Restricted cell functions on micropillars are alleviated by surface-nanocoating with amino groups

29 Pages·2017·1 MB·English

by  Mörke, Caroline

©2018.PublishedbyTheCompanyofBiologistsLtd|JournalofCellScience(2018)131,jcs207001.doi:10.1242/jcs.207001
RESEARCHARTICLE
Restricted cell functions on micropillars are alleviated by surface-
nanocoating with amino groups
CarolineMoerke1,SusanneStaehlke1,HenrikeRebl1,BirgitFinke2andJ.BarbaraNebe1,*
ABSTRACT implantintothenativebone,whichisachievedbyacceleratingthe
onsetandtherateofimmediateosteoblasticcellattachmentaswell
Thetopographicalandchemicalsurfacefeaturesofbiomaterialsare
asproliferation(Bacakovaetal.,2011).Therefore,implantsurface
sensedbythecells,affectingtheirphysiologyattheinterface.When
designwithitstopographicalaswellaschemicalproperties,hasa
placed on titanium, we recently discovered osteoblasts attempted
huge impact on osseointegration. Implant materials have to fulfill
caveolae-mediatedphagocytosisofthesharp-edgedmicrostructures.
certain requirements concerning corrosion resistance and
This active, energy-consuming process resulted in decreased
biocompatibility. Titanium (Ti) is an inert and biocompatible
osteoblastic cell functions (e.g. secretion of extracellular matrix
material and therefore commonly used as an orthopedic implant
proteins). However, chemical modification with plasma polymerized
material (Bacakova et al., 2011). Osseointegration is enhanced on
allylamine (PPAAm) was able to amplify osteoblast adhesion and
roughertitaniumsurfacesratherthanonsmoothones,butitisalso
spreading,resultinginbetterimplant osseointegrationinvivo.Inthe
accompaniedbychangesinthecellphysiology,suchastheintegrin
present in vitro study, we analyzed whether this plasma polymer
expression, if the titanium surface features sharp edges (Lüthen
nanocoating is able to attenuate the microtopography-induced
etal.,2005;Bacakovaetal.,2011).Newimplantdesignstrategies
changes of osteoblast physiology. On PPAAm, we found cells
pursuethedevelopmentofnewbioactivesurfacestoevokecellular
showeda highercellinteraction withthe geometrical micropillarsby
responses that promote osseointegration (Wennerberg and
30min, and a less distinct reduction in the mRNA expression of
Albrektsson, 2009); these include the coating of the surface to
collagentypeI,osteocalcinandfibronectin after24hofcellgrowth.
change its chemistry. In this case, coating with polymerized
Interestingly, the cells were more active and sensitive on PPAAm-
allylamine (PPAAm), which can be deposited by a physical low-
coatedmicropillars,andreactwithasubstantialCa2+ionmobilization
pressureplasma process,leadsto an improvedosseointegration of
afterstimulationwithATP.Theseresultshighlightthatitisimportantfor
titaniumimplantsinvivo(Gableretal.,2014),whichmaybecaused
osteoblaststoestablishcellsurfacecontactforthemtoperformtheir
bytheenhancedcelladhesionandspreadinginvestigatedindetailin
functions.
vitro (Rebl et al., 2012; Finke et al., 2007; Kunz et al., 2015).
KEYWORDS:Cell–substrateinteraction,Osteoblast,Ca2+ PPAAm is a nanometer-thin, positively charged amino-
mobilization,mRNAexpression,Microtopography,Plasmapolymer functionalized polymer layer that renders the surface more
nanocoating hydrophilic (Finke et al., 2007). Regular geometric micropillar
topographies with the dimension of 5μm in pillar length, width,
INTRODUCTION height and spacing (P-5×5) have been used as artificial surfaces,
Cellsaresensitivetotheirunderlyingmicro-andnano-topography; extendingtheworkofstochasticsurfacemodelswiththeadvantage
in particular topographical features on microscale level offer cues ofconstantandrecurringtopographyvariables(Lüthenetal.,2005).
that cause diverse cell responses. These responses range from Previous studies have shown that osteoblastic cells mimic the
alteredadhesionbehavior,resultinginchangedintegrinexpression underlying geometrical micropillar structure within their actin
aswellassignaling,toadisturbedcellfunction(Lüthenetal.,2005; cytoskeleton, and we recently discovered an attempted caveolae-
Staehlke et al., 2015; Moerke et al., 2016). Because of steadily mediated phagocytosis of each micropillar beneath the cells
increasing human lifespans andtheassociatedrisingrequirements (Moerke et al., 2016). Characteristic for this process was the dot-
placedonbodyparts,temporaryandpermanentorthopedicimplants likecaveolin-1(Cav-1)proteinandcholesterolaccumulationonthe
aregainingmoreandmorerelevance.Anorthopedicimplantshould micropillarplateausafter24h.Cav-1andcholesterolarethemajor
establishamechanicallysolidinterfacewithcompletefusionofthe components of caveolae and are essential for the formation and
material within the native bone tissue. This process is called stabilization of the caveolar vesicles (Parton and del Pozo, 2013).
osseointegrationandcandetermine thesuccessfulingrowthofthe Caveolae are a specialized form of cholesterol and sphingolipid-
enriched plasma membrane subdomains, called lipid rafts,
e
distinguish themselves via the containment of the caveolin-1 c
1UniversityMedicalCenterRostock,Dept.ofCellBiology,Schillingallee69,18057 n
Rostock,Germany.2Leibniz-InstituteforPlasmaScienceandTechnologye.V. protein. These specialized plasma membrane domains are e
(INP),Felix-Hausdorff-Str.2,17489Greifswald,Germany. involved in various cellular processes, including phagocytosis ci
S
(Parton and del Pozo, 2013; Pelkmans and Helenius, 2002). The
*Authorforcorrespondence([email protected]) ll
attemptedcaveolae-mediatedmicropillarphagocytosisweobserved e
C
C.M.,0000-0001-9461-2128;H.R.,0000-0001-6000-1724;B.F.,0000-0002- wasaccompaniedbyincreasedintracellularreactiveoxygenspecies
f
5941-0376;J.B.N.,0000-0003-2662-0905 (ROS) production, reduced intracellular ATP levels and a higher o
ThisisanOpenAccessarticledistributedunderthetermsoftheCreativeCommonsAttribution mitochondrialactivity(Moerkeetal.,2016).Aconsequenceofthis al
License(http://creativecommons.org/licenses/by/3.0),whichpermitsunrestricteduse, energy-consuming process was the reduction of the osteoblast n
distributionandreproductioninanymediumprovidedthattheoriginalworkisproperlyattributed. r
marker production, namely extracellular matrix (ECM) proteins u
o
Received2June2017;Accepted5November2017 involvedinthegenerationofnewbonetissue,forexample,collagen J
1
RESEARCHARTICLE JournalofCellScience(2018)131,jcs207001.doi:10.1242/jcs.207001
Fig.1.Preparationofgeometricmicro-
pillarmodelsurface.(A)Schematic
illustrationofthedeepreactiveion
etchingprocessforthegenerationof
micropillartopographyof5×5×5µm
(width×length×height).(B)SEMimagesof
theplanarreference(Ref)andthe
micropillars(P-5×5)ofwithaschematic
sideview.
