Embed Size (px)
Citation preview
Surface analysis of human plasma fibronectin adsorbed tocommercially pure ti tani urn materials
D. E. MacDonald;.2.4 B. Markovic,1 M. Allen, 3 P. Somasundaran, 1 A. L. Boskey.
JLangmuir Center for Colloids & Interfaces, Columbia University, 911 S.W. Mudd Building, Mail Code 4711, 500West UOth Street, New York, New York 10027lV.A. Medical Center, Bronx, New York30igitaJ Instruments, Santa Barbara, California4Hospital for Special Surgery, New York, New York
Received 2 June 1997; accepted 5 November 1997
Abstract: Protein binding on metallic implant surfaces,such as titanium, is governed by the physico-chemical na-ture of the metallic surface. Human plasma fibronectin(HPF) is an important matrix glycoprotein that mediates celland protein attachment to each other or to the extracellularmatrix present during wound healing. The objective of thisstudy was to investigate the adsorption of HPF onto pol-ished commercially pure titanium (cpTi) by using atomicforce microscopy (AFM) and electron spectroscopy forchemical analysis (ESCA) and to measure the resultant sur-face contact angle before and after HPF binding. Two typesof cpTi disks, one highly polished in our laboratory (HSS)and one commercially prepared (31), were reacted with HPFsolutions of varying concentrations (1jJ.g/mL-10 ng/mL).ESCA survey spectra of samples coated with 1 jJ.g/mL offibronectin showed an increase in organic nitrogen and car-bon compared with uncoated controls. Contact angle mea-
surements of HSS and 31 cpTi disks showed no significantdifference in average contact angle (36.30 .i: 3.5 and 39.10 .i:3.1) despite differences in local root mean square (RMS) sur-face roughness (4.45 :t. 0.46 nm and 22.37 :!: 4.17 nm) asmeasured by AFM. Images obtained by AFM showed that 31specimens were more irregular, with large parallel polishinggrooves. Adsorbed HPF appeared in a globular form with anaverage length of 16.5 :!: 1.0 nm, a height of 25 .i: 0.5 nm, anda width of 9.6:to 1.2 nm. Fibronectin coating on both HSS and31 cpTi specimens resulted in a significant increase in hy-drophobicity compared to uncoated specimens. These re-sults indicate the significance of HPF on cpTi and may ex-plain how cpTi implants function in situ. @1998John Wiley& Sons, Inc. J Biomed Mater Res, 41, 120-130, 1998.
Key words: implant surface; titanium; fibronectin; atomicforce microscopy; contact angle
--
INTRODUCTION
Upon implantation, the surfaces of synthetic mate-rials invariably become coated with a thin proteina-cious film.1 Protein orientation will be dependent onthe particular binding surface, and various surfacecharacteristics may modi
121SURFACE ANALYSIS OF HPF
bronectin then were evaluated by means of ESCA, contactangle, and atomic force microscopy (AFM).
Fib ro nectin
Affinity-purified human plasma fibronectin was obtainedfrom Calbiochem (LaJolla, CA). A buffer solution (fBS) con-taining 0.05M Iris and 0.15M NaCI, pH 7.4, was preparedand sterilized by filtration through a 0.22 micron FalconEasy Flow TM filter (Becton Dickinson, Lincoln Park, NJ) un-der vacuum. Human plasma fibronectin solutions of 1 IJ.g/mL, 100 ng/mL, 10 ng/mL, and 1 ng/mL were prepared inthis TBS buffer. The cpIi disks then were placed in 1 mL ofeach of these solutions in 13 mL polypropylene tubes(Sarstedt, Newton, NC) and allowed to incubate for 1 h at3~C with gentle orbital shaking. Control specimens wereplaced in TBS buffer alone. The samples then were left insolution and stored at -70OC until ready for analysis.
The adsorption of fibronectin onto different sub-strates has been investigated by a number of differentmethods. Emch and coworkers have imaged fibronec-tin with scanning tunneling microscopy (STM) onmica shadowed with a thin platinum carbide-conductive film.15 Using scanning force microscopy(SFM) of fibronectin sprayed on mica and polymeth-ylmethacrylate (PMMA), protein morphology wasfound to be influenced by substrate surface proper-ties.16 The application of this information to in vivosituations was limited because of the nonaqeous envi-ronment used for analysis.
Atomic force microscopy (AFM) is known to pro-vide tofographical images at the fluid-solid inter-face.17.1 Its ability to view surface topography in aliquid medium, unaltered through sample preparationof tip interaction during the scanning process, makesit a valuable analytical tool. For this reason, AFM wasused in this study to investigate the adsorption of fi-bronectin onto polished commercially pure (cp) tita-nium surfaces along with electron spectroscopy forchemical analysis (ESCA) and contact angle measure-ments for characterizing the titanium substrate.
Surface analysis
Electron spectroscopy for chemical analysis (ESCA) of theuntreated and fibronectin (ljLg/ mL) precoated surfaces wasconducted using a Physical Electronics (PHI) model 5600ESCA spectrometer (Eden Prairie, MN). Titanium samples(uncoated and coated by fibronectin) were allowed to reachroom temperature in covered containers, carefully rinsedwith deionized water, dried at room temperature, and thenanalyzed. A monchromatic X-ray source equipped with analurriinum anode (A1Ka = 1486.6 eV) operated at 400 W inthe diffuse mode was used to excite photoemission. Emittedphotoelectrons were analyzed using a spherical capacitorelectron energy analyzer with an Omni Focus 1II small arealens. The analyzer was operated in the fixed analyzer trans-mission mode. The angle between the plane of the sampleand analyzer lens was 65° for all analyses. ESCA surveyspectra were obtained using an analyzer pass energy of 188eV while high resolution multiplex data of the titanium im-plant disks were obtained at a pass energy of 29 eV. Ananalysis area of 800 jLm was used for characterization of thetitanium implant surfaces. An electron flood gun operated at20 mA emission current and 10 eV electron energy was usedto neutralize sample charging during analysis. The residualvacuum inside the spectrometer was 3.1 x 1~ Torr or lowerduring analysis. Data acquisition and storage was accom-plished using an HP Apollo 425 work station running PHI-Access software. The atomic percentages of the elementspresent on the titanium surfaces were calculated using soft-ware and atomic sensitivity factors included with the instru-ment data system. The binding energies of the photoelectronpeaks were referenced to the C is line at 284.6 eV.
