Qiang Zhu,ShiruiMao

School of Pharm acy,Shenyang Pharm aceutical University,No.103,W enhua Road,Shenyang 110016,China

Keywords:Con tact len ses D rug load ing Ion ic strength Abso rp tion isotherm s D rug so lubility Boun d w ater content

A B S T R A C TLow d rug load ing eff icien cy is one of them ain obstacles h in dering the app lication of con tact len ses(CLs)as the carrier fo r ex tended ocu lar d rug de livery.Here in th is study,a sim p le and effec tive d rug load ingm ethod based on salt in du ced m odu lation w as p roposed and dem onstrated w ith m echan ism elucidation.First of a ll,using po ly(2-hyd roxyethy lm ethacry late)(p-HEMA)as the con tac t len sm ateria l,betaxo lo l hyd roch lo ride,Diclofenac Sod ium and Betaxo lo l Base as the m odel d rugs w ith d ifferen t so lubility,in f luen ce of salt con cen tration,sa lt type(sod ium sa lts of su lfate,ch lo ride,and su lfocyanate)and d rug p roperties in the load ing so lu tion on d rug load ing e ff icien cy w as investigated.M echan ism of enhan ced d rug load ing in con tact lens w as fu rther exp lo red via studying the in f luen ce of sa lt on the absorp tion isotherm,d rug solubility and w ater content of CLs.App licability of th is m ethod to o ther CLs m ateria ls w as a lso investigated.It w as dem onstrated that ad justing the ion ic strength of load ing so lu tion s resu lted in sign if ican t in crease of d rug load ing in CLs.Type and concen tration of the sa lts and so lubility of the d rug w ere them ain facto rs in f luen cing enhan cem en t ratio of d rug load ing.Them echan ism for im p roved d rug loading w as related to the reduced d rug so lubility in load ing so lu tions and the reduced boun d w ater content in con tact len ses.M odu lation of d rug load ing by ad justing ion ic strength w as a lso app licab le to other CLs and the ligh t tran sm ittan ce w as not a ffected.Th ism ethod w asm o re su itab le fo r sa lt-form d rugs w ith h igh so lubility.In sum m ary,ad justing ion ic strength of load ing solu tion is an econom ica l and e ffective w ay to im p rove d rug load ing in CLs,and th is sim p le m ethod m ay a lsof in d app lication in o ther hyd rogel based d rug delivery system s.

1. Introduction

Top ica l delivery is the m ost accep ted rou te of ocu lar d rug app lication fo r its sim p licity and a ffo rdab le p rice.It is often used to treat an terior segm en t d iseases such as glau com a,d ry eye,trachom a,although som e studies have p roved that d rugs can reach posterio r segm en t of the eye th rough top ica l rou te[1].A t p resen t app roxim ately 90%of aqueous oph tha lm ic form ulations are used in th is w ay.How ever,d rugs instilled as eye d rops usually have a low bioavailability less than 5%due to the ex isten ce of static and dynam ic barriers[2],therefore frequen t dosing is requ ired.A t th is stage a lm ost 75%of the eye d rops on the m arket are recomm ended to use 3 tim es daily o r even m o re,w h ich is a great cha llenge to the com p lian ce of patien ts.

Sofar,severa l strategies have been used to p ro long d rug reten tion in the p recorneal area,su ch as viscosity-enhancing po lym ers[3],in-situ ge ls[4,5],m u coadhesive system[3],ocular inserts[6]and soft con tact lenses[7].In order to in crease d rug perm eation across the co rnea,nanocarries[8],penetration enhan cers[9]and so lubility enhan cers[10]are comm on ly used,w ith im p roved d rug ocu lar bioavailability.Am ong them,con tact lenses(CLs)have attracted ex tensive atten tion due to greatly im p roved ocu lar reten tion tim e com pared w ith o ther m ethods.In 1965 con tact lens w as f irstly used in ocu lar d rug de livery and sin ce then num erous stud ies have p roved that con tact lens is m o re effective than eye d rops[7,11,12].However,no m edicated con tact len ses are availab le in the m arket un til now and low d rug load ing is the p rim ary obstacle h indering fu rther developm en t of CLs in the f ie ld of d rug delivery.Fo r d rug load ing in CLs,soak ing and p re-in co rpo rating are tw o comm on ly used m ethods.Pre-in corporating is to introduce d rugs be fo re CLs being fo rm ed,w h ich is often used in nanocarriers laden CLs[13].Soak ing is the conven tiona l m ethod w ith CLs im m ersed in d rug so lu tion,w h ile m ost of the d rug in the soaking so lu tion is not loaded by the CLs and d rug w aste is a p rob lem[12].To im p rove d rug load ing of CLs,severa lattem p ts have been carried ou t in c lud ing accom m odation of CLs com position,m o lecu lar im p rin ting[14],in co rpo ration of cyclodextrins[15],m u ltilayer CLs[16].Those strategies have ach ieved som e p rogress,how ever,in com p lete d rug release and lens defo rm ation in CLs copo lym erized w ith ion izab lem onom ers[17],m odu lus in crease in hyd rophobic silicone[18]and cyclodex trins CLs[15],h igh crosslinker in m o lecu lar im p rin ting CLs[19]are undesired.There fo re,a sim p le and e ffective d rug load ing m ethod w ithou t a ltering lens com position s are h igh ly desirab le.

