Muye He,Chen Zhong,Huiing Hu,Yu Jin ,Yanzuo Chen ,Kaiyan Lou,c,Feng Gaoa,,c,?

a Shanghai Key Laboratory ofFunctiona lMaterials Chem istry,East China University ofScience and Technology,Shanghai200237,China

b Departm ent of Pharm aceu tics,SchoolofPharm acy,East China University of Science and Technology,Shanghai200237,China

c ShanghaiKey Laboratory ofNew Drug Design,East China University ofScience and Technology,Shanghai200237,China

Keywords:β-cyclodex trin Ch itosan nanoparticles Ovalbum in Oral p rotein delivery In testina lm ucosal im m un ity

A B S T R A C TA nove l ora l p rotein de livery system w ith enhan ced in testina l penetration and im p roved an tigen stability based on ch itosan(CS)nanoparticles and an tigen-cyc lodex trin(CD)in clusion com p lex w as p repared by a p recip itation/coacervation m ethod.Ovalbum in(OVA)as a m odel an tigen w as f irstly en capsu lated by cyclodex trin,eitherβ-cyc lodex trin(β-CD)o r carboxym ethy l-hyd roxyp ropy l-β-cyclodex trin(CM-HP-β-CD)and fo rm ed OVA-CD in clusion com p lexes,w h ich w ere then loaded to ch itosan nanoparticles tofo rm OVA loadedβ-CD/CS or CM-HP-β-CD/CS nanopartic les w ith un iform particle size(836.3 and 779.2nm,respectively)and im p roved OVA load ing eff iciency(27.6%and 20.4%,respectively).In vitro d rug release stud iesm im ick ing o ra l delivery con d ition of OVA loaded CD/CSnanoparticles show ed low in itia l releases at pH 1.2 for 2 h less than 3.0%and a de layed release w h ich w as below to 30%at pH 6.8 fo r fu rther 72 h.M ore im portan tly,a fter o ra l adm in istration of OVA loadedβ-CD/CS nanoparticles to Ba lb/c m ice,OVA-specif ic sIgA leve ls in jejunum of OVA loadedβ-CD/CS nanoparticles w ere 3.6-fo ld and 1.9-fo ld h igher than that of OVA so lu tion and OVA loaded ch itosan nanoparticles,respec tively.In vivo eva luation resu lts show ed that OVA loaded CD/CSnanoparticles cou ld enhan ce its eff icacy fo r inducing in testina lm ucosa l im m une response.In conc lu sion,ou r data suggested that CD/CS nanoparticles cou ld serve as a p rom ising an tigen-delivery system fo r ora l vaccination.

1. Introduction

Imm un ization has been w idely used in c lin ic to cu re transm issib le d iseases.Com m ercia l p roteins are com m on ly delivered th rough in jection w ith poo r patien t com p lian ce.Ora l p rotein de livery p rovides a lternative w ay fo r im m un ization w ith good com p lian ce and e licits im m une responses at m ucosa l su rfaces of in testine,w here m em b ranous cells(M cells)ex ist fo r abou t 1 ou t of 10 m illion ep ithe lia l cells[1].Large quan tities of an tigen cou ld be possibly recogn ized by M cells non-specif ica lly,and then be tran sferred across tigh t ep ithelia l ce lls to reach b lym phocy tes(B ce lls)at Peyer's Pa tches.An tigen be up taken by Peyer's patches is an essen tia l step in oral vaccination,it can in du ce both local and system ic imm une response[2].As M cells exist on m u cosa l su rfaces of in testine,o ra l p rotein de livery is pa rticu larly su itab le fo r indu cingm ucosa l im m un ization.How ever,o ra lp rotein de livery faces m ajor lim itation of poten tial inactivation tow ard p roteins.Fo r exam p le,p ro teins can be degraded by enzym es in gastroin testina l tract,and are unstab le at low pH in stomach,w h ich leads to low er o ra l bioavailability and fu rther requ irem en t of repeated adm in istration of an tigen[2].Therefo re,an tigen p ro tection,m ain ly p ro tein p rotection,against inactivation facto rs such as low pH,enzym e,and bile sa lt,is the key to design o ra l p ro tein delivery system.In the past,po lym er nanoparticle-based delivery system m ade from po lym ethy lm ethacry late,po lyesters,po lyam ides,a lbum in,starch,ch itosan,dex tran,and etc.has show n to be p rom ising fo r ora l delivery of p roteins[3],w ith som e under consideration in clinica l tria ls in Eu rope and Un ited States[4].There is still in great need of nove l o ra lly adm in istered p rotein delivery system w ith im p roved an tigen p rotection and im m un ization e ff icien cy as w ell as safety[5].

