Tingting Liu,Xioco Wn ,Zheng Luo,Cho Liu,Peng Qun ,Dongm eiCun,?,Ling Fng,??

a Wuya college of Innovation,Shenyang Pharm aceu tical University,103W enhua Road,Shenyang 110016,China b School ofPharm acy,Shenyang Pharm aceutica l University,103Wenhua Road,Shenyang 110016,China

Keywords:Donepezil Orod ispersib le f ilm Taste-m ask ing Dynam ic p rocess E-tongue Abso rp tion

A B S T R A C TThe aim of th is stu dy w as to develop a pa latab le donepezil(DP)orod ispersib le f ilm(ODF)tofacilitate the sw a llow ing p rocess and investigate the effect of cyclodex trin on tastem ask ing based on dynam ic p rocess and in vivo d rug absorp tion.Com p lexation of DPw ith hyd roxyp ropy l-β-cyc lodex trin(HP-β-CD)w as app lied to m ask the bitter taste then the p repared com p lexes w ere in corporated in to ODF using solven t casting m ethod.The tastem ask ing eff iciency w as eva luated by e-tongue;m eanw h ile the pharm acok inetic behavior of DP/HP-β-CD ODFw as investigated by in vivo study.Resu lts show ed the op tim ized f ilm w as m o re pa latab le thandonepezilhyd roch loride(DH)f ilm and w as bioequ iva len tw ith DH.The m o lecu lar m echan ism w as revealed by phase so lubility study,Fou rier-transform in frared spec trom eter(FT-IR),Differen tia l scann ing ca lo rim eter(DSC),X-ray d iffraction(XRD)and m o lecu larm odeling.Taste-m asking w as attribu ted to the fo rm ation of DP/HP-β-CD w h ich w as due tom oderate in teraction betw een DPand HP-β-CD.The stability of DP/HP-β-CD w as decreased because of the acid environm en t in stom ach,w h ich facilitated the absorp tion of DP.These resu lts ex ten ded ou r un derstand ing abou t the app lication of cyclodex trin comp lexation and p rovided gu idan ce fo r the design of ODFespecia lly for d rugs w ith d isgusting taste.

1. Introduction

Recen tly,orod ispersible f ilm(ODF)gets grow ing atten tion because of its un ique superiorities for the target groups in cluding ch ild ren and the geriatric popu lation w ho have d iff icu lty in sw a llow ing[1-2].ODF a llow s d rug to d isperse rap id ly in o ra l cavity[3],thus p rovides m any advan tages such as rap id d isin tegration,im p roved dosing accu racy and stability compared to o rd inary o ra l dosage form s[4-5].Donepezil(DP)is a comm on ly used d rug for the treatm en t of Alzheim er's d isease.Com m ercia l o ra l p rodu cts of DP(tab lets o r capsu les)m ay resu lt in sw a llow ing d iff icu lties,w h ich causes d istress to the m a jor group of A lzheim er's patien ts,the elders.Therefore,in order to im p rove patien t com p lian ce,it is m ean ingfu l to develop donepezil ora lo rod ispersib le f ilm as an a lternative to com m ercia l DP p roducts.M o reover,the recom m ended dosages of donepezil are 5m g and 10m g per day[6],w h ich is app rop riate to develop DPODF[7].

DP has unaccep tab le taste such as bitterness,lead ing to poor com p lian ce du ring the developm en t p rocess of DPODF due to its rap id d isso lu tion in o ra l cavity.Add ing f lavo rs in to the form u lation is the sim p lest and m ost comm on ly used taste-m ask ing m ethod[8].How ever,the add ition of f lavo r is no t e ffective enough to m ask the taste of d rugs like DPw ith ex trem e ly d isgusting taste[9].Cyc lodex trin and its deriva tives are w ide ly used fo r p reparing in clusion com p lexes due to its non-tox icity,h igh biodegradability and various o ther advantages[10].It w as repo rted that w hen d rugs fo rm ed stab le inclusion com p lexes w ith cyc lodex trin,the d isgusting taste of d rugs cou ld be redu ced,even fu lly m asked[11].There fo re,cyclodex trin com p lexation w as a use fu l m ethod fo r tastem ask ing in the deve lopm en t of ODF.Add itiona lly,HP-β-CD show ed h igh so lubility(60%orm ore,w/w)in w ater[12],w h ich w as su itab le to be app lied to develop ODF.

