for transdermal delivery of differently sized FITC dextrans The effect of pulse duration, power and energy of fractional Er:YAGlaser International Journal of Pharmaceutics

The effect of pulse duration, power and energy of fractional Er:YAG laser for transdermal delivery of differently sized FITC dextrans

Barbara Zorec

^a

, Dejan Š krabelj

^b

, Marko Marin 9 cek

^b

, Damijan Miklav ci 9 c 9

^a

, Nata š a Pav š elj

^a,

*

aUniversityofLjubljana,FacultyofElectricalEngineering,Tržaška25,SI-1000Ljubljana,Slovenia

bFotonaLLC,Stegne7,1000Ljubljana,Slovenia

ARTICLE INFO

Articlehistory:

Received15July2016

Receivedinrevisedform5October2016 Accepted25October2016

Availableonline3November2016

Keywords:

Transdermaldrugdelivery Fractionallaser

Er:YAG Pulseduration Energymodulation

ABSTRACT

WestudiedfractionalEr:YAGlasertoenhancetransdermaldrugdeliveryofcompounds possessing differentmolecularweights:FITC-dextrans(orFD)withaveragemolecularweightsof4,10and20kDa.

VerticalglassFranzdiffusioncellswereusedtostudymoleculartransportthroughdermatomedporcine skinandhistologicalanalysisoflaser-treatedskinwasperformedaftertreatmentwithdifferentlaser pulseprotocols.Wewerecomparingdifferentpulsedurationsatconstantorvaryingpulseenergies.We found that the energy of the delivered pulses mostly dictates the size/depth of laser-created microchannels,whilethedurationofthepulsesdictatestheextentofthermallyalteredtissue.That is,tissueablationthresholdisloweredatshorterpulsedurationswithhigherpower,whichispreferredas itlowersthermaleffectsonviableskinlayers.Especiallyforsmallermolecules,transdermaldeliveryis increasedbyincreasinglaser-createdmicrochannelsize, butalsobymakingpartitioningintotissue easierwhenlessthermaldamageiscausedontissue.Forlargemolecules,moleculartransportthrough theremainderofskintissuebecomesincreasinglydifficultregardlessoflaserpulseparameters.

ã2016ElsevierB.V.Allrightsreserved.

1.Introduction

Inthepastdecadestheuseoflasershasbecomecustomaryin many areas of medicine, such as dermatology, ophthalmology, surgery,dentistry,oncology,eitherfordiagnosisortreatment.The earliestandcurrentlymostwidespreaduseoflasersisforvarious applicationsindermatology,suchasacneorskinhyperpigmenta- tion treatment, scar and wrinkle reduction, tattoo, hair and birthmarkremoval,skinrejuvenationbyresurfacing,skincancer diagnosisandtreatment, anddermal/transdermaldrugdelivery.

Thelatterhasbecomepopularnotonlyfortreatmentswhereskin is the target tissue (dermal) but also for systemic delivery (transdermal)becauseofitsnoninvasiveness,avoidanceof liver or gastrointestinal metabolism and good acceptance from the

patients(Zorecetal.,2013b).However,theskin’souterbarrier,the stratumcorneum,representsthegreatestresistancetomolecular transport, rendering skin practically impermeable to topically applied drugs. Transdermal drug delivery methods focus on overcoming this barrier function of the stratum corneum to increase transdermal delivery and include passive (penetration enhancers,liposomes,nanoparticles,patchtechnology)andactive methodsofenhancement(electroporation,iontophoresis,micro- needles, ultrasound, laser radiation). One of these active approachesinvolvespartialortotallaserablationofthestratum corneum,whichsubstantiallyincreasesskinpermeabilityevenfor hydrophilic molecules and large molecules such as peptides, proteinsandDNA(Sklaretal.,2014).

TheCO₂(carbondioxide)andtheEr:YAG(erbium-dopedyttrium aluminium garnet) laser are most frequently used lasers in transdermaldrugdelivery(Baronetal.,2003;Fangetal.,2004a, 2004b;Gómezetal.,2011;Hsiaoetal.,2011;Huangetal.,2013;Lee etal.,2001,2002,2003,2006,2007,2008b,2008a,2009;Shenetal., 2006;Wangetal.,2004;Yunetal.,2002).Themostfrequentlyused Er:YAGlaseremitslightinthenear-infraredpartofthespectrum,at 2.94

m

m wavelength, characterized by exceptionallyhighabsorption bywaterinbiologicaltissue,whichmeansthatthelightdoesnot penetrateverywell.Thisresultsinverysmallopticalpenetration depthsandintissueablationwithminimalthermaldamageand Abbreviations:FITC,fluoresceinisothiocyanate;FD,FITClabeleddextran;FD4,

FITClabeleddextranofaveragemolecularweight4kDa;FD10,FITClabeleddextran of average molecular weight 10kDa; FD20, FITC labeled dextran of average molecularweight20kDa.

*Correspondingauthor.

E-mailaddresses:barbara.zorec@fe.uni-lj.si(B.Zorec),

dejan.skrabelj@fotona.com(D.Škrabelj),marko.marincek@fotona.com (M.Marin9cek),damijan.miklavcic@fe.uni-lj.si(D.Miklav9ci9c), natasa.pavselj@fe.uni-lj.si(N.Pavšelj).

http://dx.doi.org/10.1016/j.ijpharm.2016.10.060 0378-5173/ã2016ElsevierB.V.Allrightsreserved.

