lable at ScienceDirect

Biomaterials 37 (2015) 134e143

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomateria ls

Abnormal arrangement of a collagen/apatite extracellular matrixorthogonal to osteoblast alignment is constructed by a nanoscaleperiodic surface structure

Aira Matsugaki a, Gento Aramoto a, Takafumi Ninomiya b, Hiroshi Sawada b,Satoshi Hata c, Takayoshi Nakano a, *

a Division of Materials and Manufacturing Science, Graduate School of Engineering, Osaka University, 2-1 Yamada-oka, Suita, Osaka 565-0871, Japanb Canon Machinery Inc., 85, Minami Yamada-cho, Kusatsu, Shiga 525-8511, Japanc Department of Electrical and Materials Science, Kyushu University, 6-1 Kasuga-koen, Kasuga, Fukuoka 816-8580, Japan

a r t i c l e i n f o

Article history:Received 30 July 2014Accepted 2 October 2014Available online 30 October 2014

Keywords:Microbeam X-ray diffractionOsteoblastCollagen structureApatite orientationLaser-induced periodic surface structures(LIPSS)

* Corresponding author. Tel./fax: 81 6 6879 7505.E-mail address: nakano@mat.eng.osaka-u.ac.jp (T.

http://dx.doi.org/10.1016/j.biomaterials.2014.10.0250142-9612/ 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

Morphological and directional alteration of cells is essential for structurally appropriate construction oftissues and organs. In particular, osteoblast alignment is crucial for the realization of anisotropic bonetissue microstructure. In this article, the orientation of a collagen/apatite extracellular matrix (ECM) wasestablished by controlling osteoblast alignment using a surface geometry with nanometer-sized peri-odicity induced by laser ablation. Laser irradiation induced self-organized periodic structures (laser-induced periodic surface structures; LIPSS) with a spatial period equal to the wavelength of the incidentlaser on the surface of biomedical alloys of Tie6Ale4V and CoeCreMo. Osteoblast orientation wassuccessfully induced parallel to the grating structure. Notably, both the fibrous orientation of the secretedcollagen matrix and the c-axis of the produced apatite crystals were orientated orthogonal to the celldirection. To the best of our knowledge, this is the first report demonstrating that bone tissue anisotropyis controllable, including the characteristic organization of a collagen/apatite composite orthogonal tothe osteoblast orientation, by controlling the cell alignment using periodic surface geometry.

2014 Elsevier Ltd. All rights reserved.

1. Introduction

The realization of in vitro tissue construction shows greatpotential not only for tissue regeneration but also for under-standing the mechanisms that regulate tissue genesis. Livingtissues and organs express specialized forms and unique physicalproperties required to carry out their specific functions. Bonetissue, in particular, has a characteristic hierarchical anisotropicstructure, from nano- to macroscale [1], which governs the me-chanical properties of bone tissue [2]. Regeneration of bone thatmaintains the original anisotropic structure requires specially-designed biomaterials that can retrieve the crystallographicorientation of bone, because even advanced tissue engineeringtechniques cannot promote anisotropic regeneration of bonetissue [3,4]. Control of the orientation of cell-produced

Nakano).

collagen/apatite in the favored direction is therefore essential inbone tissue engineering.

Patterning of extracellular matrix (ECM) deposition isconsidered to be closely related to polarized cell behaviors dur-ing morphogenesis [5], and the interplay between integrin re-ceptors and the ECM is thought to be important for a variety ofmorphogenetic events [6]. Directional deposition of matrix pro-teins corresponding to cell orientation has been demonstratedusing microgrooves [7e9] and an anisotropic mechanical envi-ronment [10], indicating that the anisotropy of the ECM isdetermined by cellular arrangement in response to externalstimuli. Indeed, that the cell-produced matrix orientation fol-lows the cellular direction is a classical belief of tissue matrixstructure formation. However, it is not known this always holdstrue.

In the present article, osteoblasts were grown on a biomimeticnanogrooved substrate, and the architecture of the resulting bonematrix was examined both qualitatively and quantitatively. Laser-induced nanometer-scale surface topography was introduced intotwo typical biomedical alloys, Tie6Ale4V and CoeCreMo. The cells

mailto:nakano@mat.eng.osaka-u.ac.jphttp://crossmark.crossref.org/dialog/?doi=10.1016/j.biomaterials.2014.10.025&domain=pdfwww.sciencedirect.com/science/journal/01429612http://www.elsevier.com/locate/biomaterialshttp://dx.doi.org/10.1016/j.biomaterials.2014.10.025http://dx.doi.org/10.1016/j.biomaterials.2014.10.025http://dx.doi.org/10.1016/j.biomaterials.2014.10.025

A. Matsugaki et al. / Biomaterials 37 (2015) 134e143 135

preferentially aligned along the nanogrooves, consistent with theresults of previous studies [11,12]. In this study, however, the ar-chitecture of the cell-produced bone matrix, which was quantita-tively determined using Raman microscopy and microbeam X-raydiffraction (XRD), showed unique anisotropy against the cellularorientation; these findings reverse the current belief that the ECMis constructed parallel to the cellular arrangement and provide newinsight into the mechanisms of bone tissue morphogenesis. Thesefindings on the relationship between osteoblasts and bone matrixorientation will facilitate innovation in the development of bio-materials appropriate to induce bone regeneration, including re-covery of the original anisotropic structure of bone, and may alsofacilitate breakthroughs in our understanding of the mechanismunderlying anisotropic tissue generation.

Fig. 1. SEM images ((a)(d)) and AFM images ((e)(h)) of the surface topography of Tie6Asectional SEM image of each alloy. The control surfaces display isotropic geometry withanisotropic periodicity with spatial periods somewhat smaller than the wavelength of the

2. Materials and methods

2.1. Fabrication of periodic surface structures

Samples of forged Ti6 mass% Al4 mass% V alloy (Tie6Ale4V; ASTM F136-02A) and Co28 mass% Cr6 mass% Mo alloy (CoeCreMo; ASTM F1537-08) witha diameter of 15 mm and a height of 5.0 mm were obtained. These samples wereground using emery paper (#120, #320, #600), then polished with diamond paste (9and 3 mm) and colloidal silica suspension (0.06 mm). Periodic structures weregenerated on the surface of each sample using a p-polarized Ti: sapphire femto-second laser (peak wavelength of 800 nm, pulse width of 250 fs, cyclic frequency of2 kHz). A circularly polarized laser was used to prepare a control substrate withoutdirectional characteristics.

