third, the vastly increased availability of genomic and mutational data from human cancer specimens allows for the comparison of such data with TMIM results, thereby distinguishing genes with relevance to human cancer that were identified in mice. TMIM remains relevant in the age of CRISPR/Cas9 screens, and together these technologies form a powerful and complementary toolbox to query the genome for the genetic causes of cancer.

Notes

Abbreviations

CIS: 

Common insertion site

CRC: 

Colorectal cancer

CRISPR/Cas9: 

Clustered regularly interspaced short palindromic repeats/CRISPR-associated protein 9

ES: 

Embryonic stem

gCIS: 

Gene-centric common insertion site

MAPK: 

Mitogen-activated protein kinase

MSCV: 

Murine stem cell virus

pA: 

Polyadenylation signal

PB: 

piggyBac

PGK: 

Phosphoglycerate kinase 1

PI3K: 

Phosphoinositide 3-kinase

PRIM: 

Poisson regression insertion model

QIseq: 

Quantitative insertion site sequencing

SA: 

Splice acceptor

SB: 

Sleeping Beauty

SD: 

Splice donor

sgRNA: 

Short guide RNA

shRNA: 

Short hairpin RNA

TSS: 

Transcriptional start site

Declarations

Acknowledgements

We apologize to investigators whose work we were unable to discuss owing to space limitations. This work was supported by a Pancreatic Cancer Action Network — AACR Pathway to Leadership grant (GMD), and an American Cancer Society fellowship (FAK). CCW is a Wellcome Trust Intermediate Clinical Fellow, and DJA is supported by Cancer Research UK and the Wellcome Trust.

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Authors’ Affiliations

(1)
Meyer Cancer Center, Weill Cornell Medical College, New York, NY 10021, USA
(2)
Experimental Cancer Genetics, Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, CB10 1HH, UK
(3)
Department of Haematology, University of Cambridge, Hills Road, Cambridge, CB2 0XY, UK