W hen d rug/cyc lodex trin in clusion com p lexes are used to m ask taste in a fo rm u lation,the e ffect of cyclodex trin becom es com p licated due to m u ltip le factors in f luen cing d rug abso rp tion.Excep t for the d rug itself,one of the facto rs is the physicochem ica l p roperties of cyclodex trins.If cyclodextrin has a low so lubility,it possib ly resu lts in a slow d issolu tion of d rug in vivo.Another factor,w h ich is considered as the m ost im portan t one,is the in vivo dynam ic p rocess of d rug/cyclodex trin in c lusion com p lexes.It is sign if ican tly in f luen ced by d rug/cyc lodex trin in teraction and the environm en t in vivo.The ex isten ce ofm oderate in teraction w ill ach ieve a su ccessfu l taste-m ask ing,bu t if the in teraction fo rce betw een d rug and cyclodex trin is too strong,it w ill lead to the decrease of free d rug con cen tration in the in testine,causing low absorp tion[13-14].Add itiona lly,d rugm o lecu les trave l th rough various physio logica l environm en t from m ou th to gastroin testina l tract.Changing of the location m ay a ffect the form of d rug or the stability of com p lexes,w h ich affects d rug abso rp tion.

In a p revious research focusing on the effect of cyclodextrin con cen tration on the o ra l bioavailability,varia tions on pharm acokinetic param eters and p rof iles w ere observed after d rug/cyclodex trin so lu tion w ith d ifferen t cyclodex trin concen tration w as dosed in rats[15].Ano ther research found the sign if ican t in f luen ce of HP-β-CD and su crose on the abso rption of m idazo lam in rabbits[16].These stud ies show ed the im portan t ro le of cyclodex trin in the enhancem en t of so lubility o r taste-m ask ing and suggested the com p licated effect of cyclodex trin on d rug abso rp tion.Neverthe less,the effect of d rug/cyc lodex trin dynam ic p rocess in vivo w as rare ly m entioned or stud ied.Therefore,the know ledge of how cyclodextrin a ffected d rug abso rp tion w h ile taste-m ask ing and how the dynam ic p rocess carried ou tw as still lim ited.

Fig.1-The stru ctu res of DP(a)and HP-β-CD(b).

In th is study,DP/HP-β-CD in clusion com p lexes w ere p repared to m ask the d isgusting taste of DP,and subsequen tly the com p lexes w ere in co rpo rated in to ODF by so lven t casting m ethod.The stru ctu res of DP and HP-β-CD w ere show n in Fig.1.The taste-m asking eff icien cy w as eva luated by etongue,and the pharm acokinetic behavior of the f ilm w as investigated by in vivo study.M o re im po rtan tly,the e ffect of cyclodex trin w as exp lo red via series of techno logies based on dynam ic p rocess and in vivo d rug abso rp tion.

2. M ateria ls and m ethods

2.1. Chem icals and anim als

Donepezil hyd roch loride(DH)w as obtained from Jinan Dexin jia Bio logica l Products Co.,Ltd.(Jinan,Ch ina).HP-β-CD w as pu rchased from Xian De li Bio logica lChem ica lCo.,Ltd.(Xi'an,Ch ina).Hyp rom ellose(HPMC,M ethoce l E5 p rem ium LV)w as a gift from Co lorcon(Shanghai,Ch ina).Polyethy lene glyco l 400 and glycero l w ere p rovided by Bod i Chem ica l Co.,Ltd,(Tianjin,Ch ina).A ll o ther chem ica ls o r so lven tw ere of the h ighest reagen t grade availab le.

W ista r rats(m a le,200±20 g)w ere p rovided by the Experim en ta lAn im a lCen ter of Shenyang Pharm aceu tica lUn iversity(Shenyang,Ch ina).The experim en tsw ere perfo rm ed in acco rdance w ith the Institu tional An im a l Care and Use Comm ittee w ith the app rova l of No.SYPU-IACUC-C2016-1121-201.

2.2. Quan titative analysis ofDP

DP in various sam p les w as determ ined using a HPLC system(Hitach i High-Techno logies Co rpo ration,Tokyo,Japan)consisting of a L-2130 pum p,a L-2420 u ltravio let abso rban ce detector and a L-2200 au tom atic in jecto r.An ODS-C18co lum n(200m m×4.6mm,5μm,Dikm a Technologies Inc.,USA)w as used w ith the tem peratu re set at 40°C.For the in vitro sam p les,them obile phase w as 0.1%g lacia lacetic acid and 0.2%triethylam ine in w ater-m ethano l(40:60,v/v).For the in vivo study,the m obile phase w as 60%m ethano l and 40%glacia l acetic acid so lu tion(0.5%)ad justed to pH 6.5 w ith triethy lam ine.The quan titative ana lysis w as perform ed un der the cond ition of the f low rate at 1.0m l/m in and the w avelength at 271 nm.The in jection vo lum e w as 20μl.A llof the ana ly tica lm ethods have been va lidated.

2.3. Developm ent ofDP/HP-β-CD ODF

2.3.1. Preparation ofDP/HP-β-CD inclusion com p lexes

DP free base w as p repared using liqu id-liqu id ex traction m ethod[17].A fter that,6.90 g of HP-β-CD and 1 g of DP w ere com p lete ly d isso lved in a lcoho l and the so lu tion w as stirred for 12 h at room tem peratu re.A fter the evaporation of alcoho l,the p rodu ctw as co llected fo r fu rther use.