type I (Col1) and fibronectin (FN). As a result, the cells on the Thecaveolin-1andactinfilamentsofcells
micropillarsshoweddiminishedosteoblastcellfunction,whichwas Acharacteristicoftheattemptedcaveolae-mediatedphagocytosison
alsofoundonstochasticallystructured,corundum-blastedtitanium microstructures isacleardot-like clusteringofCav-1 accompanied
withspikyelevations(Moerkeetal.,2016).Thisindicatesthatthe withaccumulated,shortfilamentsofactinonthemicropillarplateaus.
given surface microtopography also strongly affects the cell The question arose, whether the osteoblasts still express the Cav-1
physiology in a negative sense if surface characteristics are sharp clusters and the clustered actin filaments on the pillar tops after
edged. In this study, we wanted to shed light on the question of coating with PPAAm. Immunofluorescence labeling reveals that
whether a chemical surface modification such as PPAAm, which there was less Cav-1 cluster formation on the P-5×5+PPAAm
has a positive impact on cell spreading, adipose-derived stem cell (Fig.4).Cav-1wasfoundtobemoreconcentratedaroundthepillar
differentiation(Liuetal.,2014)andosseointegration,canalleviate wallsandinregionsofthepillaredgesafterthe24hgrowthperiod.
thismicrotopography-inducednegativecellularoutcome. The actin cytoskeleton was accumulated in a ring-like fashion
around the PPAAm-coated micropillars after 24h (Fig. 5A). The
RESULTS correlative light and electron microscopy (CLEM) is a method to
Nanocoatingandsurfacecharacteristics demonstrate the cell components (e.g. the actin filaments) as
Inthisstudy,weusedsubstratesconsistingofsiliconwithafinal observed in our experiments, in precise positions relative to the
coatingof100nmtitanium.Themicrotopographywasfabricated sampletopography(i.e.themicropillars).Thus,itisimpressivethat
by deep reactive ion etching (Fig. 1). We wanted to find out it can be visually seen that the pillars are enveloped by the actin
whether cell functions that are restricted on the periodically filaments(Fig.5B).
microtextured samples can be alleviated by surface nanocoating
with amino groups. To chemically functionalize a biomaterial Geneexpression
surface the deposited nanolayer should have a homogenous ThequestionwaswhetherthePPAAmcoating,withitsadvantage
distribution. Therefore, a surface characterization using X-ray ofenhancedcell–substratecontact,isabletopositivelyinfluencethe
photoelectronspectroscopy(XPS)todetecttheelementalsurface
composition is mandatory for the detection of a pinhole-free,
chemicallycoatedlayer.Thedensityoftheaminogroups(ratioof
NH to carbon atoms) of the plasma polymerized allylamine
2
(PPAAm)nanolayerwas∼3%andthefilmthickness∼25nmdue
totheplasmadepositiontimeof480s.AfterthePPAAmcoating,
no titanium (Ti) or silicon (Si) components were found on the
surface(Fig.2).
Nanocoatingandcellmorphology
The micropillars were coated with PPAAm to enhance the cell–
substratecontactbyincreasingthesurface-occupyingcellarea.As
e
showninFig.3,theenhancedcellspreadingafterPPAAm-coating c
n
is impressive enough to be seen visually. The scanning electron e
i
microscopy (SEM) images show widely spread-out cells that are c
S
alreadyreachingthebottomofthemicrotopographyafter30minon l
l
the PPAAm-coated pillar surfaces. On the uncoated micropillars, e
C
thecellsexhibitedamoresphericalformandsitonamaximumof
f
fourpillars,whereasonPPAAm,thecellscoveredmorethanfour o
pillars. After 24h of cultivation, the changes were no longer as Fig.2.SurfacecharacterizationofthematerialsubstratesviaX-ray al
photoelectronspectroscopy.Uncoatedsamples(P-5×5,left)andplasma n
drasticasthoseseenafter30min.TheosteoblastsonthePPAAm-
polymer-coatedpillars(P-5×5+PPAAm,right)wereanalyzed.Notethatafter r
coatedmicropillarswerestillmorespreadout,withthemicropillars u
PPAAmfunctionalization,titanium(Ti)andsilicon(Si)arenotvisible,indicating o
moreimprintedintothecells. ahomogenousnitrogen(N)-containinglayer.(XPS,AxisUltra,Kratos). J
2
RESEARCHARTICLE JournalofCellScience(2018)131,jcs207001.doi:10.1242/jcs.207001
Fig.3.SpreadingofMG-63osteoblastson
micropillars.Notethatafter30minonuncoated
samples(P-5×5)thecellisrestrictedtofourpillars,
whereasonplasmapolymer-coatedpillars(P-5×5
+PPAAm)thecellareaisincreaseddramatically.