References

  1. Vogelstein B, Papadopoulos N, Velculescu VE, Zhou S, Diaz LA, Kinzler KW. Cancer genome landscapes. Science. 2013;339:1546–58.PubMed CentralPubMedView ArticleGoogle Scholar
  2. Nusse R, Varmus HE. Many tumors induced by the mouse mammary tumor virus contain a provirus integrated in the same region of the host genome. Cell. 1982;31:99–109.PubMedView ArticleGoogle Scholar
  3. van Lohuizen M, Verbeek S, Scheijen B, Wientjens E, van der Gulden H, Berns A. Identification of cooperating oncogenes in E mu-myc transgenic mice by provirus tagging. Cell. 1991;65:737–52.PubMedView ArticleGoogle Scholar
  4. Li Y, Hively WP, Varmus HE. Use of MMTV-Wnt-1 transgenic mice for studying the genetic basis of breast cancer. Oncogene. 2000;19:1002–9.PubMedView ArticleGoogle Scholar
  5. Uren AG, Kool J, Berns A, van Lohuizen M. Retroviral insertional mutagenesis: past, present and future. Oncogene. 2005;24:7656–72.PubMedView ArticleGoogle Scholar
  6. Ivics Z, Hackett PB, Plasterk RH, Izsvák Z. Molecular reconstruction of Sleeping Beauty, a Tc1-like transposon from fish, and its transposition in human cells. Cell. 1997;91:501–10.PubMedView ArticleGoogle Scholar
  7. Burns KH, Boeke JD. Human transposon tectonics. Cell. 2012;149:740–52.PubMed CentralPubMedView ArticleGoogle Scholar
  8. Mills RE, Bennett EA, Iskow RC, Devine SE. Which transposable elements are active in the human genome? Trends Genet. 2007;23:183–91.PubMedView ArticleGoogle Scholar
  9. Copeland NG, Jenkins NA. Harnessing transposons for cancer gene discovery. Nat Rev Cancer. 2010;10:696–706.PubMedView ArticleGoogle Scholar
  10. Mann MB, Jenkins NA, Copeland NG, Mann KM. Sleeping Beauty mutagenesis: exploiting forward genetic screens for cancer gene discovery. Curr Opin Genet Dev. 2014;24:16–22.PubMedView ArticleGoogle Scholar
  11. Landrette SF, Xu T. Somatic genetics empowers the mouse for modeling and interrogating developmental and disease processes. PLoS Genet. 2011;7:e1002110.PubMed CentralPubMedView ArticleGoogle Scholar
  12. Collier LS, Carlson CM, Ravimohan S, Dupuy AJ, Largaespada DA. Cancer gene discovery in solid tumours using transposon-based somatic mutagenesis in the mouse. Nature. 2005;436:272–6.PubMedView ArticleGoogle Scholar
  13. Dupuy AJ, Akagi K, Largaespada DA, Copeland NG, Jenkins NA. Mammalian mutagenesis using a highly mobile somatic Sleeping Beauty transposon system. Nature. 2005;436:221–6.PubMedView ArticleGoogle Scholar
  14. Dupuy AJ, Rogers LM, Kim J, Nannapaneni K, Starr TK, Liu P, et al. A modified sleeping beauty transposon system that can be used to model a wide variety of human cancers in mice. Cancer Res. 2009;69:8150–6.PubMed CentralPubMedView ArticleGoogle Scholar
  15. Collier LS, Adams DJ, Hackett CS, Bendzick LE, Akagi K, Davies MN, et al. Whole-body sleeping beauty mutagenesis can cause penetrant leukemia/lymphoma and rare high-grade glioma without associated embryonic lethality. Cancer Res. 2009;69:8429–37.PubMed CentralPubMedView ArticleGoogle Scholar
  16. Bender AM, Collier LS, Rodriguez FJ, Tieu C, Larson JD, Halder C, et al. Sleeping beauty-mediated somatic mutagenesis implicates CSF1 in the formation of high-grade astrocytomas. Cancer Res. 2010;70:3557–65.PubMed CentralPubMedView ArticleGoogle Scholar
  17. Been RA, Linden MA, Hager CJ, DeCoursin KJ, Abrahante JE, Landman SR, et al. Genetic signature of histiocytic sarcoma revealed by a sleeping beauty transposon genetic screen in mice. PLoS One. 2014;9:e97280.PubMed CentralPubMedView ArticleGoogle Scholar
