Chenyao Zhao,Chan Jin,Hailing Gao,Liyuan Wang,Hongzhuo Liu,Zhonggui He

Shenyang Pharm aceu tical University,Shenyang 110016,China

Keywords:Glip izide Particle size Produc t qua lity Bioequ iva len ce study Disso lu tion

A B S T R A C TThe objective of th is study w as to understand the im pact of active pharm aceu tical ingred ien ts(API)particle size on a re-deve loped generic p roduct of g lip izide and to im p rove its fo rm u lation so that it exh ibits bioequ iva len t to that of the referen ce listed d rug(RLD).Tw o com m ercia l batches of APIs(API-1 and API-2)w ith the sam e po lym orph ism and one batch of hom e-m ade APIs(API-3)w ith super-sm a ll particle size w ere used in the p resen t study.The in vitro d isso lu tion p rof iles of the tested form u lation s w ere com pared w ith the RLD in a series of d isso lu tion m ed ia.Then,the im pact of partic le size on in vivo abso rp tion w as evaluated in Beag le dogs.Com pared w ith the RLD,form u lation A w ith larger APIsize show ed slow er d isso lu tion in pH 6.0 and 7.4m ed ium,resu lting bioinequ ivalen t w ith the RLD.Conversely,form u lation Bw ith sm a ller APIsize dem on strated sim ilar in vitro d isso lu tion p rof iles w ith the RLD and thus exh ibited bioequ ivalen t in the p resen t study.Fu rtherm o re,form u lation Cw ith su per sm a ll particle size still exh ibited iden tica l o ra l abso rp tion a lthough rap id d isso lu tion w as observed in the tested con d ition.Herein,it in d icated that 2-5μm m igh t be def ined as the“inert size range”of glip izide for en su ring the bioequ iva len ce w ith the RLD.The resu lts in the p resen t study m igh t help to ob tain a better un derstand ing of the variability in raw m aterials for oral absorp tion,develop a bioequ ivalen t p roduct and thus post-m arket qua lity con tro l.

1. Introduction s

Glip izide,an oral an tidiabetic d rug of the second-generation su lfony lu rea,is used to con tro l b lood glu cose in patien ts w ith non-insu lin-depen den t d iabetesm ellitus[1,2].It p rom o tes insu lin re lease via b lock ing ATP sensitive po tassium channel and thus reducing b lood g lu cose levels[2].Genera lly,g lip izide can be classif ied as a typically Biopharm aceu tical Classif ication System(BCS)II class d rug w ith w eak ly acid ic character(p Ka5.9)[3,4].Sim ilar w ith o ther class II d rugs,the so lubility of glip izide in gastroin testina l tract lim its the d isso lu tion,and thus o ra labso rp tion and bioavailability in vivo[5].Previous stud ies dem onstrated that several strategies in clud ing particle size redu ction[4],so lid d ispersion and cyclodex trin comp lexation[6]facilitated to im p rove the in vitro d isso lu tion and thus ora l bioavailability of g lip izide in vivo.

Clin ical stud ies ind icated that the absorp tion of glip izide in vivo w as fo rm u lation dependen t.No tab ly,the fo rm u lation p repared w ith non-m icron ized active pharm aceu tica l ingred ien ts(APIs)led to slow and in com p lete abso rp tion w ith large in ter-ind ividual variation[7].Sim ilar variation in oral absorption stem m ed from the particle size of APIsw as a lso observed fo r glybu ride,ano ther second-generation su lfony lu rea,h ighligh ting the key ro le of partic le size d istribu tion in the developm en t of a su itab le p rodu ct fo r su lfony lu rea[8].

In recen t years,d rug developers and m anu factu rers are encou raged to u tilize the ach ieved p rodu ct and p rocess understand ing to add ress the sou rces of variation to p roductquality and to obtain the app rop riate con tro lstrategies to iden tify the risk area[9].Considering the po ten tia l raw m ateria l variability due to the change in either m anu factu ring rou te or their p rodu ction sites,the relationsh ip betw een m aterial variability and p roduct qua lity needs to understand[10].A lthough p revious stud ies ind icated the critica l ro le of partic le size in the qua lity attribu te of g lip izide,few stud ies investigated the im pact of APIs variability upon the f inal p roduct quality[11].Specia lly,the particle size of APIs p rovided by even the sam e m anu factu rer bu t varied batchesm igh t resu lt in a qua lity attribu te[12].Herein,the p resen t study aim ed to understand the particle size of glip izide im pacting in vitro d isso lu tion and then in vivo ora l abso rp tion.Acco rd ingly,w e designed th ree iden tica l fo rm u lations excep t fo r the particle size of glip izide.Sin ce the d isso lu tion test has been w ide ly used to serve as an ind icato r to p red ict the qua lity of p rodu cts,ou r study a lso aim ed to eva luate the pow er of in vitro d isso lu tion test in p red icting in vivo perfo rm an ce of fo rm u lations.Such understanding is in favo r of im p roving qua lity con tro l in the p rodu ction changes.

