Translocations involving ETS family proteins in human cancer

The ETS transcription factors regulate expression of genes involved in normal cell development, proliferation, differentiation, angiogenesis, and apoptosis, consisting of 28 family members in humans. Dysregulation of these transcription factors facilitates cell proliferation in cancers, and several members participate in invasion and metastasis by activating certain gene transcriptions. ETS1 and ETS2 are the founding members of the ETS family and regulate transcription by binding to ETS sequences. Three chimeric genes involving ETS genes have been identified in human cancers, which are EWS-FLI1 in Ewing’s sarcoma, TMPRSS2-ERG in prostate cancer, and ETV6-RUNX1 in acute lymphocytic leukemia. Although these fusion transcripts definitely contribute to the pathogenesis of the disease, the impact of these fusion transcripts on patients’ prognosis is highly controversial. In the present review, the roles of ETS protein translocations in human carcinogenesis are discussed. *Correspondence to: Kazushi Inoue, Department of Pathology, Wake Forest University School of Medicine, Medical Center Blvd., Winston-Salem, NC 27157, USA, Tel: +1-336-407-1642, Fax: +1-336-765-2486, E-mail: kinoue2@ triad.rr.com


Introduction
The oncogene v-Ets was discovered as a component of a chimeric gene, along with a truncated v-Myb gene, present in the genome of E26, an avian leukosis virus [1][2][3]. Subsequently, v-Ets-related genes have been identified from different species. The Ets family has been established as one of the largest group of transcriptional factors, with diverse functions [2,3]. To date, 28 human ETS family proteins have been identified [3][4][5]. All ETS genes have a conserved amino acid sequence (the ETS domain) of approximately 85 amino acids that forms the DNA-binding domain at the C-terminus required for the recognition of the consensus core sequence GGAA/T (ETS binding site [EBS]) ( Figure 1B). Binding of ETS proteins to target genes is accelerated by the binding of other trans-acting factors to close proximity to the EBS. The pointed (PNT) domain of 65-85 amino acids is the second most conserved region found in a subset of ETS genes ( Figure 1). This domain is found in 11 of 28 human ETS genes and has been shown to participate in protein-protein interaction and oligomerization [3][4][5]. The human ETS factors are classified into 11 subgroups: ETS (ETS1/2), ERG, FLI1, ETV (PEA3, ETV1/4/5), TEL (ETV6/7), ELG (GABPα), TCF (ELK1/3/4), ELF (ELF1/2/4), SPI1 (SPI1/B/C), ERF (ERF, ETV3, ETV3L), and FEV ( Figure 1) [4]. In addition, a subset of four ETS family genes (ELF3, ELF5, EHF, SPDEF) has been placenta-specific subgroup based upon their restricted expression to tissues with high epithelial cell content (4)(5)(6); there are total of 28 ETS proteins in humans [4,5].
ETS factors are regulators of the expression of genes that are involved in various biological processes in response to various signaling cascades, cellular proliferation, differentiation, apoptosis, migration, invasion/metastasis [7]. They also control cell adhesion, tissue remodeling, extracellular matrix composition, and angiogenesis [7]. ETS protein functional activity is modulated by post-translational modification and interaction with other nuclear proteins. The importance of ETS genes in human carcinogenesis is supported by the observations that ETS genes have altered expression patterns, amplified, deleted mutated, and most importantly are located at translocation breakpoints in human leukemias and solid tumors [4][5][6]. Correlation of ETS gene expression levels with tumor progression has been found in most human neoplasias, including leukemias, thyroid, pancreas, liver, prostate, colon, lung, and breast cancers [2]. Aberrant expression of ETS1 and 2 proteins results in the altered regulation of their target response genes, which was recently reviewed [4][5][6].
The ETS transcriptions factors regulate numerous genes by binding to GGAA/T core of DNA-binding, and are involved in stem cell development, cell proliferation, differentiation, development, transformation, angiogenesis, and apoptosis ( Figure 1) [7]. ETS1 and ETS2 are representative members of the ETS family of transcription factors and are downstream effectors of the RAS/RAF/ERK pathway, which involve in cellular proliferation, differentiation, apoptosis and transformation [7].
