Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

Bone-cell material interaction on Si microchannels with bioinert coatings by bholeynathji


									                                         Acta Biomaterialia 3 (2007) 523–530

               Bone cell–materials interaction on Si microchannels with
                                   bioinert coatings
                           Russell Condie              , Susmita Bose a, Amit Bandyopadhyay                           a,*

                          W.M. Keck Biomedical Materials Research Laboratory, School of Mechanical and Materials Engineering,
                                            Washington State University, Pullman, WA 99164-2920, USA
                                      Department of Bioengineering, University of Utah, Salt Lake City, UT, USA

                             Received 20 April 2006; received in revised form 30 October 2006; accepted 7 November 2006


   Bone implant life is dependent upon integration of biomaterial surfaces with local osteoblasts. This investigation studied the effects of
various microchannel parameters and surface chemistry on immortalized osteoblast precursor cell (OPC1) adhesion. Cell–materials inter-
actions were observed within channels of varying length, width, tortuosity, convergence, divergence and chemistry. Si wafers were used to
create four distinct 1 cm2 designs of varying channel dimensions. After anisotropic chemical etching to a depth of 120 lm, wafers were
sputter coated with gold and titanium; and on another surface SiO2 was grown to vary the surface chemistry of these microchannels.
OPC1 cells were seeded in the central cavity of each chip before incubation in tissue culture plates. On days 5, 11 and 16, samples were
taken out, fixed and processed for microscopic analysis. Samples were visually characterized, qualitatively scored and analyzed. Channel
walls did not contain OPC1 migration, but showed locally interrupted adhesion. Scores for channels of floor widths as narrow as 350 lm
were significantly reduced. No statistically significant preference was detected for gold, titanium or SiO2 surfaces. Bands of OPC1 cells
appeared to align with nearby channels, suggesting that cell morphology may be controlled by topography of the design to improve
Ó 2006 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Osteoblast adhesion; Osseointegration; Microchannel; Bone tissue engineering

1. Introduction                                                                causing eventual wear, inflammation and failure [4]. How-
                                                                               ever, engineered biomaterial surfaces may improve implant
   Musculoskeletal disorders and bone deficiencies have                         stability by promoting cell adhesion, proliferation and dif-
been established as being among the most important                             ferentiation. Various methods of improving osseointegra-
human health conditions that exist today, costing more                         tion have been attempted. Meyer et al. identified several
than $16 billion in products in 2006, and afflicting one in                      material physico-chemical properties that promote osteo-
seven Americans [1]. The success and lifetime of an implant                    blast adhesion [5]. Dense, inert materials, such as Au, Ti,
is largely dependent on the degree of osseointegration at                      SiO2 and steels, have relatively low bioactivity and rely pri-
the material–bone interface, especially in load-bearing                        marily on morphological fixation or press fitting to bone
orthopedic or dental applications [2,3]. A layer of fibrous                     tissue [6–8]. Biological fixation in porous materials and
or soft tissues can accumulate at the surfaces of implant                      chemical fixation in bioactive materials reduce soft-tissue
materials [3]. This may result in a modulus mismatch and                       encapsulation [3]. Resorbable ceramics may further
allow micromotion to disrupt integration with local bone,                      improve osseointegration if the rate of dissolution can be
                                                                               matched to that of tissue formation [4].
     Corresponding author. Tel.: +1 509 335 4862.                                  Nano- and micro-scale topography has been shown to
     E-mail address: (A. Bandyopadhyay).                      influence cell–materials interactions [2,4]. Titanium sur-

1742-7061/$ - see front matter Ó 2006 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
524                                    R. Condie et al. / Acta Biomaterialia 3 (2007) 523–530

faces textured with 100 lm cavities were shown to favor
                                                                              A                                                           B
osteoblast attachment and growth, while submicron-scale
etching promoted differentiation [9]. Bone tissue engineer-