2. M ateria ls and methods

2.1. M aterials

Glip izide tab lets chosen as the re feren ce(con tain ing 5 m g g lip izide in to ta l w eigh t of 200 m g,RLD)w ere pu rchased from Pf izer Co.,Ltd.(Da lian,Ch ina).Tw o ba tches of g lip izide w ere obtained from W eihai Disu Pharm aceu tica ls Co.,Ltd.(Shandong,Ch ina).Referen ce standards of g lip izide(99.5%pu rity)and g liclazide used as in terna l standard w ere obtained from the National Institu te for Con tro l of Pharm aceu tica l and Bio logica l Produ cts(Beijing,Ch ina).Lactose m onohyd rate w as pu rchased from Jiangsu DaoN ing Pharm aceu tica l Co.,Ltd.(Ch ina).M icrocrysta lline cellu lose,starch and colloidal silicon d ioxide w ere gifted by Anhu i Sunhere Pharm aceu tica l Excip ien ts Co.,Ltd.(Ch ina).Stearic acid w as p rovided by Huzhou Zhan W ang Pharm aceu tica l Co.,Ltd.(Ch ina).M ethano l,aceton itrile and form ic acid,a llw ith HPLC grade w ere p rocu red from Fisher(USA).Am m on ium acetate(HPLC grade)w as pu rchased from Dikm a(Richm ond Hill,NY,USA).Deion ized-d istilled w ater w as used th roughou t the study.

2.2. Preparation ofglipizide pow erw ith reduced particle size

A PM p lanetary ballm ill(Nan jing Ch ishun Science&Technology Co.,Ltd.,Nan jing,Ch ina)w as used to p repare glip izide pow er w ith fu rther redu ced particle size.Zircon ium d iox ide beads(0.4-0.5m m of d iam eter)w ere served as them illingm ed ia.The coarse suspen sion(3%glip izide,w/v)w as d ispersed in w ater and the m ix tu re then w as d irectly transferred to the p lanetary ba llm illw ith the equa lvo lum e of beads.Them illing w as perform ed at a rotation speed of 35Hz for 2 h.The resu lting suspen sion sw ere then lyoph ilized using FDU-1100 freezed rier(EYELA,Tokyo Rikak ikai Co.,Ltd,Japan).Brief ly,lactose(10%,w/v)w ere added in to the fresh ly p repared suspensions as lyop ro tectan ts.

2.3. Particle size analysis

The particle d istribu tion of the APIs from the m anu factu rers w ere m easu red d irectly by the laser d iffraction beam as an aeroso lized d ry pow der w ith the RODOS d ry pow der accesso ry(Sym patec HELOS Com pact,R1 he lium-neon laser Gm bH,W indox Softw are 5.0,Claustha l-Zellerfeld,Germ any).The m easu rem en t w as conducted at air p ressu re of 3.5 bar.On the o ther hand,the particle size and po lyd ispersity index of hom e-m ade glip izide pow er w ere m easu red by ZS-90 nanoparticle size ana lyzer(Ma lvern Instrum en ts Ltd.,UK).The pow er w as d ispersed w ith d istilled w ater and each samp le w as determ ined in trip licate.

2.4. Scann ing electron m icroscopy(SEM)

The m o rpho logy of the APIs w as observed by SEM(Hitach i SU8010,Hitach i LTD.,Tokyo,Japan).Brie f ly,the APIs w ere un ifo rm ly d istribu ted on the su rface of insu lated doub le-sided carbon tape and the analysis w as carried ou t at an acceleration vo ltage of 10 KV.Spu tter coating w as not needed.M icrographs w ere recorded at the requ ired m agn if ication.

2.5. X-Ray pow der diffraction(XRPD)

XRPD patterns of the sam p les w ere determ ined using a DMAX-2400 Pow der X-ray Diffractom eter(Rigaku,Japan).The instrum en tal cond itions w ere set as follow s:Cu Kαsou rce at a generator pow er of 40 kV and 30m A;d ivergen t beam(2m m);scann ing range set from 0 to 60°.