One cha llenging p rob lem associated w ith p rotein delivery is to overcom e p ro tein inactivation inside hum an body[6].The possib le reasons of p rotein inactivation in clude con fo rm ational change,covalen t bon d break,aggregation or p recipita tion[7].Trad itiona lly,add itiona l excip ien ts su ch as su rfactan t w ere added to p reven t p rotein inactiva tion bu t causing adverse effect.Fo r exam p le,su rfactan t in creases the rate of aggregation and d rug oxidation and thus redu ces shelf-tim e[7].There fo re,nove l excip ien ts to p rotect p rotein from inactivation w ith few er o r no side e ffect are u rgen tly needed.In th is regards,bo th cyclodex trin(CD)and ch itosan(CS)have em erged as new po ten tia l and p rom ising excip ien ts fo r increasing p rotein therm ostability[8],inh ibiting p rotein aggregation[9],and stabilizing p rotein again st freeze-d rying p rocess[10].The stabilizing effect of cyc lodex trin can be rationa lized by its stru ctu re,w h ich consists of a hyd roph ilic ou ter su rface and a hyd rophobic cen tra l cavity.Pro tein residues from som e am ino acids such as pheny la lan ine,ty rosine,h istid ine,and tryp tophan can be en capsu lated in the cavity of cyclodextrin by non-cova len t in teraction s,thus p reven ting p ro tein aggregation w ithou t a ffecting its activity.Mo reover,research a lso show ed that cyclodex trin cou ld inh ibit the aggregation th rough retain ing the tertiary structu re of p rotein against denatu ration induced by chem ica l facto rs o r hea t.Other researches ind icated that cyc lodex trin fun ctions w ere sim ilar to non-ion ic su rfactan t fo r p reven ting p ro tein degeneration.In add ition,it has been dem onstrated thatβ-cyclodex trin(β-CD)can a lso im p rove d rug stability[11]and acqu ire sustained release p rof ile[12].For the above reasons,in th is study,w e use β-CD as the essen tialpart to establish in clusion com p lexes for o ra l p rotein delivery.

Besides from cyc lodex trin,ch itosan,a biodegradab le,biocom patib le natu ra l po lym er w ith good bio-adhesive p roperties,is also w idely used in pharm aceu tics[13-16].Ch itosan particles can im p rove the perm eability of m acrom o lecu les th rough the ep ithe lia l tigh t jun ction and has show n in severa l repo rts to be ab le to en capsu late an tigen and deliver to Peyer's Patches.For exam p le,Lubben[17]p repared ova lbum in(OVA)loaded ch itosan m icroparticles w ith h igher load ing eff icien cy by a p recip itation/coacervation m ethod,and the obtained ch itosan m icroparticles cou ld reach the Peyer's Patches acco rding to in vivo experim en t.In ano ther repo rt,Lubben[18]used ch itosan m icroparticles fo rm ucosa lvaccination against d iphtheria,and system ic as w ell as loca l im m une responses w ere found in Balb/c m ice.Moreover the d rug cou ld release from ch itosan particu lates in a con tro lled-release m anner[19,20].

There fo re,w e f irstly repo rted a novel ora l p ro tein delivery system based on nanoparticles constructed from cyclodextrin and ch itosan.OVA,a p rotein w ith 385 am ino acids and relative m o lecu lar m ass of 45 kDa[21],w as used as the m odel an tigen tofo rm OVA-CD in c lusion com p lexes f irstly by the stirring-freeze-d ried m ethod.The OVA-CD in c lusion com p lexes loaded ch itosan nanoparticles w ere p repared by a p recip itation/coacervation m ethod.OVA loaded nanoparticles w ere characterized by in vitro p roperty stud ies in clud ing particle size d istribu tion,en capsu lation eff icien cy,load ing e ff iciency and in vitro d rug release stud ies.OVA-specif ic secreto ry IgA(sIgA)levels in in testina lm ucosa l and IgG leve ls in serum w ere determ ined by enzym e-linked im m unoso rben t assay(ELISA)to study im m une response in Ba lb/cm ice.

2. M ateria ls and m ethods

2.1. Chem icals,reagents and anim als

Ch itosan(Mw=110 kDa,90.0%deacety lation degree)w as pu rchased from Yuhuan Ocean Biochem ica l Co.Ltd(Hangzhou,Ch ina).β-CD w as pu rchased from Sinopharm Chem ica l Reagen t Co.Ltd (Shanghai,Ch ina).Carboxym ethy l-βcyclodex trin(CM-β-CD)and hyd roxyp ropy l-β-cyc lodex trin(HP-β-CD)w ere pu rchased from Zh iyuan Biotech Co.Ltd.(Binzhou,Ch ina).Ova lbum in w as pu rchased from Sigm a A ld rich(M un ich,Germ any).The m icroBCA p ro tein assay k it w as supp lied by Beyo tim e Institu te of Bio techno logy Co.Ltd(Shanghai,Ch ina).M ouse IgG and s IgA ELISA Quan titation Kits w ere pu rchased from Elabscien ce Bio techno logy Co.Ltd(W uHan,Ch ina).A ll other reagen ts used in th is study w ere h igh perfo rm an ce liqu id ch rom atography(HPLC)o r reagen t grade.

