Oncogene (2011) 30, 2595–2609 & 2011 Macmillan Publishers Limited All rights reserved 0950-9232/11 www.nature.com/onc REVIEW Nucleophosmin and its complex network: a possible therapeutic target in hematological diseases

E Colombo1,2, M Alcalay1,2 and PG Pelicci1,2

1Department of Experimental Oncology, European Institute of Oncology, Milan, Italy and 2Dipartimento di Medicina, Chirurgia e Odontoiatria, University of Milano, Milan, Italy

Nucleophosmin (NPM, also known as B23, numatrin or alternative splicing of the NPM1 transcript, have been NO38) is a ubiquitously expressed phosphoprotein belong- reported (Wang et al., 1993) (Figure 1). In the long ing to the nucleoplasmin family of chaperones. NPM is isoform, known as B23.1, exon 9 is spliced to exon 11 mainly localized in the nucleolus where it exerts many of and the coding sequence stops at exon 12, resulting in a its functions, but a proportion of the protein continuously protein of 294 residues. In the short isoform, B23.2, shuttles between the nucleus and the cytoplasm. A growing exon 9 is spliced to exon 10 that immediately contains a number of cellular proteins have been described as stop codon, and as a result, B23.2 is a truncated protein physical interactors of NPM, and consequently, NPM is of 259 residues. The prevalent form expressed in all thought to have a relevant role in diverse cellular tissues is B23.1, although the biological significance and functions, including ribosome biogenesis, centrosome physiological function of B23.2 remain largely unclear. In duplication, DNA repair and response to stress. NPM the human EST database a further transcript variant that has been implicated in the pathogenesis of several human lacks exon 8 has been identified (GenBank accession malignancies and intriguingly, it has been described both number: NM_199185). Its functional role as well as its as an activating oncogene and a tumor suppressor, tissue distribution are currently totally unknown. depending on cell type and protein levels. In fact, NPM is biochemically defined as a histone chaperone increased NPM expression is associated with different (Frehlick et al., 2007). It belongs to the nucleoplasmin/ types of solid tumors whereas an impairment of NPM nucleophosmin family of nuclear chaperones, sharing function is characteristic of a subgroup of hematolologic the core-domain at the N-terminus of the protein. This malignancies. A large body of experimental evidence links portion of the NPM protein includes an acidic tract and the deregulation of specific NPM functions to cellular is responsible for oligomerization and chaperone activ- transformation, yet the molecular mechanisms through ity (Hingorani et al., 2000). Histone chaperones are which NPM contributes to tumorigenesis remain elusive. required for the correct assembly/disassembly of nucleo- In this review, we have summarized current knowledge somes, and therefore are necessary for DNA-dependent concerning NPM functions, and attempted to interpret its activities like transcription, replication and repair multifaceted and sometimes apparently contradictory (Eitoku et al., 2008). Accordingly, NPM has been activities in the context of both normal cellular homeo- isolated as a putative histone chaperone using a series stasis and neoplastic transformation. of in vitro assays that reveal chromatin related functions. Oncogene (2011) 30, 2595–2609; doi:10.1038/onc.2010.646; Okuwaki et al., (2001a) showed that NPM favors published online 31 January 2011 replication of adenovirus DNA in complex with viral basic core proteins (viral nucleosome). They also Keywords: nucleophosmin; leukemia; cancer demonstrated that NPM interacts with human histones, binds preferentially to histone (H3-H4)2 tetramers and is able to transfer nucleosomes to naked DNA favoring chromatin assembly. On the other hand, NPM, like Nucleophosmin (NPM) is a multi-functional cellular other histone chaperones, decondenses sperm chromatin chaperone (Okuwaki et al., 2001b). Therefore, NPM behaves in vitro as a classical histone chaperone favoring NPM is the product of the NPM1 gene, which in assembly and disassembly of chromatin templates. humans contains 12 exons and maps to chromosome Whether this is a major role of NPM also in vivo is 5q35; two isoforms of the NPM protein, derived from less clear. NPM is part of the centromere protein A protein complex, which represents the nucleosome Correspondence: Dr E Colombo or Professor PG Pelicci, Department variant that replaces H3 in centromeric nucleosomes. of Experimental Oncology, Istituto Europeo di Oncologia, Via (Foltz et al., 2006, 2009) However, if NPM is required Adamello 16, Milano 20139, Italy. for centromere-specic assembly of centromere protein A E-mail: [email protected] or [email protected] nucleosomes has not been fully investigated. Received 23 November 2010; revised 24 December 2010; accepted 31 A number of reports implicate NPM in the control December 2010; published online 31 January 2011 of DNA transcription, in most cases through the Nucleophosmin and its complex network E Colombo et al 2596 S4 S10 S70 S125 T199 T219 T234/237 S254

B231.1 NPM1 isoform1 R230 R263

B231.2 NPM1 isoform2 257 259

NPMmut

287 298 NES oligomerization domain (1-119) Nuclear Localization Signal (141-157) Nuclear Export Signals (42-49; 94-102) Moderately basic region (189-257) Acidic domains (120-132; 161-188) Aromatic region (258-294; unique to B23.1 isoform) Central region Nucleolar localization region (WW 288, 290) NPM1 isoform 2 unique region (257-259) Mutated C-terminal domain (287-298; unique NPMmut) Figure 1 Schematic representation of NPM isoforms and AML-associated NPM mutant. Relevant functional domains are highlighted together with the aminoacidic position in parentheses. Modified aminoacids (phosphorylated and sumoylated) are also indicated. S4: serine4, phosphorylated by polo-like kinase 1 (Plk1) in mitosis and by Plk2 in S-phase. S10: serine10, phosphorylated by unknown CDK, involved in G2-M transition. S70: serine70, phosphorylated by unknown CDK, involved in G2-M transition. S125: serine125, phosphorylated by casein kinase II and ATR; involved in DNA damage response, chaperone activity and NPM1 mobility. T199: threonine199, phosphorylated by CDK2 in G1 and CDK1 in mitosis; involved in centrosome duplication and DNA damage response. T219: threonine219, phopshorylated by CDK1 in mitosis; involved in RNA binding. R230: arginine230, sumoylated; involved in stability, localization, proteins interaction, apoptosis. T234/37: threonine234/37, phosphorylated by CDK1 in mitosis; involved in RNA binding. S254: serine 254, phosphorylated by unknown kinase in mitosis. R263: arginine263, sumoylated; involved in stability, localization, proteins interaction and apoptosis.

interaction with a variety of different transcription in vivo. In vitro functional studies suggest that NPM factors. NPM can either contribute to transcriptional displays chaperone-like behaviors on a wide variety of activation, as in the case of Myc-interacting zinc finger denatured substrates: (i) NPM inhibits temperature- protein 1 (Miz1) (Wanzel et al., 2008), androgen dependent and independent protein aggregation, due to receptor (Leotoing et al., 2008), ribosomal genes thermal denaturation, protecting enzymes from loss of (Murano et al., 2008), c-Myc (Li et al., 2008) and activity; (ii) NPM promotes renaturation of chemically nuclear factor of k light polypeptide gene enhancer in B- denatured proteins and it appears to preferentially bind cells (NF-kB) (Dhar et al., 2004), or to repression, as for denatured substrates through the exposure of hydro- the activating protein transcription factor 2 (AP2a) (Liu phobic regions in both NPM and substrates (Szebeni et al., 2007a). However, the exact mechanism through and Olson, 1999). Although NPM binds ATP, at least which NPM mediates these regulatory activities is not in vitro(Chang et al., 1998), its binding to substrates is completely elucidated, and it is unclear if it functions as governed by association-dissociation thermodynamics, a general co-factor during transcription or if it is and release of the substrate does not require ATP physically associated to specific promoters. In support hydrolysis (Szebeni and Olson, 1999), as seen with other of a more general role for NPM in transcriptional molecular chaperones including Hsp70 and chaperonins regulation, it has been reported that NPM interacts with (Ruddon and Bedows, 1997). Interestingly, it has been and inhibits the activity of hexamethylene bis-acetamide shown that phosphorylation of NPM by casein kinase II inducible 1 (HEXIM1), a negative regulator of Pol II- induces the release of denaturated substrates (Szebeni dependent transcription (Gurumurthy et al., 2008). et al., 2003). In the case of specific interactions between Furthermore, NPM has been involved in different ways NPM and its binding partners, which involve defined and at different levels as a modulator of chromatin protein domains, such as the human immunodeficiency status. In fact, NPM was found to enhance acetylation- virus-Rev protein, NPM phosphorylation does not dependent transcription of an artificial chromatin affect the interaction (Szebeni et al., 2003). All these template, likely by removing acetylated histones (Swa- data come from in vitro studies using recombinant minathan et al., 2005). In a different experimental proteins, therefore, we do not know if NPM shows the setting, Zou and colleagues recently showed that NPM same behavior in living cells. However, there are a blocks chromatin acetylation mediated by the activity number of reports showing that NPM is required in cells of lysine acetyltransferase 2A (GCN5), and inhibits for the correct localization and/or stability and activity GCN5 dependent transactivation in a cell-based assay. of some of its interactors. Intriguingly, NPM has been found associated with the NPM modulates the half-life and the nucleolar CCCTC-binding factor, an insulator binding protein localization of many of its most relevant binding implicated in higher order chromatin organization partners, including the tumor suppressors ARF (Yusufzai et al., 2004; Liu et al., 2008). (p19ARF in mouse, p14ARF in human) (Kuo et al., 2004; Beside its specific role as histone chaperone, NPM Colombo et al., 2005), F-box and WD repeat domain displays a broad chaperone activity both in vitro and containing 7 (Fbw7g) (Bonetti et al., 2008) and the