2.6. Form u lations

Th ree tested fo rm u lations of g lip izide w ere used fo r in vitro d isso lu tion testing and in vivo bioequ iva len ce study.The re ference form u lation of glip izide w as M INIDIAB?tablet(5m g,Pf izer,Da lian,Ch ina).The test form u lation s con tain ing API-1,API-2 o r hom e-m ade API-3 w ere p repared via the w et granu lation com p ression m ethod by the fo llow ing steps:lactose m onohyd rate,m icrocrystalline cellu lose and glip izide w ere m ixed un ifo rm ly for 5m in,passing th rough an 80m esh screen for 5 tim es and then w ere added w ith 10%starch paste to obtain the soft m aterial.Soft m aterials w ere then passed th rough a 20 m esh screen to be un ifo rm granu les w h ich a llow ed to be d ried at 60°C for 1 h.A fter arranging d ry granu les th rough 20m esh screen,2%co lloida l silicon d iox ide and 0.5%stearic acid w ere added and then blend ing un iform ly to tab let.Consequen tly,th ree iden tica l fo rm u lations w ere p repared excep t the d ifferen ce of the particle size of APIs.Form u lation A,B or C w asm ade of API-1,API-2 o r API-3 respective ly.The developed test p rodu cts con tained qualitatively the sam e ingred ien ts as the re feren ce p rodu ct(data no t show n).

2.7. Dissolu tion tests

The in vitro d isso lu tion tests w ere condu cted acco rd ing to the USP Apparatus 2(Padd le m ethod,RC806D,TDTF)in d ifferen tm ed ia at 37±0.5°C.Padd les w ere ro tating at 50 rpm and 900m lof d issolu tion m ed ium w as used.5m lof sam p les w ere w ithd raw n at 5,10,15,20,30,45 and 60m in and then f iltered th rough 0.22μm f ilter.The sam e am oun t of fresh m ed ium w as rep laced at p redeterm ined tim e in terva ls.Disso lu tion fo r each form u lation w as carried ou t in trip licate(n=3)and the average of the va lues w as ca lcu lated.The content of glip izide in w ithd raw n sam p les w asm easu red by UV spectropho tom eter at 220 nm.The sim ilarity facto r(f2)w as ca lcu la ted to com pare the d issolu tion p rof iles betw een the testand referen ce form ulations.

2.8. Design of the bioavailability studies

Bo th of in vivo pha rm acok inetic stud ies w ere condu cted w ith eth ica l perm ission,w h ich w as perm itted by Eth ica l Com m ittee in Ch ina and w as p rocessed in acco rdan ce w ith the Gu ide for the Care and Use of Laboratory An im als[13].

2.8.1. Study 1

A sing le cen ter,random ized,single dose,th ree-period,th reetrea tm en t,crossover study in n ine hea lthy m a le Beagle dogs w eigh ing 10±1.5 kg,w ith a w ashou t period betw een doses of one w eek w as used.The investigated p roducts(fo rm u lation A,B and RLD)as tab let con tain ing 5m g glip izide w ith 100m l w arm w ater w ere ora lly adm in istrated under fasting cond ition.The an im als w ere p rovided w ith a standard lun ch 4 h a fter dosing.Blood sam p les(3m l)w ere co llected by ven ipun ctu re in to a heparin ized b lood co llection tube at p re-dose,and 0.5,1,1.5,2,2.5,3,4,6,8,10,12 and 24 h postadm in istration in each study period.Blood sam p les w ere imm ed iate ly cen trifuged at 3500 rpm fo r 10m in and sto red in po lyp ropy lene tubes at-80°C till fu rther ana lysis.A ll p lasm a sam p les w ere thaw ed at room tem peratu re befo re p reparing for HPLC/MS/MS ana lysis[14].

2.8.2. Study 2

A sing le cen ter,random ized,sing le dose,tw o-period,tw otreatm en t,crossover study in six healthy m ale Beagle dogs w eigh ing 10±1.5 kg,w ith a w ashou t period betw een doses of one w eek w as used.The investigated p roducts(fo rm u lation C and RLD)as tab let con tain ing 5m g g lip izide w ith 100m l w arm w ater w ere orally adm in istrated under fasting con d ition.The b lood sam p les w ere taken at p redeterm ined tim e and the p lasm a w ere p repared as described above.

