Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/14341665
Surfaceroughnessmodulatesthelocal
productionofgrowthfactorsandcytokinesby
osteoblast-likeMG-63cells.JBiomedMater
Res32(l):55-63
ArticleinJournalofBiomedicalMaterialsResearch·September1996
DOI:10.1002/(SICI)1097-4636(199609)32:1<55::AID-JBM7>3.0.CO;2-O·Source:PubMed
CITATIONS
READS
418
548
7authors,including:
ZviSchwartz
DavidDDean
VirginiaCommonwealthUniversity
UniversityofTexasHealthScienceCentera…
539PUBLICATIONS19,477CITATIONS
230PUBLICATIONS12,329CITATIONS
SEEPROFILE
SEEPROFILE
BarbaraDBoyan
VirginiaCommonwealthUniversity
442PUBLICATIONS19,111CITATIONS
SEEPROFILE
Availablefrom:BarbaraDBoyan
Retrievedon:09May2016
Surface roughness modulates the local production
of growth factors and cytokines by osteoblast-like
MG-63 cells
K. Kieswetter,' Z. S~hwartz,',~,~
T. W. H~mrnert,',~D. L. Cochran; J. Simpson,S D. D. Dean,'
and B. D. B~yanl,*,~,"
Departments of 'Orthopaedics, 2Periodontics, and 3Biochemistry, The University of Texas Health Science Center at Sun
Antonio, San Antonio, Texas 78284; 'Hebrew University Hadassah Faculty of Dental Medicine, Jerusalem, Israel;
51nstitut Sfraurnann AG, Waldenburg, Switzerland
Titanium (Ti) surface roughness affects proliferation, differentiation, and matrix production of MG-63 osteoblast-like
cells. Cytokines and growth factors produced in the milieu
surrounding an implant may also be influenced by its surface,
thereby modulating the healing process. This study examined the effect of surface roughness on the production of
two factors known to have potent effects on bone, prostaglandin E, (PGEJ and transforming growth factor PI (TGF-PI).
MG-63 cells were cultured on Ti disks of varying roughness.
The surfaces were ranked from smoothest to roughest: electropolished (EP), pretreated with hydrofluoric acid-nitric
acid (PT), fine sand-blasted, etched with HC1 and HISO,, and
washed (EA), coarse sand-blasted, etched with HC1 and
H2S04,and washed (CA),and Ti plasma-sprayed (TPS).Cells
were cultured in 24-well polystyrene (plastic) dishes as controls and to determine when confluence was achieved. Media
were collected and cell number determined 24 h postconfluence. PGE2 and TGF-PI levels in the conditioned media
were determined using commercial radioimmunoassay and
enzyme-linked immunosorbent assay kits, respectively.
There was an inverse relationship between cell number and
Ti surface roughness. Total PGEz content in the media of
cultures grown on the three roughest surfaces (FA, CA, and
TPS) was significantly increased 1.5-4.0 times over that
found in media of cultures grown on plastic or smooth surfaces. When PGEzproduction was expressed per cell number,
CA and TPS cultures exhibited six- to eightfold increases
compared to cultures on plastic and smooth surfaces. There
was a direct relationship between TGF-P, production and
surface roughness, both in terms of total TGF-PI per culture
and when normalized for cell number. TGF-P, production
on rough surfaces (CA and TPS) was three to five times
higher than on plastic. These studies indicate that substrate
surface roughness affects cytokine and growth factor production by MG-63 cells, suggesting that surface roughness may
modulate the activity of cells interacting with an implant,
and thereby affect tissue healing and implant success. 0 1996
INTRODUCTION
Local autocrine and paracrine factors have been
shown to be of importance in the development of
b ~ n e . Mesenchymal
~-~
cells have been shown to be capable of differentiating along one of several pathways
based upon the concentration and presence of local
The presence and concentration of these factors at the implant site during the initial stages of
wound healing may therefore play a role in the type of
cells recruited to the site as well as in their phenotypic
expression, ultimately influencing the long-term biologic response to the device.
The interactions at the site of titanium (Ti) devices
are of considerable interest, given the wide clinical use
of this material.9Previous in vitro studied0 have shown
that chondrocytes grown o n Ti surfaces exhibit greater
matrix production than do chondrocytes grown o n
other surfaces, such as aluminium oxide, zirconium
oxide, and calcium phosphate. While growth on Ti
The interactions occurring at the bone-biomaterial
interface largely determine the success or failure of a
device. The outcome at this site is dependent not only
o n successful wound healing, b u t also on successful
bone formation. Bone formation o n a n implant surface
requires recruitment of osteoblast precursor cells, their
differentiation into secretory osteoblasts, production
of unmineralized extracellular matrix (osteoid), and
calcification of the extracellular matrix.' This process
is highly regulated and requires complex interplay between a wide variety of cells.
*To whom correspondence should be addressed at the
Department of Orthopaedics, The University of Texas Health
Science Center at San Antonio, 7703 Floyd Curl Drive, San
Antonio, TX 78284-7774; e-mail: MESSIERButhscsa.edu.
John Wiley & Sons, Inc.
Journal of Biomedical Materials Research, Vol. 32, 55-63 (1996)
0 1996 John Wiley & Sons, Inc.
