Micron89(2016)60–76
ContentslistsavailableatScienceDirect
Micron
jo u rn al h om ep a g e :w w w . e l s e v i e r . c o m / l o c a t e / m i c r o n
Review
Atomic force microscopy for the investigation of molecular and cellular behavior
Alper D. Ozkan, Ahmet E. Topal, Aykutlu Dana, Mustafa O. Guler, Ayse B. Tekinay
∗BilkentUniversity,UNAM-InstituteofMaterialsScienceandNanotechnology,Ankara,Turkey
a r t i c l e i n f o
Articlehistory:
Received4May2016 Accepted27July2016 Availableonline29July2016
Keywords:
Atomicforcemicroscopy Biomacromolecules Mechanicalcharacterization Cells
a b s t r a c t
Thepresentreviewdetailsthemethodsusedforthemeasurementofcellsandtheirexudatesusingatomic forcemicroscopy(AFM)andoutlinesthegeneralconclusionsdrawnbythemechanicalcharacterizationof biologicalmaterialsthroughthismethod.AFMisamaterialcharacterizationtechniquethatcanbeoper- atedinliquidconditions,allowingitsusefortheinvestigationofthemechanicalpropertiesofbiological materialsintheirnativeenvironments.AFMhasbeenusedforthemechanicalinvestigationofproteins, nucleicacids,biofilms,secretions,membranebilayers,tissuesandbacterialoreukaryoticcells;however, comparisonbetweenstudiesisdifficultduetovariancesbetweentipsizesandmorphologies,sample fixationandimmobilizationstrategies,conditionsofmeasurementandthemechanicalparametersused forthequantificationofbiomaterialresponse.AlthoughstandardprotocolsfortheAFMinvestigation ofbiologicalmaterialsarelimitedandminordifferencesinmeasurementconditionsmaycreatelarge discrepancies,themethodisnonethelesshighlyeffectiveforcomparativelyevaluatingthemechanical integrityofbiomaterialsandcanbeusedforthereal-timeacquisitionofelasticitydatafollowingthe introductionofachemicalormechanicalstimulus.Whileitiscurrentlyoflimiteddiagnosticvalue,the techniqueisalsousefulforbasicresearchincancerbiologyandthecharacterizationofdiseaseprogression andwoundhealingprocesses.
©2016ElsevierLtd.Allrightsreserved.
Contents
1. Introduction...60
2. Effectofprobemorphology,compositionandsurfacechemistry...61
3. Atomicforcemicroscopyofunicellularorganisms...62
3.1. Bacterialcellsurfaces...62
3.2. Bacterialsecretions,exudatesandbiofilms ... 64
4. Atomicforcemicroscopyofmammaliancellsandtissues...64
4.1. Cancerdiagnosisandcharacterization ... 67
4.2. Diagnosisofotherdiseases ... 69
4.3. Stemcelldifferentiation...70
4.4. Extracellularsecretionsandtissuemicroenvironments ... 71
5. Futuredirections...71
References...73
1. Introduction
Bothuni-andmulticellularorganismscoordinatetheirbehavior usinganetworkofchemical,electricalandmechanicalsignals,and employavarietyofsensorymechanismstoperceiveandrespond
∗ Correspondingauthor.
E-mailaddress:atekinay@bilkent.edu.tr(A.B.Tekinay).
tointernalorexternalregulatorycues(Riccaetal.,2013;Johnson, 2013).Inunicellularorganisms,suchsignalsmayassistinfeed- ing,attractingconspecifics,synchronizingreproductivecyclesor initiatingdefensemechanismsinahostileenvironment(Dufour andLevesque,2013);whilemulticellularlifeutilizescellsignal- ingnetworkstoregulatecellrecruitment,adhesion,differentiation, proliferationanddeath(WattandHuck,2013;Ravichandran,2003;
Zoranovicetal.,2013;JaenischandBird,2003;OwensandWise, 1997).Asthelattercategoryofprocessesareintegraltosustain complexlife,thecharacterizationofregulatorysignalsisofgreat http://dx.doi.org/10.1016/j.micron.2016.07.011
0968-4328/©2016ElsevierLtd.Allrightsreserved.
A.D.Ozkanetal./Micron89(2016)60–76 61
importancetothemedicalandbiologicalsciences,andmuchwork hasbeenperformedtoelucidatethelinksbetweenenvironmen- talcuesandcellularprocesses(Andoetal.,2013;Carvalhoetal., 2013;Dorobantuetal.,2012).However,whilethechemicaland biological environment of cells are relatively well-defined, the mechanicalpropertiesofcellsandtheirimmediateenvironment areinvestigatedonlytoalesserdegree;partlybecauseofthehigh complexityandvariabilityofthemechanicalinteractionsexhibited bycells andpartly duetolimitationsassociated withthehigh- resolutionmechanicalprobingofcellsurfacesandinteriors(Cohen andKalfon-Cohen,2013).Nonetheless,considerableefforthasbeen spenttoestablishhowcellsperceiveandactuponthephysicalchar- acteristicsofnearbysubstrates(Shaoetal.,1996),andtodetermine howthemechanicalpropertiesofcellsandtissuesarealteredin responsetodiseasestateorenvironmentalfactors,usingmaterial characterizationtoolssuchasmagnetictwistingcytometry,optical tweezers,microneedleprobes,scanningacousticmicroscopyand atomicforcemicroscopy(NeumanandNagy,2008).
