ACTBIO 3603 No.

of Pages 10, Model 5G

28 February 2015

Acta Biomaterialia xxx (2015) xxx–xxx

1

Contents lists available at ScienceDirect

Acta Biomaterialia
journal homepage: www.elsevier.com/locate/actabiomat

5
6

3 Decellularized periosteum as a potential biologic scaffold for bone tissue

4 engineering q
7 Kai Chen a,1, Xianfeng Lin b,1, Qi Zhang c, Jinhu Ni c, Jianmin Li d, Yang Wang c,
8 Yiheng Ye a, Li Chen e, Keke Jin c,⇑, Lei Chen a,⇑
9 a
Department of Orthopaedics, The First Affiliated Hospital of Wenzhou Medical University, Wenzhou 325000, China
10 b
Department of Orthopaedic Surgery, Sir Run Run Shaw Hospital, Medical College of Zhejiang University, Hangzhou 310016, China
11 c
Department of Pathophysiology, Wenzhou Medical University, Wenzhou 325000, China
12 d
Department of Pathology, The First Affiliated Hospital of Wenzhou Medical University, Wenzhou 325000, China
13 e
The Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne, Melbourne 3010, Australia

14
15
a r t i c l e i n f o a b s t r a c t
1 9
2 7
18 Article history: Bone grafting or bone substitute is typically used to bridge a bone defect that has been caused by trauma, 30
19 Received 24 October 2014 tumor resection, pathological degeneration, or congenital deformations. However, bone graft healing and 31
20 Received in revised form 16 January 2015 remodeling is always a major concern of orthopedic surgeons. Because the periosteum has a remarkable 32
21 Accepted 19 February 2015
regenerative capacity and is widely recognized to be essential for the initiation of bone graft healing and 33
22 Available online xxxx
remodeling, the present study aimed to produce a rabbit decellularized periosteum (D-periosteum) to be 34
used as a biologic scaffold for future bone tissue engineering. We obtained the D-periosteum by employ- 35
23 Keywords:
ing a combination of commonly used decellularization processes, which include physical methods as well 36
24 Decellularization
25 Periosteum
as chemical and enzymatic solutions. The cellular components were effectively removed, and this 37
26 Extracellular matrix removal was demonstrated using current decellularization criteria (H&E staining, DAPI staining, DNA 38
27 Bone tissue engineering quantification and agarose gel electrophoresis); however, there were no significant alterations of the 39
28 native extracellular matrix (ECM) properties (collagen, glycosaminoglycan (GAG), microarchitecture 40
and mechanical properties). Periosteum-derived cells (PDCs) could adhere, proliferate and infiltrate into 41
the D-periosteum in vitro. The allogenic D-periosteum was implanted subcutaneously into the backs of 42
rabbits over 28 days to study the biocompatibility in vivo. The D-periosteum did not elicit a severe 43
immunogenic response. In summary, a biologic scaffold composed of ECM from periosteum has been suc- 44
cessfully developed. The D-periosteum maintains biocompatibility in vitro and in vivo and, therefore, can 45
provide a naturally compatible scaffold for use in future bone tissue engineering. 46
Ó 2015 Published by Elsevier Ltd. on behalf of Acta Materialia Inc. 47
48

49
50
51 1. Introduction and remodeling [2]. Periosteum is structurally divided into two 58
distinct layers: an outer fibrous layer and an inner cambium layer. 59
52 Bone defects caused by trauma, tumor resection, pathological The outer layer is primarily composed of collagen, and the inner 60
53 degeneration, or congenital deformations have always been a layer serves as a reservoir of progenitor cells with osteogenic 61
54 major concern to orthopedic surgeons, and bone grafting or bone potential [3,4]. Periosteal autografts have also exhibited promising 62
55 substitutes are typically used to bridge the bone gap [1]. The results when used to promote bone repair [5–8]. Autografts have 63
56 periosteum has a remarkable regenerative capacity and is widely been the gold standard for repairing bone defects, while their clin- 64
57 recognized to be essential for the initiation of bone graft healing ical application is still limited due to inaccessibility. Therefore, 65
there have been several tissue-engineered attempts to develop 66
flexible cellular constructs that can mimic the periosteal response 67
q
Part of the Biofabrication Special Group, organized by Dr. Dimitrios Zeugolis, to facilitate osteogenic activity of seeded cells, thereby enhancing 68
National University of Ireland, Galway. allograft healing and repair [2,9–11]. 69
⇑ Corresponding authors. Tel.: +86 13868306996; fax: +86 57755578033 (K. Jin). However, no tissue-engineering approach can fabricate the 70
Tel.: +86 13957789595; fax: +86 57755578033 (L. Chen).
unique three-dimensional (3D) microenvironment that completely 71
E-mail addresses: jkk6996@qq.com (K. Jin), chenlei689595@126.com (L. Chen).
contains the periosteum-specific extracellular matrix (ECM) prop- 72
erties that are essential for bone regeneration and reconstruction. 73
1
These two authors contributed equally to this work. Recently, the application of biologic scaffolds composed of ECM 74

http://dx.doi.org/10.1016/j.actbio.2015.02.020
1742-7061/Ó 2015 Published by Elsevier Ltd. on behalf of Acta Materialia Inc.

Please cite this article in press as: Chen K et al. Decellularized periosteum as a potential biologic scaffold for bone tissue engineering. Acta Biomater (2015),
http://dx.doi.org/10.1016/j.actbio.2015.02.020
 ACTBIO 3603 No. of Pages 10, Model 5G
28 February 2015

