Hanahan and Weinberg have proposed the ‘hallmarks of cancer’ to cover the biological changes required for the development and persistence of tumours [Hanahan and Weinberg (2011) Cell 144, 646–674]. We have noted that many of these cancer hallmarks are facilitated by the multifunctional protein YB-1 (Y-box-binding protein 1). In the present review we evaluate the literature and show how YB-1 modulates/regulates cellular signalling pathways within each of these hallmarks. For example, we describe how YB-1 regulates multiple proliferation pathways, overrides cell-cycle check points, promotes replicative immortality and genomic instability, may regulate angiogenesis, has a role in invasion and metastasis, and promotes inflammation. We also argue that there is strong and sufficient evidence to suggest that YB-1 is an excellent molecular marker of cancer progression that could be used in the clinic, and that YB-1 could be a useful target for cancer therapy.
- genomic instability
YB-1 (Y-box-binding protein 1) encoded by the YBX1 gene, is a member of the cold-shock protein superfamily, all of which contain a highly conserved nucleic-acid-binding motif that binds to both DNA and RNA. This motif is located within a region of 65 amino acids termed the ‘cold-shock domain’ which shares greater than 40% identity with prokaryotic cold-shock proteins . This degree of sequence conservation supports the notion that these proteins play an essential role in both the prokaryotic and eukaryotic cell . It is interesting to speculate whether the ancestral nucleic-acid-binding domain of YB-1 has existed through the evolution of prokaryotes to eukaryotes, possibly acquiring new roles as organisms became more complex. If this were the case, it is not surprising that YB-1 has taken on many seemingly diverse roles.
YB-1 was originally identified as a factor that repressed gene transcription by binding to the Y-box (an inverted CCAAT box) of MHC class II promoters . Later the same year, protein-blotting assays using DNA probes revealed that YB-1 binds to the enhancers of the EGFR (epidermal growth factor receptor) and the ERBB2 (HER2) genes . By 1995, it became clear that YB-1 played an important role in regulating cellular proliferation and development . Since then, YB-1 has been shown to be a transcription factor of many genes (reviewed in ), but also directly affects DNA repair, RNA splicing, exon skipping, drug resistance and cancer progression [EMT (epithelial–mesenchymal transition)] in a transcription-independent manner (comprehensively reviewed by Eliseeva et al. ). In the present review we discuss the multifunctional nature of YB-1 thereby illustrating how it facilitates the cancer hallmarks of Hanahan and Weinberg .
The importance of YB-1 in cancer: a preamble
YB-1 rose to prominence following reports of elevated YB-1 protein levels being highly correlated with cancer progression and poor prognosis. Initially this came from IHC (immunohistochemistry) analyses of breast tumours by Royer's group, who showed that levels of cytoplasmic YB-1 correlated with progression in 27 breast cancers . They also showed that the levels of nuclear, but not cytoplasmic, YB-1 correlated with expression of the ABC transporter (ATP-binding cassette transporter) Pglycoprotein MDR1 (multidrug resistance protein 1) in nine cancers. These data were consistent with reports that YB-1 could transactivate the MDR1 gene [8–10] and that high MDR1 protein levels in tumours were associated with poor clinical prognosis. The importance of the correlation between nuclear YB-1 and MDR1 levels with patient prognosis was strengthened by similar observations for osteosarcoma , non-small-cell lung carcinoma [12,13], synovial sarcoma , prostate cancer , melanoma  and multiple myeloma . In addition, nuclear YB-1 and MDR1 were both observed at high levels in 9/27 breast cancers after paclitaxel treatment.
Data such as these has led to the widely accepted view that nuclear YB-1 can function as an oncoprotein which, when present at elevated levels, leads to increased tumour cell proliferation and drug resistance. Then an important study in 2003 showed that YB-1 was essential to control the growth and survival of tumour cells of different origins in vitro . A subsequent study showed this to be dependent on the nuclear translocation of YB-1 . Consistent with this, in 2005 Royer's laboratory demonstrated that sustained overexpression of YBX1 in transgenic mice invariably led to the development of invasive breast cancers, further supporting the view that YB-1 can function as an oncoprotein .
YB-1 AND THE HALLMARKS OF CANCER
In 2000, Hanahan and Weinberg proposed six regulatory pathways that must be overcome in order for a cell to become malignant: uncontrolled proliferative signalling, evading growth suppressors, resisting cell death, replicative immortality, sustained angiogenesis, and invasion and metastasis . In 2011 they proposed two additional hallmarks: deregulated metabolic pathways and avoiding immune destruction, as well as two enabling characteristics: genomic instability and tumour promoting inflammation .
The YB-1 protein is remarkably multifunctional, and in the present review we demonstrate the significant contribution that YB-1 makes to each of Hanahan and Weinberg's cancer hallmarks. We argue that the multifunctionality of YB-1 renders it a true master-regulator of malignancy and therefore it deserves the status given to other multi-potent oncoproteins such as Myc and Ras. Since we focus specifically on the cancer hallmarks of Hanahan and Weinberg, the present review is not intended to provide a comprehensive evaluation of YB-1's functions. For this we recommend the excellent reviews by Eliseeva et al.  and Brandt et al. . In the following pages the role played by YB-1 in each of the hallmarks of cancer will be discussed in turn.