Atomicforcemicroscopy(AFM)isacharacterizationtoolthat measures the topology and material properties of surfaces by recordingthedeflectionofametallicprobe(or“tip”)asitmoves overthetargetsurface.AFMcanbeoperatedunderthreeprinci- palmodes:Incontactmode,thetipisdraggeddirectlyoverthe surfaceanddeflectsawayduetoarepulsiveCoulombicinterac- tion,while in non-contact modeit is held at a short(typically
<100nm)distanceoverthesampleandoscillatesatafrequency thatdependsontheattractivevanderWaalsforcesactingupon it.Intappingorintermittentcontactmode,thetipiskeptoscil- latingabovethesample,andtheoscillationfrequencychangesas thetipapproachesthesurfaceatregularintervals(Giessibl,2003).
Contactand intermittentmodesareparticularlysuitableforthe probingofbiologicalsamples,duetotheirapplicabilityinliquid media(Danino,2008).Despiteconsiderablelossesinresolution,a liquidsampleenvironmentallowscellularimaginginanative(or pseudo-native)environmentand,moreimportantly,permitsthe directinvestigationofmechanicalchangesonalivecellsurfacein responsetoanintroducedstimulus(Liuetal.,2005).Time-lapse elastographstakeninthisfashionhavebeenutilizedforadiverse arrayofapplications,includingtovisualizetheformationofamy- loid(Harperetal.,1997)orcollagen(Revenkoetal.,1994)fibers underdifferentenvironmentalconditions,determinehowmem- braneintegrityisalteredin thepresenceofantibiotics(Fantner etal.,2010a),orrecordtheproductionanddissolutionofcytoskele- talelementsduringcellmovement(RotschandRadmacher,2000).
Inaddition,itispossibletoutilizetheAFMtipasastimulusto elicitaresponsefromthetargetcell,andtheprobeitselfcanbe functionalizedwithligandmoleculestodeterminetheaffinityof thecellmembranetoaparticularbiologicalmoiety.
DuetotheversatilityandpotentialapplicationareasofAFM, thetechniquehasattractedsubstantialinterestinbiomechanical research,andhasbeenusedinthecharacterizationofagreatvari- etyoftissues,cellsandsub-cellularstructuresinbothlivecondition andfollowingfixinganddrying.Thepresentreviewaimstocover thosestudiesthatfocusonthedifferencesinmechanicalproperties associatedwithpathologicalconditionsorchangesinenvironmen- talcues,andemphasizestheimportanceofthemechanicalUmwelt inmodulatingthebehaviorofbothsingle-celledandmulticellular systems.
2. Effectofprobemorphology,compositionandsurface chemistry
BeforediscussingtheAFMimagingofbiologicalmaterials,the importanceofAFMtipchoiceshouldbeunderlined.Thediame- ters,materials,morphologiesandcantileverlengthsofcommercial
AFMprobesshowconsiderablevariance,andoptimalperformance requirestheuseofaprobeconductivetothetaskathand.Thecom- positionofthesamplematerialshouldbetakenintoconsideration tochoosethespringconstantoftheAFMprobe,assoftermaterials, suchascells,maybedamagedoverrepeatedcontactwiththeAFM tip(Costa,2003).Inaddition,dependingontheareatobescanned, itmaybedesirabletoincreaseordecreasethetipdiameter.Larger tipsareassociatedwithlowerresolution,butcanbeutilizedtoscan largersampleareaswithoutcompromisingtipintegrity,assharper tipsmayexperiencesignificantwearoverlongscanningdistances, suchaswhenscanningcells.Ontheotherhand,sharpertipsare capableofresolvingsmallerfeaturestoagreaterextent,whichis invaluablewhenmeasuringproteinsandothernanoscalebiolog- icalmaterials.Consequently,differencesinmaterialstiffnessthat areevidentundernanoscaleinvestigationmaybeunmeasurable usingmicroscaletips(Stolzetal.,2009a).Ifadhesiondataistobe collected,thematerialandmorphologyoftheAFMtip(alongside substrateproperties)alsodeterminesthesuitablemodelforusein elasticitycalculations(Fig.1).
AFMprobescanalsobefunctionalizedinordertocharacterize theinteractionbetweentwospecifictypesofbiologicalmoieties, suchasbetweenareceptoranditsligand.Thistypeofinteraction isbestexemplifiedbybiotinandavidin,usedbyColtonetal.in theirhallmarkpapertoillustratethepossibilityofusingAFMto directlyevaluatethestrengthofmolecularinteractions(Leeetal., 1994).Mechanicalpropertiesofawidevarietyofproteinshavenow beenelucidated,includingtheinteractionsbetweenantibodiesand theircorrespondingantigens(Allenetal.,1997),actinandmyosin (Koderaetal.,2010),osteopontinandintegrin(Leeetal.,2007),and variouscelladhesionproteoglycans(Dammeretal.,1995).Such proteinscaneitherbecovalentlytetheredtothetargetmaterial (Ebneretal.,2007;Kamruzzahanetal.,2006)orattachedbydrying theproteinsampleonthesurface(Florinetal.,1995).Inaddition, themechanicalstrengthoftheconstituentdomainsofasinglepro- teincanbeevaluatedbyattachingthatproteintoasurfaceand usingtheAFMtiptostretchit(Lietal.,2003).Thisresultsinthe gradualunfoldingoftheprotein,andtheunwindingofeachdomain isassociatedwithamomentarydropinforce.Tensilecharacteris- ticsoftheimmunoglobulinandfibronectinIIIdomainsoftitinwere investigatedusingthismethod(Riefetal.,1998),andtheabilityof thebacterialribonucleasebarnasetowithstandforcewaslikewise evaluatedbyincorporatingthisproteinintoachimericconstruct consistingoffourTII27and threebarnasesubunits(Bestetal., 2001).