Ba lb/c m ice(20±2 g)used in the experim en ts w ere supp lied by the Departm en t of Experim en tal An im als at Nan jing CAVENS bio techno logy Co.Ltd(Nan jing,Ch ina).The an im a l experim en ts w ere carried ou t in accordan ce w ith the gu idelines eva luated and app roved by the eth ics com m ittee of East Ch ina Un iversity of Science and Techno logy.

2.2. Syn thesis of carboxym ethyl-hydroxypropyl-βcyclodextrin(CM-HP-β-CD)

CM-HP-β-CD w as syn thesized as described in the re feren ce w ith m ino r m od if ica tions[22].Brie f ly,CM-β-CD(4.12 g)w as d isso lved in 100m l sod ium hyd roxide(3%,w/w)in a roundbo ttom ed f lask equ ipped w ith a constan t p ressu re funne l,a re f lux condensing tube,and a therm ograph.1,2-p ropy lene oxide(86.5m M)w as added slow ly in to the f lask over a period of abou t 2 h w hen reaction m ix tu re w as stirred in ice bath.A fter that,the reaction m ix tu re w as stirred at 25°C fo r 24 h,then neu tra lized to pH 7 w ith hyd roch lo ric acid fo llow ed by f iltration.The f iltrate w as then d ialyzed and freeze d ried to obtain CM-HP-β-CD.

2.3. Characterization ofCM-HP-β-CD

1H NMR spectra of the CM-β-CD and CM-HP-β-CD w ere recorded on a Bruker AVANCE 400 spectrom eter at 400MHz using D2O as the so lven t.The chem ica l sh ifts w ere repo rted in parts per m illion(ppm)and w ere referen ced to the residua l w ater signa l(δ=4.70 ppm).Fou rier transfo rm ed in frared(FT-IR)ana lysis w as perfo rm ed on a N ico let 6700 FT-IR spectrom eter.CM-β-CD and CM-HP-β-CD sam p les w ere p ressed w ith potassium brom ide in to a pellet before being p laced in the sam p le ce ll.

2.4. Preparation and characterization ofOVA-CD inclusion com p lexes

2.4.1. Preparation ofOVA-CD inclusion com plexes

OVA-CD in clusion com p lexes in d ifferen t stoich iom etric m olar ratio w ere p repared by an aqueous so lu tion-stirring m ethod and then d ried by lyoph ilization.Brie f ly,d ifferen t am oun t ofβ-CD,CM-β-CD,HP-β-CD or CM-HP-β-CD w as w eighed and d isso lved in phosphate bu ffered sa line(PBS),then the calcu lated am oun tofOVAw asm ixed thorough ly in to so lu tion underm agnetic stirring fo r certain tim e to estab lishing equ ilib rium.The OVA-CD so lu tion w as then freeze-d ried and reserved for fu rther study.

2.4.2. Binding strength of OVA-CD com p lexes:stability constant

UV spectra of OVA-CD in clusion com p lexes w ere ob tained using an UV/VIS spectroscopy(TU-1800SPC)in the range from 250 to 400 nm.The stability constan t(Kc)w as ca lcu lated accord ing to literatu re p rocedu re[23].The Kc w as ca lcu lated as fo llow s[24]:W hereΔA rep resen ts for the d ifferen ce of abso rban ce betw een in clusion com p lexes so lu tion and OVA so lu tion,[CD]rep resen ts fo r con cen tration of CD and[OVA]rep resen ts fo r con cen tration of OVA,Kc andΔεcan be ca lcu lated by non linear least-squares regression ana lysis.

2.4.3. Stoichiom etry determ ination ofOVA-CD inclusion complexes

The stoich iom etry of OVA-CD in clusion com p lexesw as determ ined by Job's p lot.OVA and cyclodextrinm ixed so lu tion samp les w ere p repared by keep ing the to ta l con cen tration of bo th com ponen ts constan t([OVA]+[CD]=10m M),w h ile changing the m olar ratio of cyc lodex trin to OVA as fo llow ing:4,6,8,10,12,20 and 40.The so lu tions w ere a llow ed stirring for 4 h at room tem peratu re,and UV abso rban ce at 280 nm w as m easu red fo r a ll sam p les and the UV abso rban ce d ifferen ces(ΔA=A-A0)betw een each sam p le(A)and reference OVA samp le(A0,[OVA]=10m M)w ere ca lcu lated and p lo tted versus the m o lar ratio of CD/OVA.The ratio corresponds to them ax im um va lue ofΔA w as regarded as the stoich iom etric ratio of the OVA-CD inclusion com p lex.

