HDAC6, at the Crossroads Between Cytoskeleton and Cell Signaling by Acetylation and Ubiquitination

Oncogene (2007) 26, 5468–5476 & 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc REVIEW HDAC6, at the crossroads between cytoskeleton and cell signaling by acetylation and ubiquitination

C Boyault, K Sadoul, M Pabion and S Khochbin

INSERM, U823, Equipe Epige´ne´tique et Signalisation Cellulaire, Institut Albert Bonniot, Universite´ Joseph Fourier, Domaine de la Merci, Grenoble, La Tronche Cedex, France

Histone deacetylase 6 (HDAC6)is a unique enzyme with et al., 2005). Moreover, it is now clear that acetylation is specific structural and functional features. It is actively or not an exclusive modification of nuclear proteins, since stably maintained in the cytoplasm and is the only many cytoplasmic proteins, including a significant member, within the histone deacetylase family, that subset of mitochondrial proteins, have recently been harbors a full duplication of its deacetylase homology shown to bear lysine acetylation (Cohen et al., 2004; region followed by a specific ubiquitin-binding domain at Dihazi et al., 2005; Iwabata et al., 2005; Kovacs et al., the C-terminus end. Accordingly, this deacetylase func- 2005; Hallows et al., 2006; Kim et al., 2006; Schwer tions at the heart of a cellular regulatory mechanism et al., 2006). The regulation of these acetylations and capable of coordinating various cellular functions largely the determination of their functional significance relying on the microtubule network. Moreover, HDAC6 now constitute a real challenge for biologists. In fact, action as a regulator of the HSP90 chaperone activity while the list of cytoplasmic acetylated proteins is adds to the multifunctionality of the protein, and allows us rapidly growing, basic information on the involved to propose a critical role for HDAC6 in mediating and enzymatic machinery, HATs and HDACs, as well as on coordinating various cellular events in response to their functions, is still missing. different stressful stimuli. The identification of HDAC6 as the first HDAC Oncogene (2007) 26, 5468–5476; doi:10.1038/sj.onc.1210614 actively maintained in the cytoplasm (Verdel et al., 2000) opened the way for the identification of its sub- Keywords: microtubule; actin; HSP90; virus; aggresome; strates and cellular functions controlled by its catalytic transcription activity in this compartment. Further analyses of HDAC6 functions revealed the involvement of the protein in cellular processes dependent and independent of its catalytic activity. Indeed, the discovery of an Introduction ubiquitin-binding domain in HDAC6 (Seigneurin-Berny et al., 2001) led to the unraveling of its participation in Protein lysine acetylation is now emerging as a widely cellular functions depending on cell signaling through occurring post-translational modification and constitu- protein ubiquitination (Kawaguchi et al., 2003; Boyault tes a genuine cellular signaling system involved in the et al., 2006). HDAC6 therefore appears as a protein control of various functions in different cellular functioning at the crossroads between at least two compartments (reviewed in Kouzarides, 2000; Sterner cellular signaling systems, respectively, involving protein and Berger, 2000; Caron et al., 2003; Yang, 2004; lysine acetylation and ubiquitination. Glozak et al., 2005). Originally, lysine acetylation, HDAC6 is not the only cytoplasmic deacetylase. mainly that of histones and transcription factors, was In fact, early investigations showed that under specific regarded as a powerful mean of gene expression circumstances, other HDACs of different classes regulation. Rapid progress in the discovery and func- could also be found in the cytoplasm. However, their tional analysis of an increasing number of nuclear cytoplasmic functions remain elusive and except for enzymes involved in protein lysine acetylation and SIRT2, a class III HDAC member (North et al., 2003), deacetylation, histone acetyltransferases (HATs) and none of the HDAC6 activities seems to be shared by histone deacetylases (HDACs), respectively, have these enzymes (Hubbert et al., 2002; Matsuyama et al., demonstrated that these enzymes, independently of 2002). Among them, HDAC10 deserves a special atten- transcription, can control some of the basic cellular tion since it is the closest relative of HDAC6. Indeed, processes such as protein stability (reviewed in Caron the HDAC10 catalytic domain shows the highest homology to those of HDAC6 and, moreover, its pseudo-duplication is reminiscent of HDAC6 cata- Correspondence: Dr S Khochbin, INSERM, U823, Equipe Epige´ ne´ - lytic domain duplication. This characteristic of HDAC10 tique et Signalisation Cellulaire, Institut Albert Bonniot, Universite´ Joseph Fourier, Domaine de la Merci, Grenoble, La Tronche Cedex, justifies a close examination of the known HDAC10 France. activities in the light of data available on HDAC6 E-mail: [email protected] functions. HDAC6, a coordinator of cell responses to stressful stimuli C Boyault et al 5469 This review will therefore essentially consider the HDAC activity of HDAC6 relies either on the integrity important body of data now available on HDAC6 since of both HDAC domains (Zhang et al., 2003, 2006) or its discovery in 1999 (Grozinger et al., 1999; Verdel and is mediated by its second catalytic domain (Zou et al., Khochbin, 1999), aiming to draw a synthetic scheme of 2006). The discovery of a-tubulin as an HDAC6 sub- its cellular functions and highlight its possible involve- strate (Hubbert et al., 2002; Matsuyama et al., 2002; ment in human pathologies. Zhang et al., 2003) allowed researchers to also include acetylated tubulin in these assays (Haggarty et al., 2003; Zhang et al., 2003, 2006; Zou et al., 2006). Here again, some of the investigators found that the integrity Domain organization and structural features of HDAC6 of both HDAC6 catalytic domains was indispensable for the whole tubulin deacetylase (TDAC) activity of HDAC6 is the only member of the cellular deacetylase the enzyme (Zhang et al., 2003, 2006), whereas others family containing a full duplication of the large class I/II showed that the whole TDAC activity could be HDAC-homology domain (Figure 1), first revealed in attributed to the second domain (Haggarty et al., the mouse HDAC6 (Verdel and Khochbin, 1999) and 2003; Zou et al., 2006). then in its human ortholog (Grozinger et al., 1999). This Additional investigations revealed the existence of feature could be considered as a unique signature other structural requirements for an efficient HDAC allowing the identification of HDAC6 orthologs in and TDAC activity of HDAC6. Indeed, the spacer other species including invertebrates, Drosophila mela- region between the two catalytic domains of the protein nogaster and Caenorhabditis elegans as well as in plants was found to play a crucial role in the total activity of such as Arabidopsis thaliana (Barlow et al., 2001; HDAC6, and amino acid addition or deletion in this reviewed in Yang and Gregoire, 2005). region dramatically affected the TDAC activity of The conservation of the double catalytic domain HDAC6. The HDAC activity of HDAC6 seemed, organization during evolution strongly argues in favor however, to be less dependent on the length of this of a critical role for this domain duplication in HDAC6 linker region than its TDAC activity (Zhang et al., functions and prompted the investigators to uncover 2006). These studies also evidenced an important role its role in the whole deacetylase activity of the enzyme. for a conserved region present in HDACs and some One obvious way to tackle this issue consisted in the HATs, known as substrate recognition site or Esa1- measurement of the deacetylase activity of HDAC6 Rpd3 (ER) motif (Adachi et al., 2002). Both HDAC and mutants containing inactivating mutations in each TDAC activities of HDAC6 rely on this motif present in domain individually. Unexpectedly, these experiments each domain, but the ER in the second HDAC domain generated conflicting results. The first attempt showed has a particularly critical role in TDAC activity of the that each domain possessed an independent catalytic enzyme (Zhang et al., 2006). activity (Grozinger et al., 1999). This result was, Overall, despite these contradictory data, the second however, challenged by other groups showing that the domain of HDAC6 seems to have a determinant role in at least the total TDAC activity of the enzyme (Haggarty et al., 2003; Zhang et al., 2006; Zou et al., HDAC6 2006). NES SE14 In addition to HDAC6 catalytic activities, the control of its intra-cellular localization was also found to be an important issue in the understanding of its functions. DD1 DD2 ZnF-UBP Early investigations considering mouse HDAC6 showed that the protein is actively retained in the cytoplasm and Nuclear Cytoplasmic could, only under specific circumstances, be partially exclusion anchoring Catalytic activity found in the nucleus (Verdel et al., 2000). Indeed, a − α-tubulin strong nuclear export signal (NES) located N terminus - HSP90 Ubiquitin to the first catalytic domain prevents the accumulation binding of the protein in the nucleus (Figure 1). Interestingly, although this N-terminus NES is conserved in human Figure 1 Functional domain organization of HDAC6. HDAC6 is the only member of the histone deacetylase family containing HDAC6 (hHDAC6), another region of the protein was tandem catalytic domains. So far two in vivo substrates, a-tubulin found to ensure a stable anchorage of the protein in the and HSP90, have been identified. In human HDAC6, two different cytoplasm (Figure 1). This domain, which has not so far domains ensure a stable maintenance of the protein in the been found in other HDAC6 orthologs, is characterized cytoplasm. A conserved nuclear export signal (NES), functional by eight consecutive tetradecapeptide repeat motifs, in mouse and human HDAC6, mediates an active export of the protein from the nucleus. Only in human HDAC6, an additional named SE14 (Bertos et al., 2004). domain, named SE14, has been found to stably anchor the protein The multiplication of domains and motifs, such as in the cytoplasm. In the C-terminus part of the protein a domain, NES and the SE14, involved in the active and stable conserved in HDAC6 orthologs in many species, ZnF-UBP, maintenance of hHDAC6 in the cytoplasm, suggest that constitutes a high affinity ubiquitin-binding motif. DD1and DD2 stand for deacetylase domains 1and 2, respectively, SE14for during evolution a pressure appeared on HDAC6 to Ser-Glu-containing tetradecapeptide and ZnF-UBP for ubiquitin exclude the protein from the nucleus and to reinforce its C-terminus hydrolase-like zinc finger. cytoplasmic localization.

