Research Article

Molecular Dynamics Guided Design of Tocoflexol:A New Radioprotectant Tocotrienol with

Enhanced BioavailabilityCesar M. Compadre,1* Awantika Singh,1,2 Shraddha Thakkar,2,3 Guangrong Zheng,1

Philip J. Breen,1 Sanchita Ghosh,4 Mahmoud Kiaei,5,6 Marjan Boerma,1

Kottayil I. Varughese,3 and Martin Hauer-Jensen1,7

1Department of Pharmaceutical Sciences, University of Arkansas for Medical Sciences, LittleRock, AR 72205, USA

2UAMS/UALR Joint Bioinformatics Graduate Program, University of Arkansas for MedicalSciences, Little Rock, AR 72205, USA

3Department of Physiology and Biophysics, University of Arkansas for Medical Sciences, LittleRock, AR 72205, USA

4Armed Forces Radiobiology Research Institute, Uniformed Services University of the HealthSciences, Bethesda, MD 20889, USA

5Department of Neurobiology and Developmental Sciences, University of Arkansas for MedicalSciences, Little Rock, AR 72205, USA

6Center for Translational Neuroscience, University of Arkansas for Medical Sciences, Little Rock,AR 72205, USA

7Surgical Service, Central Arkansas Veterans Healthcare System, Little Rock, AR 72205, USA

Strategy, Management and Health Policy

EnablingTechnology,Genomics,Proteomics

PreclinicalResearch

Preclinical DevelopmentToxicology, FormulationDrug Delivery,Pharmacokinetics

Clinical DevelopmentPhases I-IIIRegulatory, Quality,Manufacturing

PostmarketingPhase IV

ABSTRACT There is a pressing need to develop safe and effective radioprotector/radiomitigator agentsfor use in accidental or terrorist-initiated radiological emergencies. Naturally occurring vitamin E familyconstituents, termed tocols, that include the tocotrienols, are known to have radiation-protection proper-ties. These agents, which work through multiple mechanisms, are promising radioprotectant agents havingminimal toxicity. Although α-tocopherol (AT) is the most commonly studied form of vitamin E, thetocotrienols are more potent than AT in providing radioprotection and radiomitigation. Unfortunately,despite their very significant radioprotectant activity, tocotrienols have very short plasma half-lives andrequire dosing at very high levels to achieve necessary therapeutic benefits. Thus, it would be highlydesirable to develop new vitamin E analogues with improved pharmacokinetic properties, specificallyincreased elimination half-life and increased area under the plasma level versus time curve. The shortelimination half-life of the tocotrienols is related to their low affinity for the α-tocopherol transfer protein(ATTP), the protein responsible for maintaining the plasma level of the tocols. Tocotrienols have less affinityfor ATTP than does AT, and thus have a longer residence time in the liver, putting them at higher risk formetabolism and biliary excretion. We hypothesized that the low-binding affinity of tocotrienols to ATTP isdue to the relatively more rigid tail structure of the tocotrienols in comparison with that of the tocopherols.

*Correspondence to: Cesar M. Compadre, Department of Pharmaceutical Sciences, University of Arkansas for Medical Sciences,Little Rock, AR 72205, USA.E-mail: cmcompadre@uams.edu

Published online in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/ddr.21162

DRUG DEVELOPMENT RESEARCH 75 : 10–22 (2014)

DD

R

© 2013 Wiley Periodicals, Inc.

Therefore, compounds with a more flexible tail would have better binding to ATTP and consequentlywould have longer elimination half-life and, consequently, an increased exposure to drug, as measured byarea under the plasma drug level versus time curve (AUC). This represents an enhanced residence of drugin the systemic circulation. Based on this hypothesis, we developed a new class of vitamin E analogues,the tocoflexols, which maintain the superior bioactivity of the tocotrienols with the potential to achieve thelonger half-life and larger AUC of the tocopherols. Drug Dev Res 75 : 10–22, 2014. © 2013 WileyPeriodicals, Inc.

Key words: radiation protection; tocoflexol; tocotrienol; countermeasures; molecular dynamics; bioavailability

INTRODUCTION

Protection from radiation hazards is an urgentnecessity. In the event of a radiological/nuclear terroristattack, three groups will require a radiation counter-measure to ameliorate the deleterious effects of ioniz-ing radiation exposure: first responders, remediationworkers, and the resident population. Thus, developingradiation countermeasures, applicable prior to orduring exposure to prevent or limit injury, is a high-priority research area [Pellmar and Rockwell, 2005].