2.4.4. M olecu lar docking

The x-ray crysta l structu re of ova lbum in(PDB ID:1OVA)w as retrieved from Pro tein Data Bank(w w w.rcsb.org),w h ile the structu re ofβ-CD w as extracted from the crystal structu re of β-CD com p lex obtained from Protein Data Bank(PDB ID:3CGT).The w aterm o lecu les have been rem oved from the 1OVA.pdb crysta l stru ctu re p rior to dock ing study and hyd rogen atom s w ere added to ova lbum in crysta l stru ctu re.A single m o lecu le β-CD structu re w as geom etrica lly op tim ized using the sem iem p irica l PM 3 m ethod acco rd ing to lectu re[25].The op tim ized singleβ-CD structu re w as then used as the ligand in a ll dock ing p rocesses.M o lecu lar dock ing ofβ-CD to the insu lin m odel w as carried ou t using the Au toDock v4.2 softw are package[26,27].Gasteiger charges w ere com pu ted for each of the m olecu les.The grid m aps w ere p repared using a 40×40×40 grid box w ith the d istan ce betw een tw o grid poin ts set at 0.375?A cen tering on ova lbum in.A ll ro tationa l bonds inβ-CD w ere set free,w h ile those of ova lbum in w ere held rigid.The docking stud iesw ere carried ou tusing the em p irical free energy fun ction and app lying the standard p ro toco lof the Lam arck ian Genetic A lgo rithm(LGA).A to ta l of 100 in dependen t dock ing run sw ere carried ou t fo r each docking p rocesses w ith a m axim um num ber of 27 000 generations,a m u tation rate of 0.02,a crossover rate of 0.08 and an e litism va lue of 1.The f ina l low est docked energy con fo rm a tion is then used for the nex t run w ith anotherβ-CD and the dock ing stud ies w ere repeated un til them axim um num ber ofβ-CD yield a positive bin d ing free energy va lue.

2.4.5. SEM images ofOVA-CD inclusion com plexes

M o rpho logica l ana lysis of OVA-CD in clusion com p lexes w as perform ed by JEM-6360 LV scann ing electron m icroscope(SEM)(JASCO,Japan).Sam p les w ere w eigh ted and p laced on b rass p late,and observed a fter p lated w ith a th in layer of go ld.

2.5. Preparation ofOVA loaded CD/CS nanoparticles

Ch itosan nanoparticles w ere p repared by a p recip itation/coacervation m ethod as described in the re feren ce w ith m inor m od if ications[18].Brief ly,to a 0.25%(w/v)ch itosan glacia l acetic acid so lu tion,Tw een-80 w as added to a f ina l con cen tration of 1%.Then,sod ium su lfate so lu tion(10%,w/v,0.2m l)w as added d ropw ise to the ch itosan so lu tion w ith m agnetic stirring and u ltrason ic for 30 m in toform ch itosan nanoparticles.

OVA loaded ch itosan nanoparticles and CD/CS nanoparticles w ere p repa red by the fo llow ing p rocedu re.Brief ly,ch itosan nanoparticles w ere resuspended in PBS so lu tion,then OVA or OVA-CD(β-CD or CM-HP-β-CD)in clusion com p lexes w as added in to the so lu tion at 25°Cw ith stirring at 50 rpm for 3 h.The nanopartic le so lu tion w as freeze-d ried to obtain d ry nanoparticles.

2.6. M orphologica l characterization ofOVA loaded CD/CS nanoparticles

The particle size and zeta po ten tia l of ch itosan nanoparticles in aqueous solu tion w ere m easu red using Delsa Nano dynam ic ligh t sca ttering(DLS)(Beckm an Cou lter In c,USA)equ ipped w ith 4m W He-Ne laser at a w ave length of 633 nm at room tem peratu re.The m o rpho logica l eva luation of OVA loaded CM-HP-β-CD/CS nanoparticles w as perform ed w ith transm ission e lectron m icroscope(TEM).TEM im ages w ere taken on JEM-2010(JEOL,Japan).The sam p le w as stained w ith 2%(w/v)phospho tungstic acid be fo re d ropped on a copper grid.

2.7. En trapm en t efficiency,loading efficiency and stability study

To determ ine OVA en trapm en t and load ing e ff icien cy,OVA loaded CD/CS nanoparticles w ere cen trifuged a fter p reparation and the supernatan tw as co llected.The con cen tration of OVA in supernatan t w as determ ined by m icroBCA acco rd ing to supp lier's instructions.The en trapm en t e ff icien cy(EE)and load ing eff icien cy(LE)w ere calcu lated as follow s:En trapm en t efficien cy(%)

The ch itosan nanoparticles,OVA loaded CS nanoparticles and OVA loaded CD/CS nanopa rticles w ere fresh ly p repared at pH 6.0 w ith op tim ized reaction cond ition and p reserved at 4°C.Sam p lesw ere co llected at d ifferen t tim e poin ts up to one m on th.The size d istribu tion of nanopa rticles w asm easu red by DLS.The stability of ch itosan nanoparticles and OVA loaded CSnanoparticlesw ere a lso investigated at pH 1.2 at room temperatu re for 2 h.

