Hindawi Publishing Corporation
`Chemotherapy Research and Practice
`Volume 2011, Article ID 965826, 9 pages
`doi: 10.1155/2011/965826
`
`Research Article
`
`Development of an Oral Form of Azacytidine:
`2'3'5' Triacetyl-5-Azacytidine
`
`Amy Ziemba,' Eugene Hayes,' Burgess B. Freeman III,” Tao Ye,’ and Giuseppe Pizzorno!
`
`! Nevada Cancer Institute, One Breakthough Way, Las Vegas, NV 89135, USA
`*St. Jude Children’s Research Hospital, Memphis, TN 38105, USA
`>The Hong Kong Polytechnic University, Kowloon, Hong Kong
`
`Correspondence should be addressed to Giuseppe Pizzorno, gpizzorno@nvcancer.org
`
`Received 12 July 2011; Revised 14 September 2011; Accepted 21 September 2011
`
`Academic Editor: G. J. Peters
`
`Copyright © 2011 Amy Ziembaetal. ‘his is an open accessarticle distributed under the Creative CommonsAttribution License,
`which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.
`
`Myelodysplastic syndromes (MD&Ss) represent a group of incurable stem-cell malignancics which are predominantly treated by
`supportive care. Epigenetic silencing through promoter methylation of a number of genes is present in poor-risk subtypes of
`MDS and often predicts transformation to acute myelogenous leukemia (AML). Azacitidine and decitabine, two FDA-approved
`DNA methyltransferase (DNMT) inhibitors, are able to improve overall response although their oral bioavailability complicates
`their clinical use. This study evaluated 2', 3’, 5'-triacetyl-5-azacitidine (TAC) as a potential prodrug for azacitidine. The prodrug
`demonstrated significant pharmacokinetic improvements in bioavailability, solubility, and stability over the parent compound.
`In vivo analyses indicated a lack of general toxicity coupled with significantly improved survival. Pharmacodynamic analyses
`confirmed its ability to suppress global methylation in vivo. These data indicate that esterified nucleoside derivatives may be
`effective prodrugs for azacitidine and encourages further investigation of TAC into its metabolism, activity, and possible clinical
`evaluation.
`
`1. Introduction
`
`Currently, it is estimated that between 12,000 and 20,000 new
`cases of MDS are diagnosed each year in the United States.
`Although MDS can affect all ages, the highest prevalence
`occurs in those over 60 years of age [1, 2]. Much of the pop-
`ulation has indolent forms of MDS, making it one of the
`most prevalent hematologic malignanciesof older adults.
`MDS represents a heterogeneous group of hematopoietic
`disorders which are derived from an abnormal multipotent
`progenitor cell and are characterized by hyperprolifera-
`tive bone marrow, cellular dysplasia, and ineffective hem-
`atopoiesis [3]. Morbidity and mortality result from anemia,
`bleeding, and infection, along with transformation to acute
`myelogenous leukemia (AML) in approximately one-third
`of patients
`[4, 5]. The basis of therapy is supportive
`care, including red blood cell or platelet transfusions and
`treatment of infections. Stem-cell transplantation remains
`the only chance for cure, but it is associated with significant
`treatment-related morbidity and mortality and is generally
`
`restricted to patients <60 years of age [6]. Similar limitations
`exist for the use of high dose chemotherapy. Given the
`limitations of existing therapies, there is a clear need for
`additional therapeutic options for patients with MDS.
`Cancer cells are characterized by abnormal DNA methy-
`lation patterns, and DNA hypermethylation is suspected
`of being involved in MDS progression and leukernoge-
`nesis. Therefore,
`inhibitors of DNA methylation repre-
`sent a useful approach to revert
`these epigenetic chan-
`ges. 5-azacitidine (Vidaza) and its derivative 5-aza-2’-
`deoxycytidine/decitabine (Dacogen) are nucleoside analogs
`with DNA hypomethylating activity that have been FDA-
`approved during the past 5-6 years for MDS treatment
`(7, 8]. hey appear to induce re-expression of key tumor
`suppressor genes in MDS [9]. Compared to supportivecare,
`both agents show an improved overall response (60% versus
`5%), a longertime to progression to AML ordeath, but still
`with limited overall survival advantage [10]. Azacitidine has
`been developed for the treatment of acute leukemia and is
`currently being evaluated in a variety of other disorders [11].