2.3.2. Form u lation design ofDP/HP-β-CD ODF

One-facto r op tim ization m ethod w as used to ob tain the f ina l op tim ized fo rm u lation.Film fo rm ing m ateria ls in c lud ing hyd roxyp ropy lm ethy l cellu lose(HPMC),p regelatin ized starch(PS),po lyethy lene oxide(PEO)and low-substitu ted hyd roxyp ropy l cellu lose(L-HPC)w ere op tim ized acco rd ing to the standards of f ilm-fo rm ing capacity and appearan ce of f ilm s.Plasticizers su ch as po lyethy lene glyco l(PEG),glycerin and p ropy lene glyco l,w ere op tim ized w ith the standards of d issolu tion w ith in 3m in and fo ld ing endu ran ce va lue,w h ich w as determ ined by repeated ly fo ld ing the f ilm at the sam e poin t un til broken[18].Moreover,the d rug content w as con trolled by ad justing the d rug load ing.

2.3.3. Preparation ofDP/HP-β-CD ODF

So lven t castingm ethod w as used to p repare DP/HP-β-CD ODF.Certain am oun t of po lym er,p lasticizer and DP/HP-β-CD in clusion com p lexes w ere d issolved in d istilled w ater and stirred for 1 h at room tem peratu re to obtain the casting so lu tion.A fter degassing,the casting so lu tion w as casted on to a p lastic p late using a laborato ry coating un it(SLT200;KaikaiCom pany,Ltd.,Shanghai,Ch ina)w ith casting heigh t of 1m m,and then the fo rm ed f ilm w as d ried at 54°C fo r 1 h to ob tain the DP/HPβ-CD ODF.Then the f ilm w as cu t in to 2×3 cm2.

As con tro l sam p les,DH ODF,the f ilm con tain ing DH and sucralose,blank f ilm(pu re po lym er)w ere p repared using the sam e m ethod.

2.4. Critica lquality attribu tes ofDP/HP-β-CD ODF

Som e critica l qua lity attribu tes w ere con tro lled in the deve lopm en t of ODF,in c lud ing th ickness,w eigh t,d rug content and m echan ica l p roperties using m icrom eter(M itu toyo,Japan)(n=6),electron ic ana ly tica l ba lan ce CP225D(Sarto rius Instrum en ts system,Ltd.,Beijing,Ch ina)(n=20),HPLC(n=20)and au to stripp ing tester(Labth ink instrum en ts Co.,Ltd.,Ch ina)(n=6)[19],respectively.

2.5. In vitro dissolu tion test

In vitro d isso lu tion test w as perfo rm ed to investigate the d isso lu tion behavior of f ilm in differen t cond itions.Accord ing to the Un ited Sta tes Pharm acopeia(USP),the padd lem ethod w as app lied,and d istilled w ater as w e ll as 0.1m o l/l HClw ere used as d isso lu tion m ed ium.DP/HP-β-CDODF(2×3 cm2,n=6)w as p laced at the bottom of the disso lu tion cup(250m l),then d istilled w ater of 100m l at 37±1°C w as pou red in to the cup and stirred at a ro tation speed of 50 rpm.At p rese lected tim e interva ls of 1,3,5,10,15,20,30 and 60m in,d isso lu tion m ed ium of 5m lw as w ithd raw n and rep laced by equal vo lum e of fresh d istilled w ater im m ed iate ly.A fter cen trifuging at 16 000 rpm,the w ithd raw n sam p les w ere ana lyzed by the HPLC.Fo r the d isso lu tion m ed ium of 0.1m o l/l HCl,a ll the operations w ere the sam e as that of d istilled w ater.

2.6. E-tongue assessm en t

E-tongue assessm en t w as condu cted to eva luate the taste of sam p les using a Taste Sensing System SA402B(In te lligen t Sen sor Techno logy,Ltd.,Japan).Fou r kinds ofm ed icinal senso rs,BT0(for specif ic bitterness of hyd roch lo ride d rugs),AN 0(fo r a lka line bitterness),C00(fo r acid ic bitterness)and AE1(for astringen cy)w ere used separate ly.Befo re the assessm en t,the sam p les w ere d issolved,and the concen tration of DPw as at 0.05m g/m l.Po tassium ch lo ride so lu tion of 10m M w as used as the referen ce and rinse so lu tion.

Acco rd ing to relative senso r ou tpu ts,the d ifferen ce betw een the test sam p le and reference so lu tion in poten tial(Vs-Vr)and CPA va lues w h ich w as def ined as the d ifferen ce(Vr-Vr)betw een the po ten tia ls of the re feren ce so lu tion be fo re and a fter sam p le assessm en t(change of m em b rane po ten tia l caused by adsorp tion of sam p les)w ere ob tained[20-21].Prin cipa l com ponen t ana lysis(PCA)w as used to ana lyze these data[22-23].M eanw h ile,Eu clideand istance from sam p les to DH w as calcu lated accord ing to the fo llow ing Eq.1:

W here p and q w ere the Vs-Vrva lue of sam p les and DH,respectively,i rep resen ted d ifferen t sensors.