MATERIALS AND METHODS
Surface preparation
Two sets of implant specimens were prepared and ana-lyzed, one set custom prepared at the Hospital for SpecialSurgery (HSS) and another obtained commercially from 31Implant Innovations, West Palm Beach, FL.
The titanium disks (12 mm in diameter) prepared at HSSwere from a 1.5 mm thick, commercially pure titanium(cpTi) sheet (President Titanium,. Hanson, MA). The diskswere wet-ground with 320 through 400 grit silicon carbidepaper (smooth samples), deburred, and further polishedwith 85FNEXL and 7S Scotch Brite nonwoven nylon meshwheels (3M, Indianapolis, IN) and then with coarse and fineAluminum oxide compound (Matchless Metal Polish Com-pany, Chicago, IL) on clean Grade NWY 86/87 buff wheels(Divine Brothers Co., Utica, NY). The specimens were rinsedwith distilled water at each step. The metallic samples wereprecleaned ultrasonically in Alkanox followed by a washcycle in a Chemcrest 200 cleaning console unit (Crest Ultra-sonic Corp., Trenton, NJ) at 66°C and spray washed twice indistilled water. Other cpTitanium disks (rough surface) wereprovided by 31 Implant Innovations (West Palm Beach, FL)and prepared according to the manufacturer's specifications.
To ensure cleanliness of all the specimens, the disks (boththe specially fabricated and the commercial) were washedsuccessively in isopropanol, acetone, xylene, acetone, and1M ammonium hydroxide, and, finally, with deionized wa-ter. The samples then were passivated in 40% nitric acid andrinsed three times with deionized water, dried, and storedcovered under a UV light in a cell culture hood (FormaScientific, Marietta, OH). The Ti disks with and without fi-
Contact angle determination
Contact angle analysis was performed on metallic surfacesprior to and after fibronectin binding. Drops of distilled wa-
122 MACOONALD ET AL.
areas per uncoated specimen and computed with a rough-ness analysis program (Digital Instruments, Santa Barbara,CA).
ter were placed on the specimens and contact angles mea-sured using a NRL-l00 contact angle goniometer automatedwith image analysis software (Rame-Hart, Mountain Lakes,NJ). A total of 15 HSS and 15 31 specimens for control andcoated surfaces was analyzed. Eight timed measurementsover a 15 s interval were made for each surface type, and allanalyses were performed at the same temperature andhumidity.
Statistical analysis
The contact angle data were summarized and analyzedusing a two-way analysis of variance (ANOY A) for bothsurfaces, coated or not, and for both HSS and 31 specimens.The alpha level was set at the 0.01 level.Atomic force microscopy
RESULTS
ESCA surface analysis
HSS disks
Carbon, oxygen, titanium, nitrogen and aluminumwere detected on the surface of the HSS prepared cpTidisks [Fig. l(a), Table I]. The carbon concentration wasless than that for the commercially prepared disk (31),indicating that less organic material is present on thesurface. The carbon species detected were similar tothose found on the commercial disk. Nitrogen waspresent in two chemical forms: organic nitrogen (76%)and N+ (24%). Titanium was present primarily asTiO2, with less than 5% titanium metal, suggestingthat the disk was covered with a TiO2 layer. Alumi-num was present on the disk as AI3+, predominantlyas Al2O3'
Titanium samples slowly were defrosted to room tem-perature, carefully handled at the disk edges, and rinsedwith an indirect flow of several milliliters of ultrapure dis-tilled water. Any excess fluid then was wicked off the sidesof the disks using tissue adsorbent paper, and the samplesimmediately were loaded into the AFM (Digital Instru-ments, Santa Barbara, CA). The samples were analyzed atroom temperature and humidity. The MultiMode AFM andNanoScope IlIa controller equipped with an Extender Elec-tronics Module for phase-lag detection were used for the
analysis.In order to reduce the possibility of smearing the protein
molecules on the metallic surface, AFM was performed in aTapping ModeTM. A 125 JLm long, single-arm silicon canti-lever with end-mounted probe tips 5-10 nm in radius ofcurvature (TESP probes) were oscillated at cantilever res0-nance (300 kHz) with the initial amplitude corresponding to1.25 volts. The initial amplitude setpoint was decreased byapproximately JOO/o following engagement to allow goodprobe tip tracking of the sample surface. The integral andproportional gains typically were set to 0.4 and 0.6, respec-tively. Scanning frequencies of 0.3-3 Hz were used.
To visualize the underlying cpTitanium surface after theapplication of a fibronectin coating, an AFM probe tip wasused to carefully "scrape" off the proteinacious layer, creat-ing a window-like effect. A 450 JLm long, single-arm siliconcantilever with end-mounted probe tips 5-10 nm in radius ofcurvature (ESP probes) were used. During probe-tip engage-ment on the sample, cantilever deflection resulted in a 1.0volt change in the vertical difference signal for the photode-tector. For imaging, the "setpoint" then was decreased to-1.0 volt to minimize the probe-tip force on the sample. For"scraping" experiments, the "setpoint" temporarily was in-creased to 5-10 volts. Integral and proportional gains typi-cally were set to 6 and 8, respectively. Scanning frequenciesof 2-10 Hz were used. Scan rotation was 0 degrees for allimaging, with the scan direction parallel to the long axis ofthe cantilever.
31 disks
Carbon, oxygen, nitrogen, and titanium were de-tected on the surface [Fig. 1(c), Table I]. Carbon waspresent in several forms, primarily as CHx, hydrocar-bon. Carbon also was present in functional groupshaving carbon-nitrogen or carbon-oxygen bonds. Ni-trogen was present in two chemical forms, as organicnitrogen (amine or amide) and as singly charged ni-trogen, N+. Approximately 12% of the total nitrogenwas present as N+. Titanium was present almost en-tirely as TiO2. Less than 5% of the total titanium wasmetallic Ti, which would further suggest that the tita-nium disk is covered with a TiO2layer.
