Detection of Histone Deacetylase Inhibition by Noninvasive Magnetic Resonance Spectroscopy

1325

Detection of histone deacetylase inhibition by noninvasive magnetic resonance spectroscopy

Madhuri Sankaranarayanapillai,1 William P. Tong,1 can be used to monitor HDAC activity in cells.In principle, David S. Maxwell,1 Ashutosh Pal,1 Jihai Pang,2 this could be applied in vivo to noninvasively monitor William G. Bornmann,1 Juri G. Gelovani,1 HDAC activity.[Mol Cancer Ther 2006;5(5):1325–34] and Sabrina M. Ronen1

1Experimental Diagnostic Imaging and 2Pharmaceutical Introduction Development Center, The University of Texas M.D. Anderson Acetylation and deacetylation of nucleosomal core histones Cancer Center, Houston, Texas play an important role in the modulation of chromatin structure and the regulation of gene expression. The Abstract acetylation status of histones is controlled by the activities Histone deacetylase (HDAC) inhibitors are new and of histone acetyltransferases and histone deacetylases promising antineoplastic agents.Current methods for (HDAC), which, respectively, catalyze removal or addition q monitoring early response rely on invasive biopsies or of acetyl groups onto the -amino of lysine residues in the indirect blood-derived markers.Our goal was to develop a histone tail. Disruptions in HDACs and histone acetyl- magnetic resonance spectroscopy (MRS)–based method transferases have been associated with cancer development to detect HDAC inhibition.The fluorinated lysine derivative (1, 2). Conversely, it has been shown that HDAC inhibitors Boc-Lys-(Tfa)-OH (BLT) was investigated as a 19F MRS (HDACI) lead to differentiation, growth arrest, and/or molecular marker of HDAC activity together with 31P MRS apoptosis in treated cells and tumors (1, 3–6). Consequent- of endogenous metabolites. In silico modeling of the BLT- ly, HDACIs are currently in clinical trials and show HDAC interaction and in vitro MRS studies of BLT promising results in several different tumor types (1, 2, cleavage by HDAC confirmed BLT as a HDAC substrate. 7–12). The exact mechanism of action of HDACIs is not BLT did not affect cell viability or HDAC activity in PC3 entirely clear. Reactivation of silenced tumor suppressor prostate cancer cells.PC3 cells were treated, in the genes occurs following HDAC inhibition in some cases presence of BLT, with the HDAC inhibitor p-fluoro- (13, 14). However, HDACIs can lead not only to gene suberoylanilide hydroxamic acid (FSAHA) over the range stimulation but also to gene repression (15, 16). Nonethe- of 0 to 10 Mmol/L, and HDAC activity and MRS spectra less, many of the modulated genes mediate proliferation, were monitored.Following FSAHA treatment, HDAC cell cycle progression, or apoptosis and include p21/ activity dropped, reaching 53% of control at 10 Mmol/L WAF1, caspases, p53, vascular endothelial growth factor, neu FSAHA.In parallel, a steady increase in intracellular BLT HER-2/ , and bcr/abl (17–21). In addition, acetylation of from 14 to 32 fmol/cell was observed.BLT levels nonhistone proteins is likely involved in the activity of negatively correlated with HDAC activity consistent with HDACIs, and HDAC substrates include pRb, E2F, and heat higher levels of uncleaved BLT in cells with inhibited shock protein (Hsp) 90 (22). HDAC.Phosphocholine, detected by 31P MRS, increased As mentioned, several HDACIs are currently in clinical from 7 to 16 fmol/cell following treatment with FSAHA and trials. However, at present, there is no direct noninvasive also negatively correlated with HDAC activity.Increased means to measure drug delivery to the tumor tissue, drug- phosphocholine is probably due to heat shock protein target interaction, or molecular response. Response to 90 inhibition as indicated by depletion of client proteins. HDACIs in clinical trials is correlated with acetylation In summary, 19F MRS of BLT, combined with 31P MRS, of peripheral blood mononuclear cells or acetylation of histones in tumor biopsy specimens (7, 9, 10, 12). Blood tests are well tolerated by patients but provide only an indirect indicator of drug delivery and activity at the tumor Received 11/28/05; revised 1/20/06; accepted 3/8/06. site. Biopsies reliably assess drug action but are surgically Grant support: National Cancer Institute core institutional grant P30 invasive. A further difficulty is that response in many cases CA016672 30 for the support of Core NMR Facility and Core Analytical is associated with tumor stasis, rather than shrinkage (7), Facility (The University of Texas M.D. Anderson Cancer Center). limiting the use of traditional imaging methods. Determin- The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked ing the appropriate, biologically relevant, drug dose and advertisement in accordance with 18 U.S.C. Section 1734 solely to assessing drug action at the tumor site, in vivo, present a indicate this fact. challenge (23, 24). A noninvasive method of assessing drug Requests for reprints: Sabrina M. Ronen, Experimental Diagnostic Imaging, delivery to, and the effect on, the intended molecular target The University of Texas M.D. Anderson Cancer Center, 57-3D, 1515 Holcombe Blvd., Houston, TX 77030-4009. Phone: 713-563-4617; is therefore needed. Fax: 713-563-4894. E-mail: [email protected] Magnetic resonance spectroscopy (MRS) presents a Copyright C 2006 American Association for Cancer Research. noninvasive nondestructive method, which can provide doi:10.1158/1535-7163.MCT-05-0494 longitudinal pharmacokinetic and pharmacodynamic

