W eiQin ,Yuanzhi He,Zhen Guo,Liu Zhang,LiWu ,

Xianzhen Yinb,Shailendra Shakyab,d,AbiM aharjan b,d,Yan Tanga,b,W eifeng Zhu a,?,Jiw en Zhanga,b,??

aM inistry ofEducation,JiangxiUniversity ofTraditiona lChinese M edicine,Nanchang 330004,China

b Shanghai Institu te ofMateria M edica,Chinese Academ y ofSciences,Shanghai201203,China

c School ofPharm acy,Shenyang Pharm aceutica lUniversity,Shenyang 110016,China

d SchoolofPharm acy,University ofChinese Academ y ofSciences,Beijing 100049,China

Keywords:Ibup rofen m icrosphere Sp ray congea ling In terna l structu re Syn ch rotron rad iation X-ray m icro-com pu ted tom ography

A B S T R A C TThe m icrosphere w as a p rim ary particu late system fo r taste-m asking w ith un ique structu ra l featu res def ined by p roduc tion p rocess.In th is article,ibup rofen lip id m icrospheres of octadecanol and glycerin m onostearate w ere p repared to m ask the undesirable taste of ibup rofen via th ree kin ds of sp ray congea ling p rocesses,nam e ly,air-coo ling,w ater-coo ling and citric acid so lu tion-coo ling.The stereoscop ic and in terna l stru ctu res of ibup rofen m icrospheres w ere quan titatively ana lyzed by synch rotron rad iation X-ray m icro-com pu ted tom ography(SR-μCT)to estab lish the relation sh ip betw een the p reparation p rocess and m icrosphere arch itectu res.Itw as found that them ic rostructu re and m o rpho logy of them icrospheresw ere sign if ican tly in f luenced by p reparation p rocesses as the p rim ary fac tors to determ ine the release p rof iles and taste-m asking e ffects.The sphericity of ibup rofen m icrospheres congealed in citric acid so lu tion w as h igher than that of other tw o and itsm orpho logy w asm o re regu lar than thatbeing congea led in air o r d istilled w ater,and the con tac t ang les betw een congea lingm ed ia andm elted ibup rofen in octadecano land glycerin m onostearatew elldem on strated the structu re d ifferen ces am ongm icrospheres of th ree p rocesses w h ich con tro lled the release charac teristics of them icrospheres.The structu re param eters like po rosity,sphericity,and rad ius ratiofrom quan titative ana lysis w ere co rrelated w ell w ith d rug re lease behavio rs.The resu lts dem onstrated that the ex terior m o rpho logy and in ternal stru ctu re ofm icrospheres had considerable in f luences on the d rug release behavio rs as w ell as taste-m asking effects.

1. Introduction

Taste-m asking had great im portan ce for active pharm aceu tica l ingred ien ts w ith an unp leasan t taste in ped iatric m ed icines as w ell as fo r the sen sitive pa tien t popu lation.Recen tly,m any taste-m asking techno logies had been developed,such as sp ray congealing[1],hot-m elt ex trusion[2-4],m ixing w ith f lavo rs and sw eeteners[5-7],com p lexation w ith ion exchangers and cyc lodex trins[8-10],en capsu la tion in to m icrospheres and m icrocapsu les[11],physica l barriers o r coating[12-14],so lid d ispersions[15],and granu lation[16].Many m icroen capsu lation p rocessesw erem od if ications of th ree basic techn iques,nam ely so lven t ex traction/evaporation,phase separation(coacervation)and sp ray congealing.How ever,so lven t ex traction/evaporation and coacervation w ere im paired by residua l so lven ts and coacervation agen ts found in the p roducts ran sa fety risks for their app lications[17].Therefo re,sp ray congealing that w as relatively sim p le and possess h igh th roughpu t w as usua lly m u ch sa fer.

Basica lly,the aim of taste-m asking w as to redu ce o r inh ibit the in teraction betw een the active ingred ien ts and the taste buds in tongue.Sin ce the in teraction of the d isso lved d rug m o lecu les w ith the taste recep to rs in the hum an taste buds caused taste reaction,them easu rem en t of them oiety of d rug re leased or d isso lved in sim u lated o ra l cavity cond ition s w as an ind icator to the taste-m asking effects[18].