Oncogene HDAC6, a coordinator of cell responses to stressful stimuli C Boyault et al 5470 Another remarkable aspect of HDAC6 is the presence Transcriptional of a conserved cysteine- and histidine-rich domain in its response

C-terminus part (Figure 1), which is also present in a Proteasome group of ubiquitin-speciﬁc proteases (USP), named V

ZnF-UBP (Amerik et al., 2000; Seigneurin-Berny Ub et al., 2001). This domain has the particularity to VCP speciﬁcally bind mono- (Seigneurin-Berny et al., 2001; VCP HSP90 Boyault et al., 2006) and poly-ubiquitin chains (Hook Protein et al., 2002; Boyault et al., 2006). More detailed studies HDAC6

allowed for the modeling of the structure of this domain HDAC6 in HDAC6 and to predict its organization in three zinc ﬁngers. This particular organization of HDAC6 ZnF- UBP domain allows its binding to monomeric ubiquitin V with a measured Kd of 60 nM, which is the highest known afﬁnity for ubiquitin binding among all known Autophagy ubiquitin-interacting proteins (Boyault et al., 2006). Aggresomes

Ubiquitin-dependent functions of HDAC6 Figure 2 HDAC6 is an essential factor in coordinating the cell HDAC6 not only binds mono- and poly-ubiquitin responses to cytotoxic protein aggregate formation. The cellular chains, but it also forms a complex, at least in the concentration of HDAC6 and its partner, p97/VCP, determines the fate of poly-ubiquitinated proteins. An excess of HDAC6 favors cytosol of cells from mouse testis, with two proteins that the accumulation of ubiquitinated, and mainly misfolded, proteins have obvious links to cellular ubiquitin-dependent and leads to the formation of aggresomes, whereas an excess of functions. Purification of an endogenous HDAC6 p97/VCP over HDAC6 facilitates the release of ubiquitin-bound complex from mouse testis cytosolic extracts revealed HDAC6 and the delivery of ubiquitinated proteins to the proteasome. An excessive accumulation of ubiquitinated proteins the presence of two components identified as p97/VCP, elicits HDAC6 to mediate their transport along the microtubules the mouse ortholog of yeast Cdc48, and phopholipase and their accumulation in an aggresome. Under these conditions, A2 activating protein (PLAP), the ortholog of yeast HDAC6 controls the recruitment of the autophagic machinery to Ufd3 (Seigneurin-Berny et al., 2001). resorb the aggregates. The recruitment of HDAC6 to ubiquitinated A recent work in yeast showed that both Cdc48 and proteins leads to the induction of an HSP90-dependent chain of events to optimize the protective cell response (C Boyault and S Ufd3 are at the heart of an important decision center, Khochbin, unpublished data). which may determine whether a multi-ubiquitinated protein would be targeted for degradation to the proteasome or be deubiquitinated and eventually The ubiquitin-binding activity of HDAC6 has also functionally recycled (Rumpf and Jentsch, 2006). been shown to be critical for its recently discovered Although this mechanism has not yet been evidenced function as an adapter, mediating the transport of in higher eukaryotes, the investigation of the ubiquitin- ubiquitinated proteins along microtubule tracks to dependent functions of HDAC6 has also recently pericentriolar structures called aggresomes (Kawaguchi revealed its involvement in the determination of the et al., 2003). Through its simultaneous interaction with fate of ubiquitinated proteins. High-affinity binding of ubiquitin and dynein motors, HDAC6 is thought to HDAC6 to ubiquitin was shown to hinder the recogni- help in building up a cellular protective response to tion of cellular ubiquitinated proteins by other ubiqui- the accumulation of cytotoxic protein aggregates tin-binding factors and to subsequently delay their scattered in the cytoplasm by mediating their transport processing by the proteasome or USPs (Boyault et al., to aggresomes (Kawaguchi et al., 2003). Importantly, 2006). The HDAC6 partner p97/VCP is a chaperone this ubiquitin-dependent function of HDAC6 is also involved in the control of a variety of cellular functions, linked to its catalytic activity, which is discussed in more many of them relying on its ‘segregase’ activity detail below. disassembling various complexes, including those containing ubiquitinated proteins (Wang et al., 2003; Romisch, 2005). p97/VCP, upon its binding to HDAC6, Deacetylase-dependent functions of HDAC6 is able to extract HDAC6 bound to ubiquitinated proteins and therefore allows their further processing. Despite a strong in vitro histone-deacetylase activity A finely tuned equilibrium of cellular concentrations of HDAC6, there has been no evidence for its activity of HDAC6 and p97/VCP could therefore be critical in in vivo (Haggarty et al., 2003; Zhang et al., 2003). In the determination of the fate of ubiquitinated cellular the cells, HDAC6, therefore, appears to catalyse the proteins, mostly consisting of misfolded proteins removal of acetyl groups from substrates other than (Goldberg et al., 2002). An excess of HDAC6 would histones. The identification of two physiological sub- favor their accumulation in cells, while an increase of the strates of HDAC6, a-tubulin and HSP90, has opened the p97/VCP concentration would accelerate the processing way for the identification of a whole set of new cellular of these proteins (Boyault et al., 2006; Figure 2). functions associated with HDAC6 catalytic activity.