2.2. Surface characterization

The periodic surface structures produced by laser irradiation were observedusing scanning electron microscopy (SEM; VE-7800, Keyence, Osaka, Japan). The

le4V ((a), (b), (e), (f)) and CoeCreMo ((c), (d), (g), (h)). The inset outlines the cross-a fine-dot structure ((a), (c), (e), (g)), whereas the laser-irradiated substrates displayincident laser, at 800 nm ((b), (d), (f), (h)).

A. Matsugaki et al. / Biomaterials 37 (2015) 134e143136

surface geometry was characterized using a laser microscope (VK-9700, Keyence,Osaka, Japan); 10 images (177 mm2) were captured for each sample. Three-dimensional imaging and quantification of surface topographic features were per-formed using an atomic force microscope (Nano-R2, Pacific Nanotechnology, SantaClara, CA). Topographical images of 10 10 mm2 sections of the substrate surfaceswere obtained in tapping mode.

2.3. Osteoblast isolation and culture

Primary osteoblasts were isolated from the calvariae of newborn mice. Briefly,calvariae from newborn C57BL/6 mice were excised under aseptic conditions, placedin ice-cold a-MEM (Invitrogen, Carlsbad, CA), and then the fibrous tissues around thebone were gently removed. The calvariae were then subjected to a series of colla-genase/trypsin (collagenase: Wako, Osaka, Japan; trypsin: Nacalai Tesque, Kyoto,Japan) digests at 37 C for 15 min each. The first two digests were discarded, andsupernatants 3, 4, and 5 were neutralized with a-MEM, pooled, and filtered using a200-mm mesh (BD Biosciences, San Jose, CA). The filtrates were centrifuged and theresulting pellets were resuspended in a-MEM containing 10% fetal bovine serum(FBS) for cell culture. Cells were then diluted to 4000 cells/cm2 and seeded onto thefabricated specimens. The medium was changed twice a week, and after culturingfor 7 days, the media was supplemented to achieve final concentrations of 50 mg/mLascorbic acid (Sigma, St. Louis, MO), 10 mM b-glycerophosphate (Tokyo Kasei, Tokyo,Japan), and 50 nM dexamethasone (MP Bioscience, Solon, OH). All animal experi-ments were approved by the Osaka University Committee for AnimalExperimentation.

2.4. Fluorescence imaging

After culturing for 3 days, the cells were incubated in PBS-0.05% Triton X-100(PBST) containing 1% normal goat serum (NGS; Invitrogen) for 30 min to block non-specific antibody binding sites. The cells were then incubated with mouse mono-clonal antibodies against vinculin (SigmaeAldrich) at 4 C for 12 h. This step wasfollowed by incubation with Alexa Fluor 546 conjugated anti mouse IgG (MolecularProbes, Invitrogen) and Alexa Fluor 488 conjugated phalloidin (Molecular Probes,Invitrogen). Finally, the cells were washed with PBST and mounted in Prolong Gold

Fig. 2. (A) Morphological analysis of cells cultured on the fabricated substrates. An immunmation (c) was applied. The major; a, and the minor axes of cell shape; b, were determinosteoblasts cultured on the control ((a), (c)) and nanogrooved substrates ((b), (d)). Cell oriensurface, the distribution of cell alignment was concentrated around 0 , whereas no peak reScale bar: 50 mm. The insets show magnified images of the cells. Elongated focal adhesiorelatively small and round focal adhesions disconnected orthogonal to the nanogrooves. Whinanogrooves. Scale bar: 20 mm. (C) The angular standard deviation of the orientation angle otext). The SD values of the cell distribution on nanogrooved substrates were significantly smclear effect on cell alignment parallel to the periodicity. (For interpretation of the references

antifade reagent with DAPI (Molecular Probes, Invitrogen). Fluorescent images wereobtained using a fluorescence microscope (Biozero, Keyence, Osaka, Japan) andprocessed using Adobe Photoshop 10.0 software.

For visualization of collagen fibers, rabbit polyclonal antibodies against collagentype I (Abcam, Cambridge, MA) were applied, and the above-described procedurewas performed.

2.5. Cell orientation

The orientation of the cells on the substrates was examined relative to thegrating structure by taking photographs of the fluorescent phalloidin staining im-ages of the cells using a fluorescence microscope (Biozero, Keyence, Osaka, Japan).Cell orientation was quantitatively analyzed using the Cell Profiler software (BroadInstitute, Cambridge, MA) (Fig. 2A). The degree of cell orientation was characterizedby the angular standard deviation, SD (s), for a wrapped normal distribution [13,14].Here, the probability distribution function was adapted from Fisher [15] for a peri-odicity of p radians, where m is the mean angle and r is the mean resultant length.These parameters were determined from a set of n measured cell orientations, qi,using the following equations:

f q 1p

0@1 2X

p1rp

2cos2pq p

1A; (1)

r 1n

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Xni1

cos 2qi

!2 Xn

i1sin 2qi

!2vuut ; (2)

m tan1Pn

i1 sin 2qiPni1 cos 2qi

: (3)

The angular standard deviation (s) was determined using the followingequation:

s 12

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2 ln r

p: (4)

ocytochemical image of an osteoblast (a) was binarized (b), and an elliptical approxi-ed. The cellular angle was also determined q (d). (B) Immunocytochemical images oftation against surface periodicity was quantified at 10 ((e)(h)). On the nanogroovedgion was observed on the control substrate. Green: F-actin, red: vinculin, blue: nuclei.ns were found in the cells aligned along the nanogrooves. Black arrows indicate thete arrows indicate the enlarged, elongated fibril adhesions constructed in parallel to thef the cells, SD (s), calculated from the wrapped normal distribution (equation (4) in thealler than those on the control surfaces, indicating that the surface nanogrooves have ato color in this figure legend, the reader is referred to the web version of this article.)

Fig. 2. (continued).

A. Matsugaki et al. / Biomaterials 37 (2015) 134e143 137

mailto:Image of Fig. 2|eps

0

10

20

30

40

50

60

70

80

Sta

ndar

d de

viat

ion

of o

rient

atio

nan

gle

of c

ell (

deg.

)

Ti6Al4V CoCrMo

control controlnanogrooved nanogrooved

* *

C

Fig. 2. (continued).

Table 1The surface topographical properties of the fabricated substrates.