Fu rtherm o re,the m o rpho logy w as one of the p rim ary facto rs govern ing the rate and m echan ism of the d rug re lease from so lid dosage form s,su ch as su rface and shape of the dosage fo rm.Fo r instan ce,the rough su rface ofm icrospheres often show ed faster release rate than the sm oo th m icrospheres[19].The strong re lationsh ip betw een the m o rpho logy of the pellets and the value of them eand isso lu tion tim e(MDT)had been estab lished[20].

Besides d rug re lease tests,o ther conven tiona l m ethods had been em p loyed fo rm icro-particle eva luations,such as opticalm icroscopy for size analysis[21],scann ing electron m icroscopy(SEM)fo r m orpho logy detection[22],transm ission e lectron m icroscopy(TEM)for po rosity ana lysis[23],d ifferentia l scann ing ca lorim etry(DSC)fo r therm odynam ic determ ination,and X-ray pow der d iffraction(XRPD)for crystal eva luation[24].These m ethods cou ld on ly eva luate taste-m asked p roducts to som e ex ten t and w ere d iff icu lt to iden tify the interna l m icrostru ctu re's ro les in taste-m asking.Therefo re,a quan titative stru ctu re characterization m ethod had especial im po rtan ce to co rrelate the detail stru ctu ra l in fo rm ation of m icrospheres w ith the d rug release behavio rs,p rovid ing new m echan istic know ledge fo r taste-m asking techno logies.

Syn ch rotron rad iation X-ray m icro-com pu ted tom ography(SR-μCT),a pow erfu l non-invasive investigation techn ique for the in terna l th ree d im ension(3D)structu res fo r various ob jects,had great poten tial for quan titative evaluation and Ibup rofen w ith pu rity of 99.45%w as p rovided by Hunan Erkang Pharm aceu tica l Co.,Ltd.(Ch ina).Glycerin m onostearate(GMS)and octadecano l,as the excip ien ts fo r tastem ask ing w ere supp lied by Hunan Erkang Pharm aceu tica l Co.,Ltd.(Ch ina).Sod ium phosphatem onobasic dehyd rate,sod ium hyd rogen phosphate and phosphoric acid w ere obtained from Sinopharm Chem ica l Reagen t Co.,Ltd.(Ch ina).design of so lid d rug de livery system s.The in terior po rous channe ls and irregu lar stru ctu res cou ld be ob tained by SRμCT,w h ich cou ld be fu rther correlated w ith the d rug release k inetics of fe lod ip ine osm otic pum p tab lets[25].On one hand,the su rface m o rpho logy and the in terna l 3D stru ctu re of felod ip ine osm o tic pum p tab lets cou ld be visua lized;on the other hand,the in trin sic d rug release kinetics and 3D param eters,su ch as su rface areas of the rem ain ing co re inside the f ilm,cou ld be quan tita tive ly e lu cidated using SR-μCT[26].Com bined w ith statistica lm ethod,SR-μCT w as used to non-destructively investigate the m ixing and segregation of granu lar m ateria ls in 3D[27].Add itiona lly,SR-μCT w as a lso app lied fo r structu ra lana lysis of po lym er-coated granu les and spherica l ho llow excip ien ts on tab leting[28].

As a w idely used non steroidal an ti-in f lam m atory d rug,ibup rofen is rap id ly and com p lete ly absorbed in the gastroin testina l tract a fter ora l adm in istration.Strong pungen t taste of ibup rofen m ust be m asked fo r the sensitive patien ts.In th is study,ibup rofen lip id m icrospheres p repared by sp ray congea ling in d ifferen t coo ling m ed ia w ere examined by SR-μCT to un cover the in terna l 3D structu res in line w ith taste-m asking.The p rim ary objectivesw ere to study the m icrostru ctu ra l basis of lip id m icrospheres for taste-m asking and to p rovide new eva luation m ethod fo r p rodu ction p rocess op tim ization du ring taste-m ask ing.