Oncogene Nucleophosmin and its complex network E Colombo et al 2597 SUMO1/sentrin/SMT3 specific peptidases 3 and 5 anti-apoptotic response involves the interferon inducible (SENP3-5) (Yun et al., 2008), and influences the stability double-stranded RNA-dependent protein kinase PKR. of the cell cycle inhibitor p21/WAF/CIP (p21) (Xiao NPM binds PKR and inhibits its pro-apoptotic activ- et al., 2009). Moreover, the chaperone activity of NPM ities both in mouse embryo fibroblasts (MEFs) and in has been linked to its ability to assist the transport of Fanconi-anemia derived lymphoblasts (Pang et al., other proteins across the nuclear membrane. Indeed, 2003). Dhar and St Clair (2009), recently showed that, NPM itself is a protein that shuttles between nucleus although NPM overexpression favors stabilization of and cytoplasm (Borer et al., 1989) and, accordingly, it the tumor suppressor p53 and, consequently, results in contains both a nuclear localization signal and a nuclear increased levels of its targets p21 and BCL2-associated export signal (Wang et al., 2005; Yu et al., 2006). X protein (BAX), it simultaneously prevents the pro- In vitro, NPM facilitates the nuclear import of other apoptotic accumulation of p53 in the mitochondria, thus nuclear localization signal containing proteins (Szebeni limiting the apoptotic response (Dhar and St Clair, et al., 1997). On the other hand, NPM is required for 2009). On the other hand, it has been reported that, in ribosomal protein L5 nuclear export (Yu et al., 2006) neuronal cells, NPM, following apoptotic stimuli (such and, more in general, for the export of pre-ribosomal as staurosporine in vitro and ischemia in vivo), translo- nuclear particles (Maggi et al., 2008). cates from the nucleolus to the cytoplasm where it binds The correlation between NPM and ribosome biogen- to BAX and p53 favoring mitochondrial cytochrome-c esis is even more complex. The NPM long isoform B23.1 release and apoptosis (Kerr et al., 2007). These data is manly localized in the granular region of the nucleolus suggest that NPM, as nuclear-cytoplasmic molecular (while the B23.2 isoform is prevalently nuclear diffuse) chaperone, may change cell fate depending on the (Wang et al., 1993), and this specific localization stimulus, cell type and activated pathway. correlates with its reported activity in the stimulation A number of reports have demonstrated that over- of ribosomal DNA transcription (Murano et al., 2008) expression of NPM in different cellular contexts and processing of ribosomal RNA precursors (pre- correlates with increased cellular survival upon stress. rRNA) (Savkur and Olson, 1998). More recently, it was In fact, NPM overexpressing cells were more resistant to shown that NPM binds the second internal transcribed ultraviolet irradiation (Wu et al., 2002a, b; Maiguel spacer (ITS2) of the pre-rRNA leading to the release of et al., 2004), ionizing radiation (Dalenc et al., 2002), 28S ribosomal RNA. Accordingly, it has been reported hypoxia (Li et al., 2004), oncogenic overexpression (Li that inhibition of NPM expression leads to a decrease in et al., 2007) and actinomycin D treatment (Yao et al., 28S rRNA levels, and accumulation of the 32S inter- 2010). A possible explanation lies in the observation that mediate (Itahana et al., 2003). overexpression of NPM has been associated to reduced However, the RNA binding activity of NPM does not levels of damaged DNA in cells (Li et al., 2006, 2007) or seem to be confined to ribosomal RNA (Herrera et al., to increased DNA repair activity (Wu et al., 2002a, b). 1995). In fact, it has been shown that NPM is selectively Although the molecular mechanisms are largely un- deposited on mRNA during polyadenylation, but the known, it has been shown that in response to hypoxia biological significance of this binding remains unknown (Li et al., 2004) and UV (Maiguel et al., 2004), NPM (Palaniswamy et al., 2006). More recently, NPM has limits the phosphorylation of p53 on serine 15, a post- been reported to bind the 30 untranslated region of the translational modification that stabilizes p53 and thus product of the chicken connective tissue growth factor favors the p53-dependent apoptotic response (Appella gene, ccn2, triggering its degradation, and this mechan- and Anderson, 2001). These data have been interpreted ism contributes to modulate chondrocyte differentiation as NPM being a negative regulator of p53 stabilization, (Mukudai et al., 2008). therefore limiting apoptosis; however, based on avail- able data, one can envision an alternative scenario in which NPM, by limiting the extent of DNA damage NPM controls cell cycle progression and cell survival (either directly preventing damage and/or favoring DNA repair), can attenuate the DNA-damage response NPM is generally reported as a protein that promotes (DDR) and, consequently, diminish the levels of DDR- cellular growth and survival upon apoptotic stimuli and dependent p53 phosphorylation and inhibiting apopto- it appears to regulate these functions in a cell-type sis. In support of this view, NPM was shown to suppress specific manner. Furthermore, in hematopoietic cells, oncogene-induced apoptosis through mechanisms that NPM expression decreases during differentiation and involve diminished levels of DNA damage. In fact, NPM overexpression limits maturation rate both in vitro MEFs that are either wild type (WT) or defective for (Hsu and Yung, 2000; Chou et al., 2007) and in vivo (Li DNA repair pathways but overexpressing NPM showed et al., 2006). Intriguingly, in hematopoietic stem cells increased survival upon overexpression of c-Myc, which NPM overexpression also increases self-renewal, repo- correlated with a decrease of c-Myc induced DNA pulating ability and survival to DNA damage (Li et al., damage and, therefore, attenuated p53 stabilization 2006). In a neuronal cell line (PC12) NPM was shown to (Li et al., 2007). be required for nerve growth factor-dependent anti- On the other hand, it has been shown that NPM apoptotic activity by inhibiting the DNA fragmentation directly interacts with p53 and its overexpression in capacity of the caspase-activated DNase (Ahn et al., primary diploid fibroblast leads to p53 stabilization and 2005). Another pathway through which NPM exerts an senescence. Downregulation of NPM reduces p53

Oncogene Nucleophosmin and its complex network E Colombo et al 2598 stabilization in response to different stress stimuli cellular events. Itahana et al. (2003) reported that ARF and therefore reduces p53-dependent cell cycle arrest negatively regulates NPM stability in a sort of feed-back (Colombo et al., 2002). The effect of NPM on the cell loop, leading to decreased pre-rRNA processing, cycle may partly be due to its reported role in thus suggesting that NPM is the main target of ARF controlling the nuclear translocation of growth arrest in its ability to inhibit ribosomal RNA processing. On and DNA-damage-inducible a-protein (GADD45a), a the other hand, Brady et al. (2004) did not confirm target of p53 and of the tumor suppressor BRCA1 that either ARF-dependent NPM degradation and/or ARF- is required for G2-M cell cycle arrest, apoptotic dependent inhibition of pre-rRNA processing, but response and DNA repair (Gao et al., 2005). Along showed that ARF blocks NPM shuttling, likely inter- the same line, Kurki et al. showed that in response to fering with ribosome biogenesis, and that NPM UV, NPM undergoes nucleoplasmic redistribution and competes with Mdm2 for ARF binding thus favoring binds to HDM2 (Mdm2 in mouse). HDM2 functions as p53 degradation. It must be noted that all these reported an E3 ligase that ubiquitinates p53 contributing to its ARF activities on NPM stability and shuttling require rapid turnover in cycling cells. Binding of NPM to high levels of exogenous ARF overexpression in the HDM2 prevents its interaction with p53 and leads to cells; however, in endogenous conditions, their stoichio- p53 stabilization (Kurki et al., 2004). metry is quite different and, although most of ARF is It appears, therefore, that upon UV treatment, NPM bound to NPM, a large proportion of endogenous NPM has the same function described for ARF in response to protein is not bound to ARF (Bertwistle et al., 2004). oncogenic stimuli (Weber et al., 1999). ARF is a Therefore, it is still not clear if these reported and nucleolar protein that triggers oncogene-induced cell contradictory mechanisms may have a role in physiolo- cycle arrest mainly through the inhibition of Mdm2- gical conditions. mediated p53 degradation. It was originally proposed It has been shown that MEFs lacking both NPM and that ARF sequesters Mdm2 in the nucleolus where it p53 have an increased proliferative rate compared can not bind and ubiquitinate p53 (Weber et al., 1999). with MEFs lacking only p53. ARF expression is induced More recently, other reports showed that p53 is also by oncogenic stimuli both in vitro (Lowe and stabilized in the absence of nucleolar Mdm2 and that Sherr, 2003) and in vivo (Ventura et al., 2007). In the non-nucleolar forms of ARF can also activate p53 and double knockout cells, ARF levels are lower, thus suppress growth, thus suggesting that nucleolar locali- suggesting that in these cells ARF no longer contributes zation is not required for ARF activity (Llanos et al., to the control of cell cycle progression. Accordingly, 2001; Korgaonkar et al., 2002). Moreover, it has become ectopic expression of oncogenes in MEFs lacking both clear that ARF can induce cell cycle arrest in the absence p53 and NPM failed to induce high levels of ARF of p53 (Weber et al., 2000), and this activity has been expression leading to increased oncogene-dependent related to the ability of ARF to inhibit ribosomal RNA cellular growth and transformation (Colombo et al., processing (Sugimoto et al., 2003). NPM has been 2005). reported as the main interactor of ARF in the cell In conclusion, beside the difficulties in establishing a (Itahana et al., 2003; Bertwistle et al., 2004). ARF unified model, available data clearly indicate that the mutants that do not bind NPM are highly unstable, get NPM-ARF interaction has a central role in controlling rapidly degraded by the proteasome and display very their cellular functions. Further studies are required low anti-proliferative activity (Kuo et al., 2004). These to establish their relative contribution in cancer data, together with the highly unstable structure of ARF development. (Bothner et al., 2001), strongly suggest that NPM is the molecular chaperone that favors the correct folding of ARF in cells. Accordingly, NPMÀ/À cells express very Regulation of NPM activities through post-translational low levels of ARF, which is rapidly degraded, and fails modifications to localize in the nucleolus (Colombo et al., 2005). Indeed, only the nucleolar pool of ARF is stable (Rodway et al., 2004). In summary, NPM binds ARF Different post-translational modifications have been reported for NPM, which correlate with a specific and is required for its correct nucleolar localization and activity and/or subcellular localization (Figure 1). stability in the cell. The biological consequences of the NPM-ARF interaction are a matter of debate. It has been reported Phosphorylation that downregulation of NPM leads to ARF relocation NPM was originally defined as a phosphoprotein and, to the nucleoplasm and p53 stabilization, thus suggest- indeed, many kinases have been shown to phosphorylate ing that NPM sequesters ARF in the nucleolus NPM. As it will become clear, phosphorylation of NPM preventing its binding to Mdm2 (Korgaonkar et al., at different sites seems to be important in regulating cell 2005). However, NPM downregulation leads to p53 cycle progression at multiple steps. Threonine 199 activation and cell cycle arrest even in the absence of (T199) is phosphorylated by the Cdk2-cyclinE complex ARF, due to accumulation of DNA damage in the cells during G1-S transition and this phosphorylation event is (Colombo et al., 2005). Therefore, it is generally difficult required for centrosome duplication (Okuda et al., to investigate p53-dependent ARF pathways in the 2000). In fact, cells expressing a T199A NPM mutant absence of NPM due to overlapping but independent display defects in centrosome number leading to