CCC 0021-9304/96/010055-09
KIESWETTER ET AL.
56
may be more conducive to extracellular matrix synthesis than growth on other surfaces, it is clear that there
can be considerable variation in cellular response to
Ti as well, depending on the state of maturation of the
chondrocytes.
Phenotypic expression of osteoblasts is sensitive to
the topography and topology of Ti surfaces. Martin et
al." showed that human osteoblast-like MG-63 cells
grown on Ti disks with different surface roughnesses
exhibited differential responses with respect to cell
morphology, cell proliferation, alkaline phosphatase
specific activity, RNA synthesis, and protein and proteoglycan production. In general, on the three rougher
surfaces there were lower cell numbers, decreased rates
of cellular proliferation, and increased matrix production in comparison with the two smoother surfaces
and polystyrene controls.
Differences in the cytokine and growth factor profile
present at the bone-implant interface may also result
in differences in the quality, extent, and rate of bone
formation. Two local factors produced by osteoblasts
that are important in both wound healing and bone
formation are prostaglandin E2 (PGE,) and transformAt lower concening growth factor PI (TGF-&).1,3,4,12-15
trations PGE2enhances osteoblastic activity,but at high
concentrations it inhibits osteblasts and stimulates osteocla~ts.~,'~
PGE2has been proposed as having a role
in the response of osteoblasts to 1,25-(OH)2D3,acting
as a second messenger of the nongenomic action of
1,25-(OH)2D3
on the cell membrane.I6
The effects of TGF-PI are wide-ranging and vary
according to cell type.17Bone formation appears to be
strongly related to the presence of TGF-PI in the bone
matrix." TGF-& is believed to be important in osteoblast proliferation, differentiation, and matrix production necessary for bone formation.' In vivo, TGF-PI
enhances the osteoinductive properties of bone morphogenic proteins.18When TGF-PI is injected subperiosteally, it promotes bone and cartilage formati~n.'~
In MG-63 cultures, TGF-PI has been shown to elicit
dose-dependent increases in alkaline phosphatase, an
enzyme indicative of differentiated osteoblasts. Synergistic increases in alkaline phosphate levels occur when
TGF-PI and 1,25-(0H),D3are simultaneously added to
c ~ l t u r e sTGF-P,
.~
has also been shown to inhibit 1,25(OH),D3-dependent production of oste~calcin.~
These
results indicate that while TGF-0, plays a role in the
initial stages of bone formation, it may actually inhibit
the later stages associated with mineralization.
We propose that the milieu of cytokines and growth
factors surrounding a device in bone, including those
produced by the osteoblasts themselves,may be modulated by the characteristics of the implant. The characteristics of Ti surfaces currently implanted vary with
respect to surface texture and roughness. The aim of
these studies was to determine whether surface rough-
ness affects the production of two local factors, PGE2
and TGF-P1, associated with the regulation of bone
formation by osteoblast-like cells. In an effort to attain
some insight into the differential response of osteoblasts to surface roughness, these studies were conducted on the Ti disks with the same five surface
roughnesses previously employed to study osteoblast
proliferation, differentiation, and matrix production."
MATERIALS AND METHODS
Titanium surfaces
Cells were cultured on Ti disks (15 mm in diameter)
prepared from '1-mm-thick sheets of grade 2 unalloyed
Ti (American Society for Testing Materials no. 67) by
Institut Straumann AG (Waldenburg, Switzerland).
The disks were processed according to one of five different treatment regimes. Briefly, all disks were pretreated with hydrofluoric acid-nitric acid and washed
(PT). PT disks were also electropolished (EP), fine
sand-blasted, etched with HC1 and H2S04,and washed
(FA);coarse sand-blasted, etched with HC1 and H2S04,
and washed (CA); or Ti plasma-sprayed (TI'S). The
different treatments resulted in distinct differences in
surface appearance when examined by light microscopy. Disks in the El' group were very smooth and
regular in appearance, while those in the PT group
had a similar appearance but the surface was rougher.
FA disks had an appearance which was similar to that
of the untreated side. Disks in the CA group had surfaces that were very rough but regular in appearance,
while those in the TPS groups were extremely rough
and irregular in appearance. Scanning electron microscopy (SEM) confirmed the light microscopic observations. However, PT and FA were found to be similar in
roughness (seebelow and Martin et al.I1),and scanning
electron microscopic analysis showed the materials to
differ considerably with respect to morphology. Laser
confocal scanning microscopy was used to characterize
the roughness of the disks, energy-dispersion X-ray
analysis was used to confirm that the disks were essentially 99.5% pure Ti, and Auger electron spectroscopy
was used to determine the elemental composition of
the disks. Taken together, the analyses indicated that
the rank order of material surface roughness, from
smoothest to roughest, was El', PT, FA, CA, and TPS.
The materials and their surface analysis are discussed
in greater detail in Martin et al.