  18. Rahrmann EP, Collier LS, Knutson TP, Doyal ME, Kuslak SL, Green LE, et al. Identification of PDE4D as a proliferation promoting factor in prostate cancer using a Sleeping Beauty transposon-based somatic mutagenesis screen. Cancer Res. 2009;69:4388–97.PubMed CentralPubMedView ArticleGoogle Scholar
  19. Starr TK, Allaei R, Silverstein KAT, Staggs RA, Sarver AL, Bergemann TL, et al. A transposon-based genetic screen in mice identifies genes altered in colorectal cancer. Science. 2009;323:1747–50.PubMed CentralPubMedView ArticleGoogle Scholar
  20. Wong CC, Martincorena I, Rust AG, Rashid M, Alifrangis C, Alexandrov LB, et al. Inactivating CUX1 mutations promote tumorigenesis. Nat Genet. 2014;46:33–8.PubMedView ArticleGoogle Scholar
  21. Keng VW, Villanueva A, Chiang DY, Dupuy AJ, Ryan BJ, Matise I, et al. A conditional transposon-based insertional mutagenesis screen for genes associated with mouse hepatocellular carcinoma. Nat Biotechnol. 2009;27:264–74.PubMed CentralPubMedView ArticleGoogle Scholar
  22. Starr TK, Scott PM, Marsh BM, Zhao L, Than BLN, O’Sullivan MG, et al. A Sleeping Beauty transposon-mediated screen identifies murine susceptibility genes for adenomatous polyposis coli (Apc)-dependent intestinal tumorigenesis. Proc Natl Acad Sci U S A. 2011;108:5765–70.PubMed CentralPubMedView ArticleGoogle Scholar
  23. March HN, Rust AG, Wright NA, ten Hoeve J, de Ridder J, Eldridge M, et al. Insertional mutagenesis identifies multiple networks of cooperating genes driving intestinal tumorigenesis. Nat Genet. 2011;43:1202–9.PubMed CentralPubMedView ArticleGoogle Scholar
  24. Pérez-Mancera PA, Rust AG, van der Weyden L, Kristiansen G, Li A, Sarver AL, et al. The deubiquitinase USP9X suppresses pancreatic ductal adenocarcinoma. Nature. 2012;486:266–70.PubMed CentralPubMedGoogle Scholar
  25. Mann KM, Ward JM, Yew CCK, Kovochich A, Dawson DW, Black MA, et al. Sleeping Beauty mutagenesis reveals cooperating mutations and pathways in pancreatic adenocarcinoma. Proc Natl Acad Sci U S A. 2012;109:5934–41.PubMed CentralPubMedView ArticleGoogle Scholar
  26. O’Donnell KA, Keng VW, York B, Reineke EL, Seo D, Fan D, et al. A Sleeping Beauty mutagenesis screen reveals a tumor suppressor role for Ncoa2/Src-2 in liver cancer. Proc Natl Acad Sci U S A. 2012;109:E1377–86.PubMed CentralPubMedView ArticleGoogle Scholar
  27. Lastowska M, Al-Afghani H, Al-Balool HH, Sheth H, Mercer E, Coxhead JM, et al. Identification of a neuronal transcription factor network involved in medulloblastoma development. Acta Neuropathol Commun. 2013;1:35.PubMed CentralPubMedView ArticleGoogle Scholar
  28. Zanesi N, Balatti V, Riordan J, Burch A, Rizzotto L, Palamarchuk A, et al. A Sleeping Beauty screen reveals NF-kB activation in CLL mouse model. Blood. 2013;121:4355–4358.PubMed CentralPubMedView ArticleGoogle Scholar
  29. Vassiliou GS, Cooper JL, Rad R, Li J, Rice S, Uren A, et al. Mutant nucleophosmin and cooperating pathways drive leukemia initiation and progression in mice. Nat Genet. 2011;43:470–5.PubMed CentralPubMedView ArticleGoogle Scholar
  30. Dorr C, Janik C, Weg M, Been RA, Bader J, Kang R, et al. Transposon mutagenesis screen identifies potential lung cancer drivers and CUL3 as a tumor suppressor. Mol Cancer Res. 2015;13:1238–47.PubMedView ArticleGoogle Scholar