ContentslistsavailableatScienceDirect

International Journal of Pharmaceutics

j o u r n al h o m ep a g e: w w w . el s e v i e r . c o m / l o c at e / i j p h a r m

controlled coagulation zones, making it perfect for controlled removalofstratumcorneum(BerlienandMüller,2003).Conven- tionalCO2andEr:YAGlaserswereusedtoablateouterskinlayers with uniform beam profiles, therefore the treated surface was irradiateduniformly.However,inthepast5yearsuniformbeam profilesusedinskintreatmentsincludingtransdermaldrugdelivery arebeingreplacedbytheconceptoffractionalphotothermolysis,in whichthelaserbeamissplitintosmallerbeams.Theskinisnolonger ablatedhomogeneously,insteadonlysmallpartsofskinsurfaceare exposed to intense laser microbeams, resulting in microscopic channels through outer skin layers, surrounded by undamaged tissue, which allows for much faster skin healing and faster restoration of skin impermeability (Farkas et al., 2010; Graber etal.,2008;Sklaretal.,2014).

Lasertreatment for transdermal drugdelivery enhancement takesadvantageoftheablativeeffectsoflaserlightontissue.The Er:YAGlaser isveryeffectiveforcontrolledremovalofthethin deadouterlayeroftheskin,thestratumcorneum,asitexertsits effects ontissue withminimal penetration depthand minimal thermaldamage,thussparingviableunderlyinglayers(epidermis, dermis)fromanysubstantialthermaldamage.Inaddition,iflaser beam is split into microbeams (fractional lasers), even larger portion of viable skin tissue is spared, which offers a very successful yet safe transdermal drug delivery enhancement method.A literature survey confirms increased interest in this modality during the last five years, with a number of studies showingeffectivenessoffractionallasers,eitherCO2(Chenetal., 2012;Haaketal.,2012;Haedersdaletal.,2010,2011;Hsiaoetal., 2012;Leeetal.,2013,2014b)orEr:YAG(Bachhavetal.,2010,2011, 2013;Forsteretal.,2010;Geninaetal.,2011;Leeetal.,2010,2011, 2014a;Onietal.,2012;Terentyuketal.,2012;Weissetal.,2012;Yu etal.,2011),fortransdermaldeliveryofvariouscompounds.The applicationsincludeskinrejuvenation(vitaminC,hydroquinone), delivery of anesthetics (lidocaine), anti-inflammatory drugs (diclofenac), treatments for skin conditions and neoplasms (imiquimod, photosensitizers for photodynamic therapy), even vaccination,aslargermoleculessuchasproteinsorDNAcanbe deliveredsuccessfully.Pulsesofdifferentpulseenergies,fluences and durationshavebeenused, withdifferentlyfractionedlaser beam creating microchannels of various sizes and numbers.

However, we have not found any study directly comparing differentpulsedurationsat constant or varyingpulseenergies, assessing the effect of different pulse shapes on skin and consequentlyonmoleculardelivery.Tissueablationthresholdis loweredatshorter pulsedurationswithhigherpower,whichis preferredasit lowersthermal effects onviableskinlayers.We experimented with Fotona LightWalker Er:YAG laser system (FotonaLLC,Ljubljana,Slovenia),withafractionalhandpiecePS- 01,abletomodulatepulsedurationatconstantdeliveredenergy, thereby changing the instantaneous power. The transdermal delivery after treatment with different laser protocols was assessedwithFITC-labeleddextransofdifferentsizes.

2.Materialsandmethods 2.1.Skinpreparation

Dermatomed porcine ear skin was used in the study, as described previously (Zorecet al., 2013a).Briefly, the skinwas obtained from local slaughterhouse immediately post-mortem (FarmeIhanLLC,Šentjur,Slovenia),beforewashingwithsteamand detergentinordernottocompromiseskinintegrity.Skinsamples werepreparedwithinthreehourspost-mortem.Excesssubcuta- neousfatwasremoved,andtheskinwasdermatomedto350

m

^m

thickness(TCM3000BL,Nouvag¹).Preparedskinsampleswere frozenuntiluse,fornolongerthanonemonth.

2.2.Chemicals

Sodiumchlorideandfluoresceinisothiocyanatelabeleddextran (FITC-dextranorFD)withaverageMWsof4(FD4),10(FD10)and 20kDa(FD20)werepurchasedfromSigma-Aldrich(St.Louis,MO, USA).WithregardtomolecularweightsofFITC-dextranswewould liketoemphasizethateventhoughaverageMWisreported,there is actually a size distribution associated with each dextran.

However,themanufacturerstatesthatlowerMWdextransexhibit amorenarrowrangeofMWdistribution,whiledextranswithMW greaterthan10000aremorebranchedandhavesomewhatwider sizedistribution.Potassiumchloride,di-sodiumhydrogenphos- phateandpotassiumdihydrogenphosphatewerepurchasedfrom Merck(Darmstadt,Germany).PBSbuffersolutionwaspreparedin bi-distilled water (8g NaCl, 0.2g KCl, 1.44g Na2HPO4, 0.24g KH2PO4in1lofwater),aswellasphosphatebuffersolution(7.38g NaH2PO4,4.77gNa2HPO4in1lofwater).

2.3.Generalexperimentalsetup

VerticalglassFranzdiffusioncellswereusedtostudymolecular transport through excised and dermatomed porcine skin. The temperature of the chamber was regulated at 37C by water circulation. A piece of porcine dermatomed skin was placed betweentwocompartmentsrightbeforelasertreatmentwiththe stratumcorneumfacingthedonorcompartment.Theareaofskin availablefordiffusionwas0.785cm².Thereceivercompartment (3.1ml)wasfilledwithPBS(pH7.4,150mM),inordertomaintain constantpHandosmolarityaswellastomatchionconcentration tothatofhumanbody.Thedonorcompartmentcontained1mlof FDsolution(0.1mM)inphosphatebuffer(pH6.5,100mM).

2.4.Laserdevice

Er:YAG (2940nm, FotonaLightWalker, FotonaLLC,Ljubljana, Slovenia;seeFig.1a)laserwithafractionalhandpiecePS-01was usedtopretreatporcineskinsamples.Laserbeamwasdividedinto 33squaregridpoints(thesideofthesquaremeasured3.1mm) andthediameterofeachmicrobeamwasapproximately400

m

^m

(Fig.1b).Fractionaldistributionofenergywas achievedusinga diffractive optical element (DOE) and the condenser lens. The durationofthepulsecanbesettothenominaldurationsof50

m

^s,

100

m

^{s, 250}

m

^{s, 500}

m

^{s and 1000}

m

^{s. Commercial names of}

different duration laser pulses are: SSP (super short pulse:

50

m

^{s), MSP (medium short pulse: 100}

m

^{s), SP (short pulse:}

250

m

^{s)LP(longpulse:500}

m

^{s)andVLP(verylongpulse:1000}

m

^s).