2 K. Chen et al. / Acta Biomaterialia xxx (2015) xxx–xxx

75 from decellularized tissue has become increasingly popular in 1% SDS (Sigma, USA) for 4 h, were processed with 100 U/ml DNase 133
76 regenerative medicine and tissue engineering both in preclinical I (Sigma, USA) at 37 °C for 12 h, were washed in ultrapure water 134
77 studies and in clinical applications [12]. The decellularization pro- for 2 days (3 exchanges per day) and were rinsed in PBS for 1 day. 135
78 cess has been shown to efficiently remove the antigenic cellular All steps outlined above were performed by submerging the tissue 136
79 components while preserving the native 3D ultrastructure and in the solutions and incubating the samples with agitation 137
80 composition of the ECM [13]. To date, the feasibility of decellular- (120 rpm). The samples were stored in PBS containing 100 U/ml 138
81 ization has been demonstrated in a number of tissues and organs, penicillin and 100 lg/ml of streptomycin (Gibco, USA) at 4 °C. 139
82 such as skeletal muscle [14], adipose tissue [15], intervertebral
83 disks [16], cornea [17], dermis [18], bladder [19,20], heart [21] 2.4. Histological and immunohistochemical (IHC) analysis 140
84 and heart valves [22], lung [23,24], trachea [25,26], liver [27],
85 and blood vessels [28], with various degrees of success. Neverthe- Native periosteum (N-periosteum) (n = 4) and D-periosteum 141
86 less, there has been a lack of genuine attempts at developing a (n = 4) were fixed for 24 h in 10% neutral buffered formalin solu- 142
87 biomimetic scaffold made of ECM that has been derived from the tion in PBS at room temperature, were embedded in paraffin and 143
88 periosteum [29], and none of the previous studies have fabricated were sectioned into 5 lm slices. The sections were deparaffinized, 144
89 and evaluated a decellularized periosteum (D-periosteum). rehydrated and washed in distilled water. The slides underwent 145
90 We postulate that the decellularization can also be applied to histological evaluation using hematoxylin and eosin (H&E) staining 146
91 the periosteum tissue using a combination of commonly used to observe the cellular components and general structure of the D- 147
92 decellularization processes, including physical methods as well as periosteum. Safranin O staining was carried out for a qualitative 148
93 chemical and enzymatic solutions [12]. The objective of the pre- analysis of the GAG. Masson trichrome staining was used to visu- 149
94 sent study was to obtain a D-periosteum for use in future bone tis- alize the collagen distribution and orientation. Polarized light 150
95 sue engineering applications. Firstly, we developed a protocol to microscopy (Abrio Imaging System, USA) was further utilized to 151
96 derive rabbit D-periosteum and provided a detailed characteriza- image the collagen in unstained sample sections. 152
97 tion of the ECM. Secondly, to evaluate the biocompatibility of the IHC staining was performed by placing the slides in 95 °C anti- 153
98 resulting D-periosteum, periosteum-derived cells (PDCs) were gen retrieval citrate buffer in a steamer for 10 min. Endogenous 154
99 integrated within the scaffold in vitro, and the D-periosteum was peroxidases were blocked by incubation with peroxide block, and 155
100 subcutaneously implanted into the backs of rabbits to study the nonspecific binding was blocked with 1% bovine serum albumin. 156
101 host response to the scaffold in vivo. Sections were incubated with primary antibody against type I col- 157
lagen at 4 °C overnight. Following the overnight incubation, the 158
slides were washed three times in PBS. The secondary antibody 159
102 2. Materials and methods
was then applied for 30 min and then subjected to three more 160
washes in PBS, which was followed by treatment with streptavid- 161
103 2.1. Overview of study design
in–horseradish peroxidase complex, diaminobenzidine (DAB) solu- 162
tion, and counterstaining with hematoxylin. The slides were 163
104 Six-month-old female New Zealand White rabbits were used for
dehydrated, mounted, and imaged using Nikon ECLIPSE 80i micro- 164
105 this study. Native periosteum (N-periosteum) was harvested and
scopy (Nikon, Japan). The primary antibody used was mouse anti- 165
106 decellularized by physical, chemical, and enzymatic methods. The
Collagen I antibody (Abcam, Cambridge, MA) at a dilution of 1:50. 166
107 resulting D-periosteum was then evaluated for the removal of cel-
The secondary antibody used was goat anti-mouse IgG1 (Abcam, 167
108 lular components, major biological components and 3D structure
Cambridge, MA) at a dilution of 1:100. 168
109 of the D-periosteum, and biocompatibility in vitro and in vivo
110 (Fig. 1). All animal materials were obtained from the Animal
2.5. DNA assessment and quantification 169
111 Experimental Center of Wenzhou Medical University. All animal
112 experiments were approved by the Animal Experimental Ethics
Both native and decellularized tissue sections were deparaf- 170
113 Committee of Wenzhou Medical University and the animals were
finized, rehydrated and underwent DAPI (Sigma, USA) staining to 171
114 treated according to the approved experimental protocols.
identify the presence of any residual intact nuclei; an absence of 172
staining using a Nikon ECLIPSE 80i microscopy (Nikon, Japan) indi- 173
115 2.2. Periosteum harvesting cated the absence of DNA and, thus, cells. The remaining cells were 174
further quantified by measuring the DNA content. The total genomic 175
116 A total of 70 rabbits were anesthetized using an intravenous DNA in the samples was extracted using a Genomic DNA Extraction 176
117 injection of 3% sodium pentobarbital (30 mg/kg). Under sterile con- Kit (TaKaRa, China) according to the manufacturer’s instructions. 177
118 ditions, the periosteum was harvested from rabbits using the fol- Briefly, both the N-periosteum (n = 8) and the D-periosteum 178
119 lowing surgical procedures. Briefly, the skin and the overlying (n = 8) samples were lyophilized to achieve a constant weight and 179
120 fascia from the anteromedial site of the proximal tibiae of both then were weighed, cut into thin strips and digested with Proteinase 180
121 sides were opened to expose the periosteum. The periosteum tis- K and RNase for 4 h until no visible material remained. The digested 181
122 sue was carefully removed using sharp subperiosteal dissection samples were then centrifuged and purified with two phenol/chlo- 182
123 and then immediately rinsed twice with phosphate-buffered saline roform/isoamyl alcohol (25:24:1 v/v) extractions. The remaining 183
124 (PBS) supplemented with 100 U/ml penicillin and 100 lg/ml of DNA was collected with an elution buffer. The size of the extracted 184
125 streptomycin (Gibco, USA). The harvested tissue was delivered DNA fragments was determined after separation using 2% agarose 185
126 for further fabrication of the D-periosteum and isolation of PDCs. gel electrophoresis. The extracted genomic DNA was also quantified 186
by measuring the absorbance in a spectrophotometer (Bio-Rad, 187
127 2.3. Decellularization process USA). The DNA quantity was normalized to the initial dry weight 188
of the tissue and expressed as ng/mg. 189
128 The samples were frozen at 80 °C and then thawed at 37 °C. This
129 procedure was repeated for 3 cycles, and the samples were then 2.6. Glycosaminoglycan (GAG) content analysis 190
130 washed in ultrapure water overnight at room temperature with
131 agitation (120 rpm). After the freeze–thaw processing, the samples GAG content was determined by the DMMB assay as described 191
132 were then treated with 2% Triton X-100 (Sigma, USA) for 12 h and [30,31]. Briefly, both the N-periosteum (n = 8) and the D-perios- 192

Please cite this article in press as: Chen K et al. Decellularized periosteum as a potential biologic scaffold for bone tissue engineering. Acta Biomater (2015),
http://dx.doi.org/10.1016/j.actbio.2015.02.020
 ACTBIO 3603 No. of Pages 10, Model 5G
28 February 2015

K. Chen et al. / Acta Biomaterialia xxx (2015) xxx–xxx 3

Fig. 1. Schematic flow chart of this study.