                                                                                                  2.5 mm [--------]
ing on patterned collagen films coated with calcium phos-
phates and fibrinogen showed preferential osteoblast
alignment and orientation along the axis of parallel
grooves [10]. Similar experiments have shown osteoblasts
alignment, elongation and migration along parallel grooves
on other surfaces as well [11,12]. However, research on
                                                                              C                                                           D
bone cell–materials interaction along complex channels                                                                                    500 μm |-|

                                                                                                  [-------------- 1.0 cm -------------]
does not appear to have been studied well.
   Our current work investigates the hypothesis that
immortalized osteoblast precursor cells (OPC1) may prolif-
erate along microchannels with varying parameters such as
length, width, tortuosity, convergence, divergence and sur-
face chemistry. As bone cell–materials interaction is quite
complex, we have created a few simplified structures to                                                (i)
understand these interactions, which is relevant in dental
                                                                                                                                          OPC1 cells suspended in media
and orthopedic applications. Secondly, our work examines                                                                                  seeded in 120 micron deep pit
the null hypothesis of a lack of detectable preference for                 150 nm bio inert                                               with 54.74º sidewalls
OPC1 adhesion or proliferation on gold-, titanium- and sil-                coating
                                                                                                                                                       380 micron Si wafer
ica-coated surfaces. Finally, it reviews evidence that OPC1
cells may align with nearby channels, and that inclined
sidewalls inhibit OPC1 adhesion. To study these hypothe-
ses, OPC1 cells were cultured in various Si microchannels
and analyzed by scanning electron microscopy (SEM). A
semi-qualitative analysis revealed no significant preference                                       (ii)
towards materials chemistry. However, narrow channels               Fig. 1. (i) Top surfaces of four chips that are designed to test cell–
were shown to inhibit adhesion and banding patterns.                materials interactions along diverse microchannels. Design A varies
                                                                    channel length, B varies channel width, C varies convergent and divergent
2. Materials and methods                                            patterns, and D varies channel tortuosity. (ii) Cross-sectional view of the

2.1. Design and microfabrication

   To investigate OPC1 behavior along various channels,             for chip fabrication. Wafers were subjected to high temper-
1 cm2 silicon chips were etched with four designs consisting        ature wet oxidation at 1050 °C in mixed oxygen and nitro-
of several channels branching out from a central cavity.            gen environment. A 45 min ramp to 1050 °C was followed
Design A varied channel length, B varied channel width,             by 80 min soak time to grow 500 nm oxide layer on silicon.
C varied channel divergence and convergence, and D var-             Oxide layer on polished side of the wafer was then stripped
ied channel tortuosity. Chips were designed using Corel             using a buffered oxide etchant (BOE) solution of HF,
Draw and micro-machined using anisotropic Si etching,               NH4F, and H2O (10:1) for 10 min, while the back-side of
and then coated with gold, titanium, and SiO2. Each                 the wafer was protected using semiconductor tape. This
1 · 1 cm square chip has a central cavity of 2500 ·                 was followed by boron diffusion on the bare silicon side
2500 lm branching out into channels of standard width               at 1125 °C for 110 min to result in 2.3 lm boron depth.
and separation 500 lm. Fig. 1 shows all four designs.               The oxide layer on the back-side prevents boron to diffuse
Design A varied channel length from 500 to 4500 lm.                 onto silicon on this side. Boron diffusion results in borosil-
Design B varied channel width as follows: 25, 50, 75,               icate glass formation on the surface, which was stripped off
100, 200, 300, 400, 500, 600, 800 and 1000 lm. Design C             to reduce stresses in the wafer. Borosilicate glass was
had interrupted and uninterrupted convergent and diver-             removed with a 20 min etching in BOE, followed by grow-
gent channels 2500–3000 lm in length. Design D had chan-            ing a sacrificial oxide layer by low temperature oxidation
nels of standard length 6000 lm (except channel 2) and              (LTO) at 850 °C for 2 h. The sacrificial LTO layer is then
varied the channel tortuosity, or number of 90° bends,              removed in BOE for 10 min and the final LTO layer is
including 0, 1, 3, 4, 5 and 10 bends. Six copies of each of         grown, which acts as silicon dioxide windows for aniso-
the four designs were arranged on a 76.2 mm diameter Si             tropic silicon etching to fabricate silicon microchannels.
wafer.                                                              Positive photolithography using AZ400K as photoresist
   Single side polished, p-type (1 0 0) silicon wafers with         was carried out to create oxide mask on back-side of the
76.2 mm diameter and 380 ± 20 lm thickness were used                wafer. The exposed silicon was then etched away using
                                                R. Condie et al. / Acta Biomaterialia 3 (2007) 523–530                                 525