Cell culture
MG-63 osteoblast-like cells (American Type Culture
Collection no. CRL-1427) were used for these experi-
SURFACES MODULATE LOCAL FACTOR PRODUCTION
57
ments. This well-characterized cell line was originally
isolated from a human osteosarcoma and has been
shown to exhibit numerous traits characteristic of osteoblasts. Among these traits are high levels of alkaline
phosphatase production and inhibition in cell number
in response to stimulation by 1,25-(OH)2D3.19*20
The cells were seeded onto the Ti disks in 24-well
plates (CorningGlass Works, Corning, NY) at a plating
density of 9300 cells/cm2 in Dulbecco’s modified Eagle’s medium (DMEM; Gibco BRL, Grand Island, NY)
containing 10% fetal bovine serum (FBS) and 1%
penicillin-streptomycin. When placed in the 24-well
plates, the Ti disks entirely covered the bottom of the
wells. Because it was impossible to determine the extent of confluence of cells grown on the Ti disks, confluence was determined by cell growth on the culture
plate plastic surface. Each experiment contained six
cultures on each of the five Ti surfaces and six plastic
control cultures. Cultures were maintained in a 5%
C02 atmosphere at 37°C and 100% humidity. Media
were changed 24 h after plating and then at 72-h intervals. Fresh media with 10% FBS and 1%antibiotics
were added to the cultures at confluence. Cells were
harvested 24 h later.
Conditioned cell culture media obtained at harvest
were equally divided into two aliquots, one for PGE2
analysis and one for TGF-PI analysis. Cell culture media for PGE2analysis were acidified to pH 3.5 by the
addition of 0.5M HC1 to prevent PGE2degradation. All
samples were stored at -20°C.
used in preliminary studies showed that following two
trypsinizations, all cells and matrix were removed from
the smooth surfaces (data not shown). Cell suspensions
from both trypsinizations were combined and centrifuged at 500 X g for 10 min. The supernatant was
decanted, and the cell pellet was washed with PBS and
resuspended in physiologic saline. Cell number was
determined using a Coulter Counter (Hialeah, FL).
Cell morphology
To obtain a quantitative assessment of growth factor
production, the levels of total TGF-PI in the conditioned media were determined using a commercially
available enzyme-linked immunosorbent assay kit
(Promega Corp. Madison, WI) specific for human TGFPI. Immediately prior to the assay, the media were
acidified by the addition of 1M HC1 for 15 min, followed by neutralization with the addition of 1M
NaOH. The assay was conducted by placing the treated
samples in microtiter plates coated with monoclonal
antibody to TGF-PI for 90 min. After the unbound
proteins were removed, the wells were incubated
with a polyclonal rabbit anti-TGF-P1 antibody for
2 h, washed, and then incubated with a horseradish
peroxidase-conjugated anti-rabbit immunoglobulin G
antibody for 2 h. This step was followed by a wash
and color development. Color development was
stopped by the addition of 1M phosphoric acid. Intensity measurements were conducted at 450 nm using a
BioRad Model 2550 EIA reader (Hercules,CA).Sample
concentrations were determined by comparing the absorbance value to a known standard curve. The amount
of TGF-P1in the cell layer was not examined because
of difficulties associated with quantitatively extracting
this cytokine from the matrix.
To determine whether cell morphology varied as a
function of surface roughness, the cultures were examined by SEM. At harvest, the culture media were
removed. Samples were rinsed three times with
phosphate-buffered saline (PBS) and fixed with 1%
Os04 in O.1M PBS for 15-30 min. After fixation, the
disks were rinsed with PBS, sequentially incubated for
30-45 min in 50, 75, 90 and 100% tertbutyl alcohol,
and vacuum-dried. A thin layer of palladium-gold
was sputter-coated onto the samples prior to examination in a JEOL 6400 FEC cold field emission SEM
UEOL USA, Inc., Peabody, MA).
Cell number
At harvest, cells were released from the culture surface by the addition of 0.25% trypsin in Hank‘s balanced salt solution (HBSS)containing l mM ethylenediamine tetraacetic acid (EDTA) for 10 min at 37°C.
The reaction was terminated by the addition of DMEM
containing 10% FBS. A second trypsinization was performed to ensure that any remaining cells had been
removed from the surface. SEM examination of disks
PGE,
The level of PGE2 in the media was assessed using
a commerciallyavailable competitive binding radioimmunoassay (NEN Research Products, Boston, MA). In
this assay, unlabeled PGE2in the sample was incubated
overnight with radiolabeled PGEzand unlabeled antiPGE2 antibody. Antigen-antibody complexes were
separated from free antigen by precipitation with polyethylene glycol. Sample PGE2concentrations were determined by correlating the percentage bound over
unbound counts to a standard curve.12Previous studies
have shown that the majority of PGE2synthesized by
osteoblast-like cells is immediately released into the
media rather than stored in the cell layer.I2r2’For this
reason, we did not examine the PGE, content of the
cell layer.
TGF-PI
KIESWETTER ET AL.
58
Statistical analysis
Experiments were conducted at least twice, and the
data shown are from one experiment. The values given
are the mean t SEM of six individual cultures. Data
were first analyzed by analysis of variance (ANOVA);
when differences were detected, a Student’s t test for
multiple comparisons using Bonferroni’s modification
was used. Differences were considered significant at p
< 0.05.