193 teum (n = 8) samples were lyophilized to a constant weight, and chondroitin sulfate sodium salt (Sigma, USA). Final values were 201
194 then were digested in papain buffer (125 mg/ml papain, 5 mM cys- expressed as lg of GAG per dry weight. 202
195 teine/HCl, 5 mM disodium EDTA in PBS) at 60 °C for 12 h. Then, the
196 absorbance values of the samples were read at 525 nm on an Epoch 2.7. Collagen content analysis 203
197 Microplate Spectrophotometer (BioTek, USA) immediately on the
198 addition of the 1,9-dimethylene blue solution (Sigma, USA). The The collagen content was determined based on the content of 204
199 amount of GAG content was calculated with reference to a stan- hydroxyproline (Hyp), which was quantified using a Hyp assay 205
200 dard curve made by serial dilutions of different concentrations of kit (Keygen, China) according to the manufacturer’s instructions. 206

Please cite this article in press as: Chen K et al. Decellularized periosteum as a potential biologic scaffold for bone tissue engineering. Acta Biomater (2015),
http://dx.doi.org/10.1016/j.actbio.2015.02.020
 ACTBIO 3603 No. of Pages 10, Model 5G
28 February 2015

4 K. Chen et al. / Acta Biomaterialia xxx (2015) xxx–xxx

207 Briefly, after being lyophilized to a constant weight, both the N-pe- confluence, the cells were trypsinized (0.05% trypsin–0.02% EDTA, 265
208 riosteum (n = 8) and the D-periosteum (n = 8) samples were acid- 37 °C for 5 min), centrifuged (1000 rpm, 5 min) and resuspended 266
209 hydrolyzed with hydrochloric acid (HCl) at 100 °C for 20 min and in the same medium and were maintained in culture at 37 °C in 267
210 neutralized with sodium hydroxide (NaOH). Then, the absorbance a humidified atmosphere with 5% CO2. The floating cells were 268
211 was measured at 570 nm on an Epoch Microplate Spectrophotome- removed, and the medium was exchanged every 3 days. The cells 269
212 ter (BioTek, USA). The Hyp levels were determined based on the were obtained with successive cycles of trypsinization and reseed- 270
213 absorbance of standard samples from the assay kit. The amount ing. Third-passage cells were used in the subsequent experiments. 271
214 of total collagen content per mg dry weight of the samples was cal-
215 culated using a Hyp-to-collagen ratio of 1:7.2. 2.12. Cytotoxicity assay of D-periosteum 272

216 2.8. Scanning electron microscopy (SEM) The cytotoxicity of the D-periosteum was assessed using a 273
method similar to a previous report [31]. Briefly, the D-periosteum 274
217 To qualitatively evaluate the matrix structure after decellular- was incubated in standard DMEM/F12 for 72 h, and the super- 275
218 ization, both the N-periosteum (n = 4) and the D-periosteum natant was collected as an extract for subsequent use. The cells 276
219 (n = 4) were fixed using 3% glutaraldehyde in PBS for 1 h followed were seeded in 96-well cell culture plates with DMEM/F12 at a 277
220 by 0.1 M sodium cacodylate buffer (applied three times for 5 min concentration of 5  103 cells/well. After 24 h of incubation, the 278
221 each). After rinsing in cacodylate buffer, the specimens were dehy- medium was removed and replaced with an extract of varying con- 279
222 drated through an ethanol gradient (35%, 50%, 70%, 80%, 95% and centration (25%, 50% and 100%). Cells cultured in standard medium 280
223 100% ethanol for 10 min each) and were dried in hexam- served as the control group. The metabolic activities of the cells in 281
224 ethyldisilazane (HMDS, Sigma, USA). Subsequently, these dry sam- each group were assessed with the aid of Cell Counting Kit-8 (CCK- 282
225 ples were coated under vacuum with platinum alloy at thickness of 8, DOJINDO, Japan) each day for 7 days. The absorbance was mea- 283
226 25 nm and were immediately flash carbon coated under vacuum. sured at 450 nm with the use of an Epoch Microplate Spectropho- 284
227 The samples were observed with a S-3000N scanning electron tometer (BioTek, USA). Five replicates were conducted per sample. 285
228 microscope (HITACHI, Japan). Additionally, the cell viability after the treatment using standard 286
DMEM/F12 and 100% extract was evaluated using the Live-Dead 287
229 2.9. Water imbibition measurement Cell Staining Kit (BioVision, USA) on days 1, 3 and 7 according to 288
the manufacture’s instruction. Briefly, the adherent cells were 289
230 Water imbibition properties were used to reflect the porosity of washed with PBS and then were exposed to staining solution for 290
231 N-periosteum and D-periosteum. Both N-periosteum (n = 8) and D- 15 min at 37 °C. The cells were observed immediately using a 291
232 periosteum (n = 8) were immersed in PBS at 4 °C for 24 h to achieve Nikon ECLIPSE 80i microscopy (Nikon, Japan) equipped with a 292
233 fully swollen. Then, the samples were lyophilized to a constant band-pass filter. The vital cells appear to be fluorescing green. 293
234 weight, and the weight before and after lyophilizing was mea-
235 sured. The swelling ratio (%) was expressed as (Ws Wd)/Wd. 2.13. Evaluation of PDCs integration into D-periosteum in vitro 294
236 Ws is the weight of swollen samples, and Ws is the weight of lyo-
237 philized samples [31]. The whole piece of D-periosteum was immersed in DMEM/F12 295
containing 10% FBS and 1% antibiotics for 24 h and was dried using 296
238 2.10. Mechanical tests sterile filter paper. Third-passage PDCs were harvested and seeded 297
onto the D-periosteum at a density of 1  106 cells/cm2 in a 24- 298
239 The mechanical properties of the N-periosteum (n = 5) and the well plate. The cell-containing constructs were incubated for 2 h 299
240 D-periosteum (n = 5) were determined using uniaxial tensile tests before the supplemented culture medium was slowly added. The 300
241 performed using a SMT1-50N Force Transoucer (Interface, USA). culture medium was changed every 3 days. Cell viability on the 301
242 Samples with a length of 10 mm were immersed in PBS until test- D-periosteum was assessed using the Live-Dead Cell Staining Kit 302
243 ing began. Samples were clamped to the grips in the mechanical (BioVision, USA). The cellular constructs were stained as described 303
244 apparatus, and the initial specimen width and thickness were above. Cellular constructs were washed with PBS, fixed for 24 h in 304
245 recorded. The samples were then stretched to tensile failure at a 10% neutral buffered formalin solution, embedded in paraffin and 305
246 rate of 10 mm/min. The stress–strain curves were collected. The sectioned into 5 lm slices. The sections were stained with H&E 306
247 ultimate stress (UTS) was calculated by dividing the maximum staining to observe the cell distribution in the D-periosteum. Sam- 307
248 load by the cross-sectional area of the specimen. The strain at fail- ples were examined on day 3, 7, 10 and 14 after cell seeding (n = 6 308
249 ure was calculated by dividing the change in length by the initial at each time point, 3 for Live-Dead Cell Staining and 3 for H&E 309
250 length of the specimen. The elastic modulus (E) was calculated staining). 310
251 from the slope of the ascending linear region of the stress–strain
252 curve. 2.14. Examination of host response to D-periosteum in vivo 311