2.5.Experimentalprotocols

Weexperimented withSSP,MSPand SPpulses(50–250

m

^s).

Outputenergiesrangedfrom80to380mJ/pulse(all9microbeams combined).With9microbeamsof400

m

^{mdiameter,weachieved}

fluences(definedaslaserenergypersurface)rangingfrom7.1to 33.6J/cm².Inthefirstpartofthestudy,wevariedpulseduration while keeping pulse energy constant: shorter pulses of equal energymeanshigherpeakpower.RelativepeakpowersofSSP,MSP andSPpulsesassumingequalenergiesare1:0.5:0.2.Atmaximum energysettingofthelaserthatweused–380mJ–thepeakpowers ofthepulseswere:SSP:7.6kW,MSP:3.8kW,SP:1.52kW(peak powerspermicrobeamhavetobedividedby9:SSP:0.84kW,MSP:

0.42kW,SP:0.168kW).Inthesecondpartofthestudy,wekeptthe durationofthepulsesconstantandvariedthedeliveredenergy.

We experimented with the shortestpulse setting, the SSP. All protocolsarelistedinTable1.Weperformed8parallelrepetitions foreachprotocol.

2.6.Transdermaltransportmeasurements

The concentrationofFITC-dextransin receivercompartment after laser treatment was measured with spectrofluorometer (Jasco, FP-6300).Ateach sampling time, a sample of 300

m

^lof

receiversolutionwas takenforconcentrationmeasurementand replacedwithfreshPBS.Concentrationwascalculatedaccordingly.

Wemonitoredthemoleculardeliveryinthereceiversolutionevery hourforfivehours,andmorefrequentlyduringthefirsthour.

2.7.Skinhistologicalexamination

The region of the skin exposed tothe laser treatment was excised, fixed in formalin then stored in 70% ethanol until embeddinginparaffin. Subsequently, 5

m

^{mthicksections were}

cutinthedirectionperpendiculartoskinlayersandstainedwith hematoxylinandeosin.Thepreparedslides wereobservedwith BX-51microscope(Olympus,Hamburg,Germany)equippedwitha digitalcameraDP72(Olympus).

2.8.Dataanalysis

Theresultsareexpressedasthemeanstandarddeviationof themean(normalitytestpassedinallinstances).Thestatistical

analysisofdifferencesbetweenvarioustreatmentswasperformed usingTukeytest.A0.05levelofprobabilitywastakenasthelevel of significance. Statistical significance between different laser pulse protocols was analyzed for the 5-h time point after the treatment.

3.Resultsanddiscussion

Theeffectsofdifferentlasersonthebiologicaltissuedependon thewavelength,averagepower,energy,spotsizeandtheexposure timeandaredictatedbytissueopticalproperties(theydetermine distributionoflightinsidethetissue)andthermalproperties(after lightenergyisconvertedintoheat,theydetermineitsconduction in tissue). Whenhitting biological tissue, a partof thelight is directlyreflectedatthesurfaceorindirectlyafterscatteringinside thetissue.Thedistributionoflightinsidethetissueisdetermined by three main processes:absorption(byvarious tissue-specific chromophores),scattering(deviation fromstraight trajectory at sitesofparticlesandotherirregularities)andrefraction(changein direction of light propagation due to a change in propagation medium). Tissue reactions to light can be classified into non- thermal,thermal,ablativeandoptomechanical.Forlongerexpo- suretimes(500msorlonger)atmoderatepowerdensities(1W/

cm²range),tissueiscoagulated(tissuetemperaturesover50C).

Table1

Experimentalprotocols.

EXPERIMENTALMOLECULES LASERPROTOCOLS

(pulseduration;energyofthewholebeam;powerofthewholebeam;fluence) FITC-dextrans:

4kDa, 10kDa, 20kDa

varyingpulsedurationat constantpulseenergy

SSP:50m^{s;380mJ;7.6kW;33.6J/cm2} MSP:100m^{s;380mJ;3.8kW;33.6J/cm2} SP:250m^{s;380mJ;1.52kW;33.6J/cm2} varyingpulseenergyatconstantpulseduration SSP:50ms;80mJ;1.6kW;7.1J/cm²

SSP:50ms;230mJ;4.6kW;20.3J/cm² SSP:50m^{s;380mJ;7.6kW;33.6J/cm2} Fig.1.a)FotonaLightWalkerfractionalEr:YAGlaser(FotonaLLC,Ljubljana,Slovenia);b)handpieceandthegridofthefractionatedlaserbeam:thediameterofeach microbeamisapproximately400m^{m,arrangedintoa33formation(thesideofthesquaremeasures3.1mm).}

Whenexposure timeis reduced tomillisecond range athigher power(10kW/cm²range),tissuetemperaturereaches100Cand morewhichcausesvaporizationduetolocaloverheating.Ablative effectoflaserlightoccuratmicrosecondrange(theprocessstarts after a few microseconds up to 500

m

^{s) at even higher power}

output(107W/cm²)andturnintoplasmaformationand optical breakdownoftissueattheextremeendofparameterranges(ns exposuretimeandover107W/cm²powerdensity)(Berlienand Müller,2003).

A significant advantage of the use of lasers to enhance transdermaldrugdelivery istheablative effectoflaser lighton tissue with minimal laser beam penetration depth and conse- quentlyminimalthermaldamageofviableunderlyingskinlayers (epidermis,dermis),whileensuringmaximumtransdermaldrug transport.Forfuture clinicaluse, thismeanshighefficacy,good safetyprofileandpotentialacceptancebythepatients.