2.8.3. Bioanalyticalm ethod

The glip izide p lasm a con cen trations w ere determ ined by a h igh perfo rm an ce liqu id ch rom atography-tandem m ass spectrom etry(HPLC-MS/MS)m ethod a fter p ro tein p recip itation ex traction.An aliquot of 50μl IS so lu tion(gliclazide,100 ng/m l)w as p ipetted in to a 1.5m l eppendo rf m icro centrifuge tube.A fterw ards,100μl of p lasm a and 50μl of d iluen t(m ethano l/w ater=10/90,v/v)so lu tion w ere added.The samp le w as vortex-m ixed fo r 1m in.And then 300μL ofm ethano l w as added.The m ix tu re w as m ixed fo r 1m in again.A fter being cen trifuged at 13 000 rpm fo r 10m in,the superna tan t liqu id w as for direct in jection to the HPLC-MS/MS system.The ch rom atography w as carried ou t on the Acqu ity UPLCTMsystem(W aters Corp.,M ilfo rd,MA,USA)w ith a coo ling autosam p ler.Ch rom atograph ic separation w as ach ieved on a Phenom enex C18co lum n(50mm×2.1mm,2.6μm)w ith a guard co lum n(C18,4m m×3m m,Phenom enex Ltd).The m obile phase A consisted of m ethano l and m obile phase B w as com posed of 10m m o l/l am m on ium acetate and 0.2%fo rm ic acid(A:B=60:40,v/v).The f low ratew as setat0.2m l/m in.The in jection vo lum e w as 10μl.Mass spectrom etric detection w as perfo rm ed on a tandem quad rupo le detecto r equ ipped w ith an electron sp ray ion ization(ESI)sou rce.The ESIsou rce w as set in positive m ode.The tw o com pounds w ere detected using m u ltip le reaction m on ito ring(MRM)of the transition of m/z 446.3→321.1 for glip izide and 324.1→127.1 fo r IS,w ith a scan tim e of 0.20 s per tran sition.The op tim a lMS param eters w ere set as fo llow s.The cap illary vo ltage w as 3.2 kV,the cone vo ltage w as 30V fo r g lip izide and 30V fo r IS,the sou rce temperatu re w as kep t at 120°C and the deso lvation tem peratu re w as kep t at 500°C.The op tim ized co llision energy w as 16 eV for g lip izide and 20 eV fo r IS.Nitrogen w as used as the deso lvation and cone gas w ith a f low ra te of 550 and 50 l/h respectively.

2.8.4. Plasm a concentration data processing and statistical analyses

A ll data co llected w ere carried ou t by using the M asslynxTMNT 4.1 softw are w ith the Quan LynxTMp rogram(W aters Co rp.,M ilfo rd,MA,USA).Pharm acok inetic param eters w h ich using non-com partm en talm odelana lysis to calcu late,inc lud ing the m axim um con cen tration(Cmax)and the tim e of the m axim um p lasm a con cen tration(Tmax),w ere ob tained d irectly from the m easu red da ta.The fo rm u la T1/2=0.693kew as used for ca lcu lating the e lim ination ha lf-life(T1/2).The linear regression of the term ina l poin ts in the sem i-log p lo t of p lasm a con cen tration against tim e w as fo r the pu rpose of ca lcu lating the elim ination rate con stan t(ke).The linear trapezoidal ru le w as used fo r the area under the p lasm a con cen tration-tim e cu rve(AUC0-t)to the last m easu rab le p lasm a con cen tration(Ct).The fo rm u la AUC0-∞=AUC0-t+Ct/kew as used fo r f igu ring the area under the p lasm a concen tration-tim e cu rve to tim e in f in ity(AUC0-∞).The paired t tests w ere used on the logarithm-transfo rm ed data for the Cmaxand AUC to eva luate w hether the 90%con f iden ce in terva ls w ere w ith in the range of 80-125%.

Tab le 1-Resu lts fo r pa rticle sizing.

3. Resu lts and d iscussion

3.1. Particle size analysis ofglipizide

The d ry pow der laser particle size ana lyzer w as used to analyze the partic le d istribu tion of the APIs from them anu facto ry.As show n in Table 1,API-1 had w ider partic le size d istribution due to the ex isten ce of large particles as eviden ce that D90and vo lum e m ed iand iam eter(VMD)of API-1 w ere 11.70 and 11.66μm respectively.Indeed,the size d istribu tion show ed the pow der of API-1 p resen ted tw o d ifferen t popu lation s(Fig.S1),one for ind ividua l particles and one fo r agglom erates.On the con trary,API-2,w h ich had a D90va lue of 5.65μm and a VMD va lue of 2.58μm,w asm o re un ifo rm and sm a ller com pared to API-1.