RESULTS
Cell morphology
The appearance of the cells on the surfaces varied
with surface roughness (Fig. 1).The cells grown on
the smooth EP surface were well spread and appeared
to form a relatively thin, continuous monolayer [Fig.
l(A)]. Cells grown on PT disks [Fig. l(B)] had a morphology similar to those on EP, but exhibited greater
ruffling on their surface. In addition, they appeared to
form multiple layers on the surface. The cells grown
on the FA, CA, and TI’S disks [Fig. 1(C-E)] did not
cover the entire surface and were more multilayered
in nature than those on EP and PT. Many cells on the
three rougher surfaces had extensions that covered
distances in excess of 10 pm. Thin, long extensions
resulted in cells that were elongated in appearance on
the FA and CA surfaces. Similar morphologies were
also evident on TPS. In general, however, the cells on
the TI’S surface had more focal attachments, causing
them to appear to be more spread out, thereby covering
larger areas than those cells on FA and CA.
Cell number
Cell number was affected by surface treatment. In
general, cell number was inversely related to surface
roughness. The smoothest surface, EP, had significantly higher cell counts than did the plastic controls
(Table I). While CA and TI‘S both had cell counts that
were lower than plastic, the differences were only significant for CA. No differences in cell number were
evident between the PT, FA, TPS and plastic groups.
PGE2than the plastic controls (FA: 1.8X; CA: 3.9X; TPS:
3.3X). Differences among the three roughest surfaces
were also noticeable. Total PGE2 production by cells
grown on the CA and TPS surfaces was 1.8 and 2 times
greater, respectively, than that of cells grown on the
FA surfaces.
When the amount of PGE2production was normalized to cell number (Fig. 3),no differences were evident
between the EP, PT, and plastic surfaces. In contrast,
the level of PGE2produced by cells growr on FA, CA,
and TPS was 3.4, 7.6, and 5.9 times, respectively, that
produced by cells grown on plastic. The relationship
between the FA, CA, and TPS surfaces was essentially
the same on either a per-cell or per-well basis. In both
cases, the CA and TI‘S cultures had significantly
greater PGE, levels than the FA cultures.
TGF-PI
The level of TGF-PI in the conditioned media was
also related to surface roughness. Total TGF-PI in conditioned media from cultures grown on EP and PT
surfaces was slightly, but not significantly, greater than
that seen in the plastic control cultures, whether the
data were calculated per well (Fig. 4) or per cell (Fig.
5). However, total TGF-PI production by MG-63 cells
cultured on FA, CA, and TI’S surfaces was approximately 1.5 times that of plastic control cultures (Fig.
4). When assessed as a function of cell number, cells
on FA surfaces exhibited higher TGF-PI production
than did the MG-63 cells on plastic, but this was not
significant. In contrast, TGF-PI produced by cells
grown on CA and TPS was 3.0 and 4.2 times that of
cells grown on plastic.
Normalization of the total TGF-PI levels to account
for cell number altered the relationship between FA,
CA, and TPS surfaces. On a per-well basis, the three
materials were essentially equivalent. However, on a
per-cell basis, the level of TGF-P1 produced by cells
grown on CA and TI‘S surfaces was 2.2 and 3.0 times,
respectively, that of cells grown on FA. Moreover, the
differences between CA and TPS became more pronounced. On a per-cell basis, TGF-P, production on
TPS was 1.4 times that on CA.
DISCUSSION
PGE,
The level of PGE, production by the cells was affected by surface roughness (Fig. 2). Total PGE2 production by cells grown on the EP and PT surfaces was
slightly, but not significantly, higher than that of cells
grown on plastic (Fig. 2). The conditioned media from
the rougher surfaces had significantly greater levels of
Although material properties such as surface energy,
composition, roughness, and topography2’are believed
to be of critical importance at the implant-tissue interface, the subsequent steps in bone healing around the
implant may depend more on the cells at the surface
than on the surface itself. The process of bone formation and regulation involves a variety of hormones,
SURFACES MODULATE LOCAL FACTOR PRODUCTION
59
Figure 1. Scanning electron micrographs of MG-63 cells cultured on Ti with one of the five following surfaces, ranked from
smoothest to roughest: electropolished (A), pretreated surface (B), fine grit-blasted (C), coarse grit-blasted (D), and Ti plasmasprayed (E). The bar indicates 10 pm.
cytokines, and growth factors. Many of these factors
are produced by the osteoblast and act in either an
autocrine or paracrine manner. The osteoblasts are
therefore potentially capable of regulating the local
response to the biomaterial surface. The studies reported here indicate that Ti surface roughness affects
the production of local growth factors and cytokines
by osteoblasts, and that enhanced production on the
rougher surfaces may be related to the more differentiated cellular morphology on these substrates.
The appearance of the cells on the FA, CA, and TPS
surfaces differs considerably from those on the
smoother EP and PT surfaces. The decreased levels of
proliferation, as assessed by [3H]thymidineincorporation of cells grown on the FA, CA, and TPS seen by
Martinet al.," in conjunction with the distinct morphology and lower cell numbers on these surfaces, suggest
that relative to their counterparts on the EP and PT
surfaces, the cells on the rougher surfaces are at a more
advanced stage of development and that this more
differentiated state was influenced by the surface. Lian
and Stein2,23,24
hypothesized that expression of a differentiated osteoblast phenotype in culture is preceded
by a decrease in proliferation.