Oncogene HDAC6, a coordinator of cell responses to stressful stimuli C Boyault et al 5471 The first identified physiological substrate of HDAC6 essential regulatory processes. Among other functions, was a-tubulin (Hubbert et al., 2002; Matsuyama et al., HSP90 seems to be specialized in stabilizing metastable 2002; Zhang et al., 2003). Although this finding was an regions of specific factors by keeping them in a ‘holding’ important step forward to understand the mechanisms position. This is the case of glucocorticoid receptor controlling tubulin acetylation, its functional impli- (GR) whose hormone-binding activity depends on its cations remained obscure. The major reason was that, association with HSP90, which keeps the protein in an up to very recently, no function for tubulin acetyla- inactive form in the cytoplasm (Pratt and Toft, 2003). tion could be evidenced. Interestingly, a recent work The HDAC6-regulated acetylation of HSP90 was shown has reported that the tubulin binding and motility of to induce the dissociation of its cochaperone p23 and the kinesin-1is controlled by a-tubulin acetylation (Reed accumulation of GRs defective in hormone binding et al., 2006). Kinesins, like dynein, are involved in the (Kovacs et al., 2005; Murphy et al., 2005). Interestingly, transport of cargos along microtubule tracks. This study the acetylation of HSP90 was also a major cause of showed that inhibition of HDAC6 by specific inhibitors instability of some of its client proteins with critical roles induces the transport of a kinesin-1cargo protein, JIP1 in cell growth and survival (reviewed in Caron et al., (JNK-interacting protein 1), to neurites and its accu- 2005). However, importantly, HSP90 probably contains mulation in neurite-tips (Reed et al., 2006). This is the multiple acetylation sites (Scroggins et al., 2007) and not first evidence for a functional consequence of the TDAC all are deacetylated by HDAC6. Indeed, although the activity of HDAC6. Indeed, the preferential binding of HDAC inhibitor FK228 is not able to inhibit HDAC6 kinesin-1to acetylated a-tubulin is proposed to mediate activity (Furumai et al., 2002), it induces an increase in the effect of HDAC6 inhibition on JIP1transport. HSP90 acetylation (Blagosklonny, 2002). Furthermore, Likewise, the acetylation-dependent binding of kinesins an HDAC6-dependent acetylation of HSP90 may also or other molecular motors and microtubule-associated act only on a subset of the HSP90 client proteins, since proteins to tubulin may provide an explanation for no correlation was observed between HSP90 acetylation the reported effects of HDAC6 catalytic activity on in HDAC6 KO cells and the activity of one of its client the transport of protein aggregates to aggresomes proteins, the heat shock factor protein 1(C Boyault and (Kawaguchi et al., 2003), on the recruitment of the S Khochbin, unpublished data). autophagic machinery to aggresomes (Iwata et al., 2005), on cell motility (Hubbert et al., 2002; Haggarty et al., 2003), on the organization of the immune synapse Nuclear targets of HDAC6 in T cells (Serrador et al., 2004) and on the polarized release of cytokine-containing secretory lysosomes Despite the active and stable maintenance of mouse, (Carta et al., 2006). human (Verdel et al., 2000; Bertos et al., 2004) and It is important, however, to note that not all the probably Drosophila (Barlow et al., 2001) HDAC6 in the cytoskeleton-dependent functions of HDAC6 could be cytoplasm, several reports have evidenced its impact on attributed to its TDAC activity. Indeed, Cabrero et al. the transcriptional activities of various factors. Some (2006) have recently found that the cellular levels of of these effects may be indirect and attributed to the HDAC6 and not its catalytic activity are critical for activity of HDAC6 as a regulator of HSP90 chaperone lymphocyte chemotaxis. functions (Kovacs et al., 2005; Murphy et al., 2005; It is also noteworthy that the involvement of the actin Hurst et al., 2006; Kong et al., 2006). cytoskeleton in cell migration is well established Beside these indirect effects, HDAC6 also seems to (Pantaloni et al., 2001; Pollard and Borisy, 2003) and directly control the transcriptional repressor activities of that HDAC6 has also the potential to link actin several transcriptional regulators. Indeed, HATs p300 filaments and microtubule dynamics through its inter- and CBP, in addition to their well-known transcrip- action with formin homology proteins, mDia1and tional activator functions, show the ability to repress mDia2, controlling actin polymerization (Destaing transcription. Sumoylation of the CRD1domain et al., 2005; Bershadsky et al., 2006). Therefore, in of p300 was shown to mediate this repressor activity. addition to microtubules, HDAC6-dependent control of In vitro and in vivo studies suggested that a direct bind- cell migration could involve the actin cytoskeleton and ing of HDAC6 to SUMO-CRD1could be responsible their functional interconnections. for the observed transcription repression by the CRD1 Overall, these data point to HDAC6, and more domain of p300 (Girdwood et al., 2003), and further specifically to its TDAC activity, as an important deter- investigations suggested that this mechanism could be minant of microtubule-dependent intracellular traffick- involved in the repression of at least two cellular genes ing and consequently of various cellular functions (Ling and Lobie, 2004; Ma et al., 2005). depending on this process. HDAC6 was also found to be associated with The second HDAC6 substrate is the well-character- transcriptional corepressors such as LCoR, which is ized chaperone HSP90 (Bali et al., 2005; Kovacs et al., involved in a ligand-dependent repressor activity of 2005; Murphy et al., 2005). A recent thorough study nuclear receptors (Fernandes et al., 2003) as well as in showed that in yeast, HSP90 physically or genetically that of ETO-2, a component of N-CoR, SMRT and interacts with at least 10% of the yeast proteome (Zhao mSin3A complexes (Amann et al., 2001). The repressor et al., 2005). It therefore appears obvious that HSP90 activity of individual transcription factors has also been should be considered as a master regulator of many attributed to their association with HDAC6. This is the