Vitamin E components have been reported to beradioprotective in various models [Berbée et al., 2009;Ghosh et al., 2009; Kulkarni et al., 2010; Satyamitra

et al., 2011; Berbée and Hauer-Jensen, 2012]. VitaminE comprises eight compounds, present in differentratios depending on the source [Müller et al., 2013].These compounds are distributed in two chemicalcategories, the tocopherols and the tocotrienols. Thetocopherols have a saturated side chain, phytyl, and thetocotrienols have three nonconjugated unstaturationson their side chain, farnesyl. All vitamin E componentshave the same chromanol head. Depending on thedegree of methylation of the head, tocols can be cat-egorized as α, β, γ, and δ (Fig. 1).

Studies on radioprotection provided by vitamin Ecomponents have focused mainly on α-tocopherol (AT)and its acetate and hemisuccinate esters. One of the

Fig. 1. Chemical structures of vitamin E components.

DESIGN OF BIOAVAILABLE RADIOPROTECTANTS 11

Drug Dev. Res.

earliest indications that AT provided radioprotectionwas reported by Prince and Little [1973] who showedthat AT protects erythrocytes from radiation-inducedhemolysis. AT subcutaneously administered (400 IU/kg) to mice, 24 h before cobalt-60 irradiation (10.5 Gy),demonstrated significant survival benefit comparedwith AT administration at 12 h preirradiation [Kumaret al., 2002].

Compared with the vast number of studies donewith AT, radioprotective studies conducted withtocotrienols are limited. Recent research at the ArmedForces Radiobiology Research Institute (AFRRI) andthe University of Arkansas for Medical Sciences(UAMS) has focused on γ-tocotrienol (GT3) andδ-tocotrienol (DT3). When tested in the CD2F1 mousemodel, AT or GT3 were administered (400 mg/kg)24 h before 11 Gy Cobalt-60 γ-radiation. The GT3-treated group showed a 40% higher survival than theAT-treated or vehicle groups. Systematic dose optimi-zation, time optimization, and dose-reduction factor(DRF) studies were conducted at AFRRI for boththe tocotrienols, GT3 and DT3. Ghosh et al. [2009]reported the highest efficacy (100% survival at 11 Gycompared with vehicle) by a single dose of 200 mg/kgof GT3 in CD2F1 mice when administered 24 hbefore radiation as compared with three other dosesused (50 and 100 mg/kg, Fig. 2A). A dose-dependent(up to 300 mg/kg s) survival enhancement was observedin CD2F1 mice dosed with either DT3 or vehicle(PEG-400/5% Tween-80) 24 h before 9 Gy radiation(Fig. 2B). When administered 24 h prior to total bodyirradiation, 200 mg/kg of GT3 and 300 mg/kg of DT3exhibited comparable DRFs of 1.29 and 1.26, respec-tively. Both compounds produced enhanced hemato-poietic recovery over 30 days after whole body radiationat sublethal doses. Both GT3 and DT3 also protectedgastrointestinal crypt cells and prevented bacterialtranslocation after exposure to a lethal dose of radiation[Ghosh et al., 2009; Satyamitra et al., 2011].

Preliminary data from AFRRI indicate that GT3administered p.o. was only minimally effective toprotect from lethal radiation doses, in contrast to theefficacy of GT3 given subcutaneously, which protectedagainst even a supra-lethal dose of radiation [Ghoshet al., 2009]. This suggests that the limited residence oforally administered GT3 in the systemic circulation isnot sufficient to provide radioprotection.

The limited systemic residence of the tocotrienolsis due largely to their limited binding to α-tocopheroltransfer protein (ATTP), which is responsible for trans-porting tocols from the liver into the systemic circulation(Fig. 3) [Meier et al., 2003; Singh et al., 2013]. This stepis considered to be the major determinant of tocolplasma levels in humans. GT3, DT3, and other toco-

trienols have a much lower affinity for ATTP than doesAT. Thus, they are trapped in the liver, where they aresubject to clearance by metabolism and biliary excretion[Meier et al., 2003]. On the other hand, AT is activelysecreted back into the circulation by ATTP. The keyrole of ATTP in regulating the tocol pharmacokineticshas been demonstrated elegantly by Traber [2007],who showed that a vitamin E component that does notbind efficiently to ATTP has a reduced plasmahalf-life.