It was reported that the ETS2 gene was located approximately 17 cM from the breakpoint of common t(8;21) translocation of M2 subtype of acute myelogenous leukemia (AML) [8,9]. They demonstrated that ERG was situated just proximal to ETS2. However, AML1/ETO was cloned and shown to be responsible for t (8;21) in acute myelogenous leukemias [10,11]. Rao et al. performed in situ hybridization and somatic cell hybridization that the ERG gene is located at chromosome 21q22.3 [12,13]. They found that the ERG gene was translocated from chromosome 21 to chromosome 8 in the t(8;21)(q22;q22). It was then found that ERG2 is a nuclear phosphoprotein that bound to purinerich sequences [14]. ERG is considerably more stable than the short-lived ETS1 and ETS2 proteins with a half-life of 21 hours [7]. ERG2 is a sequence-specific, DNA-binding protein and is expressed at higher levels in early myeloid cells than in mature lymphoid cells, acting as a regulator of genes required for maintenance and differentiation of early hematopoietic cells.
It has been reported that the TMPRSS-ERG translocation is found in 50% of prostate cancer (PCa), and is associated with prognosis ( Figure 1B) [17][18][19]. Studies suggest the involvement of another member of the ETS family, known as ETV6, in human cancer. ETV6 is frequently activated in human leukemia with ETV6-RUNX1 (TEL-AML1) [20,21]. The roles of these ETS proteins in causing human cancer, especially acute leukemia, Ewing's sarcoma, and PCa are discussed in this review.

ERG Gene cloning and expression
The Reddy's group isolated a cDNA that comes with the complete coding sequence of ERG and predicted 363-residue protein. They predicted that ERG share about 40% and 70% homology with the 5' and 3' regions of the v-Ets oncogene, respectively, suggesting that ERG belongs to the ETS oncogene family ( Figure 1B)   The structure of ETS family proteins. The human ETS factors are classified into 11 subgroups: ETS (ETS1/2), ERG, FLI1, ETV (PEA3, ETV1/4/5), TEL (ETV6/7), ELG (GABPα), TCF (ELK1/3/4), ELF (ELF1/2/4), SPI1 (SPI1/B/C), ERF (ERF, ETV3, ETV3L), and FEV [4,5,165,166]. There are totally 28 ETS proteins in humans. ETS translocations involving Fli1, ERG, and TEL have been discussed in this review. The ETS domain that is essential for DNA-binding is shown in dark box. The DNA-binding by the ternary complex factor (TCF) subfamily of ETS-domain transcription factors is tightly regulated by intramolecular and intermolecular interactions [167][168][169] . These data indicate that p16 Ink4a is the target in the pathogenesis of EWS. In ESFT, hypoxia has been shown to contribute to resistance of apoptosis via HIF1α [74]. EWS-FLI1 expression increases transiently under hypoxic conditions in a HIF1α-dependent manner; colocalization of HIF1α and necrotic areas in an ESFT tissue array has been reported [74].
Transgenic mouse models have shown their ability to closely reproduce the conditions of transgenes in mice [75,76]. However, no genetically engineered mouse models (GEMM) have been created for Ewing's sarcoma so far because transgenic expression of the EWS-Fl1 fusion gene causes embryonic lethality [52]. The differences in signaling pathways, Ets-binding sites, co-regulators between mice and humans mice cause additional issues. To make GEMM for EWS, EWS-FLI1 should be expressed exclusively in mesenchymal stem cells (MSCs) to avoid its expression in hematological tissues [77]. Torchia et al. therefore took an approach to express EWS-FLI1 in mesenchymal tissues. They created mice with Cre-inducible expression of EWS-FLI1 from the ubiquitous Rosa26 locus [78]. When crossed with Mx1-Cre mice (interferon-inducible the most commonly used "deleter strain" in experimental hematology that is interferon-inducible) [79], activation of EWS-FLI1 resulted in the rapid development of myeloid/erythroid leukemia characterized by expansion of primitive mononuclear cells causing hepatosplenomegaly, severe anemia, and death [78]. Gene expression profiles of primary and transplanted animals were highly similar, suggesting that activation of EWS-FLI1 was the primary event leading to disease in this model. The Cre-inducible EWS-FLI1 mice provided a novel model system to study the contribution of this oncogene to malignant disease in vivo [78].