                                                                             gluteraldehyde + 0.1 M cacodylate [Na(CH3)2 AsO2Æ3H2O]
                            Silicon Wafer
                                                                             buffer at pH 7.2. Following three 10 min rinses with caco-
                                                                             dylate buffer, samples were immersed in 2% OsO4. Sample
                                                                             chips were rinsed in deionized H2O, then in 30%, 50%,
                                                                             70%, 95% and 100% ethanol, followed by 1:1 ethanol/ace-
                           Boron Diffusion
                                                                             tone, acetone and hexamethyldisilazane (HMDS). Samples
                                                                             were dried overnight, then mounted with carbon tape onto
                                                                             an SEM sample holder and sputtered with Au for 6 min to
                                                                             $200 nm in preparation for SEM. The process was repeated
                      Low temperature oxidation
                                                                             with the remaining samples on days 11 and 16. One sample
                                                                             from each duplicate (of matching design, coating and cul-
                                                                             ture time) was selected for SEM analysis. Approximately
                                                                             500 SEM images were taken, and samples were character-
                          Photolithography                                   ized and scored based on adhesion and proliferation. Scores
                                                                             were compared to show the effects of material coatings and
                                                                             channel parameters on OPC1 behavior.
                       Anisotropic Si Etching
                                                                             2.3. SEM analysis and statistical analysis

Fig. 2. Schematic of microchannel fabrication on Si using anisotropic           OPC1 cell–materials interactions were observed using a
etching.                                                                     field emission scanning electron microscope (Serion FEI,
                                                                             type 8206/02). Of the 72 chips (2 (duplicates) · 3 coat-
ethylenediamine pyrocatechol (EDP) at 110 °C for 2 h to                      ings · 4 designs · 3 time points), 36 were analyzed; the bet-
create microchannels 120 lm deep. Fig. 2 shows a sche-                       ter of each duplicate was selected. Channel beginnings,
matic of the microchannel fabrication steps. For sputter                     central sections and endpoints, as well as cavity and surface
coating, we used a DC/RF sputtering machine to deposit                       areas, were examined. Overall OPC1 conditions on each
Ti and Au on etched microchanneled Si wafers. For Au                         sample chip were described in order of increasing cell den-
layers, a 20 nm Ti was used as an adhesive layer prior to                    sity, adhesion and proliferation, as depicted in Fig. 3.
Au deposition. The thickness of these layers was estimated                   Scores were assigned as follows: bare (density score = 0),
based on established calibration data for this machine                       sparse (1), semi-confluent (2), confluent (3), banded (4)
using deposition time.                                                       and layered (5). Bad cell adhesion or peeling was assigned
                                                                             a score of À1. An average score was assigned for each chip
2.2. Cell–materials interactions                                             in cases of variation. Scores were tabulated, and a one-way
                                                                             analysis of variance (ANOVA) with a 95% confidence level
   For the cell–materials interaction study, human osteo-                    was conducted to determine whether a significant difference
blasts cells were used. Cells were derived from an immortal-                 was present among scores for each material coating.
ized osteoblastic precursor cell line (OPC1) established from
human fetal bone tissue [13]. Samples were sterilized by auto-               3. Results
claving (Amerex Instruments Ltd., CA) for 20 min at 121 °C
before cell culture. Cells were plated at a density of $20,000                  OPC1 cells proliferated over the entire floor and top sur-
cells in the central cavity of each chip and were cultured in                faces of chips, uninhibited by sidewalls. Adhesion was
McCoy’s 5A medium (with L-glutamine, without phenol                          characterized and scored for each sample and each channel
red and sodium bicarbonate). The 20,000 cells per specimen                   of Design B. No significant difference was found among
were seeded in the central cavity and allowed to flow into                    scores of Au-, Ti- or SiO2-coated surfaces. Narrow chan-
(during overflow in seeding) the channels. Then 5% fetal calf                 nels inhibited OPC1 adhesion. Bands of cells appeared to
serum and 5% bovine calf serum, 2.2 g lÀ1 sodium carbonate,                  align with nearest channels.
100 mg lÀ1 streptomycin and 8 lg mlÀ1 Fungizone (GibcoTM
Labortories, Grand Island, NY) were added in the media.                      3.1. Materials interactions
Cells were maintained at 37 °C under an atmosphere of 5%
CO2 and 95% air. Culture medium was changed every 48 h                          SEM images showed numerous OPC1 cells adhered to
for the duration of experiment. Since overflow still remains                  the material surface of 35 of the 36 chips examined. Cells
an issue for cell seeding, future experiments will focus on                  were attached to channel floors and walls, the central floor
developing an improved approach.                                             and the top surface. Material coating made no visibly
   On day 5, 24 samples (two copies of each design and                       apparent difference in OPC1 adhesion or morphology, as
coating combination) were removed, fixed, processed and                       is evident in Fig. 4. Samples cultured for 5 days generally
sputtered coated with gold in preparation for SEM. Cells                     appeared more globular and were attached to the material
were fixed for 1 h at 22 °C in 2% paraformaldehyde + 2%                       by thin extensions. Samples fixed on days 11 and 16
526                                           R. Condie et al. / Acta Biomaterialia 3 (2007) 523–530