Tocotrienols have short elimination half-lives andthus limited oral bioavailability. The apparent elimina-tion half-lives of exogenously administered tocotrienols

Fig. 2. (A) Radioprotection provided by GT3. Thirty-day survival ofmice (n = 16 per group) treated 24 h before receiving 11 Gy ofcobalt-60 gamma radiation, with a single s.c. injection of vehicle (5%Tween 80) or GT3 at doses of 50, 100, or 200 mg/kg body weight[adapted from Ghosh et al., 2009]. (B) Radioprotection provided byDT3. Thirty-day survival of mice (n = 16 per group) treated 24 h beforereceiving 9 Gy of cobalt-60 gamma radiation, with a single s.c. injec-tion of vehicle (PEG-400/5% Tween-80) or DT3 at doses of between18.75 and 300 mg/kg body weight [adapted from Satyamitra et al.,2011]. [Color figure can be viewed in the online issue which isavailable at wileyonlinelibrary.com]

COMPADRE ET AL.12

Drug Dev. Res.

in humans have been estimated to be between 2 and4.4 h; whereas the half-life for AT is longer than 70 h[Yap et al., 2001; Schwedhelm et al., 2003]. The shortelimination half-life of tocotrienols requires them to beadministered more frequently or in large doses by thesubcutaneous route. A radioprotective agent in an oralformulation with longer half-life and larger area underthe plasma drug level versus time curve (AUC) will bemore suitable for use for the general population. Inaddition, an oral formulation will be easier to stockpileand can be made readily available in case of emergen-cies [Pellmar and Rockwell, 2005]. Thus, it would behighly desirable to develop a new vitamin E analoguewith improved pharmacokinetic properties, specificallyincreased elimination half-life, and increased AUC. Inthis paper, we report the development of one suchcompound, tocoflexol, which maintains the superiorlevel of bioactivity of the tocotrienols and has the poten-tial to achieve a half-life and AUC comparable with thatof the tocopherols.

The process of drug development is a nearly over-whelming task, in which failure is much more likelythan success. In this regard, we have optimized thechances of success by designing tocoflexol via a rigorouscomputer-aided drug design process. The very specificbinding requirements of ATTP cause AT to bind more

efficiently than tocotrienols. A prominent reason forthis selectivity is the flexible tail of the AT moleculewhich allows it to bend in a conformation that can beenclosed in the binding pocket of the ATTP protein. Incontrast, the rotationally restricted tail of GT3 cannotbend in the appropriate conformation for binding.These observations prompted us to design a series ofnovel compounds with enhanced tail flexibility. Accord-ingly, we have named them “tocoflexols.” In tocoflexols,the tridienyl chain of the farnesyl tail of the correspond-ing tocotrienol is substituted with a mono- or dienylchain. Depending on the position and the number ofthe double bonds in the tail, there can be 10 groupsof tocoflexols. Taking into consideration the number ofpossible variations on the head, we can have hundredsof different compounds that could bind to ATTP withvarious degrees of efficiency.

To select the tocoflexols with better possibilitiesto bind to ATTP, potential candidates were screenedusing molecular dynamics simulations [Friedman et al.,2013]. This approach allowed us to reject numerouscompounds with lesser potential, and to focus on themore promising candidates. From this analysis we wereable to select the candidates with the highest potentialto bind ATTP. It follows that these tocoflexols shouldhave enhanced transport from the liver and improved

Fig. 3. Schematic representation of the fate of vitamin E in the body. By selectively recirculating tocols, ATTP plays a central role in maintainingthe plasma levels of the tocols. ATTP, α-tocopherol transfer protein. [Color figure can be viewed in the online issue which is available atwileyonlinelibrary.com]

DESIGN OF BIOAVAILABLE RADIOPROTECTANTS 13

Drug Dev. Res.

distribution via the systemic circulation throughout thebody to the various sites of action. By overcoming therate-limiting step of transport from the liver to the sys-temic circulation, tocoflexols will have extended elimi-nation half-lives, and thus larger AUCs, in comparisonwith the existing tocotrienols.

MATERIALS AND METHODS

Structure Preparation

The initial geometry of AT was taken from thecrystal structure of the complex of AT with human ATTP[Min et al., 2003], entry 1R5L from the Protein DataBank [Bernstein et al., 1977]. The coordinates of ATwere extracted from the complex with the transferprotein, and hydrogens were added. α-Tocotrienol(AT3) was built by combining the chroman ring of AT,prepared as described previously, with the hydrocarbonchain taken from the SQUALN structure from the Cam-bridge Structural Database. The final charges, geometryand energy of the molecules, were determined using theprogram Gaussian03 [Frisch et al., 2004]. The B3LYPhybrid function and the 3–21 G basis set were used forapproximation of the geometry. Other tocol structures inthis study were built based on the structure of AT or AT3using standard bonds and angles.