Pitch [nm] Depth [nm] Ra [nm]

Tie6Ale4V Control e e 39 4Nanogrooved 532 132 252 96 103 16

CoeCreMo Control e e 52 2Nanogrooved 658 50 176 37 66 10

A. Matsugaki et al. / Biomaterials 37 (2015) 134e143138

2.6. Raman microscopic analysis

The preferential orientation of the secreted collagen matrices was determinedusing Raman microscopy (NRS-5100, JASCO, Tokyo, Japan) by rotating the sampleswith respect to the polarization of the incident laser beam. Collagenmolecules showcharacteristic optical features which can aid in identifying the orientation ofcollagen fibers, by referring the intensity between certain Raman bands, e.g., theamide I (~1664 cm1), and CeH bending (~1451 cm1) bands [16]. The spectra werenormalized to the Raman band at approximately 1451 cm1, which is an isotropicRaman band with no preferential orientation. The peak intensities of amide I n(C]O)/d(CeH) for a total rotation of 360 were approximated by the following ellipticpolynomial function using the least-square method:

Px cos2x c

a2 sin

2x cb2

12 d (5)

where a, b, c and d are the fitting parameters and c is the angle at which the intensitypeaks. The rotation angle with the highest peak intensities means the preferredorientation of the C]O bonds, which are oriented perpendicular to the backbone ofcollagen molecules.

2.7. Microbeam X-ray diffraction analysis

The character and preferred orientation of apatite crystals formed by osteoblastswere analyzed using a micro-beam X-ray diffraction system (Bruker AXS D8,Discover with GADDS, Germany) using Cu-Ka radiation operated at 45 kV and110 mA. An incident beam was collimated into a circular spot of 50 mm in diameter.The specimens were fixed in 10% formaldehyde, followed by measurement andanalysis. The preferred orientation of the apatite c-axis was evaluated as the relativeintensity ratio of the (002) diffraction peak to the (211) peak measured in parallelwith or perpendicular to the surface patterning. Because the thickness of themineralized nodules formed on the surface of the specimens was too small fornormal diffraction analysis, in-plane diffraction, in which the incident and diffractedbeams are nearly parallel to the sample surface, was applied.

2.8. Transmission electron microscopy

To clarify the microstructural orientation of apatite crystals, the selected areaelectron diffraction of the mineralized nodules was measured using a transmissionelectron microscope (TEM; Tecnai F20, FEI) operated at 200 kV. The specimens wererinsed in 0.1 M sodium cacodylate buffer three times and fixed with 2% glutaralde-hyde in cacodylate buffer at 4 C for 30 min, followed by washing with the samebuffer and post-fixation in 1% osmium tetroxide for 1 h. Prior to resin embedding,the samples were dehydrated in a graded ethanol series from 50 to 100% at 5-minintervals, with a final 100% wash for 30 min. The ethanol was replaced withethanol/Epon 812 (TAAB, Berkshire, UK) (1:1), following which the specimens wereembedded in Epon by incubating at 65 C for 12 h. The specimens were thinnedusing a focused ion-beam (FIB) milling technique (FB-2000 K, Hitachi, Japan) at anaccelerating voltage of 30 kV for the preparation of the cross-sectional TEM speci-mens. The electron diffraction patterns were recorded with the observation of acircular area of 3 mm in diameter.

2.9. Statistical analysis

Statistical significance was tested using the Student's t-test or Welch's t-test. Asignificance of P < 0.05 was required for rejecting the null hypothesis.

3. Results

3.1. Surface characterization

AFM and SEM images revealed that periodic surface structureswere successfully obtained by irradiation with the p-polarizedfemtosecond laser, and non-directional isotropic surface structuresas control were obtained by irradiationwith the circularly polarizedlaser (Fig. 1). The periodicity and depth of the nanogroovesdepended on the composition of the alloys. Quantitative analysisshowed that the surface structure of the Tie6Ale4V andCoeCreMo alloys had a pitch of 532 132 nm, 658 50 nm and adepth of 252 96 nm, 176 37 nm, respectively (Table 1).

3.2. Cell arrangement

The periodic surface architecture obtained using p-polarized laserirradiation induced osteoblast elongation and alignment along thenanogrooves on both alloys, Tie6Ale4V and CoeCreMo, whereasthe isotropic surface architecture produced by the circularly polar-ized laser did not induce a preferential osteoblast orientation oneither alloy (Fig. 2B). Quantitative evaluation of cell orientation usingfluorescent images of the actin cytoskeleton and vinculin clearlyrevealed that the periodic structure of the substrate had a significanteffect on cell alignment. The histograms in Fig. 2B (e)(h) show thedistribution of cell orientation on each substrate. To quantitativelyevaluate the degree of cell alignment, the distribution of the orien-tation angles, SD (s), was determined using equation (4) (Fig. 2C).The angular SD was much lower in cells cultured on the laser-induced periodic structure than in cells cultured on control sub-strates, indicating that the degree of cell alignment was stronglycontrolled by the periodicity of the substrate.

3.3. Anisotropic bone matrix organization

To visualize the collagen matrix deposited by the osteoblasts,immunocytochemical analysis was performed. Unexpectedly, thecollagenmatrix showed fibril formation perpendicular to the surfaceperiodicityonthegroovedsubstrates,whereascollagenmatrixgrownon the control substrates showednopreferential direction of collagenfiber spread (Fig. 3). In addition, the directional organization of thecollagen matrix was quantitatively evaluated using Raman micro-scopic analysis. When the angle of the substrate was rotated againstthe polarization of the incident laser light (0 indicates that laserpolarization was perpendicular to surface periodicity), the intensityratio of n(C]O) to d(CeH) exhibited the highest values at rotationangles of 90 and 270, indicating that the C]O bonds were prefer-entially aligned parallel to the direction of the nanogrooves. This in-dicates that thebackboneof thecollagenmoleculeswaspreferentiallyoriented perpendicular to the periodicity of the substrate (Fig. 4).

The formation of mineralized nodules on the substrate surfacewas confirmed (Fig. 5 (a)). Use of the in-planemethod ofmicrobeamX-ray diffraction enabled identification of the nodules deposited onthe surface of the alloy substrates. Peaks from (002) and (211),which are typical of apatite crystals as a calcified constituent, weredetected, indicating the presence of apatite crystallites (Fig. 5 (b)). To

mailto:Image of Fig. 2|eps

Fig. 3. Immunocytochemical analysis of the structure of collagen type I secreted by the osteoblasts. Collagen fibers were secreted in a random orientation on both control substrates((a), (c)), whereas they were secreted orthogonal to the grooves on the nanogrooved substrates ((b), (d)). Scale bar: 100 mm.

