A novel small molecule hydroxamate preferentially inhibits HDAC6 activity and tumour growth.

1.4 MB·English

by  Kaliszczak, M

FULL PAPER
British Journal of Cancer (2013) 108, 342–350 | doi: 10.1038/bjc.2012.576
Keywords: tumour; HDAC; PET; biomarkers; pharmacology
A novel small molecule hydroxamate
preferentially inhibits HDAC6 activity
and tumour growth
M Kaliszczak1, S Trousil1, O Åberg1, M Perumal1, Q-D Nguyen1 and E O Aboagye*,1
1Comprehensive Cancer Imaging Centre, Department of Surgery and Cancer Faculty of Medicine, Imperial College London,
Hammersmith Hospital,Room 240,MRCCyclotron Building,DuCaneRoad, London W120NN, UK
Background:Thisstudyinvestigateswhetherahistonedeacetylasesubtype6(HDAC6)inhibitorcouldbeusedinthetreatmentof
solid tumours.
Methods:Weevaluatedtheeffectofanovelinhibitor,C1A,onHDAC6biochemicalactivityandcellgrowth.Wefurtherexamined
potential of early noninvasive imaging of cell proliferation by [18F]fluorothymidine positron emission tomography ([18F]FLT-PET)
todetect therapyresponse.
Results:C1AinducedsustainedacetylationofHDAC6substrates,a-tubulinandHSP90,comparedwithcurrentclinicallyapproved
HDAC inhibitor SAHA. C1A induced apoptosis and inhibited proliferation of a panel of human tumour cell lines from different
originsinthelowmicromolarrange.Systemicadministrationofthedruginhibitedthegrowthofcolontumoursinvivoby78%.The
drugshowedrestrictedactivityongeneexpressionwitho0.065%ofgenesmodulatedduring24hoftreatment.C1Atreatment
reducedtumour[18F]FLTuptakeby1.7-foldat48h,suggestingthatmolecularimagingcouldprovidevalueinfuturestudiesofthis
compound.
Conclusion: C1A preferentially inhibits HDAC6 and modulates HDAC6 downstream targets leading to growth inhibition of a
diverse set of cancer cell lines. This property together with the favourable pharmacokinetics and efficacy in vivo makes it a
candidatefor further pre-clinical andclinical development.
Histone deacetylase (HDAC) enzymes that affect the acetylation substrates. Class II members shuttle between the nucleus and the
status of histones and various cellular proteins have been cytoplasm, target primarily non-histone substrates and their
recognised as therapeutic targets in central nervous system expression is tissue specific. In parallel with the explosion of
disorders and cancer (Xu et al, 2007; Kazantsev and Thompson, HDAC research, the number of HDAC inhibitor (HDACI) drug
2008;Haberlandetal,2009;MarksandXu,2009).Twoinhibitors, candidates has grown exponentially in the last decade. Most
vorinostat (SAHA) and romidepsin, have been approved by the HDACIs reported to date inhibit multiple HDAC isoforms,
FDAfortreatmentofcutaneousTcelllymphoma(DuvicandVu, particularly HDAC1 and HDAC3, and although they have
2007; Piekarz et al, 2009). There are 18 HDACs, which are desirable antiproliferative effects, a number of these also elicit
classified structurally into class I (HDAC1, HDAC2, HDAC3, profound side effects, includingbone marrowsuppression, weight
HDAC8),classIIa(HDAC4,HDAC5,HDAC7,HDAC9),classIIb loss, fatigue and cardiac arrhythmias (Bruserud et al, 2007;
(HDAC6, HDAC10), class III (sirtuins; SIRT1-7) and class IV Marsoni et al, 2008). The discovery of HDACIs with different
(HDAC11) groups. Class I HDACs are mostly nuclear, ubiqui- isoformprofilesorindeedisoformselectiveinhibitionisimportant
tously expressed and display enzymatic activity towards histone to elucidate the mechanism of action of specific HDAC enzymes,
*Correspondence:DrEOAboagye;E-mail:[email protected]
Revised23November2012;accepted28November2012;
publishedonline15January2013
&2013CancerResearchUK.Allrightsreserved0007–0920/13
342 www.bjcancer.com|DOI:10.1038/bjc.2012.576
HDAC6inhibitor withantitumour activity BRITISH JOURNALOF CANCER
and may hold greater promise than their non-selective counter- activity of caspase-3/7 was measured using a TopCount NXT
parts, by minimising toxicity. microplate luminescence counter (PerkinElmer, Waltham, MA,
HDAC6 has recently emerged as an attractive target for the USA) and normalised to protein content.
treatment of cancer (Boyault et al, 2007; Lee et al, 2008; Kawada
PK profiling. C1A was prepared in 10% DMSO, 10% solutol
et al, 2009; Rivieccio et al, 2009). HDAC6 was shown to be the
HS15 (BASF) and 10% Tween 20 in PBS, and administrated at
deacetylaseforadiversesetofsubstrates involvedintumourigen-
20mgkg(cid:2)1intraperitoneally(i.p.)tofemaleBALB/cmice(Harlan,
esis, such as HSP90, a-tubulin, cortactin and peroxiredoxins, but
Bicester, UK). Full details are described in Supplementary
unlike other HDACs, inhibition of HDAC6 is believed not to be
Information. All animal experiments were done by licensed
associatedwithseveretoxicity;HDAC6knockoutinmicedoesnot
investigators in accordance with the United Kingdom Home
lead to embryonic lethality (Hubbert et al, 2002; Haggarty et al,
Office Guidance on the Operation of the Animal (Scientific
2003; Bali et al, 2005; Zhang et al, 2007; Parmigiani et al, 2008;
Procedures) Act 1986 (HMSO, London, UK, 1990) and within
Witt et al, 2009; Kaluza et al, 2011). To date, HDAC6 selective
guidelines set out by the United Kingdom National Cancer
inhibitors(tubacin,tubastatinA)haveonlycontributedtovalidate
Research Institute Committee on Welfare of Animals in Cancer
HDAC6 as a target, but their unfavourable pharmacokinetic (PK)
Research (Workman et al, 2010).
profileshavepreventedthemfromfurtherpre-clinicalandclinical
development, making them only good research tools (Haggarty Tumour xenografts. HCT-116 cells (5(cid:3)106) were injected
et al, 2003; Butler et al, 2010). subcutaneously in 100ml volumes into the flanks of female nu/
Here we report the development of a novel HDAC6 inhibitor, nu-BALB/c athymic nude mice (Harlan). Tumour measurements
C1A.Weshowthat,C1Ainducescelldeathandsignificantgrowth were performed daily and volumes were calculated using the
inhibition in a panel of cancer cells, and because of its favourable formula (length (mm))(cid:3)(width (mm))(cid:3)(depth (mm))(cid:3)p/6.
PK profile, is efficacious in solid tumours in vivo. When tumours reached a volume of 50–100mm3, treatment with
different compounds was initiated. Throughout the 14-day
treatment period, animal weights and tumour volumes were
MATERIALS AND METHODS determined daily.
[18F]Fluorothymidine positron emission tomography imaging.