The δ-analogues were built by modifying thechromanol head of AT by deleting the methyl groups atthe 5 and 7 positions. The tocoflexol structures weregenerated by removing the double bond and thenadding hydrogens at specific positions (Fig. 4). R/Sstereoisomers were generated using the invert routinein SYBYL 8.1 and geometry optimization was done foreach structure using Gaussian03 as described earlier.

Molecular Dynamics Simulations

Dynamics simulations were performed with theTRIPOS force field using SYBYL 8.1 (Tripos, St. Louis,MO, USA). The Gaussian-generated geometry filesdescribed previously were used for the simulations. Thesimulations were run in vacuum with an explicit dielec-tric constant of 4 using the NTV ensemble, with theSHAKE algorithm option applied to all the bondsinvolving hydrogens. The initial system was heated for100 ps at 300 K, with initial atom velocities set torandom with a 32759 seed, and no scaling and thenequilibrated for 1 ns using the velocities from the pre-vious set with the scale velocities option turned on.Finally, for the production run, the system was allowedto evolve for 20 ns at 300 K with no scaling of thevelocities. The structures generated during the produc-tion step were saved at 200 fs intervals to produce sets

of 100 000 conformations for every molecule. Thesesets were considered to be representative of the uni-verse of possible conformations that the molecules canadopt at room temperature.

Binding Analysis

The AT structure in the binding pocket of ATTPwas used as a guideline to fit other tocols. We generatedthe topology files for AT and tocoflexol using thePRODRG server [Emsley et al., 2010]. Energy minimi-zations of tocols and the protein were performed usingINSIGHT-II. Tocoflexol was incorporated in thebinding cavity of ATTP by aligning to AT using Coot[Schüttelkopf and van Aalten, 2004] and the topologyfiles generated by PRODRG.

Pharmacokinetic Modeling

Curves were constructed using the one compart-ment pharmacokinetic model 103 included in theprogram WinNonlin (Pharsight Corporation, MountainView, CA, USA) in the simulation mode. AT and GT3parameters were taken from values reported in theliterature (K = 0.0157/h and F = 0.70 for AT andK = 0.1612/h and F = 0.70 for GT3); for δ-tocoflexol,the parameters were estimated. The simulated dose was100 mg and the time course was 100 h for all threecompounds.

Chemicals

Wistar rat liver microsomes were purchased fromBD Biosciences (San Jose, CA, USA), dimethyl sulfoxide(DMSO) from Fisher Scientific (Waltham, MA,USA), tert-butyl hydroperoxide solution (TBHP) fromSigma-Aldrich (St. Louis, MO, USA), kits to test forthiobarbituric acid reactive substances (TBARS),from Cayman Chemicals (Ann Arbor, MI, USA), andAT and GT3 from Yasoo Health (Jonesborough, TN,USA). δ-Tocoflexol was synthesized through a series ofchemical transformations starting from the correspond-ing δ-tocotrienol (DT3), which was obtained fromannatto seeds. The structure of δ-tocoflexol was con-firmed nuclear magnetic resonance (NMR) and gaschromatography/mass spectrometry (GC/MS) analyses(G. Zheng and CM Compadre, unpublished results].

Inhibition of Lipid Peroxidation

The antioxidant potential of tocoflexol was evalu-ated and compared with that of AT and GT3 in

COMPADRE ET AL.14

Drug Dev. Res.

Fig. 4. Structures of δ-tocoflexol and various representative analogues analyzed by molecular dynamics for potential binding to ATTP. ATTP,α-tocopherol transfer protein.

Fig. 5. General structure of the tocoflexols showing the position of key torsion angles A and B. In this general strcuture R1, R2, and R3 can beH or CH3 groups.