The early association made between YB-1 and proliferation has focussed research on dysregulation of proliferation and the cell cycle by YB-1, which is arguably the most important hallmark of a tumour cell. Reduction of YB-1 expression causes growth inhibition or apoptosis in a broad range of cancer cells both in vitro and in vivo (Table 1). These data suggest that YB-1 plays a critical and non-redundant role in regulating cell proliferation. As a consequence, several studies have investigated the mechanism by which YB-1 regulates cell proliferation. One approach has been to search for YB-1-binding sites or ‘Y-boxes’. However, YB-1 can bind to a variety of DNA sequences that have little sequence similarity [24,25], making it difficult to predict a canonical YB-1-binding site. Thus YB-1 targets must be empirically determined. Given this, several studies have performed large-scale analyses to determine the key transcription targets of YB-1. Using gene expression arrays, changes in transcript abundance were determined following YB-1 knockdown with siRNAs (small interfering RNAs) in ovarian, colorectal, lung and ER (oestrogen receptor α)-positive breast cancer cell lines [26,27]. Another approach has been to identify the promoters bound by YB-1 in colorectal cancer or ER-negative breast cancer cells using ChIP (chromatin immunoprecipitation) promoter arrays [called COC (ChIP on chip)] [28,29] and ChIP sequencing . A comparison of the transcriptional targets of YB-1 in all of these studies showed few in common (results not shown). For example, gene expression data of three cancer cell lines (colorectal, breast and lung) following YB-1 siRNA treatment showed that although a few hundred transcripts were downstream targets of YB-1, only 25 were in common in all lines . Cumulatively, these results suggest that the downstream targets of YB-1 are cell-type-specific. This raises the possibility that any one cell type may utilize only a subset of YB-1's transcriptional capability, since it expresses only a subset of YB-1's potential transcriptional targets. This may be due to the presence of different transcriptional co-factors or YB-1-binding partners in each cell type (e.g. [31,32]). Interestingly however, there appears to be strong common themes to YB-1's transcriptional targets across lineages; rather than being a random set of genes, the downstream targets of YB-1 in different cell types, even if only slightly overlapping, are all enriched for E2F-regulated genes .
The E2F pathway
A seminal paper in 2005 by Rhodes and colleagues studied gene expression data from almost 7000 microarray experiments in the Oncomine database . They observed that transcriptional targets of the E2F family were over-represented in many tumour types, which led them to state, “These results reaffirm that activation of the E2F pathway is a prevalent event in human cancer”. As indicated, it has recently been shown that YBX1 mRNA levels are associated with the E2F1 pathway in breast, colorectal and lung tumours , which was confirmed using ChIP assays, where YB-1 bound directly to the promoters of several canonical E2F1-regulated genes [e.g. CDC6 (cell-division cycle 6), CCNA (cyclin A) and TOP2A (topoisomerase II α 170 kDa)]. Furthermore, bioinformatic analysis of COC data revealed that more than 4000 of the 6000 promoters bound by YB-1 in ER-negative breast cancer cells also have E2F1/E2Fbinding sites (see results published in ), and that 57 of the 88 YB-1 target genes identified by COC analysis in HCT116 colorectal cancer cells had E2F1/E2F-binding sites within their promoters (see results published in ). Therefore multiple studies suggest that YB-1 co-regulates the expression of genes in the E2F1/E2F pathway.
Not only does YB-1 co-regulate E2F target genes, importantly, it also controls expression of different E2F family members, in what appears to be a cell-type-specific manner . For example, YB-1 binds to the E2F2 and E2F5 promoters in MCF7 cells and activates transcription of the E2F1 promoter in A549 cells. Using different experimental systems, in the ER-negative breast cancer cell line SUM149, YB-1 was shown to bind to the E2F2, E2F3 and E2F7 promoters (see results published in ), and in the SKOV-3 ovarian cell line E2F7 expression was increased following YB-1 knockdown (see results published in ). These results suggest that the expression of ‘activator’ and ‘repressor’ members of the E2F family may be turned on and off respectively by YB-1 in a cell-type-specific manner (Figure 1).
In summary, we suggest that YB-1 has evolved to regulate the activity of E2F pathways through two synergistic mechanisms. First, YB-1 transcriptionally regulates several members of the E2F family to promote expression of the ‘activator’ E2Fs and inhibit expression of the ‘repressor’ E2Fs. Secondly, it co-regulates the expression of thousands of E2F target genes by binding to their promoters. These important observations suggest that YB-1 may in fact be an Achilles’ heel of the E2F cancer cell proliferation pathway, which could provide an opportunity to target this pathway by therapeutically targeting YB-1.
PI3K (phosphoinositide 3-kinase)/Akt/mTOR (mammalian target of rapamycin) pathway
In addition to the E2F pathway YB-1 also appears to regulate other cell proliferation pathways. The PI3K class I enzymes and the pathways driven by them are dysregulated in the majority of cancers (reviewed in ). Multiple cell-surface receptors, in particular growth factor receptors (e.g. EGFR and ERBB2), relay growth-promoting signals through the activation of this pathway (reviewed in ). The PI3KCA gene encoding the catalytic subunit p110α is frequently amplified or acquires activating mutations in cancers to further enhance the activity of the pathway . The PI3K pathway signals through the multifunctional protein kinase Akt which, in turn regulates mTOR to influence a number of oncogenic functions, including proliferation, survival, metabolism and metastasis (reviewed in ). The PI3K pathway also cross-talks with the E2F pathway discussed above, by modulating the pro-apoptotic functions of E2F1 .
YB-1 appears to be linked to, and plays an integral role within, the PI3K/Akt/mTOR pathway (Figure 2). It transcriptionally activates the expression of PIK3CA in basal-like breast cancer cells  and YB-1 depends on Akt for its nuclear translocation following phosphorylation at Ser102 in a number of cell lines, including basal-like breast cancer cells , ovarian cancer cells  and melanoma cells . Interestingly, PIK3CA expression is activated by YB-1 irrespective of whether the PIK3CA gene is mutated or amplified, leading to further dysregulation of the PI3K pathway . This results in the subsequent modulation of a number of downstream components of the PI3K pathway, which includes molecules that can phosphorylate YB-1, phospho-RSK (ribosomal S6 kinase) (Ser360) and phospho-Akt(Ser473) [16,39], leading to further YB-1 nuclear localization and further activation of PIK3CA transcription. It has recently been shown in melanoma cells that inhibitors of the PI3K pathway can modestly reduce expression from a cloned YBX1 promoter, suggesting that the activation of this pathway can also promote transcription of YBX1 .
Downstream of Akt is mTOR. mTOR forms a complex with raptor (regulatory associated protein of mTOR) called mTORC1 (mTOR complex 1), which appears to be an important hub that controls many pathways that affect protein and lipid synthesis, autophagy, production of inflammatory cytokines, glycolysis and angiogenesis (reviewed in ). In addition to regulating PI3K/Akt signalling, YB-1 may also regulate mTOR. Reduction of YB-1 with siRNAs in a number of ER-negative breast cancer cells and a paediatric glioblastoma cell line was shown to cause a marked decrease in mTOR protein levels . This was not accompanied by a reduction in mRNA expression, suggesting that YB-1 controls the translation or affects the mRNA stability of mTOR.
