Haptotaxis is Cell Type Specific and Limited by Substrate Adhesiveness

14 Pages·2015·3.27 MB·English

by  Jessica H. Wen

UC San Diego
UC San Diego Previously Published Works
Title
Haptotaxis is Cell Type Specific and Limited by Substrate Adhesiveness
Permalink
https://escholarship.org/uc/item/4g84t7rh
Journal
Cellular and Molecular Bioengineering, 8(4)
ISSN
1865-5025
Authors
Wen, JH
Choi, O
Taylor-Weiner, H
et al.
Publication Date
2015-12-01
DOI
10.1007/s12195-015-0398-3

Peer reviewed
eScholarship.org Powered by the California Digital Library
University of California
CellularandMolecularBioengineering((cid:2)2015)
DOI:10.1007/s12195-015-0398-3
Haptotaxis is Cell Type Specific and Limited by Substrate Adhesiveness
JESSICAH.WEN,1ONKIUCHOI,2HERMESTAYLOR-WEINER,1ALEXANDERFUHRMANN,1JEROMEV.KARPIAK,3
ADAH ALMUTAIRI,3,4 and ADAM J. ENGLER1,5
1Departmentof Bioengineering,UC SanDiego, LaJolla, CA92093, USA; 2Department ofBiomedical Engineering, National
Yang-MingUniversity,BeitouDistrict,TaipeiCity112,Taiwan;3DepartmentofBiomedicalSciences,UCSanDiego,LaJolla,
CA92093,USA;4SkaggsSchoolofPharmacyandPharmaceuticalSciences,UniversityofCalifornia,SanDiego,9500GilmanDr,
LaJolla,CA92093,USA;and5SanfordConsortiumforRegenerativeMedicine,LaJolla,CA92037,USA
(Received23April2015;accepted22May2015)
AssociateEditorMichaelR.Kingoversawthereviewofthisarticle.
Abstract—Motilecellsnavigatethroughtissuebyrelyingon these cues are often found together in vivo via normal
tactile cues from gradients provided by extracellular matrix tissue variation or pathological conditions, e.g., the
such as ligand density or stiffness. Mesenchymal stem cells
post-infarct myocardial scar that is stiffer than,
(MSCs) and fibroblasts encounter adhesive or ‘haptotactic’
and compositionally different from, healthy cardiac
gradientsattheinterfacebetweenhealthyandfibrotictissue
as they migrate towards an injury site. Mimicking this muscle.4,36,38 These gradients are thought to aid in
phenomenon, we developed tunable RGD and collagen directional migration through tissue; for example,
gradients in polyacrylamide hydrogels of physiologically mesenchymal stem cells (MSCs) whose fate can be
relevant stiffness using density gradient multilayer polymer- regulated by ECM stiffness14,42,45 also durotax or mi-
ization to better understand how such ligand gradients
grate in the direction of an increasing stiffness gradi-
regulatemigratorybehaviors.Independentofligandcompo-
sition and fiber deformation, haptotaxis was observed in ent.43 Dermal fibroblast migration into cutaneous
mouse 3T3 fibroblasts. Human MSCs however, haptotaxed wound spaces is also well known and critical for
onlywhencell-substrateadhesionwasindirectlyreducedvia assemblyofnewECM.27,29However, spatialvariation
addition of free soluble matrix ligand mimetic peptides.
in tissue stiffness in vivo is often accompanied by var-
Under basal conditions, MSCs were more contractile than
iationinadhesiveliganddensity.Duringfibrosis,ECM
fibroblasts. However, the presence of soluble adhesive
peptides reduced MSC-induced substrate deformations; in- accumulates and may present several ligand gradients
creasedcontractilitymaycontributetolimitedmigration,but at the interface of healthy and pathological tissue.18
modulatingcytoskeletalassemblywasineffectiveatpromot- BecauseinvivospatialdistributionsofECMproteins18
ing MSC haptotaxis. When introduced to gradients of and stiffness4 have been shownto change with fibrosis
increased absolute ligand concentrations, 3T3s displayed
and animal models have documented directed migra-
increasedcontractilityandnolongerhaptotaxed.Thesedata
suggest that haptotactic behaviors are limited by adhesion tion,27,29 it is of interest to develop systems to inves-
andthatalthoughbothcelltypesmayhometotissuetoaid tigate cell migratory behaviors in response to spatial
in repair, fibroblasts may be more responsive to ligand gradients in vitro. Although we and others have
gradients thanMSCs. developedsimplifiedsystemstoprobedurotaxis,30,31,43,46
few systems with the ability to separately modulate
Keywords—Migration, Mesenchymal stem cell, Fibroblast, directedmigrationduetoadhesiveligandgradients,6i.e.,
Ligandgradient, Surface modification, Elasticity.
haptotaxis,andmigrationduetosubstratestiffnesshave
been developed.