`
`(cid:38)(cid:40)(cid:47)(cid:42)(cid:40)(cid:49)(cid:40)(cid:3)(cid:21)(cid:20)(cid:22)(cid:21)
`CELGENE 2132
`(cid:36)(cid:51)(cid:50)(cid:55)(cid:40)(cid:59)(cid:3)(cid:89)(cid:17)(cid:3)(cid:38)(cid:40)(cid:47)(cid:42)(cid:40)(cid:49)(cid:40)
`APOTEX v. CELGENE
`(cid:44)(cid:51)(cid:53)(cid:21)(cid:19)(cid:21)(cid:22)(cid:16)(cid:19)(cid:19)(cid:24)(cid:20)(cid:21)
`IPR2023-00512
`
`

`

`DNA methylation results in the addition of a methyl
`group at the carbon 5 position of the cytosine ring in CpG
`dinucleotides, which is critical to chromatin structure and
`genomic stability [12]. Since 5-methyl cytosine can be deam-
`inated to thymidine, DNA hypermethylation also facilitates
`gene mutations in human cancers [13]. The transferring
`of methyl groups from S-adenosylmethionine to cytosine
`is catalyzed by DNA methyltransferases, the best known of
`which is DNMT1. The inactivation of DNM1Ts has been
`shown to be the most effective method of inhibiting DNA
`methylation although it is recognized that this approach
`lacks specificity. However, inhibiting the activity of DNMTs
`has resulted in the abrogation of tumorigenicity in murine
`cancer models [14].
`DNMT-inhibiting nucleoside analogs require metabo-
`lism by kinases into nucleotides before their incorporation
`into DNA and/or RNA in order to inhibit DNA methylation.
`‘The modification at C5 prevents the release of DNMITs by
`forming a covalent complex, preventing further DNA meth-
`ylation and consequently the DNA of the progenycells is not
`methylated [15]. Azacitidine and Decitabine are extremely
`potent in inhibiting DNA methylation, but their short half-
`lives in aqueous solutions [16, 17] and low oral bivavail-
`ability complicate their delivery. For years,
`these drugs
`used as antineoplastic agents in leukemias were escalated to
`maximum tolerated doses (MTD)
`[18]; however, recent
`clinical trials have confirmed that low-dose exposures lead
`to greater responses and are associated with less toxicity
`[19, 20].
`In an effort to overcome the stability and pharmacoki-
`netic limitations, a numberof groups are working to develop
`oral DNMTinhibitors, including oral forms of decitabine
`and azacitidine. As a proof of principle we have synthesized
`an acetylated derivative of azacitidine, 2’,3’,5'-triacetyl-5-
`azacitidine (TAC)
`to evaluate as an oral prodrug. While
`this is not a totally new approach, it was never utilized
`for this derivative and in this set of diseases. Our cur-
`rent study demonstrates that 2’,3',5’-triacetyl-5-azacitidine
`(TAC) effectively inhibits methylation and improves life
`expectancies in murine models while demonstrating lower
`toxicity when comparedto its parent compound,azacitidine.
`
`2. Materials and Methods
`
`2.1. Preparation of 2',3',5'-'lriacetyl-5-Azacitidine (IAC)
`[21]. TAC was prepared through condensation of trimeth-
`ylsilylated-5-azacytosine
`and
`1,2,3,3-tetra-~O-acetyl-B-D-
`ribofuranose. In a 150 mL, 3-necked flask, a mixture of 5-
`azacytosine, hexamethyldisilazane, and ammonium sulfate
`was heated at reflux for 2 hours. A fresh amount of
`ammonium sulfate was added, and the reflux was continued
`for 6 hours. The excess hexamethyldisilazane was removed
`under
`reduced vacuum to afford trimethylsilylated 5-
`azacytosine as an off-white residuc.
`‘letra-O-acetyl-D-ribofuranose was prepared by adding
`acetic anhydride dropwise to D-ribose dissolved in dry
`pyridine. The solvent was removed by heating the solution
`at 50°C for 15 minutes, and tetra-O-acetyl-D-ribofuranose
`wascrystallized from isopropanol |22].
`
`Chemotherapy Research and Practice
`
`Trimethylsilylated 5-azacytosine and 1,2,3,5-Tetra-O-
`acetyl-$-D-ribofuranose were dissolved in acetonitrile, and
`the mixture was cooled at 0°C in an ice-water bath. Tri-
`methylsily] trifluoromethanesulfonate (TMSOT) was added
`slowly at 0°C then stirred at room temperature for 3
`hours. The reaction mixture was poured into a solution
`of NazCO; and NaHCO; in ice water and then extracted
`with dichloromethane. The combined organic layer was
`washed with cold H20, cold brine, and dried over anhydrous
`NazSO,. The residue, after removal of the solvent, was re-
`crystallized from a mixture of dichloromethane and hexane
`to provide the desired compound with an 85%yield.