Unexpected ly,du ring ou r p re lim ina ry experim en t,f ivefo ld in crease of d iclofenac sod ium load ing in po ly(2-hyd roxyethy l m ethacry late)(pHEMA)lens w as found w hen sim u lated tear f luid(STF)w as used as the soaking solven t com pared to that using d istilled w ater as so lven t.Therea fter,d ifferen t sa lt so lution s w ere used as load ing so lven t and itw as foun d that d rug load ing in a ll of those sa lt so lu tions w ere im p roved com pared w ith thatof d istilled w ater.How ever,w hether th is is a general ru le fo r o ther types of CLs o r d rugs,and the m echan ism for in creased d rug load ing by sa lt add ition are un c lear and needs to be e lu cidated.Meanw h ile,detailed characterization of sa lt in f luen ce on the p roperties of CLs is essen tial.

Therefo re,them ain ob jectivesof th is study are(1)to understand the m echan ism s of d rug load ing in crease via in creasing ion ic strength of load ing so lu tion;(2)to investigate the in f luen ce of salts add ition on the p roperties of CLs.Both hyd roph ilic and hyd rophobic d rugs,Dic lofenac Sod ium(DS),Betaxo lo l Base(BB)and Betaxo lo l Hyd roch lo ride(BH)w ere selected as d rug m odel in th is study.Acco rd ing to Hofm eister series,sod ium sa lts of su lfate,ch lo ride,and su lfocyanate w ere chosen to investigate the effect of sa lts on d rug so lubility and CLs load ing.

2. M ateria ls and methods

2.1. M aterials

2-hyd roxyethy lm ethacry late(HEMA),m ethacry lic acid(MAA),4-Viny lpy rid ine(VP), N-viny l py rro lidone 4-viny lpy rid ine(NVP),3-m ethacry loxyp ropy l tris (trim ethy lsiloxy)silane(TRIS)w ere pu rchased from TCI(Tokyo,Japan).Ethy lene glyco l d im ethacry late(EGDMA)w as ob tained from Fo rtuneibo-tech Co.,Ltd.(Shanghai,Ch ina).2,2′-Azobis(2-m ethy lp rop ion itrile)(AIBN)w as supp lied by Qingdao Kex in M ateria ls and Techno logy Co.,Ltd.(Qingdao,Shangdong,Ch ina).Diclofenac Sod ium(DS)w as from Hubei Prosperity Ga laxy Chem ica l Co.,Ltd.(W uhan,Hubei,Ch ina).Betaxo lo l hyd roch loride(BH)and Betaxo lo l Base(BB)w ere from BioChem Partner Co.,Ltd.(Shanghai,Ch ina).A ll o ther chem ica ls w ere of ana ly tica l grade.

2.2. Syn thesis of con tact lenses

CLs w ere syn thesized by free rad ica l so lu tion po lym erization of the m onom er as reported p reviously[20].Brie f ly,an app rop riate am oun t of the crosslinker EGDMA w as d isso lved in HEMA to obtain a concen tration of 0.5%(w/w).Then,the m ix tu re w as bubb led w ith n itrogen fo r 15 m in to rem ove d isso lved oxygen be fo re the add ition of 0.3%(w/w)AIBN(in itiato r).Bo th m ix tu res w ere in jected in to CLs m o lds to keep a sw e lling th ickness of 100μm.The m o lds w ere then p laced in an oven fo r 12 h at 50°C fo llow ed by 24 h at 70°C.A fter po lym erization,CLsw ere im m ersed in d istilled w ater overn igh t tofacilitate the separation of CLs from them o lds.Thereafter,the CLs w ere im m ersed in d istilled w ater fo r 3 d,rep lacing w ater 3 tim es daily to rem ove residua lm onom ers.Fina lly,the CLs w ere d ried a t 40°C under vacuum overn igh t to reach a constan tw eigh t.

2.3. Preparation and characterization ofdrug loaded con tact lens

D rug loaded con tact lens w ere p repared using soak ing m ethod.Brief ly,d ry CLs w ere im m ersed in 1 m l aqueous d rug so lu tions and w ere kep t at 25°C in dark for 3 d.D rug con cen tration before and a fter soak ing p rocedu re w as determ ined spectropho tom etrica lly at 276 nm for DS(A=0.0284C+0.0108,r=0.9994),272 nm for BH(A=0.0035C+0.0171,r=0.9995),272 nm fo r BB(A=0.0039C+0.0172,r=0.9994)(UV-2000,UNIC Instrum en t Co rp.,Shanghai,Ch ina).The d rug load ing am oun t(Qloading)and load ing e ff icien cy(LE)w as ca lcu lated,respectively,by the m ass ba lan ce acco rd ing to the fo llow ing equations:

w here Qloadingis the am oun t of d rug adso rbed per gram of CLs(m g/g),C0is the in itia l d rug con cen tration(g/l),Ceis the equ ilib rium d rug con cen tration(g/l),m is the m ass of CLs(g)and V is the volum e of load ing so lu tion(l).

Tofacilitate the illustration of sa lt effect on d rug load ing of CLs,Ei,enhan cem en t facto r of sa lt(Eq.3)w as def ined as the ratio of d rug load ing in salt so lu tions to d rug load ing in d istilled w ater.

w here Qloading,saltand Qloading,waterare the am oun t of d rug adsorbed per gram of CLs(m g/g)using sa lt so lu tion and d istilled w ater as so lven t,respective ly.