  31. Rad R, Rad L, Wang W, Cadinanos J, Vassiliou G, Rice S, et al. PiggyBac transposon mutagenesis: a tool for cancer gene discovery in mice. Science. 2010;330:1104–7.PubMed CentralPubMedView ArticleGoogle Scholar
  32. Rad R, Rad L, Wang W, Strong A, Ponstingl H, Bronner IF, et al. A conditional piggyBac transposition system for genetic screening in mice identifies oncogenic networks in pancreatic cancer. Nat Genet. 2014;47:47–56.PubMedView ArticleGoogle Scholar
  33. Liang Q, Kong J, Stalker J, Bradley A. Chromosomal mobilization and reintegration of Sleeping Beauty and PiggyBac transposons. Genesis. 2009;47:404–8.PubMedView ArticleGoogle Scholar
  34. de Jong J, Akhtar W, Badhai J, Rust AG, Rad R, Hilkens J, et al. Chromatin landscapes of retroviral and transposon integration profiles. PLoS Genet. 2014;10:e1004250.PubMed CentralPubMedView ArticleGoogle Scholar
  35. Luo G, Ivics Z, Izsvák Z, Bradley A. Chromosomal transposition of a Tc1/mariner-like element in mouse embryonic stem cells. Proc Natl Acad Sci U S A. 1998;95:10769–73.PubMed CentralPubMedView ArticleGoogle Scholar
  36. Geurts AM, Collier LS, Geurts JL, Oseth LL, Bell ML, Mu D, et al. Gene mutations and genomic rearrangements in the mouse as a result of transposon mobilization from chromosomal concatemers. PLoS Genet. 2006;2:e156.PubMed CentralPubMedView ArticleGoogle Scholar
  37. Riordan JD, Drury LJ, Smith RP, Brett BT, Rogers LM, Scheetz TE, et al. Sequencing methods and datasets to improve functional interpretation of sleeping beauty mutagenesis screens. BMC Genomics. 2014;15:1150.PubMed CentralPubMedView ArticleGoogle Scholar
  38. Mann MB, Black MA, Jones DJ, Ward JM, Yew CCK, Newberg JY, et al. Transposon mutagenesis identifies genetic drivers of BrafV600E melanoma. Nat Genet. 2015;47:486–95.PubMedView ArticleGoogle Scholar
  39. Mikkers H, Allen J, Knipscheer P, Romeijn L, Hart A, Vink E, et al. High-throughput retroviral tagging to identify components of specific signaling pathways in cancer. Nat Genet. 2002;32:153–9.PubMedView ArticleGoogle Scholar
  40. Suzuki T, Shen H, Akagi K, Morse HC, Malley JD, Naiman DQ, et al. New genes involved in cancer identified by retroviral tagging. Nat Genet. 2002;32:166–74.PubMedView ArticleGoogle Scholar
  41. de Ridder J, Uren A, Kool J, Reinders M, Wessels L. Detecting statistically significant common insertion sites in retroviral insertional mutagenesis screens. PLoS Comput Biol. 2006;2:e166.PubMed CentralPubMedView ArticleGoogle Scholar
  42. Brett BT, Berquam-Vrieze KE, Nannapaneni K, Huang J, Scheetz TE, Dupuy AJ. Novel molecular and computational methods improve the accuracy of insertion site analysis in Sleeping Beauty-induced tumors. PLoS One. 2011;6:e24668.PubMed CentralPubMedView ArticleGoogle Scholar
  43. Sarver AL, Erdman J, Starr T, Largaespada DA, Silverstein KAT. TAPDANCE: An automated tool to identify and annotate transposon insertion CISs and associations between CISs from next generation sequence data. BMC Bioinformatics. 2012;13:154.PubMed CentralPubMedView ArticleGoogle Scholar
  44. de Ridder J, Kool J, Uren A, Bot J, Wessels L, Reinders M. Co-occurrence analysis of insertional mutagenesis data reveals cooperating oncogenes. Bioinformatics. 2007;23:i133–41.PubMedView ArticleGoogle Scholar
  45. Bergemann TL, Starr TK, Yu H, Steinbach M, Erdmann J, Chen Y, et al. New methods for finding common insertion sites and co-occurring common insertion sites in transposon- and virus-based genetic screens. Nucleic Acids Res. 2012;40:3822–33.PubMed CentralPubMedView ArticleGoogle Scholar