`
`2.2. LC/MS/MS Analysis of TAC and Its Metabolites Mono-
`and Di-Acetyl-5-Azacitidine and ACT. To monitor the in
`vivo absorption, distribution, and metabolism of TAC, a
`highly sensitive LC/MS/MSanalytical method was developed
`that allows the quantitation of concentrations as low as
`10 ng/mL of TAC andits metabolites. The samples were pre-
`pared by adding 200uL of plasma to 400uL of cold
`acetonitrile, vortexed, and centrifuged at 14,000 rpm for 10
`minutes. The supernatant was placed into a 10 x 75 mm glass
`test tube and evaporated to 50uL under nitrogen at 37°C.
`The residue was then diluted with 1mL of 2% phosphoric
`acid and vortexed for 10 seconds at 3,000 rpm.
`Solid phase extraction was conducted using a Varian
`Bond Elute PLEXA PCX 1 mLcolumnsplaced on a vacuum
`manifold. Columns were activated by adding 1 mL of MeOH
`followed by |1mL H.O and ImLof 2% phosphoric acid.
`All samples were allowed to drip by gravity flow. ‘he
`extraction columns were then washed sequentially with 1 mL
`2% phosphoric acid, 1mL H2O, and 1mL MeOH/AcCN
`(1/1) and eluted into a 10 x 75 mmglasstest tube containing
`50uL of 25% formic acid in MeOH with 2 x 600uL of
`2% NH4OH in MeOH/AcCN (1/1). The liquid was then
`evaporated to dryness with nitrogen at 37°C, and 100 uL of
`mobile phase A was added and vortexed for 10 seconds at
`3,000 rpm and transferred to LC autosamplervials at 4°C.
`HPLC separation was conducted on a reverse phase
`C18 column (Varian Pursuit C18, 3 micron particle size;
`100mm x 2.0mm) using a Shimadzu SIL-HTc controller
`with dual LC-20AD pumps and DGU-20A3 degasser. The
`separation was achieved using a gradient of Mobile Phase A:
`0.1% Formic Acid and Mobile Phase B: 0.1% Formic Acid,
`90% MeOH with a flow rate of 400 ¢L/min at 40°C over
`10 minutes. MS/MS fragmentation was performed on an
`Applied Biosystems API 3200 instrument with a Turbo V Ion
`Spray ion source and Positive ion polarity.
`
`2.3. Drug Stability. ACY and ‘TAC stabilities were determined
`after incubation at pH1 (2% Phosphoric acid), pH3 (0.1%
`Formic Acid), pH5 (K3PO,4), and pH7.4 (Dulbecco’s PBS).
`100 ng of each compound was resuspended in 1.2 mL total
`volume, and 200 pL samples taken at 1, 3, 6, 29, and 48 hr.
`Each pH point was repeated 6 times in 2 dram vials with
`‘Teflon-lined screw caps at room temperature. Samples were
`extracted by solid-phase extraction using 100ng decitabine
`as an internal standard and analyzed by tandem mass spec-
`troscopy as described above.
`
`

`

`Chemotherapy Research and Practice
`
`2.4, Log P Determination. The OSIRIS Property Explorer
`software was used to generate initial cLogP values. The par-
`tition coefficient was determined by using an analyte concen-
`tration near the solubility limit for the compound in water
`and incubating in a water: octanol mix ranging from 0.5: 1,
`1:1, to 2:1. The pH of the water was 2 units above the pKa
`of the compound, preventing ionization. The experiment
`was conducted using the shake-flask method, after allowing
`octanol and water to equilibrate for 24 hours at 25°C.
`The initial drug concentration in water was 0.2mM. Tubes
`were rotated at a 45 degree angle for 10 minutes with a
`3 second rotation. The test was performed in a constant
`environment of 20°C-25°C. Tubes were centrifuged at 1000 ¢
`for 10 minutes, and the octanol phase was extracted. The
`UV/Visible absorption spectrum was evaluated from 550 nm
`to 210 nm to determine the absorption maximum. The ana-
`lyte concentration was determined and the LogP calculated:
`Log = log(concentration in octanol/concentration in water).
`
`2.5. Permeability across Caco-2 Monolayers. Caco-2 cells
`(ATCC) were maintained under standard conditions in Dul-
`becco’s modified Eagle’s medium (DMEM). Approximately 1
`x 10° cells (passage number 25-35) were seeded onto poly-
`carbonate cell culture inserts (Transwell, 0.45 um, 12mm
`diameter). The cells were allowed to growand differentiate
`for three weeks changing the medium every 2-3 days.
`Lucifer yellow was used to ensure appropriate paracellular
`permeability levels [23].