253 2.11. Isolation and culture of the periosteum-derived cells (PDCs) The D-periosteum samples (n = 20) were implanted subcuta- 312
neously into the backs of a total of 10 six-month-old female New 313
254 The harvested periosteum tissue (n = 6) were immediately cut Zealand white rabbits (each rabbit with 2 samples). The rabbits 314
255 into several small flaps measuring 2  2 mm respectively. The were under general anesthesia achieved using 3% sodium pento- 315
256 minced explants were placed in a 60-mm culture dish with their barbital (30 mg/kg, IV) and all rabbits received an intramuscular 316
257 cambium (bone surface) side downwards and were allowed to injection of antibiotics (Sefazoline, 70 mg/kg) intraoperatively. 317
258 adhere to the tissue culture plastic (10 min, 37 °C) before the cul- Postoperative pain reduction was achieved by administering 318
259 ture medium was added [32]. The periosteal explants were cul- buphenomorphine (0.05 mg/kg) intramuscularly for 2 days. The 319
260 tured in Dulbecco’s modified Eagle’s medium:HAMF12 (DMEM/ animals were housed individually in cages and the housing envi- 320
261 F12, Gibco, USA) containing 10% fetal bovine serum (FBS, Gibco, ronment was maintained at a temperature of 68–76 °F. The wound 321
262 USA) and 100 U/ml penicillin and 100 lg/ml of streptomycin was evaluated daily for the duration of the study, including signs of 322
263 (Gibco, USA) at 37 °C in a humidified atmosphere with 5% CO2. swelling and discoloration. The dietary habits and daily activity 323
264 After migrating out of the explanted tissue and reaching the 90% were also observed to discern any possible abnormalities resulting 324

Please cite this article in press as: Chen K et al. Decellularized periosteum as a potential biologic scaffold for bone tissue engineering. Acta Biomater (2015),
http://dx.doi.org/10.1016/j.actbio.2015.02.020
 ACTBIO 3603 No. of Pages 10, Model 5G
28 February 2015

K. Chen et al. / Acta Biomaterialia xxx (2015) xxx–xxx 5

325 from the implantation. Two rabbits of each group were sacrificed at in D-periosteum, P < 0.05) (Fig. 5A). The content of Hyp was then 383
326 3, 7, 14, 21, and 28 days after surgery. Specimens were cut at the detected to calculate collagen content. The results showed that 384
327 explant sites including adjacent tissue (fibrous encapsulation) the content of Hyp detected in the N-periosteum 385
328 and then fixed in 10% formalin. Slides cut from paraffin-embedded (13.39 ± 1.36 lg/mg) and the D-periosteum (13.00 ± 0.56 lg/mg, 386
329 samples underwent H&E staining. Host response to the scaffold P > 0.05) samples were not significantly different, which demon- 387
330 was evaluated by an experienced pathologist with respect to cellu- strated that the content of collagen in the periosteum 388
331 lar infiltration, the presence of multinucleate giant cells, vascu- (99.96 ± 10.13 lg/mg vs. 96.98 ± 4.18 lg/mg, P > 0.05) did not 389
332 larity, connective tissue organization, encapsulation, and decrease after the decellularization process (Fig. 5B). 390
333 degradation [33].
3.3. Mechanical tests 391
334 2.15. Statistical analysis
Fig. 6 shows two representative uniaxial tensile stress–strain 392
335 Statistical analysis was performed using SPSS (17.0, USA). Data curves for the N-periosteum and the D-periosteum. The mechani- 393
336 were expressed as mean ± standard deviation (SD). Significant dif- cal quality of N-periosteum (E = 55.22 ± 6.78 MPa) decreased com- 394
337 ferences between groups were estimated using one-way ANOVA, pared with the D-periosteum (E = 48.08 ± 6.63 MPa), but these 395
338 followed by Scheffe or Tamhane’s T2 tests for multiple compar- differences were not significantly different (P > 0.05). The UTS 396
339 isons. Differences in DNA, GAG, the collagen content and the and the strain at failure were not obviously affected 397
340 mechanical parameters were analyzed using a Mann–Whitney U (38.98 ± 1.22 MPa vs. 38.12 ± 0.90 MPa and 1.01 ± 0.105 vs. 398
341 test. P values of less than 0.05 were considered significant. 1.02 ± 0.14, respectively, P > 0.05). 399

342 3. Results 3.4. Cytotoxicity analysis of D-periosteum 400

343 3.1. Confirmation of the removal of cellular components in the The cytotoxicity assay was performed to test whether the D-pe- 401

344 D-periosteum riosteum retained toxic chemical residues from the decellulariza- 402
tion processes. The CCK-8 test demonstrated that there was no 403

345 The macroscopic images showed that the N-periosteum and D- significant difference in the cellular metabolic activity between 404

346 periosteum (Fig. 2A and B). The periosteum samples turned from the cells cultured in standard medium as compared with cells cul- 405

347 red to white after the decellularization processes. H&E staining tured in the extracts of different concentrations (P > 0.05) (Fig. 7). 406

348 revealed the absence of cell nuclei in the D-periosteum as com- The Live-Dead cell staining also verified the viability of the cells 407

349 pared with the N-periosteum (Fig. 2C and D). Nuclear material cultured in the extracts (Fig. 8). 408