The Lozano lab created conditional transgenic GEMM for Ewing's sarcoma by inducing EWS-FLI1 expression in the primitive MSC of the embryonic limb buds by making use of specific recombinases that allow spatio-temporal control of the EWS-FLI1 gene (EF mice) [52]. When crossed to the Prx1-Cre transgenic mouse, which expresses Cre recombinase in the primitive mesenchymal cells of the embryonic limb bud, the EF mice had number of developmental defects of the limbs [52]. These included shortening of the limbs, muscle atrophy, cartilage dysplasia, and immature bone. Under the condition of p53 deletion, EWS-FLI1 accelerated the formation of sarcomas from a median time of 50 to 21 weeks, with poorly differentiated phenotype [52]. Taken together, the results suggest that EWS-FLI1 inhibits normal limb development and accelerates the formation of poorly differentiated sarcomas.
One important issue we have to remember to reproduce Ewing's sarcoma in GEMM is the influence of the reciprocal chimeric protein FLI1-EWS generated from the locus [80]. Elzi et al. reported that FLI1-EWS was frequently expressed in Ewing's sarcoma and presented evidence that endogenous FLI1-EWS was required for Ewing's sarcoma growth, and that it cooperated with EWS-FLI1 in human mesenchymal stem cells through abrogation of the proliferation arrest induced by EWS-FLI1. Thus, creation of double transgenic GEMM for EWS-FLI1 and FLI1-EWS under specific promoter might be necessary to reproduce Ewing's sarcoma [80].

TMPRSS2-ERG fusion gene in prostate cancer (PCa)
By using a bioinformatics approach (Figure 3), Tomlins et al. [81,82] reported candidate oncogenic chromosomal aberrations in PCa. Two ETS transcription factors, ERG and ETV1, were identified ( Figure 1). They demonstrated that 23 of 29 PCa samples (79%) had rearrangements in ERG or ETV1 by FISH. Later experiments have shown that the TMPRSS2-ERG translocation was found in ~50% of Caucasians, with a lower frequency in African Americans, and even less common in Asians [83]. Cell line experiments suggested that the androgen-responsive promoter elements of TMPRSS2 mediate the overexpression of ETS family members in PCa [81]. Using chromatin immunoprecipitation coupled with sequencing, Yu et al. found that ERG disrupts androgen receptor (AR) signaling by inhibiting AR expression, binding to and inhibiting AR activity, and inducing repressive epigenetic programs via direct activation of the H3K27 methyltransferase EZH2 [84]. In normal tissues, the TMPRSS2 and ERG genes are localized tandemly on chromosome 21q22 ( Figure 3) [81]. The TMPRSS2-ERG fusion joins TMPRSS2 exons 1 or 2 to ERG exons 2, 3, or 4, which results in activation of the ERG transcription factor [81,85].
The mechanism of translocation was studied by Mani et al. [86]. By studying human PCa cells with fluorescence in situ hybridization, they showed that androgen signaling induces proximity of the TMPRSS2 and ERG genomic loci, both located on chromosome 21q22.2 ( Figure  3). Subsequent exposure of the cells to γ-irradiation, which causes DNA double-strand breaks, facilitates the formation of the TMPRSS2-ERG gene fusion. Their results explain why TMPRSS2-ERG gene fusions are restricted to the prostate, which is dependent on androgen signaling [86].

Transgenic models for TMPRSS2-ERG gene fusion in mice
Tomlins et al. [87] studied the role of TMPRSS2-ERG gene fusion product using in vitro and in vivo model systems. Transgenic mice expressing the ERG gene fusion product under androgen-regulation developed PIN. Introduction of the ERG gene fusion product into primary or immortalized benign prostate epithelial cells induced an The TMPRSS2-ERG or ETV chimeric gene has been found in ~80% of prostate cancer samples of radical prostatectomy [81]. The former translocation has been found in ~50% of Caucasians, with a lower frequency in African Americans, and even less common in Asians [83]. invasion-associated transcriptional program, but did not increase cellular proliferation or anchorage-independent growth [87]. ERG knockdown in VCaP cells (PCa cell line established from a vertebral metastatic lesion) induced a transcriptional program consistent with prostate differentiation [87]. Importantly, VCaP cells and benign prostate cells overexpressing ERG directly engage components of the plasminogen activation pathway to mediate cellular invasion, potentially representing a downstream ETS target susceptible to therapeutic intervention. Thus, TMPRSS2-ERG fusions mediate invasion, consistent with the histologic distinction between PIN and PCa [87].