DNAandRNAcanalsobeimmobilizedandcharacterizedina similarmanner,andthemechanicalinvestigationofDNAmolecules ofvaryinglengthsandconfigurationshasbeenperformedusing AFM(Maoetal.,1999; Hansmaetal.,1995).In additiontothe determinationofcovalentbondstrengthinssDNAordsDNA,itis alsopossibletoevaluatethestrengthofthebondsbetweencom- plementarystrandsinashortdsDNApiece,ortodeterminethe forcesnecessarytostretchanintactDNAmolecule(Hansmaetal., 1996).High-resolutionAFMimagingcanalsobeusedtocharacter- izethephysicalstructureofaDNAhelix(Fig.2),andmorecomplex DNAarchitecturesandDNA–proteininteractionscanbevisualized andcharacterizedusingatomicforcemicroscopy.Yanevaetal.,for example,confirmedthatDNA-dependentproteinkinase(DNA-PK) canbindtoDNAwithouttheassistanceofKuproteins,andthatthe lattershowsatime-dependentpreferenceforstrandends,byvisu- alizingDNA-KuandDNA-DNA-PKinteractionsusingAFM(Yaneva etal.,1997).Theaffinitybetweencellsandspecificproteinscanalso beassessedbyindentingthecellofinterestwithanAFMtipfunc- tionalizedwiththeproteinofinterest(Hanetal.,1995).Gaubetal.
reportedamethodtodistinguishbetweenindividualredbloodcell originsinamixtureofA-andO-grouperythrocytes,usinganAFM tipfunctionalizedwithHelixpomatialectin(Grandboisetal.,2000).
62 A.D.Ozkanetal./Micron89(2016)60–76
Fig.1. Comparisonofmicro-andnanoindentationfortheidentificationofchangesassociatedwithaginginthecartilageofC57BL/6mice.Microindentationresultscould detectnodifferencebetween1,10and19-montholdindividuals,whilenanoindentationwasabletodeterminethatthecartilageelasticityofnon-arthriticmicechanges withage.ReplicatedwithpermissionfromStolzetal.(2009b).
Fig.2. AFMtopographyofaplasmid,showingthegeneralappearanceoftheDNA helixundervaryingpeakforces(a–d).Minorandmajorgroovescanbeobserved inAFMimages(a–d,insets),andtheDNAstructurecanbecompressedunderhigh peakforces(e,f).Whitearrowdenotesadislocationinaplasmidloopcreatedby highloadforces.ReplicatedwithpermissionfromPyneetal.(2014).
Thislectindisplaysstrongaffinitytoglycolipidsthatarepresentin themembranesofA-groupbutnotO-grouperythrocytes,resulting inhigherruptureforcesassociatedwiththeformer.Thedifferences
inadhesiveforcesarethenutilizedtocreateamapwhereindividual A-andO-groupcellscanbeidentified.
3. Atomicforcemicroscopyofunicellularorganisms
ArepresentativeselectionofAFMstudiesonthestiffnesschar- acterizationofbacteria,yeastsandcellularsecretionsisprovided inTable1.Itisreadilyevidentthatseveralmeansofsampleprepa- ration are available for measurement, and that values suchas tip-sampleadhesion,F-dcurveslopesandcellularspringconstants canallbeusedtocomparethemechanicalintegritiesofbiological samples;consequently,onlythesestudiesdetailingthefullrange ofmeasurementconditionswereincludedintothetable.Bothair andliquidimaginghavebeenperformedformechanicalinvesti- gations;however,biologicalmaterialsareoftenviscoelasticand maydisplaylargechangesinelasticbehaviordependingonenvi- ronmentalhumidity.Assuch,samplesinairtendtohavemuch largerYoung’smodulicomparedtosamplesimagedinliquids(e.g.
a10-folddifferencewasobservedinbetweentheelasticmoduliof air-driedandrehydratedmurinesacculifromEscherichiacoli(Yao etal.,1999)).Given thedifferences inmeasurementtechniques andsamplepreparationmethods,aswellasthenaturalvariance inthematerialpropertiesofbacterialcellsandtheirsecretions,it isusuallypreferabletocompareresultswithinstudiesratherthan assumingagivenstiffnessvaluewillapplyunderotherexperimen- talconditions.
3.1. Bacterialcellsurfaces
Unlikemanyvertebratecelllines,bacterialcellsarenotdepen- dentonahighlyspecificsetofenvironmentalconditionstosurvive, andcantolerateextendedAFMimagingsessionswithoutdetrimen- taleffects(Ramanetal.,2011;Franciusetal.,2008).Theireaseof procurement,non-demandinggrowthconditionsandthefactthat manylaboratoryspecieseitherare,orserveasmodelsfor,common pathogensmakebacteriapopulartargetsforAFMimaging.Bacte- riamustbeimmobilizedpriortoimaginginliquidmedia,astheir mobilityotherwisemakesitimpossibletoimage,andevenses- silebacteriacanbelaterallypushedbytheAFMtip(Doktyczetal., 2003).Immobilizationcanbeperformedbydryingandrehydrating, electrostaticbindingtoapositivelychargedsurface(e.g.gelatinor
A.D.Ozkanetal./Micron89(2016)60–7663 Table1
Mechanicalcharacterizationofmembranes,secretionsandsingle-celledorganismsbyAFM.