Adhesive ligand gradients have been shown to in-
duce directed cell migration in human dermal fibrob-
INTRODUCTION lasts in vitro.10 However, cell type-specific differences
in the degree of haptotaxis could be simultaneously
Anchorage dependent cells sense and respond to
influenced by physiologically-relevant substrate stiff-
properties of the extracellular matrix (ECM), e.g.,
ness, which is a major regulator of stem cell fate.14,45
stiffness30,43,46 and ligand density.13,17,34 Gradients of
Moreover, haptotactic mechanisms remain unclear;
this is in part due to the fact that systems commonly
Address correspondence to Adam J. Engler, Department of used to study haptotaxis are often limited in
Bioengineering,UCSanDiego,LaJolla,CA92093,USA.Electronic
gradient resolution, and typically utilize substrates of
mail:[email protected]
(cid:2)2015BiomedicalEngineeringSociety
WENet al.
non-physiologically relevant stiffness. Furthermore, 3 min, washed with ethanol, and dried. All chemicals
haptotaxis studies often lack direct comparisons wereobtainedfromSigma-Aldrichandotherproducts
between cell and adhesive ligand types. For example, from ThermoFisher unless otherwise indicated.
human melanoma cells undergo haptotaxis when
exposed to gradients of collagen type IV, fibronectin,
DGMP Hydrogel Fabrication
and laminin using Boyden chambers; these chambers
however, can create only step gradients.2 Myoblast DGMP involves layering pre-polymer hydrogel
haptotaxis has been observed with more gradual nan- solutionsintorubbermoldscuttofitbetweentwoglass
odotadhesivegradients;however,thesegradientswere slides, followed by UV polymerization. 5 9 5 mm
printed on rigid glass.35 Literature suggests that cell molds cut from super-soft 35A silicone rubber sheets
forces exerted on adhesive matrix proteins contribute (McMaster-Carr)wereclampedbetweenaglassslide,a
to cellular mechanical signaling41,45 and that in vivo functionalized glass coverslip to facilitate gel attach-
gradients are gradual.4,42 Reductionist systems ment, and a dimethylchlorosilane- (DCDMS-) treated
employing compliant hydrogels deformable by cells, glass slide to facilitate gel detachment. Pre-polymer
e.g., polyacrylamide20,22 and polyethylene glycol,10 solutions composed of acrylamide, bis-acrylamide,
with gradual haptotactic gradients are thus needed to acrylated poly(ethylene) glycol conjugated to RGD
determinewhichsignalingnetworkswithinavarietyof (aPEG-RGD) or acrylic acid, 2,2¢-Azobis(2-methyl-
celltypesregulatehaptotaxis.Althoughnotyetstudied propion0amidine) dihydrochloride photoinitiator
in detail, mechanisms regulating haptotaxis likely (Sigma), and serial dilutions of iodixanol (Sigma), an
involve adhesive and/or cytoskeletal proteins given inert density modifier, were layered into the silicone
their roles in other migratory control mechanisms in molds in order of decreasing density. Hydrogels with
processes such as chemotaxis,1 and the establishment gradients of adhesive ligands were fabricated by
of cell polarity and shape.21,44 allowing two 12 lL layers of pre-polymer solution of
Gradients of protein attachment to compliant distinct composition (Fig. 1a) to diffuse and settle for
hydrogelshavebeenachievedviamicrocontactprotein 10 s creating a gradient at the mixing interface
printing, self-assembled monolayers, and microfluidic (Fig. 1b). Layered hydrogels were subsequently poly-
mixing devices.10,26,33 Despite the use of compliant merized via exposure to 350 nm UV light for 4.5 min.