`The transport experiments were performed in Hanks’
`balancedsalt solution (HBSS) containing 25 mMHepes (pH
`7.4) as thoroughly described [24]. The drugs were dissolved
`in transport buffer to a final concentration of 0.2 mM. The
`monolayers were washed in prewarmed transport buffer for
`30min. The drug solutions were added to the donor side of
`the monolayers, and fresh transport buffer was added to the
`receiver side. The plates were gently shaken (50-100 rpm)
`at 37°C. At time intervals ranging from 5 minutes up to
`2h, 0.2mL samples were collected from the receiver cham-
`bers and replenished with fresh transport buffer. Sarnples
`containing the transported drug was collected for LC/MS
`analysis by immediately freezing at -80°C and subsequently
`extracted 2:1 with acetonitrile and analyzed as indicated in
`the LC/MS/MS methodologysection.
`
`‘To
`2.6. Pharmacokinetics and Metabolism of TAC in Mice.
`evaluate the in vivo pharmacokinetic properties of TAC,
`C57/BL6 female mice of 20g body weight were dosed with
`38 mg/kg ‘LAC p.o. (equimolar dose to 25 mg/kg of ACI’) or
`25 mg/kg ACT iv. via tail vein injection. TAC and ACT were
`solubilized in LX PBS immediately before administration and
`dosed at 0.01 ml/g fasted body weight. Blood samples (two
`time points per animal, three animals per time point) were
`obtained via retro-orbital bleeding with Natelson pipettes
`at 0.5, 1, 2, 4, 8, and 24 hours after drug administration.
`The collected blood was then centrifuged for 4 minutes
`at 13,000g. Plasma was immediately extracted with cold
`acetonitrile and stored at ~ 80°C until LC/MS/MSanalysis.
`
`Noncompartmental plasma pharmacokinetic parameters
`for TAC prodrug and ACT after oral TAC or iv ACT
`administrations were estimated using the sparse sampling
`option in WinNonlin Software version 4.1 (Pharsight Corp.,
`Mountain View, Calif, USA). Briefly, the WinNonlin sparse
`sampling option applies the linear trapezoidal rule for area
`under the concentration-time curve (AUC) calculations and
`generates a standard error for the mean AUC estimate
`that accounts for correlations in the data resulting from
`repeated sampling of individual mice. Parameters of interest
`included the AUC from time 0 to the last measurable
`concentration (AUCs), observed peak concentration after
`oral administration (Cmax), estimated peak concentration
`after iv. administration (Co), mean residence time to the
`last measurable concentration (MRT), and the half-life of
`the terminal phase (T1/.). To assess the multiphasic decline
`in plasma concentrations, an alpha (T12.) half-life was also
`estimated using a log-linear regression of the mean ACT
`concentrations through 2 hoursafter dosing.
`
`2.7, Global Methylation Detection Assay. 24 hours follow-
`ing drug administration, snap-frozen mouse tissues were
`homogenized and genomic DNA isolated using Promega’s
`Wizard Genomic DNA Purification Kit. 100ng of EcoRI
`linearized genomic DNA wasspotted onto PVDF membrane.
`‘The membrane wasair-dried and repermeabilized according
`to manufacturer’s instructions. Nonspecific binding was
`blocked using 5% milk in PBS-T and washed twice in
`PBS-T for 5 minutes.he membrane was incubated with
`primary antibody (1: 1000 anti-5-methylcytosine in PBS-T,
`Calbiochem) for 1 hour then washed four timesfor five min-
`utes each with PBS-T. The membrane was incubated with the
`HRP-conjugated secondary antibody for 1 hour and washed
`four times for five minutes each with PBS-T. Signalintensity
`following ECL detection was quantitated using KODAK
`Image Station software. Assay results were confirmed using
`EpiGentek’s Methylamp Global DNA Methylation Quantifi-
`cation Kit per manufacturer’s instructions.
`
`2.8 pISINK4B Methylation Status. Mouse leukemia cells
`(L1210) were purchased from ATCC. The published [C50
`value for ACT (0.2uM) was first confirmed using the
`Promega CellTiter MTS assay. Cells were treated with ACT
`(0.1 or 1 uM) or TAC (1 or 100 uM) tor 48 hours, with tresh
`drug added daily. Genomic DNAwasisolated after 48 hours
`using the Sigma GenElute Genomic DNA Kit, including
`the RNase treatment step to ensure no RNA carryover. The
`methylation status of plISINK4B was determined using a
`predesigned primer set and the EpiTect Methyl qPCR Assays
`MethylScreen technology (Qiagen, Frederick, Md, USA).
`
`2.9. Antileukemic Activity of Triacetyl-5-Azacitidine. The in
`vivo L1210 lymphocytic leukemia model
`in C57BL/6 x
`DBA/2 FI(BDF1) female mice was used to evaluate the
`in vivo antiproliferative effect of TAC. L1210 cells were
`carried in BDF1 mice by weekly ip. (intraperitoneal) pas-
`sages. Leukemia cells from ascites fluid were diluted ap-
`propriately in RPMI
`1610 medium and injected ip.