A. Matsugaki et al. / Biomaterials 37 (2015) 134e143 139

assess the preferential alignment of anisotropic apatite crystallites,the integrated intensity ratio of the (002) peak to the (211) peak wasanalyzed parallel (0) and perpendicular (90) to the direction of thesubstrate's periodic structure. The apatite orientation produced bycells aligned on nanogrooved substrates showed significantly highervalues in a perpendicular direction, indicating that the apatitecrystals deposited on the grating structure showed preferentialalignment of the c-axis perpendicular to the substrate periodicity. Incontrast, mineralized nodules formed on the control substratesshowed no preferential alignment (Fig. 5 (c)). TEM imaging andselected area diffraction patterns demonstrated that the bone ma-trix nodules are formed inside the grooves and show an arc-shapeddiffraction pattern of (002) in a direction orthogonal to the surfaceperiodicity (Fig. 6). These results indicate that the deposited apatitecrystals exhibit an alignment orthogonal to the osteoblast orienta-tion, as well as the collagen matrix orientation.

4. Discussion

ECM assembly is an important event that determines structuraland biological functions during both development and remodelingprocesses. In particular, the cellular systems regulating the orien-tation of the collagen/apatite microstructure of bone tissue play akey role in determining the mechanical properties of bone. Fordecades, it has been generally believed that ECM organization fol-lows cell direction [7e9,17,18]. However, it remains unclearwhether this always holds true. In particular, how osteoblasts areregulated by the biomimetic nanometer-sized topography duringthe mineralization process is still not fully understood. This report

is the first, to the best of our knowledge, to show that the bonematrix is oriented orthogonal to the osteoblast alignment, inducedby nanometer-scale periodicity.

4.1. Periodic surface structure produced by laser irradiation

Nanometer-sized grating structures were successfully obtainedby irradiation with a p-polarized femtosecond laser, as shown inFig. 1. The geometrical patterning of the fabricated surface of bothalloys, Tie6Ale4V and CoeCreMo, exhibited spatial periodssomewhat smaller than the wavelength of the incident laser (peakwavelength of 800 nm), as reported previously [19,20]. This mayhave been caused by an increase in the real part of the effectiverefractive index of the air-metal interface when nanostructuresdevelop on a metal surface [21]; the structural differences observedin the nanogrooves created on the two different alloys could also beexplained by variation in the refractive indices of the alloys.

In this study, the behavior of cells cultured on nanogrooved alloysubstrates was compared with the behavior of those cultured oncontrol substrates with a fine-dot structure. Thus, any differencesobserved between control and nanogrooved substrates shouldresult only from differences in the surface topography of the sub-strates, and not from the different levels of chemical modificationinduced by laser ablation.

4.2. Osteoblast alignment along the periodic surface structure

Cell orientation was significantly affected by the periodic surfacestructure. Osteoblasts aligned along the direction of the nanogrooves

0

100

200

300

400

500

600

700

800

11001200130014001500160017001800

).u.a(ytisnetnI

nama

R

Raman shift (cm-1)

ECM 0degECM 90deg

(C-H)1451

Amide III1247

(C=O)1664

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

1.6

0 90 180 270 360

P=

Int.[

C=O

]/Int

.[C-H

]

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

0 90 180 270 360

P=

Int.[

C=O

]/Int

.[C-H

]

Rotation angle [degree] Rotation angle [degree]

a

b cTi6Al4V CoCrMo

0

180

90270

Nanogrooveddirection

Laser polarization direction

d

Fig. 4. (a) Raman spectra of the secreted matrix oriented at 0 and 90 to the incident laser beam. The intensities are normalized to the Raman band at approximately 1451 cm1

(CeH vibration). (b) A schematic illustration of the relationship between the direction of polarization of the incident laser beam and the direction of the nanogrooves in thesubstrate. (c, d) The fitted elliptical distribution of the normalized Raman intensity ratio (amide I/CeH) for a total rotation of 360 . Two intensity peaks at approximately 90 and270 indicate the alignment of the collagen fibers orthogonal to the nanogrooves of both of the alloys, Tie6Ale4V (c) and CoeCreMo (d).

A. Matsugaki et al. / Biomaterials 37 (2015) 134e143140

on both the Tie6Ale4V and CoeCreMo alloys, with elongated,mature focal adhesions parallel to the nanogrooves, as shown inFig. 2. Previous studies have demonstrated that a separation of73 nm allows effective integrin clustering and activation [22e24].The threshold for cellular response to groove depth has been re-ported to be ~34 nm [25]; surface patterning with a depth less than34 nm leads to smoothening with serum proteins, and integrinscannot recognize the surface patterning. Moreover, nanometer-sizedgrooves can limit cell penetration into them and do not allow theadhesions throughout the cellegroove interface [26]. Taken together,the grating structure used in the present study, with a spacing pitchof 500 nm and a depth of 180 nm, is considered to induce focaladhesion maturation on the top of ridges parallel to nanoscalegrooves. In other words, integrin clustering and activation prefer-entially proceed parallel to the direction of the grooves, with stressfibers organized parallel to the nanopatterns, resulting in cellalignment. Cellular interactions with anisotropic topography are alsoknown to be related to tyrosine phosphorylation at focal adhesions[27,28]. The elongated, mature focal adhesions in the direction of thegrating on the top of ridges may therefore activate the downstreamsignaling events in a manner dependent on cellular orientation. Theformation of mature fibrillar adhesions (>5 mm), also known assupermature adhesions are closely related to integrin signaling anddownstream protein phosphorylation levels are regulated in amanner dependent on focal adhesion size [29,30]. Taken together,the formation of abnormally elongated fibrillar adhesions is involved

not only in the control of cell alignment but also in the subsequentsignaling cascade that regulates the osteoblast functions, includingbone matrix formation.

4.3. Organization of the collagen/apatite matrix orthogonal toosteoblast alignment

In this study, the production of the collagenous matrix by cellsaligned along the nanometer-scale topography was quantitativelyrevealed; a novel orthogonal relationship between cells and the cell-produced collagen matrix was revealed by qualitative immunocy-tochemical analysis (Fig. 3), quantitative analysis using Raman mi-croscopy (Fig. 4), and microbeam XRD (Fig. 5). TEM analysis of thediffraction peaks of apatite crystals also demonstrated the orthog-onal alignment of the (002) diffraction peak to surface periodicity(Fig. 6). As mentioned above, osteoblast alignment is crucial foranisotropic construction of the microstructure of bone tissue [10,31],and we revealed the quantitative relationship between osteoblastalignment and crystallographic orientation of apatite in vitro [32]. Onthe other hand, the structure of cell-produced ECM has mostly beenevaluated using qualitative observations thus far, and the generalunderstanding is that the cell-secreted collagen matrix follows thecellular alignment [7e9]. These previous studies revealed that cellsaligned along the microgrooves produce an oriented collagenousmatrix parallel to the cell direction. Our new findings indicate thatmatrix orientation may be dependent on pattern spacing, and that

groovescontrol

20 25 30 35 40 45

Inte

nsity

(a.u

.)