350 was also not detected by DAPI staining (Fig. 2E and F). There were
351 also no visible bands of DNA after separation on a 2% agarose gel 3.5. Evaluation of seeded PDCs integration with D-periosteum in vitro 409
352 compared with the N-periosteum (Fig. 2G). Moreover, the DNA
353 quantification suggested that more than approximately 95% of Overall, our findings demonstrated that PDCs could grow, pro- 410
354 the nuclear material was eliminated by the decellularization pro- liferate and infiltrate into D-periosteum, which confirmed the bio- 411
355 cesses (683.09 ± 51.61 ng/mg for the N-periosteum and compatibility of D-periosteum. Live-Dead cell staining was utilized 412
356 31.09 ± 4.99 ng/mg for the D-periosteum, P < 0.05) (Fig. 2H). These to demonstrate that D-periosteum could support PDCs survival and 413
357 results all confirmed the significant removal of the cellular proliferation (Fig. 9). PDCs attached well to the D-periosteum after 414
358 components. seeding and proliferated from day 1 to day 14. The cells became 415
more and more confluent over time. Moreover, the H&E staining 416

359 3.2. Major biological components and 3D structure of the D- revealed the dense distribution of PDCs in the inner part of the 417

360 periosteum D-periosteum sample after 14 days of incubation (Fig. 10). 418

361 The H&E staining showed the two-layer structure (cambium 3.6. Evaluation of the host response to implanted D-periosteum in vivo 419
362 layer and fibrous layer) of the periosteum was well preserved
363 (Fig. 2C and D). The SEM images showed the microarchitecture of All implanted samples were encapsulated slightly by small 420
364 cambium layer and fibrous layer of the N-periosteum and the D- amounts of fibrous connective tissue at the periphery from three 421
365 periosteum (Fig. 3A–H). The D-periosteum had an uneven and days until the end of the study. At three days of implantation, 422
366 highly porous surface consisting of long bundles of fibrils with a the host response to the D-periosteum was characterized by an 423
367 network pattern suggestive of collagen. The D-periosteum showed increased infiltration of neutrophil and mononuclear cells at the 424
368 a higher capacity to absorb water than N-periosteum (4.798 ± 0579 tissue interface (Fig. 11A and B). At seven days, the cellular respon- 425
369 vs. 2.757 ± 0.319 mg water/mg sample dry weight respectively, se for D-periosteum consisted primarily of a dense accumulation of 426
370 P < 0.05). Safranin O staining showed the GAG was identifiable both neutrophil and mononuclear cells, with scattered multinucle- 427
371 after decellularization (Fig.4A and B). Masson trichrome staining ate giant cells emerging at the periphery of the scaffolds occasion- 428
372 indicated that the porous structures of the D-periosteum observed ally (Fig. 11C and D). By fourteen days, the vascularity at the 429
373 in SEM images were primarily composed of collagen bundles interface between the fibrous encapsulation and the D-periosteum 430
374 (Fig. 4C and D). Polarized light microscopy provides a unique was very prominent, accompanied by decreased amount of infil- 431
375 opportunity to analyze the alignment and structural form of colla- trated cells throughout the scaffold (Fig. 11E and F). The D-perios- 432
376 gen without using exogenous dyes or labels [34] (Fig. 4E and F). In teum architecture was slightly disrupted at the edges at twenty- 433
377 contrast with the N-periosteum, the D-periosteum revealed that one days and poorly organized fibrous connective tissue deposition 434
378 the distributions of collagen remained alike. The IHC analysis was also noted at this time point (Fig. 11G and H). By twenty-eight 435
379 showed the distribution of type I collagen (Fig. 4G and H). The days, limited scaffold degradation with almost no evidence of con- 436
380 results of the DMMB assay showed the content of GAG decreased structive tissue remodeling was observed and the originally 437
381 significantly after decellularization (4.55 ± 1.17 lg/mg dry weight implanted biomaterials were still distinct (Fig. 11I and J). Unlike 438
382 in N-periosteum compared to the 1.51 ± 0.45 lg/mg dry weight those biologic materials with cellular components eliciting the 439

Please cite this article in press as: Chen K et al. Decellularized periosteum as a potential biologic scaffold for bone tissue engineering. Acta Biomater (2015),
http://dx.doi.org/10.1016/j.actbio.2015.02.020
 ACTBIO 3603 No. of Pages 10, Model 5G
28 February 2015

6 K. Chen et al. / Acta Biomaterialia xxx (2015) xxx–xxx

Fig. 2. Macroscopic images of N-periosteum (A) and D-periosteum (B), and confirmation of the removal of cells by H&E staining (C and D), DAPI staining (E and F), agarose gel
electrophoresis (G) and DNA quantification (H). (A and B) Scale bar 15 mm; (C–F) Scale bar 100 lm.

440 deposition of dense connective tissue and/or scarring [35], the D- sues after decellularization using Triton-X100 [31]. The addition 469
441 periosteum did not elicit a severe immunogenic response. of SDS can make the decellularization more efficient. SDS is typical- 470
ly more effective in cell removal but is associated with greater dis- 471

442 4. Discussion ruption of the ECM ultrastructure as compared with Triton-X 100 472
[19,31,37]. Additionally, DNase I was used to remove DNA frag- 473

443 Tissue decellularization techniques aim to remove almost all ments completely to minimize any adverse immune response. 474

444 the cellular components which may cause immunogenicity while Considering the intrinsic characteristics of periosteum, we utilized 475

445 simultaneously preserving the native ultrastructure and composi- the combination of the aforementioned techniques to decellularize 476

446 tion of the ECM [12]. However, all commonly used decellulariza- periosteum but preserve the ECM relatively well. 477

447 tion processes will inevitably alter the distinct microenvironment The N-periosteum and the resulting D-periosteum were then 478

448 of the ECM and cause some degree of ECM composition deteriora- comprehensively evaluated with respect to the removal of cellular 479

449 tion [12]. Therefore, it is highly desirable to minimize the disrup- components, the retention of GAG and collagen, mechanical prop- 480

450 tion rather than to avoid it completely. In this study, we used erties and biocompatibility in vitro and in vivo. Our results demon- 481

451 current decellularization principles to develop a promising proto- strate that the protocol is efficient in decellularizing the 482

452 col (freeze–thaw processing, Triton-X100, SDS and DNase I) to fab- periosteum. It has been shown that less than complete removal 483

453 ricate a rabbit-derived D-periosteum while minimizing significant of the cell remnants and an increased DNA fragment size are asso- 484

454 alterations to the native ECM properties. ciated with a greater risk of a proinflammatory or immune respon- 485

455 Physical processes used in this study included freeze–thaw pro- se that could elicit an adverse remodeling outcome [38,39]. 486

456 cessing and mechanical agitation. Multiple freeze–thaw processing However, considering the manner in which cells are embedded 487

457 circles can effectively cause periosteum-derived cell lysis with within the surrounding ECM, it is unlikely that complete removal 488