Carver et al. [88] reported that aberrant expression of ERG is a progression event in prostate tumorigenesis. Interestingly, PCa specimens containing the TMPRSS2-ERG rearrangement were significantly enriched for loss of the tumor suppressor PTEN, consistent with the report by King et al. [89]. Similarly, transgenic overexpression of ERG in mouse prostate tissue promotes marked acceleration and progression of high-grade prostatic PIN to prostatic adenocarcinoma in a Pten-heterozygous background [88]. In vitro overexpression of ERG promotes cell migration, a property necessary for tumorigenesis, without affecting proliferation. They showed that ADAMTS1 and CXCR4, two candidate genes strongly associated with cell migration, were upregulated in the presence of ERG overexpression [88]. Thus, ERG has a distinct role in PCa progression and cooperates with PTEN haploinsufficiency to promote progression of high grade PIN to invasive adenocarcinoma.
Recently, Mounir et al. compared the transcriptional effects of TMPRSS2-ERG expression in a transgenic mouse model [90]. ERG repressed the expression of a previously unreported set of androgen receptor (AR)-independent neuronal genes that were indicative of neuroendocrine (NE) cell differentiation. Cell sorting and proliferation assays performed after ERG knockdown indicated that ERG drove proliferation and blocked the differentiation of prostate cells to both NE and luminal cell types [90]. They also provided evidence that these NE cells were resistant to pharmacological AR inhibition, and can revert to the phenotype of parental cells upon restoration of AR/ERG signaling. Thus ERG may have a direct role in preventing resistance to anti-androgen therapy [90].

Expression of the TMPRSS2-ERG fusion protein in PCa
Although the TMPRSS2 (chr 21q22.3, exon 1) -ERG (chr 21q22.2, exon 4) fusion with intra-chromosomal deletion of 3MB on chromosome 21 is quite common (Figure 3), other expressed variants have been reported [91,92] differing in the exon break points in TMPRSS2 (exons 1-5) and ERG (exons 2-5). Depending on the exon break point, protein translation is predicted to start either (1) at the genuine ATG of ERG (exon 3), (2) the native ATG of TMPRSS2 (exon 2), or (3) an in-frame ATG within exon 4 or 5 of ERG; in the third cases the N terminus of ERG would be truncated (Figure 3). This is important since the expression of selected TMPRSS2-ERG break point variants has recently been linked to specific clinicopathological features of PCas.
Wang et al. [93] reported that the expression of TMPRSS2-ERG fusion mRNAs and correlated the isoforms expressed and expression levels with clinical outcome in cancers from men undergoing radical prostatectomy. Overall, 59% of clinically localized PCas expressed the TMPRSS2-ERG fusion gene, confirming the initial observations of high frequency expression of this mRNA in PCa [81]. Expression of an isoform, in which the native ATG in exon 2 of the TMPRSS2 gene was in frame fused with exon 4 of the ERG gene, was associated with clinical and pathologic manifestations of aggressive disease [93]. Expression of other isoforms, in which the native ERG ATG in exon 3 was used, was associated with seminal vesicle invasion with poor outcome following radical prostatectomy [93]. Cancers not expressing these isoforms expressed higher levels of fusion mRNAs, which was present in PCa with early prostate-specific antigen recurrence. Thus, both isoforms of TMPRSS2-ERG fusions expressed and expression levels might affect PCa progression with poor prognosis [93]. Further studies are needed to elucidate the roles of the known and variant ERG fusions in the pathogenesis and androgen dependency of PCa.