`
`

`

`4
`
`Chemotherapy Research and Practice
`
`TaBLeE1: Stability of TAC and its metabolites at various pH.
`
`Sample
`TAC (%)
`DAC (%) MAC(%)
`ACT (%)
`
`Control
`100
`
`0
`0
`1
`99
`pH i—1hr
`0
`4
`22
`74
`pH 1—3hr
`0
`Al
`41
`38
`pH 1—6hr
`46
`3]
`22
`1
`pH 1—29 hr
`0
`0
`3
`97
`pH 3—1 hr
`0
`0
`4
`96
`pH 3—3hr
`0
`2
`6
`92
`pH 3—6 hr
`1
`2
`4
`93
`pH 3—29 hr
`0
`0
`2
`98
`pH 5—1 hr
`0
`0
`6
`94
`pH 5—3hr
`0
`0
`2
`98
`pH 5—6 hr
`1
`1
`4
`94
`pH 5—29 hr
`0
`0
`4
`96
`pH 7.4—1 hr
`0
`2
`7
`91
`pH 7.4—3 hr
`0
`3
`12
`85
`pH7.4—6 hr
`1
`13
`Vw
`69
`pH 7.4—29 hr
`TAC:tri-acetyl-azacitidine; DAC: di-acetyl-azacitidine; MAC: mono-acetyl-
`azacitidine; ACT:azacitidine.
`
`TAC (26%) to the mono- and diesters in the first 3 hours
`of incubation increasing to 60% after 6 hrs. However, after
`24, 99% of TAC was hydrolyzed, including 46% to the parent
`compound, ACT.
`A similar experiment to establish the stability of azaci-
`tidine in solution resulted in significant degradation of the
`nucleoside analog throughout the pH range with maximum
`stability at pH 1. As previously determined,
`[25]
`the
`initial degradation to N-formylguanylribosylurea followed
`bythe irreversible formation of guanylribosylurea resulted in
`two derivatives with no pharmacological activity (data not
`shown).
`
`3.2. LogP Determination. The logP value of a compound is
`a well established measure of a compound’s hydrophilicity.
`The theoretical physical-chemical properties of TAC were
`evaluated by measuring the partition coefficient of the un-
`ionized molecules in two immiscible phases of water and
`octanol. Azacitidine resulted in an experimental LogP value
`of —2.32 compared to the calculated cLogP values of —2.17
`while the changes to the molecule by acetylation resulted in
`an improved LogP value for TAC of —0.61 versus a cLogP of
`—0.85.
`
`3.3. Permeability across Caco-2 Monolayers. The human
`colon carcinoma Caco-2 cell model was used to evaluate the
`oral absorption of the prodrug across membranes anddeter-
`minethe permeability coefficient (absorption rate constant).
`A sigmoidal relationship exists between drug absorption
`rates as measured with the in vitro Caco-2 model and human
`absorption [26]. Permeability coefficients typically range
`from 5 x 10°* to 5 x 107° cm/s. Drugsthat are well absorbed
`in humans have permeability coefficients greater
`than
`
`QL
`
`NAN
`aaisenyy?OIMS
`H
`1
`
`H2NV
`
`ZN
`
`ey °
`‘ Y 1
`Q
`<4 Tt
`>Q_p NEN
`y-6 a
`oO
`2 T 5
`oO
`TMSOT£ MeCN
`‘-5 a
`o ;
`
`Ficure 1: Chemical synthesis of 2’,3’,5’-triacetyl-5-Azacytidine.
`Results of analytical IH NMR (CDCI3, 500 MHz): d 8.18{s, 1H),
`7.69(s, 1H), 6.33(s, 1H), 5.82(d, 1H, J = 3.0Hz), 5.54(t, 1H, J =
`4.0 Hz), 5.41(t, 1H, J = 6.0Hz), 4.30-4.41(m, 3H), 2.11{s, 3H),
`2.10(s, 3H), and 2.08(s, 3H) ppm. 13C NMR (CDCI3, 125 MHz):
`d 170.3, 169.6, 169.5, 166.0, 156.0, 153.1, 89.7, 79.9, 73.7, 69.9, 62.8,
`20.7, 20.4, and 20.3 ppm.
`
`(110° cells/0.1 mL/mouse) into recipient mice. A daily oral
`administration of TAC (38 mg/kg via oral gavage) or ACT
`(25 mg/kg administered i.p.) diluted in PBS was given to
`leukemic BDF1 mice starting 24 hours from the inoculation
`of L1210 cells for a total of 5 days. Compounds were prepared
`fresh daily. A group of untreated leukemic mice received
`sterile water for injections and served as controls.