2. M ateria ls and methods

2.1. M aterials

Ibup rofen m icrospheres w ere p repared using a sp ray ing techn ique(Sp ray gun,W-71,Iw ata,Japan).The d isso lu tion test w as perform ed using a d isso lu tion apparatus(ZRS-8G,Tian jin Haiyida Co.,Ltd.,Ch ina).The SR-μCT scansw ere carried ou t at the Shanghai Syn ch rotron Rad ia tion Facility(SSRF)in Shanghai Institu te of App lied Physics,Ch inese Academ y of Scien ces.Data w ere analyzed using the com m ercially availab le softw are VGStud io Max(version 2.1,Vo lum e Graph ics Gm bH,Germ any),Im age-p ro Plus(version 6.0,M ed ia Cybernetics,In c.,Bethesda,MD,USA),Am ira(version 5.4.3,Visage Im aging,In c.,Australia)and Im age Pro analyzer 3D(version 7.0,Med ia Cybernetics,In c.,Bethesda,MD,USA).

2.2. Preparation of ibuprofen m icrospheres

The fo rm u las of ibup rofen m icrospheres w ere selected by screen ing the ratio of ibup rofen,octadecano l,and GMS.Acco rd ing to them o rpho logy,f low ability,and d isso lu tion in m ed ia of m icrospheres,the ratio of ibup rofen,octadecano l,and GMSa t3:6:1(w/w/w)w as the best fo rm u la.M icrospheresw ere p repared by heating app rop riate am oun t of GMS and octadecanol to 65-70°C un tilm elted.Then,ibup rofen w as gradually added and m elted w h ilst stirring at 500 rpm by a m agnetic heated stirrer(ETS-D5,IKA,Germ any)to get a un ifo rm m ixtu re.The m ix tu re w as im m ed iate ly pou red in to the reservoir of sp ray gun p reheated to 70°C in a d ry oven and sp rayed.The air-coo ling m icrospheres w ere co llected from stain less stee l trays.The w ater and citric acid so lu tion(pH=1.2)w ere a lso used as coo ling m ed ia,w h ile the coo led m icrospheres w ere qu ick ly form ed in 1000m l of correspond ingm ed ium by a position of sp raying at 5 cm above the m ed ium.It w as then f iltered and d ried in a vacuum oven at 40°C.The m icrospheres w ere co llected and sto red in desiccator for fu rther study.The m icrospheres w ere nam ed as ACM,WCM and CCM for aircoo ling m icrospheres,w ater-coo ling m icrospheres,and citric acid so lu tion-coo lingm icrospheres,respectively.

2.3. Con tact anglem easurem en t

The con tact angles betw een w ater,citric acid so lu tion(pH=1.2)and ibup rofen m icrospheres w ere determ ined under am bien t tem peratu re by a gon iom etry m ethod.The sam e as the p reparation of ibup rofen m icrospheres,app rop riate am oun ts of GMSand octadecano lw ere heated to 65-70°C and ibup rofen w as gradua lly added and m e lted w h ilst stirring at 500 rpm to get a un ifo rm m ix tu re.The m ix tu re w as im m ed iate ly pou red on to a w atch-g lass and w as coo led tofo rm a un iform f ilm.An app rop riate size of f ilm w as selected and w as pu t on a ho rizon ta l tab le.App roxim ate ly 20μl of the congea lingm ed ia,nam ely,w ater o r citric acid so lu tion w ere deposited on the f ilm from a consisten t heigh t.A fter the d rop reached a quasi-equilibrium shape,photosw ere taken imm ed iately by a cam era and w ere ana lyzed using Im age-p ro Plus(version 6.0,M ed ia Cybernetics,In c.,Bethesda,MD,USA)to ca lcu late the con tact ang le.A ll m easu rem en ts w ere repeated in trip licate by th ree ind ividua l f ilm s.

2.4. Therm ogravim etric analysis

The therm ogravim etric ana lysis(TGA)m easu rem en ts w ere perfo rm ed by a Perk in Elm er Therm ogravim etric ana lyzer Pyris 1 TGA at 10°C/m in in the n itrogen atm osphere.Samp les(app rox.5m g)w ere w eighed in a hanging a lum inum pan and the w eigh t loss percen tage of the sam p lesw asm on itored from 25 to 500°C.In o rder to com pare the TGA p rof iles of va rious sam p les,the raw data of TGA w ere re-d raw n.