Thisadvantageintheinitialphaseofoccupyinga
surfaceisnearlyrecoveredafter24h.Toprow,
3000×magnification,scalebars4µm;bottomrow,
1000×magnification,scalebars10µm.
osteoblast marker expression. The uncoated micropillars led to hyaluronan plays a key role in initial interface interactions and is
significantly reduced gene expression of collagen-I (Col1), essentialforthefirstencounterofcellswiththesubstrate(Fig.7).We
fibronectin (FN) and osteocalcin (OCN) (Fig. 6A), whereas, after found that the initial attachment, as well as the contact-mediated
PPAAmcoating,thesegenesshowedonlyatrendtowarddecrease spreading on titanium, is hampered if hyaluronan is cleaved by the
or no difference to the planar reference. However, the gene enzymehyaluronidase(600U).
expression of alkaline phosphatase (ALP) was significantly
diminished on uncoated as well as PPAAm-coated micropillars Ca2+signaling
relative to on the planar reference (Fig. 6A). In our previous InlivingosteoblastsonPPAAm,wecouldidentifyhighlyactivecells
experiments we observed a significantly increased Cav-1 gene by analysis of the Ca2+ mobilization capacity. The basal level of
expressiononthemicropillars(Moerkeetal.,2016).ThePPAAm intracellularCa2+ions(0–180s)wasalreadysignificantlyhigherin
coatinginfluencedthisprocesssuchthattheCav-1geneexpression osteoblastsestablishedfor24honP-5×5+PPAAmthanonuncoated
showedonlyatrendtowardelevation(Fig.6B). P-5×5withmeanfluorescenceintensityvalues(MFI)of42.03±0.05
versus 26.18±0.14 (mean±s.e.m.), respectively (Fig. 8A). It is
Celladhesion notable that these 24h-adherent cells on PPAAm react intensively
An important question remains: how doesthe PPAAm nanocoating afterATPstimulationwithahighincreaseofintracellularCa2+ions.
enhance the initial cell spreading? The pericellular matrix substance ThedifferenceintheCa2+dynamicsoffluo-3-stainedcellsonP-5×5
Fig.4.Localizationofthemembrane
componentCav-1inosteoblastson
micropillars.Notethelowerlevelsofcluster
formationforCav-1onplasmapolymer-coated
pillars(P-5×5+PPAAm)incontrasttoonthecontrol
samples(P-5×5)after24h.Imagesareofz-stacks,
magnifiedviewsarea5×zoomoftheareamarked
withdashedlines,ora10×zoomofthearea
markedwithcontinuouslines.Scalebars:left,
20µm;5×zoom,5µm;10×zoom,2µm.
e
c
n
e
i
c
S
l
l
e
C
f
o
l
a
n
r
u
o
J
3
RESEARCHARTICLE JournalofCellScience(2018)131,jcs207001.doi:10.1242/jcs.207001
Fig.5.Actincytoskeletonofosteoblastson
micropillars.(A)Onplasmapolymer-coated
pillars(P-5×5+PPAAm)theactincytoskeletonis
organizedinatypicalring-likestructure(arrows)
comparedtothemoreclusteredphenotypeon
controls(P-5×5)after24h.Imagesareofz-stacks.
Scalebars:left,20µm;middle,5µm;right,2µm.
(B)Correlativelightandelectronmicroscopy
(CLEM)asamethodtodemonstratethecell
components–theactinfilaments–inprecise
positionstothesamplestructures–the
micropillars.Thearrowspointtotheactin
cytoskeletonsurroundingthemicropillars.Inset:
thereflectionmodeofthemicroscopeallows
visualizationofthepillars(green,falsecolor),
whichareenvelopedbytheactinfilaments(red).
Scalebars:10µm;5µm(inset).
+PPAAm (MFI peak at 188–246s, 96.45±5.36; maximum MFI functional surface groups (PPAAm) was fabricated. The deposited
value, 130.35±13.11) is significant compared to cells grown on P- PPAAm coating of titanium renders the surface more hydrophilic
5×5 (MFI peak at 188–246s, 44.43±1.84; maximum MFI value with water contact angles for PPAAm-coated titanium versus
53.36±10.78)(mean±s.e.m.)(Fig.8A).Thesamecellularreactivity untreated titanium (planar) of 48° and 78°, respectively (Finke
was found in human primary osteoblasts (HOB) after 24h etal.,2007;Liuetal.,2014),andcreatespositivelychargedsurfaces
cultivation: the Ca2+ mobilization capacity on P-5×5+PPAAm (theζpotentialofPPAAm-coatedtitaniumversusuntreatedtitanium
(MFI peak at 188–246s, 91.81±1.13; maximum MFI value, is+7.7mVand−3.4mV,respectively)(Finkeetal.,2007).Owingto
100.46±6.47) was significantly higher compared to HOB cells on thesesurfacecharacteristics,enhancedcellspreadingafterPPAAm-
P-5×5 (MFI peak at 188–246s: 66.62±2.52, maximum MFI value coating was to be seen. We observed previously that the PPAAm
82.49±7.89). coatingpromotestheinitialcellspreadingafterseedingofthecells
After 30min of cell cultivation on P-5×5+PPAAm, MG-63 andincreasedthecellmobility(Finkeetal.,2007;Rebletal.,2016).
osteoblastsshowednotonlyamorespreadmorphologycomparedto The PPAAm coating has proven to be a cell adhesive layer and
thatseenonuncoatedpillarsasshowninFig.3(30min)butalsoa osteoblasts can overcome topographical restrictions (e.g. contact
significantly higher Ca2+ mobilization after stimulation with ATP guidance due to machined grooves) (Rebl et al., 2012, 2016). We
(Fig. 8B) (P-5×5+PPAAm, MFI peak at 188–246s, 81.40±2.89, observed that the osteoblastic cells try to internalize surface-fixed
e
with maximum MFI value 107.30±40.90; P-5×5, 49.3±0.99, with micropillars (Moerke et al., 2016) and hypothesized that cells c
n
maximum MFI value 53.88±19.1). The basal Ca2+ signal had possiblyincreasetheircell-surfacecontact,becausethiscellsurface e
i
already increased after 30min of adherence on P-5×5+PPAAm contactisessentialforthemaintenanceoftheosteoblast-specificcell c
S
(MFImean±s.e.m.:55.34±0.14)butthiswasnotsignificantversus function(SaldañaandVilaboa,2010).Enhancedspreadingaswellas l
l
P-5×5 (MFI mean±s.e.m.: 37.78±0.13) as determined by a direct moldingintothetopographyisalsoobservedonstochasticsurfaces e
C
statisticalcomparisonusingamultiplet-test (Fig.8B). afterPPAAmcoating(e.g.onmachinedtitanium;Rebletal.,2012;
f
Finkeetal.,2007). o
l
DISCUSSION a
n
Cellmorphology Caveolin-1andactinfilaments r
The elemental distribution, as measured by XPS, indicated that a Theattemptbythecelltophagocytizemicrostructuresresultsina u
o
homogeneous,pinhole-freelayeroftheplasmapolymerwithamino dot-like clustering of Cav-1 that is accompanied by short actin J
4
RESEARCHARTICLE JournalofCellScience(2018)131,jcs207001.doi:10.1242/jcs.207001
Fig.6.RelativemRNAexpressionofosteoblastsafter
24honmicropillars.(A)Expressionlevelstheosteoblast
markerproteinsalkalinephosphatase(ALP),collagentypeI
(Col1),fibronectin(FN)andosteocalcin(OCN).Notethatall
showrestrictedexpressiononuncoatedcontrols(P-5×5),
whichisalleviatedbytheplasmapolymernanolayer(P-5×5
+PPAAm),exceptforALP.(B)Cav-1,asamembrane
componentthatisinvolvedinphagocyticprocessesis
adaptedtotheplanarreferencesampleifthepillarsare
coatedwiththeplasmapolymer.Inallexperiments,Ref
valuesarenormalizedon1andn=4.*P<0.05(Mann–
WhitneyUtest).