2.4. Characterization of con tact lenses

2.4.1. Water content ofcontact lenses

Dry CLs w ere w eighed accu rately and then im m ersed in d istilled w ater or d ifferen t salt so lu tions,at 25°C for 3 d and the in crease in w eigh tw asm easu red.Tota lw ater content(WCtotal)of CLs w as ca lcu lated acco rd ing to Eq.4.

w here W0is the in itia lw eigh t of the d ry CLs;Wtis the w eigh t of the sw o llen CLs.

Them ass of bound w aterw as ob tained from the d ifferen ce betw een the m ass of to ta lw ater and that of free w ater as described in Eq.5.To m easu re the content of free w ater in contact lens,DSC stud ies w ere perform ed using a Mettler To ledo m odel w ith the STARe System(Mettler,Zu rich,Sw itzerland).The sam p les(3-5m g)w ere scanned in sea led a lum inum pans under n itrogen atm osphere.DSC cu rves w ere stud ied over a tem peratu re range of-20-10°C at a constan t rate of 3°C/m in.The content of free w ater w as obtained acco rd ing to Eq.6.

w here Q is the heat abso rbed du ring the m e lting p rocess;ΔH is them elting en tha lpy ofw ater,W0is the w eigh t of d ry lenses[21].

Fo r both w ater content ca lcu lations,w eigh t of d ry lens,o ther than w eigh t of sw o llen lens w as used as denom inato r to re f lect the relationsh ip betw een hyd roge l and so lven t in tuitively.

2.4.2. Light transm ittance

Wet CLs w ere cu t w ith a cork borer and the d iscs w ere pu t in to a 96-w e ll p late.The transm ittan ce of CLs w asm easu red on a SpectraM ax M 3 m icrop la te reader ranging from 400 nm to 900 nm using SpectraM ax Pro Softw are(M o lecu lar Devices,Sunnyvale,USA)and w ater w as used as blank con tro l.

2.5. Absorption isotherm s

Differen t con cen trations of DS(0.1-2 g/l),BH(0.5-5 g/l)and BB(0.1-0.5 g/l)in d istilled w ater o r 0.15m o l/lNaClso lu tionsw ere used as load ing so lu tions to ob tain abso rp tion iso therm in d ifferen t con tact lens.Equ ilibrium isotherm dataw ere f itted in to Linear(Eq.7),and Freun d lich(Eq.8)abso rp tion iso therm m ode ls.

w here Qloadingis the am oun t of d rug adsorbed per gram of CLs(m g/g),Ceis the equ ilibrium d rug con cen tra tion(g/l).K,and KFare the Linear and Freund lich isotherm constan ts,n is the Freund lich exponen t factor[22].

2.6. Solubility study

Excess d rugs w ere p laced in 5 m l d istilled w ater o r sa lt so lutions w ith d ifferen t ion ic strength(NaCl,Na2SO4,NaSCN at 0.005-1m o l/l)in a 10m l test tube.Then the tube w as pu t in to a shaker w ater bath at 25°C and agitated at 100 rpm.A t specif ied tim e poin ts,the suspension w as cen trifuged and the supernatan tw as f iltered th rough a 0.45μm M illiporem em brane.The f iltratew as then d ilu ted and assayed for d rug content.The em p irica l Setschenow equation w ith littlem od if ication(Eq.9)w as used to describe the in f luen ce of sa lts on d rug so lubility,as fo llow s:

w here S0is the d rug solubility in w ater,Sappis its solubility in sa lt so lu tion s,Ksis the Setschenow constan t,I is the ion ic strength of the so lu tion,ciis the m o lar con cen tration of ion(m o l/l)and ziis the charge num ber of that ion.The Ksva lues are positive for sa lts that decrease so lubility(i.e.,sa lting-ou t)and negative for sa lts that in crease so lubility(i.e.,sa lting-in)[23].

2.7. In vitro release

Drug release from the con tact lens w as investigated in vitro cond ition.STF(6.78 g/l NaCl,2.18 g/l NaHCO3,1.38 g/l KCl,0.084 g/l CaCl2·2H2O)w as used as the releasem ed ium.Brie f ly,the CLs w ere rinsed w ith d istilled w ater a fter load ing and d ried w ith absorben t paper in order to rem ove residua l d rug on the su rface.Then the CLs w ere im m ersed in 10 m l of release m ed ium at 35°C,un der stirring(100 rpm).A t p redeterm ined tim e in terva ls,2 m l a liquo ts of the supernatan t w ere w ithd raw n for d rug content m easu rem en t and m eanw h ile,the sam e vo lum e of fresh release m ed ium w as added.Drug release stud iesw ere ca rried ou t in trip licate fo r each fo rm u lation tested and standard deviations w ere calcu lated.The d ifferen ce in d isso lu tion p rof iles w as eva lua ted using sim ilarity factor(f2),w h ich can be ca lcu lated using Eq.11[24].

where n is the num ber of tim e poin ts,Rtis the d isso lu tion value of the reference at tim e t,and Ttis the disso lu tion value of the test at tim e t.The release p rof iles are sign if ican tly d ifferen t if f2＜50.