OverseveralyearsofexperimentationEr:YAGlaserwasfound effectiveforcontrolledandsaferemovalofthestratumcorneum, leadingtoenhanced transdermaldelivery ofvariousmolecules, evenlargeonessuchasproteins orDNA (Bachhav etal.,2013;

Weissetal.,2012; Yuetal., 2011).Inaddition,when fractional lasersareused,evenlargerportionofviableskintissueisspared (Farkas et al., 2010; Graber et al., 2008; Sklar et al., 2014). In previous studies, pulses of different energies, fluences and durationswereassessed(Bachhavetal.,2010,2011,2013;Forster etal.,2010;Geninaetal.,2011;Leeetal.,2010,2011,2014a;Oni etal.,2012;Terentyuketal.,2012;Weissetal.,2012;Yuetal., 2011);however,wehavenotfoundanystudydirectlycomparing differentpulsedurationsat constant or varyingpulseenergies.

Namely, pulse energy required for the ablation of the stratum corneumistheoreticallyloweredatshorterpulsedurationswith higherpower,whichalsolessensthermal effects onviable skin layers. Our aimwas to assess how changing these parameters affectsmoleculardeliveryofthelasertreatedskin.

Our studyconsistsof two parts(twotypes of protocols):a) protocolsamongwhichpulsedurationwasvariedwhilekeeping pulse energy constant; and b) protocols among which pulse durationwaskeptconstantbuttheirenergywasvaried.

3.1.Varyingpulsedurationatconstantpulseenergy

Toassesstheeffectofpulseduration/poweratconstantenergy onlaser-enhancedtransdermalpenetrationofpermeants,weused differently sized fluorescently-labeled macromolecules: FITC- dextrans (FD): 4kDa, 10kDa and 20kDa. The energy of the deliveredpulseswasfixedat380mJ(fluence:33.6J/cm²).Pulsesof three different parameters were used: SSP (super short pulse:

50

m

^{sduration),MSP(mediumshortpulse:100}

m

^{sduration)and}

SP (short pulse: 250

m

^{s duration), with their peak powers (at}

380mJ energy) amounting to: SSP: 7.6kW, MSP: 3.8kW, SP:

1.52kW.

Fig.2presentscumulativeamountofallFITC-dextransinthe receivereveryhourfor5hofpassivediffusionafterthetreatment.

TheresultsinFig.2a(FD4),b(FD10)andc(FD20)aregiveninpmol/

cm²,however,pleasenotedifferentrangesforeachpanelofFig.2.

Statisticalsignificance(p0.05)betweendifferentlasertreatment protocolsatthe5-htimepointafterlaserexposureismarkedwith anasterisk.Ascanbeinferredfromthefigures,theamountofthe FDinthereceiverat5hofpassivediffusionaftertreatmentwas significantlyhigherinthelaser-treatedgroupswhencomparedto control(passivediffusiononly,nolasertreatment),forallpulse durationsandFDsizes(asterisksdenotingstatisticalsignificance between laser treatment protocols and the control (passive diffusion)wereleftoutforclarity).Asuntreatedstratumcorneum representsasignificantbarriertotransportoflargemoleculessuch as FD, this is not surprising.Further, as couldbe expected,FD

deliverythroughtreatedskindropswithFDsize(amountinthe receiver::4kDa>10kDa>20kDa).

However,themostinterestingarethedifferencesinFDdelivery whendifferentpulsedurations/powersareused.LookingatFig.2a, itisevidentthattheamountofdeliveredFD4isthelowestwhen thelongestpulse(SP:250

m

^{sduration)isused,whencomparedto}

shorter pulses (MSP: 100

m

^{s, SSP: 50}

m

^{s), even at the same}

delivered pulse energy. Comparison of protocols at 5h after treatmentshowednostatisticalsignificancebetweenSSPandMSP protocols,whiletheefficacyofSPprotocolissignificantlylower thanboth,SSPand MSPprotocols.Further,thedeliveryofFD10 (Fig.2b)isthelowestwhenthelongesttwopulseprotocols(SP:

250

m

^{s and MSP: 100}

m

^{s duration) are used (no} statistically Fig.2. Cumulativeamount(inpmol/cm²)ofallFITC-dextrans(FD)inthereceiver for5hofpassivediffusionafterthelasertreatment:panela)FD4;panelb)FD10;

panelc)FD20.Theenergyofthedeliveredpulses(ofthewholebeam)wasfixedat 380mJ,whiletheirduration/peakpowerdiffered:SSP(supershortpulse:50ms duration, 7.6kW power), MSP(medium short pulse: 100m^{s duration,3.8kW} power)andSP(shortpulse:250m^{sduration,1.52kWpower).Notedifferentranges} foreachpanel.Statisticalsignificancebetweendifferentlasertreatmentprotocols atthe5-htimepointafterlaserexposureismarkedwithanasterisk.A0.05levelof probabilitywastakenasthelevelofsignificance.Asalllasertreatmentprotocols were significantly more efficient than passive diffusion, asterisks denoting statisticalsignificancebetweenlasertreatmentprotocolsandthecontrol(passive diffusion)wereleftoutforclarity.

significantdifferencebetweenthetwoprotocols),andsignificantly higherafterSSPprotocol(50

m

^s).

Thishigherefficacyofshorterduration/higherpowerprotocols maybetheirlowerthermalinfluenceonthetissue.Namely,typeof tissuereactionastheeffectoflaserpulsesdependsontheamount and time courseof the temperaturerise thesepulses produce.

Generally,differentreactionzonescanbeobservedinthetissue afterlaserirradiation.Thetissueisremovedintheablation(non- thermal effect) and vaporization zones (thermally removed at temperatures over 300C) that will befollowed by carbonized tissue zone (temperatures over 150C), then coagulated tissue (over60C)(BerlienandMüller,2003).Aspermeantpartitioning intothermally-alteredtissue islesser,thesezonesneedtobeas narrowaspossible.Whenshorter pulseswithhigherpowerare used, higher ablative action is achieved, reducing unwanted thermal effects on tissue thus increasing molecular delivery troughlaser-createdchannels.