2.7. In vivo pharm acokinetic study

2.7.1. Adm inistration and sam p ling

Eigh teen W istar ratsw ere d ivided in to th ree groups random ly,No.1,No.2 and No.3(6 rats fo r each group),w h ich w ere fastened and on ly allow ed free access to w ater for 12 h p rior to the experim en t.

DH aqueous so lu tion w as adm in istered to the ra ts of group No.1 in travenously(3m g/kg).The rats of the group No.2 w as given DH aqueous solu tion via in tragastrical adm in istration rou te(10m g/kg),and fo r the group No.3,DP/HP-β-CD ODFs strips w ere set on the tongue of the rat(10m g/kg).A fter adm in istration,b lood sam p les of 0.2m l w ere taken from orbit at 5,10,15,30,60,90,120,180,240,300,360,420,480,600 and 720m in fo r group No.1 and at 0.25,0.5,1.0,1.5,2.0,3.0,4.0,6.0,8.0,12.0 and 24.0 h fo r group No.2 and No.3.

2.7.2. Plasma sample and pharm acokinetic analysis

Plasm a w as separated im m ed iate ly by cen trifuging at 16,000 rpm and stored at-60°C before analysis.The W in-Non lin(Version 5.2.1;Pharsigh t Co.,Mou tain View,USA)w as used to ca lcu late pharm acok inetic param eters based on the non-com partm en ta lm ode l.

2.8. Statisticalanalysis

Resu ltsw ere exp ressed as them ean±standa rd deviation.Statistica l ana lysis of data w as carried ou t based on tw o-w ay analysis of varian ce(ANOVA)and 90%con f idence in terval ana lysis.The leve l of sign if ican ce w as taken as p＜0.05.Fo r param eters of an im a l study,AUC and Cmaxw ere logarithm ic,and then sim p ly used to eva luate the bioequ iva len ce of these tw o p reparations.

2.9. Exp loration ofdrug/cyclodextrin in teraction

2.9.1. Phase solubility study

In order to investigate the effect of HP-β-CD on the solubility of DP and exp lo re the stability of in clusion com p lexes in d ifferen t cond itions,phase so lubility study w as condu cted acco rd ing to the m ethod estab lished by Higu ch i and Conno rs[24].Aqueous so lu tion and 0.1m o l/l HCl so lu tion of HP-β-CD w ere used as m ed ium,respectively.Excess am oun ts of DP w ere added in to 10m l of aqueous HP-β-CD so lu tion w ith d ifferen t con cen trations(from 0 to 100m M).The obtained suspen sion sw ere shaken at 37°C fo r 48 h.The sam p les w ere co llected and f iltered th rough a M illipo re f iltration of 0.22μm to rem ove the un d isso lved DP.The am oun t of DP in f iltrate w as then determ ined by HPLC and all experim en ts w ere carried ou t in trip licate.The so lubility d iagram w as p lotted w ith the m o la r con cen tration of DP as a fun ction of them o lar con centration of HP-β-CD.The va lue of bind ing constan t(K)w as ca lcu lated th rough the Eq.2.

W here S0w as the so lubility of DP in w ater.Fo r them ed ium of 0.1m o l/l HCl so lu tion of HP-β-CD,a ll the operations w ere the sam e as aqueous HP-β-CD so lu tion.

2.9.2. Fourier-transform infrared spectroscopy(FT-IR)

FT-IR study w as condu cted to exp lo re the in teraction betw een DP and HP-β-CD.The FT-IR spectra of DP,HP-β-CD,physica l m ix tu res and DP/HP-β-CD in clusion com p lexesw ere obtained in the region of 400-4000 cm-1by Bruker Vertex 70 spectrom erter(Billerica,USA).A ll the pow ders w ere m easu red by the KBr disk techn ique.

2.9.3. Therm alanalysis(DSC)

Therm a l ana lysis w as conducted using a DSC-1 system(Mettler To ledo In ternationa l In c.,Sw itzerland).The sam p les of DP,HP-β-CD,physicalm ix tu res and DP/HP-β-CD in c lusion comp lexes w ere accu rately w eighed(abou t 5m g)and p laced in an a lum inum pan w ith the cover sea led.The heat p rocedu re w as at a scann ing rate of 10°C/m in from 25 to 150°C un der a n itrogen pu rge.