2 [deg.]

20 30 40

(211)

(002)

background

0.1

0.11

0.12

0.13

0.14

0.15

0deg 90deg 0deg 90deg

nanogroovedcontrol

Inte

grat

ed in

tens

ity ra

tioof

(002

)/(21

1)

*

a

b

c

Fig. 5. (a) Mineralized nodules produced by the osteoblasts on the surface of thefabricated alloys. (b) XRD profile of deposited minerals on the Tie6Ale4V alloys withnanogrooves. The background profile of the Tie6Ale4V alloy is shown in the inset.Typical diffraction peaks of apatite, (002) and (211), were detected. (c) Integrated in-tensity ratio of (002)/(211) oriented at 0 and 90 to the nanogrooves. The preferentialorientation of apatite crystals was significantly higher at 90 to the nanogrooves thanat 0 to the nanogrooves, indicating that the preferential orientation of c-axis of theapatite crystals was parallel to the secreted collagen fibers, which are producedperpendicular to the cells, in turn. *; P < 0.05.

Fig. 6. TEM observations of the cross-section of the mineralized matrix produced bythe osteoblasts on the Tie6Ale4V alloys with nanogrooves. (A) Bright-field imageshowing a cross-sectional view of the mineralized area above the nanogroovedstructure. Formation of a mineralized matrix was confirmed inside the grooves. Scalebar: 500 nm. (B) An arc-shaped diffraction pattern of (002) orthogonal to the nano-grooves is shown in (b), indicating that the c-axis of apatite crystals was preferentiallyaligned perpendicular to the nanogrooves.

A. Matsugaki et al. / Biomaterials 37 (2015) 134e143 141

there is a threshold for the parallel or perpendicular organization ofbone matrix produced by aligned osteoblasts.

The classical belief that the ECM is constructed parallel toosteoblast alignment is considered to be derived from the collagen-secretion processes involving the fiber assembly in the extracellularspace [33e35] or cellular contraction force [36]. However, theorthogonally-constructed bone matrix structure reported in thispaper indicates a cell-autonomous matrix organization perpen-dicular to groove direction. As mentioned above, focal adhesionmaturation on top of the ridges and the consequent formation ofelongated fibrillar adhesions parallel to the groove direction wereobserved in this study. Integrin clustering inmature focal adhesionsis involved in multiple signaling pathways regulating matrix con-struction, including activation of matrix metalloproteinases(MMPs), which regulate collagen I processing [37,38]. In particular,fibrillar adhesions, which mature into larger structures and stablyattach to the ECM, are involved in ECM remodeling [39]. Indeed,transgenic mice with altered integrin function in osteoblasts

express nonpolarized cells and a disorganized bone matrix [40].Collectively, the organization of the collagen/apatite matrixorthogonal to osteoblast orientation appears to be the result of anosteoblast-autonomous remodeling procedure that more stablymaintains the cell-ECM adhesion area by enlarging the ridge-topadhesion sites that cross the nanogrooves (Fig. 7). That is, theremust be a cell-autonomous procedure for stabilization of the sur-rounding ECM architecture into a comfortable bed matrix to ac-quire sufficient adhesion sites to the nanometer-sized topography.While the molecular mechanisms underlying this quite uniqueperpendicular organization of bonematrix by osteoblasts remain tobe completely elucidated, our findings indicate that the orientationof the bone matrix is controllable by regulating the spacing patternof substrate periodicity. The present study revealed a perpendicularrelationship between osteoblasts and the bonematrix they produceby using specific surface patterning obtained by a laser-inducedself-organization process. To clarify the spacing patterns criticalfor determining bone matrix orientation, an exhaustive analysisincluding a wide range of surface periodicities from molecular-recognition size of nanometer-to micrometer-scales of surfacepatterning, will be required. Clarification of the nanogroove width

Fig. 7. A schematic illustration of ECM organization by osteoblasts. On the controlsubstrate, the collagen/apatite matrix produced by the randomly-spread osteoblastsshowed no preferential orientation. On the other hand, the osteoblasts aligned alongthe nanogrooved structure produced a bone matrix organized perpendicular to thesurface periodicity of the nanogrooved substrates. There must be an osteoblast-autonomous procedure for stabilization of the secreted collagen/apatite architectureto produce a comfortable bed matrix to acquire the enough adhesion sites to thenanometer-sized topography.

A. Matsugaki et al. / Biomaterials 37 (2015) 134e143142

and depth thresholds that determine matrix orientation will becrucial for regeneration of appropriately anisotropic bone tissue;this research is currently in progress and the findings will bepublished in subsequent reports.

5. Conclusion

In this article, abnormal construction of bone matrix orthogonalto the osteoblast alignment was discovered for the first time. Os-teoblasts cultured on the nanometer-sized periodicity alignedparallel to the nanogrooves as previously reported. On the otherhand, the following ECM assembly showed extraordinary align-ment against the cell direction; collagen/apatite assembly orthog-onal to the osteoblast orientation. The present novel findingreverses the conventional belief that the ECM is constructed par-allel to the cellular arrangement and provides a new insight into themechanisms underlying anisotropic tissue construction. In addi-tion, this in vitro realization of an oriented collagen/apatiteconstruct has the potential to facilitate the development of bio-mimetic implantable materials for appropriate tissue regeneration,including recovery of the original apatite orientation.

Acknowledgments

This work was supported by Grants-in-Aid for ScientificResearch (S) from the Japan Society for Promotion of Science (GrantNo. 25220912).

References

[1] Landis WJ. The strength of a calcified tissue depends in part on the molecularstructure and organization of its constituent mineral crystals in their organicmatrix. Bone 1995;16:533e44.

[2] Nakano T, Kaibara K, Tabata Y, Nagata N, Enomoto S, Marukawa E, et al.Unique alignment and texture of biological apatite crystallites in typicalcalcified tissues analyzed by microbeam X-ray diffractometer system. Bone2002;31:479e87.

[3] Nakano T, Kaibara K, Ishimoto T, Tabata Y, Umakoshi Y. Biological apatite(BAp) crystallographic orientation and texture as a new index for assessingthe microstructure and function of bone regenerated by tissue engineering.Bone 2012;51:741e7.

[4] Ishimoto T, Nakano T, Umakoshi Y, Yamamoto M, Tabata Y. Degree of bio-logical apatite c-axis orientation rather than bone mineral density controlsmechanical function in bone regenerated using rBMP-2. J Bone Miner Res2013;28:1170e9.

[5] Goto T, Davidson L, Asashima M, Keller R. Planar cell polarity genes regulatepolarized extracellular matrix deposition during frog gastrulation. Curr Biol2005;2615:787e93.