Mol Cancer Ther 2006;5(5). May 2006

biomarkers of drug delivery and action at defined anatomic structure of BLT was drawn into Sybyl using the Sketch locations in individual cancer patients. 19F MRS has been module, and types were modified to correspond to the used in studies of fluorinated chemotherapeutic agents in protonated state at pH 7. Charges were not assigned to cells, animal models, and patients (25–27) and also the molecule, because FlexX uses formal charges that are provides a tool to assess different physiologic variables, assigned during the actual run. The crystal structure of including oxygenation, pH, and gene expression (28–30). HDAC8 complexed with suberoylanilide hydroxamic acid In addition, MRS can monitor changes in cellular metab- (SAHA) was retrieved from the RCSB (ID code 1T69; ref. olites that are associated with clinical response to tradi- 44). The binding site was then defined as 6.5 A˚ around the tional chemotherapy or radiotherapy. An increase in SAHA ligand. In the customized setting, the zinc atom was choline-containing metabolites, as detected using either added as a template. For H142 and H143, the histidines 31Por1H MRS, is associated with cell transformation, and a were selected to be in the y protonated state. This particular drop in those metabolites is typically associated with protonation state was found to be necessary for proper response to treatment (31–33). More recently, 31PMRS docking of the hydroxamic acid–based inhibitors. The has been used to identify biomarkers of response to novel CSCORE method (45) was used in the ranking of the targeted therapies. We have shown that MRS can detect 30 requested configurations. several metabolic changes associated with inhibition of the MRS Studies of BLT Cleavage mitogen-activated protein kinase pathway (34, 35) and the To confirm that MRS can be used to monitor cleavage of phosphatidylinositol 3-kinase pathway (36) as well as the HDAC substrate, 0.6 mmol/L BLT (Advance Chem- response to the Hsp90 inhibitor 17-allylamino-17-deme- Tech, Louisville, KY) was incubated alone or with 28 units thoxygeldanamycin (17AAG; refs. 37, 38). recombinant HDAC-8 (Biomol, Plymouth Meeting, PA) in a Several of the genes and proteins modulated by HDAC total volume of 500 AL HDAC assay buffer (Biomol; inhibition are expected to lead to MRS detectable changes. composed of 25 mmol/L Tris-HCl, 137 mmol/L NaCl, These include down-regulation of receptor tyrosine kinases 2.7 mmol/L KCl, 1 mmol/L MgCl2, formulated to maintain and their downstream effector molecules (16, 17, 20, 21, 39), HDAC activity). 19F MRS was used to monitor the decrease which have been associated with a drop in phosphocholine in BLT and buildup of the cleavage product TFA by (34, 36, 40); modulation of p53 (19), which could affect acquiring 1H-decoupled spectra at 10-minute intervals on phosphocholine levels (41); and Hsp90 acetylation and an Avance DPX300 Bruker spectrometer (Bruker Biospin, inhibition (22, 42, 43), which could lead to increased Rheinstetten, Germany) using a 30j flip angle, 3-second phosphocholine and glycerophosphocholine (37). Thus, relaxation delay, and 128 scans. A sealed insert containing 31 P MRS could be used to monitor metabolic changes C6F6 served as a quantification and chemical shift reference À associated with inhibition of HDAC. However, any ( 164.9 ppm relative to CFCl3). metabolic changes observed in the 31P spectrum represent Cell Culture, Cell Proliferation, and HDAC Activity indirect and often nonspecific downstream events. Specific PC3 human prostate cancer cells were routinely cultured direct indicators of drug activity on the intended molecular in DMEM/F-12 (Life Technologies, Grand Island, NY) target are therefore needed to complement the downstream supplemented with 10% FCS (Hyclone, Logan, UT) and metabolic changes. 10,000 units/mL penicillin, 10,000 Ag/mL streptomycin, We show here for the first time the use of a fluorinated and 25 Ag/mol amphotericin B (Life Technologies) at 37jC HDAC substrate, the lysine derivative Boc-Lys-(Tfa)- in 5% CO2. OH (BLT), as a specific spectroscopic indicator to directly To assess the effect on cell proliferation of BLT, the monitor HDAC inhibition. We show that intracellular BLT colorimetric WST-1 cell proliferation assay (Roche, Indian- levels, as determined by 19F MRS, are negatively correlated apolis, IN) was used following the manufacturer’s instruc- with HDAC activity. In addition, using 31P MRS, we show tions. Briefly, 20 Â 103 cells per well were seeded in 96-well that phosphocholine levels also negatively correlate with plates. Cells were then treated for 24 hours with BLT at HDAC activity. Our results indicate that MRS can be used 5, 10, 20, 50, 100, 200, 500 Amol/L and 1, 2, 5, and as a method for assessing HDAC inhibition and its 10 mmol/L or with matched DMSO (1:20,000 to 1:10). downstream signaling and metabolic effects. This method- Subsequently, cells were incubated for 4 hours with the ology could eventually be translated to the clinic where WST-1 cell proliferation reagent and absorbance was read drug delivery and molecular activity would be monitored at 440 nm using a Tecan Freedom Evo liquid handler noninvasively over time leading to optimized drug equipped with the SAFIRE monochromator-based micro- scheduling and dosing. plate reader (Tecan U.S., Research Triangle Park, NC). The effect on cell proliferation of p-fluoro-SAHA (FSAHA; synthesized in-house based on a previously Materials and Methods described method; ref. 46), the fluorinated derivative of In silico Modeling of BLT Docking into HDAC the clinically relevant HDACI SAHA, was also determined Docking was done using the FlexX 1.13.5 software as using the WST-1 assay as described above. Cells were provided in Sybyl7.1 (Tripos, Inc., St. Louis, MO) running treated with 2, 5, and 10 Amol/L FSAHA either in the on a 4-Processor R16000 SGI Tezro. Unless otherwise noted, presence of 1 mmol/L BLT or in the presence of DMSO all defaults were used in the docking experiment. The (in which BLT had to be dissolved first due to its low