HDAC deacetylase activity. HDAC enzyme inhibition was HCT-116 cells (5(cid:3)106) cells were injected on the back of female
assessed as described elsewhere, using recombinant enzymes and
nu/nu-BALB/cmice.At24or48hposttreatmentwithC1Agiven
7-amino-4-methylcoumarin (AMC)-labelled substrates from p53 at40mgkg(cid:2)1QD,theanimalswerescannedonadedicatedsmall
residues 379–382: the activities of HDACs 1, 2, 3, 6 and 10 were
animal PET scanner (Siemens Inveon PET module) following a
assessed using the substrate RHKKAc; HDAC8 activity was bolus intravenous injection of B3.7MBq of [18F]fluorothymidine
assayed using a specific substrate RHKAcKAc (Schultz et al, ([18F]FLT) as previously described (Leyton et al, 2006a,b, 2008;
2004; Butler et al, 2010). Activities of the Class IIa HDACs
Nguyen et al, 2009). The normalised uptake value at 60min post
(HDAC4,5,7,9)wereassessedusingaClassIIa-specificsubstrate
injection(NUV )wasusedforcomparisons(Nguyenetal,2009).
(R, Acetyl-Lys(trifluoroacetyl)-AMC; Lahm et al, 2007). 60
Ki-67 immunostaining. Tumours treated with C1A or vehicle
Growthinhibitoryassay. Drugconcentrationsthatinhibited50%
wereexcisedafterimaging,fixedinformalin,embeddedinparaffin
ofcellgrowth(GI )weredeterminedusingasulphorhodamineB
50 and cut into 5.0mm sections, and tumour proliferation was
techniqueasdescribedelsewhere(VichaiandKirtikara,2006).All
determined as previously described (Leyton et al, 2006a). Data
cell lines were treated for 72h on day 2 unless otherwise stated.
from10randomlyselectedfieldsofviewpersectionwerecaptured
Immunoblotting. Cells were cultured for 24h and subsequently usinganOlympusBX51microscope(Olympus,Southend-on-Sea,
treated with different compounds: C1A (synthesised in-house), UK) at (cid:3)400 magnification. Ki-67-positive cells were counted
SAHAand 17-AAG (LC Laboratories, Woburn,MA, USA) at the using ImageJ software (NIH, Bethesda, MD, USA) and expressed
indicated concentrations for 4 or 24h. Protein samples were as a percentage of total cells counted.
subsequently prepared by lysing cells in RIPA buffer (Invitrogen,
Active caspase-3 and TUNEL immunohistochemistry assay.
Paisley, UK) supplemented with protease and phosphatase
Tumour sections were processed for active caspase-3 and DNA
inhibitor cocktails (Sigma, St Louis, MO, USA). Tumour samples
degradation with terminal deoxynucleotidyl transferase dUTP nick
were prepared as follows: excised and snap-frozen tumour
endlabelling(TUNEL)aspreviouslydescribed(Nguyenetal,2009).
xenografts were homogenised in RIPA lysis buffer with the
PreCellys 24 homogeniser and CK14 bead-containing tubes (2 Gene array. Animals were treated with C1A at 40mgkg(cid:2)1 (or
cycles of 25s at 6500r.p.m.; Bertin Technologies, Montigny le vehicle) and the tumours excised and snap-frozen 24h after
Bretonneux, France). For immunoprecipitation experiments, cell injection. Total RNA was extracted from tumours using RNeasy
lysates were incubated with primary antibody for 1h at 41C and mini Kit (Qiagen, Crawley, UK) and hybridised to Affymetrix
subsequently incubated with protein A/G Plus-Agarose beads human genome U219 microarray (Affymetrix, Santa Clara, CA,
(Santa Cruz Biotechnologies, Santa Cruz, CA, USA) overnight, USA). Studies were performed under contract by AlphaMetrix
centrifuged, washed, resuspended in lysis buffer and subject to Biotech (Ro¨dermark, Germany). A differential expression of
standard western blot procedures. 1.5-fold was selected as a cut-off value.
Cellcycleanalysis. CellsweretreatedwithC1AorSAHAfor24h. Statisticalanalyses. StatisticalanalysesweredoneusingGraphPad
Collectedcellswerefixedwithethanolandstainedwithpropidium prism software. Differences between groups were analysed by two
iodide in PBS for 3h. Cell cycle distributions were determined tailedStudent’st-tests.Differencesintumourgrowthinnudemice
using flow cytometry (FACS canto, Becton Dickinson, Oxford, were analysed by repeated measures two-way ANOVA.
UK)andanalysedusingtheflowJosoftware(TreestarInc,Ashland, P-valueso0.05 using a 95% confidence intervals were considered
OR, USA). In each analysis, 10000 events were recorded. significant. Data are reported as mean±s.e.m. of at least three
independent experiments unless otherwise stated. *Po0.05,
Caspase-3/7 assay. Caspase-3/7 activity was determined using
**Po0.005, ***Po0.0001, NS, not significant.
Promega’s caspase-3/7 assay according to the manufacturer’s
instructions (Promega, Madison, WI, USA). In brief, cells were Additional methods. Additional methodology is described in
incubated for 1h with Caspase-Glo reagent, and the enzymatic Supplementary Information—Materials and Methods.
www.bjcancer.com|DOI:10.1038/bjc.2012.576 343
BRITISHJOURNAL OF CANCER HDAC6 inhibitor withantitumouractivity
after1mMC1Atreatment(Figure2AandB).TheeffectofC1Aon
RESULTS
acetylation of a-tubulin was drug concentration-dependent at
4and24h(Figure2BandC).Treatmentwiththeclinicallylicensed
C1AisaneffectiveHDAC6inhibitor. C1Awasdevelopedbased broad-spectrum inhibitor, SAHA, also increased acetylation of
onthestructureofnaturallyoccurringpanHDACIstrichostatinA a-tubulin. Unlike C1A, however, SAHA increased acetylation of
and SAHA. Although maintaining the hydroxamate part of the histoneH3,abiomarkerforinhibitionofclassIHDACs(Figure2A
molecule, responsible for the binding to the Zn2þ pocket of and C). When treated continuously for 4h (10mM) and drug was
HDAC in general, the aromatic nitrogen mustard moiety, subsequently removed by washing to mimic intermittent dosing
originally introduced to enhance duration of drug action, was in vivo, acetylation of a-tubulin was maintained for 3h after
found to enhance preferential binding to HDAC6 (Figure 1A; washout of C1A but not with SAHA (Figure 2D). These results
Finnin et al, 1999). A molecular modelling approach was used to show that C1A is associated with a sustained effect on HDAC6
illustratethepreferentialaffinityofC1AtoHDA6catalyticdomain response that would allowthedrugto be dosed less frequently.
(cd)II(Figure1BandC)relativetoHDAC1cd(Figure1DandE). The loss of chaperone activity of HSP90 is a functional
Inmanyaspects,thebindingmodeforthetwoisoforms,HDAC1 consequence of its acetylation (Scroggins et al, 2007). CDK4 is a
and HDAC6, is very similar. Both are bound to the Zn2þ ion in recognisedclientproteinoftheHSP90chaperoneandisdegraded
the bottom of the pocket in a bidentate fashion. There are uponHSP90inhibition(Banerjietal,2005).BothC1AandSAHA
hydrophobic interactions between the conjugated carbon chain of wereassociatedwithadeclineofCDK4expression,consistentwith
C1AandthesidechainsofPhe150andPhe205inHDAC1cdand HSP90inhibition(Figure2E).Asacontrol,treatmentofcellswith
the side chains of Phe620 and Phe680 in HDAC6cdII. However, theHSP90inhibitor,17-AAG,alsodecreasedCDK4expressionin
there are important differences as can be seen by comparing the these cells (Figure 2E). Treatment with positive control tubastatin
wateraccessiblesurfaces.ForHDAC1cd,theanilinemustardbinds A, a HDAC6 inhibitor tool compound, was also associated
to the surface of the protein at an unfavourable solvent-exposed with a decline of CDK4, concomitant with a drug concen-
area.TheHDAC6cdIIpocketismuchwiderandcanmuchbetter tration-dependent increase of the acetylated form of a-tubulin
accommodatetheinhibitor,andtheanilinemustardmoietybinds (Supplementary Figure 1).