filamentsonthepillarplateaus(Moerkeetal.,2016).Tobesurethat uncoatedmicropillarsseenby72h(butnotearlier)andatupto96h
this is a common phenomenon of cells on micropillars, we also cultivation time (Moerke et al., 2016). Comparing our MG-63
observedtheCav-1distributioninhumanprimaryosteoblastsand osteoblastic cells with human primary osteoblasts, we previously
humanprimaryfetalosteoblasts,aswellastheosteoblasticcelllines recognized that the same accumulation of the actin filaments that
Saos-2andU-2Osandfoundthesameaccumulatedpattern(Moerke wasseeninMG-63osteoblasticcellswasalsoseenintheprimary
et al., 2016). The Cav-1 cluster formation on the P-5×5+PPAAm osteoblasts on pillared microstructures except that primary cells
after24hwascomparablewithCav-1onuncoatedmicropillarsafter additionallyhavesomelongeractinfibers(Moerkeetal.,2016).
thelongercellcultivationperiodofupto96h(Moerkeetal.,2016). Plasma membrane reorganization works in concert with the
The accumulated actin cytoskeleton around the PPAAm-coated underlying actin cytoskeleton to mediate morphological changes
micropillars after 24h was similar to that of the cells on the and modify cell surface contacts (Head et al., 2014). Actin
Fig.7.Schemashowingtheimportanceofthehyaluronancoat
ofthecellsoninitialcellspreading.Left:cellareasofnormal
e
osteoblasts.Right:cellareasofhyaluronidase(HAdase)-treated
c
osteoblasts.Normalcellscanspreadwellafter30minonpure n
e
titanium,whereasHAdase-treatedcellsarehamperedintheir
i
surfaceoccupationprocessat30min.Note:theschemaare Sc
presentedfromanoriginalSEMimageofMG-63osteoblastson l
l
glass-blastedtitanium,HAdase600Utreatment(formethodsee e
C
Finkeetal.,2007).
f
o
l
a
n
r
u
o
J
5
RESEARCHARTICLE JournalofCellScience(2018)131,jcs207001.doi:10.1242/jcs.207001
Fig.8.IntracellularCa2+-mobilizationafterATP
stimulation(180s)inlivingMG-63osteoblastson
micropillars.(A)Influenceoftheplasmapolymer
nanolayer(P-5×5+PPAAm)onthetime-dependentCa2+
signalafterATPstimulationofcellsallowedtoadherefor
24h.Notethesubstantial(significantatP<0.05,multiple
t-test)increaseinintracellularCa2+onP-5×5+PPAAm.In
addition,thebasalCa2+levelisalreadysignificantlyhigher
comparedtothatseenforcellsontheuncoatedcontrol
(P-5×5).(B)Influenceoftheplasmapolymernanolayer
(P-5×5+PPAAm)onthetime-dependentCa2+signalafter
ATPstimulationofcellsallowedtoadherefor30min.
TheincreaseinintracellularCa2+onP-5×5+PPAAmisclear
andsignificant(P<0.05,multiplet-test)intherangeof
244–316s.Cellswerestainedwithfluo-3,three
independentexperimentsfor10cellspertimepoint(240
analyses).Resultsaremean±s.e.m.
remodelingisespeciallysensitivetosignalsthataregeneratedatthe favorable;thegeneexpressionofALPwassignificantlydiminished
membrane–cytoplasm interphase (Janmey and Lindberg, 2004). also on PPAAm-coated micropillars. ALP is involved in the
Theactinreorganizationatthemicropillartopographywasshownto mineralizationoftheboneECM.Thisleadstotheassumptionthat
becharacteristicforanattemptatphagocytosis.Ourpreviousstudy the ALP gene expression is not induced by lessened cell surface
revealedahigh-energyrequirement(ATPlossof1.45-fold)thatwas interaction, but more by the phagocytic process itself. If a process
causedbythecellattemptingphagocytosisofeverysinglepillarina requires high energy demands, such as phagocytosis, other energy
fixedposition(Moerkeetal.,2016).Thisleadstothequestion:why expendituresmustbeabated,forexample,theexpressionofproteins,
dothecellsundergothenuisanceofpillarphagocytosiswhenthisis suchasALP,whichareredundantinthecasethatcellsproduceless
accompanied by high energy requirements and negative cellular ECMorareoccupiedwiththeestablishmentofsufficientcell–ECM
outcome?Apossibleexplanationforthiscellbehaviorwouldbethe contact(AtaullakhanovandVitvitsky,2002).Inparticular,itistobe
factthatosteoblasts areattachment-dependentcellswhichwantto taken into account that high energy demands are placed on
ensure the highest cell surface contact. The maintenance of the osteoblasts not only during phagocytosis, but also during the
osteoblastic function relies on this contact; therefore acertain cell productionoftheECM,byexpressingandsecretingECMproteins
surface contact is needed for adequate osteoblastic cell function (Komarova et al., 2000). However, in experiments with polished
(Saldaña and Vilaboa, 2010). Surfaces, such as the geometrical titanium (roughness average Ra=0.19μm) the PPAAm coating
micropillarsaswellasroughcorundum-blastedtitanium,whichisa promoted the ECM expression of MG-63 cells, as is seen by the
commerciallyusedorthopedicimplantdesign,offertheosteoblasts middle-to-late-stage differentiation-related ALP and bone sialo
insufficient surface interaction area and induce internalization protein(BSP)mRNAproduction(Nebeetal.,2007).