KIESWETTER ET AL.
60
TABLE I
Effect of Surface Roughness on MG-63 Cell Number
No. of Cells
Surface
Plastic
EP
PT
FA
CA
TPS
X
lo5
Effect of Titanium Disk Surface
on PGE, Production
___
._
#
*
2.66 5 0.09
4.00 2 0.08*
3.15 2 0.65
3.11 i 0.12
1.73 5 0.15*
1.94 i 0.33
T
MG-63 cells were cultured on plastic, or Ti with one of the
five following surfaces, ranked from smoothest to roughest:
electropolished (EP), pretreated surface (PT), fine gritblasted (FA), coarse grit-blasted (CA), and Ti plasmasprayed (TPS). Data are from one of two replicate experiments, yielding similar results, and are expressed as mean
? SEM, n = 6 cultures.
* p < 0.05, Ti surfaces vs. plastic.
Plastic
EP
PT
FA
CA
TPS
Surface Treatment
The enhanced production of local factors by cells on
the rougher surfaces (FA, CA, and TPS) supports the
hypothesis that these cells are more differentiated than
those on the smooth surfaces. Although a smaller number of cells is present on these rougher surfaces, the
total levels of PGE2 and TGF-P1 are greater. The increased activity of these cells is clearly evident when
PGE2 and TFG-PI production are normalized to cell
number. Because the total levels of PGEz and TGF-0,
are greater and cell numbers smaller in cultures grown
on rougher surfaces, the amount of PGE2and TGF-P,
produced per cell is greater on the rougher surfaces
than on the smoother surfaces. Validity of the cell count
Effect of Titanium Disk Surface
on PGE, Production
100
__
~~
Figure 3. Prostaglandin E2 (PGE?)production per lo5 cells
cultured on tissue culture plastic, or Ti with one of the five
following surfaces, ranked from smoothest to roughest: electropolished (EP), pretreated surface (PT), fine grit-blasted
(FA), coarse grit-blasted (CA), and Ti plasma-sprayed (TPS).
Data are from one of two replicate experiments, each yielding
similar results, and are expressed as mean 2 SEM, n = 6
cultures. *p < 0.05, Ti surfaces vs. plastic; ' p < 0.05 vs. FA.
data was verified by examining the surfaces with SEM
during preliminary studies. Because cells were no
longer present on the surfaces after two trypsinizations, the possibility that the cell numbers for the
rougher surfaces were artificially low was eliminated.
Effect of Titanium Disk Surface
on TGF-O, Production
- - ~-
~~
#
*
-
- ~-
*
*
T
FA
CA
*
#
80
*
h
0,
60
m
Q
v
w"
40
(3
a
20
0
Plastic
EP
PT
FA
CA
TPS
Surf ace Treat ment
Figure 2. Prostaglandin Ez (PGE2) production by MG-63
cells cultured on tissue culture plastic, or Ti with one of the
five following surfaces, ranked from smoothest to roughest:
electropolished (EP), pretreated surface (PT), fine gritblasted (FA), coarse grit-blasted (CA), and Ti plasmasprayed (TPS). Data are from one of two replicate experiments, each yielding similar results, and are expressed as
mean ? SEM, n = 6 culture. *p < 0.05, Ti surfaces vs. plastic;
p < 0.05 vs. FA.
Plasttc
EP
PT
IPS
Surf ace Treatment
Figure 4. Transforming growth factor PI (TGF-PI)production by MG-63 cells cultured on tissue culture plastic, or Ti
with one of the five following surfaces, ranked from smoothest to roughest: electroyolished (EP), pretreated surface (PT),
fine grit-blasted (FA), coarse grit-blasted (CA), and Ti
plasma-sprayed (TPS). Data are from one of two replicate
experiments, each yielding similar results, and are expressed
as mean 4 SEM, 11 = 6 cultures. *p < 0.05, Ti surfaces vs.
plastic.
SURFACES MODULATE LOCAL FACTOR PRODUCTION
Effect of Titanium Disk Surface
on TGF-O, Production
#
*
:
c
~~~~~
0
I - 0
Plastic
EP
PT
FA
CA
TPS
Surf ace Treatment
Figure 5. Transforming growth factor PI (TGF-PI) produc
tion per lo5cells cultured on tissue culture plastic, or Ti wit1
one of the five following surfaces, ranked from smoothest
to roughest: electropolished (El'), pretreated surface (PT),
fine grit-blasted (FA), coarse grit-blasted (CA), and Ti
plasma-sprayed (TI'S). Data are from one of the two replicate
experiments, each yielding similar results, and are expressed
as mean ? SEM, n = 6 cultures. * p < 0.05, Ti surfaces vs.
plastic; #p < 0.05 vs. FA.
The exact role of PGE2 in bone formation is still
under investigation. It has been shown to have both
inhibitory and stimulatory effects in z ~ i t r o . ~ Recent
,'~,~~
studies indicate that local PGE2 concentrations may
affect the replication and differentiation of osteoblast
precursor^.^ At lower concentrations, osteoblasts respond to the addition of PGE2 by an increase in
t3H]thymidine incorporation at 24 h and an increase
in collagen synthesis at later times. In the presence of
cortisol
PGE2 increased the extent of [3H]proline incorporation into collagenase-digestible and
noncollagenase-digestible protein. At higher concentrations, radiolabel incorporation was inhibited in a
dose-dependent manner.