Taken together, these results suggest that, similar to its interaction with multiple points of the E2F pathway, YB-1 interacts with multiple points of the PI3K/Akt/mTOR pathway to increase the activity of this pathway in cancer cells.
MAPK (mitogen-activated protein kinase) pathways
There are at least six different molecular pathways associated with MAPKs, but only the Ras/Raf/MEK [MAPK/ERK (extracellular-signal-regulated kinase) kinase]/ERK signalling pathways will be discussed in the present review. MAPK signalling is initiated by growth factor receptors on the cell surface activating Ras, and then Raf, which activates the MAPK kinases MEK and ERK, that in turn activate several downstream pathways, which converge to promote cellular proliferation (reviewed in ). There is complex cross-talk between the Ras/Raf/MEK/ERK and PI3K/Akt pathways (Figure 2), with both pathways able to regulate the other at multiple points (reviewed in [35,42]).
YB-1 activates several members of the Ras/Raf/MEK/ERK pathway (Figure 2). As described above, YB-1 is a transcriptional activator of genes encoding the EGFR and ERBB2 cell-surface receptors, which transmit initial signals to intracellular MAPK pathways [4,43]. YB-1 appears to also regulate a number of genes downstream in the MAPK pathway. COC studies in colorectal cancer cells  and ER-negative breast cancer cells [28,30] have identified that YB-1 binds to the promoters of a number of MEK/ERK pathway genes.
The effects of this signalling pathway are modulated by a family of dual-specific MAPK phosphatases . Interestingly, microarray analysis of cancer cell lines where YB-1 levels were reduced by siRNA treatment showed significantly altered expression of transcripts encoding a number of MAPK phosphatases including MKP2 (MAPK phosphatase 2; see data published in ). This suggests yet another role for YB-1 in the regulation of this pathway.
Interestingly the Ras/Raf/MEK/ERK pathway also activates YB-1 in a positive feedback loop. RSK1  and ERK2  have been shown to phosphorylate YB-1 to promote its transcriptional activity (Figure 2), and a MEK inhibitor has been shown to reduce YB-1 protein abundance .
In summary, several pathways that promote cancer cell proliferation are activated by YB-1. These include the E2F, PI3K/Akt/mTOR and Ras/Raf/MEK/ERK pathways. These three pathways converge and overlap with one another [38,42,47], and YB-1 regulates several of members of each pathway. By doing so, YB-1 promotes cancer cell proliferation through several parallel signalling cascades involving many effector molecules. Again, this makes YB-1 an attractive target for therapies to control cancer cell growth.
Evading growth suppressors and cell-cycle checkpoints
For a cancer cell to undergo sustained proliferation, the two cell-cycle checkpoints regulated by the RB (retinoblastoma) gatekeeper protein and the tumour suppressor p53 must be overcome. YB-1 appears to override both of these checkpoints.
(a) The RB pathway
The critical role of RB as a tumour suppressor is evidenced by the fact that RB or the RB pathway is inactivated in almost all human tumours . In normal cells RB exerts its tumour suppressor activity by interacting with multiple proteins . Arguably the most important of these proteins are ‘activator’ members of the E2F family, which RB inhibits to suppress cell-cycle progression. Following CDK (cyclin-dependent kinase)-mediated hyperphosphorylation of RB (pRB), the activator E2Fs are released from the RB inhibition and can transcriptionally activate numerous genes promoting cell-cycle progression from G1- to S-phase (Figure 3) [50,51]. In cancers, the normal control of cell-cycle progression by RB pathways is reduced, in part, through YB-1.
YB-1 appears to reduce the tumour suppressive activities of RB in several ways. First, it has been shown that YB-1 controls the expression of upstream regulators of RB, namely cyclin D1 [17,27,52], CDK1 and CDK2 . Secondly, as described above, YB-1 is a transcriptional activator of several ‘activator’ E2Fs and is a transcriptional repressor of several ‘repressor’ E2Fs, as well as a co-regulator of S-phase genes, including E2F-1 (Figure 3). It seems possible that the elevated levels of YB-1 in cancer cells strongly activate the expression of ‘activator’ E2Fs so that RB binding becomes insufficient to fully inhibit these activator E2F molecules. Further research is required to fully understand the complex inhibition of the RB tumour suppressor pathway by YB-1.
The p53 pathway
TP53 (tumour protein p53) is renowned as a tumour suppressor as it is inactivated more frequently in cancers than any other gene as yet identified . The p53 protein that TP53 encodes functions as a transcription factor to control expression of a number of genes involved in cell survival and proliferation. p53 is normally present at very low levels in cells due to constant degradation by the E3 ligase MDM2 (murine double minute 2) . However, after stress, particularly DNA damage, p53 becomes phosphorylated preventing interaction with MDM2, thereby allowing p53 levels to increase dramatically . When this occurs, p53 transactivates genes to cause cell-cycle arrest allowing DNA repair, permanent arrest of cell division (senescence) or apoptosis, thereby preventing the accumulation of lesions that could otherwise go on to initiate malignancy (reviewed in [55,56]). The p53 protein is disabled in many tumours. Although this occurs mostly by mutation, p53 can also be disabled by direct interaction with other proteins (e.g. [57,58]).