Oncogene HDAC6, a coordinator of cell responses to stressful stimuli C Boyault et al 5472 case of Runx2, capable of both transcriptional activa- stimulates HIV-1infection (Valenzuela-Fernandez et al., tion and repression. A direct Runx2-dependent recruit- 2005). ment of HDAC6 from cytoplasm to chromatin could HDAC6 activity not only counteracts viral infection mediate the specific repression of p21CIP/WAF (Westendorf but its functions also appear crucial for the activity of et al., 2002). Nuclear factor-kB (NF-kB) p50 and p65 virus-specific cytotoxic T lymphocytes (CTL) to clear also seem to recruit HDAC6 to repress the expression of the infected cells. For instance, the specific inhibition of a gene encoding a subunit of an H þ -K þ -ATPase (Zhang HDAC6 has been shown to greatly reduce the efficiency and Kone, 2002). In contrast, the knockdown of of CD8 þ CTLs in clearing human T-lymphotropic virus HDAC6 had little effect on the double-stranded RNA type-1(HTLV-1)-infected cells (Mosley et al., 2006). (dsRNA)-induced activation of NF-kB, but severely Finally, the positive role of HDAC6 in the activation interfered with the activity of IRF3, suggesting a of the IRF3 transcription factor and the induction of positive role for HDAC6 in dsRNA/virus-dependent b-interferon gene expression in response to viral infec- activation of IRF3 and the b-interferon gene response tion (Nusinzon and Horvath, 2006) could constitute a (Nusinzon and Horvath, 2006). Here, however, it is not third category of mechanisms through which HDAC6 clear whether the action of HDAC6 is direct or mediated ensures its anti-viral activity. indirectly by its cytoplasmic activities. HDAC6-mediated accumulation of misfolded and It would be interesting to know, despite its active or ubiquitinated proteins (Boyault et al., 2006) and stable maintenance in the cytoplasm, how various aggresome formation (Kawaguchi et al., 2003) may also transcriptional regulatory factors described above could be determinant events in the so-called ‘protein con- recruit HDAC6 to chromatin. The cytoplasmic localiza- formational diseases’. The pathogenic aggregation of tion of HDAC6 is, however, not exclusive. Indeed, in the proteins in non-native conformation is generally asso- mouse a fraction of HDAC6 has been reported to enter ciated with a whole series of neurodegenerative diseases the nucleus under specific circumstances, for instance including Alzheimer’s disease, Parkinson’s disease, after butyrate-induced B16 cell differentiation (Verdel Huntington’s disease, amyotrophic lateral sclerosis and et al., 2000). Additionally, although the SE14 domain of prion diseases. human HDAC6 is responsible for a stable cytoplasmic The first evidence for the involvement of HDAC6 in anchoring of the protein (Bertos et al., 2004), an such a disease came from the colocalization of HDAC6 unknown cellular mechanism might modulate the with Lewy bodies of Parkinson’s disease. The HDAC6- activity of this domain and allow the nuclear localiza- mediated formation of aggresomes may constitute a tion of a fraction of HDAC6. Accordingly, an in situ protective cell response to cytotoxic effects of misfolded analysis of hHDAC6 intracellular localization showed protein accumulation (Kawaguchi et al., 2003). In the dependence of this localization on the state of cell addition to promoting aggresome formation, HDAC6 malignancy in mammary epithelial cells: the occurrence participates in the aggregate degradation by autophagy. of a strong cytoplasmic staining in cancer and a This has been shown using a model system expressing preferential nuclear staining in normal cells was aggregation prone huntingtin (htt) involved in the observed (Yoshida et al., 2004). Huntington disease, where these activities of HDAC6 It is also noteworthy that HDAC6 has been found to were shown to mediate the autophagic clearance of htt interact with two other HDACs. One of them is SIRT2. aggregates (Iwata et al., 2005). HDAC6 therefore Like HDAC6, SIRT2 is a TDAC (North et al., 2003). appears as a key element in cell protection against the The other HDAC6-interacting HDAC, HDAC11, is of deleterious effects of pathologic protein aggregation in unknown function but may act as transcription repres- neurodegenerative diseases. The HDAC-dependent func- sor (Gao et al., 2002). All together, these data may tions of HDAC6 on aggresome formation cannot, suggest a concerted action of HDAC6 and its interacting however, be generalized since, in a model for SOD1 HDACs. mutant aggresome assembly involved in amyotrophic lateral sclerosis, HDAC inhibitors prevented the accumulation of scattered mSOD1aggregates in aggresomes, HDAC6: a therapeutic target? but HDAC6 neutralization by tubacin did not inhibit this process. Moreover, trapoxin, an HDAC inhibitor which Several investigations designed HDAC6 as a cellular does not inhibit HDAC6, showed an efficient inhibition factor with potential anti-viral activities. This could of aggresome formation (Corcoran et al., 2004). There depend on its function as a regulator of the cyto- are, therefore, yet unknown HDAC-dependent mechan- skeleton, or on its direct or indirect involvement in isms involved in the control of aggresome formation. transcriptional regulation. Indeed, cytoskeleton-depen- The HDAC6 partner p97/VCP, by modulating dent functions of HDAC6 have been shown to be crucial HDAC6–ubiquitin interaction, may also play an in HIV-1infection of CD4 þ cells. An HDAC6- important role in preventing the pathological accumu- dependent increase in tubulin acetylation occurs follow- lation of misfolded proteins. Accordingly, p97/VCP ing HIV-1Env-CD4 receptor interaction in the CD4 þ mutations have been identified as causing ‘inclusion cells, which could be an important event in HIV-1cell body myopathy associated with Paget disease of bone fusion and infection (Valenzuela-Fernandez et al., 2005). and frontotemporal dementia’ (Watts et al., 2004). Accordingly, HDAC6 downregulation or the inhibition HDAC6 and p97/VCP, therefore, appear as excellent of its catalytic activity by a specific drug significantly target molecules in fighting neurodegenerative diseases.