Sample Tipproperties Imagingconditions Elasticproperties Reference
Sulfate-reducingbacteria Siliconnitride,k=0.06N/m(nominal) Contactmode,air,sampleonmica (Adhesion)−3.9to−4.3nNatcellsurface,−5.1 to−5.9nNatcell-substrateboundary,−6.5to
−6.8nNatcell–cellboundary
Fangetal.(2000)
EnteroaggregativeEscherichiacoli, wild-typeanddispersinmutant
Silicon,k=2.8N/m(nominal) Contactmode,liquid(distilledwater)andair, sampleongelatin-treatedmica
(F-dslope)0.133forwild-typestrainonagar, 0.069forwild-typestraininbroth,0.81for dispersinmutantonagar,0.78fordispersin mutantinbroth
Beckmannetal.(2006)
Bacillussubtilis,Micrococcusluteus, Pseudomonasputida,twostrainsof Escherichiacoli
Siliconnitride,k=0.32N/m(nominal) Contactmode,liquid(HMbuffer),sampleon APTEScoverslip
(Springconstant)variesbetween0.16±0.01to 0.41±0.01,higherinGram-positivecells
Volleetal.(2008)
Pseudomonasaeruginosa Siliconnitrideandsiliconnitridewithsilicon oxidetips,k=0.07±0.01(calibratedbythe thermalmethod)
Contactmode,liquid(milliQwater),sampleon poly-l-lysine-coatedglass
(Springconstant)0.044±0.002N/mfor unfixed,0.11±0.03N/mfor
glutaraldehyde-fixedcells;creepdeformation behavioralsoinvestigated
Vadillo-Rodriguezetal.(2008)
Klebsiellaterrigena Siliconnitride,k=0.06N/m(nominal) Contactmode,liquid(potassiumphosphate bufferatpH6.8),sampleonpolycarbonate membranefilter/poly-l-lysine-coated glass/immobilizedontipbygluteraldehyde fixation
(Adhesion)−0.26±0.05nNformembrane filter,−0.5±0.2forpoly-l-lysine,−35±2nN forgluteraldehydefixation,otheradhesion parametersalsomeasured
Vadillo-Rodriguesetal.(2004)
TwoStreptococcussalivariusstrains Siliconnitride,k=0.03N/m(nominal) Contactmode,liquid(deionizedwateror0.1M KClsolution),sampleonpolycarbonate membranefilter
(Adhesionandrepulsion)Fibrillatedstrain showsalargerrepulsionrange,interpretedto reflectthelayeroffibrils;retractionresultsin threeadhesionforcespotentially
correspondingtothreedifferentlengthsof fibrilsobservedbyelectronmicroscopy
vanderMeietal.(2000)
Desulfovibriodesulfuricans, Pseudomonassp.andan unidentifiedlocalmarineisolate
Siliconnitride,k=0.12±0.02N/m(calibrated bythethermalmethod)
Contactmode,liquid(artificialseawater), samplecoatedontipandbroughtinto interactionwithmetals
(Adhesion)Allthreeisolatesadhereto aluminumbetterthanmildsteel,stainless steel316andcopper;Desulfovibrioand Pseudomonasadherebetterthanthemarine isolate.
Shengetal.(2007)
B.mycoides Silicon,k=0.064N/mand0.4N/m Contactmode(constantheight),liquid (0.145MNaCl),samplecoatedontipand broughtintointeractionwithglass
(Adhesion)7.4±3.7nNofadhesionto hydrophilicglasssurface,49.5±14.42nNof adhesiontohydrophobic-coatedglasssurface
Bowenetal.(2002)
Marinebacterialdepositions Silicon,k=45.7N/m(calibratedbytheadded massmethod)
Tappingmode,air,sampledepositedon fluoridatedandnon-fluoridatedpolyurethane
(Young’smodulus)between1.5and2.2GPa Bakkeretal.(2003)
Bacterialdepositions,suspectedto beextracellularpolymeric substances
Siliconnitride,kvariesfrom0.03to0.5N/m (nominal)
Contactmode,liquid(MilliQwater),sample depositedonpolystrene
(Adhesion)Forcesof0.8±0.2nNobservedover barepolystyrene,asopposedto0.2±0.2nN aftercellattachment
vanderAaandDufrene(2002)
P.aeruginosapili Siliconnitride,0.008±0.004N/m Contactmode,liquid(water),sampleattached topoly-l-lysine-coatedtipsandbroughtinto interactionwithmicasurface
(Adhesion)Ruptureforcesof95pNduring retraction.