Mol Cancer Ther 2006;5(5). May 2006

solubility in growth medium). Note that controls were internal methylene diphosphonic acid reference, normaliz- treated with DMSO to clearly identify effects that are due to ing to cell number and correcting for saturation effects the compound being investigated rather than its vehicle (correction factors were determined as above by acquiring a (DMSO). FSAHA dose was based on previous findings (4). fully relaxed quantitative spectrum using a 90j flip angle The effect on HDAC activity of BLT and the HDACIs and a 30-second relaxation delay). FSAHA and SAHA (also synthesized in-house based on the To monitor the fluorinated metabolites, the lipid phase A previously described method; ref. 46) was determined was reconstituted in 500 LCDCl3. To monitor the using the Fluor-de-Lys fluorometric assay (Biomol) follow- fluorinated metabolites in cellular protein, the protein ing the manufacturer’s instructions. Briefly, 20 Â 103 cells pellet obtained during cell extraction was dissolved in per well were seeded in 96-well plates and incubated for 1 mL of 0.5 mol/L NaOH and heated to 60jC for 1 hour. 24 hours with (a) 1 mmol/L BLT, (b) 2, 5, and 10 Amol/L Samples were then analyzed by 19F MRS as above. FSAHA, (c) 2, 5, and 10 Amol/L SAHA, (d)2,5,7,8,9,and Analysis of TFA in Extracellular Medium 10 Amol/L FSAHA in the presence of 1 mmol/L BLT, and To assess buildup of TFA in medium, samples of extra- (e)2,5,7,8,9,and10Amol/L FSAHA in the presence of cellular medium were collected and analyzed using gas vehicle (DMSO). The Fluor-de-Lys substrate was then chromatography-mass spectroscopy. To be able to use stan- added for 1 hour, medium was removed, cells were rinsed dard capillary gas chromatography-mass spectroscopy, with PBS and incubated for 10 minutes with the Fluor-de- TFA had to be derivatized. This was done using a modi- Lys developer, and fluorescence was read at 460 nm using fication of the procedure of Scott et al. (48) as follows. the Tecan microplate reader as above. Results were Sample (1 mL) or TFA standard in medium was treated normalized to cell density as determined using the WST-1 with 20 mg NaCl, 60 AL HCl, and 30 AL each of 100 mmol/L assay in the same 96 well plate. 1,3-dicyclohexylcarbohiimide (Sigma-Aldrich Co., St. Louis, MRS Studies of HDAC Activity MO) and 2,4-difluoroaniline (Sigma-Aldrich) and adjusted For MRS studies, PC3 cells were treated for 24 hours with to a final volume of 1.320 mL in ethyl acetate. Mixture was FSAHA at 2, 5, 7, 8, 9, and 10 Amol/L in the presence of vortexed for 60 minutes at room temperature, an additional 1 mmol/L BLT or with 1 mmol/L BLT alone. Cells (f1 Â 50 mg NaCl was added, and sample was briefly vortexed. 107 –1.5 Â 107) were then extracted using the dual-phase Following phase separation, the organic layer was removed extraction method as described previously (37, 47). Briefly, and stored. The aqueous phase was further extracted twice cells were extensively rinsed with ice-cold saline to remove with 200 AL ethyl acetate. All three ethyl acetate extracts any residual extracellular BLT and medium. Cells were were combined, treated with 50 AL of 3 mol/L HCl and then fixed in 10 mL ice-cold methanol, scraped off the 50 AL saturated anhydrous sodium sulfate, and vortexed surface of the culture flask, collected into glass tubes, and and another 20 mg anhydrous sodium sulfate was added vortexed. Ice-cold chloroform (10 mL) was then added and mixed. The organic phase was then evaporated to followed by 10 mL ice-cold deionized water. Following dryness under a dry nitrogen gas stream. The residue was phase separation and solvent removal, the water-soluble dissolved in 250 AL toluene and transferred into sample fraction was reconstituted in 250 AL deuterium oxide and vials for analysis. Sample extract (3 AL) was analyzed 250 AL DMSO for 19F magnetic resonance measurements. using an Agilent (Wilmington, DE) 6890N GC coupled to To perform the 31P magnetic resonance measurement, 5973N MSD in splitless injection mode. TFA was resolved 100 AL EDTA and 50 AL methylene diphosphonic acid in using a Supleco (St. Louis, MO) Omegawax 250 capillary deuterium oxide were added to a final concentration of column (30 m Â 0.25 mm). Column temperature was 100jC 10 and 0.35 mmol/L, respectively. The number of cells for 1 minute, ramping to 230jCat25jC/min, and then held extracted was determined by counting a separate flask of at 230jC for 2 minutes. Detection was done in EI-positive cells. 19F MRS of the water-soluble metabolites were mode monitoring the m/z 225 ion. recorded as above. Metabolite concentrations were deter- Western Blot Analysis of Protein Levels mined by integration and comparison with the area of the PC3 cells were lysed using cell lysis buffer [0.1% NP40, external C6F6 reference, normalizing to cell number and 50 mmol/L HEPES (pH 7.4), 250 mmol/L NaCl, 1 mmol/L correcting for saturation effects. Correction for saturation DTT, 1 mmol/L EDTA, 1 mmol/L NaF, 10 mmol/L effects was achieved by also acquiring a quantitative h-glycerophosphate, 0.1 mmol/L sodium orthovanadate, inverse gated fully relaxed spectrum on two different 1 AL/mL protease inhibitor cocktail set III (Calbiochem, La samples and calculating the correction factors that need to Jolla, CA), 1 mmol/L phenylmethylsulfonyl fluoride]. be applied to the partially relaxed spectra. We determined Lysates were centrifuged at 12,000 rpm for 10 minutes T j that the 1 relaxation of BLT is 1 second and that of C6F6 is (4 C), the protein supernatant was collected, and total 3 seconds. The fully relaxed spectrum was therefore protein concentrations were determined using Bio-Rad DC acquired using a 90j flip angle and a 15-second relaxation protein assay reagents (Bio-Rad, Hercules, CA). Proteins T 31 delay (5 times the longest 1). P MR spectra were were separated by SDS-PAGE using 10% gels and recorded on an Avance DRX500 Bruker spectrometer transferred electrophoretically to 0.45 Am nitrocellulose (Bruker Biospin) using a 30j flip angle and 3-second membranes. Membranes were blocked in blocking buffer relaxation delay. Metabolite concentrations were deter- containing 5% nonfat dry milk in TBS (pH 7.6) and 0.1% mined by integration and comparison with the area of the Tween 20 and incubated overnight at 4jC with primary

Mol Cancer Ther 2006;5(5). May 2006

antibodies as follows: c-Raf (1:1,000; Cell Signaling Tech- The known structure of HDAC8 was used (44). The top- nology, Danvers, MA), cyclin-dependent kinase 4 (cdk4; ranking consensus scored configuration of BLT docked with 1:2,000; Cell Signaling Technology), Hsp70 (1:2,000; Stress- HDAC8 is shown in Fig. 1. The interaction of the carbonyl gen, Victoria, British Columbia, Canada), and glyceralde- group with the Y306 is consistent with the proposed model hyde-3-phosphate dehydrogenase (1:5,000; Stressgen). of how acetylated lysine would interact in the catalytic site This was followed by 1-hour incubation with horseradish (49). In the docked structure, there is a hydrogen bond to peroxidase–conjugated secondary anti-rabbit (Cell Signal- the backbone carbonyl of G151 and a hydrogen bond ing Technology) and anti-mouse (Cell Signaling Technol- between side chain NH and D101. The aliphatic side chain ogy) antibodies at dilutions of 1:1,000 and 1:2,000, interacts with two phenylalanine groups (F152 and F208) respectively. Membranes were washed with enhanced and the interaction is seen in the complex of the SAHA as chemiluminescence reagents (LumiGLO & Peroxide, Cell well (44). The Boc group consistently selected to position Signaling Technology) for 1 minute and exposed to nearby F207 and F208 rather than the area occupied in the Hyperfilm (Amersham Biosciences, Piscataway, NJ), which reported complex with SAHA. This may result from the was developed on a Konica SRX-101 automatic developer asymmetry that is present within the BLT and not within (Konica, Tokyo, Japan). SAHA, which forces the choice between optimizing the Statistical Analysis orientation of the carboxyl group and the Boc group. All results represent the average of at least three Next, 19F MRS was used to confirm that recombinant experiments and are expressed as mean FSD. Data were HDAC8 does indeed cleave BLT in vitro. As illustrated analyzed by Wilcoxon-Mann-Whitney rank test and P < in Fig. 2, a drop in BLT levels was detected over time 0.05 was considered significant. A Spearman’s rank accompanied by an increase in TFA levels, consistent with correlation was used to analyze correlations. KaleidaGraph cleavage of BLT by HDAC8 to form TFA and the 19F (Synergy Software, Essex Junction, VT) and Statistica magnetic resonance invisible Boc-Lys. No changes in BLT (Statsoft, Tulsa, OK) software were used. levels or any buildup of TFA could be detected in the control sample, which contained no HDAC8. This confirmed that BLT is indeed a substrate of HDAC8 and that Results its cleavage by HDAC can be monitored by MRS. BLT Is a HDAC Substrate BLT Does Not Affect CellViabilityor HDAC Activity First, it was necessary to test the hypothesis that a Before using BLT as a marker of HDAC activity in cells, it fluorinated compound composed of a modified lysine could was necessary to rule out its toxicity. The WST-1 assay was serve as a MRS-detectable substrate of HDAC and could used to investigate the effect on PC3 cells of a range of BLT therefore be used to assess HDAC inhibition. We investi- concentrations from 5 Amol/L to 10 mmol/L compared gated the commercially available BLT and first performed with matched DMSO controls. BLT did not significantly computer modeling to estimate the BLT-HDAC interaction. affect cell proliferation compared with controls up to a