to a well-defined hydrophobic groove, which contributes to the Similarly, HCT-116 cells transfected with HDAC6 shRNA
observed isoform selectivity. showed increased acetylation of a-tubulin concomitant with a
C1A inhibited class I/II HDACs, and sirtuins, with highest decrease of CDK4 (Supplementary Figure S2).
affinityforHDAC6(IC50¼479nM)versusHDAC1(IC50¼14mM;
Figure 1F and Supplementary Table S1). HDAC8 was also C1A does not promote non-specific DNA alkylation. We
inhibited by C1A at submicromolar concentrations (Supplemen- wondered if the presence of a nitrogen mustard moiety in C1A
tary Table S1). Initial reports suggesting that HDAC8 drug promotednon-specificDNAalkylation.AsDNAalkylatingagents
inhibition had little effect in solid tumour-derived cell lines (e.g., will generally be more active in cell lines deficient in the DNA
HCT-116)triggeredourattentiontowardsHDAC6inhibitionand repair machinery, we tested the potency of the drug in cell lines
its application in our experimental models of cancer deficient in DNA repair proteins. We showed that, in contrast to
(Balasubramanian et al, 2008). the DNA alkylating agent chlorambucil, the growth inhibitory
potency of C1A was not affected by DNA repair defects
(Figure 2F), suggesting that the nitrogen mustard moiety in C1A
C1AtreatmentaffectsknownHDAC6substrates. Weevaluated
does not induce non-specific DNA alkylation.
the effect of C1A on HDAC6 substrates. In human colon cancer
HCT-116 cells, we observed an increase in the acetylation of HDAC6 inhibition is associated with antiproliferative activity
a-tubulinandHSP90,knownsubstratesofHDAC6,asearlyas4h andapoptosis. C1Ainhibitedthegrowthofapanelof17human
O
OH
O N Tyr782 Asp742
O H
S Pro608 Asp649
N
H
Phe620 His610His651
His611
N Cys618 Phe680
CI Ser568
Pro681
CI
Phe566 Met682
Pro29 Tyr303 120 HDAC1
His28 Leu271 100 HDAC6
Asp264 %) 80
Asp176 on ( 60
PChey1s15501 His141His178 nhibiti 40
His140 I 20
Phe205
0
0.0010.01 0.1 1 10 1001000
C1A ((cid:2)M)
Figure1.EnhancedspecificityofC1AtowardsHDAC6versusHDAC1.(A)ChemicalstructureofC1A.(B-E)Representativestructuresofthe
homologymodelsofHDAC6catalyticdomain,(HDAC6cdII;3Dsolventaccessiblesurfacearea(B);3Dribbondiagramshowingkeyaminoacid
interactions(C)andHDAC1cd(3Dsolventaccessiblesurfacearea(D);3Dribbondiagramshowingkeyaminoacidinteractions(E)withC1Abound
totheZn2þ intheactivesite.ElectrostaticsurfacecolouringinBandDarered(negativecharge)andblue(positivecharge).(F)Inhibitionof
recombinantHDAC6andHDAC1byC1A.
344 www.bjcancer.com|DOI:10.1038/bjc.2012.576
HDAC6inhibitor withantitumour activity BRITISH JOURNALOF CANCER
ControlC1A SAHA 4 h C1A ((cid:2)M) SAHA ((cid:2)M)
0 0.5 1 2.5 5 7.510 0 0.050.10.5 1 5
Acetyl (cid:3)-tubulin
(cid:3)-tubulin Acetyl (cid:3)-tubulin
I.P. Ac Lys
I.B. HSP90 (cid:3)-tubulin
I.B. HSP90
C1A ((cid:2)M) SAHA ((cid:2)M)
Acetyl H3 24 h
0 0.5 1 2.5 5 7.510 0 0.050.1 0.5 1 5
H3
Actin Acetyl (cid:3)-tubulin
(cid:3)-tubulin
C1A SAHA Acetyl H3
Acetyl (cid:3)-tubulin
H3
(cid:3)-tubulin
C1A ((cid:2)M) SAHA ((cid:2)M) 17-AAG (nM)
0 0.5 1 2.5 5 7.510 0 0.050.10.5 1 2.5 5 7.5 10 0 5 10 50100500
CDK4
(cid:3)-tubulin
C1A
125 AA8 (WT) 125 CHO-K1 (WT) 125 V79 (WT)
Survival1075050 UIUrsVV129S36F (( XE(XPRRBCC–C)C13––)) Survival1075050 Xrs5 (XRCC5–) Survival1075050 Irs1 (XRCC2–)
% % %
25 25 25
0 0 0
–7 –6 –5 –4 7 –6 –5 –4 –7 –6 –5 –4
Log drug conc (M) Log drug conc (M) Log drug conc (M)
Chlorambucil
125 125 125
100 AA8 (WT) 100 CHO-K1 (WT) 100 V79 (WT)
Survival 7550 IUUrsVV129S36F (( XE(XPRRBCC–C)C13––)) Survival 7550 Xrs5 (XRCC5–) Survival 7550 Irs1 (XRCC2–)
% % %
25 25 25
0 0 0
–6 –5 –4 –3 –6 –5 –4 –3 –6 –5 –4 –3
Log drug conc (M) Log drug conc (M) Log drug conc (M)
Figure2.C1AisaselectiveinhibitorofHDAC6associatedwithasustainedeffect.(A)InhibitionofHDAC6substratesbyC1A(1mM;4h)relativeto
SAHAinHCT-116cells.(B)Drugconcentration-dependentacetylationofa-tubulinat4h,and(C)at24hrelativetoacetylhistoneH3expression.
(D)Effectof10mMC1AorSAHAonacetylationofa-tubulininHCT-116cellsfollowing4hdrugincubationandwashout(cellscollected3hafter
washout).(E)CognateeffectofC1AandSAHAandC1AonHSP90clientproteinCDK4following24htreatment.TheHSP90inhibitor,17-AAG,
wasusedasapositivecontrol.(F)EffectofC1AonthegrowthofDNArepair-deficientandrepair-competentcellsrelativetochlorambucil.