processesforthesurfacefeatures(Moerkeetal.,2016).Theselead
toimpairedosteoblastcellfunction(Lüthenetal.,2005)aswellas Celladhesionmechanism
biomaterial–tissueinteractionsinthefirstphaseofosseointegration. Thepericellularmatrixsubstancehyaluronanplaysakeyroleininitial
AcoatingsuchasPPAAm,whichenhancesthecellattachmentand interfaceinteractions(Zimmermanetal.,2002).Hyaluronanisahigh
spreadingandtheinteractionwiththetopography,hasbeenshown molecular mass (nearly 7000kDa) linear polysaccharide originally
e
toreversethesetopography-inducednegativeeffects(Moerkeetal., foundinsofttissuethatconsistsofN-acetyl-D-glucosamineandD- c
n
2017) and may also be beneficial for the improvement of glucuronicacids,responsibleforthenegativechargeofthemolecule e
i
commerciallyusedroughsurfaces. (Chopraetal.,2014).Theglycosaminoglycanhyaluronanisproduced c
S
bythecellsitself,whichcausesanegativelychargedsphericalshell l
l
Geneexpression [e.g.completelyenvelopingchondrocyteswith4.4μminthicknessas e
C
Inourpreviousexperimentsonuncoatedmicropillars,theexpression measuredbyCohenetal.usingSEM(Cohenetal.,2004;Zimmerman
f
ofosteoblastmarkergeneswasdrasticallyreducedafter24hdueto et al., 2002)]. Nebe and Luethen (2008) showed that a positively o
the attempt to phagocytize these fixed structures (Moerke et al., chargedPPAAm-coatingonatitaniumsurfacecouldboosttheinitial al
n
2016).Inour newexperiments,we revealedthatthisrestrictedcell cell adhesion, including fast formation of adaptor proteins in focal
r
function on micropillars is alleviated by surface-nanocoating with adhesions. They hypothesized that a positively charged surface u
o
PPAAm.However,forALPexpressionthePPAAmcoatingwasnot providesthebasisforhyaluronan-mediatedattachmentandspreading. J
6
RESEARCHARTICLE JournalofCellScience(2018)131,jcs207001.doi:10.1242/jcs.207001
Hyaluronan mediates the immediate adhesion of osteoblasts to the controls,planewafers(Ref)wereemployed.Thesampleswerefabricatedby
materialsurfacewithinseconds,andthefirstcontactishamperedon deep reactive-ion etching (DRIE) (Center for Microtechnologies ZFM,
different surfaces, for example, titanium or even collagen-I, if UniversityofTechnologyChemnitz,Germany)onsilicon(Si)wafersand
hyaluronan is cleaved by the enzyme hyaluronidase. For instance, coatedwithanadditional100nmtitanium(Ti)layer,asreportedpreviously
thecelladhesion(at5min)issignificantlyreducedby∼5-foldifcells (Staehlkeetal.,2015;Moerkeetal.,2016).
have to adhere to pure, polished titanium without their hyaluronan
Plasmapolymernanocoatingandsurfacecharacterization
coat,andstillreducedby1.7-foldonacollagen-I-coatedsurface(Nebe
Plasma polymerized allylamine (PPAAm) functionalization was
andLuethen,2008;Finkeetal.,2007).
performed in a low-pressure plasma process reactor V55G (plasma
Inourstudy,theenergyconsumptionofcellscould,unfortunately, finish, Germany, V=60l) in a two-step procedure with the precursor
notbemeasuredduetothecharacteristicsofthePPAAmnanolayer. allylamine(Rebletal.,2012).Theplasmadepositiontimeforthepulsed
PPAAm exhibits a slight autofluorescence, especially in the green plasma process was 480s gross. The elemental chemical surface
channel.Thiswouldinterferewithquantitativeimmunofluorescence composition was determined by high-resolution scanning X-ray
analysisofthesestructures.Additionally,thePPAAmcoatingdoes photoelectron spectroscopy (XPS). The Axis Ultra (Kratos, UK) ran
notallowfor thetrypsinizationofthecellsinavitalstatefromthe withthemonochromaticAlKαlineat1486eV(150W).Thespotsizefor
high-resolution measurements was limited to 250μm. Spectra were
micropillaredsurfacesduetoitshighcell-attractivecharacteristicsas
recordedatapassenergyof80eV(Finkeetal.,2007).
measuredpreviouslybyaspinningdiscdevice(Fritscheetal.,2010).
MG-63cellsonPPAAmwere1.5timesmoreresistanttoshearstress
Osteoblastcellculture
compared to cells on uncoated titanium alloys (Ti6Al4V) (Gabler
The human osteoblast-like cell line MG-63 (American Type Culture
et al., 2014). For this reason, further experiments, such as Collection ATCC®, CRL1427™) was used throughout the experiments.
determination of the cellular cholesterol amount, intracellular ATP MG-63 cells have an integrin subunit profile similar to that of primary
as well as reactive oxygen species (ROS) generation could not be humanosteoblastsandwereconsideredapplicableforstudyinginitialcell
investigatedinrelationtouncoatedmicropillars.Instead,weobserved attachment processes to surfaces (Czekanska et al., 2012). In our own
theCa2+dynamicsinlivingcells. experiments, we have found similarities in the formation of β1- and β3-
integrinreceptorsinMG-63osteoblastsgrownonstructuredtitanium(e.g.
Ca2+signaling glass-blasted,corundum-blasted)thatwerecomparablewithhuman-derived
primary osteoblasts (Lüthen et al., 2005). The cells were grown in
Staehlkeetal.(2015)demonstratedpreviouslythenegativeinfluence
Dulbecco’s modified Eagle’s medium (DMEM, Life Technologies
of the cubic pillar topographies on Ca2+ signaling and dynamics
GmbH, Darmstadt, Germany) with 10% fetal calf serum (FCS)
comparedtothatseenwithaplanartitaniumsurface.Here,wecould
(Biochrom FCS Superior, Merck KGaA, Darmstadt, Germany) and 1%
show that cells on these pillars coated with an additional plasma-
gentamicin(RatiopharmGmbH,Ulm,Germany)at37°Cinahumidified
nanolayer (PPAAm) with positively charged amino groups can atmosphere with 5% CO . Human primary osteoblasts (HOB, C-12720,
2
overcome the negative influence of the pillared topography and PromoCell GmbH, Heidelberg, Germany) were cultured in osteoblasts
increasetheiractivitydespitethesharpedgedsurfacestructure.This growth medium with SupplementMix (PromoCell) and 1% antibiotic-
may also contribute to the positive outcome in cell function antimycotic (Anti-Anti 100×, Life Technologies). In the experiments,
parameters typical for osteoblasts, for example, collagen-I and 15,000cells/cm² were seeded onto the samples placed in NUNC 4-well
osteocalcin expression (see Fig. 6). The local physical dishes(ThermoFisherScientific,NUNCGmbH&Co.KG,Langenselbold,
Germany).Beforeuse,theuncoatedtitaniumarrayswerewashedin70%
microenvironmentcannotonlymodulatecellshapeandmechanics
ethanol for 10min and rinsed three times in phosphate-buffered saline
(BellasandChen,2014),butalsotheactivityofcells,asshownhere
(PBS)(SigmaAldrich,Munich,Germany).