Oncogene Nucleophosmin and its complex network E Colombo et al 2599 abnormal mitosis (Tokuyama et al., 2001). Interestingly, DNA repair activity in cells overexpressing NPM, as in p53À/À MEFs, which display an abnormal number previously discussed. of centrosomes, either the absence of CDK2/CDK4 activity or the presence of a T199A NPM mutant is Acetylation sufficient to revert this phenotype (Adon et al., 2009). Shandilya et al. recently reported that NPM is A functional nuclear export signal and the capacity to acetylated by the p300 acetyltransferase enzyme and bind to the Ran-Crm1 nuclear export signal-dependent de-acetylated by the NAD-dependent deacetylase sir- nuclear export complex, are required to localize NPM to tuin-1 (SIRT1) both in vitro and in living cells. centrosomes and control their duplication (Wang et al., Acetylated NPM co-localizes in the nucleoplasm with 2005). T199 phosphorylation has been also linked to RNA PolII, likely inducing its transcriptional activity. specific NPM localization in nuclear speckles where it Interestingly, acetylated NPM is increased in oral represses pre-mRNA processing (Tarapore et al., 2006) carcinomas and is enriched on the promoter of genes and, more recently, T199 phosphorylated NPM has implicated in tumor progression (Shandilya et al., 2009). been shown to accumulate at sites of DNA damage favoring DNA repair (Koike et al., 2010). T199 together with T219, T234 and T237 are phosphorylated before Sumoylation mitosis by the Cdk1-cyclinB complex, abolishing the Two sumoylated lysines have been mapped in the NPM RNA binding activity of NPM (Peter et al., 1990; protein: K230 and K263, the latter representing the Okuwaki et al., 2002). main sumoylation site (Liu et al., 2007b). The K263R The exact role of NPM in mitosis is not known but NPM mutant does not correctly localize to the nucleolus lack of phosphorylation or lack of NPM expression has or at centrosome, leading to altered regulation of been reported to result in spindle check-point activation centrosomal duplication and likely interfering with due to severe mitotic defects such as improper mitotic NPM dependent cell cycle regulation (Liu et al., spindle formation and defects in kinetokore-microtu- 2007b). However, Haindl et al. found that the NPM bule attachment (Zhang et al., 2004; Amin et al., 2008). K263R mutant is still sumoylated and does not alter its During mitosis, NPM is mainly dispersed throughout nucleolar localization. They alternatively propose that the cell with a minor fraction located at the spindle poles sumoylation of NPM prevents its activity in regulating and its phosphorylation is required for this localization pre-rRNA processing (Haindl et al., 2008). Moreover, (Yao et al., 2004; Negi and Olson, 2006). Several sites NPM sumoylation has been reported to control the are phosphorylated in mitosis: S4 by polo-like kinase 1 binding of NPM to the tumor suppressor retinoblasto- (Zhang et al., 2004) (but also in S-phase by Plk2 (Krause ma protein (pRb), leading to pRb nucleolar sequestra- and Hoffmann, 2010)), and S70 and S254 by unknown tion and activation of the transcription factor E2F1 (Liu kinases (Nousiainen et al., 2006). Moreover, NPM is a et al., 2007b), thus contributing to cell cycle control. target of the NimA-like protein kinase (Nek2A), NPM is also associated with the two isoforms of the although the phosphorylation sites have not been ErbB-3 binding protein 1, a RNA binding protein mapped (Yao et al., 2004). Recently, it has been shown involved in the control of cell proliferation and that the NPM S10A-S70A phospho-mutant blocks apoptosis. In particular, the p42 isoform of ErbB-3 cellular proliferation at G2-M, whereas the phosphomi- binding protein 1 binds only to unsumoylated NPM in metic NPM mutant (S10E/S70E) overrides stress- response to growth factors, whereas the p48 isoform induced G2-M arrest favoring the interaction between constitutively binds sumoylated NPM. Impairment of CDK1 and Cdc25c through Cdc25c nuclear retention the association between NPM and ErbB-3 binding (Du et al., 2009), further strengthening the relevance of protein 1 isoforms leads to decreased cell proliferation NPM phosphorylation in the control of mitosis. and increased apoptosis (Okada et al., 2007). Another relevant NPM phosphorylation site is serine The sumoylation status of NPM may also have a role 125. This site is a target of casein kinase II and is in controlling its stability. Sumoylation protects NPM required for the release of NPM substrates during from proteolytic cleavage upon staurosporine-induced in vitro chaperone tests (Szebeni et al., 2003). Interest- apoptosis, (Liu et al., 2007b) both in HeLa and PC12 ingly, S125-phosho NPM is more mobile, although the cells. However, following growth factor stimulation, as biological consequences have not been investigated in PC12 cells treated with nerve growth factor, (Negi and Olson, 2006); more recently, casein kinase II unsumoylated NPM strongly interacts with the protein dependent NPM phosphorylation has been linked to kinase Akt in the nucleoplasm and this interaction NPM nuclear retention and matrix association (Wang protects NPM from proteolytic cleavage and increases et al., 2010). Serine 125 has also been reported to be cell survival (Lee et al., 2008). Therefore, sumoylation phosphorylated by ataxia telangiectasia and Rad3 may have different effects on NPM stability that depend related kinase (ATR) in response to UV irradiation by on protein localization and cellular state. Maiguel et al. The authors suggest that NPM phos- NPM is sumoylated upon ARF overexpression phorylation by ATR competes with p53ser15 phosphor- through an indirect mechanism that involves SENP3, a ylation, as NPM overexpression leads to a decrease in nucleolar SUMO2-3 deconjugating protease (Tago p53ser15 phosphorylation upon UV(Maiguel et al., et al., 2005). ARF induces the phosphorylation, 2004). This result, however, also may be ascribed to ubiquitination and degradation of SENP3, which de- decreased levels of damaged DNA and/or increased sumoylates NPM among other substrates (Nishida and

Oncogene Nucleophosmin and its complex network E Colombo et al 2600 Yamada, 2008). Therefore, ARF and SENP3 compete involved in a number of cellular functions overlapping to control the level of NPM sumoylation (Haindl et al., with reported NPM activities such as DNA repair and 2008; Kuo et al., 2008). However, ARF requires NPM replication, chromatin modulation, insulation, tran- to induce SENP3 degradation and, on the other hand, scription, apoptosis and differentiation (Kraus, 2008), NPM is required for the correct nucleolar localization it is not known at the moment if and how NPM and and stability of both ARF and SENP3 (Kuo et al., 2008; PARP1/2 may influence each other’s activities. Yun et al., 2008). ARF and SENP3 do not seem to compete for NPM binding suggesting that NPM may provide a molecular platform in which ARF and SENP3 NPM is required in vivo for DNA integrity and can meet and counteract each other’s activities. chromosome stability Although SENP3 downregulation prevents pre-rRNA processing similarly to what is observed upon NPM NPM has been widely implicated both in the main- downregulation (Haindl et al., 2008), it is unlikely that tenance of DNA integrity, either preventing or repairing NPM is the main target of the ARF-SENP3 pathway in DNA damage, and in the regulation of chromosomal controlling the cell cycle. In fact, in cells lacking NPM stability through the control of centrosome duplication the level of SUMO2/3 modified proteins (other than and mitosis. However, little is known about the NPM) is increased in the nucleolus, and downregulation molecular mechanisms underlying these functions of of SENP3 causes cell cycle arrest in NPM null cells as NPM and moreover, it is not clear if in vivo absence of well as in cells expressing NPM (Kuo et al., 2008). NPM has any impact on DNA integrity or chromosome number. Interesting insights emerge from the study of NPM knock out models. Ubiquitination NPM protects DNA integrity not only following NPM is a very stable protein and little is known about external stimuli (Li et al., 2006, 2007) but also in steady- the proteasome-dependent regulation of its turnover. state conditions. NpmÀ/À mouse embryos die in uterus However, NPM is a target of ubiquitination and this at 10.5 dpc, display widespread apoptosis, p53 upregu- modification may modulate different NPM functions. lation and accumulation of g-H2AX foci (a marker of NPM is targeted by the BRCA1/BARD1 ubiquitin damaged DNA and DDR) (Colombo et al., 2005). ligase complex, which catalyzes an unusual Lys-6-linked These observations were also confirmed in NpmÀ/À polyubiquitin chain that does not induce proteasome- murine hematopoietic precursors derived from the yolk dependent degradation but, on the contrary, stabilizes sac, where reintroduction of NPM was sufficient to NPM (Sato et al., 2004). NPM and the BRCA1/BARD1 revert this phenotype (Colombo et al., 2005). Available complex co-localize in mitosis and the ubiquitin-ligase data do not allow to discriminate if the accumulation of activity of BRCA1/BARD1 is required in mitosis for a g-H2AX foci is due to a direct involvement of NPM in correct mitotic spindle assembly, in accordance with the the DNA repair machinery (Borggrefe et al., 1998; reported role of BRCA1 in maintaining chromosome Koike et al., 2010), or if it is the indirect consequence of stability (Joukov et al., 2006). NPM is involved in other NPM activities such as chromatin organization centrosome duplication and spindle pole formation (Frehlick et al., 2007) and/or DNA replication (Take- although further investigations are required to establish mura et al., 1999; Okuwaki et al., 2001a). Increased if NPM is a critical target of BRCA1/BARD1 during proliferation imposed by non-physiological growth mitosis. promoting signals or oncogene expression leads to It was reported that ARF mediates NPM poly- replication stress and accumulation of damaged DNA ubiquitination and degradation upon ectopic overex- (Bartkova et al., 2006; Di Micco et al., 2006). Interest- pression in cells (Itahana et al., 2003), although these ingly, NpmÀ/À embryos show an increased proliferative data have not been confirmed (Brady et al., 2004). More rate (measured as increased number of cells that recently, it has been shown that NPM is de-ubiquiti- incorporate Bromodeoxyuridine) (Colombo et al., 2005). nated by the nucleolar pool of the ubiquitin specific On the other hand, Grisendi et al. (2005) have shown peptidase 36 (USP36) favoring NPM stabilization the presence of multiple centrosomes and abnormal (Endo et al., 2009b) and, on the other hand, NPM is mitosis in both NpmÀ/À embryonic tissue and in required for the nucleolar accumulation of USP36 embryo derived cells, confirming a still molecular (Endo et al., 2009a). Therefore, in the absence of undefined role for NPM in controlling centrosome NPM, there is an increased poly-ubiquitination of duplication and/or mitotic spindle formation. Indeed, nucleolar fibrillarin (another USP36 target) thus sug- the observed alterations are also compatible with a gesting that NPM, through USP36, might control the specific role of NPM in either (or both) of the two level of protein ubiquitination in the nucleolus. pathways.

Poly-(ADP-ribosyl)ation NPM binds to Poly-(ADP-ribose) polymerases 1 and 2 NPM and cancer (PARP1 and 2) (Meder et al., 2005) and is poly-(ADP- ribosyl)ated following ionizing radiation (Ramsamooj Many reports correlate NPM to cancer pathogenesis. et al., 1995), although it is not clear if PARP1 and/or This may not be surprising, as many of the activities 2 catalyze the reaction. Although PARP1 has been ascribed to NPM can be related to what is currently

Oncogene Nucleophosmin and its complex network E Colombo et al 2601 known about tumor development. However, the specific tion) imposed by oncogenes, and it reduces the contribution of NPM is far from being clearly defined, oncogene-dependent apoptotic or senescence response. as most of the conclusions drawn so far are still This result is even more evident in MEFs defective prevalently based on indirect evidence using in vitro in DNA repair proteins such as ATM and FANCC models, whereas in vivo data on suitable mouse models (Li et al., 2007). of carcinogenesis are still missing. In particular, an open On the other hand, Grisendi et al., (2005) have shown issue in the field remains: is NPM a proto-oncogene or a that Em-myc mice in an Npm þ /À background, which tumor suppressor gene? As it will become clear, NPM expresses lower NPM levels, show a significant accel- behaves both ways and, at the moment, it is still difficult eration in lymphoma onset. Moreover, Npm þ /À to reconcile all the available data. animals may develop hematological neoplasms late in NPM overexpression has been reported in carcinomas life (with higher incidence of myeloid malignancies) of different origin: ovarian, prostate, colon and bladder significantly more frequently that their Npm þ / þ (Yung, 2007) and often correlates with mitotic index and counterparts, suggesting that NPM is a haploinsufficient proliferation rate. In bladder cancer, NPM overexpres- tumor suppressor for the hematopoietic compartment sion is a prognostic marker for recurrence and progres- (Sportoletti et al., 2008). They also showed that Npm þ /À sion (Yung, 2007). The correlation between increased MEFs display a slower growth rate compared with wild- NPM levels and cancer has been mainly linked to the type MEFs between passages 2 and 6, after which they role of NPM in ribosome biogenesis and protein overcome senescence, become immortalized and grow synthesis, functions that are both increased in cancer faster than WT controls, show increased genomic cells (Roussel and Hernandez-Verdun, 1994). Amplifica- instability and, when infected with oncogene expressing tion of the NPM1 locus has never been reported in viruses, are transformed at higher frequency. Consis- cancer specimens; however, as NPM1 is a c-Myc target tently, Npm þ /À MEFs overexpressing oncogenes at gene its overexpression may be the result of increased lower passage, show decreased cellular transformation c-Myc activity. Moreover, there are no studies that in vitro compared with Npm þ / þ MEFs. investigate if NPM is mutated in these tumors, or if All these data suggest a model in which NPM can there are alterations in the post-translational modifica- contribute to tumor development in different ways, tions of the protein that may influence its activity, depending on the time when NPM haploinsufficiency or localization and stability. No transgenic models of NPM oncogene activation arises in the cell (Figure 2). If overexpression are available to date that could provide oncogene activation arises as the first genetic event in a in vivo evidence that NPM overexpression is per se normal cell, the consequent accumulation of DNA enough to trigger the tumorigenic process. damage should increase the chance to select mutations NPM overexpression in immortalized cell lines leads leading to DDR escape, before DDR itself induces cell to increased proliferation and transformation (Yung, death or senescence (Bartkova et al., 2006; Di Micco 2007), and NPM cooperates with oncogenes in inducing et al., 2006) (Figure 2a). NPM may favor this process by cellular transformation in MEFs (Li et al., 2007, 2008). keeping the level of damage and DDR response below a Interestingly, overexpressed NPM strongly reduces the threshold that cells can tolerate, therefore enabling the level of DNA damage (measured as g-H2AX accumula- selection of the necessary cooperating mutation. In light

DNA damage response increased survival

high transformation oncogene NPM increased activation level proliferation normal NPM level

senescence/apoptosis

Increased damaged DNA/ genomic instability

senescence

loss DDR NPM transformation allele oncogene activation

escape/immortalization Figure 2 Relative contribution of NPM1 expression to cellular transformation. (a) Oncogene activation in a normal cell leads to unscheduled DNA replication and DNA damage. Cellular check points are activated favoring apoptosis or senescence as response to uncontrolled proliferative signals. The presence of high NPM levels may limit DNA damage and DNA damage response, and therefore support cell survival and eventually transformation. (b) Reduced NPM1 levels due to loss of one WT allele leads to genomic instability and increased DNA damage. As a consequence, the DNA damage response blocks cellular proliferation. Few cells may escape the block and establish an immortalized clone prone to be transformed by oncogene activation.