UV23andUV96celllinesaredeficientinXPBandERCC1,respectively;proteinsinvolvedinthenucleotideexcisionrepairpathway.Irs1and
Irs1SFaredeficientinXRCC2andXRCC3,respectively;proteinsinvolvedinthehomologousrecombinationrepairpathway.XRS5isdeficient
inXRCC5;proteininvolvedinthenon-homologousend-joiningpathway.Growthcurvesinparentalrepair-competentcelllinesAA8
(forUV23,UV96andIrs1SF),CHO-K1(forXRS5)andV79(forIrs1)arealsopresented.
cancer cell lines, including cell lines derived from 8 different mechanism. The characteristic increased sub-G1 fraction was also
histological types of solid tumours and one type of B-cell demonstrated inA2780humanovariancancercelllinebutnotin
malignancy, with a mean growth inhibitory effect (GI ) of the caspase-3-deficient human breast cancer cell line, MCF7
50
3.1±2.2mM following 72h continuous exposure (Table 1). When (Supplementary Figure S3; Janicke, 2009). In contrast, SAHA
the cells were treated for 6h, washed and allowed to grow for an increasedthefractionofcellsarrestedinG2/M(from30–41%)in
additional72hindrug-freegrowthmedium,thegrowthinhibitory the HCT-116 cells. SAHA also induced a sub-G1 population, but
effect of C1A was more than 300-fold greater than for SAHA unlike C1A this effect reached a plateau at 22% from a low drug
(Figure3A).Underwashoutconditions,thelatterdidnotshowany concentration.Weconfirmedthedrug-inducedincreaseinsub-G1
antiproliferative effect at the concentrations tested. The cellular asoccurringviaapoptosisbymeasuringcaspase-3/7 activity:both
effect of C1A was further characterised in HCT-116 cells to C1A and SAHA induced a drug concentration-related increase in
investigate the mechanism of growth inhibition. Flow cytometry caspase-3/7activity(Figure3D);increasedcaspase-3/7activitywas
studiesshowedthattreatmentofcellswithC1Afor24hincreased further confirmed in HCT-116 cells by flow cytometry (FLICA
the sub-G1 population in a drug concentration-related manner positive/SYTOX red negative; Supplementary Figure S4). Drug
(from 4% in untreated cells to 64% at the highest concentration concentration-dependent increase in caspase-3/7 activity was also
tested, i.e. 10mM; Figure 3B and C), suggesting an apoptotic seen in these cells with tubastatin A (Supplementary Figure S5).
www.bjcancer.com|DOI:10.1038/bjc.2012.576 345
BRITISHJOURNAL OF CANCER HDAC6 inhibitor withantitumouractivity
Table1.GrowthinhibitoryeffectofC1AcomparedwithSAHAinapanel A B SAHA C1A
ofcancercelllines(mean±s.d.)
Control
125
Celltype Cellline C1A(lM) SAHA(lM) 0.5 (cid:2)M
100
Breast MDAM-CMFB7-435 73..73±±00..215 0.39.65±±00..065 vival 75 1 (cid:2)M
ECnodloonmetrial HIsCTh4iTk7-a1Dw1a6 435...536±±±010...06033 00..164.782±±±000...0125 % Sur 5205 SAHA 2.55 (cid:2)(cid:2)MM
Epidermal A431 4.1±0.6 1.1±0.1 C1A 7.5 (cid:2)M
0
Lung A549 5.8±2 1.8±0.6 –7 –6 –5 –4 10 (cid:2)M
Myeloma ARH-77 3±0.55 0.48±0.24 Log drug conc (M) 50 k 150 k 50 k 150 k
KMS-11 0.48±0.24 0.27±0.03
PE PE
Neuroblastoma SH-SY5Y 0.28±0.01 0.92±0.3 C
Kelly 0.18±0.01 0.4±0.12 80 Sub-G1 C1A 60 G1 C1A
S-K-NAS 3.3±1.5 0.49±0.02 n SAHA SAHA
BE2C 0.66±0.19 0.63±0.08 atio 60 40
ul 40
Ovarian IGAR27O8V01 0.19.61±±00..244 0.18.22±±00..21 % Pop 20 20
0 0
Prostate LNCAP 3.7±1.5 1.7±1.2 0 2.5 5 7.5 10 0 2.5 5 7.510
DU-145 4.6±0.6 0.54±0.01 Drug concentration Drug concentration
((cid:2)M) ((cid:2)M)
Caspase-3/7 activation by C1A also occurred in A2780 cell line S G2/M
b(HuSunDutlipAknpeColet6dmrisuenhngRttaNhtrreAyeacFtwamisgapesuanristeneh-nS3i3ob-)di.tdeePfidifrcfoebierlyinefnet2rc4aMe%tiCoiannFft7ceoarfbs3prceeadalslasesty-3stcr/oaa7nfncsasfeeceretcidvtcieinetdlygl,wwlbiintuahest % Population 221150505 CS1AAHA % Population 5432100000 CS1AAHA
detectable (Supplementary Figure S6). Flow cytometry studies 0 0
0 2.5 5 7.5 10 0 2.5 5 7.5 10
revealed that, cells transfected with HDAC6 shRNA had an
Drug concentration Drug concentration
increased sub-G1 population (from 1.2% for shRNA-scramble to ((cid:2)M) ((cid:2)M)
31.8%forshRNAHDAC6),suggestingthatthepeakofcaspase-3/7
D E
may have occurred at an earlier time point.
asbersheutnrmogoaTgaiwlanh,edincsen-2trs0estecpod0aete0shlbcule;atencrtWouycatfchmhicllateeoshnionnipegnnshreehecdieqbfyitbeucfirtioaeltoeilnilnnlr,oocs-tn2eCwita0aDoicl0nrafa6gKguhc;estttiFeiisrivnndtiegiohagtubnytibmlryeaoeirtehSfog3nAyCerEtplH1y)pew.ArA2HiaIt1nchoDwentcAtCahyosHa1ClnatAaDttCcriw.oAcalTaonasCstsmh,si6enpppsIdrie2ane(un1Rvhscitiiieeloubcedduhdivtsoiebbeellndyyyss olf increase of caspase7 activity (normalised tototal protein content) 432100000 CSA1AHA ControlC1A SAHPA(cid:4)2-a1ctin
cell growth and induced cell death in a variety of cell lines. F3/ 0 2.5 5 7.5 10
We then tested whether different levels of HDAC6 protein Drug concentration ((cid:2)M)
expression may have an impact on C1A-induced growth inhibi-
Figure3.EffectofC1Aonproliferationandapoptosis.(A)Growth
tion. HDAC6 was found to be ubiquitously expressed and no
inhibitoryeffectofC1AandSAHAinHCT-116cellsfollowing6hdrug
significantdifferenceswereseenbetweenthecancercelllinestested
incubation,washandincubationindrug-freemediumforanadditional
compared with HDAC1 (Supplementary Figure S7). There was
72h.(B-C)HCT-116cellsweretreatedwithC1AorSAHAfor24hand
no relationship between the growth inhibitory effect of C1A
DNAprofileanalysedbyflowcytometry.(D)EffectofC1AorSAHA
and HDAC6 nor HDAC1 protein expression (Supplementary
treatment(24h)oncaspase-3/7activity.(E)EffectofC1AorSAHA
Figure S7).
(24h–10mM)onp21expressioninHCT-116cells.
HDAC6 inhibition results in antitumour activity. Encouraged
by our results in vitro, and against the background that HDAC6 miceandmicetreatedwithC1Aat40mgkg(cid:2)1q.o.d.,20mgkg(cid:2)1
inhibitorsreportedtodatehavepoorPKs,wethenwantedtoknow q.d.and20mgkg(cid:2)1b.i.d.,respectively;Figure4C).C1Atreatment
whether C1A could achieve sufficient plasma levels in vivo. The was associated with a tumour growth delay (TGD ) of 4.0±0.8,
2x
PKsofC1Afollowedatwo-compartmentmodelwithanapparent 3.8±1.2and5.7±1.4daysandatumourgrowthinhibition(TGI)
terminal half-life of 10h (Figure 4A). Maximum drug concentra- of 57, 69 and 78% compared with vehicle at 40mgkg(cid:2)1 q.o.d.,
tion was rapidly achieved (30min) when given i.p. Notably, drug 20mgkg(cid:2)1 q.d. and 20mgkg(cid:2)1 b.i.d., respectively; there was no
levels equivalent to two-fold GI in HCT-116 cells could be general toxicity in any of the treated cohorts as determined by
50
achieved following administration of C1A at 20mgkg(cid:2)1. C1A changes in body weight (Figure 4D).