`
`2.10. TAC Toxicology. The in vivo toxicity of TAC waseval-
`uated in CD-1 mice using 3 groups of 3 male and 3 female
`CD-1 mice per group. CD-1 is the classical mouse strain
`chosen for toxicology studies including safety and efficacy.
`On Days 1-5 and Days 8-12, animals were administered
`vehicle (sterile water for injection, Group 1), 38 mg/kg/day
`TAC (Group 2), or 76 mg/kg/day 'TAC (Group 3) via oral
`gavage. Animals were not dosed on Days 6 and 7 and were
`euthanized on Day 13 for necropsy. Doses were prepared
`fresh daily from the drug stock. Criteria for evaluation
`included clinical observations, body weights, limited serum
`chemistry and hematology parameters, and histopathology
`evaluation.
`
`3. Results
`
`3.1. Synthesis and In Vitro Characterization of Triacetyl-5-
`Azacitidine. 2',3',5'-triacetyl-5-azacitidine has been success-
`fully synthesized through condensation oftrimethylsilylated-
`5-azacytosine and 1,2,3,5-tetra-O-acetyl-$-D-ribofuranose
`(Figure 1). An HPLC detection assay was optimized for TAC
`and ACTas well as somepotential analytes.
`The solubility of TAC in comparison to ACT was
`evaluated at various pH. ‘TAC and ACT were equally soluble
`at pH1 (40 mg/mL) and pH3 (25 mg/mL), while TAC was
`more soluble than ACT at pH5 (30 mg/mL versus 15 mg/mL)
`and pH7 (30 mg/mLversus 10 mg/mL). Thestability of the
`prodrug across a range of pHs is indicated in Table 1. TAC
`shows remarkable stability at pH 3, 5, and 7.4 with no
`degradation to the parent compound and 30% hydrolysis to
`the mono- anddiesters only at pH 7.4 after a 29 hr incubation
`at 25°C. At pH 1, we observed a moderate hydrolysis of
`
`

`

`100
`
`— Oo
`
`
`
`Concentration(uM)
`
`0.1
`
`0
`
`6
`
`12
`
`18
`
`24
`
`Time(hours)
`
`38 mg/kg p.o.
` [tin|6s]2|NE
`
`(a)
`ACT
`25 mg/kg iv.
`
`TAC
`
`Chemotherapy Research and Practice
`
`1 x 107° cm/s, while drugs and peptides that are absorbed
`to less than 1% have permeability coefficients less than
`1 x 10°’ cm/s [27, 28]. Between these ranges, compounds
`are considered to have intermediate absorption rates. In
`these studies, ACT was determined to have a permeability
`coefficient between 1-3 x 10°°, compared to 5~7 x 107°
`for ‘TAC. This indicates a potential 2-3 fold improvement
`in intestinal absorption. The efflux ratio is defined as the
`quotient of the secretory permeability and the absorptive
`permeability. ‘The efflux ratio was near 1.0, indicating these
`compoundsare not beingactively effluxed. No hydrolysis of
`the prodrug to azacitidine was observed during the course of
`this experiment.
`
`3.4. Pharmacokinetics and Metabolism of TAC in Mice. TAC
`was administered at 38 mg/kg to C57/BL6 mice via oral gav-
`age, an equimolar dose equivalent to 25 mg/kg of azacitidine.
`TAC andits active metabolite ACT were detectable in plasma
`at 30 minutes after the oral administration (Figure 2(a)).
`'TAC appearedto be rapidly deacetylated leading to a minimal
`accumulation of the prodrug resulting in plasma concen-
`trations below the limit of quantitation (10 ng/mL) at time
`points beyond 1 hour after administration. T[AC-derived
`azacitidine reached a peak concentration of approximately
`5,000 ng/mL (~20 uM) at 30 minutes with a pharmacologi-
`cally relevant concentrationofat least 0.5 uM being sustained
`for 24 hours after oral 'TAC dosing. The mean TAC and
`‘TAC-derived ACT plasma concentrations versus time plots
`after oral administration of TAC prodrug are depicted in
`Figure 2(a) with the pharmacokinetic parameters reported in
`Figure 2(b). ACT was also administered i.v. as a control to
`emulate the standard meansof administration in the clinical
`setting. The 'T1/2 and MRT: values for 'TAC-derived azaciti-
`dine were 9.2 hours and 7.7 hours, respectively, whereas iv.
`administration of ACT resulted in Ty and MRTias values
`of 6.8 hours and 1.1 hours, respectively. The AUC;values
`for plasma azacitidine were 73.5 hr-uM and 126 hr-uM after
`oral administration of TAC prodrug or iv. administration
`of ACT, respectively. While the absolute bioavailability of
`derived azacitidine after oral TAC administration is 58%, the
`prolonged Ty/2 and MRTast values for azacitidine observed
`in plasmaafter oral TAC administration suggest a protracted
`absorption and conversion of the TAC prodrug at
`the
`gastrointestinal or presystemic level (Figure 2(b)). The log-
`linear regression analysis of ACTconcentration data through
`2 hours revealed an increased Ti.q value for ACT after p.o.