2.5. X-ray pow der diffraction

The crysta llin ity of the sam p les w as characterized by Xray pow der d iffraction(XRPD).The d iffraction patterns w ere detected w ith a Bruker D8-ADVANCE X-ray d iffractom eter(Bruker,Germ any)using Cu-Kɑrad iation(λ=1.54056?A)at ambien t tem peratu re(20°C).The vo ltage and cu rren tw ere 40 KV and 40 m A,respectively.Ref lection m ode w as used in the 2θ range 3°-40°w ith a scan speed of 15°/m in(step size 0.025°,step tim e 0.1 s)by a lynxEye detecto r.

2.6. Drug release in d ifferen tm edia

For the in vitro d isso lu tion test to im itate pH cond ition s in vivo,m icrospheres con tain ing 200 m g ibup rofen in fraction w ith particle size betw een 125 and 212μm w ere sieved by 70 to 115m esh screen and w ere assessed in sex tup le for each batch ofm icrospheres.Disso lu tion tests w ere perfo rm ed using the padd lem ethod acco rd ing to the Ch inese Pharm acopoeia 2015 w ith the rotation speed of 100 rpm.A vo lum e of 900m l for the d istilled w ater,hyd roch lo ric acid so lu tion(pH=1.2)o r phosphate bu ffer(pH=6.80 and 7.20)w ere used as the d isso lution m ed ia at 37±0.5°C.A liquo ts of 2m L of so lu tion sam p les w ere w ith d raw n at 10,30,60,90,120,180,240,300,600 s and an equa l vo lum e of the fresh m ed ium w ith sam e tem peratu re w as added im m ed iately.

The a liquo t sam p les w ere im m ed iately f iltered th rough a 0.22μm m em brane and quan tif ied on a HPLC-UV system that consisted of an Agilen t 1260 series HPLC system(Agilen t Techno logies In c.,Ch ina)coup led to an Agilen t 1290 UV detecto r.Sam p les w ere exam ined by a Dikm a Spu rsil C18co lum n(250m m×4.6m m,5μm,Phenom enex,USA)at 25°C.Them obile phase w as consisted of sod ium acetate bu ffer(pH=2.5)and aceton itrile(30:70,v/v).Elu tion w as perfo rm ed at 25°C at a f low rate of 1m l/m in.The detection w avelength w as set at 263 nm.

2.7. Conven tionalm orphology characterization

A ll the kinds of m icrospheres collected du ring d isso lu tion tests at 0,5,10m in w ere investigated by conven tiona l m o rpho logy characterization of scann ing electron m icroscope(SEM,S-3400,Hitach i).The specim ens w ere im m obilized on a m etalstub w ith double-sided adhesive tape.Them icrosphere sam p les at d isso lu tion tim e of 5m in and 10m in w ere grinded gen tly befo re m easu rem en t,so as to co llect the im ages of interna l su rfaces from the broken m icrospheres.

2.8. Synchrotron radia tion X-ray m icro-com pu ted tom ography m easurem en t

To characterize the m icrostru ctu re of m icrospheres,SR-μCT im ages w ere acqu ired w ith beam line BL13W 1 a t SSRF.The p ropagation-based im aging(PBI)m ethod w as used fo r the phase con trast im aging to investigate the d istribu tion of low Z m aterials in the m icrospheres.Sam p les w ere scanned at 13.0KeV fo r the h igher f lux to redu ce the im aging tim e.A fter penetration th rough the sam p le,the X-rays w ere converted in to visib le ligh t by a Lu2SiO5:Ce scin tillator(10m m th ickness).Pro jections w ere m agn if ied by d iffraction-lim ited m icroscope op tics(10 m agn if ications)and d igitized by a h ighreso lu tion 2048 p ixel×2048 p ixe l sCMOS cam era(ORCA Flash 4.0 Scien tif ic CMOS,Ham am atsu K.K,Sh izuoka Pref.,Japan).The pixel size w as 0.65μm,the sam p le-to-detector d istance w as 12.0 cm and the exposu re tim e w as 1 s.Fo r each acqu isition,1440 p ro jection im ages w ere cap tu red.Flat-f ie ld and dark-f ield im ages w ere also co llected du ring each acqu isition p rocedu re to co rrect the electron ic noise and varia tions in the X-ray sou rce b righ tness.