YB-1 can disable the p53 pathway in cancers by regulating both the activity and the expression of p53 (Figures 4A and 4B). Several studies have demonstrated that YB-1 interacts directly with p53 [59–62] and interferes with the ability of p53 to transactivate genes [59,61,63,64]. For example, YB-1 reduced the p53-driven transcriptional activation of apoptosis-associated genes APAF1 (apoptotic peptidase activating factor 1), NOXA (NADPH oxidase activator) and BAX (Bcl2-associated X protein), but had little effect on the promoter of the CDKN (CDK inhibitor) 1 gene encoding the cell-cycle inhibitory protein p21CIP1 . It was also observed that p53 had greater affinity for the promoters of cell-cycle-associated proteins than those of apoptosis-associated proteins, potentially making it more difficult for YB-1 to override the regulation of cell-cycle-associated gene promoters by p53 . In addition to directly affecting the activity of p53, YB-1 also represses transcription of the TP53 gene . The importance of YB-1 in controlling this pathway was confirmed by the observation that the reduction in YB-1 led to an increase in p53 protein levels and triggered p53-dependent apoptosis in cancer cell lines with wild-type p53 [18,63].
In summary, YB-1 appears to help cancer cells escape cell-cycle checkpoints by inhibiting both the RB and p53 pathways.
Resisting cell death
Cancer cells have evolved to evade the normal apoptotic pathways that would otherwise remove damaged cells. Upon a ‘cellular cue’, a series of proteins transduce a signal to activate an apoptosis pathway that culminates in DNA degradation and systematic disassembly of the cell by activation of a series of proteolytic enzymes called caspases . YB-1 is involved in protecting tumour cells from apoptosis in several ways. p53 probably plays the most important role here to detect damaged DNA and initiate apoptosis if the DNA cannot be repaired. Elevated levels of YB-1 enable cells to subvert the p53-driven apoptosis pathway (Figures 4A and 4B).
Another apoptotic pathway involves the cell-surface death receptor Fas (CD95). Upon binding of Fas ligand (CD95L), an apoptotic signal is transduced from Fas into the cytoplasm through multiple downstream effector molecules, including the executioner caspases, CASP3 and CASP7, to promote orderly cellular disassembly, including PARP [poly(ADP-ribose) polymerase]-mediated cleavage of DNA and finally fragmentation of the cell into apoptotic bodies . YB-1 appears to inhibit the Fas-mediated apoptosis pathway at several points. YB-1 is a transcriptional repressor of the FAS promoter (Figure 4C)  and consistent with elevated levels of YB-1 in tumours, FAS is often down-regulated in cancers [68,69]. YB-1 also inhibits the expression of the gene encoding the pro-apoptotic protein BAX [18,63]. YB-1 may also repress transcription of CASP7 (Figure 4C), since ChIP analysis showed that YB-1 binds to the CASP7 promoter (see results published in ) and reduction of YB-1 levels increased CASP7 expression (see results published in ). In addition to quelling the pro-apoptotic signals from death receptors, the inhibition by YB-1 of BAX and CASP7 may also act to suppress intrinsic apoptotic signals from DNA or mitochondrial damage (reviewed ).
As well as playing a role in activating cellular proliferation, under the control of the PI3K/Akt  and MAPK pathways , E2F1 can also initiate an apoptotic pathway, by promoting the expression of pro-apoptotic genes. This has recently been the subject of considerable interest [72,73]. In response to growth promoting signals, such as fetal bovine serum, there is a PI3K-dependent repression of E2F1′s transcription of apoptosis genes, whereas E2F1 continues to drive the transcription of genes involved in proliferation . As described above, YB-1 is an integral component of the PI3K pathway, driving proliferation by activating many pathway components and being part of positive feedback loops. Therefore it seems probable that YB-1 promotes the expression of E2F1-dependent proliferative genes, but not E2F1-dependent apoptotic genes. We hypothesize that, as observed for p53-regulated cell cycle and apoptosis pathways , YB-1 preferentially co-activates expression of E2F1 proliferation-associated genes and not those driving apoptosis pathways (Figure 4D). In support of this, we have studied the relationship between YBX1 mRNA levels and the inferred activity of the E2F1-driven apoptotic and proliferative transcriptional programmes  using PCA (principle component analysis) of microarray data from breast tumours. The results suggest that YBX1 mRNA levels are associated with the transcription of the proliferative E2F1 target genes, but not apoptotic E2F1 target genes (results not shown).
Replicative immortality and genomic instability
The immortalization of cells is stimulated by: (i) activation of telomere maintenance mechanisms ; (ii) loss of the RB checkpoint; and (iii) loss of p53 function, which together lead to lifespan extension and genomic instability. YB-1 can promote all three of these mechanisms.
Studies of MEFs (mouse embryonic fibroblasts) generated from E13.5 (E is embryonic day) Ybx1−/− mice demonstrated that these cells prematurely senesced compared with control MEFs . It was noted that the Ybx1−/− MEFs had increased levels of Cdkn2a (p16Ink4a) and Cdkn1a (p21Cip1), which would promote senescence. Furthermore, Cdkn2a and Cdkn1a were also elevated at the mRNA level, suggesting that YB-1 may be required for their transcription or mRNA stability. The involvement of these proteins was supported as the induction of senescence could be partially bypassed by knockdown of both of these transcripts . YB-1 thus appears to play a role in controlling replicative senescence.
The loss of genomic stability, leading to alterations in the genome that include amplifications, deletions, translocations or even aneuploidy, is a characteristic of solid tumours. In hereditary cancers, genomic instability is frequently a result of mutations within DNA repair genes, however what promotes this in sporadic tumours is not well understood (reviewed in ). The key proteins driving the response to DNA damage, p53 and ATM (ataxia telangiectasia mutated), are frequently mutated in cancers, but despite the volume of sequence information on human cancers, very few mutations in DNA repair genes have been identified in sporadic cancers . Instead the data point towards an involvement of oncogene-driven replicative stress leading to genomic instability.
Recently, interest has focussed on the RB/E2F pathway in protecting the integrity of the genome, since the loss of RB leads to genomic instability (reviewed in ). An elegant study, performed in non-immortalized cells, tested the ability of both viral and cellular oncoproteins to aberrantly activate the RB/E2F pathway . This led to enhanced cellular proliferation, but without a concurrent increase in nucleotide metabolism, leading to a depletion of the nucleotide pool. The outcome of this was replicative stress, leading to incomplete progression of replication forks and thereby DNA damage.