Touhamietal.(2006)
E.colibiofilms Siliconnitride,k=0.07-0.4N/m Contactmode,air,sampledepositedonglass (Adhesion)Pull-offforcesof122.65pNfor cell-tipinteractionand51.79pNforglass-tip interaction
Ohetal.(2007)
Bacterialcellulosefibers Siliconnitridek=1.03±0.05N/m(calibrated bythethermalmethod,nominalkof0.5N/m notusedduetolargediscrepancy)
Contactmode,air,sampleonsilicon nitride-coatedsilicongrating
(Young’smodulus)78±17GPa Guhadosetal.(2005)
E.coliandE.colispheroplasts Siliconnitride,k=0.1and0.01N/m(nominal, actualspringconstantscalibratedbythe thermalmethod)
Contactmode,liquid(TBS2buffer),sampleon APTES/Glutmica
(Springconstant)0.194N/mforintactcells, 0.571N/mforfixedspheroplasts
Sullivanetal.(2007)
Saccharomycescerevisiae Siliconnitride,k=0.008±0.4N/m(calibrated bythethermalmethod)
Contactmode,liquid(milliQwater),sampleon polycarbonatemembranefilter
(Young’smodulus)6.1±2.4MPaonbudscar, 0.6±0.4MPaonsurroundingcellwall
Touhamietal.(2003)
Aspergillusnidulans Siliconnitride,k=0.47±0.06N/m(calibrated bythethermalmethod)
Contactmode,liquid(PBS),sampleon poly-l-lysine-coatedglass
(SpringconstantandYoung’smodulus) 0.29±0.02N/mand110±10MPafor wild-typehyphaeincompletemedium, decreasesformutantstrainlackingachitin synthesisgene,aswellasinthepresenceof 0.6MKCl.
Zhaoetal.(2005)
64 A.D.Ozkanetal./Micron89(2016)60–76
poly-l-lysine),entrapmentinadhesiveproteins,covalentbinding toamino-orcarboxyl-functionalizedsurfacesorbyphysicalcon- finementinmicrowellsorporousmembranes(Doktyczetal.,2003;
Suoetal.,2008).Itshouldbekeptinmindthattheimmobilization methodmayalterthesurfacepropertiesoftheentrappedcells,e.g.
bydirectlyalteringthebacterialsurfacechemistryortriggeringa defensemechanismagainsttheenvironmentalstressesassociated withtheimmobilizationtechnique.Duetothesmallsizesofbacte- ria,itisalsopossibletofunctionalizeAFMtipswithbacterialcells, whichcanthenbeusedtotesttheinteractionbetweenthebac- teriumandmaterialsurfaces.Tingetal.,forexample,usedsuchtips toshowthattheGram-negativebacteriaMassiliatimonaeandPseu- domonasaeruginosaadherebettertostainlesssteelsurfacesthan doestheGram-positiveBacillussubtilis(Harimawanetal.,2011).
AFMstudiesofbacteriafrequentlyfocusonthemechanismsby whichcertainmoleculesinhibitthegrowthofpathogens. Many antibiotics(e.g.beta-lactamantibiotics,polymyxinsandglycopep- tideantibiotics)actbyinhibitingthesynthesisofbacterialcellwalls ormembranes,andtherebyalterthemembraneintegrityofthe affectedbacteria.Otherantibiotic-mediatedeffects,suchasmem- branethinning(Meckeetal.,2005)orporeformation(Mulleretal., 1999),canalsobeobservedbyAFM,andtheeffectsoflesscon- ventionalantibiotics,suchasplantextracts(Perryetal.,2009)and peptidesequences(daSilvaandTeschke,2005;Meinckenetal., 2005),canalsobeassessed(Fig.3).TheantimicrobialpeptidePGLa hasbeendemonstrated tolowerthestiffness ofEscherichiacoli membranesandcreatemicelle-likestructuresaroundcellmem- branespriortotheireventualrupture,whilegarlicextractwasalso associatedwithmembranedisruption.Chitosanandchitooligosac- charides(COSs)werealsotestedfortheirantimicrobialeffecton E.coliandStaphylococcusaureus,andthethickpeptidoglycanwallof S.aureuswasfoundtoallowthisbacteriumtobetterretainitsover- allmorphology,despiteexperiencingasignificantdecreaseincell rigidity(Fernandesetal.,2009).Whiletheeffectsofantibacterial moleculesmaybeapparentevenunderdryimaging,itisalsopossi- bletoquantifytheresultingchangesinatime-dependentmanner bycharacterizingthemechanicalpropertiesofbacterialcellsinliq- uidbeforeandafterintroducingtheantibioticinquestionintothe medium(Fantneretal.,2010a).Thistechniquehastheadvantageof ensuringthatthechangesobservedarefullyduetotheantibiotic,as opposedtoacombinationofitandthedryingprocess,asbacterial cellwallsareviscoelasticandmaygreatlyaltertheirmechanical propertiesinresponsetorelativehumidity(ThwaitesandSurana, 1991).
Avarietyofsubcellularelementscanalsobeidentifiedonmem- branesurfacesusingAFM.Bothbacterialandeukaryoticcellscan beusedintheseefforts,andlipidbilayerscanbesubstitutedas simplifiedmodelsofcellmembranes(Buttetal.,1990;Kuznetsov andMcPherson,2011).Thestructureandfunctionofmembrane proteins(Buttetal.,1990;Fotiadisetal.,2003;Yuanetal.,2002), porecomplexes(Stoffleretal.,1999),gapjunctions(LalandJohn, 1994),amyloidaggregates(Connellyetal.,2012)andavarietyof membrane-componentlipids(Gyorvaryetal.,2003a),aswellas themodeofactionofcholeratoxin(Mouetal.,1995),wereinves- tigatedbyatomicforcemicroscopy.Ofparticularinterestarethe recentdevelopmentsinAFM-basedrapid,high-resolutionimaging methods,whichhavegrantedsubstantialinsightintothenature ofsmallsurfaceelements.Processessuchastheself-assemblyof bacterialS-layerproteins(Gyorvaryetal., 2003b), protein fold- ing(Mulleretal.,2002),misfolding(Oberhauseretal.,1999)and crystallization(Reviakineetal.,1998)events,themotionof“walk- ing”proteins(Preineretal.,2014)anddrug-membraneinteractions (Berquandetal.,2004)havebeenthesubjectofreal-timeimaging studies.