Oncogene HDAC6, a coordinator of cell responses to stressful stimuli C Boyault et al 5473 The cytoskeleton-dependent functions of HDAC6 is no obvious ubiquitin-binding domain in HDAC10 could also constitute a good target for the treatment and no TDAC activity could be observed for this of inflammatory disorders depending on interleukin-1b enzyme (Guardiola and Yao, 2002; Matsuyama et al., (IL-1b), which is a potent proinflammatory cytokine. 2002). Indeed, a treatment with HDAC inhibitors, specifically Because of an important lack of knowledge on with an HDAC6 inhibitor, has been shown to signifi- the cellular functions of HDAC10, it is difficult to cantly reduce the exocytosis of IL-1b-containing secre- conclude on the functional relationship between the tory lysosomes (Carta et al., 2006). two HDACs. However, at least one report shows that HDAC6 expression and functions may also be linked both proteins interact with the major cellular phos- to cell transformation. Depending on the type of cancer, phatase, PP1, and could therefore be involved in the the HDAC6 expression level could be considered either same regulatory networks (Brush et al., 2004). A very as a good prognosis indicator (Zhang et al., 2004) or, in recent study also led to the identification of two factors contrast, as a predictor of poor prognosis and tumor involved in the pre-mRNA 30-end processing, the aggressiveness (Osada et al., 2004; Hayashi and Yama- cleavage factor CFIm25 and poly-A polymerase, as guchi, 2006; Sakuma et al., 2006). in vivo HDAC10 substrates, suggesting a probable Indeed, HDAC6 detection has been reported to function for this protein in the regulation of the correlate with survival in a subset of tamoxifen-treated 30-end processing machinery (Shimazu et al., 2007). It ER-positive breast cancer patients (Saji et al., 2005). would now be interesting to know if these proteins Another report identified positive HDAC6 staining with are also HDAC6 substrates. It should, however, be decreased survival in ER-positive breast cancer patients, noted that some other HDACs, that is HDAC1, are also using HDAC6 as a prototypical estrogen-regulated gene capable of deacetylating these substrates (Shimazu et al., (Yoshida et al., 2004). HDAC6 levels were also found 2007). elevated in primary acute myeloid leukemia blasts compared to normal adult cells (Bradbury et al., 2005). HDAC6 may also be involved in cell transformation Conclusions and tumorigenesis through its action on several critical regulatory factors. Accordingly, HDAC6 appears to HDAC6-dependent modulation of acetylation of its play a role, directly or indirectly via HSP90, in the two known substrates, a-tubulin and HSP90, has the control of the stability of HIF-1a, a transcriptional potential to initiate a variety of cellular processes with regulator involved in tumor angiogenesis (Qian et al., widespread consequences on microtubule-dependent 2006), the breast cancer metastasis suppressor 1, events and HSP90-dependent functions, respectively. BRMS1(Hurst et al., 2006), Bcr-Abl, FLT-3, c-Raf Moreover, linked to its ubiquitin-binding activity, these and AKT (Bali et al., 2005). properties of HDAC6 create additional levels of The specific inhibition of HDAC6 catalytic activity regulation at the crossroads between lysine acetylation was also shown to synergistically increase the cytotoxi- and ubiquitination signaling pathways. city induced by the proteasome inhibitor bortezomib, These properties point to HDAC6 as a particularly therefore potentiating its anti-tumor activity in multiple well-adapted molecule to coordinate cell responses to myeloma (Hideshima et al., 2005). HDAC6 TDAC stressful stimuli. Its involvement in the management of activity also contributes to the antiproliferative and misfolded protein stress, triggering cellular ubiquitin- anti-mitotic effect of a treatment combining farnesyl- and microtubule-dependent functions, strongly supports transferase inhibitor and taxanes, mainly in taxane- this hypothesis. resistant cancer patients (Marcus et al., 2005, 2006). It is also very tempting to link these activities of HDAC6 to its HSP90-regulatory capability. Accordingly, the HDAC6-HSP90 complex could be An HDAC6-related deacetylase, HDAC10 considered as a sensor of cell response to a variety of stimuli. In fact, dissociation of the complex has been HDAC10 was discovered in 2002 (Fischer et al., 2002; observed in response to different stressful conditions, Guardiola and Yao, 2002; Kao et al., 2002; Tong et al., which lead to HSP90 acetylation (C Boyault and S 2002), and the first sequence-based analysis of its Khochbin, unpublished results). The released HDAC6 structural features revealed similarities with HDAC6. may then trigger all the microtubule-mediated cell Indeed, among the class II members, the catalytic responses and, depending on cellular concentrations of domain of HDAC10 showed the best homology to that p97/VCP, could participate in the management of of HDAC6. Moreover, like HDAC6, although partial, ubiquitinated proteins, to ensure either their transport there is a duplication of its HDAC homology domain. to aggresomes, or their delivery to the proteasome and Finally, HDAC10 presents the same sensitivity regard- facilitate protein degradation through the autophagy ing HDAC inhibitors as HDAC6; it is resistant to the pathway (Figure 2). inhibitory effects of trapoxin B and butyrate, a property It is therefore expected that future investigations will that clearly specifies these two members (Guardiola and demonstrate a crucial role for HDAC6 in protective cell Yao, 2002; Gurvich et al., 2004). responses to many cellular stresses and point to this Early investigations showed, however, no functional protein as a valuable therapeutical target for stress- overlap between these two HDACs. For instance, there related pathologies.