The interactions between PGE2and other hormones
have also been believed to play a role in the differentiation of o~teoblasts.~
In addition, PGE2 has been proposed to be a second messenger, mediating the effects
of other regulatory agents. For example, some of the
effects of 1,25-(OH)2D3on osteoblasts have been hypothesized to occur via autocrine or paracrine action
of PGE2 on the cell membranes.16 This autocrineparacrine interaction may serve to stimulate the production of cytokines, such as insulin-like growth
associated with cellular proliferation and differentiation. In the present study, the autocrine effect
of PGE2 production on the cells was not examined
but would be expected to have a significant effect on
cell behavior.
Sporn et al?7suggested that the interactionsbetween
TGF-P1and other mediators in the media may serve to
61
modulate the effects of the growth factor. For example,
TGF-P, induces osteoblasts to produce extracellular
matrix and increases alkaline phosphatase a ~ t i v i t y . ~ , ~ ~
Similarly, 1,25-(OH)2D3stimulates the differentiation
of osteoblast-like
When 1,25-(OH),D3is combined with TGF-P1, a synergistic increase in alkaline
phosphatase activity and collagen synthesis occurs, but
1,25-(OH)2D3-dependentincreases in osteocalcin production, an indicator of terminal differentiation of osteoblasts, are id~ibited.~,~
The observation that 1,25(OH)2D3regulates PGE2 production by osteoblasts'6
suggests that there may be an interrelationship between PGE2and TGF-PI as well.
The effects of surface roughness on the production
of local factors is not straightforward. The results of
the present study indicate that production of cytokines
and growth factors by osteoblasts as a function of surface roughness varies for each agent examined. Normalization of the PGEpand TGF-PI data to cell number
showed the production on FA surfaces to be elevated
with respect to EP and PT surfaces for PGE2,but not
for TGF-P,. The differences in the local concentration of
the individual factors may account for the similarities
between FA and the two smoother surfaces with respect to cell proliferation and differentiation, but not
matrix production noted by Martin et al."
The concentration of PGE2produced on these surfaces is slightly lower than those associated with stimulation of osteoblasts in v i t r ~ .While
~ ~ , ~this
~ regulatory
factor alone may not enhance osteogenic activity, together with other factors produced by the cells, it may
stimulate osteoblastic differentiation. In contrast, the
level of TGF-PI produced by the MG-63 cells on the
different Ti surfaces was significantly higher than that
found in conditioned media from cells grown on plastic
and was comparable to those associated with stimulation of osteoblasts in z ~ i t r o .The
~ , ~increased
~
local levels
of TGF-PI could account for the increased matrix production and enhanced cellular differentiationobserved
by Martin et a1.I1 for MG-63 cells on the same Ti surfaces.
Topography, as well as surface roughness, influenced cellular response. In previous studies," the average profile height of the PT and FA surfaces, as determined by laser confocal scanning microscopy, were
statistically indistinguishable; SEM examination of the
surfaces, however, showed considerable differences in
their topography. The PT surface was characterized
as a smooth, undulating surface with distinct grain
boundaries. The FA surface appeared to be uniformly
covered with pits 1-2 pm in appearance. In short, the
texture of the PT surface closely resembled that of the
EP surface, whereas the FA surface texture was similar
to that of CA. The cellular response to these two surfaces appears to differ considerably. The results of the
present study indicate that the response of cells grown
on the PT surfaces is very similar to those grown on
KIESWETTER ET AL.
62
EP surfaces with respect to cell morphology, as well
as total local factor production. With the exception of
the TGF-PI levels per cell, the response of the cells to
the FA material consistently falls in the same group as
the rougher CA and TPS surfaces. This suggests that
at a particular surface roughness, PGE2production on
a per-cell basis is more sensitive to topography than
is TGF-PI production.
It is possible that some of the differences in cellular
response to surface roughness may have been due to
variations in the degree of confluence that was attained. Because of the opacity of the Ti disks, confluence was approximated using cell behavior on polystyrene (a smooth plastic surface) as the indicator. The
rough-surfaced disks presented a greater surface area,
and consequently, cells may still have been preconfluent with respect to the entire disk at the time they were
examined. It is clear, however, that the cells formed
multilayered foci on the rougher disks, in effect simulating a postconfluent situation. Previous studies in
our lab have shown that preconfluent cells exhibit a
different phenotype when compared to confluent cells,
including response to regulatory factor^.'^ One can
vary seeding density to modulate the state of confluence at a given time; in the present study, we elected
not to consider this variable and chose a seeding density that had demonstrated differential responsiveness
to surface roughness for a number of parameters.” It
should be noted that the cells cultured on the rougher
surfaces did not exhibit the phenotype of preconfluent
MG-63 cells cultured on
indicating that
we assessed effects of surface roughness and not just
cell density in the present study.