ESCA survey spectra obtained of cpTitanium coatedwith 1 J1,g/mL fibronectin (Fig. 1(b,d), Table I] showedan increase in organic nitrogen and carbon when com-pared with uncoated controls. High resolution spectraof the Cis and Nis regions (Fig. 2) also showed theincreased presence of characteristic protein functionalgroups (C-N, C-o, N-C=O).
Surface roughness determination
Quantitative measurements of the local root mean square(RMS) surface roughness were detem\ined using both 3 ~mx 2 ~m and 500 nm x 500 nm AFM scans. The RMS rough-ness is defined as the height fluctuations in a given area.RMS roughness measurements were made in five random
123SURFACE ANALYSIS OF HPF
.-;(c)(a)
-;;c=.§.?;-~...s
""N
,:
:.~
-: ~
4-
..
~ -
(.)(b)
~
c~
~?;-!
~
w
..;~..~
/'./ ~~
c; ~;; .A
~.~
1000 0 1000 800 600 400
Binding energy (eV)
200800 600 ~
Binding energy (eV)
200
Figure 1. ESCA survey spectra of: (a) HSS prepared cp titanium sample; (b) HSS prepared cp titanium sample coated withfibronectin at a concentration of llLg/mL; (c) commercial 31 titanium sample; (d) commercial 31 titanium sample coated withfibronectin at a concentration of llLg/mL.
strated that the presence of a protein coating resultedin a more hydrophobic surface (P < 0.001, Table 11). Nodifference in contact angle measurements could be de-tected in the range of 1 ~g/mL and 10 ng/mL fibro-nectin concentrations for coating (P = 0.2).
Contact angle measurements
Contact angle measurements of the control HSS and31 cpTidisks showed an average contact angle of 36.3:t 3.5 and 39.1 :t 3.1, respectively (Table II). No differ-ence was noted between the receding and advancingcontact angles of either of the control specimens. Acomparison of contact angle measurements made be-tween native and coated titanium specimens demon-
Atomic force microscopy
TABLE IESCA Results for HSS Prepared and Commercial 31
Titanium Samples Uncoated and Coated With HumanPlasma Fibronectin
Fibronectin-CoatedH55 Prepared cpTi Sample
Element Concentration (%)
In AFM, polished surfaces of HSS disks demon-strated areas of polishing grooves with small regionsof surface pitting [Fig. 3(a)]. Commercially prepared 31specimens showed a more irregular topography [Fig.3(b)] with large parallel grooves.
When the same surfaces were exposed to 10 ng/mLof human plasma fibronectin in a buffered TBS solu-tion, the surfaces appeared coated by many smallcompact protein clusters [Fig. 3(d)}. The control s~i-mens were void of any protein material (Fig. 3(c)]. Theprotein clusters did not appear to show a preferencefor areas containing polishing grooves and wereevenly dispersed on the HS5-prepared samples. Onthe commercially prepared 31 specimens, the proteinappeared to aggregate within the deeper surface-finishing grooves. At a higher magnification with ascanned area of 74.4 nm, the fibronectin molecule ap-peared in globular form, with an average length of16.5:t. 1.0 nrn [Fig. 3(e)}. The average height of tenseparate molecules was measured to be 2.5 :t 0.5 nrn,
HSS ~ cpTi ~pleConcentration (%)Element
CisOt5
Ti2pNt.
Al2p
19.9054.3520.76
1.763.23
CIs
OIS
Ti2pNIS
Al2p
25.5449.5417.644.342.78
Fibronectin-Coated31 cpTi Sample
Element Concentration (%)31 cpTi SampleElement Concentration (Ofo)
Cl.
OlS
Ti2pN1s
23.9953.1721.131.40
Cts
Ots
Ti2p
N1s
31.8348.0517.452.67
/'-" .,-1 .' .'~J~~-!; ~:I... . C
MACOONALD ET AL.124
Figure 2. High resolution ESCA spectra of HSS cp titanium sample uncoated (a and c) and coated (b and d) with fibronectin.Cts and Nt. regions are presented.
3 ~m x 2 ~m and 3.87 nrn j: 2.6 at a scanned area of 500nrn x 500 nrn. The 31 specimens revealed numerousscratch marks from wet grinding with a correspond-ing increase in RMS scores in comparison with HSS
prepared disks.
and the width measured was 9.6:i: 1.2 nm. Quantifi-cation of protein size was not possible on the 31 speci-mens due to the irregular nature of the surface.
By increasing the concentration of fibronectin to 1JA,g/mL, a mat-like surface coating was observed withAFM [Fig. 3(f)] on the HSS specimen. When the AFMstylus purposely was pulled across the coating, thecrosslinked fibronectin layer rolled back upon itself,revealing the underlying metallic surface [Fig. 3(g)]. DISCUSSION
Surface roughness detem\ination
The RMS roughness measurement for commerciallyprepared 31 disks was 22.37 nm :t 4.17 at a scannedarea of 3 JLIn x 2 JLm and 12.23 nm 1: 3.42 whenscanned at SOO nm x 500 nm. Extensively finished sur-faces prepared at HSS were considerably smootherwith RMS values of 4.45 nm j: 0.46 at a scanned area of
The excellent biocompatibility of titanium implantsis due to the favorable titanium oxide-tissue interac-tion that occurs in vivo. An important protein involvedin cell binding is fibronectin. Plasma fibronectin, a gly-coprotein present in blood, has been shown to pro-mote cell adhesion, wound healing, and embryonicdevelopment. 19 In this study, the appearance of fibro-nectin binding onto cp Titanium surfaces of differentcharacteristics was visualized using tapping modeatomic force microscopy.