Fig.1-In f luen ce of sa lt type on DS load ing.DS load ing of p-HEMA CLs from 0.5 g/l DS in sa lt so lu tion s w ith ion ic strength ranging from 0.005 to 1 m o l/l using NaCl(A),Na2SO4(B),NaSCN(C)as ad justo r and the linear co rre lation betw een the loga rithm s of the ion ic streng th and DS load ing am oun ts of CLs in NaCl,Na2SO4 and NaSCN so lu tion s(D).Each po in t rep resen ts the average of th ree m easu rem en ts.

2.8. Sta tisticalanalysis

A ll resu ltsw ere exp ressed asm ean±SD.M icrosoftO ff ice 2010(M icrosoft Co.,Redm ond,Wash ington State,USA)and Origin Pro 8.0(Origin Lab Co.,Northam p ton,Massachusetts,USA)w ere em p loyed for statistica l ana lysis.T-test and One-w ay ana lysis of varian ce w ere used to com pare load ing capacity of CLs,w ater content of CLs and d rug solubility in d ifferen t salt so lu tions.The d ifferen cesw ere considered to be sign if ican tat P＜0.05.

3. Resu lts and d iscussion

3.1. Sa lt effect on drug loading ofcon tact lenses

In p revious stud iesm o re atten tion has been paid to in crease d rug load ing via stru ctu ralm od if ication s of CLs[17-20],the poten tial in f luence of salts on d rug load ing has been ignored.Here,in f luen ce of sa lts con cen tration and sa lt type on d rug load ing in CLs w as stud ied in detail.To e lim ina ting the in f luen ces of so lu tion pH on d rug load ing,in add ition to NaCl,tw o other sod ium salts of strong acids,Na2SO4and NaSCN w ere used in fo llow ing stud ies and the pH of the d rug load ing so lution sw erem easu red and kep t constan t(pH=6.5,DS;pH=5.5,BH;pH=9.5,BB).

3.1.1. Inf luence ofsalt concentration

To elu cidate the effect of sa lt con cen tration on d rug loading of CLs,f irst of a ll,tak ing p-HEMA CLs as an exam p le and DS as the m odel d rug,in f luen ce of NaCl concen tration on d rug load ing w as investiga ted and the resu lts are p resen ted in Fig.1A.W hen DS-w ater w as used,the d rug load ing e ff icien cy w as on ly 7.58%,im p lying a lm ost 92%of d rug w as un loaded and th isw as ineff icien t.As NaClw as added to the load ing solu tion,excep t fo r the 0.005 and 0.01 m o l/l group,a ll the o ther NaCl con cen tration groups enhan ced DS load ing sign if ican tly com pared w ith the d istilled w ater group(P＜0.05).Drug loading eff icien cy of 0.15m ol/l NaCl group w as 37.17%.Moreover,there existed a good linear co rrelation betw een the logarithm s of the ion ic strength and DS load ing am oun t in CLs(r=0.9834)(Fig.1D),im p lying that itw as possible to ad just the d rug loading ability of CLs via changing ion ic strength of the soaking so lu tion.

3.1.2. Inf luence ofsalt type

In add ition to ion ic strength,w hether sa lt type can in f luen ce d rug load ing in CLs at the sam e ion ic strength needs to be clarif ied.There fo re,tak ing p-HEMA CLs as an exam p le,the e ffect of sa lt type on DS load ing in CLs w as fu rther investigated by se lecting tw o o ther sa lts,Na2SO4and NaSCN,to ad just the ion ic strength in the sam e range of 0.005-1 m o l/l,and compared w ith NaCl based system.Here DS con cen tration w as kep t at 0.5 g/l.As show n in Fig.1B and C,sim ilar to NaCl,the load ing am oun t of DS in Na2SO4and NaSCN so lu tions increased w ith in creasing ion ic strength bu t w ith d ifferen t exten t.Linear co rre lation a lso ex isted betw een logarithm of the ion ic strength and DS load ing in CLs in Na2SO4(r=0.9726)and NaSCN(r=0.9839)based system s(Fig.1D).In terestingly,itw as no ticed that the in f luen ce of sa lt type on d rug load ing w as a lso ion ic strength con cen tration dependen t.W hen the ion ic strength w as under 0.01m o l/l,no rem arkab le in f luen ce of sa lt type w as found,and the d rug load ing on ly in creased sligh tly w ith the in crease of ion ic strength irrespective of sa lt type.How ever,w hen the ion ic strength w as equa l o r greater than 0.01 m o l/l,the in f luen ce of sa lt type w asm o re apparen t.The effect of Na2SO4and NaClon d rug load ing in CLsw as com parab le in the ion ic strength range of 0.01-1m o l/l(P＞0.05),bu t sign if ican tly low er d rug load ing w as found in NaSCN based system com pared w ith that of Na2SO4and NaCl con tain ing system(P＜0.05).Therefore,the in f luen ce of salt type ism ore apparen t at h igher ion ic con cen tra tion.