2.9. Three-dim ension structure reconstruction

The tw o-d im ension(2D)p ro jection CT scann ing transm ission im ages w ere acqu ired and ana lyzed a fter op tim ization,noise reduction,shad ing and segm en tation to evaluate p relim inarily the effect of im aging and the im aging param eters w ere op tim ized.At the sam e tim e,the abso rp tion of X-ray of d ifferen t sam p les w as ob tained to qua litative ly com pare the stru ctu ral arch itectu re of sam p les.Based on a series of 2D p ro jection im ages of d ifferen t angles,slice im ages w ere acqu ired by reconstructing th rough X-TRACT softw are(CSIRO,Com m onw ea lth Scien tif ic and Industria l Resea rch Organ ization,Australia,h ttp://w w w.ts-im aging.net)to perform a d irectly f iltered back p ro jection a lgo rithm.Slice im ages w ere noise-redu ced and op tim ized to im p rove the d iscrim ination of d ifferen t substan ces and density region.From the ana lysis of gray va lues,SR-μCT data from a ll of the m icrospheres in the sam p les w ere exam ined.High ly reso lved tom ograph ic im ages of voids w ith in m icrospheres w ith h igh-qua lity phase con trastw ere then derived after 3D reconstruction s.

In to ta l,30 m icrospheres w ere random ly se lected from each kinds of m icrospheres,w h ich w ere 3D reconstru cted using Im age-p ro Plus(version 6.0,M ed ia Cybernetics,In c.,Bethesda,MD,USA),Am ira(version 5.4.3,Visage Im aging,In c.,Austra lia)and Im age Pro Ana lyzer 3D(version 7.0,M ed ia Cybernetics,In c.,Bethesda,MD,USA)to quan titatively characterize and visua lize the details of in terna l voids w ith in m icrospheres.

2.10. Statistical data ana lysis

Analysis of varian ce(ANOVA)w as perform ed using SPSS version 21.0 fo r w indow s(SPSS In c.,Ch icago,the Un ited States)for em p loying an erro r p robability at P≤0.05 for experim enta l data d ifferen ces.For each param eter,WCM and CCM w ere com pared w ith ACM by Studen t's t-Test.

3. Resu lts and d iscussion

3.1. Con tact angle

App rox im ate ly 20μl w ater o r citric acid so lu tion w ere deposited on the f ilm,there w as sign if ican t d ifferen ce in the quasi-equilibrium shape of d rops.The d rop of w ater had a sm a ller con tact ang le than that of citric acid so lu tion w ith con tact ang le of 71.94±1.20°,80.97±2.22°,respectively.The su rface of f ilm rep resen ting the su rface of ibup rofen m icrospheresw as hyd roph ilic due to the con tactangle less than 90°.It a lso ind icated that ibup rofen m icrospheres had a betterw ettability in w ater ow ing to low er con tactang le.Acco rd ing to the equation Yang,the greater con tact angle m eans larger in terfacial tension.A h igh in terfacial tension m ay generate m ore typ ica l spherica lm orpho logy and h igh po rosity in ibup rofen m icrospheres congea led in citric acid so lu tion.

3.2. Therm algrav im etric ana lysis

GMS show ed th ree-stage w eigh t loss,the f irst w eigh t loss at around 120-215°C relating to the w eigh t loss ofm oistu re,second w eigh t loss at the tem peratu re range of 220-325°C correlating to the decom position of ester bond of GMS and f ina l w eigh t loss at 350-420°Cm igh t be due to the carbon ization of in term ed iate p roducts.Ibup rofen and octadecano l perfo rm ed single-step w eigh t loss at around 120-228°C relating to the therm a l oxidation reaction.Fo rm icrospheres,the sing le-step m ass loss w as at the tem peratu re range of 120-226°C,120-258°C and 120-254°C fo r ACM,WCM and CCM,respectively.Overall,itw as in d icated that the therm aldecom position of all the sam p les occu rred above 120°C,w h ich p roved the stability of constitu tions ofm icrospheres at low er tem peratu re du ring the p reparation p rocess(Fig.1).

3.3. X-ray pow der diffraction

It w as ind icated that ibup rofen,octadecano l,and GMS w ere crysta ls(Fig.2).The characteristic peaks fo r ibup rofen w ere no rm a lized to invariab le peaks of octadecano land GMSw h ich w ere inso lub le in d isso lu tion m ed ia.The characteristic peak a t 2θof 16.66°fo r ibup rofen(dashed fram e area in Fig.2)w as selected for relatively qualitative analysis because the rest characteristic peaksw ere overlapped w ith that of octadecano land GMS.