The ESCA survey spectrum of both HSS and 31 pre-pared cpTi disks demonstrated the presence of C, N,0, and Ti [Fig. l(a,c»). Higher resolution of the Ti2pspectrum for both control specimens identified threepeaks whose positions and shapes were consistent forTi02 as reported in the literature.3,2G-22 The ESCAspectra of samples immersed in a fibronectin solutionof different concentrations showed increased carbonand nitrogen signals characteristic of protein deposi-tion [Fig. 1(b,d)].23,24 Nitrogen was present in twoforms: organic nitrogen and singlet N+. This chargedmoiety is probably contributed by amino acids withinthe protein under the particular condition of the
T ABLE IIMeasurements of Water Contact Angle for H55 and 31Titanium Samples Uncoated and Coated with Human
Plasma Fibronectin
Sample Contact Angle (O)
39.1 :!53.3i55.9 i36.3.i54.4:153.0j
Commercial TiCommercial Ti + HFN (10 ng/mL]Commercial Ti + HFN (1 ILg/mL)HSS TiHSS Ii + HFN (10 ng/mL)HSS Ti + HFN (llLg/mL)
Measurements are mean:t standard deviation. -Statisti-cally significant at P ~O.Ol.
:3.1:4.7":3.0-:3.5:2.~t6.5-
SURFACE ANALYSIS OF HPF 125
4.00 PM00 5.00 JIMAMPlitude
'1. 00 ,
Dat~ typeZ ranqe
Data type2 range
AMp. itl.de5.00 nM ba
1.67 JIM 0 1.50 UN0AMplitude
5.00 nMData type2 range
AMplitude5.00 nM
Data type2 rangedc
Figure 3. Atomic force microscopy (AFM) of cp titanium samples uncoated and coated by hwnan plasma fibronectin: (a)Tapping ModeTM AFM image of HSS prepared cp titanium sample (scan size = 5 IJ.m); (b) Tapping ModeTM AFM image of31 prepared cp titanium sample (scan size = 4IJ.m); (c) Tapping ModeTM AFM image of HSS prepared cp titanium sample (scansize = 1.671J.1n); (d) Tapping ModeTM AFM image of HSS prepared cp titanium sample coated with hwnan plasma fibronectinat concentration of 10 ng/mL (scan size = 1.5 IJ.m); (e) High resolution Tapping ModeTM AFM image showing globular formof hwnan plasma fibronectin on HSS prepared cp titanium sample. Amplitude (left) and phase (right) mode (scan size = 74.4nm); (f) Contact mode AFM image of HSS prepared cp titanium sample preincubated with 1 IJ.g/mL of hwnan plasmafibronectin before intentionally raised force scraping of protein layer. (deflection mode, scan size = 4.13 1J.In); (g) Contact modeAFM image of the same area in Figure 3{f) showing the rolled mat-like fibronectin surface coating exposing the metallicundersurface on HSS prepared cp titanium sample (deflection mode, scan size = 4.13 1J.In).
126 MACDONALD fiT At
0 74.4 nM 74.4 nM0Data type2 range
AMplitude2.00 nM
Data type2 range
Phase40.0 dee
0 4.13 JIM 0
9
4.13 JIMData type2 range
Deflection0.500 nM
Deflection0.500 nM
Data type2 rangef
Figure 3. Continued.
analysis. The carbon species detected on the fibronec-tin-treated disks also were consistent with the pres-ence of protein as C-N, C-O, and N-C=O functional-ities were observed (Fig. 2). Slight amounts of polish-ing residue might account for the aluminum signalobserved. A comparison of the titanium signal be-tween control and coated specimens showed a reduc-tion for the latter. This is best explained by the depththe beam penetrates. Because only the top 10 nm of thesurface of each specimen is analyzed, the presence of
a proteinacious coating would reduce the titanium sig-nal due to the increased thickness the beam has topenetrate. The observed reduction in oxygen signalcould be for a similar reason.
The wettability of any biological material dependson the magnitude of intermolecular forces present atthe solid and liquid interface, expressed here in termsof contact angle 8.25 Wettability of particles appears tobe surface roughness dependent. With increasing sur-face roughness the receding angle decreases and the
127SURFACE ANALYSIS OF HPF
ules with an average length of 15.5 :!: 1.3 nm inlength.36.37 In contrast, rotary shadowing of fibronec-tin on mica: has revealed lengths of 60 :!: 7 nm to arange of 120 to 160 nm.38.J9 This variability of fibro-nectin chain length has been attributed to the extremeflexibility of the protein.39 Fibronectin has been re-ported to have globular and filamentous forms de-pending on substrate and solution conditions.4() Theglobular form of a particle may represent collapsedfibronectin molecules, which is the form foW1d in so-lution..1 In our study, the average height of ten sepa-rate molecules was measured to be 2.5 :!: 0.5 nm, whichis in accord with the reported value of 2.0 nm for thediameter of the fibronectin molecule based on electronmicroscopic analysis on mica.39
The conformation of the HPF molecule will varydepending on whether or not the molecule is in solu-tion or is surface boW1d. Moreover, factors such astemperature, pH, and ionic strength of the local envi-ronment also will playa role in determining the mo-lecular shape of fibronectin. In solution W1der normalphysiologic conditions, fibronectin has been reportedto assume a globular or compact form.42.43 This com-pact form of fibronectin is stabilized by interdomainelectrostatic interactions.« At temperatures above40°C, extreme pH levels, or increased ionic strength,fibronectin assumes a more extended form.«-48 Fibro-nectin is extremely flexible, and this extension due tolocal environmental changes does not result from de-naturation of the molecule.46 Based on the above con-siderations, it can be assumed that our experimentalconditions of pH 7.4 and a temperature of 37OC wouldfavor plasma fibronectin in a globular-like conforma-tion when in solution. Unfortunately, little is knownabout the effect of these two factors on the structuralnature of this protein molecule once it is adsorbed ona substrate. As indicated above, aside from tempera-ture and pH, fibronectin conformation also may beionic strength dependent, with the molecule being acompact form in a low salt environment and in anextended form in a high salt environment.49 This in-terpretation has been questioned based on fluores-cence energy studies which, in high salt solutions, de-scribes a fibronectin molecule in which both the aminoand carboxyl termini are juxtaposed whereas the cen-tral regions expand, creating a "balloon-like" effect. 50
The conformation of fibronectin changes when it isadsorbed on a variety of surfaces. Erickson and Carrellreported that at high ionic strength (0.2M NaCI), fi-bronectin appeared as an extended molecule when vi-sualized by sedimentation analysis on a mica sur-face.51 At low ionic strengths (0.02M NaCI), fibronec-tin assumed a more compact structure. Half moleculesand proteolytic fragments of fibronectin missing theterminal NH2 or COOH groups underwent ionicstrength-dependent changes in conformation similarto unaltered molecules. These conformational changes
advancing angle increases!6 It has been reported inthe past that totally hydrophobic materials can be con-verted to hydrophilic ones by roughening their sur-faces.26-28 A comparison of contact angle measure-ments on sterilized polished and unpolished cpTi-coined flats showed larger contact angles for theunpolished sterilized Specimens.29 Although the con-tact angles on the rougher commercially prepared 31specimens were larger when compared with the pol-ished HSS prepared group, the difference was not sta-tistically significant (P > 0.05). This is in agreementwith reports that mechanically polishing or buff f;!l-ishing metals has little effect on contact angle. 31
Some degree of surface roughness will result from me-chanical polishing, and it is possible that the magni-tude of roughness determines the overall contactangle. Our results, along with this, suggest that sur-face roughness will contribute to the wettability of asurface, yet a more important factor may be changes insurface chemical composition due to various environ-mental interactions and surface oxidation states.