[6] Marsden M, DeSimone DW. Integrin-ECM interactions regulate cadherin-dependent cell adhesion and are required for convergent extension in Xen-opus. Curr Biol 2003;13:1182e91.

[7] Wang JH, Jia F, Gilbert TW, Woo SL. Cell orientation determines the alignmentof cell-produced collagenous matrix. J Biomech 2003;36:97e102.

[8] Wu L, Lee LA, Niu Z, Ghoshroy S, Wang Q. Visualizing cell extracellular matrix(ECM) deposited by cells cultured on aligned bacteriophage M13 thin films.Langmuir 2011;27:9490e6.

[9] Guillemette MD, Cui B, Roy E, Gauvin R, Giasson CJ, Esch MB, et al. Surfacetopography induces 3D self-orientation of cells and extracellular matrixresulting in improved tissue function. Integr Biol 2009;1:196e204.

[10] Matsugaki A, Fujiwara N, Nakano T. Continuous cyclic stretch induces osteo-blast alignment and formation of anisotropic collagen fiber matrix. Acta Bio-mater 2013;9:7227e35.

[11] Hamilton DW, Oates CJ, Hasanzadeh A, Mittler S. Migration of periodontalligament fibroblasts on nanometric topographical patterns: influence of filo-podia and focal adhesions on contact guidance. PLoS One 2010;5:e15129.

[12] Andersson AS, Olsson P, Lidberg U, Sutherland D. The effects of continuousand discontinuous groove edges on cell shape and alignment. Exp Cell Res2003;288:177e88.

[13] Bashur CA, Dahlgren LA, Goldstein AS. Effect of fiber diameter and orientationon fibroblast morphology and proliferation on electrospun poly (D,L-lactic-co-glycolic acid) meshes. Biomaterials 2006;27:5681e8.

[14] Fujita S, Ohshima M, Iwata H. Time-lapse observation of cell alignment onnanogrooved patterns. J R Soc Interface 2009;6:S269e77.

[15] Fisher NI. Statistical analysis of circular data. New York: Cambridge UniversityPress; 1993.

[16] Janko M, Davydovskaya P, Bauer M, Zink A, Stark RW. Anisotropic Ramanscattering in collagen bundles. Opt Lett 2010;35:2765e7.

[17] Jones SJ, Boyde A. Is there a relationship between osteoblasts and collagenorientation in bone? Isr J Med Sci 1976;12:98e107.

[18] Jones SJ, Boyde A, Pawley JB. Osteoblasts and collagen orientation. Cell TissueRes 1975;159:73e80.

[19] Oliveira V, Ausset S, Vilar R. Surface micro/nanostructuring of titanium understationary and non-stationary femtosecond laser irradiation. Appl Surf Sci2009;255:7556e60.

[20] Vorobyev AY, Guo C. Femtosecond laser structuring of titanium implants. ApplSurf Sci 2007;253:7272e80.

[21] Vorobyev AY, Makin VS, Guo C. Periodic ordering of random surface nano-structures induced by femtosecond laser pulses on metals. J Appl Phys2007;101:034903.

[22] Arnold M, Cavalcanti-Adam EA, Glass R, Blmmel J, Eck W, Kantlehner M, et al.Activation of integrin function by nanopatterned adhesive interfaces. Chem-physchem 2004;5:383e8.

[23] Cavalcanti-Adam EA, Volberg T, Micoulet A, Kessler H, Geiger B, Spatz JP. Cellspreading and focal adhesion dynamics are regulated by spacing of integrinligands. Biophys J 2007;92:2964e74.

[24] Geiger B, Spatz JP, Bershadsky AD. Environmental sensing through focal ad-hesions. Nat Rev Mol Cell Biol 2009;10:21e33.

[25] Lamers E, van Horssen R, te Riet J, van Delft FC, Luttge R, Walboomers XF, et al.The influence of nanoscale topographical cues on initial osteoblastmorphology and migration. Eur Cell Mater 2010;20:329e43.

[26] Kim DH, Lipke EA, Kim P, Cheong R, Thompson S, Delannoy M, et al. Nanoscalecues regulate the structure and function of macroscopic cardiac tissue con-structs. Proc Natl Acad Sci 2010;107:565e70.

[27] Franco D, Klingauf M, Bednarzik M, Cecchini M, Kurtcuoglu V, Gobrecht J, et al.Control of initial endothelial spreading by topographic activation of focaladhesion kinase. Soft Matter 2011;7:7313e24.

[28] Schernthaner M, Leitinger G, Wolinski H, Kohlwein SD, Reisinger B, Barb RA,et al. Enhanced Ca2 entry and tyrosine phosphorylation mediatenanostructure-induced endothelial proliferation. J Nanomater 2013;2013:251063.

[29] Biggs MJP, Richards RG, McFarlane S, Wilkinson CDW, Oreffo ROC, Dalby MJ.Adhesion formation of primary human osteoblasts and the functional