DESIGN OF BIOAVAILABLE RADIOPROTECTANTS 15

Drug Dev. Res.

microsomes by measuring the inhibition of lipidperoxidation. Microsomes were diluted in 0.1 M phos-phate buffer (pH 7.4) at a final protein concentration of1 mg/mL. Microsomes were incubated with 200 μMTBHP in the presence and absence of tocoflexol, AT, orGT3, and were tested at concentrations between 0.5and 7.5 μM. Stock solutions of the test compounds wereprepared in DMSO, stored at −20°C, and used within aweek. Stock solutions of TBHP were prepared imme-diately before each experiment. After incubation withTBHP, samples and standards were mixed with 200 μLacetic acid and 200 μL TBARS reagent and incubatedfor 45 min above 90°C. TBARS reagent contained0.67% TBA in 10% NAOH (w/v). Samples were thencentrifuged for 20 min at 10 000 g at 4°C. Aliquots of200 μL of each sample were loaded in duplicate in a96-well plate and the absorbance was recorded at535 nm using a Spectramax Microplate reader(Molecular Devices, Sunnyvale, CA, USA). The con-centration of malondialdehyde (MDA), a product oflipid peroxidation, was calculated for each sample froma standard curve.

Statistical Analysis

Data are presented as mean ± SEM. InhibitionThe concentration for 50% inhibition (IC50) values fortocoflexol, AT, and GT3 were determined usingWinNonlin pharmacodynamic model 104 and plottedusing R software (Free Software Foundation, Bostoan,MA, USA).

RESULTS

Dynamics Simulations

An analysis of the torsion angles of each set ofconformations was performed to determine whetherthe molecules could adopt conformations capable to fitin the binding cavity of ATTP. This analysis is based onthe assumption that AT is the optimal ligand for ATTP,and that molecules which can adopt conformationssimilar to that of AT in the binding cavity of ATTPwould be able to bind effectively. An inspection of thestructure of AT in the binding pocket of ATTP suggeststhat rotations around the bonds between C(6') andC(7') (torsion A) and C(7') and C(8') (torsion B) of theside chain are key for determining the feasibility ofeffective binding (Fig. 5). In the crystal structure of ATbound to ATTP, the values of torsion angle A andtorsion angle B are 61° and 58°, respectively. Thus, todetermine whether a molecule has possibility of bindingwith ATTP, we made plots of torsion angle A againsttorsion angle B for each one of the molecular dynamics-

derived sets (100 000 conformations for each analogue)(Figs 6–8). Using the values of AT torsion angles [61°,58°] as a reference, a threshold was specified with arange ± 10° (represented with red circles). The confor-mations in the range were considered as allowed con-formations. These conformations can bend to fit insidethe ligand-binding pocket of ATTP in a manner similarto AT.

Molecular dynamics-based screening was able todiscriminate accurately between strong ATTP bindingmolecules such as AT and δ-tocopherol, and weakerbinding molecules such as AT3 and DT3. As shown inFigure 6, the stronger binding molecules have substan-tial populations of conformations with combination oftorsion angles A and B in the allowed conformationarea, while the weaker binding molecules have noconformations in that region. These computationalresults agree very well with the experimentally deter-mined relative affinity of AT and AT3 for ATTP[Hosomi et al., 1997].

Molecular dynamics-based screening predictedthat the number and position of double bonds in theside chain of the tocotrienols is critical for strongbinding to ATTP. As shown in Figure 7, of the variouspossible tocoflexol analogues, only the ones in whichthere is no double bond between C(7') and C(8') havepopulations of conformations in the allowed conforma-tion area.

Molecular dynamic simulation analysis also pre-dicted that the stereochemistry of the tocoflexols iscritical for strong binding to ATTP. Figure 8 suggeststhat tocoflexols lacking unsaturation in position 7' arecapable of strongly binding to ATTP only if theirstereochemistry is equivalent to that of naturally occur-ring AT. Thus, while δ-tocoflexol and δ-toco-11-enylshow substantial populations on the allowed region,their epimeric counterparts show no conformations inthe same region.

Binding Analysis

The crystal structure from PDB entry 1R5Lshows that AT fits completely inside the binding pocketof ATTP in a bent conformation that was largely unfore-seen before the elucidation of the crystal structure. Theflexibility of the tail region of AT allows the tail to bendand fit entirely inside the binding pocket of ATTP. Tovisualize and further analyze the predictions of thedynamics simulations, structural and computationalbinding analyses were conducted to predict the fit oftocoflexol in the binding pocket of ATTP. The crystalstructure of AT in complex with ATTP was used as thescaffold to fit tocoflexol in the binding pocket (Fig. 9).

COMPADRE ET AL.16

Drug Dev. Res.