forintracellularCa2+levels.Althoughcellsonpillars‘feel’thesame
microtopography(range5μm)theadditionalnanolayer(PPAAm)is
Real-timequantitativePCRformRNAexpression
decisivetotriggerthecellsintoamoreactivestate. Total RNA was isolated after 24h cell growth on the samples using a
We could shed light on how plasma polymer nanocoatings can NucleoSpin® RNA II kit (Macherey-Nagel GmbH & Co KG, Dueren,
stronglyalleviatetherestrictionsofamicrotopographyconcerning Germany),whichincludesastepfortheeliminationofanygenomicDNAby
cell functions and cell signaling. This could have importance for DNase(Macherey-Nagel)treatment.Thepurityandquantityoftheresulting
implant surface design and production in cases where mechanical RNAsweredeterminedviameasurementoftheabsorbanceat280nmand
aspectsofanimplanthavetobeconsidered. 260nmwiththeNanoDrop1000(ThermoScientific).50ngoftotalRNA
wasusedforfirst-strandcDNAsynthesisbyusingSuperscript®IIreverse
transcriptaseandrandomprimers(InvitrogenAG,Carlsbad,CA,USA).The
Conclusion
real-timequantitativepolymerasechainreaction(RT-qPCR)wasperformed
Thepresentstudyconfirmsthehypothesisthatosteoblastswantto
withTaqMan®UniversalPCRMasterMixandTaqMan®geneexpression
create the highest possible cell surface contact to maintain their
assaysforalkalinephosphatase(ALP)(Hs00758162_m1),caveolin-1(Cav-
osteoblast-specific function, as displayed by the production and 1)(Hs00971716_m1),collagentype1(Col1)(Hs0016404_m1),fibronectin
secretion of ECM proteins. Moreover, this analysis demonstrates (FN) (Hs00900054_m1), glyceraldehyde-3-phosphate dehydrogenase e
thatchemicalsurfacemodificationwithaplasmapolymernanolayer (GAPDH) (Hs99999905_m1) and osteocalcin (OCN) (Hs01587813_g1) c
n
is able to attenuate the microtopography-mediated cell alterations. (allAppliedBiosystemsbyLifeTechnologiesGmbH,Darmstadt,Germany) e
This PPAAm-coating process could contribute to improved cell followingthemanufacturer’sinstructions.TheTaqMan®PCRassayforeach ci
S
acceptance of rough surfaces with extraordinary topographies, for targetgenewasperformedintriplicateforfourindependentexperiments.The l
examplecorundum-blastedtitanium. PCRwasperformedwithaRT-PCRAppliedBiosystem7500machineand el
the data were collected and analyzed by the 7500 System SDS Software C
(AppliedBiosystems).EachexpressionwascalculatedrelativetoGAPDH f
o
MATERIALSANDMETHODS (ΔCt)andthenrelativetothereferences(ΔΔCt).
l
Titaniumsurfaces a
n
Periodicallymicrotexturedsamples(size1cm²)witharegularcubicpillar Immunofluorescencestaining r
u
geometry on the surface having a dimension of 5×5×5μm in MG-63 osteoblastic cells were cultured on the titanium arrays for 24h, o
width×length×heightand5μminspacing(P-5×5)wereused(Fig.1).As washed three times with PBS and then fixed with 4% paraformaldehyde J
7
RESEARCHARTICLE JournalofCellScience(2018)131,jcs207001.doi:10.1242/jcs.207001
(PFA) (10min; room temperature) (Sigma-Aldrich). After washing three Living cell observations were performed as follows. The analysis of
times with PBS, the cells were permeabilized with 0.1% Triton X-100 intracellularCa2+mobilizationwasperformed ontheconfocalmicroscope
(10min,roomtemperature,Merck),washedagainthreetimeswithPBSand LSM780(CarlZeiss)equippedwithanargon-ionlaser(excitationat488nm)
blocked with 2% bovine serum albumin (BSA) (Sigma-Aldrich) in PBS usingaZeiss40×waterimmersionobjective(CApochromat,1.2WKorr
(30min,roomtemperature). M27). The measurement of the mean fluorescence intensity (MFI) of the
Foractinfilamentstaining,cellswereincubatedwithphalloidincoupled globalCa2+signalwascarriedoutbyZen2011(SP4,black)software(Carl
totetramethylrhodamine(TRITC)(5μg/mlinPBS,Sigma-Aldrich).Nuclei Zeiss)withthemode‘timeseries’. Images(512×512pixels)fromatleast
werestainedfor10minwithHoechst33342dye(1mg/ml;SigmaAldrich, threeorfour(30minand24h,respectively)independentexperimentsfor10
1:2000inPBS). cellspertimepoint(for240cycles)weremeasuredatfullpinhole(13.5μm
Caveolinwasstainedwiththepolyclonalrabbitanti-caveolin-1(cat.no. section)andthesamesettings.Duringrecordingofthetimeseriesover240
3238,NewEnglandBiolabsGmbH,1:400)asprimaryantibodyfor1hat cycles (2s each) adenosine 50-triphosphate (ATP, 10μl, SERVA
room temperature followed by the secondary antibody anti-rabbit-IgG Electrophoresis GmbH, Heidelberg, Germany) was added after the 90th
conjugatedtoAlexaFluor488(LifeTechnologies,diluted1:200inPBS) cycleforstimulatingtheintracellularCa2+mobilization.Theevaluationof
appliedatroomtemperaturefor30min.Thesampleswereembeddedwith fluorescencesignalsindefinedareasofindividualsinglecellsovertimewas
FluoroshieldTMmountingmedium(Sigma-Aldrich). carriedoutbyusingZEN2(blueedition,CarlZeiss)usingthefunction‘mean
ROI’.