458 minor disruptions of the ECM ultrastructure and mechanical prop- of all the DNA fragments is possible even with the most rigorous 489

459 erties, which facilitates the infiltration of decellularization solu- processing methods [40]. The resulting D-periosteum in our study 490

460 tions to the cells and the removal cellular components [36]. meets the objective criteria for evaluating the efficacy of the decel- 491

461 Mechanical agitation can also assist in the diffusional transport lularization. These objectives included the following: (1) lack of 492

462 of decellularization agents. Both non-ionic (Triton-X100) and ionic visible nuclear material in the tissue sections stained with DAPI 493

463 (SDS) detergents have been utilized in succession to solubilize the and H&E, (2) total quantified DNA of less than 50 ng/mg 494

464 cell membrane in our study. Triton-X100, the most widely used (31.09 ± 4.99 ng/mg in the D-periosteum) dry tissue weight, and 495

465 non-ionic detergent in decellularization processes, is generally (3) DNA fragment sizes smaller than 200 bp [12], which demon- 496

466 thought to have a relatively mild effect on cell removal with minor strated highly efficient cell removal. 497

467 disruption of the ECM [12]. However, the efficacy of Triton-X100 is The preserved ultrastructure and composition of the ECM con- 498

468 controversial as nuclear material has been observed in some tis- tribute to the cell adhesion, migration and proliferation [13]. Our 499

Please cite this article in press as: Chen K et al. Decellularized periosteum as a potential biologic scaffold for bone tissue engineering. Acta Biomater (2015),
http://dx.doi.org/10.1016/j.actbio.2015.02.020
 ACTBIO 3603 No. of Pages 10, Model 5G
28 February 2015

K. Chen et al. / Acta Biomaterialia xxx (2015) xxx–xxx 7

Fig. 5. Effect of decellularization on (A) the GAG content, and (B) collagen content.

flow of the solution within the tissues; this increase in swelling 506
ratio may promote cell penetration deep into the scaffold. The 3D 507
framework of the D-periosteum was composed of type I collagen, 508
which plays a key role in the cascade of events leading to the for- 509
mation of new bone from periosteal cells [41]. Collagen content 510
was also calculated and was found to be not significantly different 511
from the native samples. Additionally, the mechanical properties of 512
Fig. 3. SEM images showed the 3D microarchitecture of cambium layer and fibrous the D-periosteum, which are closely associated with the collagen 513
layer of the N-periosteum (A, C, E and G) and the D-periosteum (B, D, F and H). The content and alignment, were also not significantly affected after 514
D-periosteum had an uneven and porous surface. Scale bar 20 lm. decellularization. The loss of GAG may be due to its high sensitivity 515
to the decellularization solutions. GAG was a determinant for the 516
preservation of biological growth factors in the decellularized tis- 517
500 SEM images showed that the D-periosteum still maintained an
sues [19]. Overall, the resulting D-periosteum retained the major- 518
501 irregular fibrous surface architecture similar to the native architec-
ity of the defining compositions and structures of the N- 519
502 ture, and the collagen integrity remained unruptured. Compared
periosteum. 520
503 with the N-periosteum, a significant increase in the swelling ratio
To test the biocompatibility of the D-periosteum when cultur- 521
504 was observed in the D-periosteum, which could possibly be due to
ing PDCs, we examined the cytotoxicity and PDCs integration with 522
505 the increased porosity because of the removal of cells and to the
the D-periosteum, which were important features of decellularized 523

Fig. 4. Characterization of GAG and collagen distribution in N-periosteum and D-periosteum. Samples were stained with Safranin O to detect GAG (A and B); collagen was
analysis using masson trichrome staining (C and D) and polarized light microscopy (E and F), type I collagen (G and H) was identified using immunohistochemical staining.
Scale bar 100 lm.

Please cite this article in press as: Chen K et al. Decellularized periosteum as a potential biologic scaffold for bone tissue engineering. Acta Biomater (2015),
http://dx.doi.org/10.1016/j.actbio.2015.02.020
 ACTBIO 3603 No. of Pages 10, Model 5G
28 February 2015

8 K. Chen et al. / Acta Biomaterialia xxx (2015) xxx–xxx

Fig. 6. Two representative uniaxial tensile stress–strain curves for the N-perios-
teum and the D-periosteum.

Fig. 10. H&E staining of paraffin sections of D-periosteum seeded with PDCs on
days 3, 7, 10, and 14. Scale bar 50 lm.

scaffolds for tissue engineering. PDCs have excellent osteogenic 524

capacity and, thus, are becoming promising candidates in the field 525
of bone tissue engineering [42,43]. In our study, after decellulariz- 526
ing the periosteum with various chemical and enzymatic solutions, 527
the residual cellular components might be toxic to the cells. The 528
CCK-8 test revealed that the extracts did not affect the cell 529
Fig. 7. Cytotoxicity analysis of D-periosteum. The cellular metabolic activity of the metabolic activity, and the live-dead staining results also con- 530
cells cultured in standard medium and the cells cultured in the extracts of different firmed the viability of PDCs cultured in the extracts and in the 531
concentrations.
standard medium. These results demonstrate that there are no 532

Fig. 8. Fluorescence images showed that PDCs proliferated on days 3 and 7 of culture in standard medium and extracts. The vital cells appeared to fluorescing green. Scale bar
25 lm.

Fig. 9. Fluorescence images showed that D-periosteum could support PDCs survival and proliferation on days 3, 7, 10 and 14. The vital cells appeared to fluoresce green. Scale
bar 30 lm.

Please cite this article in press as: Chen K et al. Decellularized periosteum as a potential biologic scaffold for bone tissue engineering. Acta Biomater (2015),
http://dx.doi.org/10.1016/j.actbio.2015.02.020
 ACTBIO 3603 No. of Pages 10, Model 5G
28 February 2015

K. Chen et al. / Acta Biomaterialia xxx (2015) xxx–xxx 9

Fig. 11. H&E staining showing host response to D-periosteum on days 3 (A and B), 7 (C and D), 14 (E and F), 21 (G and H), and 28 (I and J). Scale bar 50 lm. Images with higher
magnification (40) represent the area within the black box in lower magnification images (20).