Pharmacokinetic Modeling

Figure 10 shows the estimated time course ofδ-tocoflexol plasma concentration, as compared withthe AT and GT3 plasma concentrations that wouldoccur for equal oral doses of each compound. Theδ-tocoflexol curve is based on a half-life estimated to be

38% of that of AT and an oral bioavailability F estimatedto be midway between AT and GT3. The former esti-mate is based on δ-tocoflexol showing 38% of theallowed conformations of AT, as shown in Figure 7;while the latter estimate is based on the expectation thatδ-tocoflexol will have an oil/water partition coefficientintermediate to that of AT and GT3.

Fig. 6. Molecular dynamics-based screening accurately discriminates between strong ATTP binding molecules and weak ATTP bindingmolecules. The figure shows the distribution of conformations around torsion angles A and B for 100 000 conformations of each molecule. Inmolecules with favorable conformations for binding to ATTP, such as the tocopherols, both torsion angles must be in the region delineated bythe circles in the figures. Tocotrienols show no conformations in this region. . ATTP, α-tocopherol transfer protein.

DESIGN OF BIOAVAILABLE RADIOPROTECTANTS 17

Drug Dev. Res.

Comparative Antioxidant Activity of Tocoflexol,GT3, and AT in Microsomes

The antioxidant efficacy of tocoflexol was mea-sured for its potential to inhibit lipid peroxidation inmicrosomal membranes and also compared with ATand GT3 using the thiobarbituric assay with MDA asstandard. Tocoflexol, AT, and GT3 were dissolved inDMSO and added to the microsomes at micromolar

concentrations, and their effects on TBHP-inducedMDA formation were then measured at different con-centrations for 30 min. The extent of lipid peroxidationwas calculated by the level of TBARS produced duringincubation with TBHP, as measured at 535 nm. Thetocols used in the study reduced lipid peroxidation in adose-dependent manner (Fig. 11).

TBARS production increased several fold fol-lowing the addition of TBHP in the absence of

Fig. 7. Molecular dynamics-based screening predicts that the number and position of double bonds in the side chain of the tocotrienols arecritical for strong binding to ATTP. The figure shows the distribution of conformations around torsion angles A and B for 100 000 conformationsof each molecule. The analysis suggests that tocols lacking an unsaturation in position 7 are capable of strong binding to ATTP. ATTP,α-tocopherol transfer protein.

COMPADRE ET AL.18

Drug Dev. Res.

tocols. The presence of tocoflexol, AT, and GT3 signifi-cantly reduced lipid peroxidation in a dose-dependentmanner. In micro-molar concentrations, all testedcompounds were able to inhibit lipid peroxidationin microsomes. Tocoflexol (IC50 = 1.4 μM) was moreeffective at quenching oxidative damage in microsomesthan AT (IC50 = 6.7 μM) and equivalent in activity toGT3 (IC50 = 1.7 μM). These findings clearly demon-

strate that tocoflexol is a powerful antioxidant with anIC50 value 4.9-fold lower than that of AT.

DISCUSSION

The increasing threat of radiological terrorism andnuclear accidents creates new demands for discovering

Fig. 8. Molecular dynamics-based screening predicts that stereochemistry of the tocoflexols is critical for strong binding to ATTP. The figureshows the distribution of conformations around torsion angles A and B for 100 000 conformations of each molecule. The analysis suggests thattocoflexols lacking an unsaturation in position 7 are capable of strong binding to ATTP, only if their stereochemistry is equivalent to that ofnaturally occurring AT. AT, α-tocopherol; ATTP, α-tocopherol transfer protein.

DESIGN OF BIOAVAILABLE RADIOPROTECTANTS 19

Drug Dev. Res.

and developing agents that can provide effective pro-tection against ionizing radiation with minimal sideeffects. In recent findings, tocotrienols have shownvarious biological activities transcending traditionalvitamin E antioxidant activity. In the current study, wehave addressed the problem of limited systemic resi-dence of tocotrienols resulting from their lower affinityfor the ATTP.