Figure 1. HDAC8 crystal structure with docked BLT into catalytic site. Hydrogen bonds are designated with a line. Red, oxygen; blue, nitrogen; pink, fluorine. The element zinc is gold. Inset, BLT.

Mol Cancer Ther 2006;5(5). May 2006

of HDAC activity (58 F 14% at 5 Amol/L and 41 F 8% at 10 Amol/L; P < 0.03). Again, the presence of BLT did not significantly alter the effect of FSAHA treatment (down to 88 F 12%, 69 F 9%, and 53 F 5% for 2, 5, and 10 Amol/L, respectively; P > 0.1 relative to FSAHA). Because SAHA is being investigated in clinical trials, it was necessary to confirm that the biological effects of its fluorinated derivative FSAHA were comparable with those of SAHA. The effects of SAHA and FSAHA on HDAC activity were compared at 2, 5, and 10 Amol/L. The effect of SAHA on HDAC activity was comparable with that observed with FSAHA at 2 and 5 Amol/L (84 F 10% at 2 Amol/L and 51 F 3% at 5 Amol/L; P > 0.7 relative to FSAHA). However, treatment with 10 Amol/L SAHA resulted in a greater inhibition of HDAC activity compared with 10 Amol/L FSAHA (29 F 3%; P < 0.03 relative to FSAHA). Nonetheless, we reasoned that if MRS can detect Figure 2. Sequential 19F MRS spectra of BLT in the presence of recombinant HDAC8 in vitro. Spectra are the result of 128 1H-decoupled the effect of FSAHA on HDAC activity, it would also be scans acquired using a 30j flip angle and a 3-second relaxation delay. able to detect the potentially larger effect of SAHA. Spectra show a drop in BLT as it is cleaved by HDAC8 to form TFA. Inset, 19F MRS of BLT Can Be Used to Assess HDAC Inhibi- full time course of the experiment. tion in Cells Figure 4A illustrates the 19F MRS spectra recorded from concentration of 10 mmol/L (P < 0.03 for 10 mmol/L and extracts of PC3 cells. Control cells cultured in the presence P > 0.4 for all other concentrations from 5 Amol/L to 5 mmol/L; data not shown). We therefore chose to perform MRS experiments with a BLT concentration of 1 mmol/L, which was expected to lead to a MRS detectable signal. Cell numbers following 24-hour treatment with 1 mmol/L BLT represented 95 F 4% of controls (P > 0.1). Next, it was necessary to confirm that 1 mmol/L BLT did not affect HDAC activity in cells. Using the Fluor-de-Lys assay, we determined that incubation of PC3 cells for 24 hours with 1 mmol/L BLT resulted in HDAC activity levels of 102 F 9% (P > 0.3) relative to DMSO-treated controls, indicating no statistically significant effect of 1 mmol/L BLT on HDAC activity. Inhibition of HDAC Activityand Cell Proliferation by HDACI Treatment Is Not Affected bythe Addition of BLT Before using MRS of BLT to assess the effect of HDACIs, it was necessary to confirm that the addition of BLT will not modify the biological effects of the HDACI. To this end, we investigated both cell proliferation and HDAC activity. Figure 3A illustrates our findings. Treatment with FSAHA, the fluorinated derivative of the HDACI SAHA, resulted in a significant drop in cell proliferation relative to control at all three doses investigated down to 87 F 5% at 2 Amol/L, 79 F 3% at 5 Amol/L, and 66 F 6% at 10 Amol/L (P <0.03for all three doses). Importantly, the presence of BLT did not further affect cell proliferation. In the presence of BLT, cell proliferation dropped to 86 F 1% at 2 Amol/L, 77 F 7% at 5 Amol/L, and 63 F 2% at 10 Amol/L (P < 0.03 relative to controls and P > 0.5 relative to FSAHA). Figure 3B illustrates the effect on HDAC activity of FSAHA and FSAHA in Figure 3. Effect on cell proliferation (A) and HDAC activity (B) of HDACI the presence of 1 mmol/L BLT. The lower concentration of treatment. PC3 cells were treated with 2, 5, and 10 Amol/L FSAHA or with FSAHA BLT FSAHA (2 Amol/L) did not lead to a statistically significant FSAHA in the presence of 1 mmol/L BLT ( + ). Cell proliferation F was determined using the WST-1 assay and HDAC activity was drop in HDAC activity (93 16%), but the higher determined using the Fluor-de-Lys assay. Bars, SD. *, P < 0.05, concentrations of FSAHA resulted in significant inhibition compared with DMSO-treated controls. n.s., not significant (P > 0.05).