Given the manner in which YB-1 modulates the RB and E2F pathways (Figure 3) it appears possible that elevated levels of YB-1 may promote genomic instability through replicative stress, without concurrent induction of apoptosis. Indeed, overexpression of YBX1 does appear to promote genomic instability. In a study by Bergmann et al.  the overexpression of YBX1 was associated with genomic instability when expression was targeted to the mammary gland of transgenic mice. All of these mice ultimately developed mammary tumours after 52 weeks. Furthermore, in human mammary epithelial cells, prolonged expression of YB-1 induced a loss of cell-cycle control, genomic instability and centrosomal amplification .
In summary, the results of both YB-1 inactivation and overexpression suggest that YB-1 can promote replicative immortality and genome instability.
(v) Inducing angiogenesis
Aberrant angiogenesis is a hallmark of many solid tumours. Once the tumour reaches a size where nutrients and oxygen becomes limiting, a pro-angiogenic pathway is initiated. This event is termed the ‘angiogenic switch’ [81,82]. There are several proteins involved in this process, with perhaps the best known being VEGF-A (vascular endothelial growth factor A), but also includes PDGF-β (platelet-derived growth factor β), ANG-1 (angiopoietin 1), PGF (placental growth factor), TGF-β (transforming growth factor β), Notch and Wnt pathway proteins . Each of these proteins plays a different role in promoting and regulating blood vessel development. However, in developing tumours, angioregulatory pathways are not as tightly controlled as occurs in normal development, so that dysfunctional vascular beds are formed, often with irregular structure and are poorly synchronized with the needs of the tissues they supply . Because of this, the tumours become more hypoxic, driving further aberrant angiogenesis, and resulting in decreased drug delivery and vascular dissemination of cancer cells. It is therefore not surprising that angiogenesis and tumour invasiveness are closely linked . The association between YB-1 and TGF-β, the Notch and Wnt pathways will be discussed below.
In endothelial cells, YB-1 has been shown to activate expression of pro-angiogenic PDGF-β following thrombin treatment . Studies in epithelial-derived cancer cell lines suggest that YB-1 may up-regulate the expression of other pro-angiogenic genes in a hypoxia-dependent manner. However, under normoxic conditions, YB-1 appears to inhibit the expression of a number of pro-angiogenic genes. For example, in A549 lung cancer cells, reduction of YB-1 led to increased levels of the transcripts encoding the pro-angiogenic chemokines IL-8 (interleukin 8) and CXCL2 [chemokine (C-X-C motif) ligand 2] (see data published in ), suggesting that YB-1 is a transcriptional repressor of these genes (Figure 5). In support of this, YB-1 has been shown to bind to the IL-8 promoter in ER-negative breast cancer cells (see results published in ). More compellingly, YB-1 has been shown to repress transcription of VEGFA by binding to the hypoxia-response region in the VEGFA promoter in normoxic conditions [45,86]. The authors proposed that YB-1 bound to single-stranded DNA would prevent binding of the double-stranded DNA binding HIF-1 (hypoxia-inducible factor 1) complex. YB-1-mediated repression of the VEGFA promoter was considerably enhanced after phosphorylation of YB-1 by activated GSK3β (glycogen synthase kinase 3β) . However, in tumours under hypoxia GSK3β is not activated , potentially reducing the binding of YB-1 to the VEGFA promoter, thereby allowing greater access of HIF-1 to activate the expression of VEGFA [45,86]. These tantalizing data suggest that YB-1 inhibits angiogenesis in epithelial-derived tumour cells under normoxic conditions, however, under hypoxia repression by YB-1 is relieved allowing expression of pro-angiogenic factors. Therefore the induction by YB-1 of genes encoding pro-angiogenic proteins such as PDGF-β and VEGF-A, although requiring further investigation, suggests that YB-1 may play an important role in the angiogenic switch.
Invasion and metastasis
There are many steps involved in the invasion and metastasis of tumour cells to distant sites. Tumour cells initially constrained by BMs (basement membranes), must first dissociate from the tumour mass and cross the BM before invading the adjacent stromal tissue. This migratory behaviour is facilitated by loosening the connections between adjacent cells of the BM and also of the ECM (extracellular matrix) to allow the passage of tumour cells into blood vessels and lymphatic system for dissemination [88,89].
There are several lines of evidence linking YB-1 with a role in invasion and metastasis (Figure 6). For example, in vitro studies have shown that reducing YB-1 levels inhibits the invasive properties of a number of cancer cell lines [39,90,91], and overexpression of YB-1 promotes invasion of MCF-7 breast cancer cells  and Ras-transformed ‘normal’ mammary epithelial cells . Furthermore, analysis of tumour data showed that high YBX1 mRNA levels are associated with lower distant metastasis-free survival rates in breast cancer .
The cadherins are involved in maintaining cell–cell adhesion within the tumour mass, particularly E-cadherin (CDH1), which is frequently inactivated in metastatic cancers [94–96]. One mechanism by which this can occur is via the transcriptional repression of CDH1 by a number of transcription factors including Snail (SNAI1), LEF1 (lymphoid enhancer-binding factor 1) and TWIST1 [97–99]. Interestingly YB-1 has been shown to promote the translation of the mRNAs encoding these CDH1-repressing factors [91,93]. The tumour cell must then become motile, which often appears to involve EMT . EMT is driven by many molecules and pathways that have been linked to YB-1. For example, YB-1 appears to regulate the translation of TGFβ1 , which although a tumour suppressor in normal cells, plays an important role in driving EMT . Both the Wnt and Notch pathways are also involved in EMT  and YB-1 has been shown to bind to the promoters of a number of Wnt pathway proteins . Interestingly a potential secreted fragment of YB-1 has been identified as a ligand for Notch3 receptors .