`‘TAC administration compared with i.v. ACT administration
`(0.73 hr versus 0.32 hr, resp.), also suggesting protracted
`absorption and both presystemic and systemic conversion of
`TAC.
`
`3.5. Pharmacodynamic Effect of Triacetyl-5-Azacitidine. MTS
`assays (Promega, Madison, Wis, USA) were performed to
`determine the in vitro effects of ACT and TAC on L1210
`cell proliferation. TAC had no effect on L1210 cell viability
`even at doses 100-fold higher than the known IC50 of ACT
`(0.2 uM) [29], and this is an indication that this compound
`does notget activated in vitro. This result was predicted,since
`
`(b)
`
`Figure 2: Pharmacokinetics of oral TAC versusi.v. 5-azacitidine in
`C57BL/6 mice. (a) Mean plasma concentration versus time for tri-
`acetyl azacytidine (TAC) (4) and resultant ACT (*) in nontumor
`bearing mice after a 38mg/kg dose of oral TAC. Mean plasma
`concentration versus time for ACT in nontumor bearing mice after
`25 mg/kg ACTi.v. (c). Error bars indicate standard deviation. (b)
`Noncompartmental plasma pharmacokinetic parameters for azac-
`itidine (ACT) andtri-acetyl azacitidine (TAC). AUCys values are
`mean (standard error). NE—not estimated. NC—notcalculated.
`
`cells in culture lack the necessary esterase activity to convert
`the prodrug to ACT that remains below thelimit of detection
`using our LC-MS method.
`Gene-specific methylation PCR was performed in order
`to confirm whether TAC could affect the methylation level of
`a specific target even in the absenceofcellular toxicity. ACT,
`which is the active derivative of TAC, has been extensively
`studied by numerous research groups to determine specific
`cellular methylation targets and mechanism of action [30,
`31]. PISINK4B is a classic target known to be hyper-
`methylated in AML and MDSandattributes to a poor
`prognosis in those patients [32, 33]. 0.1uM and 10uM ACT
`decreased P15 promoter methylation by 39% (+3) and 49%
`(+17), respectively. In comparison, 1uM and 100 uM ‘TAC
`decreased P15 promoter methylation by only 19% (+8) and
`23% (+16), respectively.
`
`

`

`ACT
`
`TAC
`Average
`
`Gut.2
`100
`14
`30
`
`Gut.3
`100
`34
`41
`
`Gut.4
`100
`33
`25
`
`Average
`100
`29
`38
`
`Spleen.1
`100
`54
`48
`
`Spleen.2
`100
`71
`31
`
`Spleen.3
`100
`48
`62
`
`Spleen.4
`100
`29
`22
`
`100
`50
`41
`
`(b)
`
`Ficure 3: Effect of DNMT inhibitors on global DNA methylation
`in tissues of mice exposed for 5 days to ACT or TAC. (a) Dot blot
`of mouse tissues isolated 24 hours after treatment with 25 mg/kg
`iv. ACT or 38 mg/kg p.o. TAC. Groups were compared to untreated
`animals (unt). (b) Densitometry of methylation levels represented
`in (a).
`
`In an in vivo experiment designed to evaluate the effect
`ofa prolonged administration of TAC on global methylation,
`TAC (38me/kg p.o.) or ACT (25mg/kg ip.) diluted in
`PBS was administered to C57BL6 mice every 24 hours for
`5 consecutive days (four animals per group). Tissue methyla-
`tion levels are reported in Figure 3. A significant reduction
`in global DNA methylation was detected in the gut and
`spleen after the administration of TAC, comparable to the
`methylation decrease detected after the administration of
`ACT. We did not observe a significant change in global
`methylation status in kidneyorliver tissue (data not shown).
`The positive results in gut and spleen were confirmed
`using the EpiGentek’s Methylamp Global DNA Methylation
`Quantification Kit. This quantitative analysis confirmed a
`comparable 50%-60% decrease in global methylation by
`both TAC and ACT (data not shown).
`
`3.6. Antileukemic Activity of Triacetyl-5-Azacitidine. The
`L1210 lymphocytic leukemia model was used to determine
`the efficacy of TAC in an animal model. The L1210 model
`is a classical model used to predict for compounds that
`are effective against leukemias and lymphomas. A daily oral
`administration of TAC (38 mg/kg equimolar to 25 mg/kg
`of ACT and a higher dose of 50mg/kg) or azacitidine i.p.
`(25 mg/kg) was given to leukemic mice for a total of 5 days.