2.8. In vitro drug release ofOVA loaded CD/CS nanoparticles

The in vitro d rug release from the OVA loaded CS nanoparticles and OVA loaded CD/CS nanopartic lesw ere stud ied in solu tion sequen tia lly at pH 1.2 and pH 6.8 to sim u la te the pH environm en t in the fasting stom ach and then the sm a ll intestine.Brie f ly,the p repared nanoparticles w ere resuspended in solu tion at a con cen tration of 0.1%(w/v),then the so lu tion w as in cubated in 37°C at 100 rpm.A fter shaken fo r 0.5,1 and 2 h at pH 1.2,and then 0.5,1,2,4,8,12,24,48 and 72 h at pH 6.8 in the vessel,a 200μla liquotw as sam p led and ref illed w ith an equal vo lum e of fresh m edium correspond ingly.Each aliquot sam p le w as then cen trifuged(10 000 rpm fo r 5m in)in a p lastic tube to sepa rate the supernatan t for fu rther determ ination of released OVA con cen tration in supernatan t by m icroBCA assay.

2.9. In vivo imm unization experim en ts and determ ination ofserum IgG and siga

Tw en ty m a le Ba lb/c m ice w ere random ly d ivided in tof ive groups,w h ich w ere treated w ith OVA so lu tion,OVA loaded CS nanoparticle,OVA loadedβ-CD/CS nanoparticle,OVA loaded CM-HP-β-CD/CSnanoparticle and sa line as con tro lgroup.OVA con tain ing so lu tions w ere p repared by d isso lving the OVA o r OVA loaded nanoparticles in PBS(w/v,200μg OVA-equ iva len t)and w ere adm in istered after fasting for 24 h to m ice orally(0.2m l/m ouse)for im m un ization[28].Each m ouse received im m un iza tion on day 0 and day 7 w ith d ifferen t OVA form ulations show n above.The im m un ization dosage on day 7 w as doub led to ensu re positive im m une response.Blood sam p les(200μl)w ere taken from the o rbita lsinus on day 14 and day 21,and the serum w as iso lated by cen trifugation(10 000 rpm,4°C)and sto red at-20°C fo r IgG leve lana lysis.Them ice w ere then sacrif iced after the blood sam p les w ere taken on day 21,then 5 cm of je junum or ileum w as cu t as sam p les,supernatan tw as co llected to test the OVA-specif ic sIgA by tissue hom ogenation,cen trifugation(10 000 rpm,4°C)and sto red at-20°C until s IgA levels in in testine w as ana lyzed.The OVA-specif ic IgG in serum and OVA-specif ic s IgA w ere exam ined by ELISA acco rd ing to supp lier's instructions.

2.10. Sta tisticalanalysis

M u ltip le group com parisons w ere conducted using one-w ay analysis of variance(ANOVA).A ll data analysis w as execu ted using the IBM SPSSStatistics 17.0.A lldata w ere p resen ted as a m ean va lue w ith its standard deviation ind icated(m ean±SD).P-va lues less than 0.05 w ere considered to be statistica lly sign if ican t.

3. Resu lts and d iscussion

3.1. Syn thesis and characterization ofCM-HP-β-CD

Fig.1A included the1H NMR spectrum and the com p lete peak assignm en ts fo r the CM-β-CD.The so lven t peak of D2O w as found at 4.79 ppm.The peaks at 5.07 ppm and 3.40-4.30 ppm w ere assigned to the p rotons of carbon1-6 inβ-CD segm en ts,respective ly.The peaks at 5.28 ppm w ere attribu ted to-CH2-of the carboxym ethy l(CM)group in CM-β-CD.Compared w ith CM-β-CD,characteristic peak of CM-HP-β-CD atδ(ppm)=1.13-1.15(m ethy l group(-CH3)in hyd roxyp ropy l(HP))(Fig.1B)w as observed,suggested that HPhad been in trodu ced to CM-β-CD.It cou ld be ca lcu lated from1H NMR in tegration that average m o lar substitu tion of HP and CM w as 2.9 and 5.7,respectively.The FTIR spectra of CM-β-CD and CM-HPβ-CD are show n in Fig.1C.The strong absorp tion peaks at 1594 cm-1,1419 cm-1,and 1328 cm-1w ere from ether group,and absorp tion at 3440 cm-1w as due to the a lcoho lic hyd roxy l group in the CM-β-CD.In the IR spectrum of CM-HP-β-CD,the in tensity of C-H stretch band at 2929/cm-1in creased,w h ich cou ld be attribu ted to the m ethy lene group in the CM side chain.Eviden ces from1H NMR and IR spectra dem onstrated the successfu l syn thesis of CM-HP-β-CD.

Fig.1-The 1H NMR spectrum of CM-β-CD(A)and CM-HP-β-CD(B)and FT-IR spectrum of CM-β-CD and CM-HP-β-CD(C).