hydrogels, these methods are often difficult to scale 10/0.15% acrylamide/bis-acrylamide solutions
down, tune, and require significant quantities of pro- composed of0.1or10 mMaPEG-RGD with1or 4%
teinorpeptide,muchofwhichdoesnotbind.Here,we iodixanol respectively, were used to obtain PA-PEG-
used density gradient multilayer polymerization RGDhydrogelswithadhesiveRGDSligandgradients
(DGMP), to construct compliant culture substrates of approximately 10 mM/mm that span an 800 lm
with spatial gradients of composition.22,23 DGMP, a region across the center of the hydrogel (Fig. 1c). A
simple method that does not require specialized hydroxycoumarin dye (excitation 350 nm) pre-conju-
equipment, ingredients, or expertise, relies on density- ated to aPEG-RGD allowed us to visualize serial
mediated phase separation to construct layers of dis- dilutions of adhesive peptide; differences in cell
tinct composition. By using this method to fabricate spreading and fluorescent intensity confirm that sub-
gradients of adhesive peptide RGD or collagen type I strate adhesiveness scales with aPEG-RGD concentra-
on a polyacrylamide (PA) hydrogel platform, we tion(Fig. 1d).aPEG-RGD,oracryl-PEG3400-GRGD-
demonstrate that while fibroblasts undergo haptotaxis amide and acryl-PEG3400-[K-7-hydroxycoumarin-3-
in response to both gradient types under basal condi- OH]GGGRGD-amide were custom ordered (21st
tions, MSC haptotaxis appears to be limited by cell- Century Biochemicals).
substrate adhesiveness, and is observed only when cell To investigate the migration behavior of cells in
affinity for the substrate is reduced. response to solely a ligand gradient on a substrate of
physiologically relevant stiffness, AFM was used to
characterize the hydrogel stiffness in order to ensure
MATERIALS AND METHODS thatnosimultaneousstiffnessgradientswerecreatedas
a consequence of the haptotactic gradient fabrication
Functionalized Coverslips
process (Fig. 1e). Iodixanol and aPEG-RGD did not
To covalently attach hydrogel substrates to glass, significantly alter hydrogel stiffness at the concentra-
25 mm square glass coverslips were cleaned and oxi- tions used in this study, as stiffness was maintained at
dized by exposing both sides to UV/ozone (BioForce 10 kPa across the substrate.
Nanoscience) for 3 min. Samples were subsequently To obtain collagen gradients, 10/0.1% acrylamide/
functionalized with 20 mM 3-(trimethoxysilyl)propyl bis-acrylamide, acrylic acid, and serial dilutions of
methacrylate in 10% acetic acid (AA) in ethanol for iodixanol were used. Solutions of 0.005 and 0.1%
Haptotaxis is Cell TypeSpecific and Limited bySubstrateAdhesiveness
FIGURE1. FabricationofpolyacrylamidehydrogelswithRGDandcollagengradients.(a)Pre-polymersolutionsofacrylamide,
bis-acrylamide,water,photoinitiator,acrylicacidoracrylated-PEG-RGD,aredissolvedwithserialdilutionsofiodixanol.(b)Two
pre-polymersolutionsarelayeredinorderofdecreasingdensityintoasiliconerubbermoldthatisclampedwithamethacrylated
glass coverslip in between an untreated glass slide, and a DCDMS-treated glass slide. After a settling time of 10s, layered
hydrogels are polymerized with UV light for 4.5min. The iodixanol is rinsed away subsequent to polymerization. (c) A
hydroxycoumarindye(excitation350nm)pre-conjugatedtoaPEG-RGDallowsforgradientcharacterization;incorporating0.1and
10mMaPEG-RGDintotwolayersyieldagradientofapproximately10mM/mm.(d)Imagesofstaticpolyacrylamidegelswith0.1,
0.5, and 2.5mM aPEG-RGD-dye under UV light (left); brightfield images indicate that MSC spreading correlates with RGD con-
centration(scalebar,100lm).(e)Elasticmodulusmeasuredevery500lmviaAFMacrosstheRGDgradienthydrogelsubstrate
(n>3;mean6SD).(f)Elasticmodulusmeasuredevery500lmviaAFMacrosstheacrylicacid/collagengradienthydrogelsub-
strate(n>3;mean6SD).(g)Antibody-proteinruptureeventsdetectedviaAFS(left)andcorrespondingmeasuredrupturelengths
(right) on low (0.005%) and high (0.1%) acrylic acid hydrogels activated with collagen I. (n>500; mean6SEM; *p<0.05,
***p<0.005).
acrylic acid, with 1 and 4% iodixanol respectively, then subsequently functionalized with type I collagen
wereusedtofabricatePA-AAhydrogelswithgradients using EDC/NHS crosslinking to create a gradient of
ofacrylicacidthatspanasimilar800 lmregionacross collagen on the substrate surface. Hydrogels were
the hydrogel at the mixing interface. Hydrogels were incubatedin5 mg/mLEDC(AlfaAesar)and9 mg/mL
WENet al.