To avoid the in terferen ce of lactose in the determ ination by d ry pow der laser partic le size ana lyzer,the hom e-m ade API-3 w as in itially d ispersed in suspension and then ana lyzed.As a lso show n in Fig.S1,the average particle size of fresh p repared suspensions w as 2μm w ith a PDI of 0.202,ind icating un ifo rm d istribu tion of sm a ll APIparticles.

3.2. SEM ana lysis

SEM w as used to determ ine the su rface m o rpho logy and partic le size of APIs.As show n in the Fig.1 a ll of APIs w ere rod-like shaped w ith a sm ooth su rface,w h ich w as consisten t w ith the p revious reports[15].Consisten tly w ith the resu lts from laser pa rticle size ana lyzer,API-1 had larger partic le size w ith around 10μm in length,w hereas API-2 w as sm a ller w ith around 5μm in length.Fu rtherm ore,hom e-m ade API-3 exh ibited super-sm a ll particles as com pared w ith API-1 and API-2.

Fig.1-SEM im ages of APIs of glip izide(A):API-1;(B):API-2;(C):API-3 be fo re ba llm ill;(D):API-3 a fter ba llm ill.

Fig.2-XRPD stud ies of APIs of g lip izide(A:API-1;B:API-2).

3.3. X-ray diffraction stud ies

XRPD w as used to determ ine the iden tity of crysta lline form s p resen ting in tw o com m ercia l batches of APIs.The XRPD resu lts exh ibited no d istingu ishab le d ifferen ces in the d iffraction patterns betw een API-1 and API-2(Fig.2),ind icating the sam e crysta lline fo rm.Com pared w ith the p revious repo rts,the iden tica l peaks and their strength in the d iffraction patternsw ere consisten tw ith those of them ost stab le crysta lline form[15].PXRD analysis of API-3 and the RLD failed in the p resen t study due to ra ther low dose of glip izide and abundan t crysta lline stated lactose in the pow der of API-3 and M INIDIAB?.

3.4. In vitro dissolu tion and in vivo pharm acokinetic study

To add ress the risk of APIs raw m ateria l,w e designed tw o runs to iden tify the d ifferences in the form u lations.The f irst run em p loyed tw o com m ercia l batches of g lip izide from one supp lier to p repare the test p roducts w ith the aim to investigate the im pact of raw m ateria l variability on regard ing qua lity in p ractice.The second one u tilized the hom e-m ade APIs w ith super-sm a ll particles size to determ ine the con tro lled sa fety range of raw m ateria l for a robust fo rm u lation.Acco rdingly,the com parison of the d isso lu tion p rof iles and their co rrespond ing bioavailabilities w ere perform ed to consider the re levan t p roperties in the re-developm en t of fo rm u lations.Tofu rther eva luate the in vitro and in vivo perfo rm an ce of the developed fo rm u lations,the hardness and d isin tegration tim e of tab lets w ere com pared and the resu lts w ere show n in Tab le 2.Consequen tly,a ll of the tested fo rm u lations exh ibited rap id d isin tegration(＜2m in)and no sign if ican t d ifferen ce(P＞0.05)betw een either of test tablets w as observed in the hardness.

Tab le 2-The characte rization s of RLD and se lf-m ade tab lets.

Fig.3-The d isso lu tion p rof ile com parison s of fo rm u lation A and Bw ith the re feren ce tab lets(M IN IDIAB?)in bu ffered m ed ia.