ThetrendsofenhancementefficiencyforFD4andFD10seemto suggestthat,ingeneral,shorterpulses(higherpeakpulsepower) aremoreefficientthanlongerpulses(lowerpeakpulsepower):

SSP (super short pulse: 50

m

^{s)>MSP (medium short pulse:}

100

m

^{s)>SP(shortpulse:250}

m

^{s),however,thesetrendsarenot}

alwaysstatisticallysignificant.Interestingly,forthedeliveryofthe smallest molecule – FD4 – SSP and MSP pulses were equally efficient,whileforthelargerFD10,theshortestpulses(SSP)were more efficient than both, SP (the longest), as well as MSP (medium);thedifferencewasstatisticallysignificant.Thereason forthis differenceis probablyeasierpartitioningof thesmaller molecule(FD4)throughthethermallycarbonatedandcoagulated partsofthelaser-createdmicrochannels,comparedtothelarger FD10.At50

m

^{spulseduration(SSP)thisthermallymodi}fiedpartis very thin and does not hinder molecular delivery. Conversely, whenlonger,100

m

^{spulses(MSP)areused,thiscoagulatedzoneis}

deeper.Wepostulatethatthisprobablymeanslessobstructionfor thepenetrationofthesmallermolecule(FD4)and moreforthe largerFD10

LookingatFig.2c,this“theshorter,thebetter”trendcannotbe observed for the largest molecule, FD20. While there is no statisticallysignificantdifferencebetweenSSPandSP,themedium duration–MSP–pulsesseemtobethemostefficient forFD20 delivery(somestatisticallysignificantimprovementoverboth,SSP andSP).However,we couldfindnologicalexplanationfor this somewhat higher efficiency of MSP protocol. The overall low deliveryofFD20canbeattributedtothelargemolecularsizeof FD20thatmakes partitioninginto intactorthermally modified tissueequallydifficult.

Astheshortestduration/highestpowerpulseparametersetting atconstantdeliveredenergyseemstobethebestpickfortherange ofmolecularsizesweexperimentedwith, wechoseSSP (50

m

^s

duration)pulsesforthenextpartofourstudy,whereweassessed theeffectofdifferentpulseenergiesonmoleculardelivery.

3.2.Varyingpulseenergyatconstantpulseduration

Inthissecondpartofthestudyweusedsupershortpulses(SSP;

50

m

^{s)ofthreedifferentlaseroutputenergies:80,230and380mJ/}

pulse (corresponding to 1.6, 4.6 and 7.6kW peak powers, respectively). Fig. 3 presents cumulative amount of all FITC- dextrans in the receiver for 5h of passive diffusion after the treatment.The concentrationsin Fig.3a (FD4),b (FD10) and c (FD20)aregiveninpmol/cm²(pleasenotedifferentrangesforeach panel).Statisticalsignificance(p0.05) betweendifferentlaser treatmentprotocolsatthe5-htimepointafterlaserexposureis markedwithanasterisk. Again,moleculardeliverywas signifi- cantly higher in the laser-treated groups when compared to control(passivediffusiononly,nolasertreatment),forallpulse

energiesand FDsizes(asterisksdenotingstatisticalsignificance between laser treatment protocols and the control (passive diffusion) were left out for clarity). Looking at Fig. 3a and Fig.3b,thereseemstobeaprevailingtrend(thedifferencesare statisticallysignificant,seeasterisksinthegraphs):thehigherthe energy,thelargerthemoleculardelivery,whichisnotsurprising.

However,increaseddeliverywithincreasedpulseenergyseemsto bemoreproportionalforthesmallermolecule,FD4thanthelarger one,FD10.Inotherwords,whentheenergyofthedeliveredpulse isincreasedfrom80to230–380mJ,thedeliveryofFD4fromthe donorintothereceiversolutionisincreasedbyapproximatelythe sameamount.ThisproportionalityislesserforFD10wherelarger enhancementinmoleculardeliveryisobservedwhentheenergyis increasedfrom230to380mJthenwhenenergyisincreasedfrom Fig.3.Cumulativeamount(inpmol/cm²)ofallFITC-dextrans(FD)inthereceiver for5hofpassivediffusionafterthelasertreatment:panela)FD4;panelb)FD10;

panelc)FD20.Thedurationofthedeliveredpulseswasfixedat50msduration, whiletheirenergy/peakpower(ofthewholebeam)differed.Threesettingswere used:80mJ,230mJ,380mJ(correspondingto1.6,4.6and7.6kW,respectively).

Notedifferentrangesforeachpanel.Statisticalsignificancebetweendifferentlaser treatmentprotocolsatthe5-htimepointafterlaserexposureismarkedwithan asterisk.A0.05levelofprobabilitywastakenasthelevelofsignificance.Asalllaser treatment protocols were significantly more efficient than passive diffusion, asterisksdenotingstatisticalsignificancebetweenlasertreatmentprotocolsand thecontrol(passivediffusion)wereleftoutforclarity.

80to230mJ.For thelargestmolecule– FD20–increasingpulse energyfrom80to230–380mJhasnoeffectonthedeliveryasthere arenostatisticallysignificantdifferencesamongtheprotocols,or evenanoticeabletrend.

Asthesizeofthelaser-createdchannelsinskinincreaseswith amountofdeliveredenergy,largermoleculardeliveryafterhigh- energypulseswasexpected.However,theresultsseemtosuggest thatmoleculardeliveryisnotsimplyproportionalwithpathway size. Instead, different pulse energy thresholds (and with that channel sizes) are required for differently sized molecules to achieveoptimaldeliverythroughthelaser-createdchannels.This isparticularlyevidentfromtheresults forthelargestmolecule, FD20(Fig.3c)where,althoughthedeliverythroughlaser-treated skinisenhancedandthepermeantisabletopassthroughlaser- createdmicrochannels,thesizeofthesedeliverypathwayshasno effectonthemagnitudeofdeliveryenhancement.