Mol Cancer Ther 2006;5(5). May 2006

Figure 4.A, representative 1H- decoupled 19F MR spectra of PC3 cell extracts. Cells were treated for 24 h with 10 Amol/L FSAHA in the presence of 1 mmol/L BLT (top)or with 1 mmol/L BLT alone (bottom). Average BLT content in the FSAHA- treated cells was 32 fmol/cell compared with 14 fmol/cell in controls. Spectra are the result of 128 scans acquired using a 30j flip angle and a 3-second relaxation delay. Reference (Ref) was C6F6. B, intracellular BLT levels as a function of HDAC inhibition. BLT levels were determined by 19FMRSasinA. HDAC inhibition was determined using the Fluor- de-Lys assay. Bars, SD. Line, best linear fit. Spearman’s rank correlation indicates that BLT levels negatively correlated with HDAC activity (q = À0.75; P < 0.05).

of 1 mmol/L BLT contained 14 F 4 fmol/cell BLT. This 31P MRS Can Be Used to Assess HDAC Inhibition in value increased significantly to 32 F 4 fmol/cell in cells Cells treated with 10 Amol/L FSAHA in the presence of BLT (P < Because HDAC inhibition leads to modulation of several 0.0002), consistent with intracellular uncleaved BLT levels genes that are associated with MRS detectable metabolic being higher when HDAC is inhibited. No signal could be changes, we questioned whether response to treatment detected from FSAHA, which was expected to resonate at with HDACIs is also detectable in the 31P MRS of treated À121 ppm. In addition, no 19F signal was observed from the cells. 31P MRS was used to investigate the same PC3 lipid phase or the protein pellet of the cells. samples investigated by 19F MRS. As illustrated in Fig. 5A, Interestingly, TFA levels observed in both treated and treatment with 10 Amol/L FSAHA resulted in a significant control cells remained, within experimental error, un- increase in phosphocholine levels from 7 F 1to16F 2 changed. The average TFA concentration observed in fmol/cell (P < 0.01). None of the other metabolites control cells was 1.1 F 0.6 versus 1 F 1 fmol/cell in cells observed in the 31P MRS spectrum were altered. As in treated with 10 Amol/L FSAHA. We therefore speculated BLT, phosphocholine levels for cells treated with lower that TFA produced inside the cell was removed into the concentrations of FSAHA also showed an increase relative extracellular compartment. To test this hypothesis, gas to controls, reaching statistical significance when HDAC chromatography-mass spectroscopy was used to determine activity dropped to 74% relative to controls. Phosphocho- the levels of TFA present in cellular growth medium. In line levels also negatively correlated with HDAC activity control cells, the level of TFA was 5 F 0.4 Ag/mL but was (q = À0.86; P < 0.02; Fig. 5B). only 3.4 F 0.7 Ag/mL in the medium obtained from 10 31P MRS Changes Are Consistent with Depletion of Amol/L FSAHA-treated cells (P < 0.03). Hsp90 Client Proteins cdk4 and c-Raf-1 To confirm that the intracellular BLT levels detected by The increase in phosphocholine observed in our spectra MRS are indicative of inhibition of cellular HDAC, we next following response to HDACI treatment is an unusual monitored HDAC activity and cellular BLT levels in cells observation and only previously reported following treat- treated with a range of FSAHA concentrations from 2 to 10 ment with the Hsp90 inhibitor 17AAG (37). Hsp90 Amol/L. HDAC activity dropped significantly for all inhibition has also been reported following HDACI concentrations greater than 2 Amol/L. An increase in BLT treatment. To test the hypothesis that the 31P MRS changes levels was observed for all FSAHA concentrations investi- observed here were associated with inhibition of Hsp90, we gated and reached statistical significance when HDAC monitored the levels of the Hsp90 client proteins c-Raf-1 activity dropped to 74% relative to controls. Furthermore, a and cdk4. Figure 6 indicates depletion of these two Hsp90 negative correlation was observed between HDAC activity client proteins following treatment with FSAHA. However, and the level of intracellular BLT (q = À0.75; P < 0.05; induction of Hsp70, which has also been reported following Fig. 4B). inhibition of 17AAG, was not observed in our cells (Fig. 6).

Mol Cancer Ther 2006;5(5). May 2006

Discussion nously administered fluorine-containing compounds are The recent shift to development of antineoplastic therapies observed without interference (28). By introducing a that target specific alterations in cancerous cells has been fluorinated HDAC substrate, it is therefore straightforward fueled by the clinical successes of tyrosine kinase inhibitors, to monitor its fate and thus assess HDAC activity directly such as imatinib, gefinitib, and trastuzumab (50, 51). in the target tissue. 31 HDACIs are a novel class of anticancer drugs that target We have shown previously that P MRS provides a HDACs. Modulations in gene expression resulting from the noninvasive method for the detection of metabolic bio- activities of HDACs and their counterpart histone acetyl- markers associated with response to targeted therapies (34, transferases have been associated with cell transformation 37). Here, we have applied this methodology to monitor the (52, 53). Conversely, response to HDACIs has been downstream metabolic effects correlated with HDAC 19 associated with alterations in expression of several onco- inhibition, complementing the use of F MRS to monitor genes and results in induction of differentiation, growth drug activity. We show that phosphocholine levels increase arrest, or apoptosis (1–3, 16, 54). Response to HDACIs in following HDAC inhibition and are correlated with the clinical trials has been promising (1, 2, 7, 11, 12). level of this inhibition. This provides a downstream Unfortunately, current methods to monitor response metabolic biomarker of tumor response to HDACI treat- in vivo rely on either surgically invasive biopsies or ment, further confirming activity of the drug on its target. blood-derived markers (7, 12). There is therefore a need MRS is a noninvasive method that can be readily for a direct, reliable, and noninvasive method to monitor translated to the clinic. Our investigations therefore drug delivery and response to HDACIs at the tumor site. concentrated on a derivative of the clinically relevant Here, we have investigated the fluorinated lysine HDACI SAHA (10). We chose to concentrate our studies derivative BLT as a MRS marker of HDAC activity. BLT on FSAHA, rather than SAHA, because we reasoned that if is detectable by 19F MRS and its cleavage by HDAC was a significant level of FSAHA accumulates intracellularly it expected to produce TFA and Boc-Lys. In silico and in vitro would be possible to simultaneously monitor both the studies confirmed that BLT is a substrate of HDAC8 and delivery of FSAHA and its effect on HDAC activity by 19F therefore most likely a substrate of other class I and II MRS. FSAHA could not be detected in any of our spectra. HDACs. BLT did not affect cell viability or HDAC activity. Therefore, we conclude that the intracellular level of Most importantly, we show here that the intracellular levels FSAHA is below MRS detection level (f0.1 fmol/cell). of BLT, as measured by 19F MRS, are correlated with Current phase I trials (10) found that the mean plasma cellular HDAC activity. concentration of SAHA in vivo was <600 mg/mL, which is Fluorine in the body is in the form of solid fluorides with also expected to be below detection level. T very short 2 relaxation times, producing wide and The inhibitory effect of FSAHA was slightly lower at 10 virtually nondetectable MRS peaks. In vivo 19FMRS Amol/L than that of SAHA. We have not investigated the therefore presents the advantage that there are no naturally reasons for this difference. Nonetheless, this observation observable fluorinated molecules. Consequently, exoge- does not affect the value of our MRS findings. Because we

Figure 5.A, representative 1H- decoupled 31P MR spectra of PC3 cell extracts. Cells were treated for 24 h with 10 Amol/L FSAHA in the presence of 1 mmol/L BLT (top)or with 1 mmol/L BLT alone (bottom). Average phosphocholine (PC) content in the FSAHA-treated cells was 15 fmol/cell compared with 7fmol/ cell in controls. Spectra are the result of 3,000 scans acquired using a30j flip angle and a 3-second relaxation delay. Reference was methylene diphosphonic acid. B, phosphocholine levels as a function of HDAC inhibition. Phosphocholine levels were determined by 31PMRS as in A. HDAC inhibition was determined using the Fluor-de-Lys assay. Bars, SD. Line, best linear fit. Spear- man’s rank correlation indicates that phosphocholine levels negatively correlated with HDAC activity (q = À0.86; P = 0.02).