Oncogene HDAC6, a coordinator of cell responses to stressful stimuli C Boyault et al 5474 Acknowledgements Dr Sophie Rousseaux for critical reading of the manuscript. SK laboratory was supported by CLARA cance´ ropoˆ le, We gratefully acknowledge the association ‘vaincre la mucov- EpiMed and INCa EpiPro and Association pour la Recherche iscidose’ for supporting CB PhD fellowship for 4 years and sur le Cancer (ARC) ARECA programs.

References

Adachi N, Kimura A, Horikoshi M. (2002). A conserved motif Ku70 by CBP and PCAF controls Bax-mediated apoptosis. common to the histone acetyltransferase Esa1and the Mol Cell 13: 627–638. histone deacetylase Rpd3. J Biol Chem 277: 35688–35695. Corcoran LJ, Mitchison TJ, Liu Q. (2004). A novel action of Amann JM, Nip J, Strom DK, Lutterbach B, Harada H, histone deacetylase inhibitors in a protein aggresome disease Lenny N et al. (2001). ETO, a target of t(8;21) in acute model. Curr Biol 14: 488–492. leukemia, makes distinct contacts with multiple histone Destaing O, Saltel F, Gilquin B, Chabadel A, Khochbin S, deacetylases and binds mSin3A through its oligomerization Ory S et al. (2005). A novel Rho-mDia2-HDAC6 pathway domain. Mol Cell Biol 21: 6470–6483. controls podosome patterning through microtubule acetyla- Amerik AY, Li SJ, Hochstrasser M. (2000). Analysis of the tion in osteoclasts. J Cell Sci 118: 2901–2911. deubiquitinating enzymes of the yeast Saccharomyces Dihazi H, Kessler R, Muller GA, Eschrich K. (2005). Lysine 3 cerevisiae. Biol Chem 381: 981–992. acetylation regulates the phosphorylation of yeast 6- Bali P, Pranpat M, Bradner J, Balasis M, Fiskus W, Guo F phosphofructo-2-kinase under hypo-osmotic stress. Biol et al. (2005). Inhibition of histone deacetylase 6 acetylates Chem 386: 895–900. and disrupts the chaperone function of heat shock protein Fernandes I, Bastien Y, Wai T, Nygard K, Lin R, Cormier O 90: a novel basis for antileukemia activity of histone et al. (2003). Ligand-dependent nuclear receptor corepressor deacetylase inhibitors. J Biol Chem 280: 26729–26734. LCoR functions by histone deacetylase-dependent and Barlow AL, van Drunen CM, Johnson CA, Tweedie S, Bird A, -independent mechanisms. Mol Cell 11: 139–150. Turner BM. (2001). dSIR2 and dHDAC6: two novel, Fischer DD, Cai R, Bhatia U, Asselbergs FA, Song C, Terry R inhibitor-resistant deacetylases in Drosophila melanogaster. et al. (2002). Isolation and characterization of a novel Exp Cell Res 265: 90–103. class II histone deacetylase, HDAC10. J Biol Chem 277: Bershadsky AD, Ballestrem C, Carramusa L, Zilberman Y, 6656–6666. Gilquin B, Khochbin S et al. (2006). Assembly and Furumai R, Matsuyama A, Kobashi N, Lee KH, Nishiyama mechanosensory function of focal adhesions: experiments M, Nakajima H et al. (2002). FK228 (depsipeptide) as a and models. Eur J Cell Biol 85: 165–173. natural prodrug that inhibits class I histone deacetylases. Bertos NR, Gilquin B, Chan GK, Yen TJ, Khochbin S, Yang Cancer Res 62: 4916–4921. XJ. (2004). Role of the tetradecapeptide repeat domain of Gao L, Cueto MA, Asselbergs F, Atadja P. (2002). Cloning human histone deacetylase 6 in cytoplasmic retention. J Biol and functional characterization of HDAC11, a novel Chem 279: 48246–48254. member of the human histone deacetylase family. J Biol Blagosklonny MV. (2002). Hsp-90-associated oncoproteins: Chem 277: 25748–25755. multiple targets of geldanamycin and its analogs. Leukemia Girdwood D, Bumpass D, Vaughan OA, Thain A, Anderson 16: 455–462. LA, Snowden AW et al. (2003). P300 transcriptional Boyault C, Gilquin B, Zhang Y, Rybin V, Garman E, Meyer- repression is mediated by SUMO modiﬁcation. Mol Cell Klaucke W et al. (2006). HDAC6-p97/VCP controlled 11: 1043–1054. polyubiquitin chain turnover. EMBO J 25: 3357–3366. Glozak MA, Sengupta N, Zhang X, Seto E. (2005). Acetyla- Bradbury CA, Khanim FL, Hayden R, Bunce CM, White DA, tion and deacetylation of non-histone proteins. Gene 363: Drayson MT et al. (2005). Histone deacetylases in acute 15–23. myeloid leukaemia show a distinctive pattern of expression Goldberg AL, Cascio P, Saric T, Rock KL. (2002). The that changes selectively in response to deacetylase inhibitors. importance of the proteasome and subsequent proteolytic Leukemia 19: 1751–1759. steps in the generation of antigenic peptides. Mol Immunol Brush MH, Guardiola A, Connor JH, Yao TP, Shenolikar S. 39: 147–164. (2004). Deactylase inhibitors disrupt cellular complexes Grozinger CM, Hassig CA, Schreiber SL. (1999). Three containing protein phosphatases and deacetylases. J Biol proteins deﬁne a class of human histone deacetylases related Chem 279: 7685–7691. to yeast Hda1p. Proc Natl Acad Sci USA 96: 4868–4873. Cabrero JR, Serrador JM, Barreiro O, Mittelbrunn M, Guardiola AR, Yao TP. (2002). Molecular cloning and Naranjo-Suarez S, Martin-Cofreces N et al. (2006). Lym- characterization of a novel histone deacetylase HDAC10. phocyte chemotaxis is regulated by histone deacetylase 6, J Biol Chem 277: 3350–3356. independently of its deacetylase activity. Mol Biol Cell 17: Gurvich N, Tsygankova OM, Meinkoth JL, Klein PS. (2004). 3435–3445. Histone deacetylase is a target of valproic acid-mediated Caron C, Boyault C, Khochbin S. (2005). Regulatory cross- cellular differentiation. Cancer Res 64: 1079–1086. talk between lysine acetylation and ubiquitination: role in Haggarty SJ, Koeller KM, Wong JC, Grozinger CM, the control of protein stability. Bioessays 27: 408–415. Schreiber SL. (2003). Domain-selective small-molecule Caron C, Col E, Khochbin S. (2003). The viral control of inhibitor of histone deacetylase 6 (HDAC6)-mediated cellular acetylation signaling. Bioessays 25: 58–65. tubulin deacetylation. Proc Natl Acad Sci USA 100: Carta S, Tassi S, Semino C, Fossati G, Mascagni P, Dinarello 4389–4394. CA et al. (2006). Histone deacetylase inhibitors prevent Hallows WC, Lee S, Denu JM. (2006). Sirtuins deacetylate and exocytosis of interleukin-1beta-containing secretory lyso- activate mammalian acetyl-CoA synthetases. Proc Natl somes: role of microtubules. Blood 108: 1618–1626. Acad Sci USA 103: 10230–10235. Cohen HY, Lavu S, Bitterman KJ, Hekking B, Imahiyerobo Hayashi S, Yamaguchi Y. (2006). Basic research for hormone- TA, Miller C et al. (2004). Acetylation of the C terminus of sensitivity of breast cancer. Breast Cancer 13: 123–128.