The dissociation of the BM/ECM occurs via several mechanisms, one of which is proteolysis. Multiple proteolysis pathways appear activated in metastasis including the uPA (urokinase-type plasminogen activator) system and the MMPs (matrix metalloproteinases) [104,105]. uPA promotes degradation of the BM/ECM by plasmin and also activates the MMPs . YB-1 appears to be an activator of uPA as reduction of YB-1 expression led to decreased levels of uPA . The MMPs are a family of proteases with multiple specificities. Given that they can degrade almost all proteins in the BM/ECM, their expression is tightly regulated . Several studies have shown that YB-1 transcriptionally regulates a number of MMPs, including MMP-2 , MMP-12  and MMP-13 . One of these studies showed that whether YB-1 is an MMP activator or repressor, perhaps not surprisingly, depended on the cellular context . However, reduction of YB-1 expression in invasive melanoma cells led to decreased expression of MMP2 . YB-1 has also been shown to increase the levels of the membrane-associated MT1-MMP (membrane-type 1 MMP; also known as MMP-14), which plays a critical role in metastasis [109,110]. In ER-positive breast cancer cells, YB-1 performs this task by subverting the endocytic mechanism and directing MT1-MMP back to the cell membrane where it can interact with and degrade the ECM . CD44 is another membrane-bound protein that plays an important role in metastasis. A recent report has shown that this occurs via interaction with MT1-MMP . YB-1 is a transcriptional activator of the CD44 gene , and also promotes alternate splicing of the transcript to include exon 4 leading to the CD44v4 variant . Isoforms of CD44 mRNA containing exon 4 promote increased invasion of cancer cells (e.g. ).
YB-1 can also regulate a member of the integrin family. These proteins are adhesion receptors classically associated with cell adhesion, migration, differentiation, proliferation and cancer metastasis [115,116]. Experiments have shown that YB-1 expression is linked to that of integrin α6 (also known as CD49f) in both mammary progenitor and breast cancer cells [28,112] and may be involved in the modulation of proliferation and differentiation.
In conclusion, YB-1 regulates multiple proteins and pathways involved in invasion and metastasis.
Studies on tumour cell metabolism in recent years have confirmed that tumour cells gain a selective advantage by generating energy not only from mitochondrial-driven oxidative respiration, but also in the presence of oxygen, fermenting glucose into lactate (termed the Warburg effect , reviewed in ). A number of pathways can promote this ‘aerobic glycolysis’ to enable the use of glucose as a fuel for cancer cells. Once again, it seems a number of the usual players and pathways are involved, frequently those associated with controlling tumour growth, suggesting a close link between proliferative and metabolic pathways. Potentially the most important is the PI3K/Akt1/mTOR pathway, which regulates many components of the glycolytic pathway. Akt1 regulates both the expression and membrane translocation of a number of glucose transporter molecules, contributing to increased glucose uptake [119,120]. mTOR, when part of mTORC1, acts as a hub to link growth factor signalling with activation of metabolic pathways for growth (reviewed in ). Activated mTOR has recently been identified as a key driver of aerobic glycolysis, via increasing the levels of the crucial enzyme in this pathway, PKM2 (pyruvate kinase M2) . mTOR promotes this via the control of HIF1 and MYC transcription factors [122–124], which then increase the expression of genes encoding the splicing proteins PTB (polypyrimidine tract-binding protein), hnRNPA (heterogeneous nuclear ribonucleoprotein A) 1 and hnRNPA2 . This results in alternative splicing of the pyruvate kinase transcript, leading to generation of the pro-glycolytic PKM2 splice form .
YB-1 may affect energy metabolism through the regulation of PI3K/Akt1/mTOR pathways as described above (Figure 2). Interestingly, YB-1 has also been shown to regulate the translation of MYC mRNA via binding to an internal ribosome entry site in the MYC 5′-untranslated region . Therefore, together with mTOR, YB-1 may increase Myc protein levels thereby promoting glycolysis. Furthermore, reducing the expression of YB-1 in rapidly proliferating A549 and HCT116 cells led to a decrease in PKM2 RNA expression (see results published in ).
In addition, YB-1 may affect energy metabolism through the regulation of E2F and RB activity. A recent study of E2f1−/− mice showed that E2F1 is also able to influence metabolic pathways. In conditions where energy demand was high, the RB/E2F1 pathway blocked oxidative respiration to drive the expression of genes involved in glycolytic metabolism .
Another pathway inactivated by YB-1, p53, has also been associated with energy metabolism. p53 has been shown to prevent aerobic glycolysis and promote oxidative phosphorylation , partly through impeding the PI3K/Akt/mTOR pathway by transcriptional activation of the Akt inhibitor PTEN (phosphatase and tensin homologue deleted on chromosome 10) .
Collectively, these findings suggest that YB-1 modulates tumour cell energy metabolism to promote aerobic glycolysis by regulating several molecules and pathways including PI3K/Akt1/mTOR, Myc, PKM2, the RB/E2F1 pathway and p53.
Evading immune destruction
Cancer cells have acquired several ways to evade immunosurveillance mechanisms to proliferate unhindered. These involve both intrinsic and extrinsic mechanisms that either allow tumour cells to avoid detection or removal by the immune cells, or by the secretion of factors that affect immune cell function respectively . A number of molecules associated with YB-1 are involved in intrinsic evasion. For example, to resist immune detection and killing, MHC class II and Fas/CD95 have been observed at lower levels in tumours than in healthy cells ([131,132] and reviewed in ). The genes encoding both of these proteins are transcriptionally repressed by YB-1 [3,67] (Figure 4C).
YB-1 also affects extrinsic mechanisms, including the TGF-β pathway that promotes immune response evasion through multiple mechanisms (reviewed in ). Although a link between YB-1 and this pathway has not been studied in cancer cells, YB-1 has been shown to regulate translation of TGFβ (TGFB1) mRNA in proximal tubule cells, where lower levels of YB-1 inhibit translation . This would be consistent with an established cancer cell, where YB-1 is expressed at elevated levels, which would promote the translation of TGFB1 leading to the activation of the TGFβ pathway. Consistent with this model, we noted that siRNA-mediated reduction of YB-1 expression in MCF7 breast cancer cells reduced the levels of TGFB1 and TGFB3 mRNA (see results published in ).
In summary YB-1 promotes the escape of tumour cells from the immune system by intrinsic mechanisms such as the regulation of the genes encoding MHC class II and Fas. YB-1 may possibly also contribute to extrinsic mechanisms such as regulation of the TGF-β pathway.