  46. Tuveson DA, Shaw AT, Willis NA, Silver DP, Jackson EL, Chang S, et al. Endogenous oncogenic K-ras(G12D) stimulates proliferation and widespread neoplastic and developmental defects. Cancer Cell. 2004;5:375–87.PubMedView ArticleGoogle Scholar
  47. Olive KP, Tuveson DA, Ruhe ZC, Yin B, Willis NA, Bronson RT, et al. Mutant p53 gain of function in two mouse models of Li-Fraumeni syndrome. Cell. 2004;119:847–60.PubMedView ArticleGoogle Scholar
  48. Karreth FA, Tay Y, Perna D, Ala U, Tan SM, Rust AG, et al. In vivo identification of tumor- suppressive PTEN ceRNAs in an oncogenic BRAF-induced mouse model of melanoma. Cell. 2011;147:382–95.PubMed CentralPubMedView ArticleGoogle Scholar
  49. Perna D, Karreth FA, Rust AG, Pérez-Mancera PA, Rashid M, Iorio F, et al. BRAF inhibitor resistance mediated by the AKT pathway in an oncogenic BRAF mouse melanoma model. Proc Natl Acad Sci U S A. 2015;112:E536–45.PubMed CentralPubMedView ArticleGoogle Scholar
  50. Tang JZ, Carmichael CL, Shi W, Metcalf D, Ng AP, Hyland CD, et al. Transposon mutagenesis reveals cooperation of ETS family transcription factors with signaling pathways in erythro-megakaryocytic leukemia. Proc Natl Acad Sci U S A. 2013;110:6091–6.PubMed CentralPubMedView ArticleGoogle Scholar
  51. Takeda H, Wei Z, Koso H, Rust AG, Yew CCK, Mann MB, et al. Transposon mutagenesis identifies genes and evolutionary forces driving gastrointestinal tract tumor progression. Nat Genet. 2015;47:142–50.PubMedView ArticleGoogle Scholar
  52. Koudijs MJ, Klijn C, van der Weyden L, Kool J, ten Hoeve J, Sie D, et al. High-throughput semiquantitative analysis of insertional mutations in heterogeneous tumors. Genome Res. 2011;21:2181–9.PubMed CentralPubMedView ArticleGoogle Scholar
  53. Moriarity BS, Otto GM, Rahrmann EP, Rathe SK, Wolf NK, Weg MT, et al. A Sleeping Beauty forward genetic screen identifies new genes and pathways driving osteosarcoma development and metastasis. Nat Genet. 2015;47:615–24.PubMedView ArticleGoogle Scholar
  54. Wu X, Northcott PA, Dubuc A, Dupuy AJ, Shih DJH, Witt H, et al. Clonal selection drives genetic divergence of metastatic medulloblastoma. Nature. 2012;482:529–33.PubMed CentralPubMedView ArticleGoogle Scholar
  55. Mumert M, Dubuc A, Wu X, Northcott PA, Chin SS, Pedone CA, et al. Functional genomics identifies drivers of medulloblastoma dissemination. Cancer Res. 2012;72:4944–53.PubMed CentralPubMedView ArticleGoogle Scholar
  56. Rahrmann EP, Watson AL, Keng VW, Choi K, Moriarity BS, Beckmann DA, et al. Forward genetic screen for malignant peripheral nerve sheath tumor formation identifies new genes and pathways driving tumorigenesis. Nat Genet. 2013;45:756–66.PubMed CentralPubMedView ArticleGoogle Scholar
  57. Ni TK, Landrette SF, Bjornson RD, Bosenberg MW, Xu T. Low-copy piggyBac transposon mutagenesis in mice identifies genes driving melanoma. Proc Natl Acad Sci U S A. 2013;110:E3640–9.PubMed CentralPubMedView ArticleGoogle Scholar
  58. Genovesi LA, Ng CG, Davis MJ, Remke M, Taylor MD, Adams DJ, et al. Sleeping Beauty mutagenesis in a mouse medulloblastoma model defines networks that discriminate between human molecular subgroups. Proc Natl Acad Sci U S A. 2013;110:E4325–34.PubMed CentralPubMedView ArticleGoogle Scholar