Titanium surface roughness has been shown to affect
production by osteoblasts of local factors involved in
bone formation, suggesting that the complement of
autocrine and paracrine factors produced by cells at the
bone-biomaterial interface can be directed by altering
implant surface roughness; this, in effect, could thereby
direct the type of interface that forms at the implant
site. It is unlikely that effects on PGE2 and TGF-Pl
production alone can account for differences in bone
formation observed in vim adjacent to rough or smooth
surface^.^^,^^ However, to understand the mechanisms
of bone formation at the implant surface, the effects of
the material on the surrounding cells and on the profile
of cytokines, growth factors, and other local mediators
must be understood.
The authors thank Monica Luna, Maricella Alderete, and
Ruben Gomez for their expert technical assistance, and Sandra Messier for her assistance in the preparation of the manuscript. The research was supported by NSF EEC 9209612 and
the Center for the Enhancement of the Biology/Biomaterials
Interface, as well as by NIH Grants DE-05937 and DE-08603.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
References
1. S. I. Dworetzky, E. G. Fey, S. Penman, J. B. Lian, J. L.
Stein, and G. S. Stein, “Progressive changes in the pro-
16.
tein composition of the nuclear matrix during rat osteoblast differentiation,” Proc. Natl. Acud. Sci. 87, 46054609 (1990).
G. S. Stein, J. B. Lian, and T. A. Owen, “Relationship
of cell growth to the regulation of tissue-specific gene
expression during osteoblast differentiation,” F A S E B J.,
4, 3111-3123 (1990).
L. G. Raisz and P. M. Fall, ”Biphasic effects of prostaglandin E, on bone formation in cultured fetal rat calvariae: Interaction with cortisol,” Endocrinology, 126,16541659 (1990).
L. F. Bonewald, M. B. Kester, Z . Schwartz, L. D. Swain,
A. Khare, T. L. Johnson, R. J. Leach, and B. D. Boyan,
“Effects of combining transforming growth factor /3 and
1,25-dihydroxyvitamin D3on differentiation of a human
osteosarcoma (MG-63),” J. Biol. Chem., 267, 8943-8949
(1992).
R. T. Ingram, S. K. Bonde, B. L. Riggs, and L. A. Fitzpatrick, ”Effects of transforming growth factor beta
(TGFP) and 1,25 dihydroxyvitamin D3 on the function,
cytochemistry, and morphology of normal human
osteoblast-like cells,” Diflerentiution, 55, 153-163 (1994).
H. Nakahara, J. E. Dennis, S. P. Bruder, S. E. Haynesworth, D. P. Lennon, and A. I. Caplan, ”In vitro differentiation of bone and hypertrophic cartilage from periosteal-derived cells,” Exp. Cell Res., 195, 492-503 (1991).
A. E. Grigoriadis, J. N. M. Heersche, and J. E. Aubin,
“Differentiation of muscle, fat, cartilage, and bone from
progenitor cells present in a bone-derived clonal cell
population: Effect of dexamethasone,” J. Cell. Biol., 106,
2139-2151 (1988).
A. E. Grigoriadis, J. E. Aubin, and J. N. M. Heersche,
”Effects of dexamethasone and vitamin D3on cartilage
differentiation in a clonal chondrogenic cell population,” Endocrinology, 125, 2103-2110 (1989).
J. C. Keller and E. P. Lautenschlager, ”Metals and
alloys,“ in Handbook of Biomaterials Evaluation: Scientific,
Technical, and Clinical Testing of Implantable Materials,
A. F. von Iiecum (ed.), Macmillan, New York, 1986,
pp. 3-21.
J. Hambleton, Z . Schwartz, A. Khare, S. W. Windeler,
M. Luna, B. P. Brooks, D. D. Dean, and B. D. Boyan,
“Culture surfaces coated withvarious implant materials
affect chondrocyte growth and metabolism,” J. Ortkop.
Res., 12, 542-552 (1994).
J.Y. Martin, Z. Schwartz, T. W. Hummert, D.M.
Schraub, J. Simpson, J. Lankford, Jr., D. D. Dean, D. L.
Cochran, and B. D. Boyan, “Effect of titanium surface
roughness on proliferation, differentiation, and protein
synthesis of human osteoblast-like cells (MG63),” J. Biomed. Muter. Res., 29, 389-401 (1995).
Z. Schwartz, L. D. Swain, D. W. Kelley, B. Brooks, and
B. D. Boyan, ”Regulation of prostaglandin E2 production by vitamin D metabolites in growth zone and resting zone chondrocyte cultures is dependent on cell maturation,” Bone, 13, 395-401 (1992).
L. F. Bonewald, Z. Schwartz, L. D. Swain, and B. D.
Boyan, ”Stimulation of matrix vesicle enzyme activity
in osteoblast-like cells by 1,25(OH),D, and transforming
growth factor P (TGFP),” BoneMiner. 17,139-144 (1992).
M. E. Joyce, A. 8. Roberts, M. B. Sporn, and M. E. Bolander, “Transforming growth factor-p and the initiation of chondrogenesis and osteogenesis in the rat femur,” J. Cell Biol., 110, 2195-2207 (1990).