Aberrant TMPRSS2-ERG signaling in PCa
It has been reported that increased expression of ETS proteins in adult murine prostate epithelial cells is sufficient to induce epithelial hyperplasia and focal prostatic PIN lesions, but not enough in progression to carcinoma [94,95], which are consistent with the findings by Tomlins et al. for the fusion protein PCa [81,82]. ERG interacts with activations in PI3K signaling, such as PTEN inhibition or AKT1 upregulation, to cause the development of a well-differentiated adenocarcinoma. Thus, loss of PTEN and the presence of the TMPRSS2-ERG gene fusion are events significantly associated with PCa [96]. Bismar et al. proposed the hypothesis that PCa development might be driven initially by PTEN hemizygous loss that caused high grade PIN lesions, which leads to genomic instability, chromosomal rearrangement, and progression to cancer [97]. Subsequent biallelic PTEN inactivation characterizes a particularly aggressive subset of metastatic and hormone-refractory PCa [98]. HDAC1 upregulation is common in PCa, and was found to be increased in tumors with ERG rearrangement [99]. ERG overexpression in PCa is highly implicated in promoting motility and invasiveness, high levels of HDAC1, and subsequent down-regulation of HDAC1-targeted genes, activation of WNT/β-catenin signaling pathway, and inhibition of apoptosis. Activation of the AR through the WNT/β-catenin signaling results in increase in AR expression, enhanced transcription of TMPRSS2-ERG, and high levels of ERG [100]. Elevated ERG, in turn, modulates the growth of PCa cells by upregulating the MYC oncogene, and by abrogating the differentiation of prostate epithelium [101]. Elevated MYC expression in primary PCa is biologically relevant and may be a predictor of future biochemical recurrence [102]. In summary, activation of the HDAC1-WNT/β-catenin-MYC pathway by ERG arrangement is associated with invasive behavior in PCa [103,104].
In the first group, Attard et al. reported that patients with two or more copies of the TMPRSS2-ERG fusion gene due to the interstitial deletion had worse survival rates than patients without TMPRSS2-ERG rearrangement [106]. This is consistent with the view that ERG overexpression is responsible for driving cancer progression [118], and that the 2.8 Mb deletion (containing genes with tumor suppressor activity) may contribute to the oncogenic potential of the TMPRSS2-ERG fusion product [81]. To understand molecular signatures associated with poor outcome in PCa, Markert et al. [107] analyzed a microarray data set characterizing 281 PCa from a Swedish cohort.
They identified a subset of tumors manifesting stem-like signatures together with p53 and PTEN inactivation, which had very poor survival outcome; a second group characterized by the TMPRSS2-ERG fusion with poor survival outcome, composed of 18% of the tumors in the dataset [107]. Font-Tello et al. reported that the expression levels of the TMPRSS2-ERG fusion and ERG mRNA were associated with high Gleason scores, aggressive PCa phenotypes, low PTEN expression, and poor prognosis [108,109]. Finally Deplus et al. reported that TMPRSS2-ERG increased bone tropism of PCa cells and metastasis development [110]. Overall, these results suggest that TMPRSS2-ERG could be molecular markers of aggressiveness for PCa.
In the second group, Saramäki et al. [111] studied the presence of the TMPRSS2-ERG rearrangement by reverse transcription-PCR and fluorescence in situ hybridization in 19 PCa xenografts and 7 PCa cell lines. The expression of ERG was studied in the xenografts, cell lines, and in 49 freshly frozen clinical prostate samples by quantitative reverse transcription-PCR. Seven of 19 (37%) of the xenografts overexpressed ERG and had TMPRSS2-ERG rearrangement. In clinical tumor specimens, the overexpression of ERG was associated with the rearrangement. Fifty of 150 (33%) of the prostatectomy specimens and 28 of 76 (37%) of the hormone-refractory PCa on tissue microarrays carried the TMPRSS2-ERG rearrangement [111]. It was associated with longer progression-free survival in patients treated by prostatectomy, and was an independent predictor of favorable outcome. The fusion was not associated with other clinical markers such as Gleason score, clinical stage, prostate-specific antigen.