Oncogene HDAC6, a coordinator of cell responses to stressful stimuli C Boyault et al 5475 Hideshima T, Bradner JE, Wong J, Chauhan D, Richardson P, CD8+ cell antiviral function in HTLV-1infection. Blood Schreiber SL et al. (2005). Small-molecule inhibition of 108: 3801–3807. proteasome and aggresome function induces synergistic Murphy PJ, Morishima Y, Kovacs JJ, Yao TP, Pratt WB. antitumor activity in multiple myeloma. Proc Natl Acad (2005). Regulation of the dynamics of hsp90 action on the Sci USA 102: 8567–8872. glucocorticoid receptor by acetylation/deacetylation of the Hook SS, Orian A, Cowley SM, Eisenman RN. (2002). chaperone. J Biol Chem 280: 33792–33799. Histone deacetylase 6 binds polyubiquitin through its zinc North BJ, Marshall BL, Borra MT, Denu JM, Verdin E. finger (PAZ domain) and copurifies with deubiquitinating (2003). The human Sir2 ortholog, SIRT2, is an NAD+- enzymes. Proc Natl Acad Sci USA 99: 13425–13430. dependent tubulin deacetylase. Mol Cell 11: 437–444. Hubbert C, Guardiola A, Shao R, Kawaguchi Y, Ito A, Nixon Nusinzon I, Horvath CM. (2006). Positive and negative A et al. (2002). HDAC6 is a microtubule-associated regulation of the innate antiviral response and beta deacetylase. Nature 417: 455–458. interferon gene expression by deacetylation. Mol Cell Biol Hurst DR, Mehta A, Moore BP, Phadke PA, Meehan WJ, 26: 3106–3113. Accavitti MA et al. (2006). Breast cancer metastasis Osada H, Tatematsu Y, Saito H, Yatabe Y, Mitsudomi T, suppressor 1(BRMS1)is stabilized by the Hsp90 chaperone. Takahashi T. (2004). Reduced expression of class II histone Biochem Biophys Res Commun 348: 1429–1435. deacetylase genes is associated with poor prognosis in lung Iwabata H, Yoshida M, Komatsu Y. (2005). Proteomic cancer patients. Int J Cancer 112: 26–32. analysis of organ-specific post-translational lysine-acetyla- Pantaloni D, Le Clainche C, Carlier MF. (2001). Mechanism tion and -methylation in mice by use of anti-acetyllysine and of actin-based motility. Science 292: 1502–1506. -methyllysine mouse monoclonal antibodies. Proteomics 18: Pollard TD, Borisy GG. (2003). Cellular motility driven 4653–4664. by assembly and disassembly of actin filaments. Cell 112: Iwata A, Riley BE, Johnston JA, Kopito RR. (2005). HDAC6 453–465. and microtubules are required for autophagic degradation Pratt WB, Toft DO. (2003). Regulation of signaling protein of aggregated huntingtin. J Biol Chem 280: 40282–40292. function and trafficking by the hsp90/hsp70-based chaper- Kao HY, Lee CH, Komarov A, Han CC, Evans RM. (2002). one machinery. Exp Biol Med (Maywood) 228: 111–133. Isolation and characterization of mammalian HDAC10, a Qian DZ, Kachhap SK, Collis SJ, Verheul HM, Carducci MA, novel histone deacetylase. J Biol Chem 277: 187–193. Atadja P et al. (2006). Class II histone deacetylases are Kawaguchi Y, Kovacs JJ, McLaurin A, Vance JM, Ito A, Yao associated with VHL-independent regulation of hypoxia- TP. (2003). The deacetylase HDAC6 regulates aggresome inducible factor 1alpha. Cancer Res 66: 8814–8821. formation and cell viability in response to misfolded protein Reed NA, Cai D, Blasius TL, Jih GT, Meyhofer E, Gaertig J stress. Cell 115: 727–738. et al. (2006). Microtubule acetylation promotes kinesin-1 Kim SC, Sprung R, Chen Y, Xu Y, Ball H, Pei J et al. (2006). binding and transport. Curr Biol 16: 2166–2172. Substrate and functional diversity of lysine acetylation Romisch K. (2005). Endoplasmic reticulum-associated degra- revealed by a proteomics survey. Mol Cell 23: 607–618. dation. Annu Rev Cell Dev Biol 21: 435–456. Kong X, Lin Z, Liang D, Fath D, Sang N, Caro J. (2006). Rumpf S, Jentsch S. (2006). Functional division of substrate Histone deacetylase inhibitors induce VHL and ubiquitin- processing cofactors of the ubiquitin-selective Cdc48 independent proteasomal degradation of hypoxia-inducible chaperone. Mol Cell 21: 261–269. factor 1alpha. Mol Cell Biol 26: 2019–2028. Saji S, Kawakami M, Hayashi S, Yoshida N, Hirose M, Kouzarides T. (2000). Acetylation: a regulatory modification Horiguchi S et al. (2005). Significance of HDAC6 regulation to rival phosphorylation? EMBO J 19: 1176–1179. via estrogen signaling for cell motility and prognosis Kovacs JJ, Murphy PJ, Gaillard S, Zhao X, Wu JT, Nicchitta in estrogen receptor-positive breast cancer. Oncogene 24: CV et al. (2005). HDAC6 regulates Hsp90 acetylation and 4531–4539. chaperone-dependent activation of glucocorticoid receptor. Sakuma T, Uzawa K, Onda T, Shiiba M, Yokoe H, Mol Cell 18: 601–607. Shibahara T et al. (2006). Aberrant expression of histone Ling L, Lobie PE. (2004). RhoA/ROCK activation by growth deacetylase 6 in oral squamous cell carcinoma. Int J Oncol hormone abrogates p300/histone deacetylase 6 repression 29: 117–124. of Stat5-mediated transcription. J Biol Chem 279: Schwer B, Bunkenborg J, Verdin RO, Andersen JS, Verdin E. 32737–32750. (2006). Reversible lysine acetylation controls the activity of Ma H, Nguyen C, Lee KS, Kahn M. (2005). Differential roles the mitochondrial enzyme acetyl-CoA synthetase 2. Proc for the coactivators CBP and p300 on TCF/beta-catenin- Natl Acad Sci USA 103: 10224–10229. mediated survivin gene expression. Oncogene 24: 3619–3631. Scroggins BT, Robzyk K, Wang D, Marcu MG, Tsutsumi S, Marcus AI, O’Brate A M, Buey RM, Zhou J, Thomas S, Beebe K et al. (2007). An acetylation site in the middle Khuri FR et al. (2006). Farnesyltransferase inhibitors domain of hsp90 regulates chaperone function. Mol Cell 25: reverse taxane resistance. Cancer Res 66: 8838–8846. 151–159. Marcus AI, Zhou J, O’Brate A, Hamel E, Wong J, Nivens M Seigneurin-Berny D, Verdel A, Curtet S, Lemercier C, Garin J, et al. (2005). The synergistic combination of the farnesyl Rousseaux S et al. (2001). Identification of components of transferase inhibitor lonafarnib and paclitaxel enhances the murine histone deacetylase 6 complex: link between tubulin acetylation and requires a functional tubulin acetylation and ubiquitination signaling pathways. Mol Cell deacetylase. Cancer Res 65: 3883–3893. Biol 21: 8035–8044. Matsuyama A, Shimazu T, Sumida Y, Saito A, Yoshimatsu Y, Serrador JM, Cabrero JR, Sancho D, Mittelbrunn M, Seigneurin-Berny D et al. (2002). In vivo destabilization of Urzainqui A, Sanchez-Madrid F. (2004). HDAC6 deacety- dynamic microtubules by HDAC6-mediated deacetylation. lase activity links the tubulin cytoskeleton with immune EMBO J 21: 6820–6831. synapse organization. Immunity 20: 417–428. Mosley AJ, Meekings KN, McCarthy C, Shepherd D, Shimazu T, Horinouchi S, Yoshida M. (2007). Multiple Cerundolo V, Mazitschek R et al. (2006). Histone deacety- histone deacetylases and CBP regulates pre-mRNA 30-end lase inhibitors increase virus gene expression but decrease processing. J Biol Chem 282: 4470–4478.