533 cytotoxic residual reagents in the D-periosteum after decellulariza- supported the adhesion, proliferation and infiltration of PDCs in vit- 580
534 tion processes and thorough washing with PBS. Furthermore, as ro. Subcutaneous implantation of the D-periosteum further 581
535 the H&E and live-dead staining of the cell-scaffold complex indi- demonstrated its outstanding biocompatibility in vivo. Therefore, 582
536 cated, the PDCs were successfully integrated within the D-perios- this study provides a compatible scaffold that has huge potential 583
537 teum and these cells infiltrated extensively into the scaffold over for future bone tissue engineering. 584
538 the time course of the study. Cell infiltration into the deep of scaf-
539 fold was always thought to be a great challenge but is very sig- Disclosures 585
540 nificant in terms of testing the biocompatibility [15,16]. Despite
541 the slight changes of the ECM in the D-periosteum, the dense dis- No conflicts of interest exist. 586
542 tributions of PDCs within the D-periosteum support the idea that
543 the scaffold in this study maintains the N-periosteum properties Acknowledgments 587
544 that are necessary for recellularization. Due to the necessity for
545 immunogencity evaluation prior to orthotopic transplantation for This study was supported by the Zhejiang Provincial Natural 588
546 use in bone defect repairs, we also studied the host response to Science Foundation of China (Y207495 and LY14H070005). We 589
547 the D-periosteum scaffold after subcutaneous implantation. With- thank the Department of Electron Microscopy of Wenzhou Medical 590
548 out severe graft rejection which was evidenced by tissue necrosis University for SEM imaging, and the Experimental Research Center 591
549 or granulomatous inflammation [40], this bioscaffold presented a of China Academy of Chinese Medical Sciences for polarized light 592
550 low-grade inflammation, slight scaffold degradation, and thin microscopy. 593
551 fibrous encapsulation, which suggested that the D-periosteum
552 did not elicit a severe immunogenic response. Taken together these
Appendix A. Figures with essential color discrimination 594
553 results indicate that the D-periosteum is sufficiently biocompatible
554 when utilized in vitro and in vivo.
Certain figures in this article, particularly Figs. 1–4 and 8–11 are 595
555 There were several limitations in the present study. Firstly, the
difficult to interpret in black and white. The full color images can 596
556 host response to the samples was examined limited in scope to the
be found in the on-line version, at http://dx.doi.org/10.1016/j.act- 597
557 histomorphologic response and more detailed further investigation
bio.2015.02.020. 598
558 about the long-term constructive tissue remodeling is needed.
559 Secondly, despite the resulting D-periosteum providing a com-
560 patible and supportive environment for recellularization based References 599

561 on the current results, there is still no evaluation of regenerative 600

[1] Keating JF, Simpson AH, Robinson CM. The management of fractures with bone
562 capacity in remodeling process. Further studies that evaluate the loss. J Bone Joint Surg Br 2005;87:142–50. 601
563 effects of the D-periosteum scaffold applied to bone defect or bone [2] Zhang X, Awad HA, O’Keefe RJ, Guldberg RE, Schwarz EM. A perspective: 602
engineering periosteum for structural bone graft healing. Clin Orthop Relat Res 603
564 nonunion models with the goal of better evaluating the scaffold’s 604
2008;466:1777–87.
565 functionality and regeneration capacity should be conducted. [3] Squier CA, Ghoneim S, Kremenak CR. Ultrastructure of the periosteum from 605
566 Thirdly, the fabrication and testing of the D-periosteum scaffold membrane bone. J Anat 1990;171:233–9. 606
[4] Lin Z, Fateh A, Salem DM, Intini G. Periosteum: biology and applications in 607
567 should be developed in a large animal model to better translate
craniofacial bone regeneration. J Dent Res 2014;93:109–16. 608
568 into clinical application. In spite of these limitations, the effective [5] Finley JM, Acland RD, Wood MB. Revascularized periosteal grafts – a new 609
569 decellularization protocol in a rabbit model that was developed method to produce functional new bone without bone grafting. Plast Reconstr 610
570 in this study can provide useful information to further decellular- Surg 1978;61:1–6. 611
[6] Reynders P, Becker JH, Broos P. Osteogenic ability of free periosteal autografts 612
571 ize periosteums from other animals or humans due to their similar in tibial fractures with severe soft tissue damage: an experimental study. J 613
572 tissue material properties. Orthop Trauma 1999;13:121–8. 614
[7] Kanou M, Ueno T, Kagawa T, Fujii T, Sakata Y, Ishida N, et al. Osteogenic 615
potential of primed periosteum graft in the rat calvarial model. Ann Plast Surg 616
573 5. Conclusion 2005;54:71–8. 617
[8] Karaoglu S, Baktir A, Kabak S, Arasi H. Experimental repair of segmental bone 618
defects in rabbits by demineralized allograft covered by free autogenous 619
574 We have utilized the combination of physical, chemical and 620
periosteum. Injury 2002;33:679–83.
575 enzymatic decellularization techniques for the preparation of a [9] Beniker D, McQuillan D, Livesey S, Urban RM, Turner TM, Blum B, et al. The use 621
576 D-periosteum scaffold. The cellular components of the periosteum of acellular dermal matrix as a scaffold for periosteum replacement. 622
Orthopedics 2003;26:s591–6. 623
577 were effectively removed but the major extracellular biochemical
[10] Cuthbert RJ, Churchman SM, Tan HB, McGonagle D, Jones E, Giannoudis PV. 624
578 composition and the natural 3D microenvironment were well pre- Induced periosteum a complex cellular scaffold for the treatment of large bone 625
579 served. The resulting D-periosteum scaffold was non-cytotoxic and defects. Bone 2013;57:484–92. 626

Please cite this article in press as: Chen K et al. Decellularized periosteum as a potential biologic scaffold for bone tissue engineering. Acta Biomater (2015),
http://dx.doi.org/10.1016/j.actbio.2015.02.020
 ACTBIO 3603 No. of Pages 10, Model 5G
28 February 2015