In the third group, Gopalan et al. [115] analyzed TMPRSS2-ERG gene rearrangement status by fluorescence in situ hybridization in 521 cases of clinically localized surgically treated PCa with 95 months of median follow-up. 42% of primary tumors and 40% of metastases had rearrangements. 11% had copy number increase of the TMPRRS2-ERG region. Rearrangement alone was associated with lower grade, but not with stage, biochemical recurrence, metastases, or survival. They reported that a subgroup of cancers with chromosomal number instability (CNI) and rearrangement by deletion, with two or more copies of the deleted locus, tended to be more clinically aggressive. DNA index assessment revealed that the majority of tumors with CNI of TMPRSS2-ERG had generalized aneuploidy/tetraploidy in comparison to control tumors. They concluded that translocation of TMPRSS2-ERG was not associated with clinical outcome [115]. It has also been reported that TMPRSS2-ERG gene fusion was not associated with PCa recurrence, but that ERG gene copy number gain without fusion was associated with twice the risk of PCa recurrence [116,117]. They explained that the ERG gene copy number increase in PCa was the result of tumor aneuploidy, which potentially leads to chromosome 21 polysomy [118]. In this context, the association of greater probability of tumor progression with increased copy number of the ERG gene formed without fusion is not surprising given the previous evidence supporting aneuploidy as a negative prognostic indicator in PCa [119,120].
It seems that early studies tended to interpret TMPRSS2-ERG gene fusion as a marker of poor prognosis; however, later large cohort studies report that it is a marker of better prognosis. There are several explanations for the different prognostic values of TMPRSS2-ERG: they are clinical settings, size of patient cohorts, differences in sample collection, and the technique used for determination of fusion gene. It is generally believed that high expression of TMPRSS2-ERG fusion transcript in PCa tumor tissue in combination of high PSA and low PTEN levels are associated with aggressiveness of the tumor with poor prognosis.

ETV6-RUNX1 (TEL-AML1 ) in human leukemia
Acute lymphoblastic leukemia (ALL) is a hematopoietic malignancy characterized by clonal proliferation of immature blood cell precursors (blasts) that progresses rapidly into systemic disease. Signs of leukemia include fever and hemorrhage caused by reduced number of platelets. In pediatric B-cell ALL, the most frequent chromosomal lesion is t(12;21)(p13;q22), which results in the fusion of ETV6/RUNX1 (E/R fusion; also known as TEL/AML1) fusion gene (Figure 4) [121][122][123]. This genetic alteration occurs in ~25% of pediatric ALL diagnosed between the ages of 2 and 10 years [123,124]. It was originally believed that this rearrangement is a favorable prognostic indicator based on the excellent molecular response to treatment and beneficial clinical outcome [125][126][127]. However, others found late relapses occurring in up to 20% of patients [128,129], suggesting that it is a sign of ominous prognosis in some patients.
The E/R fusion gene develops as an early event in childhood ALL. The expression of E/R results in generation of a persistent pre-leukemic clone, which postnatally changes to ALL after secondary genetic events [129]. The detection of the E/R fusion sequence in identical twins and pediatric ALL indicates that this gene fusion happens in the prenatal

5' 3'
TEL, ~350kb, 12p13 period [130]. However, 1) the occurrence of ALL simultaneously is found in only 5% of twin children, 2) the postnatal latency period of the disease is highly variable, and 3) data from transgenic animal models of E/R fusion [131] suggest that additional genetic events are required for the development of overt ALL. Early studies addressing secondary genetic changes have focused on the deletion of the non-rearranged ETV6 allele which is a tumor suppressor by dimerizing with E/R to reduce its transforming activity [132,133]. Consistently, loss of ETV6 is found in as many as 70% of t(12;21) (+) ALLs [134,135]. Moreover, ~20% of t(12;21) (+) patients have additional genetic change in ETV6 or RUNX1 [136]. These genomic changes are so common that ETV6 has been attributed to the development of E/R(+) ALLs.