Confocallaserscanningmicroscopy
Fluorescentcellswereobservedusinganinvertedconfocallaserscanning Statisticalanalysis
microscope (LSM 780; Carl Zeiss AG, Oberkochen, Germany) with a Statistical analyses were carried out with GraphPad Prism5 software
ZEISSoilimmersionobjective(C-Apochromat63).Forimageacquisition, (GraphPadSoftwareInc.,LaJolla,CA).Resultsarepresentedinboxplots
theZEISSEfficientNavigation(ZEN)2011SP4(blackedition)software with medians, quartiles and an interquartile range (IQR) with whiskers
wasused. Allimages were displayedasthree dimensional z-stacksof13 representing1.5×IQR.DataanalyseswereperformedwithaMann–Whitney
slices with an interval of 1μm. High-resolution images of the actin Utest.ForCa2+analyses,themultiplet-testwasusedaswell,andthetime-
cytoskeletoninFig.5B(inset)wererecordedwiththeLSM410(CarlZeiss dependent data are presented as mean±s.e.m. P<0.05 was considered to
AG), equipped with a 63× oil immersion objective (1.25 oil/0.17). For indicatesignificantdifferences.
visualizationofthecorrespondingsubstrateregion,thereflectionmodewas
usedandthepillarswerefalsecoloredingreen. Acknowledgements
WethankDrMarcusFrank(ElectronMicroscopyCenter,UniversityMedicalCenter
Rostock,Germany)forhisassistancewithscanningelectronmicroscopy(FE-SEM
Correlativelightandelectronmicroscopy MerlinVP)andNorbertZichner(CenterforMicrotechnologiesZFM,Chemnitz
Forcorrelativemicroscopy,thefluorescenceimagesoftheactinstained,non- UniversityofTechnology,Germany)forproductionofthetitanium-coatedwafers.
embedded samples were acquired with the upright microscope
AxioScopeA1 (Carl Zeiss) equipped with a 63× water objective. The Competinginterests
calibration of the samples and image overlay was performed with the Theauthorsdeclarenocompetingorfinancialinterests.
software AxioVision v. 4.8.2 (tool: shuttle and find). Calibration of the
samples was accomplished by using the edges of the sample as point of Authorcontributions
reference. Afterwards, the samples were prepared for scanning electron Conceptualization:C.M.,H.R.,J.B.N.;Methodology:C.M.,S.S.,H.R.,B.F.,J.B.N.;
microscopyasdescribedbelow. Validation:C.M.,S.S.,H.R.,B.F.;Investigation:C.M.,S.S.,H.R.,B.F.;Datacuration:
C.M.,H.R.;Writing-originaldraft:C.M.,S.S.,H.R.,B.F.;Writing-review&editing:
J.B.N.;Visualization:C.M.,S.S.,H.R.,B.F.,J.B.N.;Supervision:J.B.N.;Project
Scanningelectronmicroscopy
administration:J.B.N.;Fundingacquisition:J.B.N.
Cellsweregrownfor30min,or24honthesamples,washedwithPBSthree
times and thenfixed with2.5%glutardialdehyde(cat.no. 814393, Merck Funding
KGaA, Darmstadt, Germany) for 1h at room temperature, dehydrated WeappreciatethefinancialsupportoftheGermanResearchFoundation(Deutsche
throughagradedseriesofethanol(30,50,75,90and100%for5,5,15, Forschungsgemeinschaft;DFG)graduateschoolwelisa(no.1505/2)forC.M.and
10minandtwicefor10min,respectively)driedinacriticalpointdryer(K S.S.aswellbytheUniversityMedicalCenterRostock.DepositedinPubMedCentral
850,EMITECH, Taunusstein, Germany) and thensamples were sputtered (PMC)forimmediaterelease.
with gold for 100s (layer of ∼20nm) (SCD 004, BAL-TEC, Wetzlar,
Germany). Scanning electron microscopy (SEM) observations were References
performed with a field emission (FE)-SEM (ZEISS Merlin VP compact, Ataullakhanov,F.I.andVitvitsky,V.M.(2002).Whatdeterminestheintracellular
ATPconcentration.Biosci.Rep.22,501-511.
CarlZeissAG).Therecoveringofthedesiredareasforcorrelativemicroscopy
Bacakova,L.,Filova,E.,Parizek,M.,Ruml,T.andSvorcik,V.(2011).Modulation
wasperformedwiththesoftwareAxioVisionv.4.8.2(tool:shuttleandfind)
ofcelladhesion,proliferationanddifferentiationonmaterialsdesignedforbody
aftersamplecalibration.
implants.Biotechnol.Adv.29,739-767.
Bellas,E.andChen,C.S.(2014).Forces,andstemcellfate.Curr.Opin.CellBiol.
Ca2+ionimaginginlivingosteoblasts 31,92-97.