Oncogene Nucleophosmin and its complex network E Colombo et al 2602 of this, it is not surprising that NPM overexpression promyelocytic leukemia (Redner et al., 1996). In the is more powerful in MEFs defective for DNA repair resulting fusion protein, the N-terminal portion of NPM proteins. On the other hand, if NPM haploinsufficiency (first 117aa) is fused to the C-terminal portion of the arises first, cells will slowly accumulate genomic retinoic receptor alpha (RARa), including its DNA instability, and thus the surviving cells would have binding domain. Also in this case, NPM provides the already selected for mutations that escape the DDR homo-dimerization interface required for RARa activity pathway. The activated oncogene would not find any and, similarly to the most frequent fusion protein found barrier to transformation in these cells, and they could in acute promyelocytic leukemia, PML-RARa, the be immediately transformed rendering NPM over- NPM-RARa chimera recruits co-repressor proteins that expression dispensable (Figure 2b). interfere with RARa-dependent transcriptional activities Loss of the second Npm1 allele is never selected in (Redner et al., 2000). The consequence is a block of the Npm þ /À animals that develop hematological tumors, myeloid differentiation program that can be released by suggesting that either loss of the second allele is not treatment with pharmacological doses of all-trans required for transformation or that tumor cells cannot retinoic acid (Okazuka et al., 2007). tolerate the complete loss of NPM expression. The latter The t(3;5)(q25;q35) is the third translocation invol- may be related to NPM activities required for cancer ving the NPM1 gene and has been isolated in o1% of cells to proliferate and survive, such as, for example, acute myeloid leukemia (AML) (Yoneda-Kato et al., mitosis and ribosome biogenesis or anti-apoptotic and 1996). The resulting fusion protein contains a larger pro-survival functions. This may explain why depletion portion of the N terminus of NPM (first 175aa) fused to of NPM or use of a truncated dominant negative NPM virtually the entire coding sequence of a previously protein can sensitize cancer cells to apoptosis (Li et al., unknown protein, MLF1 (myelodysplasia/myeloid leu- 2005; Zhou et al., 2008). Although these results open the kemia factor 1). Little is known about the biological fascinating possibility of using NPM as a target to activity of the NPM-MLF1 fusion protein, but it is induce cancer cell death (Jian et al., 2009), it must be capable of transforming MEFs in cooperation with the considered that a partial targeting of NPM in normal oncogenic mutant RASV12 (Yoneda-Kato and Kato, tissues may trigger or favor the initial stages of tumor 2008). The wild-type MLF1 protein is involved in p53 development. stabilization (Yoneda-Kato et al., 2005), an activity that correlates to its nuclear localization. Interestingly, in the presence of NPM-MLF1, the WT MLF1 protein is Role of NPM in human hematological malignancies partially retained in the nucleolus. It is worth noting that a proportion of AML/myelodysplastic syndromes The NPM1 gene is frequently involved in chromosomal (MDS) cases bearing the t(3;5) rearrangement show translocations or mutations, invariably restricted to the breakpoints located outside the NPM1 and MLF1 hematopoietic compartment and associated to the genes, with a deletion in chromosome 5 that includes development of leukemias or lymphomas (Falini et al., the NPM1 locus, similarly to that observed in myelo- 2007). Three chromosomal translocations involving dysplastic syndromes with 5q deletion (Berger et al., NPM1 have been identified so far. The t(2;5) transloca- 2006). These data support the concept that NPM tion, found in B30% of anaplastic large cell lymphomas haploinsufficiency per se may cause myeloproliferative (Morris et al., 1994), fuses the 50end of the NPM1 gene neoplasms, as observed in Npm þ /À mice (Grisendi on chromosome 5 (encoding for the first 117aa of the et al., 2005). NPM protein) to the 30portion of the ALK locus on The contribution of the NPM portion to the chromosome 2, encoding the catalytic domain of ALK, oncogenic activity of these fusion proteins is still under a receptor tyrosine kinase normally expressed only in the investigation. At least in the case of NPM-RARa and nervous system. The result is the abnormal expression in NPM-ALK, the fact that the same tumor type is also the lymphoid compartment of a fusion protein (NPM- associated to other translocations in which the NPM ALK) in which the NPM homo-dimerization interface portion is substituted by other genes, indicates that the leads to constitutive activation of the ALK kinase. ALK and RARa moieties are indispensable for tumor NPM-ALK can associate with many different adaptor formation whereas NPM may serve to provide a proteins normally involved in transducing tyrosine homo-dimerization interface plus a strong promoter to kinase receptor signals, such as SHC, Grb2 and IRS-1, drive high levels of expression of the chimera in the thereby activating intracellular pathways such as RAS- hematopoietic compartment. However, it must be MAPK, JAK-STAT, PI3K-Akt and PLCg, that in turn considered that these translocations also lead to a promote cell proliferation and provide anti-apoptotic situation of haploinsufficiency for the WT NPM1 gene signals (Amin and Lai, 2007). Therefore, constitutive and, moreover, all three fusion proteins bind to the WT ALK activation is a key feature for the oncogenic NPM protein altering its cellular localization (Naoe activity of the NPM-ALK fusion protein. Accordingly, et al., 2006). Interestingly, in vivo experiments of mice NPM-ALK transgenic mice develop T-cell tumors bone marrow transplantation with hematopoietic pro- resembling human anaplastic large cell lymphomas genitor cells infected either with viruses expressing (Chiarle et al., 2003). NPM-ALK or the variant H4-ALK (in which the The t(5;17)(q35;q21) translocation is a rare variant, NPM portion has been substituted with an inert isolated so far in only four cases of human acute dimerization interface), showed that both constructs

Oncogene Nucleophosmin and its complex network E Colombo et al 2603 led to lymphoma onset with the same frequency a portion of AML blasts, suggesting they represent (unpublished observations). However, in the case of secondary events (Haferlach et al., 2009). Accordingly, NPM-ALK, the latency was significantly reduced, thus no differences have been found in terms of clinical or suggesting that NPM may contribute to reducing the biological behavior between NPMc þ AML with interval required for the second cooperating hit to be normal karyotype or those bearing additional chromo- selected, and strengthening the putative role of NPM in somal abnormalities (Alcalay et al., 2005; Haferlach controlling genome integrity. et al., 2009). Perhaps the most suggestive indication that NPM On the other hand, a significant correlation has been activity is involved in tumor pathogenesis comes from reported between NPM mutations and another frequent the discovery that B35% of AML cases bear a mutation genetic abnormality found in AML: FLT3-ITD (internal in one of the two NPM1 alleles (Falini et al., 2005). tandem duplication of the fms-related tyrosine kinase 3 A constant immunohistochemical hallmark of AML gene) (Falini et al., 2005). NPM1 mutations likely with mutated NPM1 is the cytoplasmic NPM staining in precede the appearance of FLT3-ITD, further suggest- bone marrow-derived tumor specimens, leading to the ing they represent a primary genetic event. Interestingly, general definition of NPMc þ (cytoplasmic positive) NPMc þ AML that are FLT3-ITD negative have a AML (Falini et al., 2006b). With very few exceptions better prognosis due to higher remission rate after (Mariano et al., 2006; Albiero et al., 2007), virtually all chemotherapy (Dohner et al., 2005; Schnittger et al., the NPM1 mutations thus far described occur in the 2005; Suzuki et al., 2005; Thiede et al., 2006; Mrozek last exon of the NPM1 gene, and lead to the de novo et al., 2007; Schlenk et al., 2008). The molecular basis of formation of a new nuclear export signal in the C- this increase in response to therapy is not completely terminal part of the NPM mutated protein (Figure 1), elucidated. One possible explanation may involve which alters the normal balance in nuclear-cytoplasmic nuclear factor-kB, which is frequently upregulated in NPM shuttling and causes the characteristic cytoplasmic AML blasts (Guzman et al., 2001) and is involved in localization (Mariano et al., 2006; Falini et al., 2006a; inducing pharmacological resistance (Frelin et al., 2005) Bolli et al., 2007). by inhibiting apoptosis and promoting cell survival. The NPM1 gene mutations represent the most frequent mutant NPM delocalizes and inactivates nuclear factor- genetic alterations found in adult AML, whereas they kB, therefore increasing chemosensitivity (Cilloni et al., are much less represented in childhood AML (Cazzaniga 2008). Recently, Del Poeta et al. (2010) have shown that et al., 2005; Brown et al., 2007). NPMc þ AML have NPMc þ FLT3-ITD negative AMLs have a higher been found in all morphological FAB subtypes, except transcript ratio between the anti-apoptotic BAX and the for acute promyelocytic leukemia, and increasing pro-apoptotic BCL2 mRNAs, which may sensitize evidence suggests that they may represent a distinct AML blasts to apoptosis. AML subtype with common pathological, immuno- A further indication that NPMc þ AML may phenotypic and prognostic features. NPMc þ AMLs represent a new clinical entity comes from gene are frequently CD34 negative (Taussig et al., 2010) and expression profile studies. AML with mutated NPM1 they often show multi-lineage involvement (Pasqualucci show a specific gene expression signature characterized et al., 2006). However, it has been recently shown that by increased expression of homeo-box genes and other the rare CD34 þ blasts purified from these patients can genes involved in stem cell maintenance (Alcalay et al., engraft in immunocompromised mice and support the 2005; Verhaak et al., 2005; Wilson et al., 2006). development of a leukemia that recapitulates the Accordingly, a specific miRNA profile linked to the original patient’s disease (Martelli et al., 2010). NPM1 regulation of the HOX gene expression is also associated mutations arise, in most cases, in primary AML, they to NPMc þ leukemia (Garzon et al., 2008; Jongen- are not present in other tumor types (Falini et al., 2005), Lavrencic et al., 2008). and they have rarely been found in myeloproliferative or All the available data on NPMc þ AML point at a myelodysplastic disorders (Caudill et al., 2006; Oki crucial role of the NPM mutated protein in disease et al., 2006; Zhang et al., 2007) or in secondary AML. In development. Although a formal proof that mutant the latter cases, it cannot be excluded that they indicate NPM possesses oncogenic potential in vivo is still a de novo AML following chemotherapy (Andersen missing, a number of observations in vitro strongly et al., 2008; Falini, 2008). NPM1 mutations are always support this hypothesis. NPMc þ causes the abnormal mutually exclusive with other known recurrent genomic cytoplasmic localization of several NPM interactors, abnormalities associated to AML pathogenesis, such as leading to their inactivation and in some cases favoring t(15;17), t(8;21), t(6;9), inv(16) (Falini et al., 2008), and their degradation (den Besten et al., 2005; Colombo they are significantly associated with a normal karyo- et al., 2006; Bonetti et al., 2008; Cilloni et al., 2008; type (about 85% of NPMc þ AML). NPM1 mutations Gurumurthy et al., 2008; Wanzel et al., 2008). are often the only genetic abnormality detected in AML In particular, NPMc þ inactivates relevant tumor sup- blasts and are very stable during the course of the pressor proteins including Miz1, ARF and Fbw7g.Fbw7g disease (Meloni et al., 2008), therefore likely represent- is an E3 ubiquitin ligase involved in c-Myc degradation, ing a reliable marker for monitoring minimal residual therefore its inactivation leads to the overexpression of disease (Schnittger et al., 2009). Only 15% of NPMc þ c-Myc in cells expressing NPMc þ . Miz1 is a transcrip- AML show a complex karyotype, although in these tion factor that activates the expression of two cases chromosomal alterations are often present only in important cell cycle inhibitors, p15Ink4b and p21.