Fig. 3. Examples of qualitative characterization of OPC1 adhesion and proliferation on various samples and corresponding scores assigned for semi-
quantitative analysis.

generally appeared to be more elongated and confluent,                      the anisotropic etching, the narrowest five channels have
and were often banded together.                                            no floor, but form a ‘‘V’’ at the bottom. As shown by
   Cell density varied from bare to layered, according to                  Fig. 6, OPC1 populations were visibly less dense within
the descriptions listed in Section 2 and in Table 1. The table             narrow channels and showed evidence of poor adhesion,
also shows material adhesion scores corresponding to each                  including cracking and peeling. Narrow channels appeared
sample. The average scores for gold, silica and titanium                   to break up banding patterns of differentiating OPC1 cells.
were 2.58, 2.75 and 2.83, respectively. The average scores                 Edges and corners appeared to induce peeling and disrupt
for day 5, 11 and 16 were 2.33, 2.25 and 2.58, respectively.               banding.
Fig. 5 compares adhesion scores for gold, titanium and sil-                   Adhesion scores for each channel of nine sample chips
ica surfaces. A Kruskal–Wallis non-parametric one-way                      for days 5, 11 and 16 are displayed in Table 2. Average
ANOVA test [14] showed no significant difference among                       scores for the narrowest five channels varied from 1.17 to
adhesion scores for gold, titanium or silica.                              1.53. Average scores for the other channels generally
                                                                           increased from 1.22 to 2.94. Top surface and center floor
3.2. Narrow channel inhibition                                             average scores were 3.28 and 3.44, respectively. Fig. 7 illus-
                                                                           trates these scores. Scores generally increased with channel
   Design B varied channel width from 25 to 1000 lm at                     floor widths of 50–350 lm, and increased slightly further
the surface and from 0 to 750 lm at the floor. Because of                   with the nearly unlimited channel widths of the center floor
                                                 R. Condie et al. / Acta Biomaterialia 3 (2007) 523–530                                            527




                                                                                  Average Score





                                                                                                                 Gold      Titanium   Silica
                                                                                                   Error = Standard Error Mean

                                                                              Fig. 5. Comparison of OPC1 adhesion scores on gold, titanium and silica
                                                                              coatings. An ANOVA test showed no significant difference among mean
                                                                              scores for the three materials at the 95% confidence level.

                                                                              and top surface. The slight variance within this pattern at
                                                                              250 lm is well within one standard deviation. A trend line
                                                                              describes a mutually increasing relationship between

Fig. 4. Comparison of OPC1 morphology on: (a) gold, (b) titanium and
(c) silica after 16 days in culture from the central well.