Foreasierassessment ofhowlaser pulseparameterchanges affecttransdermaldeliveryofdifferentlysizedFITC-dextrans,their cumulativeamountsinthereceiverfor5hofpassivediffusionafter thetreatmentforallpulseprotocolsarecompiledagaininFig.4 (panela:FD4,panelb:FD10,panelc:FD20).Changingtheduration of the pulse (at constant energy) and with that the extent of thermallymodifiedtissueseemstohavelargereffectthanvarying the size of the microchannels (by changing pulse energy) for transdermal delivery of the smallest molecule, FD4 (Fig. 4a).

Namely,increasingpulsedurationfrom100to250

m

^sat380mJ

pulseenergylowersFD4transdermaldeliverymorethanreducing pulseenergyfrom380to230mJattheshortestpulseduration 50

m

^{s.Further,forFD10,threeoutof}fivepulseprotocolsleadto practically equal molecular delivery: 250

m

^{s/380mJ, 100}

m

^s/

380mJ and 50

m

^{s/230mJ. The most ef}ficient protocol for FD10 deliveryseemstobethe50

m

^s/380mJcombination,whiletheleast efficientis50

m

^s/80mJ(statisticallysignificantdifferences).Lastly, forthelargestmolecule,FD20,theresultsofallfivepulseprotocols areveryclosetooneanother,withnodiscernabletrend.Theonly protocol statistically different from the rest is 100

m

^s/380mJ

combination;again,wecouldfindnologicalexplanationforthis seeminglyhigherefficiencyofthisprotocol.

Insummary,changesintwocharacteristicsofthelaser-created microchannelsseemtobecomenoticeablewhenchangingpulse parameters:i)thesizeofthecreatedmicrochannels,forwhichthe energyofthedeliveredpulsesisthemostimportantparameter, andii)theextentofthermally-alteredtissuethatcan–atconstant energy–becontrolledpredominantlybydifferingpulseduration/

powercombination.Forsmallermolecules,bothseemtoaffectthe outcome of transdermal delivery enhancement significantly.

Namely, transdermal delivery of the permeant is increased by increasinglaser-created microchannelsize, but also bymaking partitioningintotissueeasierwhenlessthermaldamageiscaused ontissue.Increasingmolecularsize,itbecomesmoredifficultto separatetheeffectsofthesetwomechanisms.Forlargemolecules, transport through skin tissue becomes increasingly difficult regardless of the properties of thelaser-created microchannels through the outermost skin layers. However, it needs to be emphasizedthatthisholdstruefortherangesofparameterswe experimentedwith.Namely,aseachmoleculeweexperimented withexhibitsdifferentchangesintransdermaldeliverygoingfrom oneexperimentalprotocoltothenext,anoptimalpulseprotocolis probablydifferentforeach molecule, and optimalprotocolsfor each moleculemaylie outside ofparameterranges weexperi- mentedwith.Inotherwords,bychangingpulseparameterswe mayalso beable toseechanges intransdermal deliveryof the largestmolecule,FD20,however,optimalparameterrangesforthis moleculemaybedifferentthantheonesweexperimentedwith.

Further,therangesofthedeliveredamountofFD4vs.FD10are verysimilar.Specifically,theamountof deliveredFD4 ishigher

thantheamountofFD10,butthedifferenceismuchsmallerthan onemightexpect.Assurprisingasthismayseem,thisresultagrees wellwithdatapublishedbyLeeatal.(Leeetal.,2011).Theauthors also used fractional Er:YAG laser to deliver (among other molecules)FD4,FD10andFD20.Althoughabsolutevaluesofthe delivered amounts are approximately an order of magnitude higherinthisstudywhencomparedtoourresults(whichisnot surprising, becausetheskinwas treatedwithmore passes and more microchannels were created than in our study), we can observe verysimilardifferences in moleculardelivery between differentlysizedFITC-dextrans.Namely,themolecularfluxesfor FD4,FD10andFD20publishedbyLeeatal.(Leeetal.,2011)are 17.8,12.6and 1.4pmol/cm²/h,respectively,whichshowssimilar differencesindeliverybetweendifferentlysizedmolecules.This seemstosuggestthat,fortheserangesoflaserpulseparameters, Fig.4.Cumulativeamounts(inpmol/cm²)ofallFITC-dextrans(FD)inthereceiver for5hofpassivediffusionafterthelasertreatment:panela)FD4;panelb)FD10;

panelc)FD20.Allpulseprotocols–(variousduration/constantenergy+various energy/constantdurationprotocols)arecompiledinasinglegraphforeachFD.

Notedifferentrangesforeachpanel.Statisticalsignificancebetweendifferentlaser treatmentprotocolsatthe5-htimepointafterlaserexposureismarkedwithan asterisk.A0.05levelofprobabilitywastakenasthelevelofsignificance.Asalllaser treatment protocolswere significantly more efficient than passive diffusion, asterisksdenotingstatisticalsignificancebetweenlasertreatmentprotocolsand thecontrol(passivediffusion)wereleftoutforclarity.

moleculardeliveryenhancementisnotlinearwithmolecularsize but drops drastically for molecules larger than approximately 10kDa.

3.3.Histology

The region of the skin exposed tothe laser treatment was examined histologically toobserve laser-created microchannels throughupperskinlayersforallexperimentalprotocolslistedin Table1.Skinwascutinthedirectionperpendiculartoskinlayers and sections cut through the middle of microchannels were observedwithamicroscope.Micrographsofskinafter5different laserpulseprotocolsareshowninFig.5.