Y . S. Chyun and L. G. Raisz, ”Stimulation of bone formation by prostaglandin E,,“ Prostaglandins, 27, 97103 (1984).
Z . Schwartz, R. Dennis, L. F. Bonewald, L. Swain, R.
Gomez, and B. D. Boyan, ”Differential regulation of
63
SURFACES MODULATE LOCAL FACTOR PRODUCTION
17.
18.
19.
20.
21.
22.
23.
24.
25.
prostaglandin E, synthesis and phospholipase A, activity by 1,25-(OH),D3in three osteoblast-like cell lines
(MC-3T3-E1, ROS 17/28, and MG-63), Bone, 13, 51-58
(1992).
L. F. Bonewald and G. R. Mundy, ”Role of transforming
growth factor-beta in bone remodeling,“ Clin. Ortkop.
Rel. Res., 250, 261-276 (1990).
H. Bentz, A.Y. Thompson, R. Armstrong, R. J. Chang,
K. A. Piez, and D. M. Rosen, “Transforming growth
factor-beta enhances the osteoinductive activity of a bovine bone-derived fraction containing bone morphogenetic protein 2 and 3,” Matrix, 11, 269-275 (1991).
B. Boyan, Z. Schwartz, L. F. Bonewald, and L. D. Swain,
”Localization of 1,25-(OH)2D,-responsivealkaline phosphatase in osteoblast-like cells (ROS 17/2.8, MG 63, and
MC 3T3) and growth cartilage cells in culture,” 1. Biol.
Ckem., 264, 11879-11886 (1989).
R.T. Franceschi, W.M. James, and G. Zerlauth,
“l,Alpha,2S-dihydroxyvitaminD3specific regulation of
growth, morphology, and fibronectin in a human osteosarcoma cell line,” J. Cell. Phys., 123, 401-409 (1985).
R. Dziak, “Prostaglandins as mediators of bone cell metabolism,‘‘ in Calcium in Biological Systems, R. P. Rubin,
G. B. Weiss, J. W. Putney (eds.), Plenum, New York,
1985, pp. 533-539.
Z. Schwartz and B. D. Boyan, ”Underlying mechanisms
at the bone-biomaterial interface,” J. Cell. Biockem., 56,
340-347 (1994).
J. B. Lian and G. S. Stein, “The developmental stages of
osteoblast growth and differentiation exhibit selective
responses of genes to growth factors (TGFP1) and hormones (vitamin D and glucocorticoids),” J. Oral I7nplantol., 19, 95-105 (1993).
J. B. Lian and G. S. Stein, ”Concepts of osteoblast growth
and differentiation: Basis for modulation of bone cell
development and tissue formation,” Crit. Rev.Oral Biol.
Med., 3, 269-305 (1992).
L. G. Raisz and A. R. Kooleman-Beyney, ”Inhibition of
bone collagen synthesis by prostaglandin E2 in organ
culture,” Prostaglandins, 8, 377-385 (1974).
26. T. L. McCarthy, M. Centrella, L. G. Raisz, and E. Canalis,
“Prostaglandin E2stimulates insulin-like growth factor
synthesis in osteoblast-enriched cultures from fetal rat
bone,“ Endocrinology, 128, 2895-2900 (1991).
27. M. B. Sporn, A. B. Roberts, L. M. Wakefield, and R. K.
Assoian, “Transforming growth factor+: Biological
function and chemical structure,” Science, 233, 532534 (1986).
28. S. E. Harris, L. F. Bonewald, M. A. Harris, M. Sabatini,
S. Dallas, J. Q. Feng, N. Ghosh-Choudhury, J. Wozney,
and G. R. Mundy, “Effects of transforming growth factor p on bone nodule formation and expression of bone
morphogenetic protein 2, osteocalcin, osteopontin, alkaline phosphatase, and type I collagen mRNA in long
term fetal rat calvarial osteoblasts,” 1. Bone Miner. Res.,
9, 855-863 (1994).
29. D. Farr, W. Pochal, M. Brown, E. Shapiro, N. Weinfeld,
and R. Dziak, “Effects of prostaglandins on rat calvarial
bone-cell calcium,” Arch. Oral Biol., 29, 885-891 (1984).
30. Z . Schwartz, L. F. Bonewald, K. Caulfield, B. Brooks,
and B. D. Boyan, ”Direct effects of transforming growth
factor+ on chondrocytes are modulated by vitamin D
metabolites in a cell maturation-dependent manner,”
Endocrinology, 132, 1544-1552 (1993).
31. B.D. Boyan, T. W. Hummert, D.D. Dean, and Z.
Schwartz, ”The role of material surfaces in regulating
bone and cartilage cell response,” Biomaterials, 17,137146 (1996).
32. K. Thomas and S. Cook, ”An evaluation of variables
influencing implant fixation and direct bone apposiBiomed. Muter. Res., 19, 875-901 (1985).
tion,” I.
33. D. Buser, R. Schenk, S. Steinemann, J. Fiorellini, C. Fox,
and H. Stich, “Influence of surface characteristics on
bone integration of titanium implants: A histomorphometric study in miniature pigs,” J. Biomed. Muter. Res.,
25,889-902 (1991).
Received May 9, 1995
Accepted November 13, 1995