Oncogene Nucleophosmin and its complex network E Colombo et al 2604 NPMc þ delocalizes Miz1 to the cytoplasm, favoring normal karyotype AML and for complete remission uncontrolled cell cycle progression. Finally, ARF in the after intensive induction therapy. presence of NPMc þ is delocalized to the cytoplasm too Interestingly, the retrospective analysis of a cohort of and degraded, thus blocking the ARF-dependent p53 elderly AML patients treated with all-trans retinoic acid stabilization that would be induced by c-Myc over- together with conventional chemotherapy showed a expression (Zindy et al., 1998). Based on these data, beneficial effect in patients with mutant NPM (Schlenk NPMc þ is an oncoprotein that can simultaneously lead et al., 2009). These data were, however, not confirmed to oncogene overexpression (c-Myc) and deactivate when studying younger patients (Burnett et al., 2009). physiological cellular barriers (ARF, p21, p15Ink4b) The molecular basis of this response remains elusive, that would limit uncontrolled proliferation (Schmitt but it may correlate with the reported role of NPM as et al., 1999), two key events in tumor development. co-repressor of AP2a during retinoic acid-induced Interestingly, NPMc þ overexpression in MEFs, either differentiation, including the NPM1 promoter itself WT, p53À/À or ARFÀ/À, is not sufficient to induce (Liu et al., 2007a). transformation in vitro. Moreover, NPMc þ does not These findings further support the idea that mutant cooperate with RasV12 in inducing transformation, NPM may represent a good therapeutic target. The whereas it does cooperate with E1A, which exerts main problem with this approach is its specificity, its transforming activity by inactivating the pRb onco- considering that the mutant is very similar to the wild- suppressor protein, thus suggesting that NPMc þ type protein. The mutated C-terminal region seems to be may indirectly activate the Rb pathway (Cheng et al., unfolded compared with the wild type (Grummitt et al., 2007). 2008), although it is not yet clear if this region may As only one allele of NPM is mutated, all NPMc þ influence the overall structure of NPM and, in AML are characterized by haploinsufficient expression particular, the interaction between NPM mutant and of the WT NPM protein; in addition, NPMc þ interacts other proteins like ARF. It will be important in the with WT NPM through their homo-dimerization inter- future to define the 3D structure of wild type versus faces and it can alter its localization (Falini et al., 2006a; mutant NPM bound to ARF in order to understand if it Bolli et al., 2007). The extent of cytoplasmic delocaliza- is possible to design small molecules that specifically tion of WT NPM in in vitro experiments depends on the block the interaction between NPM mutant and ARF. expression levels of NPMc þ , whereas the situation Alternatively, it may be easier to treat patients with in vivo is less clear. In fact, using an antibody that compounds like nutlins that mimic ARF activity recognizes both WT and mutated NPM in AML bone inhibiting the HDM2-p53 interaction (Vassilev et al., marrow specimens, Bolli et al., (2009) found a high level 2004). Indeed, lack of ARF activity in NPMc þ AML of heterogeneity among samples. Gruszka et al. (2010), correlates with the finding that p53 is rarely mutated using antibodies that specifically recognize either the (Suzuki et al., 2005) suggesting that there is no genetic mutant or the WT NPM protein showed that, in pressure on the ARF-HDM2-p53 pathway. Accord- NPMc þ AML blasts, a sizable fraction of mutant ingly, it should be possible to activate p53 in blasts and WT NPM proteins remains in separate cellular using DNA damaging drugs and this may explain compartments (cytoplasmic and nuclear/nucleolar, re- why NPMc þ leukemia are generally very sensitive to spectively). However, it remains unclear whether partial chemotherapy. delocalization of the WT NPM protein has a crucial role Another possible approach is to target the abnormal in the pathogenic activity of the NPMc þ protein. In nuclear export of NPMc þ . It has been shown in vitro particular, we cannot exclude that NPMc þ not only that it is possible to relocate NPMc þ and ARF in the acts as a dominant-negative protein over WT NPM, but nucleus (but not in the nucleolus) using inhibitors of has also acquired new functions and/or has ‘super- proteins involved in nuclear export such as exportin-1 physiological’ activities that correlate to its increased (CRM1) like leptomycin B (Fukuda et al., 1997), and nuclear export capacity. For example, NPMc þ may that this relocation is enough to stabilize ARF affect export of pre-ribosomal particles, regulation of (Colombo et al., 2006). However, it is difficult to centrosomal duplication or other molecularly undefined understand the long-term effect of this relocation, as mitotic activities. In this respect, it is worth noting that leptomycin B is per se toxic and, indeed, cannot be used NPMc þ AML has, in most cases, a normal karyotype, for in vivo treatment (Kau and Silver, 2003). Interest- thus indicating that the activity of NPM in preserving ingly, alternative less toxic semi-synthetic leptomycin B chromosomal stability is maintained in these tumors. derivatives have been selected (Mutka et al., 2009). High-throughput, small molecule screens using mislo- calized proteins as read-out have proved to be a Concluding remarks and future perspectives powerful tool for the selection of chemicals that inhibit yet unknown pathways involved in the trafficking of Besides the ongoing efforts to define the molecular basis defined delocalized molecules; these compounds have of the pathogenic activity of NPMc þ , at the clinical more chances to be specific, and hopefully less toxic level it already represents an important prognostic and (Kau et al., 2003). predictive marker in AML. The presence of NPM1 Last, but not least, a mouse model of NPMc þ - mutations (with a wild type FLT3 gene) is a favorable dependent AML still needs to be created. Cheng and marker for relapse free survival and over-all survival in et al. have recently reported the generation of a

Oncogene Nucleophosmin and its complex network E Colombo et al 2605 transgenic mouse expressing the NPMc þ complemen- mutations, like FLT3-ITD or, as shown more recently, tary DNA under the control of the myeloid-specific DNMT3A (Ley et al., 2010). Therefore, more efforts human MRP8 promoter. These mice show increased need to be engaged in this direction, not only because myeloproliferation in bone marrow and spleen progres- this model will definitely prove the tumorigenic activity sively in life, but they do not develop overt AML (Cheng of NPMc þ but, more importantly, because it will et al., 2009). However, these data demonstrated for the provide a valuable tool to verify the in vivo efficacy of first time that mutant NPM can increase the prolifera- new therapeutic approaches. tive potential of myeloid cells. This aspect was confirmed also in the zebrafish model system, in which expression of NPMc þ causes an embryonic expansion Conflict of interest of primitive myeloid and definitive hematopoietic cells (Bolli et al., 2010). Although these models provide The authors declare no conflict of interest. evidence concerning the capability of NPMc þ to perturb normal hematopoiesis, they still do not provide a definitive NPMc þ dependent AML model system. Acknowledgements This failure may be ascribed to problems in reaching the correct expression levels, as NPMc þ has to compete This work was supported by grants from AIRC, EC and with WT NPM, and/or to the lack of cooperative Ministero della Salute to PGP, EC and MA.

References

Adon AM, Zeng X, Harrison MK, Sannem S, Kiyokawa H, Kaldis P Bolli N, Nicoletti I, De Marco MF, Bigerna B, Pucciarini A, Mannucci et al. (2009). Cdk2 and Cdk4 regulate the centrosome cycle and are R et al. (2007). Born to be exported: COOH-terminal nuclear export critical mediators of centrosome amplification in p53-null cells. signals of different strength ensure cytoplasmic accumulation of Mol Cell Biol 30: 694–710. nucleophosmin leukemic mutants. Cancer Res 67: 6230–6237. Ahn JY, Liu X, Cheng D, Peng J, Chan PK, Wade PA et al. (2005). Bolli N, Payne EM, Grabher C, Lee JS, Johnston AB, Falini B et al. Nucleophosmin/B23, a nuclear PI(3,4,5)P(3) receptor, mediates (2010). Expression of the cytoplasmic NPM1 mutant (NPMc+) the antiapoptotic actions of NGF by inhibiting CAD. Mol Cell 18: causes the expansion of hematopoietic cells in zebrafish. Blood 115: 435–445. 3329–3340. Albiero E, Madeo D, Bolli N, Giaretta I, Bona ED, Martelli MF et al. Bonetti P, Davoli T, Sironi C, Amati B, Pelicci PG, Colombo E. (2007). Identification and functional characterization of a cytoplas- (2008). Nucleophosmin and its AML-associated mutant regulate mic nucleophosmin leukaemic mutant generated by a novel exon-11 c-Myc turnover through Fbw7 gamma. J Cell Biol 182: 19–26. NPM1 mutation. Leukemia 21: 1099–1103. Borer RA, Lehner CF, Eppenberger HM, Nigg EA. (1989). Major Alcalay M, Tiacci E, Bergomas R, Bigerna B, Venturini E, Minardi SP nucleolar proteins shuttle between nucleus and cytoplasm. Cell 56: et al. (2005). Acute myeloid leukemia bearing cytoplasmic 379–390. nucleophosmin (NPMc+ AML) shows a distinct gene expression Borggrefe T, Wabl M, Akhmedov AT, Jessberger R. (1998). A B-cell- profile characterized by up-regulation of genes involved in stem-cell specific DNA recombination complex. J Biol Chem 273: maintenance. Blood 106: 899–902. 17025–17035. Amin HM, Lai R. (2007). Pathobiology of ALK+ anaplastic large-cell Bothner B, Lewis WS, DiGiammarino EL, Weber JD, Bothner SJ, lymphoma. Blood 110: 2259–2267. Kriwacki RW. (2001). Defining the molecular basis of Arf and Amin MA, Matsunaga S, Uchiyama S, Fukui K. (2008). Nucleo- Hdm2 interactions. J Mol Biol 314: 263–277. phosmin is required for chromosome congression, proper mitotic Brady SN, Yu Y, Maggi Jr LB, Weber JD. (2004). ARF impedes spindle formation, and kinetochore-microtubule attachment in NPM/B23 shuttling in an Mdm2-sensitive tumor suppressor path- HeLa cells. FEBS Lett 582: 3839–3844. way. Mol Cell Biol 24: 9327–9338. Andersen MT, Andersen MK, Christiansen DH, Pedersen-Bjergaard J. Brown P, McIntyre E, Rau R, Meshinchi S, Lacayo N, Dahl G et al. (2008). NPM1 mutations in therapy-related acute myeloid leukemia (2007). The incidence and clinical significance of nucleophosmin with uncharacteristic features. Leukemia 22: 951–955. mutations in childhood AML. Blood 110: 979–985. Appella E, Anderson CW. (2001). Post-translational modifications Burnett AK, Hills RK, Green C, Jenkinson S, Koo K, Patel Y et al. and activation of p53 by genotoxic stresses. Eur J Biochem 268: (2009). The impact on outcome of the addition of all-trans retinoic 2764–2772. acid to intensive chemotherapy in younger patients with nonacute Bartkova J, Rezaei N, Liontos M, Karakaidos P, Kletsas D, Issaeva N promyelocytic acute myeloid leukemia: overall results and results in et al. (2006). Oncogene-induced senescence is part of the tumorigen- genotypic subgroups defined by mutations in NPM1, FLT3, and esis barrier imposed by DNA damage checkpoints. Nature 444: CEBPA. Blood 115: 948–956. 633–637. Caudill JS, Sternberg AJ, Li CY, Tefferi A, Lasho TL, Steensma DP. Berger R, Busson M, Baranger L, Helias C, Lessard M, (2006). C-terminal nucleophosmin mutations are uncommon in Dastugue N et al. (2006). Loss of the NPM1 gene in myeloid chronic myeloid disorders. Br J Haematol 133: 638–641. disorders with chromosome 5 rearrangements. Leukemia 20: Cazzaniga G, Dell’Oro MG, Mecucci C, Giarin E, Masetti R, 319–321. Rossi V et al. (2005). Nucleophosmin mutations in childhood Bertwistle D, Sugimoto M, Sherr CJ. (2004). Physical and functional acute myelogenous leukemia with normal karyotype. Blood 106: interactions of the Arf tumor suppressor protein with nucleophos- 1419–1422. min/B23. Mol Cell Biol 24: 985–996. Chang JH, Lin JY, Wu MH, Yung BY. (1998). Evidence for the Bolli N, De Marco MF, Martelli MP, Bigerna B, Pucciarini A, Rossi R ability of nucleophosmin/B23 to bind ATP. Biochem J 329(Part 3): et al. (2009). A dose-dependent tug of war involving the 539–544. NPM1 leukaemic mutant, nucleophosmin, and ARF. Leukemia Cheng K, Grisendi S, Clohessy JG, Majid S, Bernardi R, Sportoletti P 23: 501–509. et al. (2007). The leukemia-associated cytoplasmic nucleophosmin