2.9.4. X-ray diffractom etry(XRD)

XRD study w as condu cted to investigate the crysta l fo rm of substan ce.XRD patterns of DP,HP-β-CD,physica l m ix tu res and DP/HP-β-CD in clusion com p lexes w ere obtained at room tem peratu re using a DX2700 X-ray d iffractom eter(Ao long Rad ia tive Instrum en t Group Co.Ltd,Dandong,Ch ina)under a vo ltage of 40 kV and a cu rren t of 40m A.A ll the sam p les w ere analyzed in an angle range of 5-50°(2θ)w ith a scann ing step w id th of 2°perm inu te.

2.9.5. M olecu larm odeling

M o lecu larm odeling w as conducted to investigate in teraction betw een HP-β-CD and DP.Au toDock 4.2 softw are w as used to ca lcu late them o lecu larm odel.The in itia lstru ctu res of DPand HP-β-CD w ere obtained using Chem BioDraw U ltra l 14.0,and then these structu res w ere op tim ized using Discovery Stud io 4.0 softw a re.Add itiona lly,the Lam arck ian Genetic A lgorithm(LGA)in Au toDock 4.0 w as used to perform the docking study in a vacuum environm en t.Fu rtherm ore,in order to p robe the DP/HP-β-CD dynam ic p rocess in vivo,the DP/HCl in teraction w as exp lo red using sam e opera tions as DP/HP-β-CD.

3. Resu lts and d iscussion

3.1. Form u lation design ofDP/HP-β-CD ODF

Resu lts of the op tim ization p rocess w ere listed in Tab les 1 and 2.HPMC had p roper f ilm fo rm ing capacity and the f ilm p repared by HPMC had a good appearance;m eanw h ile,the f ilm con tain ing glycerin of 40%(accoun ting fo r the p roportion of po lym er,16.7%of the fo rm u lation)w as excellen t in bo th m echan ica l p roperties and d isso lu tion behavior.Therefore,HPMC w as selected as the op tim ized polym er,and glycerin of 40%w as chosen as the op tim ized p lasticizer.Add itiona lly,DP content w as con tro lled as 5m g/2×3cm2by adjusting DP/HP-β-CD in c lusion com p lexes load ing as 41.4%of the form u lation.The op tim ized form u lation of DP/HP-β-CD ODF con tained DP/HP-β-CD in c lusion com p lexes(41.4%,DP accoun ted fo r 4.7%of the fo rm u lation),HPMC(41.9%),and g lycero l(16.7%).

3.2. Critical qua lity attributes ofDP/HP-β-CD ODF

The critical quality attribu tes of the op tim ized f ilm w ere listed in Tab le 3.The DP content and the w eigh t of f ilm w ere 4.96±0.21m g,129.02±6.29m g(RSD＜5%)(n=20),respectively,w h ich w ou ld be used for fu rther stud ies.

3.3. In vitro dissolu tion test

An o rod ispersib le f ilm shou ld d isso lve qu ick ly to ensu re the rap id d isso lu tion of d rug.In th is study,tw o d ifferen t d isso lution m ed ia w ere se lected to investigate the e ffect of HCl on d issolu tion behavior.Meanw h ile,0.1m ol/l HCl so lu tion w as used to sim u late the gastric environm en t.The d isso lu tion behavio rs of DP/HP-β-CD ODFw ere show n in Fig.2.DPd isso lved rap id ly from the f ilm in both d istilled w a ter and 0.1m o l/l HCl solu tion.The d isso lu tion rate in HClso lu tion w as sign if ican tly faster than that in d istilled w ater at 3m in,w h ich w as attribu ted to the effect of HCl on DP/HP-β-CD com p lexes.Comp lete d isso lu tion w as ach ieved in 15m in irrespective of the pH va lue of d isso lu tion m ed ium,w h ich suggested that DP/HP-β-CD ODF had a favo rab le d isso lu tion p rof ile[19].

Tab le 1-The resu lt of po lym er op tim ization(n=3).

Tab le 2-The resu lt of p lasticizer op tim ization(n=3).

Tab le 3-Critica l qua lity attribu tes of the op tim ized DP/HP-β-CD ODF.

Fig.2-The d isso lu tion p rof iles of DP/HP-β-CD ODF in d istilled w a ter and 0.1m o l/l HCl so lu tion.

3.4. E-tongue assessm en t

Taste-m ask ing w as the key poin t to the design of DP ODF.Electron ic tongue w as sensor array based robotic system[21],w h ich w as often app lied to assess the e ff icien cy of tastem asking effects[25].Based on the poten tiom etry techn ique,Taste Sensing System SA402B equ ipp ing d ifferen t sen so rs w as used to eva luate taste-m ask ing eff icien cy,and each lip id m em b rane sen so r w as associa ted w ith a specif ic taste[25].Liew et al.[26]screened som e f lavors and DH ODF con tain ing 7m g of sucra losew as repo rted to be superio r in term s of taste,a ftertaste,m ou th fee l and accep tan ce.In o rder to investigate the taste-m ask ing effect of HP-β-CD,the f ilm con tain ing DH and sucralose w as p repared accord ing to the p revious study.