Mol Cancer Ther 2006;5(5). May 2006

increased Hsp90 acetylation also leading to inhibition of its activity and depletion of client proteins (39, 43, 59). In the specific case of SAHA, a drop in both cdk4 and c-Raf-1 has been observed in some cases (16) but not in others (60). Our Western blot analysis indicates depletion of both cdk4 and c-Raf-1 in treated cells. However, we did not observe any up-regulation of Hsp70. Thus, it is not entirely clear if the depletion of cdk4 and c-Raf-1 is a direct result of HDAC inhibition or occurs subsequent to Hsp90 acetylation following HDAC inhibition. Nonetheless, we believe that the increase in phosphocholine observed by MRS is associated with the depletion of cdk4 and c-Raf-1. The mechanism linking cdk4 and c-Raf-1 with modulation of phosphocholine remains to be elucidated, but the results described here following HDACI treatment are entirely Figure 6. Representative Western blot analysis of c-Raf-1, cdk4, consistent with our earlier observations (37). It should be Hsp70, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). PC3 noted that we have also previously observed an increase in cells were treated with vehicle (DMSO), 1 mmol/L BLT, or 10 Amol/L glycerophosphocholine following response to 17AAG. In FSAHA in the presence of 1 mmol/L BLT. Glyceraldehyde-3-phosphate dehydrogenase served as a loading and transfer control. this study, this metabolite remained below detection level; thus, it is not clear if its levels are altered by HDACI treatment. are able to detect the effects of the less potent fluorinated In future, these methods could be applied in vivo. Full inhibitor, we believe we would also be able to detect the toxicity studies of BLT still need to be done. However, effects of the parent inhibitor SAHA or other HDACIs of preliminary experiments in our laboratory indicate that an equal or greater efficacy. i.p. injection of 100 mg/kg BLT once weekly on three Surprisingly, TFA, the cleavage product of BLT, was consecutive weeks (a schedule consistent with monitoring constant in the intracellular compartment of control and response to HDACIs in vivo) results in no detectable HDACI-treated cells. However, an analysis of the extracel- toxicity to the animal. Importantly, this dose was sufficient lular medium showed that TFA was lower by an average to produce a MRS visible BLT signal in s.c. PC3 tumors 1.6 Ag/mL in the medium of 10 Amol/L FSAHA-treated with a temporal resolution of 5 minutes at 4.7 T. The BLT cells compared with controls. On average, intracellular signal remained detectable within the tumor region for BLT increased by 18 fmol/cell in those cells. If any TFA over 2 hours. As expected, the TFA peak could not be easily produced through cleavage of BLT by HDACs is removed resolved in our preliminary in vivo studies. However, based into the extracellular medium, this would lead to TFA on the data presented here, it is expected that intratumoral levels lower by 1.3 Ag/mL in the medium of HDAC- BLT levels will be higher in HDACI-treated tumors inhibited cells. This number is consistent with the TFA compared with controls; therefore, tumoral BLT levels levels determined experimentally in our extracellular could be used to assess HDAC activity in vivo. Future work medium. We conclude that TFA produced by cleavage of in our laboratory is aimed at confirming this point. BLT is removed from the intracellular compartment into Downstream metabolic biomarkers could also be assessed the medium, in line with previous findings (55). In vivo, in our preliminary studies. We were able to acquire a 31P depending on the rate of TFA clearance, it is possible that spectrum from PC3 s.c. tumors in 30 minutes, providing a both BLT and TFA will be present in the tumor region. means for monitoring phosphocholine levels. This is However, due to the small chemical shift difference consistent with previous studies (37) in which an increase between BLT and TFA, monitoring TFA is expected to be in phosphocholine could be monitored as an indicator of difficult. response to 17AAG treatment. 1H MRS, with its greater In addition to directly assessing HDAC activity by 19F sensitivity compared with 31P, could also be used to MRS of BLT, a unique metabolic biomarker of response was monitor the total choline signal as a downstream metabolic afforded by using 31P MRS to monitor the intrinsic cellular marker of response to HDAC inhibition. metabolites. Inhibition of cell growth following chemother- In summary, we present here a dual method for apeutic treatment as well as signaling inhibition are noninvasively monitoring response to HDACIs. 19F MRS typically associated with a magnetic resonance visible drop of the targeted molecular imaging agent BLT can be used to in phosphocholine levels (31, 34–36, 56). The increase in monitor delivery and activity of HDACIs at the tumor site, phosphocholine observed here is therefore unusual and has whereas 31P MRS can be used to monitor the downstream been observed previously only following response to metabolic consequences of HDAC inhibition. Together, treatment with 17AAG (37, 38). 17AAG causes inhibition these two MRS methods provide both a direct marker of of Hsp90, which results in depletion of its client proteins, HDAC inhibition and a downstream biomarker of cellular including cdk4 and c-Raf-1, as well as up-regulation of response to the inhibition. We believe that combining both Hsp70 (57, 58). Interestingly, HDACI treatment results in indicators provides a more powerful tool than a single