Oncogene Nucleophosmin and its complex network E Colombo et al 2606 mutant is an oncogene with paradoxical functions: Arf inactivation Falini B. (2008). Therapy-related acute myeloid leukaemia with and induction of cellular senescence. Oncogene 26: 7391–7400. mutated NPM1: treatment induced or de novo in origin? Leukemia Cheng K, Sportoletti P, Ito K, Clohessy JG, Teruya-Feldstein J, 22: 891–892. Kutok JL et al. (2009). The cytoplasmic NPM mutant Falini B, Bolli N, Shan J, Martelli MP, Liso A, Pucciarini A et al. induces myeloproliferation in a transgenic mouse model. Blood (2006a). Both carboxy-terminus NES motif and mutated trypto- 115: 3341–3345. phan(s) are crucial for aberrant nuclear export of nucleophosmin Chiarle R, Gong JZ, Guasparri I, Pesci A, Cai J, Liu J et al. (2003). leukemic mutants in NPMc+ AML. Blood 107: 4514–4523. NPM-ALK transgenic mice spontaneously develop T-cell lympho- Falini B, Martelli MP, Bolli N, Bonasso R, Ghia E, Pallotta MT et al. mas and plasma cell tumors. Blood 101: 1919–1927. (2006b). Immunohistochemistry predicts nucleophosmin (NPM) Chou CC, Yung BY, Hsu CY. (2007). Involvement of nPKC-MAPK mutations in acute myeloid leukemia. Blood 108: 1999–2005. pathway in the decrease of nucleophosmin/B23 during megakar- Falini B, Mecucci C, Saglio G, Lo Coco F, Diverio D, Brown P et al. yocytic differentiation of human myelogenous leukemia K562 cells. (2008). NPM1 mutations and cytoplasmic nucleophosmin are Life Sci 80: 2051–2059. mutually exclusive of recurrent genetic abnormalities: a comparative Cilloni D, Messa F, Rosso V, Arruga F, Defilippi I, Carturan S et al. analysis of 2562 patients with acute myeloid leukemia. Haemato- (2008). Increase sensitivity to chemotherapeutical agents and logica 93: 439–442. cytoplasmatic interaction between NPM leukemic mutant and Falini B, Mecucci C, Tiacci E, Alcalay M, Rosati R, Pasqualucci L NF-kappaB in AML carrying NPM1 mutations. Leukemia 22: et al. (2005). Cytoplasmic nucleophosmin in acute myelogenous 1234–1240. leukemia with a normal karyotype. N Engl J Med 352: 254–266. Colombo E, Bonetti P, Lazzerini Denchi E, Martinelli P, Zamponi R, Falini B, Nicoletti I, Bolli N, Martelli MP, Liso A, Gorello P et al. Marine JC et al. (2005). Nucleophosmin is required for (2007). Translocations and mutations involving the nucleophosmin DNA integrity and p19Arf protein stability. Mol Cell Biol 25: (NPM1) gene in lymphomas and leukemias. Haematologica 92: 8874–8886. 519–532. Colombo E, Marine JC, Danovi D, Falini B, Pelicci PG. (2002). Foltz DR, Jansen LE, Bailey AO, Yates III JR, Bassett EA, Wood S Nucleophosmin regulates the stability and transcriptional activity of et al. (2009). Centromere-specific assembly of CENP-a nucleosomes p53. Nat Cell Biol 4: 529–533. is mediated by HJURP. Cell 137: 472–484. Colombo E, Martinelli P, Zamponi R, Shing DC, Bonetti P, Luzi L Foltz DR, Jansen LE, Black BE, Bailey AO, Yates III JR, Cleveland et al. (2006). Delocalization and destabilization of the Arf tumor DW. (2006). The human CENP-A centromeric nucleosome- suppressor by the leukemia-associated NPM mutant. Cancer Res 66: associated complex. Nat Cell Biol 8: 458–469. 3044–3050. Frehlick LJ, Eirin-Lopez JM, Ausio J. (2007). New insights into Dalenc F, Drouet J, Ader I, Delmas C, Rochaix P, Favre G et al. the nucleophosmin/nucleoplasmin family of nuclear chaperones. (2002). Increased expression of a COOH-truncated nucleophosmin Bioessays 29: 49–59. resulting from alternative splicing is associated with cellular Frelin C, Imbert V, Griessinger E, Peyron AC, Rochet N, Philip P resistance to ionizing radiation in HeLa cells. Int J Cancer 100: et al. (2005). Targeting NF-kappaB activation via pharmacologic 662–668. inhibition of IKK2-induced apoptosis of human acute myeloid Del Poeta G, Ammatuna E, Lavorgna S, Capelli G, Zaza S, Luciano F leukemia cells. Blood 105: 804–811. et al. (2010). The genotype nucleophosmin mutated and FLT3-ITD Fukuda M, Asano S, Nakamura T, Adachi M, Yoshida M, Yanagida negative is characterized by high bax/bcl-2 ratio and favourable M et al. (1997). CRM1 is responsible for intracellular transport outcome in acute myeloid leukaemia. Br J Haematol 149: 383–387. mediated by the nuclear export signal. Nature 390: 308–311. den Besten W, Kuo ML, Williams RT, Sherr CJ. (2005). Myeloid Gao H, Jin S, Song Y, Fu M, Wang M, Liu Z et al. (2005). B23 leukemia-associated nucleophosmin mutants perturb p53-dependent regulates GADD45a nuclear translocation and contributes to and independent activities of the Arf tumor suppressor protein. Cell GADD45a-induced cell cycle G2-M arrest. J Biol Chem 280: Cycle 4: 1593–1598. 10988–10996. Dhar SK, Lynn BC, Daosukho C, St Clair DK. (2004). Identification Garzon R, Garofalo M, Martelli MP, Briesewitz R, Wang L, of nucleophosmin as an NF-kappaB co-activator for the induction Fernandez-Cymering C et al. (2008). Distinctive microRNA of the human SOD2 gene. J Biol Chem 279: 28209–28219. signature of acute myeloid leukemia bearing cytoplasmic mutated Dhar SK, St Clair DK. (2009). Nucleophosmin blocks mitochondrial nucleophosmin. Proc Natl Acad Sci USA 105: 3945–3950. localization of p53 and apoptosis. J Biol Chem 284: 16409–16418. Grisendi S, Bernardi R, Rossi M, Cheng K, Khandker L, Manova K Di Micco R, Fumagalli M, Cicalese A, Piccinin S, Gasparini P, et al. (2005). Role of nucleophosmin in embryonic development and Luise C et al. (2006). Oncogene-induced senescence is a DNA tumorigenesis. Nature 437: 147–153. damage response triggered by DNA hyper-replication. Nature 444: Grummitt CG, Townsley FM, Johnson CM, Warren AJ, Bycroft M. 638–642. (2008). Structural consequences of nucleophosmin mutations in Dohner K, Schlenk RF, Habdank M, Scholl C, Rucker FG, acute myeloid leukemia. J Biol Chem 283: 23326–23332. Corbacioglu A et al. (2005). Mutant nucleophosmin (NPM1) Gruszka AM, Lavorgna S, Consalvo MI, Ottone T, Martinelli C, predicts favorable prognosis in younger adults with acute myeloid Cinquanta M et al. (2010). A monoclonal antibody against mutated leukemia and normal cytogenetics: interaction with other gene nucleophosmin 1 for the molecular diagnosis of acute myeloid mutations. Blood 106: 3740–3746. leukemias. Blood 116: 2096–2102. Du W, Zhou Y, Pike S, Pang Q. (2009). NPM phosphorylation Gurumurthy M, Tan CH, Ng R, Zeiger L, Lau J, Lee J et al. (2008). stimulates Cdk1, overrides G2/M checkpoint and increases leukemic Nucleophosmin interacts with HEXIM1 and regulates RNA blasts in mice. Carcinogenesis 31: 302–310. polymerase II transcription. J Mol Biol 378: 302–317. Eitoku M, Sato L, Senda T, Horikoshi M. (2008). Histone chaperones: Guzman ML, Neering SJ, Upchurch D, Grimes B, Howard DS, 30 years from isolation to elucidation of the mechanisms Rizzieri DA et al. (2001). Nuclear factor-kappaB is constitutively of nucleosome assembly and disassembly. Cell Mol Life Sci 65: activated in primitive human acute myelogenous leukemia cells. 414–444. Blood 98: 2301–2307. Endo A, Kitamura N, Komada M. (2009a). Nucleophosmin/B23 Haferlach C, Mecucci C, Schnittger S, Kohlmann A, Mancini M, regulates ubiquitin dynamics in nucleoli by recruiting deubiquitylat- Cuneo A et al. (2009). AML with mutated NPM1 carrying a ing enzyme USP36. J Biol Chem 284: 27918–27923. normal or aberrant karyotype show overlapping biological, Endo A, Matsumoto M, Inada T, Yamamoto A, Nakayama KI, pathological, immunophenotypic, and prognostic features. Blood Kitamura N et al. (2009b). Nucleolar structure and function are 114: 3024–3032. regulated by the deubiquitylating enzyme USP36. J Cell Sci 122: Haindl M, Harasim T, Eick D, Muller S. (2008). The nucleolar 678–686. SUMO-specific protease SENP3 reverses SUMO modification of