3.2. Bacterialsecretions,exudatesandbiofilms
Itiswell-knownthatcellsinmulticellularorganismsenhance theirsurvivalandcoordinationthroughtheproductionofanextra- cellularmatrix;however, this property isnot unique tohigher eukaryotes. Bacteria also secrete proteins, polysaccharides and quorumsensingmoleculesthatrelayinformationbetweencon- specificcellsandserveasabufferagainstenvironmentalstresses.
Inaddition,thesurfaceattachmentof bacteriaisalsomediated byextracellulardepositions,theadhesioncapacitiesofwhichcan bemeasuredthroughAFM. Colanicacidproduction andsurface lipopolysaccharidelengths,forexample,werepreviouslydemon- stratedtodeterminethestrengthofattachmentofE.colicells,as colanicacid-overproducingmutantswerefoundtoexhibitstronger attachmentwhileshortersurfacelipolysaccharideswereassoci- atedwithalackofadhesivecapacity(Razatosetal.,1998).
BacterialbiofilmsarealsoofspecialinterestwithregardstoAFM characterization.Biofilmsareextracellularpolysaccharidesthatare secretedtofacilitatetheattachmentofbacteria(orotherunicellular organisms)toasurface,andprotecttheadheringcellsagainsthos- tileenvironmentalfactorssuchasantibiotics,detergentsandheavy metals(Daviesetal.,1998).Biofilmsareundesirableelementsin manysettings,andsuitablemeanstoinhibittheirformationisnec- essarytopreventpotentialhealthhazardsinfood,agriculturaland medicalindustries.Assuch,themechanicalpropertiesofbiofilms, as wellasthe mechanismsby which antifoulingmolecules act againstbiofilmproduction,havebeeninvestigatedbyusingAFM and othermechanical characterizationtechniques(Beech etal., 2002;ArnoldandBailey,2000).Corrosiondamageandadhesive capacity of biofilms on a variety of metal surfaces have been detailed intheliterature: Holdenet al.reportthat unsaturated andliquid-grownbiofilmsofPseudomonasputidaresponddiffer- entlytodrying,suggestingthatthebiofilmcompositionisaltered foroptimalgrowthindryandwetenvironments(Auerbachetal., 2000).Inanotherreport,Tayetal.detailtheeffectofsilverions onStaphylococcusepidermidisbiofilmsandproposeamechanism throughwhichsilverdestabilizesthebiofilmstructurebybinding to the electron donor groups provided by the biofilm compo- nents,therebyweakeningthehydrogenbondsthatholdthebiofilm matrixtogether(Chawetal.,2005).
4. Atomicforcemicroscopyofmammaliancellsandtissues
AselectionofAFMstudiesonthestiffnesscharacterizationof eukaryoticcellsandtissuesisprovidedinTable2.AFMofmam- maliancellsandtissuesisoftenundertakenfordiseasediagnosis orstemcelldifferentiationstudies,asdiseasedtissuesoftendisplay mechanicalcharacteristicsdistinctfromtheirhealthycounterparts and stem cells are widely known to alter their differentiation pathwaysand mechanicalcharacteristicsdependingonexternal stimuli.Consequently,AFMofcancerandstemcellscanprovide greaterinsightintothepathwaysrequiredforeventssuchasmetas- tasisandlineagecommitment.However,measurementofthese samplesismoredifficultthansingle-cellularorganismsornon- livingbiologicalmaterialssuchasbiofilmsandothersecretions,as mammaliancellsrequireaspecificsetofenvironmentalconditions tosurviveandenvironmentalstresscanhaveamajorimpacton theirmechanicalproperties.Inaddition,whilefixativetreatment canbeusedpriortoimaging,fixationmaygreatlyalterthemechan- icalpropertiesofeukaryoticcells,andlivecellimagingisgenerally requiredtoacquirereliableelasticitydata(Fig.4).Nonetheless, cellcultureconditionscanbereplicatedtosomeextentwithinthe liquidcellofanAFM,andalargenumberofsuccessfulinvestiga- tionshavebeenmaderegardingthemechanicalcharacterofliving
A.D.Ozkanetal./Micron89(2016)60–7665 Table2
MechanicalcharacterizationofmammaliancellsandtissuesbyAFM.