3.1.3. Inf luence ofdrug type

The above study dem onstrated that DS load ing in CLs is depen den t on salt type and con cen tration in the in cubation m ed ium.To clarify w hether th is is a genera l ru le fo r d rug loading in CLs,by using the sam e CLs(p-HEMA),the load ing behavior of tw o other d rugs w ith d ifferen t so lubility characteristics,Betaxo lo lHyd roch loride(BH)(S＞1000 g/l),and BetaxololBase(BB)(S=0.67 g/l),w as stud ied in detail.As show n in Fig.2A,w hen BH w as used as the m odel d rug,the load ing am oun t in CLs enhan ced sign if ican tly com pared w ith the load ing in d istilled w ater irrespective of salt type(P＜0.05),bu t the exten t of d rug load ing in crease w as sa lt type and ion ic strength depen den t.A t physio logica l ion ic strength of 0.15 m o l/l,BH load ing in creased in the o rder NaSCN＞NaCl＞Na2SO4.BH load ing in NaSCN based system w as tw o tim es h igher than that of NaClbased system,and th ree tim es h igher than that of Na2SO4based system.BH load ing decreased w ith the in crease of ion ic strength in NaCl(r=0.8886)or Na2SO4(r=0.9330)based system.In NaSCN based system,in itia lly BH load ing in creased rem a rkab ly w ith the in crease of ion ic strength,bu t w hen NaSCN con cen tration w as above 0.15 m o l/l,BH load ing decreased w ith fu rther in crease of NaSCN con cen tration.

As show n in Fig.2B,w hen BBw as used as am ode ld rug,d ifferen t phenom enon w as observed.W hen the ion ic con cen tration is under 0.15m o l/l,no im p rovem en t in d rug load ing w as found irrespective of sa lt type.W hen the ion ic con cen tration w as in creased to 1 m o l/l,sign if ican t in crease in d rug load ing w as observed in NaCland Na2SO4based system(P＜0.05),bu t d rug load ing decreased statistica lly in NaSCN based system(P＜0.05).

Fig.2-Load ing resu lts of BH(A)and BB(B)in NaCl,Na2SO4 and NaSCN so lu tion s w ith the ion ic streng th ranging from 0 to 1 m o l/l and com pa rison of enhan cem en t facto r fo r DS,BB and BH in NaCl so lu tion s(C).Each poin t rep resen ts the average of th ree m easu rem en ts.(?P＜0.05,??P＜0.01,???P＜0.001).

Tofu rther elu cidate the in f luen ce of d rug p roperties,taking NaCl based system as an exam p le,enhancem en t facto r,Eiva lue w as ca lcu lated at d ifferen t sa lt con cen trations.As show n in Fig.2C,Eiva lues are com parab le fo r DS and BH in a ll the con cen tration ranges investiga ted and all the values are above one.In con trast,in f luen ce of ion ic con cen tration on the load ing of BB w as m argina l.Th is study ind icated that the ex ten t of enhan ced d rug load ing v ia sa lt add ition is d rug p roperty dependen t.It seem s that th is strategy is m ore effective fo r d rugs w ith h igher so lubility(BH,S＞1000 g/l;DS,S=17 g/l)than that w ith low er ones(BB,S=0.67 g/l).

Tab le 1-Resu lts of the f itting of the Linea r and Freun d lich m ode ls w ith the load ing of DS,BH,and BB.

3.2. M echan ism s ofenhanced drug loading in con tact lenses

The above stud ies dem onstrated that sa lt con cen tration,sa lt type and p roperties of d rug can im pact d rug load ing in p-HEMA con tact lenses.How ever,them echan ism beh ind is unc lea r and needs to be exp lored.

3.2.1. Inf luence ofsalts on absorption isotherm s

Abso rp tion iso therm s are com m on ly used to describe the d istribu tion of d rug betw een CLs and so lu tions in the equ ilibrium state.By using d istilled w ater and 0.15m o l/l NaCl as the in cubation solven t,the absorp tion isotherm s of th ree m odel d rugs in p-HEMA based CLs w ere stud ied and f itted in to tw o m ode ls,linear and Freund lich abso rp tion iso therm s,the resu lts are p resen ted in Fig.3 and Tab le 1.It w as found that no m atter d istilled w ater or 0.15m o l/l NaClw as used as so lven t,Freund lich isotherm got the best f it w ith experim en ta l data for a ll the d rugs,im p lying that the natu re of CLs d rug load ing m igh t be d rug abso rp tion on th ree-d im ensiona l netw o rk of the lens[25]and saltadd ition has no in f luen ce on d rug absorption m echan ism.In the case of DS and BH,KF,the Freund lich iso therm constan ts in creased app ro.5 fo ld in 0.15 m o l/l NaCl based so lu tion com pared to that in d istilled w ater,im p ly ing a h igher abso rp tion capacity[26].Sim ilarly,a ll the va lues of Freund lich exponen t factor in sa lt groups(n＞1)w ere h igher than that in w ater groups(n＜1),in d icating m o re favo rab le cond ition for d rug absorp tion in salt con tain ing solu tions[27].W hen BBw as used as them ode ld rug,d rug absorp tion behavior in bo th d istilled w ater and 0.15 m o l/l NaCl based so lu tion w as com parab le,no rem a rkab le in f luen ce of sa lt on d rug absorp tion ex ten tw as d isclosed.

3.2.2. Inf luence ofsalts on drug solubility

The abso rp tion iso therm s stud ies ind icated that the natu re of d rug load ing m igh t be abso rp tion on CLs irrespective of ions con cen tration in the in cubation so lu tion.It isw ellknow n that the abso rp tion in aqueous so lu tion is a com p lex p rocess invo lving the in term o lecu lar action betw een so lu te and so lven t,so lu te and adsorben t,adsorben t and so lven t.Drug so lubility cou ld be used as an in d icato r of so lu te-so lven t a ff in ity.It is assum ed that sa lt add ition m ay in crease so lu te-adsorben t a ff inity by decreasing so lu te-so lven t a ff in ity,there fo re lead ing to decreased d rug so lubility and enhan ced d rug load ing.