http://refhub.elsevier.com/S0142-9612(14)01072-2/sref1http://refhub.elsevier.com/S0142-9612(14)01072-2/sref1http://refhub.elsevier.com/S0142-9612(14)01072-2/sref1http://refhub.elsevier.com/S0142-9612(14)01072-2/sref1http://refhub.elsevier.com/S0142-9612(14)01072-2/sref2http://refhub.elsevier.com/S0142-9612(14)01072-2/sref2http://refhub.elsevier.com/S0142-9612(14)01072-2/sref2http://refhub.elsevier.com/S0142-9612(14)01072-2/sref2http://refhub.elsevier.com/S0142-9612(14)01072-2/sref2http://refhub.elsevier.com/S0142-9612(14)01072-2/sref3http://refhub.elsevier.com/S0142-9612(14)01072-2/sref3http://refhub.elsevier.com/S0142-9612(14)01072-2/sref3http://refhub.elsevier.com/S0142-9612(14)01072-2/sref3http://refhub.elsevier.com/S0142-9612(14)01072-2/sref3http://refhub.elsevier.com/S0142-9612(14)01072-2/sref4http://refhub.elsevier.com/S0142-9612(14)01072-2/sref4http://refhub.elsevier.com/S0142-9612(14)01072-2/sref4http://refhub.elsevier.com/S0142-9612(14)01072-2/sref4http://refhub.elsevier.com/S0142-9612(14)01072-2/sref4http://refhub.elsevier.com/S0142-9612(14)01072-2/sref5http://refhub.elsevier.com/S0142-9612(14)01072-2/sref5http://refhub.elsevier.com/S0142-9612(14)01072-2/sref5http://refhub.elsevier.com/S0142-9612(14)01072-2/sref5http://refhub.elsevier.com/S0142-9612(14)01072-2/sref6http://refhub.elsevier.com/S0142-9612(14)01072-2/sref6http://refhub.elsevier.com/S0142-9612(14)01072-2/sref6http://refhub.elsevier.com/S0142-9612(14)01072-2/sref6http://refhub.elsevier.com/S0142-9612(14)01072-2/sref7http://refhub.elsevier.com/S0142-9612(14)01072-2/sref7http://refhub.elsevier.com/S0142-9612(14)01072-2/sref7http://refhub.elsevier.com/S0142-9612(14)01072-2/sref8http://refhub.elsevier.com/S0142-9612(14)01072-2/sref8http://refhub.elsevier.com/S0142-9612(14)01072-2/sref8http://refhub.elsevier.com/S0142-9612(14)01072-2/sref8http://refhub.elsevier.com/S0142-9612(14)01072-2/sref9http://refhub.elsevier.com/S0142-9612(14)01072-2/sref9http://refhub.elsevier.com/S0142-9612(14)01072-2/sref9http://refhub.elsevier.com/S0142-9612(14)01072-2/sref9http://refhub.elsevier.com/S0142-9612(14)01072-2/sref10http://refhub.elsevier.com/S0142-9612(14)01072-2/sref10http://refhub.elsevier.com/S0142-9612(14)01072-2/sref10http://refhub.elsevier.com/S0142-9612(14)01072-2/sref10http://refhub.elsevier.com/S0142-9612(14)01072-2/sref11http://refhub.elsevier.com/S0142-9612(14)01072-2/sref11http://refhub.elsevier.com/S0142-9612(14)01072-2/sref11http://refhub.elsevier.com/S0142-9612(14)01072-2/sref12http://refhub.elsevier.com/S0142-9612(14)01072-2/sref12http://refhub.elsevier.com/S0142-9612(14)01072-2/sref12http://refhub.elsevier.com/S0142-9612(14)01072-2/sref12http://refhub.elsevier.com/S0142-9612(14)01072-2/sref13http://refhub.elsevier.com/S0142-9612(14)01072-2/sref13http://refhub.elsevier.com/S0142-9612(14)01072-2/sref13http://refhub.elsevier.com/S0142-9612(14)01072-2/sref13http://refhub.elsevier.com/S0142-9612(14)01072-2/sref13http://refhub.elsevier.com/S0142-9612(14)01072-2/sref14http://refhub.elsevier.com/S0142-9612(14)01072-2/sref14http://refhub.elsevier.com/S0142-9612(14)01072-2/sref14http://refhub.elsevier.com/S0142-9612(14)01072-2/sref15http://refhub.elsevier.com/S0142-9612(14)01072-2/sref15http://refhub.elsevier.com/S0142-9612(14)01072-2/sref16http://refhub.elsevier.com/S0142-9612(14)01072-2/sref16http://refhub.elsevier.com/S0142-9612(14)01072-2/sref16http://refhub.elsevier.com/S0142-9612(14)01072-2/sref17http://refhub.elsevier.com/S0142-9612(14)01072-2/sref17http://refhub.elsevier.com/S0142-9612(14)01072-2/sref17http://refhub.elsevier.com/S0142-9612(14)01072-2/sref18http://refhub.elsevier.com/S0142-9612(14)01072-2/sref18http://refhub.elsevier.com/S0142-9612(14)01072-2/sref18http://refhub.elsevier.com/S0142-9612(14)01072-2/sref19http://refhub.elsevier.com/S0142-9612(14)01072-2/sref19http://refhub.elsevier.com/S0142-9612(14)01072-2/sref19http://refhub.elsevier.com/S0142-9612(14)01072-2/sref19http://refhub.elsevier.com/S0142-9612(14)01072-2/sref20http://refhub.elsevier.com/S0142-9612(14)01072-2/sref20http://refhub.elsevier.com/S0142-9612(14)01072-2/sref20http://refhub.elsevier.com/S0142-9612(14)01072-2/sref21http://refhub.elsevier.com/S0142-9612(14)01072-2/sref21http://refhub.elsevier.com/S0142-9612(14)01072-2/sref21http://refhub.elsevier.com/S0142-9612(14)01072-2/sref22http://refhub.elsevier.com/S0142-9612(14)01072-2/sref22http://refhub.elsevier.com/S0142-9612(14)01072-2/sref22http://refhub.elsevier.com/S0142-9612(14)01072-2/sref22http://refhub.elsevier.com/S0142-9612(14)01072-2/sref23http://refhub.elsevier.com/S0142-9612(14)01072-2/sref23http://refhub.elsevier.com/S0142-9612(14)01072-2/sref23http://refhub.elsevier.com/S0142-9612(14)01072-2/sref23http://refhub.elsevier.com/S0142-9612(14)01072-2/sref24http://refhub.elsevier.com/S0142-9612(14)01072-2/sref24http://refhub.elsevier.com/S0142-9612(14)01072-2/sref24http://refhub.elsevier.com/S0142-9612(14)01072-2/sref25http://refhub.elsevier.com/S0142-9612(14)01072-2/sref25http://refhub.elsevier.com/S0142-9612(14)01072-2/sref25http://refhub.elsevier.com/S0142-9612(14)01072-2/sref25http://refhub.elsevier.com/S0142-9612(14)01072-2/sref26http://refhub.elsevier.com/S0142-9612(14)01072-2/sref26http://refhub.elsevier.com/S0142-9612(14)01072-2/sref26http://refhub.elsevier.com/S0142-9612(14)01072-2/sref26http://refhub.elsevier.com/S0142-9612(14)01072-2/sref27http://refhub.elsevier.com/S0142-9612(14)01072-2/sref27http://refhub.elsevier.com/S0142-9612(14)01072-2/sref27http://refhub.elsevier.com/S0142-9612(14)01072-2/sref27http://refhub.elsevier.com/S0142-9612(14)01072-2/sref28http://refhub.elsevier.com/S0142-9612(14)01072-2/sref28http://refhub.elsevier.com/S0142-9612(14)01072-2/sref28http://refhub.elsevier.com/S0142-9612(14)01072-2/sref28http://refhub.elsevier.com/S0142-9612(14)01072-2/sref28http://refhub.elsevier.com/S0142-9612(14)01072-2/sref29http://refhub.elsevier.com/S0142-9612(14)01072-2/sref29

A. Matsugaki et al. / Biomaterials 37 (2015) 134e143 143

response of mesenchymal stem cells to 330 nm deep microgrooves. J R SocInterface 2008;5:1231e42.