The predicted binding of vitamin E compo-nents to ATTP using commonly docking tools, e.g.,

AutoDock, does not correlate well with publishedresults [Upadhyay and Misra, 2009]. Therefore, in thisstudy we used molecular dynamics simulations toexplore the binding of different vitamin E componentsand novel analogues to ATTP. Based on these simula-tions, we designed the tocoflexols as novel vitamin Eanalogues with potential for enhanced residence in the

Fig. 9. Comparison of the binding of AT (green, left) and tocoflexol(right, yellow) in the binding cavity of ATTP. Coordinates of ATTP wereobtained from PDB entry 1R5L. Topology files for tocoflexol weregenerated using the program PRODRG. Tocoflexol was fitted in thecavity of ATTP using the program COOT. AT, α-tocopherol; ATTP,α-tocopherol transfer protein. [Color figure can be viewed in theonline issue which is available at wileyonlinelibrary.com]

Fig. 10. Estimated time course of δ-tocoflexol plasma concentration, as compared with the α-tocopherol and γ-tocotrienol plasma concentra-tions that would occur for equal oral doses of each compound. The δ-tocoflexol curve is based on a half-life estimated to be 38% of that ofα-tocopherol and an oral bioavailability F estimated to be midway between α-tocopherol and γ-tocotrienol. [Color figure can be viewed in theonline issue which is available at wileyonlinelibrary.com]

Fig. 11. Comparison of the antioxidant capacity of AT, GT3, andtocoflexol. Inhibition of TBHP induced lipid peroxidation in Wistar ratliver microsomes. Data are means ± SEM of triplicate experiments. AT,α-tocopherol; TBHP, tert-butyl hydroperoxide solution.

COMPADRE ET AL.20

Drug Dev. Res.

systemic circulation. Our data, based on torsion-angleanalysis, predicted that tocoflexol will have 38% moreconformations in the allowed region and is expectedto have better binding to ATTP than its parenttocotrienol. These results also support the notion thatthe novel compound tocoflexol has greater flexibility toacquire conformations similar to AT, as compared withtocotrienols. Further analysis of the structure oftocotrienol demonstrates that its tail region is not asflexible as that of AT. The tocoflexol tail, however, doesdemonstrate flexibility similar to AT. The structurallynovel analogues are likely to be more bioactive andhave longer half-lives and larger AUCs than thetocotrienols.

Radiation exposure leads to severe damage to mul-tiple cellular components such as DNA, proteins, andlipids. One of the mechanisms by which radioprotectantsexert their beneficial effect includes their antioxidantpotential. Current knowledge indicates that tocotrienolsplay a significant role in the prevention of diseases wherefree radicals are involved. Tocotrienols are powerfulnatural antioxidants, do not affect normal tissue andorgans, and have favorable toxicity profiles [Satyamitraet al., 2011; Nowak et al., 2012].

Our results support the hypothesis that tocoflexol,a novel synthetic analogue of tocotrienol, maintains itsantioxidant potential even after side-chain modification.If, as expected, tocoflexol residence time in the body isextended, this compound will be a practical counter-measure for the prevention and mitigation of radiation-induced damage.

The structural modification of drugs to takeadvantage of endogenous transport systems is a noveland intriguing concept whose potential is just starting toemerge. Successful demonstration of its usefulness inthis application is likely to encourage development ofsimilar strategies for future drug design and develop-ment in other areas of biomedical sciences.

ACKNOWLEDGMENTS

This research was supported in part by grantsfrom the National Center for Research Resourcesaward 1ULl RR029884, the IDeA Networks ofBiomedical Research Excellence (INBRE), UAMSstart-up funds, and NIH/NIAID grant AI67798. TheUniversity of Arkansas has applied for patent protectionon the tocoflexols. A potential royalty stream may occurconsistent with the University of Arkansas policy.

REFERENCESBerbée M, Hauer-Jensen M. 2012. Novel drugs to ameliorate gas-

trointestinal normal tissue radiation toxicity in clinical practice:

what is emerging from the laboratory? Curr Opin Support PalliatCare 6:54–59.

Berbée M, Fu Q, Boerma M, Wang J, Kumar KS, Hauer-Jensen M.2009. Gamma-tocotrienol ameliorates intestinal radiation injuryand reduces vascular oxidative stress after total body irradiationby an HMG-CoA reductase-dependent mechanism. Radiat Res171:596–605.

Bernstein FC, Koetzle TF, Williams GJ, Meyer EE, Brice MD,Rodgers J, Kennard O, Shimanouchi T, Tasumi M. 1977. TheProtein Data Bank: a computer-based archival file formacromolecular structures. J Mol Biol 112:535–542.

Emsley P, Lohkamp B, Scott WG, Cowtan K. 2010. Features anddevelopment of COOT. Acta Crystallogr D66:486–501.

Friedman R, Boye K, Flatmark K. 2013. Molecular modelling andsimulations in cancer research. Biochim Biophys Acta 1836:1–14.

Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA,Cheeseman JR, Montgomery JA, Vreven T, Kudin KN, Burant JC,et al. 2004. Gaussian 03, revision C.02. Wallingford, CT, Gaussian,Inc.