  59. Riordan JD, Keng VW, Tschida BR, Scheetz TE, Bell JB, Podetz-Pedersen KM, et al. Identification of rtl1, a retrotransposon-derived imprinted gene, as a novel driver of hepatocarcinogenesis. PLoS Genet. 2013;9:e1003441.PubMed CentralPubMedView ArticleGoogle Scholar
  60. Chen L, Stuart L, Ohsumi TK, Burgess S, Varshney GK, Dastur A, et al. Transposon activation mutagenesis as a screening tool for identifying resistance to cancer therapeutics. BMC Cancer. 2013;13:93.PubMed CentralPubMedView ArticleGoogle Scholar
  61. Juliano RL, Ling V. A surface glycoprotein modulating drug permeability in Chinese hamster ovary cell mutants. Biochim Biophys Acta. 1976;455:152–62.PubMedView ArticleGoogle Scholar
  62. Choi J, Landrette SF, Wang T, Evans P, Bacchiocchi A, Bjornson R, et al. Identification of PLX4032-resistance mechanisms and implications for novel RAF inhibitors. Pigment Cell Melanoma Res. 2014;27:253–62.PubMed CentralPubMedView ArticleGoogle Scholar
  63. Poulikakos PI, Persaud Y, Janakiraman M, Kong X, Ng C, Moriceau G, et al. RAF inhibitor resistance is mediated by dimerization of aberrantly spliced BRAF(V600E). Nature. 2011;480:387–90.PubMed CentralPubMedView ArticleGoogle Scholar
  64. Johannessen CM, Boehm JS, Kim SY, Thomas SR, Wardwell L, Johnson LA, et al. COT drives resistance to RAF inhibition through MAP kinase pathway reactivation. Nature. 2010;468:968–72.PubMed CentralPubMedView ArticleGoogle Scholar
  65. Shi H, Moriceau G, Kong X, Lee M-K, Lee H, Koya RC, et al. Melanoma whole-exome sequencing identifies (V600E)B-RAF amplification-mediated acquired B-RAF inhibitor resistance. Nat Commun. 2012;3:724.PubMed CentralPubMedView ArticleGoogle Scholar
  66. Pettitt SJ, Rehman FL, Bajrami I, Brough R, Wallberg F, Kozarewa I, et al. A genetic screen using the PiggyBac transposon in haploid cells identifies Parp1 as a mediator of olaparib toxicity. PLoS One. 2013;8:e61520.PubMed CentralPubMedView ArticleGoogle Scholar
  67. Bouwman P, Aly A, Escandell JM, Pieterse M, Bartkova J, van der Gulden H, et al. 53BP1 loss rescues BRCA1 deficiency and is associated with triple-negative and BRCA-mutated breast cancers. Nat Struct Mol Biol. 2010;17:688–95.PubMed CentralPubMedView ArticleGoogle Scholar
  68. Straussman R, Morikawa T, Shee K, Barzily-Rokni M, Qian ZR, Du J, et al. Tumour micro-environment elicits innate resistance to RAF inhibitors through HGF secretion. Nature. 2012;487:500–4.PubMed CentralPubMedView ArticleGoogle Scholar
  69. Wilson TR, Fridlyand J, Yan Y, Penuel E, Burton L, Chan E, et al. Widespread potential for growth-factor-driven resistance to anticancer kinase inhibitors. Nature. 2012;487:505–9.PubMed CentralPubMedView ArticleGoogle Scholar
  70. Molyneux SD, Waterhouse PD, Shelton D, Shao YW, Watling CM, Tang Q-L, et al. Human somatic cell mutagenesis creates genetically tractable sarcomas. Nat Genet. 2014;46:964–72.PubMedView ArticleGoogle Scholar
  71. Koso H, Takeda H, Yew CCK, Ward JM, Nariai N, Ueno K, et al. Transposon mutagenesis identifies genes that transform neural stem cells into glioma-initiating cells. Proc Natl Acad Sci U S A. 2012;109:E2998–3007.PubMed CentralPubMedView ArticleGoogle Scholar
  72. Chew SK, Lu D, Campos LS, Scott KL, Saci A, Wang J, et al. Polygenic in vivo validation of cancer mutations using transposons. Genome Biol. 2014;15:993.View ArticleGoogle Scholar
  73. Furushima K, Jang C-W, Chen DW, Xiao N, Overbeek PA, Behringer RR. Insertional mutagenesis by a hybrid piggyBac and sleeping beauty transposon in the rat. Genetics. 2012;192:1235–48.PubMed CentralPubMedView ArticleGoogle Scholar