exhibited linear PK characteristics with dose-proportional
increases observed in plasma concentration (Figure 4B). In vivo antitumour activity is associated with molecular and
Next, we testedif the potent in vitro activity ofC1A combined imaging biomarker changes. Biochemical target modulation
with its favourable PK features favoured efficacy in the human invivowasdeterminedbymeasuringlevelsofacetylateda-tubulin
cancer xenograft models. Different schedules and dose levels of in HCT-116 tumours. Following a single dose of C1A at
C1AweretestedintheHCT-116xenograftmodelandantitumour 40mgkg(cid:2)1, acetylated a-tubulin could be detected at 6h and
activity was found to be dose-related. Doubling times were sustainedupto24h(Figure5A).Nomodulationofacetylationof
4.4±1.1, 8.3±1.0, 7.6±1.2 and 9.5±1.5 days for vehicle control histoneH3wasdetectable,confirmingalackofinhibitoryeffecton
346 www.bjcancer.com|DOI:10.1038/bjc.2012.576
HDAC6inhibitor withantitumour activity BRITISH JOURNALOF CANCER
Cmax ((cid:2)g ml–1) 3.2
Cmax time (min) 30
(cid:2)–1g.ml 432100 1HA0Ua0lCf- li0v e–2s 02 o04f he l(iGm(cid:2)3g0Ii5n 00ma it(cid:3)ln(cid:4)i–o4 1tHtn0 h110C//)221T5-10101516001 hmin (cid:2)–1Plasma concentration (g.ml) 10864200P valu1e0 < 0.002001 30 40 Relative tumour volume to day 1 64200 22V42000e hmmmicgggl4e kkkggg–––1116 QQBIODDD8 P1 <0 01.0200114 % Weight compared with day 1 10752050500 222V4000e hmmmic4gggle kkkggg–––6111 QQBIODD8D 10 12 14
Time (min) Injected dose (mg kg–1) Time (days) Time (days)
Figure4.C1Aisassociatedwithfavourablepharmacokineticprofileandsignificantantitumouractivity.(A)PharmacokineticprofileofC1A
followingasinglei.p.injectionat20mgkg(cid:2)1(n¼3).(B)Dose-dependentrelationshipofthepharmacokineticprofile1hpostinjection(n¼3).
(C)EffectofC1AontumourgrowthinHCT-116colontumourxenograftmodel(n¼6).ANOVAwasdonetodeterminestatisticalsignificance
betweengroupcontrolandeachgrouptreatedwithC1A.(D)Shownaretheweightsoftheanimalsusedinthetimecourseofthestudy.
A B C
*
AcetylA ααc--ettuutybblu uHlliin3n Vehicle2 h 6 h –#61 h –2#42 h 2–4# 1h –#2 Ve h 24 h 48 h1000 %% Normalised NUV60 2110....00505 Veh 24 h**4*8 h Vehicle C1A100 μm KI-67 (% of positivecells per field of view)21000 Veh C*1*A*
D Δ Caspase-3 TUNEL DAPI Merge E Gene name
BST2
FHL1
CV1eAh Δ Caspase-3(% staining)2110....00505 Veh C*1A* Tunel (% staining)41032 Veh C*1*A* DAPI (% staining)6420000 Veh C1*A NBPHF8IS-DPRS1TP1IHRTA2PT/AHCXDD15IPBPP42AR25KRAO-BA3AFL21GXPCB711146
100 μm
–2 –1 0 1 2
Fold change of
mRNA levels C1A-
treated versus vehicle
Figure5.MolecularandimagingbiomarkersformonitoringtheresponsetoC1A.NudemicebearingHCT-116tumourxenograftswereinjected
witheithervehicleorC1Aat40mgkg(cid:2)1perday.(A)Expressionofacetylateda-tubulinandH3overtimefollowingasingleinjectionofC1Aat
40mgkg(cid:2)1.(B)Computedtomography(CT;top)andcorresponding[18F]-FLT-PET-CT(bottom).Forvisualisation,cumulativeimagesupto60min
aredisplayed.Thetumourisindicatedbyanarrow.Quantitative[18F]-FLTuptakeat24and48hwasexpressedasradiotraceruptakeintumours
normalisedtothatofheartat60min(NUV ;n¼4).(C)Ki-67immunostainedtumoursectionsfollowingtreatmentfor2days.Proliferativecells
60
havebrownstaining.Ki-67labellingindexwascalculatedfrom10randomfieldsofviewperslice(2slicespertumour,and3tumourspergroup).
TheKi-67-positivecellswereexpressedasapercentageoftotalcells.(D)Tumouractivecaspase-3andTUNELimmunohistochemistryanalysis.
TumourtissueswereremovedafterPETimaging(followingtreatmentfor2days)andstainedforactive(cleaved)caspase-3(Dcaspase-3)andDNA
fragmentation(TUNELassay).Representativeimagesofhistologicaltumoursectionsareshown.Stainingintensitieswereexpressedaspercent
stainingperfieldusing10randomfieldsofviewperslice(2slicespertumourand3tumourspergroup).(E)DifferentialgeneexpressionofHCT-
116tumours24hfollowingasingleinjectionofC1Aat40mgkg(cid:2)1andcomparedwithvehicle-treated(n¼3).
class I HDACs. Furthermore, tumour uptake of the proliferation (Po0.0001), respectively, together with a decrease of DNA
imaging biomarker, [18F]FLT (Shields et al, 1998; Leyton et al, content (P¼0.023; Figure 5D).
2006a),decreasedafterC1Atreatment(Figure5B).Radioactivityin
tumourregionsnormalisedtothatoftheheartofthesameanimal Gene expression signature following C1A treatment
wasusedquantitativelytocomparetheeffectsofC1A(Leytonetal, in vivo. Although HDAC6 has been described as being pre-
2006a). There was a 1.7-fold decrease in tumour [18F]FLT uptake dominantly cytoplasmic, it has also been suggested to act as a
in animals treated with C1A for 48h compared with vehicle- nuclearcofactor(Palijanetal,2009).Toassessifthemechanismof
treated mice; NUV was 1.77±0.11 before treatment and response to C1A in vivo was related at least in part to
60
decreased to 1.59±0.10 (P¼0.033) at 24h and 1.06±0.15 transcriptional activation, directly or indirectly via its substrates,
(P¼0.0001) at 48h after initiating treatment (Figure 5B). Ki-67 a gene array was performed in the HCT-116 xenograft model. Of
immunostaining decreased in these tumours in keeping with the the20000genestested,only13werederegulated(0.065%)at24h
reduction in tumour [18F]FLT uptake (1.7-fold; Po0.0001) following a single injection of C1A (40mgkg(cid:2)1) in keeping
(Figure 5C). We further evaluated if C1A induced apoptosis with the class II (HDAC6) effect of C1A (i.e., minor effect in
invivo.Activatedcaspase-3expressionandsubsequentfragmented the regulation ofgenes; Figure 5E)in contrast to SAHA(LaBonte
DNA staining (TUNEL) increased 6- (P¼0.0082) and 7-fold et al, 2009). The following genes were upregulated: BST2, FHL1,
www.bjcancer.com|DOI:10.1038/bjc.2012.576 347
BRITISHJOURNAL OF CANCER HDAC6 inhibitor withantitumouractivity
NBPF(8-11, 14-16), XAF1, PIP5KL1, BAX, SRD5A1, DHCR24, inhibitionofproliferation,increaseinsub-G1fraction,increasedof
TAPBP, PTPRG, IPO7 and RAD23B; HIST2H2AC was down- acetylation of a-tubulin and decreased CDK4 expression with
regulated. Among those, the pro-apoptotic factors XAF1 (gene unremarkable changes in acetylation of histone H3 providing
encoding XIAP-associated factor 1, an antagonist of X-linked further confidence in the specificity of C1A towards HDAC6.