NHS (Alfa Aesar) dissolved in water for 10 min at until use. Force spectrograms were taken in 10 9 10
room temperature, rinsed in water, then incubated in arrays, with indentations spaced 10 lm apart. The
100 lg/mL rat tail collagen type I (BD Biosciences) in functionalized tips were indented and retracted at
DI water for 30 min. 2 lm/s. Force curves were converted to force vs. tip
AFMindentationindicatesthathydrogelstiffnessof Z-position and analyzed for rupture events using a
5 kParemainedconstantacrossthesubstrate(Fig. 1f). previouslydescribedalgorithm.16Ruptureeventcount
Force spectrograms from microindentations on both along with corresponding event lengths and forces
low and high collagen density sides of the substrate were recorded.
were obtained via atomic force spectroscopy (AFS)
usingaprobefunctionalizedwithananti-collagentype
Cell Culture
I antibody as described below. When the tip makes
contact with the substrate, protein-antibody bonds Human MSCs (Lonza) were cultured between pas-
form.Asthetipisretractedfromthesurface,segments sages 4–7 in low glucose Dulbecco’s modified eagle
of collagen fibers unfold and stretch until antibody- medium (DMEM) (Gibco) supplemented with 5%
collagen bonds rupture. These rupture events can be fetal bovine serum (Gemini Bio-Products) and 1%
visualized within each spectrogram. A greater number antibiotic/antimycotic (Corning). NIH-3T3s were cul-
of events (Fig. 1g, left) detected on the high collagen tured in DMEM supplemented with 5% fetal bovine
density side of the substrate indicate that collagen serum (Gemini Bio-Products), 4 mM L-glutamine,
density scales with acrylic acid concentration, and 1 mM sodium pyruvate, and 1% penicillin-strepto-
shorter rupture lengths on this side of the substrate myocin (Corning). For migration experiments, MSCs
indicate a greater degree of collagen fiber-EDC/NHS and3T3swereseededonhydrogelsatdensitiesof1000
tethering to the surface (Fig. 1g, right). All hydrogels and 5000 cells/cm2, respectively. Cells were allowed to
were soaked in sterile PBS, stored at 4 (cid:3)C, and UV attach and spread for 3 h prior to migration tracking.
sterilized for 10 min prior to use in cell culture.
Migration Studies
Atomic Force Microscopy
Migration experiments were conducted using
Substrate stiffness was measured via indentation Brightfield timelapse video microscopy over a period
using an MFP-3D-Bio (Asylum Research) atomic of 48 h. Cells were placed in a temperature, CO , and
2
force microscope (AFM). Chromium/gold-coated, humidity controlled LiveCell chamber (Pathology
silicone nitride (SiN) cantilevers with pyramid tips Devices). Images were taken every 15 min, and cell
(PNP-TR; NanoWorld) with ~30 pN/nm spring con- positions were tracked in FIJI. Absolute migration
stants as determined from the MFP-3Ds built-in velocities, velocities in the gradient direction, tactic
calibrationfunctionwereused.Samplesweremounted indices,andanglehistogramswerecalculatedfromraw
on glass slides using vacuum grease and immersed in position values using custom MATLAB scripts. To
PBS. The probe was indented into the sample with minimize error in tracking, a running average was
approach and retraction velocities of 2 and 20 lm/s taken of every four velocities. Tactic index, or persis-
and a force trigger of 2 nN. AFM data was analyzed tence, iscalculatedby taking theratioofdisplacement
using custom code in Igor Pro (Wavemetrics); the (frominitialtofinalposition)tototaldistancetraveled.
substrate spring constant, i.e., Young’s Moduli, was Where indicated, cells were chosen from three regions
determined using a linearized Sneddon model.25 of the gradient substrates: the low adhesive ligand
To detect fibrous protein on the substrate surfaces, density region, the gradient region, and the high
cantilevers were functionalized with the C2456, a adhesive ligand density region.