Sample Tipproperties Imagingconditions Elasticproperties Reference
NIH3T3fibroblasts Siliconnitride,k=0.018N/m (calibratedbythermalmethod)
Contactmode,liquid(DMEM containingd-glucose(1000mg/L)and 10%fetalbovineserum,fresh,warmed Ringer’ssolutionusedasmedium replenishment),sampleon fibronectin-coatedglass
(Young’smodulus)4–100kPaoverthecellsurface,lower aroundthenucleusthanintheperiphery
Hagaetal.(2000)
Breastcancerlines(MCF-10A andMCF7)
Siliconnitride,k=0.01N/m(nominal) modifiedwitha4.5mdiameter polystyrenebead
Contactmode,liquid(culture medium),sampleonglass
(Young’smodulus)0.2–1.2kParange,malignantcellline (MCF7)1.4–1.8timessofterthanbenigncellline (MCF-10A)
Lietal.(2008)
Osteoblasts,mesenchymal stemcellsandosteosarcoma cells
Siliconnitride,k∼20N/m(calibrated usingthermalmethod)
Contactmode,liquid(25mMHEPES), fixedcellsonpolystyrene,glassor collagen-coatedglass
(Young’smodulus)0.7±0.1to2.6±0.7kParange,lower Young’smodulusforMG63osteosarcomacellsoncollagen
Dochevaetal.(2008)
Zonalarticularchondrocytes Sphericalgold-coatedborosilicatebead (5mdiameter),k∼0.065N/m (calibratedbythermalmethod)
Contactmode,liquid(DMEM),sample onpoly-l-lysine-coatedglass
(Young’smodulus)Instantaneousmoduliat0.55±0.23kPa forsuperficial,0.29±0.14kPaformiddle/deepcells;
relaxedmoduliat0.31±0.15kPaforsuperficial, 0.17±0.09kPaformiddle/deepcells;apparentviscosities at1.15±0.66kPasforsuperficial,0.61±0.69kPasfor middle/deepcells
Darlingetal.(2006)
Cardiacmuscle,skeletal muscleandendothelialcells
Siliconnitride,k=0.03to0.05N/m (calibratedusingthermalmethod)
Contactmode,liquid(growth medium),sampleonglassslide
(Young’smodulus)Moduliof100.3±10.7kPaforcardiac muscle,24.7±3.5kPaforskeletalmuscleand1.4±0.1to 6.8±0.4kPadependingontheregiontestedforepithelial cells
Mathuretal.(2001)
Cardiacmyocytes Siliconnitride,k=0.06N/m(nominal) Contactmode,liquid(culture medium),sampleonlaminin-coated petridish
(Young’smodulus)35.1±0.7kPaforcardiomyocytesfrom 4-montholdrats,42.5±1.0kPaforcardiomyocytesfrom 30-montholdrats.
Lieberetal.,(2004a)
LLC-PK1andMDCKkidney epithelialcelllines
Siliconnitride,k=0.12N/m(nominal) Contactmode,liquid(artificialurine) (Young’smodulus)1.5±0.8MPaforLLC-PK1cells, 5±1.5MPaforMDCKcells,oxalatetreatmentdecreases Young’smodulusto1.2±MPaforLLC-PK1cells(other stiffnessparametersalsomeasured)
Rabinovichetal.(2005)
Neuronalgrowthcones Siliconnitride,k=0.006N/m(nominal) Contactanddynamicmodes,liquid (L15/ASWmedium),sampleon poly-l-lysine-coatedglass
(Young’smodulus)3–7kPafortheCdomain,7–23kPafor theTdomain,10–40kPaforthePdomain
Xiongetal.(2009)
Healthyandpathological erythrocytes
Siliconnitride,k=0.03(nominal) Contact,liquid(PBS),sampleon poly-l-lysine-coatedglassandfixedby glutaraldehyde
(Young’smodulus)Moduliof26±7kPaforhealthy erythrocytes,43±21kPaforhereditaryspherocytosis, 40±24kPaforthalassemiaand90±20kPaforG6PD deficiencysamples
Dulinskaetal.(2006)
Liverendothelialcells Siliconnitride,k=0.032N/m Contact,liquid(serum-freeendothelial cellmedium),cellsoncollagen-coated petridisheswithandwithout glutaraldehydefixation
(Young’smodulus)Moduliof2kPaforlivingcellsandover 100kPaforfixedcells
Braetetal.(1998)
Oralsquamouscellcarcinoma, normalandmalignantlines
Siliconnitridetip,k=0.01to0.1N/m;
APTES-modifiedsiliconoxidesphere tip,k∼0.5N/m(calibratedusingSader method)
Contact,air,samplepre-fixedwith2%
PFAandfixedwith3.7%PFA
(Young’smodulus)Medianvaluesof6.75MPafor“normal”
and4.36MPaformetastaticcancercells.Elasticity measurementstakenusingsphere-modifiedtips.
Lasalviaetal.(2015)
PC-3prostatecancercells Siliconnitridetip,k=0.012N/m (calibratedusingthermalmethod)
Contact,liquid(culturemedium), samplestreatedwithanticancerdrugs orDMSOcontrolfor24hpriorto analysis.
(Young’smodulus)c.3kPaforuntreatedcells,increasedto c.6–12kPainadose-dependentmannerfollowingdrug treatment.Frequency-dependencyoftheelasticmodulus wasalsotestedandfoundtochangesignificantlyfor Celebrex,BAY,Totamine,TPAandVPAtreatment,butnot forDSF,MKandTaxol.Thiseffectislinkedtothefactthat theformerdrugsmayaltercrosslinkingratesof cytoskeletalfilaments,whilethelatteronlychangefiber lengthandthickness.