3.4.1. Study 1

The d isso lu tion p rof iles from the test fo rm u la tions and the RLD(M IN IDIAB?tab let)w ere eva luated in fou r recom m ended d isso lu tion m ed ia(pH 1.2,4.5,6.8 and 7.4)and severa l d issolu tion m ed ia w ith pH values around pKaof glip izide(5.9).The d isso lu tion resu ltsw ere show n in Fig.3.As show n in Fig.3,the d isso lu tion p rof iles of glip izide strong ly depen ded on its so lubility.Ou r p re lim inary study revea led that the so lubility of API in itially slow ly in creased w ith pH un til its p Ka,w here a sharp enhan cem en t w as observed.Con sequen tly,it dem onstrated that glip izide released from either of the tested fo rm u lations w ith a slow d isso lu tion rate in the bu ffered m ed ia below pH 6.0,as the eviden ce of less than 40%of the to ta l am oun t d isso lved w ith in 60m in(Fig.3A-C).The d ifferen ce of d isso lu tion rate am ong fo rm u lations in these cond itions w as no t p rom inen tbecause of rather low so lubility of API.In con trast,form ulation B show ed a sign if ican tly h igher d isso lu tion rate compared w ith the form u lation A w hen them ed ia bu ffered m ed ia above pH 6.0 w ere used,p robab ly due to sm a ll particle size of APIs used in form u lation B.In order to assess the sim ilarity betw een the test fo rm u lation and RLD,the f2w as ca lcu lated and the resu ltsw ere a lso show n in Fig.3.As dep icted in Fig.3,glipizide d isso lved from fo rm u lation A w ith a low er d isso lu tion rate as com pared w ith that from M INIDIAB?in pH 6.0,6.4 and pH 7.4 bu fferm ed ia,exh ibiting less than 50 of ca lcu la ted f2.On the o ther hand,the form u lation B exh ibited sim ilar d isso lution p rof iles of M IN IDIAB?in pH 6.0,6.4 and pH 7.4 bu fferm ed ia w ith the ca lcu lated f2above 50.How ever,the d isso lu tion rate of API from the form u lation B(～90%d isso lved in 60m in)w as sign if ican tly h igher than that of M INIDIAB?(～70%d isso lved in 60m in)in the d isso lu tion m ed ium of pH 6.8,resu lting in 30.0 of f2.On the con tra ry,the d ifferen ce of d isso lu tion rate betw een the fo rm u lation A and M INIDIAB?w as redu ced in PBS 6.8(f2of 58.0),especia lly for the in itial 20m in.

Tab le 3-Pharm acok inetic param eters obtained from the Glip izide form u lations in clud ing M INIDIAB?(RLD),form u lation A and B a fte r sing le o ra l adm in istration in beag le dogs.(da ta w ere show n as m ean±SD,n=9).

Fig.4-M ean p lasm a con cen tration-tim e cu rves of g lip izide a fter o ra l adm in istra tion of fo rm u lations A,B o r M IN IDIAB?.

Based on above in vitro d isso lu tion resu lts,bo th of test fo rm u lations w ere then eva luated their bioequ iva len ce w ith RLD in Beag le dogs.The m ean con cen tration-tim e p rof iles of glip izide after single oral adm in istration of either of test fo rm u lations or RLD w ere dem onstrated in Fig.4 and the invo lved pharm acokinetics param eters w ere show ed in Tab le 3.As show n in Fig.4,the o ra l adm in istration of fo rm u lation A gave a low er p lasm a con cen tration of g lip izide than that of RLD o r fo rm u lation B over 24 h.As a resu lt,bo th of AUCt(23149.12±6818.19 ng·h/m l)and Cmax(2255.2±629.74 ng/m l)w ere low er for form ulation A than those for RLD(27691.59±10106.13 ng·h/m l and 3139.88±964.07 ng/m l).On the con trary,the form ulation B w ith sm a ll partic le size of API dem onstrated com parab le AUCt(28417.33±11042.82 ng·h/m l)and Cmax(2803.22±833.23 ng/m l)of RLD.

Acco rd ing to the gu ide line,the test fo rm u lation is believed bioequ iva len t w ith the referen ce w hen the 90%con f iden ce in terval(CI)of the m ean ratios(test/referen ce)for the log-transform ed Cmax,AUC0-∞and AUC0-tare w ith in the range of 80.0%-125.0%.The tw o one-sided t tests and 90%CIs resu lts of AUC and Cmaxw ere sum m arized in Tab le 4.As exh ibited in Tab le 4,the 90%CIs of Cmaxand AUC0-tfo r fo rm u lation A w ere 65.5%-81.0%and 76.4%-95.0%,respectively,in dicating bioinequ iva len t w ith the RLD.Conversely,the 90%CIs of Cmaxand AUC0-tfor form u lation B w ere 81.0%-100.3%and 91.3%-113.5%,respective ly,w ith in the range of80.0%-125.0%,ind icating bioequivalen tw ith the RLD(Tab le 5).