Fig.3-Abso rp tion iso therm s of DS(A),BH(B),BB(C)in w a ter(squa re and so lid line)o r 0.15 m o l/l NaCl(d iam ond and dash line).Fitting cu rves w ere ob tained based on the tw o m odels,linear-type m odel(b lue)and Freun d lich-type m ode l(red).Each poin t rep resen ts the average of th ree m easu rem en ts.(Fo r in terp retation of the re feren ces to co lo r in th is f igu re legen d,the reader is re ferred to the w eb version of th is article.)

To test th is hypothesis,in f luen ce of sa lt add ition on d rug solubility w as investigated.Sin ce the so lubility of BH in w ater is ex trem e ly h igh(＞1000 g/l),w h ich is d iff icu lt to accu rately quan tify,on ly the so lubility of DS and BB in d ifferen t sa lt solu tionsw asm easu red.As show n in Fig.4A,w hen DSw as used as the m ode l d rug,irrespective of sa lt type,d rug so lubility decreased sign if ican tly w ith the in crease of ion ic strength.A linear co rrelation ex isted betw een the loga rithm s of the ion ic strength and DSso lubility(r=0.9880,NaCl;r=0.9899,Na2SO4;r=0.9874,NaSCN).Th is w as in good agreem en t w ith the inf luen ce of ion con cen tration on d rug load ing in CLs,as p resen ted in Fig.1B.The decrease of DSso lubility by sa lt add ition m igh t be exp lained by the comm on ion effect.For DS,w h ich w as a sa lt form of d iclofenac,the so lubility p rodu ct cou ld be described by:

w here‘a(chǎn)'is the activity,γ±is the m ean ion ic activity coeff icien t,D-and Na+are the con cen tration s of the cation and the an ion,respectively.Add ing NaCl/Na2SO4/NaSCN to the solu tion w ou ld in crease the con cen tration of the cation,Na+and in order to keep the K°spconstan t,the f inal resu lt w as to dep ress the d rug's aqueous so lubility.The Setschenow constan ts ca lcu lation,as p resen ted in Tab le 2,fu rther dem onstrated that salting ou t is them ain reason for decreased so lubility of DSby sa lt add ition.

Tab le 2-Setschenow con stan ts fo r Diclofenac Sod ium and Betaxo lo l Base in d ifferen t sa lt so lu tion s w ith the ion ic streng th ranging from 0.005 to 1 m o l/l.

W hen BB w as used as the m ode l d rug,apparen t in f luen ce on d rug so lubility w as observed on ly w hen the ion ic con centration w as above 0.15 m o l/l(Fig.4B).Above th is poin t,NaCl and Na2SO4add ition caused sa lting-ou t effect,w ith the so lubility of BB decreased sign if ican tly(P＜0.05),in con trast,NaSCN led to sa lting-in effect,w ith statistica l in crease in d rug so lubility(P＜0.05),as dem onstrated by the calcu lated Setschenow constan ts(Tab le 2),w ith negative va lue in NaSCN added system.These data are in good agreem en tw ith the inf luen ce of ion con cen tration on d rug load ing in CLs,as p resen ted in Fig.2B,w ith decreased so lubility correspond ing to in creased d rug load ing in CLs.

3.2.3. Inf luence ofsalts on w ater content ofcontact lenses

W ater content is an im po rtan t param eter re lated to eye comfort.Previous stud ies show that d rug load ing in soft CLs m ain ly depends on thew ater contentand in teraction betw een d rugs w ith po lym er com ponen ts[28].Fo r d rugs w ithou t in teraction w ith CLs,h igh content of w ater in CLs w ou ld d isso lve m ore d rugs and then a h igh d rug load ing cou ld be ach ieved[29].In add ition to d rug so lubility change,to testw hetherw ater content variation in CLs con tribu ted to d rug load ing d ifferen ce w hen sa lt so lu tions w ere used,and to get a deep understand ing of the in f luence of salts on the w ater content of CLs,bo th to ta l w ater and bound w ater of con tact lens w ere m easu red using sa lt so lu tion as so lven t.

First of a ll,to e lu cidate the e ffect of sa lt con cen tration on w ater content of CLs,in f luence of NaCl concen tration on p-HEMA CLs w ater contentw as investigated and the resu lts are p resen ted in Fig.5A.W hen NaCl con cen tra tion w as 0.005 o r 0.15 m o l/l,no statistica l d ifferen ces in both to ta l w ater conten t and bound w ater content w ere found com pared w ith the d istilled w ater group(P＞0.05),therea fter,fu rther increasing NaCl con cen tration to 1 m o l/l caused sign if ican t decrease of total w ater content(P＜0.05),w ith free w ater alm ost com p letely rem oved.As fo r boun d w ater,it decreased from 55.3%in 0.15 m o l/l NaCl group to 42.5%in 1 m o l/l NaCl group(P＜0.01).Those da ta dem on strated that the to ta l and bound w ater content in con tact lens cou ld be ad justed via changing ion ic strength of the im m ersion so lu tion.Sim ilar trend w as observed w ith Na2SO4con tain ing system(Fig.5B),w ith w ater content un changed under low er o r physio logica l ion cond ition s(0.15 m o l/l),and sign if ican tly decreased w ater content at h igher ion con cen tration,im p lying in creased d rug load ing at low er o r physio logica l ion cond itions w as no t related to the w a ter content in the CLs.In con trast,a reverse casew as observed at h igher ion ic con cen tration w hen NaSCN w as used as the sa lt fo r ion ic strength ad justm en t,w ith the bound w ater content in creased to 210.7%at 1 m o l/l NaSCN com pared to 64.3%at 0.15 m o l/l(Fig.5C).The va rious e ffects of salt type on w ater content of CLsm igh t be exp lained by the stru ctu re-m aker o r stru ctu re-b reaker ability of sa lts used.Kosm o tropes agen ts,NaCl and Na2SO4,w ou ld bind w aterm o lecu les strongly and in crease w ater stru ctu re,and then lessw aterw ou ld have the opportun ity to be adsorbed on po lym er m atrix at h igher ion ic strength.How ever,fo r chao tropes agen ts,NaSCN,the opposite phenom enon w ou ld occu r[23].