Renetal.(2015)
66A.D.Ozkanetal./Micron89(2016)60–76 Table2(Continued)
Sample Tipproperties Imagingconditions Elasticproperties Reference
Normalandcancerousbladder epitheliumcells
Siliconnitridetip,k=0.011to0.018 (calibratedusingthermalmethod)
Contact,liquid(culturemedium), sampleonglass
(Adhesionenergy)averageof8.17×10−16Jfornormal, 26.95×10−16Jforcancercells
(Young’smodulus)averageof27.57kPafornormal, 2.46kPaforcancercells
Canettaetal.(2014)
Porcinearticularcartilage Siliconnitride,k=0.06N/m(nominal) andborosilicateglassbeadswith r=2.5m,k=0.06and13N/m (nominal)
Contact,liquid(PBS),tissueson poly-l-lysine-coatedglass
(Dynamicelasticmodulus)Ontheorderof2.6MPafor borosilicateglassbeads,about100-foldlowerforsharp tips
Stolzetal.(2004)
Articularcartilageofnormal andarthritic(Col9a1−/−
knockout)mice
Siliconnitride,k=0.06N/m(nominal) andmicrospheres,k=10and12N/m
Contact,liquid(PBS),bulktissuesglued onaroundTeflondisk
(Dynamicelasticmodulus)1.3±0.4MPaformicrospheres (nochangerecordedbetweenages);22.3±1.5kPa, 36.8±1.5kPaand50.9±4.7kPaforsharptipsin1-,10- and19-montholdnormalmice;22.3±1.5kPa,25.5±kPa and27.7±1.1kPainthenon-thickened,intermediateand heavilythickenedcollagenfibersof1-montholdarthritic mice.
Stolzetal.(2009b)
Aorticintimaofrats Notlisted,tipmountedonacustom platformforinvivoAFMimaging
Anaesthetizedlivinganimals (Young’smodulus)0.4–0.5MParangeforbloodvessels withoutdruginfluence,raisedtoc.1.0MPainthepresence ofnitroglycerinanddecreasedbacktoc.0.3MPainthe presenceofnorepinephrine
Maoetal.(2009)
Anteriorhumancornealstroma Phosphorus-dopedsilicon,k=25and 33N/m(calibratedbyanoptical method,asdescribedbySaderetal.)
Contact,liquid(15%dextran),bulk tissueplacedonTefloncellwithout attachment
(Young’smodulus)Between1.14and2.63MPa,consistent acrosstheindentationdepths(between1.0and2.7m)
Lombardoetal.(2012),Sader etal.(1999)
Humancornealbasement membrane
Borosilicateglass,k=0.06N/m (nominal)
Contact,liquid(PBS),a3×3mmtissue piecedissectedandgluedontoawell onsteeldisk
(Young’smodulus)2–15KPa,meanof7.5±4.2kPaaverage fortheanteriorbasementmembrane;20–80KPa, 50±17.8kPaaveragefortheDescemet’smembrane
Lastetal.(2009)
Monkeylenses Gold-coatedtip,k=0.01N/m (nominal),calibratedbythe relationshipbetweenappliedvoltage andcantileverdeflection
Contact,liquid(BSS),lensesdissected andplacedonTeflonslide
(Young’smodulus)1.720±0.88kPa Ziebarthetal.(2007)
Humanbone Various Various (Young’smodulus)16.6±1.1to27.1±1.7fordryadult
tibiae(lowerforchildren),13.4±2.0to22.7±3.1fordry adultvertebrae(lowerforwetsamples),16.58±0.32to 26.6±2.1fordryadultfemoralmidshaft(lowerforwet samplesandinthefemoralneck),otherbone measurementsandtissuehardnessesalsonoted
Thurner(2009)
Bovineoculartendonfibers Siliconnitridetip,k=0.02N/m (calibratedusingthermalmethod)
Contact,air(samplingchamberkeptat 100%humidity),samplegluedonglass petridish
(Young’smodulus)60±2.69MPaforlateralrectus, 59.69±5.34MPaforinferiorrectus,56.92±1.91MPafor medialrectus,59.66±2.64MPaforsuperiorrectus, 57.7±1.36MPaforinferiorobliqueand59.15±2.03for superiorobliquetendons.Differencesbetweentendon elasticitiesarenotstatisticallysignificant.
Yooetal.(2014)
Breasttissuesections Borosilicateglasstip,k=0.06N/m (nominal),individualtipscalibrated usingthermalmethod
Contact,liquid(PBSsuppliedwith proteaseinhibitorsandpropidium iodide),samplessectionedby cryomicrotomy
(Young’smodulus)c.400Painhealthyandnon-invasive tumorregions,4-foldincreaseinaveragestiffnessin invasivetumorfront.Higheraveragestiffnessinthe invasivefrontiscausedbyhighlystiff(>5kPa)regionsin thisarea.Aggressivebreastcancersubtypesarealsofound toexhibithigherYoung’smoduli,quantifiedintermsof upper10%stiffness.
Acerbietal.(2015)
Benignandaggressiveprostate tumors
Siliconnitridetip,k=0.06N/m (nominal),individualtipscalibrated usingthermalmethod
Contact,liquid(physiologicalbuffer), sampleslicedbyrazorandgluedon glass
(Young’smodulus)3.03±0.64kPaforbenign,1.727±1.22 forcanceroustissues.TheaverageYoung’smodulusfor cancersampleswithGleasonscoresinthe2–7rangewas 2.07±1.30,thisvaluewas1.39±0.48forsampleswith Gleasonscoresinthe8–10range.Inaddition,metastatic tumorshadanaveragemodulusof1.06±0.58,while non-metastaticcancertissuehadanaverageelasticity valueof1.99±1.24.Thesevaluesreflectthetissue microenvironmentandcontrastmacro-scaleelastography results,inwhichmoreaggressivecancersarestiffer.
Wangetal.(2014)