The respon ses and CPA va lues of sam p les in c lud ing DH,DP/HP-β-CD in clusion com p lexes,the physica lm ix tu res,DP/HP-β-CD ODF,the f ilm con tain ing DH and su cra lose,DH ODF and the blank f ilm w ere obtained.The PCA m ap d isp laying 4 senso rs ou tpu ts w as show n in Fig.3.Acco rd ing to the sca le of every senso r in the ex tracted p rin cipa l com ponen t,the PC1(69.88%)tended to exp ress the bitterness(AN 0:0.976,C00:0.959),w h ile the PC2(29.13%)m ain ly show ed the astringen cy of the substan ces(AE1:0.997).It w as found that sam p les w ith sim ilar taste gathered together and fo rm ed a c luster.The sam p les includ ing DH,the physical m ixtu res,DH ODF and the f ilm con tain ing DH and su cra lose gathered in to a cluster,w h ich suggested that the taste w as unp leasan t,or taste-m ask ing effect w as w eak.DP/HP-β-CD inc lusion com p lexes and DP/HP-β-CD ODF clustered together,w h ich in d icated that cyclodex trin com p lexation w as sign if ican tly effective for taste-m asking and the DP/HP-β-CD in clusion com p lexesw ere stab le in the f ilm.Ano ther separate c luster form ed by the blan k f ilm show ed sligh tly astringen t taste,w h ich w as ascribed to the taste of the f ilm-fo rm ing m ateria l(HPMC)itself.In o ther sam p lesw ith d rug,th is astringen t taste w as possib ly d ilu ted by the add ition of d rug or f lavor,so they exh ibited a d ifferen t degree of astringen cy.

Fig.3-PCA m ap d isp lay ing 4 sen so rs ou tpu ts in a tw o-d im en siona l g raph.

Fig.4-Eu clideand istan ces from sam p les to the d rug substan ce(DH)(b).?in d icated a sign if ican t d ifferen ce in statistics.

Euc lideandistan ces from sam p les to DH w ere show n in Fig.4.Usua lly,larger d istan ce in d icated better taste-m asking[21].Eu clideand istan ce of in clusion com p lexes and DP/HPβ-CD ODF w ere sign if ican tly larger than that of o ther samp les.Therefore,it w as fu rther p roved that cyclodex trin comp lexation w asm o re effective than com m on f lavo rs fo r tastem ask ing in design ing o rod ispersib le f ilm of DP.

3.5. In vivo pharm acokinetic study

The p lasm a con cen tration-tim e p rof iles w ere show n in Fig.5,and pharm acokinetic param eters w ere listed in Tab le 4.Usually,orod ispersib le f ilm exh ibited a fast d isin tegration in o ra l cavity,then it w as sw a llow ed and abso rbed in gastrointestina l tract[1].In th is study,it w as found that DP w as detectab le im m ed iate ly a fter the ora l adm in istration of DP/HPβ-CD ODF(5m in),w h ich dem onstrated that DPw as disso lved from DP/HP-β-CD ODF and en tered in to b lood circu lation rap id ly.The abso lu te bioavailability and re lative bioavailability of DP/HP-β-CD ODFw ere 57.46%and 110.05%,respectively.In add ition,AUC and Cmaxof DP/HP-β-CD ODF and DH so lution show ed no sign if ican t d ifferen ce based on ANOVA and 90%con f iden ce in terva lana lysis.The resu lt ind icated that the DP/HP-β-CD ODFw as bioequ iva len tw ith DH.

Tab le 4-Pharm acok inetic param eters of DH a fter injection of DH(3m g/kg),fo llow ing o ra l gavage of DH(10m g/kg)and o ra l cav ity adm in istra tion of DP/HP-β-CD ODF(10m g/kg)(n=6,m ean±SD).

3.6. Exp loration ofdrug/cyclodextrin in teraction

The effect of HP-β-CD on DPw as exp lo red in term s of phase so lubility study,FT-IR,DSC,XRD and m o lecu lar m odeling.It cou ld be con cluded that the aqueous so lubility of DPin creased as the con cen tration of HP-β-CD in creased by the resu lts of phase so lubility study.For the m ed ium of aqueous HP-β-CD so lu tion,the regression equation w as Y=0.2039X+0.00007,r2=0.9995,w here Y w as the con cen tration(m o l/l)of DP,X w as the con cen tration(m o l/l)of HP-β-CD.The obtained d iagram w as classif ied as ALtype,w h ich suggested the inclusion com p lexes w ere fo rm ed in the stoich iom etric ratio of 1:1[24].The in clusion com p lexes w ere stab le in w ater because bond ing constan t(K)of DP/HP-β-CD in clusion com p lexes w as 3658.91 M-1accord ing to the Eq.2 w h ich w as c lassif ied as a m oderate bind ing[27].Fo r the m ed ium of 0.1m o l/l HCl HP-β-CD so lu tion,the regression equation w as Y=0.1516X+0.0372,r2=0.9991,and the bind ing constan tw as 4.80M-1,w h ich w as sign if ican t low.Genera lly,larger bind ing con stan t(K)ind icated m o re stab le com p lex[27].The d ram atic change of K from 3658.91 M-1to 4.80 M-1ind icated a decrease in stability in the p resen ce of HCl.