The effect that increasing pulse duration (hence decreasing pulsepeakpowersatconstantbeamenergy)hasonskincanbe seenwhencomparingpanelsa)b)andc)onFig.5.Asevidencedby FITC-dextrandeliveryresultspresentedinsection3.1(Fig.2),the shortest pulse creates the most favorable microchannels for moleculardelivery(panela)onFig.5).First,thecreatedpathways goasdeepas150micrometersintothedermis,andsecond,the tissueindeedseemstoberemovedthroughlaserpulseablative actionastherearenovisiblezonesofthermallyalteredtissueand no residual tissue fragments. Further, the increase in pulse duration(Fig.5panelsb)andc))somewhatdecreasesthechannel depthbut the more important differences are probably i) the presence of harder top layer residue (epidermis with stratum corneum)thatmayobstructmoleculardelivery (Bachhavetal., 2013), and ii)deeper thermallyaltered zonesin theremaining tissue,bothduetotheshiftfromablativetowardsthermalactionof

laserpulses.Althoughdifferentthermallymodifiedtissue zones (carbonized and coagulated) cannot be discerned in histology micrographs(Fig.5),moremodifiedepidermalanddermaltissue canbeobservedforlongerpulseprotocols(darkerbordersoflaser- createdmicrochannels).However,thereseemstobelessdifference between the histology slides presented on panels b) and c) (correspondingtopulsedurations100

m

^sand250

m

^s),compared

to the shortest pulse protocol (panel a). This implies that the thermallymodifiedzonesofthelasercreatedmicrochannelsmay notchangeproportionallywithpulsedurationbutinsteadmaybe asortofathresholdresponse.

Further,using theshortest duration pulses whiledecreasing laser energy from 380 to 230 and on to 80mJ/pulse notably decreasestissueablationdepthasisapparentfromFig.5:panels a),d)ande).At230mJ/pulse energy,thelaser-createdpathway traversesepidermisandendsintheverytoppartofthedermis, while 80mJ pulse is only able to ablate the top part of the epidermis,whichseemstosupportourassumptionthatchanging pulseenergypredominantlyaffectsthedepthofthelasercreated microchannels. This seems to correspond well with the skin ablationdepthsreportedintheliterature,usingfractionalEr:YAG lasers.Leeetal. (Lee etal.,2010,2014a)statethattheablation depth equals 4

m

^{m per J/cm2} fluence which agrees well with ablationdepthsthatcanbeinferredfromFig.5.Namely,at7.1J/

cm²fluence,wewerejustabletoablatethestratumcorneuminto thetoplayerofepidermis(panele)),whileat33.6J/cm²fluencewe achievedablationdepthsbetween120and150

m

^{m(asestimated}

frompanelsa),b)andc)).Further,Forsteretal.(Forsteretal.,2010) reportablationofonlystratumcorneumwhenusinglowfluences

Fig.5.Micrographsofskinsectionscutinthedirectionperpendiculartoskinsurface,throughlaser-createdmicrochannelsafterdifferentpulseprotocols:a)b)andc)The energyofthedeliveredpulseswasfixedat380mJ(fluence:33.6J/cm²),pulsedurationswerevaried:a)50m^s,b)100m^sandc)250m^{s,withtheirpeakpowers(at380mJ} energy)amountingto:SSP:7.6kW,MSP:3.8kW,SP:1.52kW,respectively.Panelsa)d)ande)Pulsedurationwasthesame:50ms,laseroutputenergywasvaried:a)380mJ/

pulse(fluence:33.6J/cm²),b)230mJ/pulse(fluence:20.3J/cm²),c)80mJ/pulse(fluence:7.1J/cm²),correspondingto7.6,4.6and1.6kWpeakpowers,respectively.

(4–8J/cm²),epidermiswhenusingfluencesof12and24J/cm²,and microchannelsreachingthedermisatfluenceshigherthan48J/

cm².ThesefindingsmatchablationdepthsseeninFig.5,especially for7.1J/cm²(panele))andalsofor20.3J/cm²(paneld)).However, at higher fluences, ourablation depthswere somewhat larger:

between120and150

m

^mat33.6J/cm2fluence(panelsa),b)and c)),whilemeandepthsinForsteretal.(Forsteretal.,2010)were 75.6and89.3

m

^{mat48J/cm2and144J/cm2}fluence,respectively.

Similarly,aseriesofarticlesbyBachhavetal.(Bachhavetal.,2010, 2011,2013)confirmbetteragreement with ourexperimentsat lowerfluencesthanathigherfluencevalues.Thatis,fluenceof90J/

cm²wasneededtoreach150

m

^{mdepthintheir}experiments,as opposedto33.6J/cm²fluencetoachievedepthsbetween120and 150

m

^minourstudy.

The reason for this discrepancy with thepublished data at higherfluencesmaybedue tothedifferencein theused pulse duration. Namely, when comparing different pulse energies (fluences),pulsedurationinourstudywasshorter–50

m

^s–than forexampleinForsteretal.(Forsteretal.,2010)orBachhavetal.

(Bachhavetal.,2010,2011,2013),wherepulsedurationwas350or 400

m

^{s.Inotherwords,ascanbeinferredfromthecomparisonof}

panelsa),b)and c)inFig.5,evenwiththesamepulseenergy (fluence)betweenthethreeprotocols,theshortestpulsesseemto producethedeepestmicrochannels.Anotherpossiblereasonfor thisdiscrepancymaybeanoverestimationofthediameterofour fractionalhandpiecePS-01microbeams,whichweroundedupto the maximum of 400

m

^{m (see Fig.1b).The actual diameter is}

somewhatsmallerand,accordingly,thefluenceofourlaserpulses maybehigher,whichbringsourresultsevenclosertothefindings ofotherresearchers.Thisseemstobeconfirmedbythediameterof the created microchannels seen in Fig. 5 that range from approximately 100

m

^{m to 300}

m

^{m. However, these numbers}

cannotbedirectlytranslatedintothelasermicrobeamdiameter astheactualareaofskinalteredbythemicrobeamdependsonthe laserpulseparameters.