mouse monoclonoal anti-collagen type I antibody
(Sigma) using a previously established method.5,8
Inhibitors and Drugs and RGD Blocking
Briefly, cantilevers were cleaned with chloroform for
30 s and immersed in 5 M ethanolamine-HCl in Lysophosphatidic acid (Enzo Life Sciences #BML-
dimethylsulfoxide (DMSO) at room temperature for LP100), bradykinin (Sigma #B3259), nocodazole
3 h, resulting in the functionalization of an amine (Sigma#M1404),cytochalasinD(Sigma#C2618),and
group to the surface of the probe. Tips were rinsed in Rac inhibitor (Millipore #553502) were included in
PBS, incubated in 25 mM bis[sulfosuccinimidyl] culturemediaatconcentrationsof20 lM,10 nM,0.5,
suberate crosslinker (BS3; Pierce) for 30 min, rinsed 0.5, and 50 lM respectively. Drugs were stored at
again in PBS, then immersed in 200 lg/mL C2456 for 1009 in DMSO at (cid:2)20 (cid:3)C until use. To reduce MSC
30 min to crosslink the antibody to the tip. Function- adhesiontothehydrogelsubstratesurfacebyblocking
alizedcantileverswererinsed,dried,andstoredat4 (cid:3)C RGD peptide-binding integrins, 0.02, 0.05, 0.1 and
Haptotaxis is Cell TypeSpecific and Limited bySubstrateAdhesiveness
0.2 mM free RGD peptide was included in the culture stained cells were manually selected for analysis. The
media throughout the duration of the timelapse. FRETratioforeachpixelwithinaselectedregionwas
calculated by dividing the pixel intensity in the accep-
tor image by the corresponding pixel intensity in the
Immunofluorescence
donor image. FRET ratios less than 0.05 and greater
Cellswerefixedin3.7%formaldehydefor15 minat than1.0wereexcludedfromanalysis.ThemeanFRET
room temperature, rinsed in PBS, and permeabilized ratio within the selected regions was calculated for
with 1% triton-X in PBS for 15 min at room tem- each cell and then averaged over all the cells in each
perature. Cells were incubated in 1:400 rhodamine condition.
phalloidin (Life Technologies) in 2% bovine serum
albumin (BSA) in PBS supplemented with MgCl for
2 Traction Force Microscopy
30 min, then incubated in 1:2000 Hoescht in DI water
for 10 min. Samples were washed with DI water and Tractionforcemicroscopy(TFM)wasperformedas
mounted with fluoromount-G (SouthernBiotech). previously described.9 Briefly, fluorescent 0.2 lm mi-
Imaging was performed using a Nikon Eclipse TI mi- crospheres(Invitrogen)wereaddedtothepre-polymer
croscope equipped with a CARV II confocal system solution to a concentration of 1% v/v. 0.1% v/v
(BD Biosciences), motorized stage and MS-2000 con- TWEEN was added to the solution to prevent micro-
troller (Applied Scientific Instrumentation), and a spheres from forming aggregates. Hydrogel substrates
Cool-Snap HQ camera (Photometrics) and controlled used in TFM experiments were cast in 35 mm glass
by Metamorph (Molecular Devices). bottom dishes (MatTek) and layered with an 18 mm
DCDMS-treated circular coverslip, then UV poly-
merizedfor4.5 minat350 nmasdescribedabove.The
Fo¨rster Resonance Energy Transfer (FRET)
microspheres directly underneath selected live cells
Fibronectin was isolated from human plasma using were imaged with a 609 confocal objective (Zeiss)
gelatin-Sepharose binding and eluted with 6 M urea. using the Nikon Eclipse TI microscope. Cells were
Isolated fibronectin was concentrated to approximately released with 2.5% trypsin and the same microspheres
3 mg/mL using an Amicon Ultra Centrifugal Filter were imaged again. Embedded microsphere displace-
(10 kDaNMWL)(Millipore)anddenaturedfor15 min ment from prior to subsequent to trypsin treatment
in 4 M guanidine hydrochloride (GdnHCl). Denatured was determined using a particle image velocimetry
fibronectin was dual-labeled with Alexa Fluor 488 script in Matlab (MathWorks). Mean displacements
(donor) and Alexa Fluor 546 (acceptor) fluorophores, were calculated for each cell, and averaged over all
as previous described.3,40 Briefly, denatured fibronectin cells within one condition.