The FT-IR spectra of DP,HP-β-CD,the physica lm ix tu re and DP/HP-β-CD in clusion com p lexes w ere show n in Fig.6.The in tense sharp abso rp tion band in the FT-IR spectrum of DP at 1685.6 cm-1and 1308.5 cm-1w ere assigned to the characteristic peak of C=O and the C-N stretch ing vibration[26].The in tense b road abso rp tion band at 3416.6 cm-1in the spectrum of HP-β-CD w as attribu ted to the stretch ing vib ration of O-H[10].These characteristic bands show ed a sim p le sum in the spectrum of the physica l m ix tu res,w h ich ind icated that HP-β-CD and DP justm ixed w ith each other.How ever,in the spectrum of DP/HP-β-CD in clusion com p lexes,the cha racteristic absorp tion of C=O at 1685.6 cm-1d isappeared,and the absorp tion of C-N at 1308.5 cm-1d im in ished greatly.It dem onstrated that som e groups of DPw ere em bedded in to the cavity of HP-β-CD successfu lly.Add itionally,it show ed abou t 2.0 cm-1w avenum ber variation s of O-H abso rp tion of HP-β-CD,w h ich ind icated the ex isten ce of in teraction betw een DP and HP-β-CD.

Fig.5-Pha rm acok inetic p rof iles of DH so lu tion a fter in tragastrica l adm in istration and DP/HP-β-CD ODF a fter o ra l cav ity adm in istration(A);Pha rm acok inetic p rof ile of DP a fter in jection(B).

Fig.6-The FT-IR spectra of sam p les:DP(a),HP-β-CD(b),the physica lm ix tu res(c)and DP/HP-β-CD in clusion com p lexes(d).

The therm ogram s of DP,HP-β-CD,physica l m ix tu res and DP/HP-β-CD in clusion com p lexes w ere show n in Fig.7.The endo therm ic peak starting from 94.33°C in the p rof ile of DP w as assigned to the m elting poin t of DP[17].The peak in the p rof ile of HP-β-CD in d icated that HP-β-CD w as am o rphous.The therm a l peak from 60°C to 110°C w as attribu ted to the evapo ration ofw aterm o lecu les.The physica lm ix tu res of DP and HP-β-CD show ed a sim p le superposition of the endo therm ic peaks of DP and HP-β-CD,w h ich suggested that there w as no in teraction betw een DP and HP-β-CD.In the cu rve of DP/HP-β-CD in clusion com p lexes,the endo therm ic peak of DP d isappeared.It w as reported that m elting poin t of guestm o lecu les sh ifted to d ifferen t tem peratu re o r d isappeared w hen they inserted in to cavities o r crysta l lattice of CD[28].Therefo re,the resu ltsm en tioned above ind icated the fo rm ation of DP/HP-β-CD com p lexes.

Fig.7-The DSC cu rves of sam p les:DP(a),HP-β-CD(b),the physica lm ix tu res(c)and DP/HP-β-CD in clusion com p lexes(d).

The XRD p rof iles of DP,HP-β-CD,the physica l m ix tu res and DP/HP-β-CD in clusion com p lexes w ere show n in Fig.8.Again,DP alone p resen ted the typ ical XRD pattern of crysta lline state,i.e.,sharp characteristic d iffraction peaks,w h ile HP-β-CD ex isted as am orphous form.The d isappearan ce of d iffraction peaks of DP in the pattern of DP/HP-β-CD in clusion com p lexes,w h ich w ere still detectable in the pattern of physica lm ix tu res,in d icated the com p lete am orph ization of DP a fter it inserted in to the cavity of HP-β-CD.

The DP/HP-β-CD in teraction w as fu rther exp lo red by m olecu lar m odeling.As show n in Fig.9,the m ost p robable m odes of DP/HP-β-CD in c lusion com p lexes w ere determ ined by the p rogram of Au todock.Hyd rogen bond ing w as found betw een the carbony l,tertia ry n itrogen groups of the guest m o lecu le and the hyd roxy l group of host m o lecu le,respectively,w h ich w as in acco rdan ce w ith the resu lts of FT-IR.Add itiona lly,the m in im um energy DP/HPβ-CD w as-6.51 kca l/m o l.How ever,the m in im um energy of DP/HClw as-23.46 kcal/m o l,w h ich w as far less than DP/HP-β-CD in clusion com p lexes.There is no de f in ite regu lation abou t the m in im um energy va lue in m o lecu la rm ode ling.Genera lly,low er energy va lue ind icated stronger stability[29].These resu lts suggested that DP/HCl in teraction w as m u ch stronger than DP/HP-β-CD in teraction,w h ich verif ied the resu lts of phase so lubility study.Therefo re,the reason for successfu l taste-m ask ing w as that DP m o lecu le w ith unp leasan t taste en tered in to the cavity of HP-β-CD in w hole or in part,and the com p lexes w as re latively stab le in m ou th due to m oderate DP/HP-β-CD in teraction.