Thisresponseoftissueintermsofmicrochanneldepth(Fig.5 panelsa),d)ande))isalsomuchmoreproportionaltolaserpulse energythantissueresponsetochangesinpulseduration(Fig.5 panelsa),b)andc)),whichagreeswellwithmoleculardelivery resultsin Figs. 2 and 3. Namely, thedelivery ofFD4 and FD10 alwaysincreaseswhenlaserpulseenergyisincreasedfrom80to 230–380mJ(Fig.3).Ontheotherhand,moleculardeliveryresults in Fig. 2 indicate that changing laser pulse duration causes a threshold effectin the delivery of the smallesttwo molecules, similartotheeffectsonmicrochannelpropertiesshowninFig.5.

Asstatedbefore,forthelargestmolecule,FD20,transportthrough skin tissue is low and fairly independent of changes in pulse parametersandthepropertiesofthelaser-createdmicrochannels.

In summary, both, molecular delivery as well as microchannel properties display more proportionality when pulse energy is variedthanwhenlaserpulsedurationisvaried.

Theeffectof microchanneldensity anddepth(controlledby fluence) was investigated by different researchers for various compounds.Thedeliveryofdiclofenac(Bachhavetal.,2011)and twotherapeuticantibodies(AntithymocyteglobulinandBasilix- imab) (Yu et al., 2011) was found to increase when either microchannel number or depth (laser fluence) was increased.

Anotherstudyby Bachhavet al.(Bachhavet al., 2013)showed increaseddeliveryofcytochromecandFITC-BSAwithincreased microchannel number, however, only the delivery of FITC-BSA increasedwithincreasingfluence,ontheotherhand,increasing microchannel depth had no effect on cytochrome c delivery.

Anotherstudybythesameauthors(Bachhavetal.,2010)reveals similarfindingsforlidocainedelivery(nodependenceonfluence, increaseddelivery withincreasingmicrochannel density).How- ever, Oni etal. (Oni et al., 2012)do report increasedlidocaine

deliverywithincreasingfluence,uptothemicrochanneldepthof 250

m

^{m.Theseresultssuggestthatmoleculartransportthrough}

laser-created microchannels depends strongly on permeant’s physicochemicalproperties(weight,partitioncoefficient,ioniza- tion),notonlyonthepropertiesofthetransportpathways.Inour study,weonlyexaminedtheeffectofmolecularweight.

Further,asevidentfrommicrographsinFig.5,skinintegrityis compromisedat micro-ablationsites, wherepartsof skintissue rangingbetween100and300

m

^{mindiameterandupto150}

m

^min

depthareremoved,whichraisessafetyconcerns.Asthiswasanex vivostudy,nofurtherexaminationintosafetyaspectsofthelaser- assistedtransdermaldrugdeliverywasconducted.However,invivo studieshaveshownthataftertheuseoffractionallasersskintissue displaysgoodepidermalbarrierfunction.Namely,duetothevery smallarea ofablation,theremaininguntreated skinservesasa reservoirforhealing,whichleadstocompletere-epithelializationof thestratumcorneumwithinonedayafterlasertreatment(Leeetal., 2010).Also inclinical setting,theuse offractional Er:YAGlaser demonstratedsignificantlylowerrateof(mild)adverseeffectsthan whenskinwastreatedwithconventional(non-fractional)Er:YAGor CO2lasers(Graberetal.,2008).Nevertheless,evenwiththisgood safety profile of fractional Er:YAG lasers, concerns regarding transientincreasedskinpermeability(thedesiredoutcomeofthe treatment)remain.Toavoidunwantedintroductionofpathogens, cautionshouldbeadvisedforaboutadayafterlasertreatmentuntil skinbarrierfunctionisfullyrestored.

4.Conclusions

Er:YAGlaserwithfractionaloutputbeamprofilewasusedas enhancement method for transdermal drug delivery of three model molecules of different sizes: fluorescein isothiocyanate labeled dextrans(FITC-dextrans or FD) with averagemolecular weightsof4(FD4), 10(FD10)and20kDa(FD20).TheEr:YAGlaseris usedforcontrolledremovalofthethindeadouterlayeroftheskin, thestratumcorneum,takingadvantageoftheablativeeffectsof laserlightontissuewhilecausingminimalthermaldamage,thus sparingviableunderlyinglayers(epidermis,dermis).Further,laser beamoffractionallasersissplitintomicrobeams,soevenlarger portionof viableskintissue is spared.Pulses ofdifferentpulse energies,fluencesanddurationshavebeentriedexperimentallyso far, however, toour knowledge, this is the first study directly comparingdifferentpulsedurationsatconstantorvaryingpulse energies,assessingtheeffectofdifferenttemporalpulseprofiles onskinandconsequentlyonmoleculardelivery.Weexperimented withFotonaLightWalkerlasersystemwithafractionalhandpiece PS-01 able to modulate pulse duration at constant delivered energy,therebychangingtheinstantaneouspower.

Inthefirstpartofthestudy,weusedprotocolsamongwhich pulsedurationwasvariedwhilekeepingpulseenergyconstant:

shorter pulses of equal energy means higher peak power.

Followingthat,wekeptthedurationofthepulsesconstantwhile varying theirenergy.Our results suggestthat two mechanisms seem toplay importantpartin thelaser-enhancedtransdermal deliverystory.Oneiscreationofdifferentlysizedmicrochannels, forwhichtheenergyofthedeliveredpulsesisthemostimportant parameter.Thesecondmechanismispartitioningofthepermeant into tissue, which differs with the extent of thermally-altered tissue due to laser pulses that can at constant energy – be controlled predominantly by differing pulse duration/power.

Namely, tissue ablation threshold is lowered at shorter pulse durations with higher power, which is preferred as it lowers thermal effects on viable skin layers. This is also shown by histologicalanalysisoflaser-treatedskinandobservationoflaser- createdmicrochannelsaftertreatmentwithdifferentlaserpulse protocols.Especiallyforsmallermolecules,transdermaldeliveryis