inhibitor of apoptosis protein) and BAX (gene encoding Bcl-2- C1A was potent in the low micromolar range across a diverse
associated X protein) were upregulated, which could explain in range of cancer cell lines. The presence of a sub-G1 population
part the antitumour activity. Of note is the upregulation together with caspase-3/7 induction suggested that the drug may
of RAD23B that was recently proposed as a predictive marker act via an apoptotic mechanism. Further studies are required to
of response to HDACIs generally (Fotheringham et al, 2009). To elucidate the exact mechanism leading to apoptosis. We speculate
confirm changes in protein expression, western blot analysis was that drug potency profile could be due, in part, to acetylation of
performed onthe excised tumours.Levels ofXAF1, PIP5KL1and cytoplasmic proteins, including HSP90 and a-tubulin, which have
RAD23B were indeed higher following treatment with C1A pivotal roles in tumour progression. HSP90 is involved in the
(Supplementary Figure S8). maturationofseveraloncogenicclientproteins,thedegradationof
which could abrogate cell growth and induce apoptosis (Maloney
and Workman, 2002). Acelylation of a-tubulin, which promotes
microtubulestabilisation,maydistortmitosisandtriggerapoptosis
DISCUSSION (Matsuyamaet al,2002); itcan alsoaffect changes incell motility
and,hence,invasionandmetastasisofcancercells(Hubbertetal,
Cancer remains an unmet medical need with one in three people 2002). Further work is required to directly link inhibition of C1A
affected by thedisease. Newtherapies withimproved efficacy and targets to growth inhibition. It is possible that the presence of an
reducedsideeffectprofileareneeded.Wehaveshownthatapotent aromatic nitrogen mustard moiety may independently induce cell
smallmoleculeinhibitorofHDAC,inparticularHDAC6,inhibits death. We examined this possibility in cell lines proficient and
proliferation and induces cell death in tumour cells lines from deficient in DNA repair. These studies (Figure 2F) suggested
diverse histological background. that the mustard moiety in C1A does not induce non-specific
There is an increasing interest in HDAC6 as a therapeutic DNAalkylationand,thus,unlikelytosignificantlycontributetoa
target. Although HDAC6 inhibitors including tubacin and DNA alkylation-dependent cell death. We wondered if HDAC6
tubastatin A (Haggarty et al, 2003; Butler et al, 2010) have been expression in the different cell lines might predict sensitivity to
reported, the lack of inhibitors with favourable PK profile has C1A. This did not appear to be the case as HDAC6 was
stifled progress within the field. More recently, a new HDAC6 ubiquitously expressed with no significant differences across cell
inhibitor ACY-1215 was reported as an inhibitor of multiple lines. Furthermore, the differences in growth sensitivity
myelomacellgrowthwhencombinedwithbortezomib(Santoetal, between the cell lines were only B10-fold. Future studies in a
2012). The chemical structure of ACY-1215 is not available for larger cohort of cell lines may provide the possibility to confirm
comparisonandtheeffectofACY-1215onsolidtumourswasnot this finding.
evaluated in that study. We report the mechanism of action and Unlike HDAC6 tool compounds like tubastatin A, C1A had
growth inhibitory properties of a new HDACI, C1A in a solid favourable PKs in vivo, and systemic administration of the drug
tumour model. HCT-116 xenograft model was chosen because it inhibited the growth of colon tumours in vivo by up to 78%.
represented the highest HDAC6 expression, has intermediate Consistent with proposed mechanism of action, we observed a
sensitivity to C1A (thus, data are not biased to high sensitivity time-dependentinductionofa-tubulinintumoursofmicetreated
tumoursonly)andithasbeenusedextensivelyasamodelforother withC1A,withnochangeinhistoneH3intumours.Futurestudies
HDACIs, including genomic and imaging studies (Leyton et al, that specifically investigate multiple tolerable dose levels in
2006a;Witteretal,2008;LaBonteetal,2009).Furthermore,there xenograftbearingmice–HCT-116andothertumourtypes–will
isaneedtodeveloptherapiesforcolorectalcancer,whichaccounts confirm or disprove the biological relevance of a-tubulin as a
for10–15%ofallcancersandisthesecondleadingcauseofcancer pharmacodynamic biomarker in tumours. Imaging biomarkers
deathsinwesterncountries.Traditionallyusedfornearly50years, alsoprovideameanstoextendpre-clinicalfindingsintothedesign
5-fluorouracil (5-FU) has led to a small benefit in survival (Buyse of innovative clinical trials of novel agents. In this regard, C1A
et al, 1988). More recently, combination regimens comprising of treatment reduced tumour [18F]FLT uptake by 1.7-fold at 48h,
5-FUleuvorinirinotecan,cetuximaboroxiloplatinhavebeenused suggesting that molecular imaging could provide value in future
to improve clinical efficacy. When tested as monotherapy in pre- studies of this compound. [18F]FLT-PET uptake was previously
clinical models of colon cancer, these compounds demonstrate showntobealteredonlyatlatertimepointswithbroad-spectrum
only modest activity. We hypothesised that treatment with C1A HDACI LAQ-824 (Leyton et al, 2006a). The imaging output was
could offer benefit over conventional chemotherapy. consistentwiththeobservedreductioninKi-67immunostainingin
C1A was designed to have sustained duration of HDAC excised tumour samples.
inhibitory activity via an aromatic nitrogen mustard moiety. In keeping with the preferential HDAC6 activity, the drug
Interestingly,wefoundthatC1AhadpreferentialHDAC6activity showed restricted activity on gene expression in vivo with
in vitro. Molecular modelling analysis suggested favourable o0.065% of genes modulated during 24h of treatment. Notably,
interaction between C1A and HDA6cdII relative to the unfavour- however, although HDAC6 modulation was seen in vivo, we
able binding of the aniline mustard of C1A to a solvent-exposed cannot wholly attribute the changes in gene expression seen after
groove within HDAC1cd. These physicochemical properties of C1AtreatmenttoanHDAC6mechanism,aswecannotcompletely
C1A probably explain preferential acetylation of known HDAC6 rule out accumulation ofC1A into thetumourcells to levels high
substrates, a-tubulin and HSP90 in vitro, compared with SAHA, enoughtoalsoinhibitotherHDACclasses.Thelackofanincrease
which is a pan-HDACI. By making comparisons with SAHA inhistoneH3acetylationtogetherwiththeabsenceoftypicalgene
(being a current clinically approved HDACI, as well as having expression profile characteristic of class I HDAC inhibition, for
sufficiently different selectivity profile (Bradner et al, 2010)), we instance activation of CDKNIA that encodes p21WAF1/CIP1,
show that C1A is mechanistically novel with a mechanism of suggest that the observed gene expression profile was not due to
action more in keeping with the HDAC6 selective inhibitor tool- inhibition of class I HDAC. As per the in vitro studies, the
compoundtubastatinA.PhenotypiceffectsofHDAC6knockdown observationthatC1Ainducedpro-apoptoticgenesincludingXAF1
by shRNA overlapped with those of C1A treatment, including and BAX requires further investigation.