Oncogene Nucleophosmin and its complex network E Colombo et al 2607 nucleophosmin and is required for rRNA processing. EMBO Rep 9: Li J, Sejas DP, Burma S, Chen DJ, Pang Q. (2007). Nucleophosmin 273–279. suppresses oncogene-induced apoptosis and senescence and en- Herrera JE, Savkur R, Olson MO. (1995). The ribonuclease activity of hances oncogenic cooperation in cells with genomic instability. nucleolar protein B23. Nucleic Acids Res 23: 3974–3979. Carcinogenesis 28: 1163–1170. Hingorani K, Szebeni A, Olson MO. (2000). Mapping the functional Li J, Sejas DP, Rani R, Koretsky T, Bagby GC, Pang Q. domains of nucleolar protein B23. J Biol Chem 275: 24451–24457. (2006). Nucleophosmin regulates cell cycle progression and stress Hsu CY, Yung BY. (2000). Over-expression of nucleophosmin/B23 response in hematopoietic stem/progenitor cells. J Biol Chem 281: decreases the susceptibility of human leukemia HL-60 cells to 16536–16545. retinoic acid-induced differentiation and apoptosis. Int J Cancer 88: Li J, Zhang X, Sejas DP, Bagby GC, Pang Q. (2004). Hypoxia-induced 392–400. nucleophosmin protects cell death through inhibition of p53. J Biol Itahana K, Bhat KP, Jin A, Itahana Y, Hawke D, Kobayashi R et al. Chem 279: 41275–41279. (2003). Tumor suppressor ARF degrades B23, a nucleolar protein Li J, Zhang X, Sejas DP, Pang Q. (2005). Negative regulation of p53 involved in ribosome biogenesis and cell proliferation. Mol Cell 12: by nucleophosmin antagonizes stress-induced apoptosis in human 1151–1164. normal and malignant hematopoietic cells. Leuk Res 29: 1415–1423. Jian Y, Gao Z, Sun J, Shen Q, Feng F, Jing Y et al. (2009). RNA Li Z, Boone D, Hann SR. (2008). Nucleophosmin interacts directly aptamers interfering with nucleophosmin oligomerization induce with c-Myc and controls c-Myc-induced hyperproliferation and apoptosis of cancer cells. Oncogene 28: 4201–4211. transformation. Proc Natl Acad Sci USA 105: 18794–18799. Jongen-Lavrencic M, Sun SM, Dijkstra MK, Valk PJ, Lowenberg B. Liu H, Huang J, Wang J, Jiang S, Bailey AS, Goldman DC et al. (2008). MicroRNA expression profiling in relation to the genetic (2008). Transvection mediated by the translocated cyclin D1 locus in heterogeneity of acute myeloid leukemia. Blood 111: 5078–5085. mantle cell lymphoma. J Exp Med 205: 1843–1858. Joukov V, Groen AC, Prokhorova T, Gerson R, White E, Rodriguez Liu H, Tan BC, Tseng KH, Chuang CP, Yeh CW, Chen KD et al. A et al. (2006). The BRCA1/BARD1 heterodimer modulates (2007a). Nucleophosmin acts as a novel AP2alpha-binding tran- ran-dependent mitotic spindle assembly. Cell 127: 539–552. scriptional corepressor during cell differentiation. EMBO Rep 8: Kau TR, Schroeder F, Ramaswamy S, Wojciechowski CL, Zhao JJ, 394–400. Roberts TM et al. (2003). A chemical genetic screen identifies Liu X, Liu Z, Jang SW, Ma Z, Shinmura K, Kang S et al. (2007b). inhibitors of regulated nuclear export of a Forkhead transcription Sumoylation of nucleophosmin/B23 regulates its subcellular locali- factor in PTEN-deficient tumor cells. Cancer Cell 4: 463–476. zation, mediating cell proliferation and survival. Proc Natl Acad Sci Kau TR, Silver PA. (2003). Nuclear transport as a target for cell USA 104: 9679–9684. growth. Drug Discov Today 8: 78–85. Llanos S, Clark PA, Rowe J, Peters G. (2001). Stabilization of p53 by Kerr LE, Birse-Archbold JL, Short DM, McGregor AL, Heron I, p14ARF without relocation of MDM2 to the nucleolus. Nat Cell Macdonald DC et al. (2007). Nucleophosmin is a novel Biol 3: 445–452. Bax chaperone that regulates apoptotic cell death. Oncogene 26: Lowe SW, Sherr CJ. (2003). Tumor suppression by Ink4a-Arf: 2554–2562. progress and puzzles. Curr Opin Genet Dev 13: 77–83. Koike A, Nishikawa H, Wu W, Okada Y, Venkitaraman AR, Ohta T. Maggi Jr LB, Kuchenruether M, Dadey DY, Schwope RM, Grisendi (2010). Recruitment of phosphorylated NPM1 to sites of DNA S, Townsend RR et al. (2008). Nucleophosmin serves as a rate- damage through RNF8-dependent ubiquitin conjugates. Cancer Res limiting nuclear export chaperone for the mammalian ribosome. 70: 6746–6756. Mol Cell Biol 28: 7050–7065. Korgaonkar C, Hagen J, Tompkins V, Frazier AA, Allamargot C, Maiguel DA, Jones L, Chakravarty D, Yang C, Carrier F. (2004). Quelle FW et al. (2005). Nucleophosmin (B23) targets ARF to Nucleophosmin sets a threshold for p53 response to UV radiation. nucleoli and inhibits its function. Mol Cell Biol 25: 1258–1271. Mol Cell Biol 24: 3703–3711. Korgaonkar C, Zhao L, Modestou M, Quelle DE. (2002). ARF Mariano AR, Colombo E, Luzi L, Martinelli P, Volorio S, Bernard L function does not require p53 stabilization or Mdm2 relocalization. et al. (2006). Cytoplasmic localization of NPM in myeloid leukemias Mol Cell Biol 22: 196–206. is dictated by gain-of-function mutations that create a functional Kraus WL. (2008). Transcriptional control by PARP-1: chromatin nuclear export signal. Oncogene 25: 4376–4380. modulation, enhancer-binding, coregulation, and insulation. Curr Martelli MP, Pettirossi V, Thiede C, Bonifacio E, Mezzasoma F, Opin Cell Biol 20: 294–302. Cecchini D et al. (2010). CD34+ cells from AML with Krause A, Hoffmann I. (2010). Polo-like kinase 2-dependent mutated NPM1 harbor cytoplasmic mutated nucleophosmin phosphorylation of NPM/B23 on serine 4 triggers centriole and generate leukemia in immunocompromised mice. Blood 116: duplication. PLoS One 5: e9849. 3907–3922. Kuo ML, den Besten W, Bertwistle D, Roussel MF, Sherr CJ. (2004). Meder VS, Boeglin M, de Murcia G, Schreiber V. (2005). PARP-1 and N-terminal polyubiquitination and degradation of the Arf tumor PARP-2 interact with nucleophosmin/B23 and accumulate in suppressor. Genes Dev 18: 1862–1874. transcriptionally active nucleoli. J Cell Sci 118: 211–222. Kuo ML, den Besten W, Thomas MC, Sherr CJ. (2008). Arf-induced Meloni G, Mancini M, Gianfelici V, Martelli MP, Foa R, Falini B. turnover of the nucleolar nucleophosmin-associated SUMO-2/3 (2008). Late relapse of acute myeloid leukemia with mutated NPM1 protease Senp3. Cell Cycle 7: 3378–3387. after eight years: evidence of NPM1 mutation stability. Haemato- Kurki S, Peltonen K, Latonen L, Kiviharju TM, Ojala PM, Meek D logica 94: 298–300. et al. (2004). Nucleolar protein NPM interacts with HDM2 and Morris SW, Kirstein MN, Valentine MB, Dittmer KG, Shapiro DN, protects tumor suppressor protein p53 from HDM2-mediated Saltman DL et al. (1994). Fusion of a kinase gene, ALK, to a degradation. Cancer Cell 5: 465–475. nucleolar protein gene, NPM, in non-Hodgkin’s lymphoma. Science Lee SB, Xuan Nguyen TL, Choi JW, Lee KH, Cho SW, Liu Z et al. 263: 1281–1284. (2008). Nuclear Akt interacts with B23/NPM and protects it from Mrozek K, Dohner H, Bloomfield CD. (2007). Influence of new proteolytic cleavage, enhancing cell survival. Proc Natl Acad Sci molecular prognostic markers in patients with karyotypically USA 105: 16584–16589. normal acute myeloid leukemia: recent advances. Curr Opin Leotoing L, Meunier L, Manin M, Mauduit C, Decaussin M, Verrijdt Hematol 14: 106–114. G et al. (2008). Influence of nucleophosmin/B23 on DNA binding Mukudai Y, Kubota S, Kawaki H, Kondo S, Eguchi T, Sumiyoshi K and transcriptional activity of the androgen receptor in prostate et al. (2008). Posttranscriptional regulation of chicken ccn2 gene cancer cell. Oncogene 27: 2858–2867. expression by nucleophosmin/B23 during chondrocyte differentia- Ley TJ, Ding L, Walter MJ, McLellan MD, Lamprecht T, Larson DE tion. Mol Cell Biol 28: 6134–6147. et al. (2010). DNMT3A mutations in acute myeloid leukemia. Murano K, Okuwaki M, Hisaoka M, Nagata K. (2008). Transcription N Engl J Med 363: 2424–2433. regulation of the rRNA gene by a multifunctional nucleolar protein,