  74. Cong L, Ran FA, Cox D, Lin S, Barretto R, Habib N, et al. Multiplex genome engineering using CRISPR/Cas systems. Science. 2013;339:819–23.PubMed CentralPubMedView ArticleGoogle Scholar
  75. Jinek M, East A, Cheng A, Lin S, Ma E, Doudna J. RNA-programmed genome editing in human cells. eLife. 2013;2:e00471.PubMed CentralPubMedView ArticleGoogle Scholar
  76. Qi LS, Larson MH, Gilbert LA, Doudna JA, Weissman JS, Arkin AP, et al. Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression. Cell. 2013;152:1173–83.PubMed CentralPubMedView ArticleGoogle Scholar
  77. Gilbert LA, Larson MH, Morsut L, Liu Z, Brar GA, Torres SE, et al. CRISPR-mediated modular RNA-guided regulation of transcription in eukaryotes. Cell. 2013;154:442–51.PubMed CentralPubMedView ArticleGoogle Scholar
  78. Konermann S, Brigham MD, Trevino AE, Joung J, Abudayyeh OO, Barcena C, et al. Genome-scale transcriptional activation by an engineered CRISPR-Cas9 complex. Nature. 2015;517:583–8.PubMed CentralPubMedView ArticleGoogle Scholar
  79. Perez-Pinera P, Kocak DD, Vockley CM, Adler AF, Kabadi AM, Polstein LR, et al. RNA-guided gene activation by CRISPR-Cas9-based transcription factors. Nat Methods. 2013;10:973–6.PubMed CentralPubMedView ArticleGoogle Scholar
  80. Maeder ML, Linder SJ, Cascio VM, Fu Y, Ho QH, Joung JK. CRISPR RNA-guided activation of endogenous human genes. Nat Methods. 2013;10:977–9.PubMed CentralPubMedView ArticleGoogle Scholar
  81. Shalem O, Sanjana NE, Hartenian E, Shi X, Scott DA, Mikkelsen TS, et al. Genome-scale CRISPR-Cas9 knockout screening in human cells. Science. 2014;343:84–7.PubMed CentralPubMedView ArticleGoogle Scholar
  82. Moffat J, Grueneberg DA, Yang X, Kim SY, Kloepfer AM, Hinkle G, et al. A lentiviral RNAi library for human and mouse genes applied to an arrayed viral high-content screen. Cell. 2006;124:1283–98.PubMedView ArticleGoogle Scholar
  83. Root DE, Hacohen N, Hahn WC, Lander ES, Sabatini DM. Genome-scale loss-of-function screening with a lentiviral RNAi library. Nat Methods. 2006;3:715–9.PubMedView ArticleGoogle Scholar
  84. Silva JM, Li MZ, Chang K, Ge W, Golding MC, Rickles RJ, et al. Second-generation shRNA libraries covering the mouse and human genomes. Nat Genet. 2005;37:1281–8.PubMedGoogle Scholar
  85. Bernards R, Brummelkamp TR, Beijersbergen RL. shRNA libraries and their use in cancer genetics. Nat Methods. 2006;3:701–6.PubMedView ArticleGoogle Scholar
  86. Sanjana NE, Shalem O, Zhang F. Improved vectors and genome-wide libraries for CRISPR screening. Nat Methods. 2014;11:783–4.PubMed CentralPubMedView ArticleGoogle Scholar
  87. Koike-Yusa H, Li Y, Tan E-P, Velasco-Herrera MDC, Yusa K. Genome-wide recessive genetic screening in mammalian cells with a lentiviral CRISPR-guide RNA library. Nature Biotechnol. 2013;32:267–73.View ArticleGoogle Scholar
  88. Wang T, Wei JJ, Sabatini DM, Lander ES. Genetic screens in human cells using the CRISPR-Cas9 system. Science. 2014;343:80–4.PubMed CentralPubMedView ArticleGoogle Scholar
  89. Gilbert LA, Horlbeck MA, Adamson B, Villalta JE, Chen Y, Whitehead EH, et al. Genome-scale CRISPR-mediated control of gene repression and activation. Cell. 2014;159:647–61.PubMedPubMed CentralView ArticleGoogle Scholar
  90. Yang X, Boehm JS, Yang X, Salehi-Ashtiani K, Hao T, Shen Y, et al. A public genome-scale lentiviral expression library of human ORFs. Nat Methods. 2011;8:659–61.PubMed CentralPubMedView ArticleGoogle Scholar