As confirmed by the ESCA results, titanium and itsalloys are covered mainly with TiO2.21,22,25.30,31 Thisphase determines both the contact angle and spread-ing behavior of a liquid drop on the surface.30 More-over, the process of passivation will affect the surfaceoxide layer .32 Although at elevated temperatures theseoxide layers undergo phase changes,33 at the tempera-tures used for sterilization (autoclave, dry heat, UV)there appears to be no effect on contact angle mea-surements for cpTi.29,34 However, the contact anglesmeasured are lower than those reported in the litera-ture,30,34 The differences in surface treatment used inthese studies account for the variations noted in con-tact angle between our specimens and those in theliterature. This suggests that passivation, surface treat-ment, and temperature are important factors in deter-mining the wettability of biological materials.29
Surface roughness is another variable. The AFM im-ages of the commercially obtained 31 specimens andHSS samples showed a consistently more scratchedsurface on the former [Fig. 3(a,b)]. The RMS roughnessmeasurements also were greater for the 31 specimens.Both the size of the scan and the area evaluated playeda role in the RMS roughness measurements. With bothsamples analyzed, the RMS roughness scores in-creased with larger areas scanned. A possible expla-nation is that greater scanned areas can visualize sur-face features that might be overlooked on smaller
scannings.35At higher magnifications with a scanned area of 74.4
nm, the fibronectin molecule appeared in the globularform with an average length of 16.5 :f: 1.0 nm [Fig.3(e)], This is comparable with the structure describedby Koteliansky and coworkers36 based on electron mi-croscopic observation of fibronectin shadowed withtungsten-tantalum. They reported fibronectin glob-
128 MACDONALD ET AL.
tingS.16 Moreover, in this work by Emch and cowork-ers,16 all the extended fibronectin proteins appear tobe oriented in the.same direction. If protein binding isarbitrary and random, it would be difficult for all theproteins to assume the same "V" shape and same di-rectional orientation. Possibly the force with which thesprayed protein contacts the surface may be sufficientto overcome some of the short range electrostatic in-teractions of the protein strand, creating an illusion ofprotein unfolding. Fibronectin in our study was al-lowed to react with the surface from a suspension, andno attempt was made to vigorously shake the reactionvessel nor to spray the metal surfaces with the pr:oteinto encourage binding. Therefore, the physical manipu-lation of the protein is proposed to have an effect onthe shape it assumes on a surface.
Emch and coworkers16 employed a volatile buffer inplace of salt buffers since, if allowed to dry, precipi-tated salt crystals would obscure the resolution of thesmaller fibronectin molecule. On titanium, we foundthis not to be the case provided the samples were keptmoist throughout the procedure and carefully wererinsed several times with triple distilled water prior toanalysis. This procedure would result in a lower ionicstrength at the metallic-water interface. Samples wereair dried to SOO/o relative humidity, thus leaving a wa-ter layer on the surface. Therefore, there was no expo-sure of the bound fibronectin to a high salt environ-ment nor was there complete drying of the buffer con-taining the protein. Our findings suggest that the localiorlic strength at the metallic-water interface duringspecimen preparation might be significant enough toaffect the resultant protein conformation, whetherglobular or extended.
Upon increasing the concentration of fibronectin to1 fLg/mL, a mat-like surface coating was observedwith AFM [Fig. 3(f)]. When the AFM stylus purposelywas pulled across the coating, the crosslinked fibro-nectin layer could roll back upon itself revealing theunderlying metallic surface [Fig. 3(g)]. These resultssupport earlier observations that fibronectin has thecapacity to form an interdigitating protein layer .55Once fibronectin-fibronectin crosslinking resulting ina fibrillar state in solution occurs, this complex dem-onstrates characteristics of an insoluble monolayer.56Thus, conformational changes appear to be necessaryfor the conversion of the soluble form of fibronectin toan insoluble form. Self assembly of fibronectin is cen-tral in the formation of the extracellular matrix. Sev-eral modules (4-12) located at the amino terminal endof each fibronectin arm are required for matrix assem-bly by fibroblasts and exhibit fibronectin-fibronectinbinding properties.57-W Other segments (module III)are buried in intact insoluble fibronectin and becomeactive only after surface binding.59.60 Matrix assemblyin vivo requires disulfide exchange in which two di-sulfide bonds of neighboring fibronectin molecules ex-
were attributed to short range electrostatic interac-tions along the strand rather than to attraction of dis-tant molecular segments. 51 In our investigation, use of
a O.15M salt solution did not result in the extendedstrands observed by Erickson and Carrell on mica.51This could be due to the different substrate, namelycpTi, and its effect upon protein conformation. Undersimilar salt and buffer conditions as in our study, Ko-teliansky and coworkers36.37 observed the fibronectinto assume a compact protein structure. At the sameinitial salt concentration as that of our experiment,Wolff and Laiso reported some separation between theamino termini of plasma fibronectin immobilized ontoCytodex dextran beads, suggesting that some confor-mational changes may occur in the binding process. Apossible explanation for the difference in appearance(folded versus unfolded) could be due to the differentsubstrates used and the geometry of the substrate:beads versus smooth polished surfaces. Total internalreflection fluorescence (TIRF) of the local tryptophanmicroenvironment when plasma fibronectin was ad-sorbed to hydrophobic silica surfaces show the tryp-tophan groups shift into a more hydrophilic microen-vironment.52 Such a shift was not observed when theprotein was tested on hydrophilic silica surfaces. Fluo-rescence energy transfer methods have been em-ployed to estimate changes in intramolecular distancebetween plasma fibronectin chains before and afteradsorption to hydrophilic Cytodex beads.47,52 The en-ergy transfer of fluorescein-labeled amino terminusgroups and coumarin-labeled sulfhydryl groups wascompletely reduced, suggesting that interchain sepa-ration occurs after protein absorption. It should benoted that there was a high signal-to-noise ratio,which would indicate that the protein was not com-pletely extended and that some variability in the re-sults exists. Of the two protein chain sulfhydrylgroups, one (SH-l) becomes exposed while the other(SH-2) remains buried when plasma fibronectin bindsto polystyrene beads.53.54 It is clear that there are anumber of variables that dictate the conformation fi-bronectin may assume upon substrate binding. Theseconformational changes indeed can be expected to in-fluence further the protein's exposed potential bind-ing sites.