To understand the genetic evolution of E/R(+) ALL without prejudice, identification of entire spectrum of genetic changes that accompany this fusion gene is necessary. The genome of E/R (+) ALL has been well-characterized at the copy number and cytogenetic level. Generally, E/R (+) ALL has an average of 2.8 (0-14) additional copy number alterations (CNAs) [137]. As predicted, deletion of 12p (39%) was the most common abnormality resulting in the loss of wild type ETV6 allele [138]. The 9p deletion containing the INK4b/ARF/INK4a locus [139][140][141] can be seen in up to 25% of E/R(+) patients, and the B-cell differentiation regulator PAX5 [126,142]. CDKN2a/b, ETV6, PAX5, deletions, chromosome 6q loss, and chromosome 16 gain are probably among the earliest genetic aberrations in E/R(+) ALL [143]. Fuka et al. addressed the issue by conducting a shRNA-mediated knockdown (KD) of the E/R fusion gene and investigated the consequences with two E/R(+) leukemic cell lines [144]. Microarray analyses identified 777 genes whose expression was significantly altered. The E/R KD -upregulated set was enriched for genes included in the "cell activation", "apoptosis", "signal transduction", "immune response", and "development and differentiation" categories, whereas in the E/R KD -downregulated set consisted only with the "PI3K/AKT/mTOR signaling" and "hematopoietic stem cells", which was confirmed with primary E/R(+) ALL samples [144]. The results suggest that the E/R chimeric protein accelerates tumorigenesis by up-regulating genes involved in cell proliferation without sending signals to the PI3K-AKT-mTOR pathway.

Pathogenesis, diagnosis, and treatment of ALL in relapse
Most ALL relapses occur during treatment, following discontinuation of treatment or within the first 2-year after treatment completion. Certain relapses have been reported to occur even after longer period of time following the initial diagnosis [125][126][127][128][129]145]. Site of relapse and length of first remission are the major criteria for the classification of patients after first relapse, which include patients with isolated or concurrent marrow, isolated central nervous system (CNS), isolated testicular and other extramedullary relapses with or without CNS involvement [145]. Leukemia relapse is the result of outgrowth of a clonal cell population not entirely eradicated by treatment. Each case of ALL comprises of a unique rearrangement of immunoglobulin or T-cell receptor (TCR) genes; therefore, detailed study on rearrangements of immunoglobulin or TCR genes typical of ALL clones may help to determine the origin of leukemia recurrence. Origin of the relapsed ALL has been addressed based on the study of genome-wide DNA copy number analysis on matched samples obtained at the time of diagnosis and relapse [146].
Standard recovery regimens for relapsed ALL are typically based on different combinations of the same agents used in frontline chemotherapy [146,147]. Many practitioners adopt therapeutic strategies involving risk-adapted treatment, alternating short term chemotherapy, in some cases together with cranial/craniospinal irradiation, and conventional maintenance therapy [146]. Salvage treatment after ALL relapse involves inducing a second remission with conventional intensive chemotherapy and consolidation, reintensification and maintenance therapy, or allogeneic stem-cell transplantation (SCT) to strengthen a treatment [148]. Effective treatment of ALL relapse very much depends on the risk-based treatment allocation of patients in order to maximize response to therapy while minimizing toxicity and adverse effects from chemotherapy [146][147][148]. Using the prognostic factors such as first remission duration; site and immunophenotype of relapse; genetic alterations; and initial response to relapse therapy, distinct subgroups of relapsed ALL can be identified, which may either be treated with chemo and radiation therapies or by additional allogeneic steam cell treatment [149]. Since responses to single agent therapy have been poor, integrating new chemotherapy agents in combination with novel approaches, new antimetabolites, and monoclonal antibodies against leukemia-associated antigens can been implemented [150,151].

Molecular mechanisms and the treatment of E/R(+) ALL in relapse
In most E/R patients, relapses occur several years after discontinuation of treatment meaning that relapse is a genetic alterations [125][126][127][128][129]. To characterize the clonal origins of E/R(+) ALL relapse, numerous studies have compared immunoglobulin/T cell receptor gene rearrangements, genomic structure of un-rearranged ETV6 allele, and CNA patterns of diagnostic and relapse samples in the same patient [152][153][154][155]. Most importantly, late relapsing cases of E/R(+) ALL are chemosensitive with long-term remissions, implying that initial survival and re-emergence have common properties [156]. Comparison of genomic boundaries of un-rearranged ETV6 allele indicated that the relapse clone derived from a sibling clone already present at diagnosis [153,154]. It is possible that this minor population shows only moderate reduction during initial therapy, but rapidly expands before relapse. After relapse, these clones are rapidly eradicated by chemotherapy [152]. In conclusion, the relapse clones of E/R(+) leukemia originate from those exist at initial presentation.