Fig.5-W ater content of p-HEMA con tact len ses in NaCl(A),Na2SO4(B)and NaSCN(C)so lu tions w ith the ion ic streng th ranging from 0.005 to 1 m o l/l.Each data poin t rep resen ts the average of th ree m easu rem en ts and all the statistica l tests a re based on the data of boun d w ater content.(?P＜0.05,??P＜0.01,???P＜0.001).

Fig.6-Re lationsh ip betw een d rug load ing am oun t of CLs w ith apparen t aqueou s so lubility of DS(A),BB(B)in NaCl,Na2SO4 and NaSCN.Each data poin t rep resen ts the average of th ree m easu rem en ts.

3.2.4. Relationship am ong drug solubility and w ater content w ith drug loading ofcontact lenses

The above study dem onstrated that salt add ition can in f luen ce d rug so lubility in so lu tion,and at h igher ion ic con cen tration,bound w ater in CLs can a lso be in f luen ced.Then,w hat is the re lationsh ip am ong d rug so lubility,w ater content of CLs and d rug load ing in CLs?

To answ er th is question,f irst of a ll,the re lationsh ip betw een d rug so lubility and d rug load ing in CLs w as characterized.As show n in Fig.6A,a good negative co rrelation betw een DS load ing in CLs and DS so lubility in d ifferen t soak ing solu tions w as estab lished irrespective ly of sa lt type(r=0.9678,NaCl;r=0.9820,Na2SO4;r=0.9486,NaSCN).Sim ilar trend w as observed w hen BBw as used as them odel d rug(Fig.6B).

The negative relationsh ip betw een d rug load ing in CLsw ith its apparen t aqueous so lubility ind icated that d rug so lubility m igh t be an im po rtan t facto r determ in ing d rug load ing in CLs.How ever,w hen the ion ic strength w as greater than or equa l to 0.15m o l/lDSso lubility in NaClw as statistica lly low er than that in Na2SO4(Fig.7C),bu t no d ifferen ce in DS load ing betw een the tw o sa lt groups w as found(Fig.7A).It im p lied that d rug so lubility m igh t not be the on ly factor determ in ing d rug load ing in CLs,especia lly at h igh ion ic con cen tration.As show n in Fig.7E,h igher bound w ater content w as foun d in NaClbased system com pared w ith Na2SO4based system w ith the in crease of ion ic strength.Due to the imm obilization of bound w ater,itm ay be d iff icu lt to d isso lve and transfer d rug bu tm igh t decrease d rug load ing[30,31].Sin ce free w ater conten t in CLs is qu ite low,its in f luen ce on d rug load ing cou ld be ignored.Therefore,the com parab le DS load ing in NaCland Na2SO4based system m igh t be a com p rom ise betw een the in f luen ce of d rug so lubility and bound w ater content in the CLs(Fig.7E).Sim ilar phenom enon w as observed w hen NaSCN based system w as com pared w ith NaCl con tain ing system,as p resen ted in Fig.7B,D and F.

Fig.7-The va riation s of DS load ing of p-HEMA CLs(A and B),so lubility(C and D)and boun d w ater content(E and F)as a fun ction of ion ic streng th using d ifferen t sa lts.Each data poin t rep resen ts the average of th ree m easu rem en ts.

Based on the resu lts and d iscussion p resen ted above,the possib le m echan ism of sa lts e ffect on d rug load ing of CLs is schem atica lly p resen ted in Fig.8.The change of ion ic strength in soaking solu tion m igh t affect d rug load ing by affecting d rug appa ren t so lubility and a ltering bound w ater am oun t.The apparen t so lubility can be used as a param eter to ref lect the a ff in ity betw een d rug and so lven t,the h igher the solubility,the better the aff in ity.Sim ilarly,content of boun d w ater can be used to rep resen t the a ff in ity betw een CLs and so lven ts.A better a ff in ity w ill lead to h igher w ater conten ts.If w hen sa lts added,bo th the apparen t so lubility of d rug and the bound w ater content of CLs decrease,it in dicates the decrease of d rug-so lven t and lens-so lven t a ff in ity,w h ich w ill enhan ce d rug load ing in CLs synergistica lly,lik ing a com bina tion of“squeezing ou t”and“l(fā)eaving space”p rocess,resu lting in h igher d rug load ing(Fig.8B).An opposite situation w ill happen if sa lt add ition causes d rug so lubility and bound w ater content in crease,lead ing to d rug load ing decrease(Fig.8C).