Scanning force micrographs of fibronectin adsorbedon hydrophilic (mica) and hydrophobic (polymethyl-methacrylate) surfaces have shown that fibronectinshape is substrate dependent.16 On the hydrophilic. mica surfaces, fibronectin exhibited a V -shaped orien-
tation whereas when sprayed on a hydrophobicPMMA surface, the protein formed a thin, densemeshwork over the surface. A consistent problemnoted with SFM using traditional contact scanningwas the destruction caused by the tip-surface interac-tion during the scanning process and destruction ofthe protein surface coating at higher resolution set-
129SURFACE ANALYSIS OF HPF
change with their binding partners.61,62 This disulfide-bonded climer structure is critical for fibronectin fibrilformation. 59 Fibronectin molecules therefore can be
proposed to line up with one another in such a man-ner as to optimize disulfide crosslinking.56 This mayenhance the interaction with the titanium surface.
CONCLUSIONS
Our results for fibronectin interaction with commer-cially pure titanium show the fibronectin protein tohave a small globular structure of 16.5 :t 1.0 nm inlength. 2.5 :t 0.5 nm in height, and 9.6 :t 1.2 nm inoverall diameter when bound to a cpTitanium surface.The average contact angles for cpTi control specimenswere 36.3 :t: 3.5 and 39.1 :t: 3.1 for the HSS and com-mercially prepared 31 specimens, respectively. In thepresence of fibronectin, the surface appears more hy-drophobic due to exposed moieties. Both survey andhigh resolution ESCA spectra have shown the pres-ence of surface groups that are characteristic for ad-sorbed protein molecules. At higher concentrations,fibronectin appears to form a thin crosslinked matrixthat can be lifted off the surface. It is important to notethe implications of these conformational changes onthe interaction with tissue cells. Future studies shouldfocus on the effect of these interactions.
Generous support was received from a Physicians CareerDevelopment A ward, V. A. Central Office, during the prepa-ration of this article and for the studies described herein.Special thanks goes to Dr. Michael Stranick (Colgate Palmo-live Company, Piscataway, Nn for the ESCA analysis and toKelly O'Hara (Rame-Hart, Inc., Mountain Lakes, NJ) for useof the NRL-lOO-OO-S automatic contact angle goniometer.Thanks also go to Dr. Timothy Wright, Hospital for SpecialSurgery (NY, NY) for preparation of the polished cpTita-nium specimens and to Dr. Richard Caudill of the 31 ImplantInnovations Corporation (West Palm Beach, FL) for furnish-ing the commercial cpTi samples.
.
References
teration in cell surface characteristics:' Dev. Bioi., 64, 31-4.7
(1978).6. W. Dessau"J. Sasse, R. Timpl, F. Jilek, and K. von der Mark,
"Synthesis 'and extracellular deposition of fibronectin in ch0n-drocyte cultures," J. Ceu Bioi., 79, 342-355 (1978).
1. R. M. Pratt and K. M. Yamada, "Enhanced cellular fibronectinaccumulation in d\Ondrocytes treated wid1 vitamin A:' Cell,
11,821-826 (1979).8. R. A. Proctor, "Fibronedin: A brief overview of its structure,
function. and physiology:' Rev. Infrct. Dis., 9, 5317-5321 (1987).9. F. Grinnell and M. K. Feld, "Flbronectin adsorption on hydro-
philic and hydrophobic surfaces detected by antibody bindingand analyzed during cell adhesion in senIm containing solu-tion," ,. Bioi. Chem., 251, 4888-4893 (1982).
10. D. D. McAbee and F. GrinnelL "Fibronectin mediated bindingand phagocytosis of polystyrene latex beads by baby hamsterkidney cells:' 1. Cell BioI., 91,1515-1523 (1983).
11. C. E. Wolff and c.-s. Lai, "Fluorescence energy transfer detectschanges in fibronectin structure upon surface binding," Arch.Biochem. Biophys., 268, 536-545 (1989).
12. C. E. Wolff and C.-S. Lai. "Inter-sulphydryl distances inplasma fibronectin determined by fluorescence energy trans-fer: Effect of environmental factors:' Biochemistry, 29, 3354-3361 (1m).
13. S.-S. Cheng, K. K. Chittur, C. N. Sukenik, LA. Culp, and K.Lewandowska, "The conformation of flbronectin on self-assembled monolayers with different surface composition: anFfIR/ATR study:' 1. Colloid Interfiu:e Sci., 162. 135-143 (1994).
14. R. W. Watkin and C. R. Robertson. "A total internal-reflectiontechnique for the examination of protein adsorption:' 1.Biomed. Milter. Res., 11, 915-938 (1977).