Fig.8-The XRD pattern s of sam p les:DP(a),HP-β-CD(b),the physica lm ix tu res(c)and DP/HP-β-CD in clu sion com p lexes(d).

A fter the DP/HP-β-CD ODFw as app lied to the tongue of rat,a series of changes occu rred in DP/HP-β-CD in clusion comp lexes and the dynam ic p rocess w as closely related to d rug abso rp tion.For the op tim ized DP/HP-β-CD ODF,the pharm acok inetic p rof ile w as sim ilar to the DH so lu tion acco rd ing to in vivo study.The reason w as exp lained from the fo llow ing tw o aspects.On one hand,the environm en t had changed d ram atica lly from m ou th to stom ach,w h ich led to a variation of DP/HP-β-CD in clusion com p lexes.As show n in phase so lubility study,the bind ing constan t(K)changed from 3658.91 M-1to 4.80 M-1in the p resen ce of HCl.From the resu lts of m o lecu le m odeling,DP w as easier to in teract w ith HCl than HP-β-CD.Meanw h ile,the f ilm exh ibited better d isso lu tion behavio r in 0.1m o l/l HCl so lu tion acco rd ing to the resu lts of in vitro d isso lu tion test(Fig.2),ind icating that m ost of hyd rogen bond ing betw een DP and HP-β-CD m ay be b roken due to the acid ic environm en t of stom ach,w h ich facilitated the DPabso rp tion.On the other hand,the abso rp tion of cyclodextrin occu rred in the co lon w here cyclodex trin w asm etabo lized by co lon ic bacteria[27].DP/HP-β-CD in clusion com p lexes d issolved qu ick ly from f ilm due to the am orphous form of DPa fter it inserted in to the cavity of HP-β-CD.A fter that,the DP/HPβ-CD in clusion com p lexes qu ick ly passed th rough the static w ater layer w ith the he lp of HP-β-CD and reached to the su rface of biof ilm.Since cyclodex trin s cou ld not pass th rough the lip id biof ilm,there w as a dynam ic ba lan ce betw een the in clusion and the free DP.A sm a ll bind ing constan t(K=4.80 M-1)w as bene f icia l fo r DP across the biof ilm[27].Therefo re,rap id d issolu tion,and the decreased stability of DP/HP-β-CD inclusion com p lexes in stom ach,exp lained w hy the DP/HP-β-CD ODFw as bioequ iva len tw ith DH so lu tion w h ile taste-m asking.

Fig.9-M ode ls of the DP/HP-β-CD in clu sion com p lexes and DP/HCl from dock ing ca lcu lation:DP/HP-β-CD in clusion com p lexes(A)and DP/HCl(B).

In summ ary,taste-m asking w as ach ieved by m oderate interaction betw een DP and HP-β-CD,w h ich m ade DP/HP-β-CD stab le in o ra l cav ity.M o reover,DP/HP-β-CD ODF show ed favo rab le pharm acok inetic behavior in rat,w h ich w as m ain ly attribu ted to the decrease of stability of com p lexes in stomach based on DP/HP-β-CD dynam ic p rocess.

4. Con clu sion s

A rap id d isso lving DP/HP-β-CD ODF w as deve loped,and the taste w as m asked successfu lly.Apart from fu lf illing the o rigina l aim of im p roving the bitter taste and so lving the sw a llow ing d iff icu lties,the effectof cyclodex trin on taste-m asking w as illustrated at m o lecu lar leve l based on dynam ic p rocess and d rug absorp tion in vivo.It w as found that tastem ask ing w as ach ieved su ccessfu lly by the m oderate DP/HPβ-CD in teraction.Add itionally,the varied stability of DP/HPβ-CD in clusion com p lexes from o ra l cavity to stom ach facilitated d rug absorp tion.These con c lusions ex tended ou r understand ing abou t the app lication of cyclodex trin comp lexation in the developm en t of ODF and p rovided gu idan ce fo r the design of ODF,especia lly fo r the d rugs needed taste-m asking.

Con f lict of In terest

The au tho rs declare that there is no con f licts of in terest.

Acknow ledgm en ts

Th is research d id not receive any specif ic gran t from fun d ing agen cies in the pub lic,com m ercia l,o r no t-for-p rof it secto rs.

Sup p lem en tary m ateria ls

Supp lem en tary m ateria l associated w ith th is article can be found,in the on line version,at doi:10.1016/j.a jps.2018.05.001.