[30] Goffin JM, Pittet P, Csucs G, Lussi JW, Meister JJ, Hinz B. Focal adhesion sizecontrols tension-dependent recruitment of alpha-smooth muscle actin tostress fibers. J Cell Biol 2006;172:259e68.

[31] Matsugaki A, Aramoto G, Nakano T. The alignment of MC3T3-E1 osteoblastson steps of slip traces introduced by dislocation motion. Biomaterials2012;33:7327e35.

[32] Matsugaki A, Isobe Y, Saku T, Nakano T. Quantitative regulation of bone-mimetic, oriented collagen/apatite matrix structure depends on the degreeof osteoblast alignment on oriented collagen substrates. J Biomed Mater Res2014. http://dx.doi.org/10.1002/jbm.a.35189.

[33] Trelstadand RL, Hayashi K. Tendon collagen fibrillogenesis: intracellular sub-assemblies and cell surface changes associated with fibril growth. Dev Biol1979;71:228e42.

[34] Birk DE, Trelstadand RL. Extracellular compartments in tendon morphogen-esis: collagen fibril, bundle, and macroaggregate formation. J Cell Biol1986;103:230e40.

[35] Canty EG, Lu Y, Meadows RS, Shaw MK, Holmes DF, Kadler KE. Coalignment ofplasma membrane channels and protrusions (fibripositors) specifies theparallelism of tendon. J Cell Biol 2004;165:553e63.

[36] Petroll WM, Ma L, Jester JV. Direct correlation of collagen matrix deformationwith focal adhesion dynamics in living corneal fibroblasts. J Cell Sci 2003;116:1481e91.

[37] Wang Y, McNiven MA. Invasive matrix degradation at focal adhesions occursvia protease recruitment by a FAK-p130Cas complex. J Cell Biol 2012;196:375e85.

[38] Segarra M, Vilardell C, Matsumoto K, Esparza J, Lozano E, Serra-Pages C, et al.Dual function of focal adhesion kinase in regulating integrin-induced MMP-2and MMP-9 release by human T lymphoid cells. FASEB J 2005;19:1875e7.

[39] Gardel ML, Schneider IC, Aratyn-Schaus Y, Waterman CM. Mechanical inte-gration of actin and adhesion dynamics in cell migration. Annu Rev Cell DevBiol 2010;26:315e33.

[40] Zimmerman D, Jin F, Leboy P, Hardy S, Damsky C. Impaired bone formation intransgenic mice resulting from altered integrin function in osteoblasts. DevBiol 2000;220:2e15.

http://refhub.elsevier.com/S0142-9612(14)01072-2/sref29http://refhub.elsevier.com/S0142-9612(14)01072-2/sref29http://refhub.elsevier.com/S0142-9612(14)01072-2/sref29http://refhub.elsevier.com/S0142-9612(14)01072-2/sref30http://refhub.elsevier.com/S0142-9612(14)01072-2/sref30http://refhub.elsevier.com/S0142-9612(14)01072-2/sref30http://refhub.elsevier.com/S0142-9612(14)01072-2/sref30http://refhub.elsevier.com/S0142-9612(14)01072-2/sref31http://refhub.elsevier.com/S0142-9612(14)01072-2/sref31http://refhub.elsevier.com/S0142-9612(14)01072-2/sref31http://refhub.elsevier.com/S0142-9612(14)01072-2/sref31http://dx.doi.org/10.1002/jbm.a.35189http://refhub.elsevier.com/S0142-9612(14)01072-2/sref33http://refhub.elsevier.com/S0142-9612(14)01072-2/sref33http://refhub.elsevier.com/S0142-9612(14)01072-2/sref33http://refhub.elsevier.com/S0142-9612(14)01072-2/sref33http://refhub.elsevier.com/S0142-9612(14)01072-2/sref34http://refhub.elsevier.com/S0142-9612(14)01072-2/sref34http://refhub.elsevier.com/S0142-9612(14)01072-2/sref34http://refhub.elsevier.com/S0142-9612(14)01072-2/sref34http://refhub.elsevier.com/S0142-9612(14)01072-2/sref35http://refhub.elsevier.com/S0142-9612(14)01072-2/sref35http://refhub.elsevier.com/S0142-9612(14)01072-2/sref35http://refhub.elsevier.com/S0142-9612(14)01072-2/sref35http://refhub.elsevier.com/S0142-9612(14)01072-2/sref35http://refhub.elsevier.com/S0142-9612(14)01072-2/sref36http://refhub.elsevier.com/S0142-9612(14)01072-2/sref36http://refhub.elsevier.com/S0142-9612(14)01072-2/sref36http://refhub.elsevier.com/S0142-9612(14)01072-2/sref36http://refhub.elsevier.com/S0142-9612(14)01072-2/sref37http://refhub.elsevier.com/S0142-9612(14)01072-2/sref37http://refhub.elsevier.com/S0142-9612(14)01072-2/sref37http://refhub.elsevier.com/S0142-9612(14)01072-2/sref37http://refhub.elsevier.com/S0142-9612(14)01072-2/sref38http://refhub.elsevier.com/S0142-9612(14)01072-2/sref38http://refhub.elsevier.com/S0142-9612(14)01072-2/sref38http://refhub.elsevier.com/S0142-9612(14)01072-2/sref38http://refhub.elsevier.com/S0142-9612(14)01072-2/sref39http://refhub.elsevier.com/S0142-9612(14)01072-2/sref39http://refhub.elsevier.com/S0142-9612(14)01072-2/sref39http://refhub.elsevier.com/S0142-9612(14)01072-2/sref39http://refhub.elsevier.com/S0142-9612(14)01072-2/sref40http://refhub.elsevier.com/S0142-9612(14)01072-2/sref40http://refhub.elsevier.com/S0142-9612(14)01072-2/sref40http://refhub.elsevier.com/S0142-9612(14)01072-2/sref40

Abnormal arrangement of a collagen/apatite extracellular matrix orthogonal to osteoblast alignment is constructed by a nano ...1. Introduction2. Materials and methods2.1. Fabrication of periodic surface structures2.2. Surface characterization2.3. Osteoblast isolation and culture2.4. Fluorescence imaging2.5. Cell orientation2.6. Raman microscopic analysis2.7. Microbeam X-ray diffraction analysis2.8. Transmission electron microscopy2.9. Statistical analysis

3. Results3.1. Surface characterization3.2. Cell arrangement3.3. Anisotropic bone matrix organization

4. Discussion4.1. Periodic surface structure produced by laser irradiation4.2. Osteoblast alignment along the periodic surface structure4.3. Organization of the collagen/apatite matrix orthogonal to osteoblast alignment

5. ConclusionAcknowledgmentsReferences