15. R. Emch, X. Oivaz, C. Taylor-Denes, P. Vaudaux, P. Descouts,"Scanning tunneling microscopy for studying the biomaterial-biological tissue interface:' 1. Vac. Sci. TechnoL, AS, 655-658
(1m).16. R. Emch, F. Zenhausern, M. Jobin, M. Taborelli, and P.
Descouts, "Morphological difference between fibronectinsprayed on mica and on PMMA," Ultramicroscopy, 42-44, 1155-
1160 (1992).17. G. BinNg, C. F. Quate, and C. Gerber, "Atomic force micro-
scope:' Phys. Rev. Left., 56, 930-933 (1986).18. C. Wyman, E. Grotkopp, C. Bustamante, H. C. M. Nelson, "De-
termination of heatshock transcription factor 2 stoichiometry atlooped DNA complexes using scanning force microscopy,"EMBO 1.,14,117-123 (1995).
19. R. O. Hynes, Fibronectins, Springer Series in Molecular Biology,A. Rich (ed.), Springer-Verlag, New York, 1m, pp. 200-364.
20. N. R. Armstrong and R. K. Quinn, "Auger and X-ray photo-electron spectroscopic and electrochemical characterization oftitanium thin film electrodes," Surf Sci., 61, 451-468 (1977).
n. K. E. Healy and P. Ducheyne, "Oxidation kinetics of titaniumfilms in model physiologic environments," 1. Colloid lnteTfilce
Sci., 150,404-417 (1992).22. K. E. Healy and P. Ducheyne, "Hydration and preferential mo-
lecular adsorption on titanium in vitro," Biomateria/s, 13, 553-
561 (1992).23. K. E. Dombrowski. S. E. Wright, J. C. Birkbeck, and W. E. Mod-
deman, "X-ray photoelectron spectroscopy of amino acids,polypeptides and simple caIbohydrates," in Met~ in ProteinSurface Analysis, M. Z. Atassi and E. Appella (eds.), PlenumPress, New York, 1995, pp. 251-260.
24. P. A. Gerin, P. B. Dengis, and P. G. Rouxhet, "Performance ofXPS analysis of model biochemical compounds," ,. Chim. Phys.,
92, 1043-100 (1995).25. R. E. Baier, R. C. Dutton, and V. L Gott, in Surface Chemistry of
Biological Systems, Plenum Press, New York, 1970, pp. 235-260.
26. J. F. Oliver, C. Huh, and S. G. Mason, "An experimental study
1. B. D. Ratner, D. C. Castner, T. A. Horbett, T. J. lenk, K. B.lewis, and R. J. Rapoza, "Siomolecules and surfaces:' J. Vac.Sci., AS, 2306-2317 (1990).
2. D. M. Brunette, "Interactions of epithelial cells with foreignsurfaces:' CRC Crit. Rev. Biocom~t., L 323-370 (1986).
3. B. Kasemo and J. Lausmaa, "Surface science aspects on inor-ganic biomaterials:' CRC Crit. Rev. BilXompat., 2., 335-380(1986).
4. A. C. Brownell, C. C. Bessem, and H. C. Slavkin, "Possiblefunction of mesenchyme cell-derived fibronectin during for-mation of basal lamina;' Proc. Nat!. Acad. Sci. USA, 78, 3711-3715 (1981).
5. C.A. lewis, R. M. Pratt, J. P. Pennypacker, and J. R. Hassell,"lnhibition of limb chondrogenesis in vitro by vitamin A: Al-
MACOONALD ET AL.130
of some effects of solid surface roughness on wetting," Coil.Surfaces, 1, 79-104 (1980).
27. R. D. Kulkarni and P. Somasundaran, "Mineralogical hetero-geneity of ore particles and its effects on the interfacial char-acteristics,' Powder Tech., 14, 279-285 (1976).
28. I. J. Lin and P. Somasundaran, "Alterations in properties of
samples during their preparation by grinding." Powder Tech., 6,171-180 (1972).
29. D. V. Kilpadi and J. E. lemons, "Surface energy characteriza-tion of unalloyed titanium implants," T. Biomed. MAter. Res., 28,1419-1425 (1994).
30. Y. Oshida. R. Sachdeva, and s. Miyazaki. "Cl1anges in contactangles as a function of time on some pre-oxidized biomateri-als," T. MAter. Sci. MAter. Med., 3, 306-312 (1992).
31. Y. Oshida, R. Sachdeva, S. Miyazaki, and J. Daly, "Effects ofshot-penning on surface contact angles of biomaterials," T.Mater. Sci. MAter. Med., 4, 443-447 (1993).
32. P. Ducheyne, "Titanium and calcium phosphate ceramic den-tal implants, surface, coatings, and interfaces," Oral Implllntol.,14, 325-340 (1988).
33. A. Wisbey, P. J. Gregson, L. M. Peter, and M. Tuke, "Effect ofsurface treatment on the dissolution of titanium-based implantmaterials," BioIIIIJterillls, 12. 470-473 (1991).
34. D. Kawahara, Y. Kimura, M. Nakamura, and H. Kawahara,"Studies on the tissue adhesive capability to titanium by dy-namic wettability test and cell attachment, in vitro," Clin.MAter., 14, 229-233 (1993).
35. C. M. Demanet, S. Shrivastava, and J. P. F. SeI1schop, "Scan-ning force microscopy analysis of the surface of ion-irradiateddiamond," Surf Interface Anal., 23,115-119 (1995).
36. V. E. Koteliansky, M. V. Bejanian. and V. N. Smirnov, "Elec-tron microscopy study of fibronectin structure," FEBS Left.,
120,283-286 (1980).37. V. E. Koteliansky, M. A. Glukhova, M. V. Bejanian, V. N.
Smirnov, V. V. Filirnonov, O. M. Zalite, and S. Y. Venyaminov,"A study of the structure of fibronectin," Eur.T. Biochem., 119,619-624, (1981).
38. J. Engle, E. Odermatt, A. Engel, J. A. Madri, H. Furthmayr, H.Rohde, and R. TlD\pl, "Shapes, domain organizations and flex-ibility of laminin and fibronectin, two multifunctional proteinsof the extracellular matrix," 1- Mol. Bioi., 150,97-120 (1981).