The spectra of ibup rofen m icrospheresw ere a lso show n in Fig.2.The relative qualitative analysis show ed that ibup rofen w as still crysta l in octadecano l and GMS based m icrospheres.As dep icted in ou r p revious study of acetam inophen m icrospheres,the d rug crysta ls w ere d isso lved in d isso lu tion m ed ia,form ing porous structu ral arch itectu re[29].Therefore,it w as in ferred that ibup rofen crysta ls exposed on the su rface ofm icrospheres d isso lved in w ater,in f luen cing the stru ctu ra l arch itectu re of m icrospheres,w h ich w as verif ied by SR-μCT resu lts in the later sections.

3.4. Instan t dissolu tion to m im ic in vivo release

The instan t d isso lu tion tests w ith in 10m in w ere investigated to m im ic the in vivo release of ibup rofen ind icating the tastem ask ing e ffects to ibup rofen m icrospheres(Fig.3).The release behaviors of raw ibup rofen pow ders and m icrospheres w ere tested in d istilled w ater and hyd roch lo ric acid so lution(pH=1.2).It w as observed that a ll m icrospheres w ere slow ly released under the sink cond ition,less than 10%w ith in 10m in,w h ich cou ld be re lated to the poor so lubility of ibup rofen in correspond ing m ed ia.In phosphate bu ffer(pH=6.80 and 7.20)m ed ia as artif icia lsa liva,the raw ibup rofen pow ders w as re leased m o re than 90%w ith in 10m in,w h ich resu lted in to unp leasan t tastes.How ever,a ll the ibup rofen m icrospheres exh ibited sustained-release characteristics,nam ely 50%,38%and 32%w ith in 10m in at pH 6.80 and 58%,41%,and 35%w ith in 10m in at pH 7.20 for ACM,W CM and CCM,respectively.Com pared w ith ibup rofen raw pharm aceu tica lm ateria l,ibup rofen m icrospheresw ere capab le of retard ing the release of ibup rofen in sim u lated hum an sa liva,w h ich cou ld e ffectively m ask the un p leasan t taste.Especially at pH 7.20 m ed ia,CCM released 0.19%,4.56%,and 9.41%at 10,60,and 120 s,w h ich w ere lesser than ACM(0.75%,4.85%,and 13.01%)and W CM(1.12%,7.2%,and 12.99%).The resu lts illustrated that CCM had the best taste-m asking effect.

Fig.1-TGA characteristics show ed the stability of con stitu tion s ofm icrospheres.GMS:g lycerin m onostea rate;ACM:air-coo ling m icrospheres;W CM:w ater-coo ling m icrospheres;CCM:citric acid so lu tion m icrospheres.

Fig.2-XRD cha racteristics illustra ted the com ponen ts in m icrospheres.GMS:g lycerin m onostea rate;ACM:air-coo ling m icrospheres;W CM:w a ter-coo ling m icrosphe res;CCM:citric acid so lu tion-coo ling m icrospheres.

3.5. Scanning electron m icroscope

Du ring the d rug d issolu tion tests in phosphate bu ffer at pH 7.20,m icrospheres w ere co llected at 0,5 and 10m in fo r SEM detection.The m icrospheres w ere rep resen ted as nonreleased,partia lly-re leased and com p lete ly-re leased,respectively.There w as no obvious d ifferen ce in m orpho logy am ong the m icrospheres be fo re d isso lu tion.It w as a lso no ted that there w ere no sign if ican t changes in the arch itectu res ofm icrospheres du ring d isso lu tion test,ow ing to the fact that ibup rofen w as m elted and un iform ly d ispersed in the m atrix by f ine recrysta llization resu lting no large voids fo rm ation du ring d rug d isso lu tion(Fig.4).