Oncogene Nucleophosmin and its complex network E Colombo et al 2608 B23/nucleophosmin, through its histone chaperone activity. Mol Roussel P, Hernandez-Verdun D. (1994). Identification of Ag-NOR Cell Biol 28: 3114–3126. proteins, markers of proliferation related to ribosomal gene activity. Mutka SC, Yang WQ, Dong SD, Ward SL, Craig DA, Timmermans Exp Cell Res 214: 465–472. PB et al. (2009). Identification of nuclear export inhibitors with Ruddon RW, Bedows E. (1997). Assisted protein folding. J Biol Chem potent anticancer activity in vivo. Cancer Res 69: 510–517. 272: 3125–3128. Naoe T, Suzuki T, Kiyoi H, Urano T. (2006). Nucleophosmin: a Sato K, Hayami R, Wu W, Nishikawa T, Nishikawa H, Okuda Y et al. versatile molecule associated with hematological malignancies. (2004). Nucleophosmin/B23 is a candidate substrate for the Cancer Sci 97: 963–969. BRCA1-BARD1 ubiquitin ligase. J Biol Chem 279: 30919–30922. Negi SS, Olson MO. (2006). Effects of interphase and mitotic Savkur RS, Olson MO. (1998). Preferential cleavage in pre-ribosomal phosphorylation on the mobility and location of nucleolar protein RNA byprotein B23 endoribonuclease. Nucleic Acids Res 26: B23. J Cell Sci 119: 3676–3685. 4508–4515. Nishida T, Yamada Y. (2008). SMT3IP1, a nucleolar SUMO-specific Schlenk RF, Dohner K, Kneba M, Gotze K, Hartmann F, Del Valle F protease, deconjugates SUMO-2 from nucleolar and cytoplasmic et al (2009). Gene mutations and response to treatment with all- nucleophosmin. Biochem Biophys Res Commun 374: 382–387. trans retinoic acid in elderly patients with acute myeloid leukemia. Nousiainen M, Sillje HH, Sauer G, Nigg EA, Korner R. (2006). Results from the AMLSG Trial AML HD98B. Haematologica 94: Phosphoproteome analysis of the human mitotic spindle. Proc Natl 54–60. Acad Sci USA 103: 5391–5396. Schlenk RF, Dohner K, Krauter J, Frohling S, Corbacioglu A, Okada M, Jang SW, Ye K. (2007). Ebp1 association with nucleo- Bullinger L et al. (2008). Mutations and treatment outcome in phosmin/B23 is essential for regulating cell proliferation and cytogenetically normal acute myeloid leukemia. N Engl J Med 358: suppressing apoptosis. J Biol Chem 282: 36744–36754. 1909–1918. Okazuka K, Masuko M, Seki Y, Hama H, Honma N, Furukawa T Schmitt CA, McCurrach ME, de Stanchina E, Wallace-Brodeur RR, et al. (2007). Successful all-trans retinoic acid treatment of acute Lowe SW. (1999). INK4a/ARF mutations accelerate lymphoma- promyelocytic leukemia in a patient with NPM/RAR fusion. Int J genesis and promote chemoresistance by disabling p53. Genes Dev Hematol 86: 246–249. 13: 2670–2677. Oki Y, Jelinek J, Beran M, Verstovsek S, Kantarjian HM, Issa JP. Schnittger S, Kern W, Tschulik C, Weiss T, Dicker F, Falini B et al. (2006). Mutations and promoter methylation status of NPM1 in (2009). Minimal residual disease levels assessed by NPM1 mutation myeloproliferative disorders. Haematologica 91: 1147–1148. specific RQ-PCR provide important prognostic information in Okuda M, Horn HF, Tarapore P, Tokuyama Y, Smulian AG, Chan AML. Blood 114: 2220–2231. PK et al. (2000). Nucleophosmin/B23 is a target of CDK2/cyclin E Schnittger S, Schoch C, Kern W, Mecucci C, Tschulik C, Martelli MF in centrosome duplication. Cell 103: 127–140. et al. (2005). Nucleophosmin gene mutations are predictors of Okuwaki M, Iwamatsu A, Tsujimoto M, Nagata K. (2001a). favorable prognosis in acute myelogenous leukemia with a normal Identification of nucleophosmin/B23, an acidic nucleolar karyotype. Blood 106: 3733–3739. protein, as a stimulatory factor for in vitro replication of adenovirus Shandilya J, Swaminathan V, Gadad SS, Choudhari R, Kodaganur DNA complexed with viral basic core proteins. J Mol Biol 311: GS, Kundu TK. (2009). Acetylated NPM1 localizes in the 41–55. nucleoplasm and regulates transcriptional activation of genes Okuwaki M, Matsumoto K, Tsujimoto M, Nagata K. (2001b). implicated in oral cancer manifestation. Mol Cell Biol 29: Function of nucleophosmin/B23, a nucleolar acidic protein, as a 5115–5127. histone chaperone. FEBS Lett 506: 272–276. Sportoletti P, Grisendi S, Majid SM, Cheng K, Clohessy JG, Viale A Okuwaki M, Tsujimoto M, Nagata K. (2002). The RNA binding et al. (2008). Npm1 is a haploinsufficient suppressor of myeloid and activity of a ribosome biogenesis factor, nucleophosmin/B23, is lymphoid malignancies in the mouse. Blood 111: 3859–3862. modulated by phosphorylation with a cell cycle-dependent kinase Sugimoto M, Kuo ML, Roussel MF, Sherr CJ. (2003). Nucleolar Arf and by association with its subtype. Mol Biol Cell 13: 2016–2030. tumor suppressor inhibits ribosomal RNA processing. Mol Cell 11: Palaniswamy V, Moraes KC, Wilusz CJ, Wilusz J. (2006). Nucleo- 415–424. phosmin is selectively deposited on mRNA during polyadenylation. Suzuki T, Kiyoi H, Ozeki K, Tomita A, Yamaji S, Suzuki R et al. Nat Struct Mol Biol 13: 429–435. (2005). Clinical characteristics and prognostic implications Pang Q, Christianson TA, Koretsky T, Carlson H, David L, Keeble W of NPM1 mutations in acute myeloid leukemia. Blood 106: et al. (2003). Nucleophosmin interacts with and inhibits the catalytic 2854–2861. function of eukaryotic initiation factor 2 kinase PKR. J Biol Chem Swaminathan V, Kishore AH, Febitha KK, Kundu TK. (2005). 278: 41709–41717. Human histone chaperone nucleophosmin enhances acetylation- Pasqualucci L, Liso A, Martelli MP, Bolli N, Pacini R, Tabarrini A dependent chromatin transcription. Mol Cell Biol 25: 7534–7545. et al. (2006). Mutated nucleophosmin detects clonal multilineage Szebeni A, Hingorani K, Negi S, Olson MO. (2003). Role of protein involvement in acute myeloid leukemia: impact on WHO classifica- kinase CK2 phosphorylation in the molecular chaperone activity of tion. Blood 108: 4146–4155. nucleolar protein b23. J Biol Chem 278: 9107–9115. Peter M, Nakagawa J, Doree M, Labbe JC, Nigg EA. (1990). Szebeni A, Mehrotra B, Baumann A, Adam SA, Wingfield PT, Olson Identification of major nucleolar proteins as candidate mitotic MO. (1997). Nucleolar protein B23 stimulates nuclear import of the substrates of cdc2 kinase. Cell 60: 791–801. HIV-1 rev protein and NLS-conjugated albumin. Biochemistry 36: Ramsamooj P, Notario V, Dritschilo A. (1995). Modification of 3941–3949. nucleolar protein B23 after exposure to ionizing radiation. Radiat Szebeni A, Olson MO. (1999). Nucleolar protein B23 has molecular Res 143: 158–164. chaperone activities. Protein Sci 8: 905–912. Redner RL, Chen JD, Rush EA, Li H, Pollock SL. (2000). The t(5;17) Tago K, Chiocca S, Sherr CJ. (2005). Sumoylation induced by the Arf acute promyelocytic leukemia fusion protein NPM-RAR interacts tumor suppressor: a p53-independent function. Proc Natl Acad Sci with co-repressor and co-activator proteins and exhibits USA 102: 7689–7694. both positive and negative transcriptional properties. Blood 95: Takemura M, Sato K, Nishio M, Akiyama T, Umekawa H, Yoshida 2683–2690. S. (1999). Nucleolar protein B23.1 binds to retinoblastoma protein Redner RL, Rush EA, Faas S, Rudert WA, Corey SJ. (1996). The and synergistically stimulates DNA polymerase alpha activity. t(5;17) variant of acute promyelocytic leukemia expresses a J Biochem 125: 904–909. nucleophosmin-retinoic acid receptor fusion. Blood 87: 882–886. Tarapore P, Shinmura K, Suzuki H, Tokuyama Y, Kim SH, Mayeda Rodway H, Llanos S, Rowe J, Peters G. (2004). Stability of nucleolar A et al. (2006). Thr199 phosphorylation targets nucleophosmin to versus non-nucleolar forms of human p14(ARF). Oncogene 23: nuclear speckles and represses pre-mRNA processing. FEBS Lett 6186–6192. 580: 399–409.

Oncogene Nucleophosmin and its complex network E Colombo et al 2609 Taussig DC, Vargaftig J, Miraki-Moud F, Griessinger E, Sharrock K, Wu MH, Chang JH, Yung BY. (2002b). Resistance to UV-induced Luke T et al. (2010). Leukemia-initiating cells from some acute cell-killing in nucleophosmin/B23 over-expressed NIH 3T3 fibro- myeloid leukemia patients with mutated nucleophosmin reside in blasts: enhancement of DNA repair and up-regulation of PCNA in the CD34(-) fraction. Blood 115: 1976–1984. association with nucleophosmin/B23 over-expression. Carcino- Thiede C, Koch S, Creutzig E, Steudel C, Illmer T, Schaich M et al. genesis 23: 93–100. (2006). Prevalence and prognostic impact of NPM1 mutations in Xiao J, Zhang Z, Chen GG, Zhang M, Ding Y, Fu J et al. (2009). 1485 adult patients with acute myeloid leukemia (AML). Blood 107: Nucleophosmin/B23 interacts with p21WAF1/CIP1 and contributes 4011–4020. to its stability. Cell Cycle 8: 889–895. Tokuyama Y, Horn HF, Kawamura K, Tarapore P, Fukasawa K. Yao J, Fu C, Ding X, Guo Z, Zenreski A, Chen Y et al. (2004). Nek2A (2001). Specific phosphorylation of nucleophosmin on Thr(199) by kinase regulates the localization of numatrin to centrosome in cyclin-dependent kinase 2-cyclin E and its role in centrosome mitosis. FEBS Lett 575: 112–118. duplication. J Biol Chem 276: 21529–21537. Yao Z, Duan S, Hou D, Wang W, Wang G, Liu Y et al. (2010). B23 Vassilev LT, Vu BT, Graves B, Carvajal D, Podlaski F, Filipovic Z acts as a nucleolar stress sensor and promotes cell survival through et al. (2004). in vivo activation of the p53 pathway by small-molecule its dynamic interaction with hnRNPU and hnRNPA1. Oncogene 29: antagonists of MDM2. Science 303: 844–848. 1821–1834. Ventura A, Kirsch DG, McLaughlin ME, Tuveson DA, Grimm J, Yoneda-Kato N, Kato JY. (2008). Shuttling imbalance of MLF1 Lintault L et al. (2007). Restoration of p53 function leads to tumour results in p53 instability and increases susceptibility to oncogenic regression in vivo. Nature 445: 661–665. transformation. Mol Cell Biol 28: 422–434. Verhaak RG, Goudswaard CS, van Putten W, Bijl MA, Sanders MA, Yoneda-Kato N, Look AT, Kirstein MN, Valentine MB, Raimondi Hugens W et al. (2005). Mutations in nucleophosmin (NPM1) SC, Cohen KJ et al. (1996). The t(3;5)(q25.1;q34) of myelodysplastic in acute myeloid leukemia (AML): association with other syndrome and acute myeloid leukemia produces a novel fusion gene, gene abnormalities and previously established gene expression NPM-MLF1. Oncogene 12: 265–275. signatures and their favorable prognostic significance. Blood 106: Yoneda-Kato N, Tomoda K, Umehara M, Arata Y, Kato JY. (2005). 3747–3754. Myeloid leukemia factor 1 regulates p53 by suppressing COP1 via Wang D, Umekawa H, Olson MO. (1993). Expression and subcellular COP9 signalosome subunit 3. EMBO J 24: 1739–1749. locations of two forms of nucleolar protein B23 in rat tissues and Yu Y, Maggi Jr LB, Brady SN, Apicelli AJ, Dai MS, Lu H et al. cells. Cell Mol Biol Res 39: 33–42. (2006). Nucleophosmin is essential for ribosomal protein L5 nuclear Wang G, Pan Y, Ahmad KA, Ahmed K. (2010). Protein B23/ export. Mol Cell Biol 26: 3798–3809. nucleophosmin/numatrin nuclear dynamics in relation to protein Yun C, Wang Y, Mukhopadhyay D, Backlund P, Kolli N, Yergey A kinase CK2 and apoptotic activity in prostate cells. Biochemistry 49: et al. (2008). Nucleolar protein B23/nucleophosmin regulates the 3842–3852. vertebrate SUMO pathway through SENP3 and SENP5 proteases. Wang W, Budhu A, Forgues M, Wang XW. (2005). Temporal and J Cell Biol 183: 589–595. spatial control of nucleophosmin by the Ran-Crm1 complex in Yung BY. (2007). Oncogenic role of nucleophosmin/B23. Chang Gung centrosome duplication. Nat Cell Biol 7: 823–830. Med J 30: 285–293. Wanzel M, Russ AC, Kleine-Kohlbrecher D, Colombo E, Pelicci PG, Yusufzai TM, Tagami H, Nakatani Y, Felsenfeld G. (2004). CTCF Eilers M. (2008). A ribosomal protein L23-nucleophosmin tethers an insulator to subnuclear sites, suggesting shared insulator circuit coordinates Mizl function with cell growth. Nat Cell Biol mechanisms across species. Mol Cell 13: 291–298. 10: 1051–1061. Zhang H, Shi X, Paddon H, Hampong M, Dai W, Pelech S. (2004). Weber JD, Jeffers JR, Rehg JE, Randle DH, Lozano G, Roussel MF B23/nucleophosmin serine 4 phosphorylation mediates mitotic et al. (2000). p53-independent functions of the p19(ARF) tumor functions of polo-like kinase 1. J Biol Chem 279: 35726–35734. suppressor. Genes Dev 14: 2358–2365. Zhang Y, Zhang M, Yang L, Xiao Z. (2007). NPM1 mutations in Weber JD, Taylor LJ, Roussel MF, Sherr CJ, Bar-Sagi D. (1999). myelodysplastic syndromes and acute myeloid leukemia with Nucleolar Arf sequesters Mdm2 and activates p53. Nat Cell Biol 1: normal karyotype. Leuk Res 31: 109–111. 20–26. Zhou Y, Du W, Koretsky T, Bagby GC, Pang Q. (2008). TAT- Wilson CS, Davidson GS, Martin SB, Andries E, Potter J, Harvey R mediated intracellular delivery of NPM-derived peptide induces et al. (2006). Gene expression profiling of adult acute myeloid apoptosis in leukemic cells and suppresses leukemogenesis in mice. leukemia identifies novel biologic clusters for risk classification and Blood 112: 2474–2483. outcome prediction. Blood 108: 685–696. Zindy F, Eischen CM, Randle DH, Kamijo T, Cleveland JL, Wu MH, Chang JH, Chou CC, Yung BY. (2002a). Involvement of Sherr CJ et al. (1998). Myc signaling via the ARF tumor suppressor nucleophosmin/B23 in the response of HeLa cells to UV irradiation. regulates p53-dependent apoptosis and immortalization. Genes Dev Int J Cancer 97: 297–305. 12: 2424–2433.

Oncogene