3.2. Characterization ofOVA-CD inclusion com p lexes

The OVA-CD in clusion com p lexes w ere p repared by the stirred-freeze-d ryingm ethod.UV spectra(250 to 400 nm)w ere m easu red at d ifferen tm o la r ratio of OVA to cyclodex trin(Fig.2).The decrease of the ratio OVA/cyc lodex trin w as accompan ied w ith in crease of UV absorban ce at aroun d 280 nm,suggesting that in creased am oun t of in clusion com p lex w as form ed betw een arom atic am ino acid residues of OVA and cyclodex trin m o lecu les.The stability constan ts(Kcva lues)of OVA-CD in clusion com p lexes are sum m arized in Tab le 1.Kcva lues of CM-β-CD and HP-β-CD cou ld no t be ob tained by th is m ethod due to too sm allUV absorban ce d ifferen ces at various OVA to cyc lodex trin ratios(Fig.2Band C),w h ich suggested that CM-β-CD and HP-β-CD have low bin d ing e ff icien cy w ith OVA.Asβ-CD and CM-HP-β-CD bound to OVA m ore strong ly,they w ere selected to p repare OVA-CD in clusion com p lexes for fu rther stud ies.

The stoich iom etry of com p lex form ation betw een OVA and eitherβ-CD or CM-HP-β-CD w as a lso determ ined.As show n in Fig.3A and B,w hen the m o lar ratio of cyclodextrin to OVA reached 10,it reached the m ax im um absorban ce d ifferen ce,w h ich ind icates form ation of 1:10 in clusion com p lexes betw een OVA andβ-CD o r CM-HP-β-CD.The m o lecu lar dock ing study w as perfo rm ed and the 3D stru ctu re of the OVA-β-CD form ation w ith bind ing sites is show n in Fig.3C.The in teraction betw een con fo rm ation of OVA andβ-CD at bind ing sites w as ana lyzed by PyMOL.The docking resu lts ind icate that the 1:9 OVA/β-CD fo rm ation p roduces the m ost stab le con fo rm ation in w h ichβ-CD form hyd rophobic in teractionsw ith a total of 9 am ino acids that are Asn86,Lys89,Lys186,Glu196,Glm 226 and Trp274 on chain A,Lys58,Glu89 and Ty r94 on chain B.Th is resu lt is in agreem en tw ith the resu lt of stoich iom etry of comp lex form ation.

Tab le 1-The resu lts of Kc betw een OVA and cy clodex trin in the PBS bu ffer.

Fig.2-UV-VIS spectra of OVA andβ-CD(A),CM-β-CD(B),HP-β-CD(C)and CM-HP-β-CD(D)a t d ifferen tm o lar ratio of OVA/cy clodex trin.

Fig.3-Changes of the d ifferen ce in the abso rban ce as the ratio ofβ-CD/OVA(A)and CM-HP-β-CD/OVA(B).The 3D stru ctu re of OVA-β-CD in clusion com p lex fo rm ation that in d icates n ineβ-CD m o lecu le(red stru ctu re)b ind ing on one OVA m o lecu le(green stru ctu re)(C).

The SEM im ages show ed m o rpho logy of OVA,β-CD,CM-HPβ-CD,OVA-β-CD in c lusion com p lex,and OVA-CM-HP-β-CD inc lusion com p lex(Fig.4).SEM im ages OVA-CD in clusion comp lexes(Fig.4D and E)p resen ted both characteristics of OVA(Fig.4C)and the cyclodex trin that they w ere derived(Fig.4A and B),suggesting OVA w as successfu lly en capsu lated.

In terestingly,UV/Vis spectra of in clusion com p lexes(Fig.2)show ed the m axim um absorp tion peak of OVA-CD in clusion com p lexes around 280 nm red-sh ifted sligh tly w ith in creasing am oun t of cyclodex trin.Th is can be con cluded that the in creasing am oun t of cyc lodex trin coup ling on the su rface of OVA,lead ing to an in creased sh ie ld ing effect on the am ino acid residue of the p ro tein[29].

3.3. Preparation and characterization ofCD/CS nanoparticles

Fig.4-The SEM pho tog raphs ofβ-CD(A),CM-HP-β-CD(B),OVA(C),OVA-β-CD in clusion com p lex(D)and OVA-CM-HP-β-CD in clusion com p lex(E).

Fig.5-The possib le fo rm u lation m echan ism of OVA loadedβ-CD/CS nanopa rticles.

Severa l series of ch itosan nanoparticles w ere p repared at d ifferen t con d itions by a p recip itation/coacervation m ethod and op tim ized by single-factor m ethod.Encapsu lation eff icien cy(EE)and load ing eff icien cy(LE)w ere then investigated by m icroBCA.The op tim ized load ing e ff icien cy of OVA loaded nanoparticles of CS,β-CD/CS,and CM-HP-β-CD/CS w ere 30.9%,27.6%and 20.4%,respectively,and the correspond ing particle sizesw ere 213.6,836.3 and 779.2 nm,respectively.Com pared w ith the OVA-loaded ch itosan nanoparticles,the size of OVA-loaded CD/CS nanopartic les in creased sign if ican tly.It can be specu lated that du ring the p reparation p rocess,som e of the OVA com p lex bound to the ou ter su rface of particles,and particles aggregated together th rough p rotein and form a stab le stru ctu re of nanopartic les.The possib le m echan ism in form ation of OVA loaded CD/CS nanoparticles w as illustrated in Fig.5.The TEM im age in Fig.6A show ed the spherica land un ifo rm m o rpho logy of the OVA loadedβ-CD/CS nanoparticles.