Table 1
Adhesion scores on Au, Ti, and silica surfaces
Chip design            Gold surface    Titanium surface     Silica surface
Day 5
Design   A             2               1                    1
Design   B             1               3                    4
Design   C             2               3                    3
Design   D             2               3                    3
Day 11
Design A               4               2                    2
Design B               3               3                    3
Design C               3               3                    1
Design D               2               0                    1
Day 16
Design A               3               5                    4
Design B               3               3                    4
Design C               2               3                    5
Design D               4               4                    3                 Fig. 6. SEM images of OPC1 cells on microchannels of Design B. (a)
                                                                              Channels of surface width 25, 50, 75 and 100 lm. Sidewalls meet at the
Mean                   2.58            2.75                 2.83              bottom of each channel. Channels 200, 300 and 400 lm wide at the surface
Standard deviation     0.90            1.29                 1.34              are shown in (b). Banding patters are broken and cells are sparsely
Qualitative descriptions of OPC1 adhesion are correlated to a score on the    scattered in channels less than 400 lm wide. Peeling frequently occurs at
left. Scores for 36 samples are tabulated on the right.                       edges.
528                                                                                   R. Condie et al. / Acta Biomaterialia 3 (2007) 523–530

Table 2
Table of adhesion scores on Design B microchannels with varied width
Path #     Floor width (lm)                                              Top width (lm)    Design B adhesion scores
                                                                                           Au 5    Au 11         Au 16       Ti 5   Ti 11     Ti 16       SiO2 5   SiO2 11        SiO2 16   Average
 1           0                                                             25              1       1             1.5         1      2         2           1        1              2         1.39
 2           0                                                             50              1       1             1.5         1      2         1           1.5      2              2         1.44
 3           0                                                             75              1       1             0           1      2         1.5         1        1.5            1.5       1.17
 4           0                                                            100              1       1             1           1      2         1           1        2              1.5       1.28
 5           0                                                            200              1       2             1           2      1         2           1.5      2              1.25      1.53
 6          50                                                            300              1       1             1.5         1      1         2           1        1              1.5       1.22
 7        150                                                             400              1       2             2           2      2         2           2        2              1.5       1.83
 8        250                                                             500              1       3             3           3      3                     3        2              3         2.63
 9        350                                                             600              1       2             3           3                            3        2                        2.33
10        550                                                             800              1       3             3.5         3      3         3           3                       4         2.94
11        750                                                            1000              1       3             3.5         3      3         3           3                       4         2.94
Center floor                                                                                1.5     3.5           4           3      3.5       4           3.5      3.5            4.5       3.44
Surface                                                                                    1       3             4           3      3         4           3        4              4.5       3.28
Average                                                                                    1.00    1.82          1.96        1.91   2.1       1.94        1.91     1.72           2.23      1.88
Qualitative descriptions of OPC1 proliferation and adhesion are correlated to a score on the left. Mean scores for nine samples for days 5, 11 and 16 are
tabulated on the right.

                                                                                                                  y = -3E-06x2 + 0.0048x + 1.297
                                                                                                                  R2 = 0.8843
                    error = Standard Error Mean


                                                  Average Score





                                                                                25    50     75    100     200         300   400    500     600     800    1000    Top    Floor
                                                  Width (μm)

Fig. 7. Plot comparing adhesion scores for channels of Design B. Channels vary from 25 to 1000 lm in width at the top surface. Channels narrower than
300 lm at the top surface had no floor or flat surface at the bottom. The area is nearly unlimited for the top surface.