39. H. P. Erickson, N. CarrelL and J. Mc~nagh, "Fibronectin mol-ecule visualized in electron microscopy: A long. thin, flexiblestrand," 1- Cell. Bioi., 91, 673-678 (1981).
40. E.G. Williams, P.A. Janmey, J.D. Ferry, and D.F. Mosher,"Conformational states of fibronectin: Effects of pH, ionicstrength, and collagen binding." 1- Bioi. Chem., 2S7, 14973-14978 (1982).
41. E. Odematt and J. Engel, "Physical properties of fibronectin,"in Fibronectin, Biology of Extractllulllr MAtrix, D. F. Mosher (ed.),Academk Press, Inc., New York, 1989, pp. 2S-45.
42. B. Sj(;berg, M. Eriksson, E. Osterlund, S. Pap, and K. Osterlund,"Solution structure of human plasma fibronectin as a functionof NaCI concentration determined by small-angle neutron dif-fraction," Eur. 1- Biophys., 17, 5-11 (1989).
43. M. J. Benecky, C. G. Kolvenbach, R. W. Wine, J. P. DiOrio, andM. W. Mosesson, "Human plasma fibronectin structureprobed by steady-state fluorescence polarization: Evidence fora rigid oblate structure," Biochemistry, 29, 3082-3091 (1990).
44. M. Y. Khan. M. s. Medow, and S. A. Newman, "Unfoldingtransitions of fibronectin and its domains, stabilization and
structural alternation of the N-terminal domain by heparin:'Biochem. 1., 270, 3~38 (1990).
45. M. J. Benecky,~. W. Wine, C. G. Kolvenbach, and M. W. Mo-sesson. "Ionic-strength- and pH-dependent conformationalstates of human plasma fibronectin:' Biochemistry, 30, 4298-4306 (1991).
46. E. Osterlund, "The secondary structure of human plasma fi-bronectin: Conformational changes induced by acidic pH andelevated temperatures; a circular dichroic study:' Biochim. Bfo-phys. Acta, 955, 330-336 (1988).
47. M. Rocco, E. Infusini, M. G. Daga, L. Gogioso, and C. Cunib-erti, "Models of fibronectin:' EMBO 1., 6, 2343-2349 (1987).
48. C.-s. Lai and N. M. Tooney, "Electron spin resonance spin la-bel studies of plasma fibronectin: Effect of temperature:' Arch.Biochem. Biophys.. 228,465-473 (1984).
49. S. s. Alexander, Jr., G. Colonna, and H. Edelhoch, "The struc-ture and stability of human plasma cold-insoluble globulin:' ,.BioI. Chem., 254, 1501-1505 (1979).
SO. C. Wolff and C.-S. Lai, "Fluorescence energy transfer detectschanges in fibronectin structure upon surface binding:' Arch.Biochem. Biophys.. 268,536-545 (1989).
51. H. P. Erickson and N. A. Carrell, "Fibronectin in extended andcompact conformations:' 1. BioI. Chem., 258, 14539-14544(1983).
52. G. K. Iwamoto, L. C. Winterton, R. S. Stoker, RA. van Wa-genen. J. D. Andrade, and D. F. Mosher, "Fibronectin adsorp-tion detected by interfacial fluorescence:' 1. CoUoid Interfact?
Sci., 106, 459-464 (1985).53. C. Narasimhan, C.-S. Lai, A. Haas, and J. McCarthy, "One free
sulfhydryl group of plasma fibronectin becomes titratableupon binding of the protein to solid substrates," Biochemistry,27, 4970-4973 (1988).
54. C. Narasimhan and C.-S. Lai, "Conformational changes ofplasma fibronectin detected upon adsorption to solid sub-strates: A spin-labeled study," Biochemistry. 28, 5041-5046
(1989).55. D. F. Mosher. "Physiology of fibronectin:' Ann. Rev. Med., 35,
561-575 (1984).56. V. Vogel, "Fibronectin in surface-adsorbed state:' in Proteins at
Interfizces. II. Fundamentals and Applications. T. A. Herbett andJ. L. Brash (eds.), American Chemical Society, Washington DC,1995, pp. 505-518.
57. J. A. McDonald, B. J. Quade, T. J. Broekelmann, R. laChance,K. Forsman. E. Hasegawa, and S. Akiyama, "Fibronectin cell-adhesive domain and an amino-terminal matrix assembly do-main participate in its assembly into fibroblast:' 1. Biol. Chem.,262. 2957-2967 (1987).
58. J. Sottile, J. Schwarzbauer, J. Selegue, and D. F. Mosher, "Fivetype I modules of fibronectin form a functional unit that bindsto fibroblasts and Staphylococcus aureus," 1. BioI. Chern., 266,12840-12843 (1991).
59. J. E. Schwarzbauer, "Identification of fibronectin sequences re-quired for assembly of a fibrillar matrix:' 1. Cell. BioI., 113,1463-1473 (1991).
60. A. Morla and E. Ruoslahti, "A fibronectin self-assembly siteinvolved in fibronectin matrix assembly: Reconstruction in asynthetic peptide:' 1. Cell BioI., 118,421-429 (1992).
61. D. F. Mosher and R B. Johnson. "In vitro formation of disul-fide-bonded fibronectin multimers:' 1. BioI. Chem., 258, 6595-fBJl (1983).
62. D. F. Mosher, F. J. Fogerty, M. A. Chernousov, and E. L. RBarry, "Assembly of fibronectin into extracellular matrix,"Ann. NY Acad. Sci., 614. 167-180 (1991).
![Td Adsorbed (Tetanus and Diphtheria Toxoids …products.sanofi.ca/en/td-adsorbed.pdfTd ADSORBED [Tetanus and Diphtheria Toxoids Adsorbed], is a sterile, cloudy, white, uniform suspension](https://img.pdfslide.net/doc/110x75/5e5ed39d07f6e0285b51c50f/td-adsorbed-tetanus-and-diphtheria-toxoids-td-adsorbed-tetanus-and-diphtheria.jpg)