Fig.6-TEM pho tog raph of the OVA loadedβ-CD/CS nanopa rticles(A).Size changes in CD/CS nanopa rticles du ring one m on th-sto rage at pH 6.0 at 4°C in PBS(B)(n=3).

Fig.7-In v itro re lease of OVA from OVA loaded CS nanopa rticles and OVA loaded CD/CS nanopa rticles in PBS at p H 1.2(A)and at p H 6.8(B)at 37°C(n=3).

3.4. Stability study

in the partic le size of OVA loadedβ-CD/CS nanoparticles a fter 20 d w as accep tab le,w h ich m ay be caused by sh ield ing e ffect ofβ-CD lead ing to low er electrostatic fo rce betw een am ino acid residue and ch itosan.M oreover,w e test the residual percen t of OVA in CD/CS nanopartic les,93.7%of OVA w as retained in nanoparticles on day 30.The resu lts suggested that OVA loaded ch itosan nanoparticles w ere m o re stab le than ch itosan nanopa rticles w ithou t OVA in acid ic m ed ium,likely due to stabilization effect of p ro tein corona th rough OVA coated on the su rface of nanoparticles by e lectrostatic in teraction,w h ich can p rotect ch itosan from degradation by gastric acid.

3.5. In vitro drug release ofOVA loaded CD/CS nanoparticles

The in vitro d rug release p rof iles ofOVA loaded CD/CSnanopartic les w ere investigated in so lu tion at pH 1.2 fo r 2 h and pH 6.8 fo r 72 h sequen tia lly at 37°C to sim u late the pH environm en t of the fasting stom ach and then the sm a ll in testine

Fig.8-Schem atic d iagram ofm ice im m un ization w ith OVA loaded CD/CS nanoparticles(A).The level of OVA-specif ic s IgA in je junum(B)and ileum(C)m u cosa 21 d,and m ou se OVA-specif ic IgG in serum 14 d(D)and 21 d(E),a fter the f irst tim e o ra lly given OVA-loaded nanopa rticles(n=4).?P＜0.05,??P＜0.01,???P＜0.001 com pa red w ith OVA so lu tion;#P＜0.05,##P＜0.01,###P＜0.001 com pared w ith OVA loaded CS nanoparticles.

The tim e-dependen t stability of ch itosan nanopa rticles,OVA loaded CS nanoparticles,OVA loadedβ-CD/CS nanoparticles and OVA loaded CM-HP-β-CD/CS nanoparticles w as investigated at 4°C,and partic le size w asm easu red in specif ic tim e poin ts by DLS ana lysis as show n in Fig.6B.No sign if ican t partic le size changes of ch itosan nanoparticles,OVA loaded CS nanoparticles and OVA loaded CM-HP-β-CD/CS nanoparticles w ere observed up to 30 d,ind ica ting that a ll these nanopartic les w ere stab le w ith in 30 d.The sm a ll am oun t of in crease(Fig.7).Bo th OVA loaded CS nanopartic les and OVA loaded CD/CS nanopartic les had sim ilar sustained release p rof ile in bo th pH environm en t,w h ich show ed an in itia l bu rst in the f irst hou r like ly due to deso rp tion of the OVA on the su rface of ch itosan nanoparticles,fo llow ed by a delayed release.The total am oun t of OVA released from the OVA loaded CS nanoparticles and OVA loaded CD/CS nanopartic les w ere less than 3%w ith in 2 h at pH 1.2,ind icating thatm ost of OVA p rotein m o lecu les w ere loaded deep inside nanoparticles that can keep OVA su rvived in the in itia l 2 h in the fasting stomach in vivo.Therefore,the relative low in itial bu rst release at pH 1.2 is a favored p roperty fo r ora l im m un ization.In the fo llow ing 72 h at pH 6.8,the am oun t of OVA re leased from the OVA loaded CS nanopa rticles,OVA loadedβ-CD/CS nanoparticles and OVA loaded CM-HP-β-CD/CS nanopa rtic les w as 31.5%,21.4%and 17.3%,respective ly,ind icating that in co rpo ration of cyclodex trin in nanoparticles eff icien tly slow ed dow n the rate of OVA release.Th is phenom enon cou ld be rationa lized by stronger en capsu lation of OVA inside nanopartic les th rough fo rm ation of OVA-CD in c lusion com p lexes.

3.6. In vivo imm une responses

The OVA-loaded nanopartic les w ere adm in istered orally to Ba lb/cm ice(Fig.8A).OVA-specif ic IgG leve ls in serum and sIgA levels in in testina l w ere determ ined by ELISA.As show n in Fig.8B and C,com pared w ith them ice group treated w ith OVA so lu tion,the OVA loaded CD/CS nanoparticle and OVA loaded CS nanopartic le m ice groups had in creased OVA-specif ic IgG level in serum a fter o ra l adm in istra tion fo r 14 and 21 d The OVA-specif ic IgG levels in creased m ore sign if ican tly on day 14 w ith abou t 1.3-fo ld,1.7-fo ld and 1.9-fo ld in crease for OVA loaded CS,β-CD/CS,and CM-HP-β-CD/CSnanopartic le treated m ice group,respectively.The resu lts ind icated that ch itosanderived nanopartic les cou ld enhan ce the bioavailability of OVA at sm a ll in testine,likely caused by p ro tection ofOVA from inactivation due to low pH and en zym es in the gastroin testina l tract.The resu lts w ere a lso in agreem en t w ith literatu re[30,31].The elevated levelof IgGw as also observed for the OVA loaded CS and CD/CS nanoparticle treated groups com pared w ith the OVA so lu tion group at statistica lly sign if ican t leve l on day 21(P＜0.05).