10 K. Chen et al. / Acta Biomaterialia xxx (2015) xxx–xxx

627 [11] Zhou Y, Chen F, Ho ST, Woodruff MA, Lim TM, Hutmacher DW. Combined [28] Funamoto S, Nam K, Kimura T, Murakoshi A, Hashimoto Y, Niwaya K, et al. The 673
628 marrow stromal cell-sheet techniques and high-strength biodegradable use of high-hydrostatic pressure treatment to decellularize blood vessels. 674
629 composite scaffolds for engineered functional bone grafts. Biomaterials Biomaterials 2010;31:3590–5. 675
630 2007;28:814–24. [29] Rapp SJ, Jones DC, Gerety P, Taylor JA. Repairing critical-sized rat calvarial 676
631 [12] Crapo PM, Gilbert TW, Badylak SF. An overview of tissue and whole organ defects with progenitor cell-seeded acellular periosteum: a novel biomimetic 677
632 decellularization processes. Biomaterials 2011;32:3233–43. scaffold. Surgery 2012;152:595–604 [605 e591; discussion 604–595]. 678
633 [13] Badylak SF, Freytes DO, Gilbert TW. Extracellular matrix as a biological scaffold [30] Farndale RW, Buttle DJ, Barrett AJ. Improved quantitation and discrimination 679
634 material: structure and function. Acta Biomater 2009;5:1–13. of sulphated glycosaminoglycans by use of dimethylmethylene blue. Biochim 680
635 [14] Wolf MT, Daly KA, Reing JE, Badylak SF. Biologic scaffold composed of skeletal Biophys Acta 1986;883:173–7. 681
636 muscle extracellular matrix. Biomaterials 2012;33:2916–25. [31] Xu H, Xu B, Yang Q, Li X, Ma X, Xia Q, et al. Comparison of decellularization 682
637 [15] Wang L, Johnson JA, Zhang Q, Beahm EK. Combining decellularized human protocols for preparing a decellularized porcine annulus fibrosus scaffold. PLoS 683
638 adipose tissue extracellular matrix and adipose-derived stem cells for adipose ONE 2014;9:e86723. 684
639 tissue engineering. Acta Biomater 2013;9:8921–31. [32] Declercq HA, De Ridder LI, Cornelissen MJ. Isolation and osteogenic 685
640 [16] Chan LK, Leung VY, Tam V, Lu WW, Sze KY, Cheung KM. Decellularized bovine differentiation of rat periosteum-derived cells. Cytotechnology 2005;49: 686
641 intervertebral disc as a natural scaffold for xenogenic cell studies. Acta 39–50. 687
642 Biomater 2013;9:5262–72. [33] Valentin JE, Badylak JS, McCabe GP, Badylak SF. Extracellular matrix 688
643 [17] Ponce Marquez S, Martinez VS, McIntosh Ambrose W, Wang J, Gantxegui NG, bioscaffolds for orthopaedic applications. A comparative histologic study. J 689
644 Schein O, et al. Decellularization of bovine corneas for tissue engineering Bone Joint Surg Am 2006;88:2673–86. 690
645 applications. Acta Biomater 2009;5:1839–47. [34] Oldenbourg R. Polarized light microscopy: principles and practice. Cold Spring 691
646 [18] Eberli D, Rodriguez S, Atala A, Yoo JJ. In vivo evaluation of acellular human Harb Protoc 2013;2013. 692
647 dermis for abdominal wall repair. J Biomed Mater Res A 2010;93:1527–38. [35] Brown BN, Valentin JE, Stewart-Akers AM, McCabe GP, Badylak SF. 693
648 [19] Yang B, Zhang Y, Zhou L, Sun Z, Zheng J, Chen Y, et al. Development of a porcine Macrophage phenotype and remodeling outcomes in response to biologic 694
649 bladder acellular matrix with well-preserved extracellular bioactive factors for scaffolds with and without a cellular component. Biomaterials 695
650 tissue engineering. Tissue Eng Part C Methods 2010;16:1201–11. 2009;30:1482–91. 696
651 [20] Bolland F, Korossis S, Wilshaw SP, Ingham E, Fisher J, Kearney JN, et al. [36] Gilbert TW. Strategies for tissue and organ decellularization. J Cell Biochem 697
652 Biomaterials 2007;28:1061–70. 2012;113:2217–22. 698
653 [21] Ott HC, Matthiesen TS, Goh SK, Black LD, Kren SM, Netoff TI, et al. Perfusion- [37] Woods T, Gratzer PF. Effectiveness of three extraction techniques in the 699
654 decellularized matrix: using nature’s platform to engineer a bioartificial heart. development of a decellularized bone-anterior cruciate ligament–bone graft. 700
655 Nat Med 2008;14:213–21. Biomaterials 2005;26:7339–49. 701
656 [22] Schenke-Layland K, Vasilevski O, Opitz F, Konig K, Riemann I, Halbhuber KJ, [38] Keane TJ, Londono R, Turner NJ, Badylak SF. Consequences of ineffective 702
657 et al. Impact of decellularization of xenogeneic tissue on extracellular matrix decellularization of biologic scaffolds on the host response. Biomaterials 703
658 integrity for tissue engineering of heart valves. J Struct Biol 2003;143:201–8. 2012;33:1771–81. 704
659 [23] Price AP, England KA, Matson AM, Blazar BR, Panoskaltsis-Mortari A. [39] Gilbert TW, Freund JM, Badylak SF. Quantification of DNA in biologic scaffold 705
660 Development of a decellularized lung bioreactor system for bioengineering materials. J Surg Res 2009;152:135–9. 706
661 the lung: the matrix reloaded. Tissue Eng Part A 2010;16:2581–91. [40] Badylak SF, Gilbert TW. Immune response to biologic scaffold materials. Semin 707
662 [24] Cortiella J, Niles J, Cantu A, Brettler A, Pham A, Vargas G, et al. Influence of Immunol 2008;20:109–16. 708
663 acellular natural lung matrix on murine embryonic stem cell differentiation [41] Ryu YM, Hah YS, Park BW, Kim DR, Roh GS, Kim JR, et al. Osteogenic 709
664 and tissue formation. Tissue Eng Part A 2010;16:2565–80. differentiation of human periosteal-derived cells in a three-dimensional 710
665 [25] Baiguera S, Del Gaudio C, Kuevda E, Gonfiotti A, Bianco A, Macchiarini P. collagen scaffold. Mol Biol Rep 2011;38:2887–94. 711
666 Dynamic decellularization and cross-linking of rat tracheal matrix. [42] Castro II S, Zambuzzi WF, de Oliveira Castro L, Granjeiro JM. Periosteal-derived 712
667 Biomaterials 2014;35:6344–50. cells for bone bioengineering: a promising candidate. Clin Oral Implants Res 713
668 [26] Macchiarini P, Jungebluth P, Go T, Asnaghi MA, Rees LE, Cogan TA, et al. Clinical 2012;23:1238–42. 714
669 transplantation of a tissue-engineered airway. Lancet 2008;372:2023–30. [43] Hutmacher DW, Sittinger M. Periosteal cells in bone tissue engineering. Tissue 715
670 [27] Uygun BE, Soto-Gutierrez A, Yagi H, Izamis ML, Guzzardi MA, Shulman C, et al. Eng 2003;9(Suppl. 1):S45–64. 716
671 Organ reengineering through development of a transplantable recellularized 717
672 liver graft using decellularized liver matrix. Nat Med 2010;16:814–20.

Please cite this article in press as: Chen K et al. Decellularized periosteum as a potential biologic scaffold for bone tissue engineering. Acta Biomater (2015),
http://dx.doi.org/10.1016/j.actbio.2015.02.020