A substantial amount of evidence now links inflammation to the development of tumours . For example, patients with inflammatory bowel disease are prone to the development of colorectal cancers, chronic pancreatitis with pancreatic cancer and haemochromatosis with liver cancer (reviewed in ). Furthermore, chronic inflammatory autoimmune diseases such as rheumatoid arthritis and Sjogren's syndrome are associated with increased rates of lymphoma.
Many molecules regulated by YB-1, either transcriptionally or post-transcriptionally, have been linked with promoting inflammation in cancer. These include EGFR, ERBB2, STAT3 (signal transducer and activator of transcription 3) and mTOR [4,41], as well as MMP-2 and CD44 [106,112] whose association with YB-1 has been discussed in earlier sections of the present review. In addition to these, YB-1 has been studied in inflammatory diseases and found to regulate several proteins involved in inflammatory pathways (reviewed in ). For example, the chemokines CCL2 [MCP-1 (monocyte chemoattractant protein 1)] and CCL5 [RANTES (regulated upon activation, normal T-cell expressed and secreted)] induce inflammatory cell infiltration, particularly macrophages, into the tumour microenvironment . This has been especially well described in breast cancer . YB-1 activates transcription from the CCL5 promoter  and may also regulate CCL2 as reduction of YB-1 led to a decrease in CCL2 mRNA levels in ER-positive breast cancer cells (see results published in ).
In summary, YB-1 regulates the expression of several genes encoding proteins known to drive inflammation, such as mTOR, STAT3, MMP-2, CD44, CCL5 and potentially also CCL2, all of which may contribute to tumour development.
YB-1 IN THE CLINIC
Given the master-regulatory role played by YB-1 in all of Hanahan and Weinberg's ‘hallmarks of cancer’, it is not surprising that YB-1 is strongly associated with clinical parameters such as tumour progression. As indicated in the opening section of the present review, IHC and genomic studies have shown that YB-1 protein and mRNA levels are frequently elevated in advanced breast cancer, have an inverse correlation with ER and PR (progesterone receptor) expression, and are associated with poor patient outcome [27,140–143]. Subsequent studies over many years have shown that YB-1 abundance is also associated with the outcome of a range of other human malignancies such as glioblastoma , melanoma , multiple myeloma , osteosarcoma , synovial sarcoma , prostate cancer , colorectal cancer , ovarian cancer [26,146] and lung cancer . Thus the detection and quantitation of YB-1 is potentially a powerful prognostic tool. Despite this the potential clinical importance of YB-1 has been largely underplayed.
One possible reason for this is due to an historical focus on nuclear YB-1. As described above, several studies have suggested that it is nuclear YB-1 that is associated with the more aggressive cancers and poor prognosis. However, a number of studies have found that the overall YB-1 level (which is essentially cytoplasmic) is a sufficient indicator of prognosis [17,147,148]. Indeed, it is very difficult in our opinion to discern nuclear localization of YB-1 using IHC, and in our analyses encompassing three different breast cancer cohorts, the proportion of tumours or cells showing clear nuclear YB-1 is very small. In one study, only three cells in 96 breast cancers showed unequivocal nuclear YB-1 staining , and in the large cohort described by Habibi et al.  only 3% of tumours had some limited nuclear YB-1 staining. However, the overall YB-1 level was found to be prognostic in both cases. That the absolute YB-1 level is sufficient for prognostication is also highlighted by another study of ~400 breast cancers, which showed that abundance of the mRNA encoding YB-1 is significantly associated with poor clinical outcome .
A second reason why YB-1 has attracted less clinical and pathological attention than expected may be due to the use of multiple antibodies for the detection of YB-1 in tumours. This is of critical importance since some antibodies detect both nuclear and cytoplasmic YB-1 , whereas others appear to detect essentially only cytoplasmic YB-1 . This point was emphasized in a comparative study of just two antibodies (raised against residues 1–12 and 299–313 of YB-1), which showed clear differences in their ability to detect nuclear YB-1 and also their ability to show an association with tumour progression . Analysis with additional antibodies against YB-1 generated yet other patterns of staining (results not shown). Thus antibodies raised against different regions of YB-1 may give quite different results. It is paramount therefore that publications state which antibodies are used to enable comparison and reproducibility between studies.
A further issue was identified when it was shown that some antibodies are not only poorly immunoreactive to YB-1, they are also cross-reactive with the nuclear protein hnRNPA1 . Figure 7 shows an example of this. In addition to the ~45 kDa YB-1 band, which is specifically reduced following transfection with YB-1 siRNAs, a second protein at 37 kDa (shown to be hnRNPA1 ) is also detected by this antibody and is unchanged following YB-1 knockdown. Worryingly, hnRNPA1 is predominantly a nuclear protein, which casts doubts on studies using these antibodies for IHC analysis of YB-1, particularly those drawing conclusions from nuclear staining. Despite these findings, one of the cross-reactive antibodies continues to be used  (http://www.abcam.com/YB1-antibody-ab12148references.html). We propose therefore that a simple validation of YB-1 antibodies is performed (e.g. as in Figure 7) before IHC studies with YB-1 antibodies are published.
Collectively, the variety of antibodies with different qualities and specificities has limited the development of antibody-based predictive and prognostic screens utilizing YB-1. However, now knowing the limitations of the area, the field can move on, developing standardized and properly validated (monoclonal) antibodies against YB-1 to enable the use of YB-1 as a powerful prognostic indicator in the clinic.
YB-1 AS A THERAPEUTIC TARGET
Given that the position of YB-1 is upstream of the molecular pathways responsible for all nine hallmarks of cancer, YB-1 is a very attractive therapeutic target. Several approaches to target YB-1 have been employed, including targeting of YB-1 directly, interfering with the activation of YB-1 or targeting the regulators that activate YB-1.
Historically, the first of these used a ‘decoy’ YB-1-binding site to sequester the YB-1 protein. This was successfully employed in cultured cancer cells of many different lineages, with the result of inhibition of tumour cell growth and p53-mediated apoptosis [18,63]. In this case, normal cells (fibroblasts and melanocytes, results not shown) were not sensitive to YB-1 inhibition. Despite the publication of this work almost 10 years ago, the use of nucleic acids as therapeutics has been hampered predominantly by issues of delivery, although rapid advances are now being made [151,152].