3.4.2. Study 2

A lthough form u lation B w as assessed as bioequ iva len t,the abso rp tion rate of glip izide adm in istrated w as still a bit low er w hen com pared w ith RLD(2803.22±833.23 ng/m l vs 3139.88±964.07 ng/m l,P＞0.05).Then form u lation C w ith super-sm a ll particle size of APIs w as then used in the second run.Fig.5 dem onstrated the com pa rative d isso lu tion p rof iles in the sim ilar cond itions of the f irst run.As estim ated,fo rm u lation C exh ibited faster d isso lu tion rate in alm ost all the investigating m ed ium s.The f2of the d isso lu tion p rof iles for the test fo rm u lation vs.RLD at pH 6.0,6.8 and 7.4 w ere 46.5,20.5 and 32.7,respective ly.The pharm acok inetic study,w h ich com pared form u lation C to the RLD,how ever,ind icated sim ilar AUC and Cmax(P＞0.05)w ith a fter o ra ladm in istration(Fig.6 and Tab le 6).Sin ce h igh w ith in-sub ject variability in AUC and Cmaxw ere obtained in Study 2,it w as no t su itab le to p roceed in to the bioequ iva len t assessm en t due to a rather sm a ll samp le of beag le dogs.

3.5. Discussion

The p resen t study ind icated that the variability of APIs can severely im pact in vitro d issolu tion and in vivo absorp tion of g lip izide.Accord ing ly,it is crucia l to add ress the sou rces of variation to set up a con tro lled strategy fo r re liab le p roduct qua lity by a design of robust fo rm u lation and p rocess.If possib le,the eva luation of APIvariability needs to be iden tif ied in an early stage of fo rm u lation deve lopm en t to m itigate risk for the f ina l p roduct qua lity[16].

Notably,several param eters of APIs cou ld be used to character physicochem ica l p roperties of raw m ateria l and then their im pacts on p rocessability and d rug p rodu ct qua lity,su ch as the particle size,po lym orph ism,agglom eration p rof ile and specif ic su rface area[9,17].Am ong these param eters,tw o m ain independen t variab les(the particle size and po lym o rph ism)cou ld be rou tinely used to understand the physicochem ica l p roperties of raw m ateria ls.Genera lly,the ana lysis of partic le size d istribu tion w ou ld m irror the agglom eration p rof ile of API p rovided the consisten cy of p rodu ction.In the sam e w ay,specif ic su rface area of raw m aterials is h igh ly depended on the crysta l habit and the particle size of APIs.As show ed in Fig.1,a ll tested APIs exh ibited sim ilarm orpho logy,w h ich w as in accordan ce w ith the p revious resu lts founded in them ost stable po lym orph ism of glip izide.Accord ingly,the particle size w ou ld be the key factor to add ress the con tro lled strategy of API as long as the sam e po lym orph ism of glipizide has been em p loyed in the p reparation of p rodu cts.To avoid them isin terp retation ofm u ltivariate param eters,a sing le m ateria l param eter(partic le size)w as a llow ed to exp lore the in f luen ce of APIvariabilities on d rug p roduct qua lity in the p resen t investigation.The reason w as because that APIvariation in particle size w as con sidered as the rep resen tative and regu lar“batch to batch”variability[18].On the o ther hand,the so lid state of APIs is another key determ in ing facto r fo r their basic physicochem ica l p roperties,in clud ing so lubility,stability and d isso lu tion rate[15,19].In th is respect,them ost stable po lym o rph ism of g lip izide w as used in the p reparation p rocess of the test fo rm u lations.W e d id no t observe any change in the perfo rm an ce characteristics of dosage fo rm,such as pow er f low,tablet com p ressibility and m echan ical strength,w hen the APIs from d ifferen t batches w as in troduced in the fo rm u lation.

Tab le 4-Tw o one-sided t-test ofm ain pa ram eters be tw een fo rm u la tion A and the RLD.

Tab le 5-Tw o one-sided t-test ofm ain pa ram eters be tw een fo rm u la tion B and the RLD.

Fig.5-The d isso lu tion p rof ile com pa risons of fo rm u lation C w ith the re feren ce tab lets(M IN IDIAB?)in bu ffered m ed ia.

Fig.6-M ean p lasm a con cen tration-tim e cu rves of glip izide a fter o ra l adm in istration of fo rm u lation s C o r M INIDIAB?.

Tab le 6-Pharm acok inetic pa ram eters fo r Glip izide fo rm u lation in clud ing M INIDIAB?,fo rm u lation C(data w ere show n as m ean±SD,n=6).