348 www.bjcancer.com|DOI:10.1038/bjc.2012.576
HDAC6inhibitor withantitumour activity BRITISH JOURNALOF CANCER
In conclusion, C1A preferentially inhibits HDAC6 and (HDAC6)-mediatedtubulindeacetylation.ProcNatlAcadSciUSA100(8):
modulates HDAC6 downstream targets, leading to growth 4389–4394.
inhibition of a diverse set of cancer cell lines. This property HubbertC,GuardiolaA,ShaoR,KawaguchiY,ItoA,NixonA,YoshidaM,
together with the favourable PKs and efficacy in vivo makes it a WangXF,YaoTP(2002)HDAC6isamicrotubule-associateddeacetylase.
candidateforfurtherpre-clinicalandclinicaldevelopment.Acetyl- Nature417(6887):455–458.
a-tubulin and [18F]FLT-PET hold promise as pharmacodynamic JanickeRU(2009)MCF-7breastcarcinomacellsdonotexpresscaspase-3.
BreastCancerResTreat117(1):219–221.
and efficacy biomarkers in the clinical testing of C1A. These
KaluzaD,KrollJ,GesierichS,YaoTP,BoonRA,HergenreiderE,TjwaM,
interestingresultsprovidearationaleforthefuturedevelopmentof
RossigL,SetoE,AugustinHG,ZeiherAM,DimmelerS,UrbichC(2011)
HDAC6 inhibitors to treat patients with solid tumours.
ClassIIbHDAC6regulatesendothelialcellmigrationandangiogenesisby
deacetylationofcortactin.EMBOJ30(20):4142–4156.
KawadaJ,ZouP,MazitschekR,BradnerJE,CohenJI(2009)Tubacinkills
ACKNOWLEDGEMENTS Epstein-Barrvirus(EBV)-Burkittlymphomacellsbyinducingreactive
oxygenspeciesandEBVlymphoblastoidcellsbyinducingapoptosis.
JBiolChem284(25):17102–17109.
This study was supported by CR-UK & EPSRC Cancer Imaging
KazantsevAG,ThompsonLM(2008)Therapeuticapplicationofhistone
Centre at Imperial College, London, in association with the
deacetylaseinhibitorsforcentralnervoussystemdisorders.NatRevDrug
MRC and Department of Health (England) grant number
Discov7(10):854–868.
C2536/A10337.
LaBonteMJ,WilsonPM,FazzoneW,GroshenS,LenzHJ,LadnerRD(2009)
DNAmicroarrayprofilingofgenesdifferentiallyregulatedbythehistone
deacetylaseinhibitorsvorinostatandLBH589incoloncancercelllines.
CONFLICT OF INTEREST BMCMedGenomics2:67.
LahmA,PaoliniC,PallaoroM,NardiMC,JonesP,NeddermannP,
SambuciniS,BottomleyMJ,LoSurdoP,CarfiA,KochU,DeFrancescoR,
The authors declare no conflict of interest.
SteinkuhlerC,GallinariP(2007)Unravelingthehiddencatalyticactivity
ofvertebrateclassIIahistonedeacetylases.ProcNatlAcadSciUSA
104(44):17335–17340.
REFERENCES LeeYS,LimKH,GuoX,KawaguchiY,GaoY,BarrientosT,OrdentlichP,
WangXF,CounterCM,YaoTP(2008)Thecytoplasmicdeacetylase
BalasubramanianS,RamosJ,LuoW,SirisawadM,VernerE,BuggyJJ(2008) HDAC6isrequiredforefficientoncogenictumorigenesis.CancerRes
Anovelhistonedeacetylase8(HDAC8)-specificinhibitorPCI-34051 68(18):7561–7569.
inducesapoptosisinT-celllymphomas.Leukemia22(5):1026–1034. LeytonJ,AlaoJP,DaCostaM,StavropoulouAV,LatigoJR,PerumalM,
BaliP,PranpatM,BradnerJ,BalasisM,FiskusW,GuoF,RochaK, PillaiR,HeQ,AtadjaP,LamEW,WorkmanP,VigushinDM,
KumaraswamyS,BoyapalleS,AtadjaP,SetoE,BhallaK(2005)Inhibition AboagyeEO(2006a)Invivobiologicalactivityofthehistone
ofhistonedeacetylase6acetylatesanddisruptsthechaperonefunctionof deacetylaseinhibitorLAQ824isdetectablewith3’-deoxy-3’-
heatshockprotein90:anovelbasisforantileukemiaactivityofhistone [18F]fluorothymidinepositronemissiontomography.CancerRes
deacetylaseinhibitors.JBiolChem280(29):26729–26734. 66(15):7621–7629.
BanerjiU,WaltonM,RaynaudF,GrimshawR,KellandL,ValentiM,Judson LeytonJ,LockleyM,AertsJL,BairdSK,AboagyeEO,LemoineNR,
I,WorkmanP(2005)Pharmacokinetic-pharmacodynamicrelationships McNeishIA(2006b)QuantifyingtheactivityofadenoviralE1ACR2
fortheheatshockprotein90molecularchaperoneinhibitor17- deletionmutantsusingrenillaluciferasebioluminescenceand3’-deoxy-
allylamino,17-demethoxygeldanamycininhumanovariancancer 3’-[18F]fluorothymidinepositronemissiontomographyimaging.
xenograftmodels.ClinCancerRes11(19Pt1):7023–7032. CancerRes66(18):9178–9185.
BoyaultC,SadoulK,PabionM,KhochbinS(2007)HDAC6,atthecrossroads LeytonJ,SmithG,LeesM,PerumalM,NguyenQD,AigbirhioFI,GolovkoO,
betweencytoskeletonandcellsignalingbyacetylationandubiquitination. HeQ,WorkmanP,AboagyeEO(2008)Noninvasiveimagingofcell
Oncogene26(37):5468–5476. proliferationfollowingmitogenicextracellularkinaseinhibitionby
BradnerJE,WestN,GrachanML,GreenbergEF,HaggartySJ,WarnowT, PD0325901.MolCancerTher7(9):3112–3121.
MazitschekR(2010)Chemicalphylogeneticsofhistonedeacetylases. MaloneyA,WorkmanP(2002)HSP90asanewtherapeutictargetforcancer
NatChemBiol6(3):238–243. therapy:thestoryunfolds.ExpertOpinBiolTher2(1):3–24.
BruserudO,StapnesC,ErsvaerE,GjertsenBT,RyningenA(2007)Histone MarksPA,XuWS(2009)Histonedeacetylaseinhibitors:Potentialincancer
deacetylaseinhibitorsincancertreatment:areviewoftheclinicaltoxicity therapy.JCellBiochem107(4):600–608.
andthemodulationofgeneexpressionincancercell.CurrPharm MarsoniS,DamiaG,CamboniG(2008)Aworkinprogress:the
Biotechnol8(6):388–400. clinicaldevelopmentofhistonedeacetylaseinhibitors.Epigenetics3(3):
ButlerKV,KalinJ,BrochierC,VistoliG,LangleyB,KozikowskiAP(2010) 164–171.