Another therapeutic approach to directly target YB-1 could be to use siRNAs. Anti-YB-1 siRNAs have been shown to suppress tumour cell invasion, proliferation, differentiation, insensitivity to chemotherapy and to promote apoptosis, as described previously [17,27,41,153,154]. The challenge will be delivering siRNAs in humans given the widely discussed problems of stability and bioavailability and the current drive to evolve better chemistry and delivery technologies .
As outlined above, the ‘transcription factor’ functions of YB-1 are activated by phosphorylation. The most widely characterized phosphorylation site is at Ser102, which stimulates nuclear translocation and DNA binding . Site-directed mutagenesis showed that S102A YB-1 mutants exhibited reduced cell proliferation  and an interference peptide was designed to serve as another type of ‘molecular decoy’ that competes with YB-1 for phosphorylation by RSK and Akt . This CPP (cell-permeable peptide) inhibited the proliferation of breast and prostate cancer cells in cell culture . Importantly, the CPP had no effect on the growth of normal mammary epithelial cells isolated from patients . Thus peptide-based delivery systems could be used to inhibit YB-1 therapeutically.
The use of ‘signal transduction’ inhibitors is another approach to suppressing YB-1 activity. A side effect of blocking kinases such as PDK-1 , Akt [19,157] and RSK  is quenching of YB-1 phosphorylation at Ser102 and thereby reduction of the transactivation activity of YB-1. This ‘side effect’ on YB-1 may in fact turn out to be responsible for a significant proportion of the activity of these inhibitors. Although Akt was originally reported to phosphorylate and activate YB-1  it was subsequently reported that relative to other kinases such as RSK, Akt might be a minor player in YB-1 activation . In support of this, RSK inhibition with siRNA, or small molecules such as BI-D1870 and SL0101, completely suppress the activation of YB-1, despite the presence of activated Akt in the same cells . Likewise, the MEK inhibitor PD098059 blocks RSK and suppresses the nuclear localization of YB-1 . Moving further upstream, PDK-1 inhibition also suppresses YB-1  and is known to directly activate RSK through phosphorylation at Ser380 . Given that RSK is the most proximal kinase that activates YB-1, and given the numerous YB-1-independent pathways that RSK also activates, it would seem reasonable to focus on blocking RSK [161–163]. The effect on YB-1 of other proteins associated with YB-1 pathways, such as mTOR or ILK (integrin-linked kinase) inhibitors [91,164], have not been fully investigated.
In summary, direct targeting of YB-1 using cell-permeable inhibitory peptides, YB-1 siRNAs or oligonucleotide decoys have shown promise in cell culture experiments. Indirect inhibition of YB-1 is a known side effect of blocking kinases such as PDK-1, Akt, MEK or RSK, and YB-1 blockade may potentially underlie a significant part of the activity of these inhibitors.
The present review has described YB-1 as a master regulator of cancer cell biology, contributing to all nine of Hanahan and Weinberg's ‘hallmarks of cancer’ (Figure 8) and is therefore a bona fide oncoprotein. YB-1 is a multitasking protein that may operate in different, but overlapping, ways in different cell types. Moreover, due to the levels of YB-1 protein and YBX1 mRNA being highly correlated with poor patient outcome, YB-1 should be regarded as a useful biomarker of cancer progression and as a novel therapeutic target. Given the strong association of YB-1 with aggressive (basal-like) breast cancers, targeting YB-1 could, for example, be of particular value for such cancers that are currently difficult to treat. Given the increasing consistency with which YB-1 is being associated with cancer progression and interest in understanding how YB-1 functions, we are confident it will feature prominently in the cancer literature in the future and hopefully, in time, be developed as a biomarker and target for therapy.
The work described in the present review was supported by the Cancer Society of New Zealand [grant numbers 10/21 (to A.L. and C.G.P.) and 11/07 (to A.G.W. and A.W.B.)], the Health Research Council of New Zealand [grant number 10/02 (to A.G.W. and A.W.B.)], the New Zealand Breast Cancer Research Trust [grant number 3621880 (to A.L. and C.G.P.)], the University of Otago Dean's Bequest (NZ), the Canadian Institutes for Health Research (to S.E.D.) and the Cancer Institute NSW [grant number RLP 05/01 (to A.W.B.)].
We thank Ms Sunali Mehta for critically reading the paper prior to submission.
Abbreviations: BAX, Bcl2-associated X protein; BM, basement membrane; CASP, caspase; CDK, cyclin-dependent kinase; CDKN, CDK inhibitor; ChIP, chromatin immunoprecipitation; COC, ChIP on chip; CPP, cell-permeable peptide; ECM, extracellular matrix; EGFR, epidermal growth factor receptor; EMT, epithelial–mesenchymal transition; ER, oestrogen receptor α; ERK, extracellular-signal-regulated kinase; GSK3β, glycogen synthase kinase 3β; HIF-1, hypoxia-inducible factor 1; hnRNPA, heterogeneous nuclear ribonucleoprotein A; IHC, immunohistochemistry; IL-8, interleukin 8; LEF1, lymphoid enhancer-binding factor 1; MAPK, mitogen-activated protein kinase; MDM2, murine double minute 2; MDR1, multidrug resistance protein 1; MEF, mouse embryonic fibroblast; MEK, MAPK/ERK kinase; MKP, MAPK phosphatase; MMP, matrix metalloproteinase; MT1-MMP, membrane-type 1 MMP; mTOR, mammalian target of rapamycin; mTORC1, mTOR complex 1; PDGF-β, platelet-derived growth factor β; PI3K, phosphoinositide 3-kinase; PKM2, pyruvate kinase M2; RB, retinoblastoma; RSK, ribosomal S6 kinase; siRNA, small interfering RNA; STAT3, signal transducer and activator of transcription 3; TGF-β, transforming growth factor β; TP53, tumour protein p53; uPA, urokinase-type plasminogen activator; VEGF-A, vascular endothelial growth factor A; YB-1, Y-box-binding protein 1
- © The Authors Journal compilation © 2013 Biochemical Society