3.6. Synchrotron radia tion X-ray m icro-com pu ted tom ography

3.6.1. Tw o-dim ension m onochrom e X-raym icro-com pu ted tom ography im ages

As dep icted in Fig.5,the large circu lar ring on the periphery w as the tube used for load ing m icrospheres.In the tube,size of the m icrospheres va ried a lo t ow ing to d ifferen t longitud ina l positions.The dark a reas in m icrosphere w ere som e sm a ll voids w ith in m icrospheres w h ich w ere believed to be air bubbles resu lted du ring coo ling p rocesses ofm icrospheres fo rm a tion.There w ere m o re voids fo r ACM because of contactw ith large quan tities of air in the coo ling m ed ium,w h ich w asm o re like ly tofo rm air voids du ringm icrospheres coo ling p rocess.

Fig.3-The in vitro d isso lu tion p rof iles of ibup rofen m icrospheres in d icated taste-m ask ing e ff icacy.●:ibup rofen;■:air-coo ling m icrospheres(ACM);◆:w ater-coo ling m icrospheres(W CM);▲:citric acid so lu tion-coo ling m icrospheres(CCM).

Fig.4-SEM characteristics dem onstrated the m orpho logy of ibup rofen m icrospheres at d ifferen t sam p ling tim es.(A,B,and C)are im ages of air-coo ling m icrospheres(ACM)at 0,5,10m in;(D,E,and F)a re im ages of w ater-coo ling m icrospheres(W CM)at 0,5,10m in;(G,H,and I)a re im ages of citric acid so lu tion-coo ling m icrospheres(CCM)at 0,5,10m in.

Fig.5-2D slice p ictu res of th ree k in ds of ibu p rofen m icrospheres.(A)air-coo ling m icrospheres(ACM);(B)w ater-coo ling m icrospheres(W CM);(c)citric acid so lu tion-coo ling m icrospheres(CCM).

Fig.6-3D im ages of agg regate cav ities of ibup rofen m icrospheres ex tracted a fter segm en tation and re lated ana lysis.(A)and(D)3D im age and cavity of sing le air-coo ling m icrosphere(ACM);(B)and(E)3D im age and cav ity of sing le w ater-coo lingm icrosphere(W CM);(C)and(F)3D im age and cavity of sing le citric acid so lu tion-coo ling m icrosphere(CCM).

A poor sphericity w as found fo r the ACM w ith rough su rface.Itw as p robab ly due to the longer freezing tim e fo rm icrospheres that w as coo led by air than W CM and CCM.The m icrospheres,thus,becam e m u ch deform ed that w as resu lted from the co llision of m icrospheres w ith each o ther o r w ith stain less stee l trays du ring coo ling p rocess.M eanw h ile,due to rap id coo ling in w ater o r citric acid so lu tion,better sphericity and sm oother su rface w ere observed for them icrospheres coo led by the co rrespond ing m ed ia.

3.6.2. Visualization of the internal3D structure

The reconstru cted 3D tom ograph ic im ages of ibup rofen m icrospheres dem onstrated that the 3D m orpho logy and d istribu tion of voids w ith in them icrospheres w ere h igh ly va riab le.As dep icted in Fig.6,the roughest su rface and the poo rest sphericity cou ld be seen in ACM.It w as a lso dem onstrated that there w as a larger void in the cen ter of m icrospheres su rrounded by m any sm a ller voids.The voids w ere sm a ll and heterogeneously d istribu ted w ith in W CM,bu tm u ch bigger in CCM.These resu lts w ere in acco rdan ce w ith those obtained from 2D m onoch rom e X-ray CT im ages.The su rface of CCM w as sm oother than them icrospheres coo led by w ater because so lubility of ibup rofen w as h igher in w ater than that in citric acid so lu tion.Th is led to parts of ibup rofen crysta ls getting exposed on the su rface ofm icrospheres and d issolved in to w ater,fo rm ing con cave stru ctu re on the su rface.

Herein,the in terna l 3D structu re ofm icrospheres w as fu rther investiga ted on the basis of structu re param eters,su ch as area of su rface,volum e,specif ic su rface area,d iam eter,box ratio,sphericity,rad ius ratio,Feret ratio,and porosity(Tab le 1),to deep ly understand the m echan ism of task-m asking.