Chopra,A.,Murray,M.E.,Byfield,F.J.,Mendez,M.G.,Halleluyan,R.,Restle,D.J.,
Fluo-3staining ofsuspendedcells wasperformed ontrypsinized MG-63
Aroush,D.R.-B.,Galie,P.A.,Pogoda,K.,Bucki,R.etal.(2014).Augmentationof
cellsthatwerewashedinPBS(containingCa2+andMg2+)andafterwards integrin-mediatedmechanotransductionbyhyaluronicacid.Biomaterials35,71-82. e
stainedwiththeCa2+ionindicatorfluo-3/acetoxymethylester(fluo-3/AM, Cohen,M.,Joester,D.,Geiger,B.andAddadi,L.(2004).Spatialandtemporal nc
5μM,LifeTechnologiesCorporation,Eugene,OR)inhypotonicHEPES sequenceofeventsincelladhesion:frommolecularrecognitiontofocaladhesion e
buffer(Staehlkeetal.,2015).Subsequently,theosteoblastswerecollected assembly.Chem.BioChem.5,1393-1399. ci
S
inisotonicHEPESbufferandseededontothemicropillars(withorwithout Czekanska,E.M.,Stoddart,M.J.,Richards,R.G.andHayes,J.S.(2012).In
l
PPAAm).After30mincultivation,thesampleswereturnedupsidedownin searchofanosteoblastcellmodelforinvitroresearch.Eur.CellsMater.24,1-17. el
aibidiμ-dish(ibiTreat,Ø35mm;ibidiGmbH,Martinsried,Germany)and Finke,B.,Luethen,F.,Schroeder,K.,Mueller,P.D.,Bergemann,C.,Frant,M., C
analyzedintheconfocalmicroscope. Ohl, A. and Nebe, B. J. (2007). The effect of positively charged plasma f
polymerization on initial osteoblastic focal adhesion on titanium surfaces. o
Fluo-3stainingofadherentcellswasperformedaftera24hcultivationof
MG-63 cells on micropillars (with or without PPAAm). The cells were Biomaterials28,4521-4534. al
Fritsche, A., Luethen, F., Lembke, U., Finke, B., Zietz, C., Rychly, J., n
markedwithfluo-3/AMasdescribedabove.Fortheanalysis,theosteoblasts Mittelmeier, W. and Bader, R. (2010). Measuring bone cell adhesion on ur
wereculturedinisotonicHEPESbuffer(Staehlkeetal.,2015)asdescribed implantsurfacesusingaspinningdiscdevice.Mat.-wiss.u.Werkstofftech.41, o
forsuspendedcells. 83-88. J
8
RESEARCHARTICLE JournalofCellScience(2018)131,jcs207001.doi:10.1242/jcs.207001
Gabler,C.,Zietz,C.,Göhler,R.,Fritsche,A.,Lindner,T.,Haenle,M.,Finke,B., Nebe, B. J. and Luethen, F. (2008). Integrin- and hyaluronan-mediated cell
Meichsner,J.,Lenz,S.,Frerich,B.etal.(2014).Evaluationofosseointegration adhesion on titanium. In Metallic Biomaterial Interfaces (ed. J. Breme, C.J.
oftitaniumalloyedimplantsmodifiedbyplasmapolymerization.Int.J.Mol.Sci.15, KirkpatrickandR.Thull),pp.179-182.Weinheim:Wiley-VCH;ISBN:978-3-527-
2454-2464. 31860-5.
Head,B.P.,Patel,H.H.andInsel,P.A.(2014).Interactionofmembrane/lipidrafts Nebe,J.G.B.,Luethen,F.,Lange,R.andBeck,U.(2007).Interfaceinteractionsof
with thecytoskeleton:Impactonsignalingandfunction:membrane/lipidrafts, osteoblastswithstructuredtitaniumandthecorrelationbetweenphysicochemical
mediatorsofcytoskeletalarrangementandcellsignaling.Biochim.Biophys.Acta- characteristicsandcellbiologicalparameters.Macromol.Biosc.7,567-578.
Biomembr.1838,532-545. Parton,R.G.anddelPozo,M.A.(2013).Caveolaeasplasmamembranesensors,
Janmey,P.A.andLindberg,U.(2004).Cytoskeletalregulation:richinlipids.Nat. protectorsandorganizers.Nat.Rev.Mol.CellBiol.14,98-112.
Rev.Mol.CellBiol.5,658-666. Pelkmans, L. and Helenius, A. (2002). Endocytosis via caveolae. Traffic 3,
Komarova,S.V.,Ataullakhanov,F.I.andGlobus,R.K.(2000).Bioenergeticsand 311-320.
mitochondrial transmembrane potential during differentiation of cultured Rebl,H.,Finke,B.,Lange,R.,Weltmann,K.-D.andNebe,J.B.(2012).Impactof
osteoblasts.Am.J.Physiol.CellPhysiol.279,C1220-C1229. plasmachemistryversustitaniumsurfacetopographyonosteoblastorientation.
Kunz,F.,Rebl,H.,Quade,A.,Matschegewski,C.,Finke,B.andNebe,J.B. ActaBiomater.8,3840-3851.
(2015).Osteoblastswithimpairedspreadingcapacitybenefitfromthepositive Rebl,H.,Finke,B.,Schmidt,J.,Mohamad,H.S.,Ihrke,R.,Helm,C.A.andNebe,
chargesofplasmapolymerisedallylamine.Eur.CellsMater.29,177-189. J.B.(2016).Acceleratedcell-surfaceinterlockingonplasmapolymer-modified
Liu,X.,Feng,Q.,Bachhuka,A.andVasilev,K.(2014).Surfacemodificationby porousceramics.Mater.Sci.Eng.C69,1116-1124.
allylamineplasmapolymerizationpromotesosteogenicdifferentiationofhuman Saldaña,L.andVilaboa,N.(2010).Effectsofmicrometrictitaniumparticleson
adipose-derivedstemcells.ACSAppl.Mater.Interfaces6,9733-9741. osteoblast attachment and cytoskeleton architecture. Acta Biomater. 6,
Lüthen,F.,Lange,R.,Becker,P.,Rychly,J.,Beck,U.andNebe,J.B.(2005).The 1649-1660.
influenceofsurfaceroughnessoftitaniumonβ1-andβ3-integrinadhesionandthe Staehlke,S., Koertge,A. andNebe, B. (2015). Intracellularcalciumdynamics
organization of fibronectin in human osteoblastic cells. Biomaterials 26, dependentondefinedmicrotopographicalfeaturesoftitanium.Biomaterials46,
2423-2440. 48-57.
Moerke, C., Mueller, P. and Nebe, B. (2016). Attempted caveolae-mediated Wennerberg, A. and Albrektsson, T. (2009). Effects of titanium surface
phagocytosisofsurface-fixedmicro-pillars.Biomaterials76,102-114. topographyonboneintegration:asystematicreview.Clin.OralImplantsRes.
Moerke,C.,Rebl,H.,Finke,B.,Dubs,M.,Nestler,P.,Airoudj,A.,Roucoules,V., 20,172-184.
Schnabelrauch, M., Koertge, A., Anselme, K., Helm, C.A. and Nebe, J.B. Zimmerman,E.,Geiger,B.andAddadi,L.(2002).Initialstagesofcell-matrix
(2017).Abrogatedcellcontactguidanceonamino-functionalizedmicro-grooves. adhesioncanbemediatedandmodulatedbycell-surfacehyaluronan.Biophys.J.
ACSAppl.Mater.Interfaces9,10461-10471. 82,1848-1857.
e
c
n
e
i
c
S
l
l
e
C
f
o
l
a
n
r
u
o
J
9