  91. Chen S, Sanjana NE, Zheng K, Shalem O, Lee K, Shi X, et al. Genome-wide CRISPR screen in a mouse model of tumor growth and metastasis. Cell. 2015;160:1246–160.PubMedView ArticleGoogle Scholar
  92. Richmond A, Su Y. Mouse xenograft models vs GEM models for human cancer therapeutics. Dis Model Mech. 2008;1:78–82.PubMed CentralPubMedView ArticleGoogle Scholar
  93. Montini E, Cesana D, Schmidt M, Sanvito F, Bartholomae CC, Ranzani M, et al. The genotoxic potential of retroviral vectors is strongly modulated by vector design and integration site selection in a mouse model of HSC gene therapy. J Clin Invest. 2009;119:964–75.PubMed CentralPubMedView ArticleGoogle Scholar
  94. Modlich U, Navarro S, Zychlinski D, Maetzig T, Knoess S, Brugman MH, et al. Insertional transformation of hematopoietic cells by self-inactivating lentiviral and gammaretroviral vectors. Mol Ther. 2009;17:1919–28.PubMed CentralPubMedView ArticleGoogle Scholar
  95. Wu X, Scott DA, Kriz AJ, Chiu AC, Hsu PD, Dadon DB, et al. Genome-wide binding of the CRISPR endonuclease Cas9 in mammalian cells. Nat Biotechnol. 2014;32:670–6.PubMed CentralPubMedView ArticleGoogle Scholar
  96. Kuscu C, Arslan S, Singh R, Thorpe J, Adli M. Genome-wide analysis reveals characteristics of off-target sites bound by the Cas9 endonuclease. Nat Biotechnol. 2014;32:677–83.PubMedView ArticleGoogle Scholar
  97. Birmingham A, Anderson EM, Reynolds A, Ilsley-Tyree D, Leake D, Fedorov Y, et al. 3′ UTR seed matches, but not overall identity, are associated with RNAi off-targets. Nat Methods. 2006;3:199–204.PubMedView ArticleGoogle Scholar
  98. Jackson AL, Bartz SR, Schelter J, Kobayashi SV, Burchard J, Mao M, et al. Expression profiling reveals off-target gene regulation by RNAi. Nat Biotechnol. 2003;21:635–7.PubMedView ArticleGoogle Scholar
  99. Jackson AL, Linsley PS. Recognizing and avoiding siRNA off-target effects for target identification and therapeutic application. Nat Rev Drug Discov. 2010;9:57–67.PubMedView ArticleGoogle Scholar
  100. Wang H, Yang H, Shivalila CS, Dawlaty MM, Cheng AW, Zhang F, et al. One-step generation of mice carrying mutations in multiple genes by CRISPR/Cas-Mediated genome engineering. Cell. 2013;153:910–8.PubMed CentralPubMedView ArticleGoogle Scholar
  101. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell. 1990;61:759–67.PubMedView ArticleGoogle Scholar
  102. Network TCGA. Comprehensive molecular characterization of human colon and rectal cancer. Nature. 2012;487:330–7.View ArticleGoogle Scholar
  103. Hodis E, Watson IR, Kryukov GV, Arold ST, Imielinski M, Theurillat J-P, et al. A landscape of driver mutations in melanoma. Cell. 2012;150:251–63.PubMed CentralPubMedView ArticleGoogle Scholar
  104. Cancer Genome Atlas Network. Genomic classification of cutaneous melanoma. Cell. 2015;161:1681–96.View ArticleGoogle Scholar
  105. Biankin AV, Waddell N, Kassahn KS, Gingras M-C, Muthuswamy LB, Johns AL, et al. Pancreatic cancer genomes reveal aberrations in axon guidance pathway genes. Nature. 2012;491:399–405.PubMed CentralPubMedView ArticleGoogle Scholar
  106. Waddell N, Pajic M, Patch A-M, Chang DK, Kassahn KS, Bailey P, et al. Whole genomes redefine the mutational landscape of pancreatic cancer. Nature. 2015;518:495–501.PubMed CentralPubMedView ArticleGoogle Scholar

Copyright

Advertisement