Ghosh SP, Kulkarni S, Hieber K, Toles R, Romanyukha L, Kao T,Hauer-Jensen M, Kumar KS. 2009. Gamma-tocotrienol, a tocolantioxidant as a potent radioprotector. Int J Radiat Biol 85:598–606.

Hosomi A, Arita M, Sato Y, Kiyose C, Ueda T, Igarashi O, Arai H,Inoue K. 1997. Affinity for α-tocopherol transfer protein as adeterminant of the biological activities of tocols. FEBS Lett 409:105–108.

Kulkarni S, Ghosh SP, Satyamitra M, Mog S, Hieber K, RomanyukhaL, Gambles K, Toles R, Kao T-C, Hauer-Jensen M, et al. 2010.Gamma-tocotrienol protects hematopoietic stem and progenitorcells in mice after total-body irradiation. Radiat Res 173:738–747.

Kumar B, Jha MN, Cole WC, Bedford JS, Prasad KN. 2002. d-alphatocopheryl succinate (vitamin E) enhances radiation-inducedchromosomal damage levels in human cancer cells, but reduces itin normal cells. J Am Coll Nutr 21:339–343.

Meier R, Tomizaki T, Schulze-Briese C, Baumann U, Stocker A.2003. The molecular basis of vitamin e retention: structure ofhuman a-tocopherol transfer protein. J Mol Biol 331:725–734.

Min KC, Kovall RA, Hendrickson WA. 2003. Crystal structure ofhuman alpha-tocopherol transfer protein bound to its ligand:implications for ataxia with vitamin E deficiency. Proc Natl AcadSci USA 100:14713–14718.

Müller M, Cela J, Asensi-Fabado MA, Munné-Bosch S. 2013.Tocotrienols in plants: occurrence, biosynthesis, and function. In:Tan B, Watson RR, Preedy VR, editors. Tocotrienols: vitamin Ebeyond tocopherols, 2nd ed. Boca Raton, FL: CRC Press, p 1–15.

Nowak G, Bakajsova D, Hayes C, Hauer-Jensen M, CompadreCM. 2012. Gamma-tocotrienol protects against mitochondrial dys-function and renal cell death. J Pharmacol Exp Ther 340:330–338.

Pellmar TC, Rockwell S. 2005. Priority list of research areas forradiological nuclear threat countermeasures. Radiat Res 163:115–123.

Prince EW, Little JB. 1973. The effect of dietary fatty acids andtocopherol on the radiosensitivity of mammalian erythrocytes.Radiat Res 53:49–64.

Satyamitra MM, Kulkarni S, Ghosh SP, Mullaney CP, Condliffe D,Srinivasan V. 2011. Hematopoietic recovery and amelioration of

DESIGN OF BIOAVAILABLE RADIOPROTECTANTS 21

Drug Dev. Res.

radiation-induced lethality by the vitamin e isoform δ-tocotrienol.Radiat Res 175:736–745.

Schüttelkopf AW, van Aalten DMF. 2004. PRODRG—a tool forhigh-throughput crystallography of protein-ligand complexes.Acta Crystallogr D60:1355–1363.

Schwedhelm E, Maas R, Troost R, Boger RH. 2003. Clinicalpharmacokinetics of antioxidants and their impact on systemicoxidative stress. Clin Pharmacokinet 42:437–459.

Singh A, Breen PJ, Ghosh S, Kumar KS, Varughese KI, Crooks PA,Hauer-Jensen M, Compadre CM. 2013. Structural modification oftocotrienols to improve bioavailability. In: Tan B, Watson RR,

Preedy VR, editors. Tocotrienols: vitamin E beyond tocopherols,2nd ed. Boca Raton, FL: CRC Press, p 359–370.

Traber MG. 2007. Vitamin E bioavailability. In: Preedy VR, WatsonRR, editors. The encyclopedia of vitamin E. Cambridge, MA:CABI, p 221–230.

Upadhyay J, Misra K. 2009. Towards the interaction mechanism oftocopherols and tocotrienols (vitamin E) with selected metaboliz-ing enzymes. Bioinformation 3:326–331.

Yap SP, Yuen KH, Wong JW. 2001. Pharmacokinetics andbioavailability of alpha-, gamma- and delta-tocotrienols underdifferent food status. J Pharm Pharmacol 53:67–71.

COMPADRE ET AL.22

Drug Dev. Res.