`The median survival time for the untreated control group
`was 8 days, while the median survival time for the TAC
`
`Chemotherapy Research and Practice
`
`— Control
`—#— 25 mg/kg AC
`—#- 38 mg/kg TAC
`
`Days
`
`(a)
`
`Toxicity curve
`
`120
`100
`
`Survivial
`
`
`
`Averageweight
`
`(%)
`(g)
`
`Days
`
`—e Control
`—=-— 25mg/kg ACT
`
`—*— 38mg/kg TAC
`—— 50me/kg TAC
`
`(b)
`
`Figure 4: (a) Antileukemic activity ofTAC and ACT in BDF1 mice
`bearing L1210 lymphocytic leukemiacells after 5 days of treatment
`with indicated concentrations of oral TAC or i-p. ACT. (b) Toxic
`effect as indicated by average weightloss of the same tumorbearing
`mice.
`
`treated groups was 12 days resulting in a 50% increased
`lifespan, demonstrating not only an epigenetic modulating
`effect, but also antileukemic activity. Azacitidine at
`the
`maximum tolerated dose of 25mg/kg administered i.p.
`resulted in a 17-day survival with a 112% increased lifespan
`(Figure 4(a)). The toxicity curve (Figure 4(b)) indicates a
`nonsignificant weight loss for the TAC-treated groups at day
`3 followed by a rapid recovery indicating the lack of toxicity
`of this prodrug coupled with a significant antileukemiceffect,
`while the azacitidine-treated animal showed a 16% weight
`reduction with a significant recovery once the treatment was
`completed.
`
`3.7. TAC Toxicology. The in vivo toxicity of TAC was eval-
`uated by orally administering repeated doses to CD-1 mice
`for two weeks. On Days 1-5 and Days 8-12, animals
`were administered vehicle, 38mg/kg/day TAC (Group 2,
`equivalent
`to 25mg/kg of ACT), or 76mg/kg/day 'TAC
`(Group 3) via oral gavage. Criteria for evaluation included
`clinical observations, body weights, limited serum chemistry
`
`

`

`Chemotherapy Research and Practice
`
`
`
`60 qr ee
`y
`
`
`
`Control(%)
`
`y
`4------- FA
`40
`39 J... 7 ee eee
`2004 _---- on toes
`-
`be
`:
`
`
`RTC WBC
`
`NEU
`
`@ 38 mg/kg TAC
`GO 76meg/kg TAC
`
`Figure 5: TAC-assuciated effect on hematological parameters
`following TAC administration to CD-1 mice. RTC: reticulocyte
`count, WBC: white blood cell count, NEU: neutrophil count, LYM:
`lymphocyte count, MNO: monocyte count, EOS: cosinophil count,
`LUC: large unstained cell count.
`
`and hematology parameters, and histopathologyevaluation
`of a limited number of tissues from vehicle-treated and
`76 mg/kg/day TAC-treated animals.
`There were no deaths and no notable 'TAC-associated
`body weight changes during the study. Additionally, there
`were no TAC-associated changes in liver or renal function
`markers, including alanine and aspartate aminotransferases,
`alkaline phosphatase, bilirubin, blood urea nitrogen, and
`creatinine.
`TAC-associated changes in various hematological param-
`eters were observed (Figure 5). This included a reduction in
`white blood cell counts, absolute lymphocytes, and absolute
`leukocytes by 41%, 31%, and 24%, respectively, for the
`38 mg/kg/day dosage group. The reduction values for the
`76 mg/kg/day group only slightly increased by 44%, 31%,
`and 34%,
`respectively. There were 'TAC-related findings
`in the bone marrow (hypocellularity) and in the lymph
`nodes (inactive germinal centers) which reflect the decreased
`overall white blood cell count reductions. An additional
`microscopic '[AC-related finding was duodenum apoptosis
`of the crypt epithelium in the 76 mg/kg/day-treated animals.
`Other observations included a decrease in myeloid andery-
`throid compartmentcellularity correlating with microscopic
`findings in varioustissues.
`
`4, Discussion
`
`Epigenetic modulators are appealing cancer therapeutics due
`to their potential reversibility and preference for highly
`proliferating cancer cells. Azacitidine and decitabine have
`demonstrated improved clinical success in hematologic ma-
`lignancies with longer or continuous infusion [34]. However,
`the aza-pyrimidine ring is unstable in aqueous solutions and
`has poor bioavailability, limiting the use of oral formulations
`
`s
`
`that would be more convenientfor thepatients, help in long-
`term dosing, and reduce local side effects when administered
`as subcutaneous depot.
`We have successfully synthesized an azacitidine pro-
`drug, 2',3’,5’ -triacetyl-5-azacitidine and shownits favorable
`physical-chemical characteristics. TAC