The stru ctu ra l param eters ofm icrospheres w ere ana lyzed by Studen t's t-Test.Based on ANOVA,a ll th ree kinds of m icrospheres of ibup rofen w ere sign if ican tly d ifferen t(P＜0.05).W CM w as greatly d istingu ished(P＜0.001)from ACM for area of su rface,vo lum e,specif ic su rface area,d iam eter,Feret ratio and po rosity.And CCM w as greatly d istingu ished(P＜0.001)from ACM for sphericity,Feret ratio and porosity,w h ile all the stru ctu ra lparam eters con tribu ted to d istingu ish each ibup rofen m icrospheres.

Tab le 1-Stru ctu re pa ram eters of the ibup rofen m icrospheres ca lcu lated by th ree-d im en siona l recon stru ction techno logy.ACM:air-coo ling m icrospheres;W CM:w ate r-coo ling m icrospheres;CCM:citric acid so lu tion-coo ling m icrospheres.n=30,m ean±SD.

The po rosity here w as def ined as the p ropo rtion of voids w ith in m icrospheres and con fo rm ed in the o rder of ACM＞CCM＞W(wǎng) CM.Box ratio w as the ratio betw een the m axim um and the m in im um side length of the bounding box(box ratio value≥1)and con form ed in the order of ACM＞W(wǎng) CM＞CCM.Rad ius ratio referred to the ratio betw een the m axim um and the m in im um d istan ce betw een the centroid p ixel position and the perim eter of m icrospheres(rad ius ratio value≥1)w h ich w as con form ed in the order of ACM＞W(wǎng) CM＞CCM.The box ratio and the rad ius ratio cou ld re f lect the ex ten tof them icrospheres app roach ing the sphere:w hen the va lues close to 1,the m icrospheres are spherica l.In sum m ary,ow ing to longer freezing tim e and air in the coo lingm ed ium,irregu lar shape,rough su rface w ith num bers of con cave stru ctu re and a large num ber of voids w ere observed in ACM.Because of h igher in terfacia l tension resu lted in stab le in ternal structu re ofm icrospheres congealed in citric acid so lu tion,CCM had regu lar and roun der m o rpho logy,the sm oo thest su rface and little vo lum e of voids inside.Genera lly speak ing,m icrospheresw ith regu lar shape and sm oo ther su rface had sm aller su rface area asw ellas slow er d issolu tion rate than that in rough particles,thus to have taste-m ask ing e ffects.

These resu lts p rovided the structu ra l eviden ces to the m echan ism of d rug release p rof iles d iscussed above.It w as con c luded that due to the irregu lar shape and rough su rface w ith num bers of con cave stru ctu re,the ACM w erem o re like ly to rem ain in con tact w ith d isso lu tion m ed ia du ring the d isso lu tion p rocess.In con trast,CCM and W CM had the slowest release ow ing to the sm oother su rface,regu lar and round m orpho logy.How ever,the d rug re lease p rof iles w ere sim ilar for th ree kinds of m icrospheres in d istilled w ater and hyd roch lo ric acid so lu tion(pH=1.2)because of poo r so lubility of ibup rofen at tha t con d ition.Com pared w ith the stru ctu ra l arch itectu re ofm icrospheres,the so lubility of ibup rofen w as the m ain factor affecting the release behavior ofm icrospheres.Overa ll,there w as a strong corre lation betw een coo ling p rocesses,structu ra l arch itectu re and re lease behavio r of m icrospheres.

4. Con clu sion

The sam e fo rm u la of o ra ldosage fo rm sm ay have d ifferen t e ffects on taste-m asking due to the p roduction p rocess.W ith the app lication of SR-μCT,the 3Dm o rpho logy as w ellas in terna l arch itectu res of ibup rofen m icrospheres w ere accu rately quan tif ied for each p reparation m ethods.The effect of coo ling p rocesses on particle m o rpho logy and in terna l stru ctu re as w e ll as their con tribu tions to the taste-m ask ing w ere revea led in th is study.Thus,the structu re eva luation p rovides a new app roach for in ternal structu re based op tim ization and rationa l design fo r taste-m ask ing.

Con f licts of in terest

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

Acknow ledgm en ts

The au thors are gratefu l for the f inancial support from the Nationa l Natu ra l Scien ce Foundation of Ch ina(No.81773645)and Nationa l Scien ce and Techno logy M a jor Pro jects fo r the M a jo r New Drugs Innovation and Deve lopm en t(2017ZX09101001-006).