Rationaldesignandsimplechemistryyieldasuperior,neuroprotective MatsuyamaA,ShimazuT,SumidaY,SaitoA,YoshimatsuY,Seigneurin-
HDAC6inhibitor,tubastatinA.JAmChemSoc132(31):10842–10846. BernyD,OsadaH,KomatsuY,NishinoN,KhochbinS,HorinouchiS,
BuyseM,Zeleniuch-JacquotteA,ChalmersTC(1988)Adjuvanttherapyof YoshidaM(2002)Invivodestabilizationofdynamicmicrotubulesby
colorectalcancer.Whywestilldon’tknow.JAMA259(24):3571–3578. HDAC6-mediateddeacetylation.EMBOJ21(24):6820–6831.
DuvicM,VuJ(2007)Vorinostat:aneworalhistonedeacetylaseinhibitor NguyenQD,SmithG,GlaserM,PerumalM,ArstadE,AboagyeEO(2009)
approvedforcutaneousT-celllymphoma.ExpertOpinInvestigDrugs Positronemissiontomographyimagingofdrug-inducedtumor
16(7):1111–1120. apoptosiswithacaspase-3/7specific[18F]-labeledisatinsulfonamide.
FinninMS,DonigianJR,CohenA,RichonVM,RifkindRA,MarksPA, ProcNatlAcadSciUSA106(38):16375–16380.
BreslowR,PavletichNP(1999)Structuresofahistonedeacetylase PalijanA,FernandesI,BastienY,TangL,VerwayM,KourelisM,
homologueboundtotheTSAandSAHAinhibitors.Nature401(6749): Tavera-MendozaLE,LiZ,BourdeauV,MaderS,YangXJ,WhiteJH
188–193. (2009)Functionofhistonedeacetylase6asacofactorofnuclearreceptor
FotheringhamS,EppingMT,StimsonL,KhanO,WoodV,PezzellaF, coregulatorLCoR.JBiolChem284(44):30264–30274.
BernardsR,LaThangueNB(2009)Genome-wideloss-of-functionscreen ParmigianiRB,XuWS,Venta-PerezG,Erdjument-BromageH,YanevaM,
revealsanimportantrolefortheproteasomeinHDACinhibitor-induced TempstP,MarksPA(2008)HDAC6isaspecificdeacetylaseof
apoptosis.CancerCell15(1):57–66. peroxiredoxinsandisinvolvedinredoxregulation.ProcNatlAcadSci
HaberlandM,MontgomeryRL,OlsonEN(2009)Themanyrolesofhistone USA105(28):9633–9638.
deacetylasesindevelopmentandphysiology:implicationsfordiseaseand PiekarzRL,FryeR,TurnerM,WrightJJ,AllenSL,KirschbaumMH,ZainJ,
therapy.NatRevGenet10(1):32–42. PrinceHM,LeonardJP,GeskinLJ,ReederC,JoskeD,FiggWD,Gardner
HaggartySJ,KoellerKM,WongJC,GrozingerCM,SchreiberSL(2003) ER,SteinbergSM,JaffeES,Stetler-StevensonM,LadeS,FojoAT,BatesSE
Domain-selectivesmall-moleculeinhibitorofhistonedeacetylase6 (2009)PhaseIImulti-institutionaltrialofthehistonedeacetylaseinhibitor
www.bjcancer.com|DOI:10.1038/bjc.2012.576 349
BRITISHJOURNAL OF CANCER HDAC6 inhibitor withantitumouractivity
romidepsinasmonotherapyforpatientswithcutaneousT-celllymphoma. WilsonAJ,ByunDS,PopovaN,MurrayLB,L’ItalienK,SowaY,ArangoD,
JClinOncol27(32):5410–5417. VelcichA,AugenlichtLH,MariadasonJM(2006)Histonedeacetylase3
RichonVM,SandhoffTW,RifkindRA,MarksPA(2000)Histonedeacetylase (HDAC3)andotherclassIHDACsregulatecoloncellmaturationand
inhibitorselectivelyinducesp21WAF1expressionandgene-associated p21expressionandarederegulatedinhumancoloncancer.JBiolChem
histoneacetylation.ProcNatlAcadSciUSA97(18):10014–10019. 281(19):13548–13558.
RivieccioMA,BrochierC,WillisDE,WalkerBA,D’AnnibaleMA, WittO,DeubzerHE,MildeT,OehmeI(2009)HDACfamily:Whatarethe
McLaughlinK,SiddiqA,KozikowskiAP,JaffreySR,TwissJL,RatanRR, cancerrelevanttargets?CancerLett277(1):8–21.
LangleyB(2009)HDAC6isatargetforprotectionandregeneration WitterDJ,HarringtonP,WilsonKJ,ChenardM,FlemingJC,HainesB,
followinginjuryinthenervoussystem.ProcNatlAcadSciUSA106(46): KralAM,SecristJP,MillerTA(2008)Optimizationofbiaryl
19599–19604. SelectiveHDAC1&2Inhibitors(SHI-1:2).BioorgMedChemLett18(2):
SantoL,HideshimaT,KungAL,TsengJC,TamangD,YangM,JarpeM,van 726–731.
DuzerJH,MazitschekR,OgierWC,CirsteaD,RodigS,EdaH,ScullenT, WorkmanP,AboagyeEO,BalkwillF,BalmainA,BruderG,ChaplinDJ,
CanaveseM,BradnerJ,AndersonKC,JonesSS,RajeN(2012)Preclinical DoubleJA,EverittJ,FarninghamDA,GlennieMJ,KellandLR,Robinson
activity,pharmacodynamic,andpharmacokineticpropertiesofaselective V,StratfordIJ,TozerGM,WatsonS,WedgeSR,EcclesSA(2010)
HDAC6inhibitor,ACY-1215,incombinationwithbortezomibin Guidelinesforthewelfareanduseofanimalsincancerresearch.
multiplemyeloma.Blood119(11):2579–2589. BrJCancer102(11):1555–1577.
SchultzBE,MisialekS,WuJ,TangJ,ConnMT,TahilramaniR,WongL XuWS,ParmigianiRB,MarksPA(2007)Histonedeacetylase
(2004)KineticsandcomparativereactivityofhumanclassIandclassIIb inhibitors:molecularmechanismsofaction.Oncogene26(37):
histonedeacetylases.Biochemistry43(34):11083–11091. 5541–5552.
ScrogginsBT,RobzykK,WangD,MarcuMG,TsutsumiS,BeebeK,Cotter ZhangX,YuanZ,ZhangY,YongS,Salas-BurgosA,KoomenJ,OlashawN,
RJ,FeltsS,ToftD,KarnitzL,RosenN,NeckersL(2007)Anacetylation ParsonsJT,YangXJ,DentSR,YaoTP,LaneWS,SetoE(2007)HDAC6
siteinthemiddledomainofHsp90regulateschaperonefunction.MolCell modulatescellmotilitybyalteringtheacetylationlevelofcortactin.
25(1):151–159. MolCell27(2):197–213.
ShieldsAF,GriersonJR,DohmenBM,MachullaHJ,StayanoffJC,Lawhorn-
CrewsJM,ObradovichJE,MuzikO,MangnerTJ(1998)Imaging This work is licensed under the Creative Commons
proliferationinvivowith[F-18]FLTandpositronemissiontomography.
Attribution-NonCommercial-ShareAlike3.0Unported
NatMed4(11):1334–1336.
License.Toviewacopyofthislicense,visithttp://creativecommons.
VichaiV,KirtikaraK(2006)SulforhodamineBcolorimetricassayfor
org/licenses/by-nc-sa/3.0/
cytotoxicityscreening.NatProtoc1(3):1112–1116.
Supplementary Information accompanies this paper on British Journal of Cancer website (http://www.nature.com/bjc)
350 www.bjcancer.com|DOI:10.1038/bjc.2012.576