was incubated with a 30-fold molar excess of Alexa
Fluor 546 C5 Maleimide (Life Technologies) for 2 h to
Statistical Analyses
labelcysteineresidueswithintheIII7andIII15domains
offibronectin.Thesingle-labeledfibronectinwasbuffer All data are expressed as mean ± standard error of
exchanged into 0.1 M sodium bicarbonate pH 8.3 and the mean. Student’s t tests were used to determine
separatedfromunreactedacceptorfluorophoresusinga statistical significance in Figs. 1g, 2, 3, 5c, and
spin desalting column (Thermo Scientific). The single- Supplementary Figs. 3 and 5. 1- or 2-way ANOVAs
labeled fibronectin was then incubated with a 40-fold wereusedinFigs. 3,4,and5basindicated.Significant
molarexcessofAlexaFluor488succinimidylester(Life differences were considered for p<0.05 as indicated;
Technologies) for 1 h to label amine residues through- n.s. denotes not significant. Multiple comparisons
out fibronectin. Unreacted donor fluorophores were Tukey’s post hoc tests were performed where
removedusingaspindesaltingcolumnanddual-labeled appropriate.
fibronectin was stored with 10% glycerol at (cid:2)20 (cid:3)C.
The average number of acceptor and donor fluor-
ophores per fibronectin dimer was 3.6 and 8.8, respec- RESULTS
tively, and was determined using published extinction
Fibroblasts Exhibit Haptotaxis
coefficients and absorbances of the dual-labeled
fibronectin at 280, 498, and 556 nm. Murine 3T3 fibroblasts and human MSCs attached
Imagesofthedual-labeledfibronectinwereacquired to and spread on RGD and collagen ligand gradients
on a Zeiss LSM 780 Confocal Microscope and ana- within 4 h of seeding. Quantifying cell density and
lyzed using a custom MATLAB script, as previously spread area and across collagen gradient substrates
described.37Briefly,imageswereaveragedwitha3 9 3 (Fig. S1) 15 h post-seeding demonstrated that density
pixel-sliding block and perinuclear regions of DAPI and spread area both scale with surface collagen den-
WENet al.
FIGURE2. Migratoryresponseofstemcellsandfibroblastsonadhesiveligandgradients.AbsolutevelocitiesandXvelocitiesof
MSCs(left) and3T3s(right) on indicatedregions of hydrogels with (a). RGD concentration gradients and (b). Surfacecollagen
densitygradients.Lowandhighindicatescellsfoundonthelowandhighstaticconcentrationregionsofthesubstrate;gradient
indicatescellsfoundatthemixinginterfaceofthesubstrate.PositiveXvelocityindicatesmigrationtowardsregionsofincreasing
adhesiveliganddensity.Representativeroseplotsshowmigratorypathsofcellsbeginningattheorigin.(n>20;mean6SEM;
*p<0.05;n.s.,notsignificant).
sity, and further confirmed the presence of a surface absolute fibroblast migration velocity inversely scaled
adhesive gradient. Migration velocity and haptotactic withRGDconcentration,i.e.,cellsmovedfasteratlow
velocity, i.e., migration in the gradient direction, were RGD concentration (Fig. 2a, right). To determine
measured. MSCs did not display any preferential whether migration was affected by the presence of a
migration within 48 h of spreading on gradients of full-length, fibrous protein compared to the short
RGD peptide. Absolute MSC migration velocity was adhesiveRGD peptide,MSC andfibroblastmigration
also insensitive to ligand density (Fig. 2a, left). How- was assessed on collagen density gradients. Consistent
ever fibroblasts exhibited haptotactic migration in the with RGD, MSCs on collagen substrates displayed
direction of higher RGD concentration. Moreover, randommigrationthatlackedpreferenceforeitherlow
Haptotaxis is Cell TypeSpecific and Limited bySubstrateAdhesiveness
FIGURE3. Influence of drug treatment on fibroblast morphology and migration. (a) 3T3 fibroblasts spread on PA-PEG-RGD
hydrogels24hafterseedingandtreatingbytheindicateddrugs(actin,red;nucleistaining,blue).Filledwhitearrowsindicatecells
thataresmallerandlessspreadthanuntreatedcontrolcells.Openwhitearrowsindicatecellsthatarelargerandmorespreadthan
untreatedcontrolcells.(scalebar,100lm).(b)Spreadareaand(c)absolutemigratoryvelocitiesandXvelocitiesofcellstreated
withtheindicateddrugsonlow,gradient,andhighregionsofRGDgradienthydrogelsubstrates.(n>20;mean6SEM;*p<0.05;
n.s.,notsignificant).(d)AbsolutevelocitiesandXvelocitiesof3T3streatedwithLPAonindicatedregionsofacrylicacidhydrogels
activated with surface gradients of collagen type I. (n>20; mean6SEM; n.s., not significant). (e) Absolute velocities and X
velocities of 3T3s treated with NSC23766 Rac inhibitor on gradient regions of acrylic acid hydrogels activated with surface
gradientsofcollagentypeI.PositiveXvelocityindicatesmigrationtowardsregionsofincreasingadhesiveliganddensity.(n>20;
mean6SEM;scalebar,100lm).