adhesion score and path width. A Kruskal–Wallis non-                                                                     of chips as well as within channels. Occasionally, bands
parametric one-way ANOVA test showed a significant dif-                                                                   appeared to radiate out from exterior corners. Sidewalls,
ference among scores of different channel widths. Channel                                                                 narrow channels and other obstacles often broke up band-
sidewalls sloped at 54.74° did not visibly inhibit the prolif-                                                           ing and alignment patterns. Peeling of banded cells fre-
eration of OPC1 cells. Fig. 8 is a characteristic SEM image                                                              quently occurred at channel edges as well.
of a design B chip showing similar cell density on the chan-
nel floors and walls, and on the surface. Because cells were                                                              4. Discussion
not confined within channels under the conditions of this
experiment, no substantive data were gathered on the                                                                         In an effort to better understand osseointegration, this
effects of channel tortuosity, divergence or convergence.                                                                 experiment investigated immortalized OPC1 adhesion and
   In many images, especially those of samples with high                                                                 proliferation on various micro-topographies and bioinert
adhesion scores, bands of OPC1 cells appeared to be elon-                                                                coatings. The relative bioactivities of gold, titanium, and
gated and uniformly oriented in local areas. Frequently,                                                                 silica coatings were compared. Adhesion and proliferation
they appeared to be aligned with the nearest channels, as                                                                behaviors based on observations of cell density along chan-
seen in Fig. 9. This behavior occurred on the top surface                                                                nels of varied width and length were observed. Fabrication
                                             R. Condie et al. / Acta Biomaterialia 3 (2007) 523–530                                529

                                                                          channel floors and walls and the top surfaces of each sam-
                                                                          ple chip by day 5 and continued to extend and differentiate
                                                                          through day 16. Some cracking and peeling that reduced
                                                                          adhesion scores may have developed during fixation and
                                                                              It was hypothesized that OPC1 would exhibit no detect-
                                                                          able preference for gold, titanium or silica surfaces. These
                                                                          materials are described as bioinert in literature and are
                                                                          often used in bone implants or other biomedical devices.
                                                                          As illustrated in Fig. 4 and confirmed by the Kruskal–Wal-
                                                                          lis non-parametric one-way ANOVA test, no significant
                                                                          difference between adhesion scores for the three material
                                                                          coatings was observed.
                                                                              The migration of osteoblasts through various porous
                                                                          materials indicates their ability to conform to a variety of
Fig. 8. SEM image of convergent channel of a Design B sample chip.        topographies [15]. As bone tissue growth along two-dimen-
Sidewalls failed to contain OPC1 proliferation and migration. Cells       sional paths does not appear to have been studied previ-
adhered to the channel floor, side walls and top surface.
                                                                          ously, it was also hypothesized that OPC1 cells may
                                                                          proliferate along and adhere to channels of varying length,
                                                                          width, tortuosity, convergence and divergence. Design B
                                                                          varied channel width from 25 to 1000 lm at the surface
                                                                          and from 0 to 750 lm at the floor. As observed in Fig. 6,
                                                                          cultured cells were present in all channels, but sidewalls
                                                                          and narrow channels broke up banding patterns and signif-
                                                                          icantly inhibited adhesion. A plot of adhesion scores versus
                                                                          channel width, shown in Fig. 7, revealed an increasing trend
                                                                          with width. A significant variation among the mean scores
                                                                          of various channel widths was confirmed by the Kruskal–
                                                                          Wallis non-parametric one-way ANOVA test. These results
                                                                          support the key finding that channels of floor width nar-
                                                                          rower than 350 lm reduce OPC1 adhesion in these bioinert
                                                                          surfaces, suggesting a minimal width for flat engineered bio-
                                                                          material surfaces for optimized integration with bone cells.
                                                                          This surface channel width requirement is somewhat larger
                                                                          than the 100 lm requirement of three-dimensional porous
                                                                          matrices quoted in the literature [16,17]. OPC1 cells have
                                                                          been shown to proliferate well in ceramic pores ranging
                                                                          from 300 to 500 lm in width [18]. Similarly, foam pore sizes
                                                                          ranging from 150 to 710 lm did not significantly affect stro-
                                                                          mal osteoblast proliferation or function [16].
                                                                              Because 120 lm deep sidewalls sloped at 54.74° failed to
                                                                          contain OPC1 cell growth within channels under these
                                                                          experimental conditions, no substantive data were accumu-
                                                                          lated on the effects of channel divergence or convergence,
                                                                          or tortuosity. However, the presence of healthy OPC1 cells
                                                                          in all areas of each chip shows their ability to adhere to
                                                                          such surfaces. Overflow during OPC1 seeding likely con-
                                                                          tributed to the broad proliferation patterns. Proliferating
                                                                          cells may have migrated over channel walls and across
Fig. 9. OPC1 cells appear to be aligned with nearby sidewalls in SEM      the top surfaces as well as along the floor, disrupting con-
images (a) and (b). Banding patterns appear to follow channels around     fidence in definite cell migration pathways. In future exper-
bends or radiate from corners.                                            iments, channel surfaces may be covered with a glass cover
                                                                          slip or removable, microfacricated top to contain cell
of effective micro-patterned substrates for cell proliferation             migration within the channel. The 2 ml suspension volume,
was accomplished by oxidization of silicon wafers, photoli-               corresponding to a 7.4 ml seeding volume, was selected as a
thography, EDP etching and sputter coating with gold or                   minimum for statistically consistent seeding of 20,000 cell
titanium or oxidation to form SiO2. Cultured cells covered                cells. Seeding fewer cells in a lower-volume suspension
530                                      R. Condie et al. / Acta Biomaterialia 3 (2007) 523–530