Mol Cancer Ther 2006;5(5). May 2006

marker alone, particularly at lower levels of HDAC 20. Park SH, Lee SR, Kim BC, et al. Transcriptional regulation of the transforming growth factor h type II receptor gene by histone acetyl- inhibition when the changes observed in either marker transferase and deacetylase is mediated by NF-Y in human breast cancer alone are relatively small. The combination of 19F and 31P cells. J Biol Chem 2002;277:5168 – 74. (or 1H) MRS could thus serve as a reliable noninvasive 21. Scott GK, Marden C, Xu F, Kirk L, Benz CC. Transcriptional repression of ErbB2 by histone deacetylase inhibitors detected by a genomically modality to assess HDAC inhibition. integrated ErbB2 promoter-reporting cell screen. Mol Cancer Ther 2002;1: 385 – 92. Acknowledgments 22. McLaughlin F, La Thangue NB. Histone deacetylase inhibitors open We thank Dr. Kumaralal K. Kaluarachchi for help in recording the nuclear new doors in cancer therapy. Biochem Pharmacol 2004;68:1139 – 44. magnetic resonance spectra, Dr. Edward Felix with the analytic methods, 23. Gelmon KA, Eisenhauer EA, Harris AL, Ratain MJ, Workman P. and Dr. Andrei Volgin and Galina Mikheeva for help with the biological Anticancer agents targeting signaling molecules and cancer cell environ- assays and Western blots. ment: challenges for drug development? J Natl Cancer Inst 1999;91: 1281 – 7. References 24. Parulekar WR, Eisenhauer EA. Phase I trial design for solid tumor 1. Marks P, Rifkind RA, Richon VM, Breslow R, Miller T, Kelly WK. studies of targeted, non-cytotoxic agents: theory and practice. J Natl Histone deacetylases and cancer: causes and therapies. Nat Rev Cancer Cancer Inst 2004;96:990 – 7. 2001;1:194 – 202. 25. Griffiths JR, Glickson JD. Monitoring pharmacokinetics of anticancer 2. Drummond DC, Noble CO, Kirpotin DB, Guo Z, Scott GK, Benz CC. drugs: non-invasive investigation using magnetic resonance spectroscopy. Clinical development of histone deacetylase inhibitors as anticancer Adv Drug Deliv Rev 2000;41:75 – 89. agents. Annu Rev Pharmacol Toxicol 2005;45:495 – 528. 26. Klomp DW, Van Laarhoven HW, Kentgens AP, Heerschap A. 19 3. Marks PA, Richon VM, Miller T, Kelly WK. Histone deacetylase Optimization of localized F magnetic resonance spectroscopy for the inhibitors. Adv Cancer Res 2004;91:137– 68. detection of fluorinated drugs in the human liver. Magn Reson Med 2003; 50:303 – 8. 4. Butler LM, Agus DB, Scher HI, et al. Suberoylanilide hydroxamic acid, 19 an inhibitor of histone deacetylase, suppresses the growth of prostate 27. Wolf W, Presant CA, Waluch V. F-MRS studies of fluorinated drugs cancer cells in vitro and in vivo. Cancer Res 2000;60:5165 – 70. in humans. Adv Drug Deliv Rev 2000;41:55 – 74. 19 5. Kuefer R, Hofer MD, Altug V, et al. Sodium butyrate and tributyrin 28. Yu JX, Kodibagkar VD, Cui W, Mason RP. F: a versatile reporter for induce in vivo growth inhibition and apoptosis in human prostate cancer. non-invasive physiology and pharmacology using magnetic resonance. Br J Cancer 2004;90:535 – 41. Curr Med Chem 2005;12:819 – 48. 6. Fronsdal K, Saatcioglu F. Histone deacetylase inhibitors differentially 29. Seddon BM, Maxwell RJ, Honess DJ, et al. Validation of the fluorinated mediate apoptosis in prostate cancer cells. Prostate 2005;62:299 – 306. 2-nitroimidazole SR-4554 as a noninvasive hypoxia marker detected by magnetic resonance spectroscopy. Clin Cancer Res 2002;8:2323 – 35. 7. Kelly WK, Richon VM, O’Connor O, et al. Phase I clinical trial of histone deacetylase inhibitor: suberoylanilide hydroxamic acid administered 30. Stegman LD, Rehemtulla A, Beattie B, et al. Noninvasive quantitation intravenously. Clin Cancer Res 2003;9:3578 – 88. of cytosine deaminase transgene expression in human tumor xenografts with in vivo magnetic resonance spectroscopy. Proc Natl Acad Sci U S A 8. Sandor V, Bakke S, Robey RW, et al. Phase I trial of the histone 1999;96:9821 – 6. deacetylase inhibitor, depsipeptide (FR901228, NSC 630176), in patients with refractory neoplasms. Clin Cancer Res 2002;8:718 – 28. 31. Negendank W. Studies of human tumors by MRS: a review. NMR Biomed 1992;5:303 – 24. 9. Chavez-Blanco A, Segura-Pacheco B, Perez-Cardenas E, et al. Histone acetylation and histone deacetylase activity of magnesium valproate in 32. de Certaines JD, Larsen VA, Podo F, Carpinelli G, Briot O, tumor and peripheral blood of patients with cervical cancer. A phase I Henriksen O. In vivo 31P MRS of experimental tumours. NMR Biomed study. Mol Cancer 2005;4:22. 1993;6:345 – 65. 10. Kelly WK, O’Connor OA, Krug LM, et al. Phase I study of an oral 33. Evelhoch JL, Gillies RJ, Karczmar GS, et al. Applications of magnetic histone deacetylase inhibitor, suberoylanilide hydroxamic acid, in patients resonance in model systems: cancer therapeutics. Neoplasia 2000;2: with advanced cancer. J Clin Oncol 2005;23:3923 – 31. 152 – 65. 11. Vigushin DM, Coombes RC. Histone deacetylase inhibitors in cancer 34. Beloueche-Babari M, Jackson LE, Al-Saffar NMS, Workman P, Leach treatment. Anticancer Drugs 2002;13:1 – 13. MO, Ronen SM. Magnetic resonance spectroscopy monitoring of MAPK signaling inhibition. Cancer Res 2005;65:3356 – 63. 12. Villar-Garea A, Esteller M. Histone deacetylase inhibitors: understand- ing a new wave of anticancer agents. Int J Cancer 2004;112:171 – 8. 35. Ronen SM, Jackson LE, Beloueche M, Leach MO. Magnetic resonance detects changes in phosphocholine associated with Ras 13. Archer SY, Meng S, Shei A, Hodin RA. p21(WAF1) is required for activation and inhibition in NIH 3T3 cells. Br J Cancer 2001;84:691 – 6. butyrate-mediated growth inhibition of human colon cancer cells. Proc Natl Acad Sci U S A 1998;95:6791 – 6. 36. Beloueche-Babari M, Jackson L, Salman-Al Saffar N, et al. Identifi- cation of magnetic resonance detectable metabolic changes associated 14. Blagosklonny MV, Robey R, Sackett DL, et al. Histone deacetylase with inhibition of phosphoinositide 3-kinase signalling in human breast inhibitors all induce p21 but differentially cause tubulin acetylation, mitotic cancer cells. Mol Cancer Ther 2006;5:187– 96. arrest, and cytotoxicity. Mol Cancer Ther 2002;1:937– 41. 15. Glaser KB, Staver MJ, Waring JF, Stender J, Ulrich RG, Davidsen SK. 37. Chung YL, Troy H, Banerji U, et al. Magnetic resonance spectroscopic Gene expression profiling of multiple histone deacetylase (HDAC) pharmacodynamic markers of the heat shock protein 90 inhibitor 17- inhibitors: defining a common gene set produced by HDAC inhibition in allylamino,17-demethoxygeldanamycin (17AAG) in human colon cancer T24 and MDA carcinoma cell lines. Mol Cancer Ther 2003;2:151 – 63. models. J Natl Cancer Inst 2003;95:1624 – 33. 16. Mitsiades CS, Mitsiades NS, McMullan CJ, et al. Transcriptional 38. Beloueche M, Payne GS, Banerji U, et al. MRS monitoring of the signature of histone deacetylase inhibition in multiple myeloma: biological response to 17AAG treatment: from the laboratory to the clinic. Proc 20th and clinical implications. Proc Natl Acad Sci U S A 2004;101:540 – 5. ESMRMB; 2003; Rotterdam. 17. Richon VM, Sandhoff TW, Rifkind RA, Marks PA. Histone deacetylase 39. Yu X, Guo ZS, Marcu MG, et al. Modulation of p53, ErbB1, ErbB2, inhibitor selectively induces p21WAF1 expression and gene-associated and Raf-1 expression in lung cancer cells by depsipeptide FR901228. histone acetylation. Proc Natl Acad Sci U S A 2000;97:10014 – 9. J Natl Cancer Inst 2002;94:504 – 13. 18. Almenara J, Rosato R, Grant S. Synergistic induction of mitochondrial 40. Podo F. Tumour phospholipid metabolism. NMR Biomed 1999;12: damage and apoptosis in human leukemia cells by flavopiridol and the 413 – 39. histone deacetylase inhibitor suberoylanilide hydroxamic acid (SAHA). 41. Mori ND, Natarajan R, Singh K, Bhujwalla ZM. Loss of p53 function in Leukemia 2002;16:1331 – 43. colon cancer cells results in increased phosphocholine and total choline. 19. Kim MS, Kwon HJ, Lee YM, et al. Histone deacetylases induce Proc 10th ISMRM; 2002; Hawaii. p. 28. angiogenesis by negative regulation of tumor suppressor genes. Nat Med 42. Fuino L, Bali P, Wittmann S, et al. Histone deacetylase inhibitor 2001;7:437 – 43. LAQ824 down-regulates Her-2 and sensitizes human breast cancer cells to