orhighliganddensity(Fig. 2b,left)andwasnotproces- (CytoD) to prevent actin polymerization and nocoda-
sive(Fig.S2)asmeasuredbyatacticratioofdisplacement zole (Noco) to prevent microtubule polymerization.
tototaldistancetravelledoveraperiodof48 h.Fibrob- Cells treated with nocodazole and cytochalasin D
lasts,on the otherhand, displayed biasednetmigration exhibited less-spread and rounded morphology com-
towardshighercollagendensity(Fig. 2b,right)thatwas pared to untreated cells (Fig. 3a), while cells treated
moreprocessivecomparedtocellsfoundonstaticregions withLPAandbradykininweremorespreadcompared
of the substrate (Fig. S3). Taken together, these data to untreated cells, especially at low RGDS concentra-
indicate that fibroblasts undergo haptotaxis on both tion (Fig. 3b). While LPA and bradykinin vs. noco-
ligand gradients of RGD and collagen under the basal dazole and cytochalasin D treatment increased and
conditionsdescribedinthisstudy, whileMSCsdonot. decreased absolute migration velocity respectively, all
drug treatments inhibited haptotaxis (Fig. 3c and
S4A). LPA was similarly effective at reducing hapto-
Fibroblast Haptotaxis is Regulated
taxis on collagen gradients (Fig. 3d and S4B). These
by a Cytoskeletal-mediated Mechanism
data suggest that while depolymerizing the cytoskele-
To provide insight into how the cytoskeleton ton prohibits both spreading and migration, promot-
potentially regulates migration and haptotaxis, we ing its assembly and contraction produces overly
treated fibroblasts with contractile agonists lysophos- spread, faster moving cells that lose their sensing
phatidic acid (LPA) and bradykinin (BK) as well as capabilities. Rac1, a GTP-binding protein that reg-
with cytoskeleton-perturbing drugs cytochalasin D ulates membrane ruffling in migrating cells19 and
WENet al.
FIGURE4. Measuring cell-induced deformations of the underlying substrate. Mean FRET intensity ratio and ratio maps of fi-
bronectinunderneath.(a)Spread3T3s(left)and(b)MSCs(right)onlowandhighacrylicacidPAhydrogelsactivatedwithEDC/
NHSandcoatedwithdual-labeledfibronectin.(n>16;mean6SEM;**p<0.01;***p<0.005).(c)Displacementmapsofembedded
fluorescentparticles(left)and(d)Meanparticledisplacement(right)resultingfrom3T3andMSCtractionsexertedonlowandhigh
acrylicacidPAhydrogelsactivatedwithEDC/NHSandcoatedwithcollagentypeI.(n=25;mean6SEM;***p<0.005).
facilitates labile adhesion,12 may be involved in the Although fibronectin FRET ratios underneath both
sensing capabilities that cause biased migration that cell types scaled with fibronectin surface density,
which LPA and bradykinin treatment reduced. While absolute FRET ratio values were similar between the
Rac-inhibitedfibroblastsstilladheredtoandspreadon two cell types on both low and high ligand density
collagengradienthydrogels,cellswerenolongercapable substrates (Figs. 4a and 4b). These observations sug-
ofhaptotaxis(Fig. 3e).Thesedataoverallsuggestthat gest that although a greater degree of attachment
appropriateprotrusivecapabilities,i.e.,membraneruf- effectively ‘‘ties’’ down fibronectin to the substrate
fling, followed by modest cytoskeletal assembly and more tightly and hinders fiber unfolding to a greater
contractionsappeartobecrucialforhaptotaxis. extent at higher ligand concentrations, protein tether-
Biased migration resulting from cell protrusions ing is perhaps unrelated to haptotaxis due to differ-
produces substrate deformations that are small and ences in migratory behaviors between MSCs and
transient43 and may be reflected in deformation of fibroblasts. Conversely, although both fibroblasts and
matrix proteins and/or the substrate.41,45 Fo¨rster MSCs exhibited adhesive ligand density-dependent
resonance energy transfer (FRET) and TFM were differences in mean substrate displacements observed
utilized to investigate the effect of adhesive ligand by TFM (Fig. 4c), MSCs displaced the substrate five-
density on cell-induced deformations of the ligand foldmorethanfibroblasts,regardlessofliganddensity
coating and substrate. The FRET ratio decreased and degree of tethering (Fig. 4d). Together these data