Fig.8-M echan ism s of d rug load ing m odu lation in CLs by ad justing the ion ic streng th of lead ing so lu tion s.(A)d rug load ing in w ate r;(B)load ing enhan cem en t caused by sa lt;(C)load ing decrease caused by sa lt.

3.3. Feasibility of im proving CLs drug loading by adjusting the ionic strength of loading solutions

Tofu rther investigate w hether th is sim p lem ethod is also app licab le w hen crosslinker am oun t or the com positions of CLs are changed,pHEMA CLs w ith d ifferen t degree of crosslinking(0.1%;0.5%;1%;2%,w/w),pHEMA CLs w ith various fun ctionalm onom ers(VP or MAA,2%,w/w),silione-hyd rogel CLs(HEMA/NVP/TRIS,20/40/40,w/w/w)w ere p repared and DS load ing via im m ersing CLs in 0.5 g/l DS-w ater or 0.5 g/l DS-0.15 m o l/lNaCl so lu tion w as investigated.As show n in Fig.9A acco rd ing to the va lue of Ei,ad justing ion ic strength of the load ing so lu tion im p roved d rug load ing sign if ican tly compared w ith d istilled w ater based system(P＜0.05),bu t the exten t of Eivalue in crease w as CL p roperty dependen t.A lthough degree of cross link ing had an in f luen ce on d rug loading,there w as no de f in ite dependen ce of d rug load ing on crosslink density and th is w as in agreem en t w ith p revious reports[18,19].

In any case,the ligh t transm ittan ce of CLs shou ld not be in f luen ced by the d rug load ing p rocess.To test the po ten tia l in f luen ce of sa lt-enhan ced d rug load ing on the app lication of CLs,the ligh t transm ittan ces of HEMA based CLs w ere tested a fter im m ersed in w ater,0.5 g/l DS,0.5 g/l DS-0.15 m o l/l NaCl,and 0.5 g/lDS-1m o l/lNaCl fo r 3 d(Fig.9B),no apparen t change in ligh t transm ittance w as observed after salt add ition.Ligh t transm ittan ce of p-HEMA CLsw ere above 95%at 500 nm in a ll cases.

Tof igu re ou t w hether changing ion ic strength of loading so lu tion had an in f luence on d rug release p rof ile,DS release from CLs p repared w ith 0.5 g/l DS so lu tion using w ater,0.15 m o l/l o r 1m o l/l NaCl as so lven ts w as investigated in STF and the resu lts are show n in Fig.9C.A lm ost 80%of DS w as released in the f irst 2 h and th is is in agreem en t w ith p revious repo rts[29].Sligh tly h igher d rug release w as found at tim e poin ts 0.5 h and 1 h fo r 0.15 m o l/l NaCl based system,bu t no statistica l d ifferen ce betw een the release p rof iles of d istilled w a ter and 0.15 m o l/l NaCl based system s w as foundSim ilarly,DS release from 1 m o l/l NaCl based system w as com parab le w ith that from w ater based system

4. Con clu sion s

In th is paper,it w as found that ad justing the ion ic strength of load ing so lu tions resu lted in sign if ican t in crease of d rug load ing in CLs.Type and con cen tration of the sa lts,so lubility of the d rug w ere the m ain factors in f luencing d rug load ing.The m echan ism fo r im p roved load ing w as likely related to the redu ced d rug so lubility in load ing so lu tions and the reduced bound w ater content in con tact lenses.M odulation of d rug loading by ad justing ion ic strength w as also app licab le to other CLs and the ligh t transm ittan ce w as not a ffected.Th is m ethod w as m o re su itab le fo r sa lt-fo rm d rugs w ith h igh so lubility.Sin ce d rugs in sa lt fo rm accoun t fo r a considerab le p ropo rtion in oph tha lm ic d rugs on the m arket,a broad app lication of th is m ethod can be expected and fu rther exp lored.In sum m ary,it w as dem on strated that ad justing ion ic strength of load ing so lu tion m ay be an econom ical and effective w ay to im p rove d rug load ing in CLs.Th is simp lem ethod can a lso be usefu l fo r o ther d rug delivery system s based on hyd roge l.The num ber of d rugs used in th is study w as qu ite lim ited and m o re d rugs shou ld be used to verify the feasibility of th ism ethod in the near fu tu re.

Fig.9-DS load ing am oun t(left Y)and Ei(righ t Y)of 0.15 m o l/l NaCl using d ifferen t com position con tact len ses(A);Ligh t transm ittan ces of p-HEMA con tact len ses using w ater;0.5 g/l DS;0.5 g/l DS-0.15 m o l/l NaCl o r 0.5 g/l DS-1 m o l/l NaCl as im m ersion so lu tion,respective ly(B);DS re lease from CLs loaded w ith 0.5 g/l DS in w ater,0.15 m o l/l NaCl o r 1 m o l/l NaCl separate ly using STF as re lease m ed ium.The in serting g raphs rep resen t d rug re lease in 2 h(C).

Con f licts of in terest

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

Acknow ledgm en ts

Th is w ork w as supported by the Distingu ished Professor Pro ject of Liaon ing Provin ce(2015).

Sup p lem en tary m ateria ls

Supp lem en tary m ateria l associated w ith th is article can be found,in the on line version,at doi:10.1016/j.a jps.2018.05.002.