In Fig.8D and E,elevated OVA-specif ic slgA levelw as alsofound in OVA-loaded nanoparticle groups.Com pared w ith the OVA so lu tion group,OVA-specif ic s IgA leve l in je junum of the OVA-loaded CS,β-CD/CS,and CM-HP-β-CD/CS nanoparticle treated groups in creased abou t 1.9-fo ld,3.6-fo ld and 3.0-fo ld on day 21,respective ly.The co rrespond ing va lues in ileum on day 21 w ere 1.4-fo ld,2.3-fo ld and 2.1-fo ld,respectively,w h ich are in accordan ce w ith p revious research[31].Moreover,the h igher levelof OVA-specif ic sIgA observed in je junum than in ileum w as likely caused by the phagocy tosis of Mce llsm ain ly occu rred in jejunum and thus p rodu ced a h igher level of OVA-specif ic s IgA[32].Besides,in creased sIgA level in groups of m ice trea ted w ith nanoparticles w ith cyclodex trin in both jejunum and ileum clearly show s the synergetic e ffect of cyc lodex trin and ch itosan.It is repo rted that cyclodextrins have no effect on the oraladm in istration of p roteins and neither have any p rom ising effects as ad juvan ts[33].Here,cyclodex trin p robab ly fun ction as stabilizer to p ro tect the loaded p rotein from acid ic and en zym atic degradation in the gastroin testinal tract th rough form ation of in clusion com p lex.The resu lts show ed OVA loaded CD/CS nanoparticles exh ibited a m o re de layed OVA releasing than OVA loaded ch itosan nanoparticles,w h ich m igh t help to m ain tain OVA h igher concen tration in bo th je junum and ileum,th is p robab ly attribu ted to the ch itosan nanoparticles m atrix erosion o r degradation f irstly,and the second re lease from cyclodex trin in c lusion com p lexes a fter OVA loaded CD/CS nanopartic le tran sferred across tigh t ep ithelia l cells to reach B ce lls at Peyer's Patches and indu ce bo th loca l and system ic im m une response[32].Overa ll,bo th the stabilization e ffect and the slow er release p rof ile due to cyclodextrin in corporation w ere responsible the in creased OVA bioavailability foun d in OVA loaded CD/CS nanoparticle treated m ice.M o reover,CD/CS nanopartic le has p reviously repo rted to be ab le to eff icien tly p ro tect OVA and retain the p rotein's activity from gastric acid in stom ach,facilitating the p ro tein abso rp tion th rough nanopartic le decom position a fter reach ing the Peyer's patches[34].The released OVA cou ld reach lym phoid tissue to induce imm une response[35,36].Besides,cyc lodex trin itselfm igh t trigger innate im m une response at je junum and ileum as the im m un ization ad juvan t[37].Overa ll,CD/CS nanopartic les repo rted herein show ed great poten tials as a novel oral p rotein delivery system to indu cem u cosa land system ic im m une response in vivo.

4. Con clu sion

A nove l ora l p ro tein de livery system,OVA-loaded CD/CS nanoparticles,w as successfu lly p repared by a p recip itation/coacervation m ethod w ith enhan ced o ra l abso rp tion,penetration and im p rovem en t of OVA stability from cyclodextrin and ch itosan.The fo rm ation of OVA-CD in clusion comp lexes w as verif ied by m o lecu lar docking.OVA loaded CD/CS nanoparticles had un iform particle size and exh ibited h igher OVA en capsu lation eff icien cy and slow er d rug re lease rate in vitro.Mo re im portan tly,in vivo eva luation resu lts show ed OVA loaded CD/CS nanoparticles cou ld enhan ce its eff icacy in inducing in testinalm ucosal imm une response.Thus,w e consider that CD/CS nanoparticles can be used as a p rom ising o ra l an tigen delivery system to im p rove the im m unogen icity of o ra l p ro tein delivery.

Con f licts of in terest

The au thors repo rt no con f licts of in terest.

Acknow ledgem en t

Th is w o rk w as suppo rted by Scien ce and Techno logy Comm ission of Shanghai M un icipa lity(No.17ZR1406600)and Nationa l Scien ce Foun da tion of Ch ina(No.21577037).Th is w o rk w as a lso sponsored by Scien ce and Techno logy Comm ission of Shanghai Mun icipality(No.10DZ2220500 and No.11DZ2260600).