Being a w eak acid(p Ka=5.9),glip izide is better abso rbed in basicm ed ium,w hereas in very acid icm ed ia,the solubility of g lip izide ism in im a l[20,21].Herein,the chosen d isso lu tion m ed ium aim ed to m im ic the in vivo gastroin testina l cond itions has been w ide ly used in the in vitro d isso lu tion investigation.Consequen tly,the eva luation of the generic fo rm u la tions at pH 1.2,4.5 and 6.8 ref lects rough ly the pH cond itions in the gastroin testina l tract,how ever,the com parative d isso lu tion in relevan t m ed ia m ay not d istingu ish sign if ican t changes in the com position and m anu factu ring p rocess.Fo r exam p le,the d isso lu tion p rof iles of fo rm u lation A in pH 1.2,4.5 and 6.8 cond itions w ere assessed as sim ilar w ith the RLD,w hereas the p roduct failed in the bioequivalen ce test.Conversely,the fo rm u lation B exh ibited faster d isso lu tion of APIin the rega rding cond itions,w h ile p roving the bioequ iva len ce to RLD.It ind icated tha t the d iscrim inato ry pow er of the classica l d issolu tion criteria w as not suitable for the p rodu ct qua lity evaluation of g lip izide.Therefore,the com parative d isso lu tion studiesw ere fu rther condu cted at pH 6.0,6.4 and 7.4 in the p resen t study w ith the aim to cover the m idd le and h igh so lubility regions of API.As an ticipated from the so lubility eva luation,rap id and com p lete d isso lu tion fo r fo rm u lation B and RLD w as observed at pH 7.4(f2of 50.8),w h ile form u lation A w ith coarse particle size distribu tion led to a slow and incom p lete d isso lution.In them idd le so lubility region of API,form u lation A and B a lsom ain tained slow and com parab le d isso lu tion rate respectively w hen com pared w ith RLD.Specia lly,the run-to-run variability in the d issolu tion of RLDw as notobserved at pH 5.5,6.0 and 6.4,w here the pH cond ition w asm ost close to p Ka of API.Overa ll,the d isso lu tion m ed ium w ith pH 6.0 w as considered to be robust enough to to lerate sligh t changes(±0.4 pH)in labo rato ry cond itions to p rovide rep roducib le d isso lu tion p rof iles of API[22].Therefo re,the d isso lu tion m ed ium w ith pH 6.0 and 7.4 w ere in itia lly p roposed to possess su itab le d iscrim inato ry natu re.

The second run of com parative d isso lu tion and pharm acok inetic study w as conduct to seek the accep tab le low er lim it of APIs'particle size.The obtained resu lts suggested overd iscrim inating d isso lu tion tests in pH 6.0 and 7.4.Specially,fo rm u lation C exh ibited faster d isso lu tion rate(f2＜50.0),however,it show ed com parab le o ra l abso rp tion ex ten t and rate w hen com pared w ith the RLD.W e specu lated that one possible reason for th is phenom enon w as less sensitivity of oral abso rp tion of glip izide given the pa rticle size of APIs in the range of 2-5μm.Therefo re,2-5μm m igh t be def ined as the“inert size range”of APIs fo r en su ring the bioequ iva len ce w ith the RLD[23].Ou r p relim inary stud ies ind icated that the granu larity of g lip izide cou ld be reduced to~4μm(D90,m easu red by d ry pow der laser partic le size ana lyzer)by the jet grind ing m ill.The pow der of APIs ob tained from the relating p rocess m igh t be used for the deve lopm en t of generic p rodu cts.

4. Con clu sion

In con c lusion,the p resen t study h igh ligh ted that the particle size d istribu tion of g lip izide w as an im po rtan t facto r to the in vitro d isso lu tion and thus in vivo bioavailability of redeve loped o ra l tab lets.The d isso lu tion p rof iles in pH 6.0 and 7.4 m ed ium s offered d iscrim inatory natu re in understand ing the im pact of API's partic le size.The form u lation s w ith API's particle sizes in the range of 2-5μm appeared to exh ibit similar o ra l abso rp tion w hen com pared w ith the re feren ce p rodu cts.In the near fu rther,the c lin ica l tria lsw illbe condu cted to assess the bioequ ivalence betw een the re-developed form u lation and the RLD.

Con f lict of in terest

The au tho rs repo rt no con f licts of in terest.The au tho rs a lone are responsib le for the content and w riting of th is article.

Acknow ledgm en ts

Th is research w as supported by National Scien ce and Techno logy M a jor Pro jects fo r“Ma jo r New Drugs Innovation and Developm en t”(No.2017ZX09101-001-005,Beijing,Ch ina).

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.2018.06.005.