may prevent overflow and enable channels to contain cell               program under grant number 0453554 and the Office of
proliferation. Analysis at earlier time points may provide            Naval Research (Grant Number: N00014-1-04-0644).
useful data as well. Moreover, other etching techniques
such as plasma etching can be used to fabricate channels              References
with nearly 90° sidewalls that may better contain cell pro-
liferation within the channels.                                        [1]
    Abrupt angles at channel walls appeared to interrupt               [2] Hedia HS, Mahmoud NA. Design optimization of functionally
                                                                           graded dental implant. Bio-Med Mater Eng 2005;14:133–43.
banding and adhesion patterns, and frequently stimulated               [3] Ratner BD, Hoffman AS, Schoen FJ, Lemons JE. Biomaterials
peeling of banded cells away from the material surface.                    science: an introduction to materials in medicine. San Diego
Therefore, straight surfaces may be recommended in engi-                   (CA): Elsevier Academic Press; 2004, p 154–5.
neered biomaterial surfaces to improve osseointegration.               [4] Dalby MJ, McCloy D, Robertson M, Wilkinson CDW, Oreffo ROC.
We have also observed the apparent alignment of OPC1                       Osteoprogenitor response to defined topographies with nanoscale
                                                                           depths. Biomaterials 2006;27:1306–15.
bands on floor or top surfaces with local channels. These               [5] Meyer U, Joos U, Wiesmann HP. Biological and biophysical
observations are in harmony with previous studies of                       principles in extracorporal bone tissue engineering. Int J Oral
osteoblast alignment with parallel grooves fabricated in                   Maxillofac Surg 2004;33:325–32.
various material surfaces, but expand the study to tortuous            [6] Daculsi G, LeGeros RZ, Deudon C. Scanning and transmission
and divergent channels [10,11,16]. Based only on visual                    electron microscopy, and electron probe analysis of the interface
                                                                           between implants and host bone. Osseo-coalescence versus osseo-
observation, aligned bands appeared to follow channels                     integration. Scanning Microsc 1990;4:309–14.
around corners and obstacles. An in-depth statistical anal-            [7] Fredel MC, Boccaccini AR. Processing and mechanical properties of
ysis of alignment is beyond the scope of this experiment,                  biocompatible Al2O3 platelet-reinforced TiO2. J Mater Sci
but further studies recording the elongation and orienta-                  1996;31:4375–80.
                                                                       [8] Mayr-Wohlfart U, Fiedler J, Gunther K-P, Puhl W, Kessler S.
tion of each cell could validate the observation that cells
                                                                           Proliferation and differentiation rates of a human osteoblast-like cell
align with the nearest obstacle. This alignment with topo-                 line (SaOS-2) in contact with different bone substitute materials. J
graphical features may allow tissue engineering of bone                    Biomed Mater Res 2001;57:132–9.
cells on implant surfaces [19]. In a recent publication sup-           [9] Zinger O, Zhao G, Schwartz Z, Simpson J, Wieland M, Landolt D,
porting this idea, bone tissue has been shown to preferen-                 et al. Differential regulation of osteoblasts by substrate microstruc-
tially form within and along grooves 110 lm wide and                       tural features. Biomaterials 2005;26:1837–47.
                                                                      [10] Ber S, Torun Koseb G, Hasırcıa V. Bone tissue engineering on
70 lm deep in oxidized titanium dental implants, providing                 patterned collagen films: an in vitro study. Biomaterials