Oncogene HDAC6, a coordinator of cell responses to stressful stimuli C Boyault et al 5476 Sterner DE, Berger SL. (2000). Acetylation of histones and Yang XJ. (2004). Lysine acetylation and the bromodomain: a transcription-related factors. Microbiol Mol Biol Rev 64: new partnership for signaling. Bioessays 26: 1076–1087. 435–459. Yang XJ, Gregoire S. (2005). Class II histone deacetylases: Tong JJ, Liu J, Bertos NR, Yang XJ. (2002). Identification of from sequence to function, regulation, and clinical implica- HDAC10, a novel class II human histone deacetylase tion. Mol Cell Biol 25: 2873–2884. containing a leucine-rich domain. Nucleic Acids Res 30: Yoshida N, Omoto Y, Inoue A, Eguchi H, Kobayashi Y, 1114–1123. Kurosumi M et al. (2004). Prediction of prognosis of Valenzuela-Fernandez A, Alvarez S, Gordon-Alonso M, estrogen receptor-positive breast cancer with combi- Barrero M, Ursa A, Cabrero JR et al. (2005). Histone nation of selected estrogen-regulated genes. Cancer Sci 95: deacetylase 6 regulates human immunodeficiency virus type 496–502. 1infection. Mol Biol Cell 16: 5445–5454. Zhang W, Kone BC. (2002). NF-kappaB inhibits transcription Verdel A, Curtet S, Brocard MP, Rousseaux S, Lemercier C, of the H(+)-K(+)-ATPase alpha(2)-subunit gene: role Yoshida M et al. (2000). Active maintenance of mHDA2/ of histone deacetylases. Am J Physiol Renal Physiol 283: mHDAC6 histone-deacetylase in the cytoplasm. Curr Biol F904–F911. 10: 747–749. Zhang Y, Gilquin B, Khochbin S, Matthias P. (2006). Two Verdel A, Khochbin S. (1999). Identification of a new family of catalytic domains are required for protein deacetylation. higher eukaryotic histone deacetylases. Coordinate expres- J Biol Chem 281: 2401–2404. sion of differentiation-dependent chromatin modifiers. Zhang Y, Li N, Caron C, Matthias G, Hess D, Khochbin S J Biol Chem 274: 2440–2445. et al. (2003). HDAC-6 interacts with and deacety- Wang Q, Song C, Li CC. (2003). Hexamerization of p97-VCP lates tubulin and microtubules in vivo. EMBO J 22: is promoted by ATP binding to the D1domain and required 1168–1179. for ATPase and biological activities. Biochem Biophys Res Zhang Z, Yamashita H, Toyama T, Sugiura H, Omoto Y, Commun 300: 253–260. Ando Y et al. (2004). HDAC6 expression is correlated Watts GD, Wymer J, Kovach MJ, Mehta SG, Mumm S, with better survival in breast cancer. Clin Cancer Res 10: Darvish D et al. (2004). Inclusion body myopathy associated 6962–6968. with Paget disease of bone and frontotemporal dementia is Zhao R, Davey M, Hsu YC, Kaplanek P, Tong A, Parsons AB caused by mutant valosin-containing protein. Nat Genet 36: et al. (2005). Navigating the chaperone network: an 377–381. integrative map of physical and genetic interactions Westendorf JJ, Zaidi SK, Cascino JE, Kahler R, van Wijnen mediated by the hsp90 chaperone. Cell 120: 715–727. AJ, Lian JB et al. (2002). Runx2 (Cbfa1, AML-3) interacts Zou H, Wu Y, Navre M, Sang BC. (2006). Characterization of with histone deacetylase 6 and represses the p21(CIP1/ the two catalytic domains in histone deacetylase 6. Biochem WAF1) promoter. Mol Cell Biol 22: 7982–7992. Biophys Res Commun 341: 45–50.

Oncogene