Then what is the impact of E/R fusion in patients' responsiveness to therapy? To gain insight into the relapse mechanisms, Kuster et al. [157] analyzed single nucleotide polymorphism arrays for CNAs in 18 matched diagnosis and relapse ALLs. They identified recurrent, mainly non-overlapping deletions associated with glucocorticoid-mediated apoptosis targeting the Bcl2 modifying factor (BMF), glucocorticoid receptor NR3C1, and components of the mismatch repair pathways [157]. Fluorescence in situ hybridization demonstrated that BMF deletions, NR3C1 and mismatch repair alterations became more common at relapse. These findings implicate that glucocorticoidassociated drug resistance in E/R(+) relapse, which is a direction for future therapies.
Gandemer et al. performed a long-term, follow-up retrospective study to address the outcome of patients with E/R(+) leukemia relapses [158]. They reported that E/R had a strong effect on overall survival after relapse. In 81% of cases the relapses that were found late, they were combined with extramedullary relapses. The 5-year survival rate of patients with E/R(+) ALL relapses reached 81% when the relapse occurred after 36 months [158]. In multivariate analysis, only the duration of first remission remained associated with outcome. Overall, they found an excellent outcome for patients with E/R(+) leukemia relapses that occurred more than 36 months after diagnosis [158]. The duration of first complete remission may be a guide to define the treatment strategy for patients with relapsed E/R(+) leukemia. Similar findings were reported by other groups [159][160][161][162][163], demonstrating that E/R fusion is a sign of favorable prognosis although relapse does happen in 20% of patients. The report from Borssen et al. is especially intriguing since t(12; 21)(p13; q22) E/R fusion is associated with hTERT promoter [164] methylation with short telomeres, and thus with more favorable prognosis of ALL.

Current problems and future directions
ETS proteins are frequently activated in human cancers by translocation. To make GEMM for EWS, the EWS-FLI1 fusion transcripts express exclusively in mesenchymal stem cells to avoid its expression in hematological tissues. When crossed with interferon -inducible Mx1-Cre mice, activation of EWS-FLI1 resulted in the rapid development of myeloid/erythroid leukemia characterized by expansion of mononuclear cells causing hepatosplenomegaly, severe anemia, and death. Gene expression profiles of animals were highly similar to those of human Ewing's sarcoma, suggesting that activation of EWS-FLI1 was the primary event leading to disease in this model. The Cre-inducible EWS-FLI1 mice provide a novel model system to study the contribution of this oncogene to malignant disease in vivo. The disease is accelerated in p16 Ink4a -null mice but not Arf-null mice suggesting the critical role of p16 INK4a in Ewing's sarcoma. In ETS family, the ERG and ETV genes are frequently activated by translocation in human cancer; since the fusion protein often has nearly normal or normal functions of ERG, we speculate that overexpression of ERG proteins is essential for prostate cancer development. The prognostic value of TMPRSS2-ERG fusion in human PCa is still controversial; it seems that it is associated with intermediate survival outcomes second to p53 and PTEN inactivation. The prognostic value of TMPRSS2-ERG fusion will be determined in future prospective studies.
In pediatric B-cell ALL, the translocation involving ETV6 and RUNX1 are very common. It is believed that this is a favorable prognostic factor. The fusion protein develops as an early event in childhood ALL. CDKN2a/b, ETV6, PAX5, deletions, -6q, and +16 are among the earliest genetic aberrations in E/R(+) ALL; they are secondary events that contribute to the leukemic disease. Although 20% of patients suffer from late relapse, studies show that relapsed leukemia responds to chemotherapy with 5-year survival rate of more than 80%, and thus it is believed that it is a favorable prognostic indicator. However, leukemia-free survival remains low for many patients after relapse and despite efforts to intensify therapeutic outcomes for these children. The development of novel therapeutic approaches such as anti-sense therapy or small molecule inhibitor to the chimeric gene holds great promise for the first-line treatment for patients with E/R(+) relapsed ALL.
In any case of human cancer with ETS translocations, specific therapy can be attempted to fusion proteins to eradicate tumor cells since they play critical roles in causing the neoplastic disease.