increased resistance to shear forces [20]. Overall, our results            2005;26:1977–86.
show that engineered surface topographies have potential              [11] Dunn GA, Brown AF. Alignment of fibroblasts on grooved surfaces
to greatly improve osseointegration.                                       described by a simple geometric transformation. J Cell Sci
                                                                      [12] Lenhert S, Meier MB, Meyer U, Chi Lifeng, Wiesmann HP.
5. Conclusions                                                             Osteoblast alignment, elongation and migration on grooved polysty-
                                                                           rene surfaces patterned by Langmuir–Blodgett lithography. Bioma-
   We have studied effects of various microchannel param-                   terials 2005;26:563–70.
eters on immortalized osteoblast precursor cell (OPC1)                [13] Winn SR, Randolph G, Uludag H, Wong SC, Hair GA, Hollinger
                                                                           JO. Establishing an immortalized human osteoprecursor cell line:
adhesion and proliferation. Microchannels were micoma-
                                                                           OPC1. J Bone Miner Res 1999;14:1721–33.
chined on (1 0 0) Si wafers using anisotropic wet etching             [14] William H Kruskal, Allen Wallis W. Use of ranks in one-criterion
with varying designs to understand the influence of channel                 variance analysis. J Am Stat Assoc 1952;47(260):583–621.
width, channel length, tortuosity, convergence and diver-             [15] Kalita SJ, Bose S, Hosick HL, Bandyopadhyay A. Development of
gence. We have also coated these channels with Au, Ti                      controlled porosity polymerceramic composite scaffolds via fused
and SiO2 to vary the bio-inert surface chemistry. Among                    deposition modeling. Mater Sci Eng, C 2003;23:611–20.
                                                                      [16] Ishaug SL, Crane GM, Miller MJ, Yasko AW, Yaszemski MJ, Mikos
all parameters, channel width is found to influence the                     AG. Bone formation by three-dimensional stromal osteoblast culture
most in terms of cell attachment. A channel floor width                     in biodegradable polymer scaffolds. J Biomed Mater Res
narrower than 350 lm has shown reduced OPC1 adhesion                       1996;36:17–28.
in these bioinert surfaces, suggesting a minimal width for            [17] Klawitter JJ, Hulbert SF. Application of porous ceramics for the
flat engineered biomaterial surfaces for optimized integra-                 attachment of load bearing orthopaedic applications. Biomed Mater
                                                                           Symp 1971;2:161–7.
tion with bone cells.                                                 [18] Bose S, Darsell J, Kinter M, Hosick H, Bandyopadhyay A. Pore size
                                                                           and pore volume effects on alumina and TCP ceramic scaffolds. Mater
Acknowledgements                                                           Sci Eng, C 2003;23:479–86.
                                                                      [19] Bandyopadhyay A, Bernard S, Xue W, Bose S. Calcium phosphate
   The authors thank Mr. Hongsoo Choi, Ms. Kakoli Das                      based resorbable ceramics: influence of MgO, ZnO and SiO2 dopants.
                                                                           J Am Ceram Soc 2006;89(9):2675–88.
and Ms. Jessica Moore of WSU for experimental assis-                  [20] Hall J, Miranda-Burgos P, Sennerby L. Stimulation of directed bone
tance. This work was supported through the National Sci-                   growth at oxidized titanium iimplants by macroscopic grooves: an
ence Foundation: Division of Materials Research REU site                   in vivo study. Clin Implant Dent Relat Res 2005;7:76–82.

To top