Mol Cancer Ther 2006;5(5). May 2006

trastuzumab, taxotere, gemcitabine, and epothilone B. Mol Cancer Ther 52. Wolffe AP. Chromatin remodeling: why it is important in cancer. 2003;2:971 – 84. Oncogene 2001;20:2988 – 90. 43. Bali P, Pranpat M, Bradner J, et al. Inhibition of histone deacetylase 6 53. Cress WD, Seto E. Histone deacetylases, transcriptional control, and acetylates and disrupts the chaperone function of heat shock protein 90: a cancer. J Cell Physiol 2000;184:1 – 16. novel basis of antileukemia activity of histone deacetylase inhibitors. J Biol 54. Di Gennaro E, Bruzzese F, Caraglia M, Abruzzese A, Budillon A. Chem 2005;280:26729 – 34. Acetylation of proteins as novel target for antitumor therapy: review 44. Berman HM, Westbrook J, Feng Z, et al. The Protein Data Bank. article. Amino Acids 2004;26:435 – 41. Nucleic Acids Res 2000;28:235 – 42. 55. Xu AS, Kuchel PW. Characterisation of erythrocyte transmembrane 45. Clark RD, Strizhev A, Leonard JM, Blake JF, Matthew JB. exchange of trifluoroacetate using 19F-NMR: evidence for transport via the Consensus scoring for ligand/protein interactions. J Mol Graph Model monocarboxylate transporter. Biochim Biophys Acta 1993;1150:35 – 44. 2002;20:281 – 95. 56. Leach MO, Verrill M, Glaholm J, et al. Measurements of human breast 46. Stowell JC, Huot RI, Van Voast L. The synthesis of N-hydroxy-N¶- cancer using magnetic resonance spectroscopy: a review of clinical phenyloctanediamide and its inhibitory effect on proliferation of AXC rat measurements and a report of localized 31P measurements of response to prostate cancer cells. J Med Chem 1995;38:1411 – 3. treatment. NMR Biomed 1998;11:314 – 40. 47. Tyagi RK, Azrad A, Degani H, Salomon Y. Simultaneous extraction of 57. Workman P. Combinatorial attack on multistep oncogenesis by cellular lipids and water-soluble metabolites: evaluation by NMR spectros- inhibiting the Hsp90 molecular chaperone. Cancer Lett 2004;206: copy. Magn Reson Med 1996;35:194 – 200. 149 – 57. 48. Scott BF, Mactavish D, Spencer C, Strachan WMJ, Muir DCG. 58. Neckers L. Development of small molecule Hsp90 inhibitors: utilizing Haloacetic acids in Canadian lake waters and precipitation. Environ Sci both forward and reverse chemical genomics for drug identification. Curr Technol 2000;34:4266 – 72. Med Chem 2003;10:733 – 9. 49. Somoza JR, Skene RJ, Katz BA, et al. Structural snapshots of human 59. Kovacs JJ, Murphy PJ, Gaillard S, et al. HDAC6 regulates Hsp90 HDAC8 provide insights into the class I histone deacetylases. Structure acetylation and chaperone-dependent activation of glucocorticoid recep- (Camb) 2004;12:1325 – 34. tor. Mol Cell 2005;18:601 – 7. 50. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100: 60. Peart MJ, Smyth GK, van Laar RK, et al. Identification and functional 57– 70. significance of genes regulated by structurally different histone deacety- 51. Sawyers C. Targeted cancer therapy. Nature 2004;432:294 – 7. lase inhibitors. Proc Natl Acad Sci U S A 2005;102:3697– 702.

Mol Cancer Ther 2006;5(5). May 2006

Madhuri Sankaranarayanapillai, William P. Tong, David S. Maxwell, et al.

Mol Cancer Ther 2006;5:1325-1334.

Updated version Access the most recent version of this article at: http://mct.aacrjournals.org/content/5/5/1325

Cited articles This article cites 57 articles, 18 of which you can access for free at: http://mct.aacrjournals.org/content/5/5/1325.full#ref-list-1

Citing articles This article has been cited by 3 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/5/5/1325.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/5/5/1325. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.