underneath both stem cells and fibroblasts, suggesting along with the observation that fibroblasts but not
that the cells exerted traction forces on the ligand MSCshaptotaxedsuggestthatalthoughtheabilityofa
coating thus deforming protein fibers, effectively cell to exert traction forces may be necessary for
decreasing the FRET ratio. Consequentially, the pro- motilityandmigration,MSCstendtosurpasstraction
tein underneath blebbistatin-treated cells that lack or adhesiveness thresholds which results in extremely
the ability to exert forces due to perturbed myosin large cell-induced deformations of the substrate. This
contractilityexhibitedincreasedFRETratios(Fig.S5). perhaps occurs to an extent where cells no longer can
Haptotaxis is Cell TypeSpecific and Limited bySubstrateAdhesiveness
FIGURE5. Reducing MSC-substrate adhesion by blocking surface integrins with free peptide. (a) MSCs spread on 10mM
PA-PEG-RGDhydrogelswithindicatedconcentrationsoffreeRGDpeptideintheculturemedia.(scalebar,100lm).(b)Absolute
andXvelocitiesofMSCsonthegradientregionofRGDgradientPAhydrogelswithindicatedconcentrationsoffreeRGDpeptidein
the culture media. (n>40; mean6SEM; *p<0.05). (c) Mean displacement of particles embedded in PA hydrogels with 10mM
aPEG-RGDresultingfromtractionsofMSCsinnormalmedia,andmediasupplementedwith0.02mMfreeRGDpeptide.(n>10;
mean6SEM;***p<0.005).
sense a directional cue such as increasing adhesive li- behaviors may be achieved by reducing cell-substrate
gand density. binding by blocking a subset of relevant integrins, if
too many integrins are blocked, cells potentially lose
their ability to sufficiently sense ligand density gradi-
Modulating Cell-Substrate Adhesion Affects
ents. This further suggests a biphasic phenomenon
Haptotactic Behavior in Both Stem Cells
where cells lose the ability to haptotax if they are too
and Fibroblasts
strongly or tooweakly adheredtothe surface, and the
Todeterminewhetherornotexceedingatractionor existence of a mid-range of adhesive ligand density
adhesiveness threshold can explain the absence of where haptotaxis is most profound.
haptotaxis in MSCs, we reduced MSCs tractions by To further support this theory, 3T3 migration was
addingfreeRGDpeptidetotheculturemediatoblock assessedonRGDgradientsofsimilarsteepness,butof
RGD binding sites on integrins to effectively reduce higher absolute peptide concentrations. Low and high
cell-substrate adhesiveness. MSCs had slightly smaller RGD concentrations were increased from 0.1 and
but still spread morphology (Fig. 5a). Including RGD 10 mM, to 5.6 and 15.8 mM respectively (Fig. 6a).
in the culture media increased absolute migration Total migration velocity and haptotactic velocity were
velocities (Fig. 5b), consistent with the observation of measured. In this system, 3T3s did not display pref-
higher migration velocities on substrates with low erential migration towards either end of the gradient
liganddensities(Fig. 2).Additionally,MSCsdisplayed substrate (Fig. 6b). Furthermore, absolute fibroblast
haptotaxis in an RGD dose-dependent manner, where migration velocity no longer scaled with RGD con-
20 lm RGD peptide resulted in the most significant centration; cells moved at similar speeds regardless of
haptotacticresponse.Thepresenceoffreepeptidealso their position relative to the gradient (Fig. 6b). Mean
decreased mean substrate displacement compared to substrate displacements induced by fibroblasts seeded
untreated MSCs cultured on high concentration RGD onthesemore-adhesivesubstratesincreasedbyseveral-
substrates (Fig. 5c). This sevenfold reduction yields fold compared to fibroblasts that haptotaxed on gra-
deformations within range of fibroblast-induced dients of lower absolute ligand concentrations; these
hydrogel deformations at surface ligand densities displacements are on par with non-haptotaxing MSCs
where directed migration is observed (Fig. 2). These (Fig. 4d). These data provide additional evidence
data also imply that while haptotactic migratory suggesting that haptotactic behaviors are limited by