A combination graft of low-molecular-weight silk fibroin with

Choukroun platelet-rich fibrin for rabbit calvarial defect

Eui-Hee Lee, DDS,a Jwa-Young Kim, DDS,b Hae Yong Kweon, PhD,c You-Young Jo, PhD,d
Soo-Kee Min, MD, PhD,e Young-Wook Park, DDS, PhD,f Je-Yong Choi, DDS, PhD,g and
Seong-Gon Kim, DDS, PhD,h Kyoungkido, Daegu, and Gangneung, Korea
HALLYM UNIVERSITY, RURAL DEVELOPMENT ADMINISTRATION, KYUNGPOOK NATIONAL
UNIVERSITY, AND GANGNEUNG-WONJU NATIONAL UNIVERSITY

Objective. The objective of this study was to determine the capabilities of silk fibroin as a biomaterial template for
bone formation when mixed with Choukroun platelet-rich fibrin (PRF) in vivo.
Study design. Ten New Zealand white rabbits were used for this study and bilateral round shaped defects were
formed in the parietal bone (diameter 9.0 mm). The silk fibroin was digested by acid and made into powder
(molecular weight ⬍1.0 kDa). The right side (experimental group) received the silk fibroin plus platelet-rich fibroin and
the left side (control group) did not receive a graft. Animals were killed at 6 weeks and 12 weeks. The specimens were
examined by microscopic computerized tomography (␮-CT). Subsequently, they underwent decalcification and were
stained for histologic analysis.
Results. There was no significant difference between groups at 6 weeks after operation. In the ␮-CT results, however,
tissue mineral content in the experimental group at 12 weeks after operation was 132.09 ⫾ 4.41 and that in the
control group was 126.42 ⫾ 6.62 (P ⫽ .011). Tissue mineral density in the experimental group was 2,088.88 ⫾
648.34, and that in the control group was 2,029.72 ⫾ 668.22 (P ⫽ .013). The results of the histomorphometric
analysis were in accordance with the ␮-CT results. The total new bone was 49.86 ⫾ 7.49% in the control group at 12
weeks after the operation and 59.83 ⫾ 10.92% in the experimental group (P ⫽ .021).
Conclusion. A combined application of Choukroun PRF with acid-digested silk fibroin showed more rapid bone
healing than unfilled control. (Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2010;109:e33-e38)

Bony defects can result from pathologic lesions or longed time of operation. Material originating from
surgical procedures. For the reconstruction of bony autogenous blood can be used as filling material for the
defects, autogenous bone graft is the treatment of bony defect instead of bone, in which case donor site
choice. However, it has many limitations, such as its morbidity and operating time can be reduced. Platelet-
limited availability, donor site morbidity, and the pro- rich plasma (PRP) has been widely used for this pur-
pose. It can be used alone or in combination with other
biomaterials.1 However, the preparation of PRP re-
Supported by the BioGreen21 Program (grant nos. 200810FTH010103002
and 200810FTH010102001), Rural Development Administration. quires several steps and chemical additives.1 Chouk-
a
Resident, Department of Oral and Maxillofacial Surgery, Sacred roun platelet-rich fibrin (PRF) is a modification of PRP
Heart Hospital, Hallym University. and can be prepared in a single step without any addi-
b
Instructor, Department of Oral and Maxillofacial Surgery, Sacred
tive.2
Heart Hospital, Hallym University.
c
Senior Researcher, National Academy of Agricultural Science, Rural Fibrin is derived from a form of plasmatic molecule
Development Administration. called fibrinogen. The fibrillary molecule is present in
d
Senior Researcher, National Academy of Agricultural Science, Ru- plasma and plays an important role in platelet aggrega-
ral Development Administration.
e
tion during hemostasis.3 Thrombin acts on fibrinogen to
Associate Professor, Department of Pathology, Sacred Heart Hospi-
tal, Hallym University. form 1 molecule of fibrin monomer that has the ability
f
Professor, Department of Oral and Maxillofacial Surgery, Gan- to polymerize with other fibrin monomer molecules to
gneung-Wonju National University. form fibrin fiber.4 The addition of fibrin sealant with
g
Professor, Department of Biochemistry and Cell Biology, Kyung- resorptive ceramic biomaterial increases bone forma-
pook National University.
h
Assistant Professor, Department of Oral and Maxillofacial Surgery, tion in the bony defects of rabbits’ tibia.5 Fibrin adhe-
Gangneung-Wonju National University. sive can be used to maintain bone graft fragments in a
Received for publication Aug 18, 2009; returned for revision Dec 9, mass.6 However, the fibrin without bone substitute ma-
2009; accepted for publication Dec 28, 2009.
terial cannot increase bone regeneration and even in-
1079-2104/$ - see front matter
© 2010 Mosby, Inc. All rights reserved. hibits normal healing.7 This fibrin can inhibit bony
doi:10.1016/j.tripleo.2009.12.043 integration into some biomaterials, such as Bio-Oss.8

e33
 OOOOE
e34 Lee et al. May 2010

Fig. 1. A, Scanning electron microscopic image of silk powder. Because the silk fibroin macromolecule was randomly degraded
by acid, the size of each particle was different. B, The blood was sampled from the ear of the rabbit. C, After centrifugation, the
blood was separated in 3 layers (arrows). D, A combination of silk powder and platelet-rich fibroin was placed on the right side
(asterisk), and the left side was left unfilled.

Therefore, caution must be exercised when selecting Institutional Animal Care and Use Committee of the Bio-
the biomaterial to combine with fibrin. venture Incubation Center, Hanbat National University,
Silk is composed of protein polymers that are produced Daejeon, Korea (no. 2009-NCT-004). The low-molec-
by spiders9 and larvae such as silkworms.10 Silk protein ular-weight silk fibroin powder (molecular weight 0.5-
has a repetitive protein sequence consisting of 4 different 1.0 kDa) was prepared by the Rural Development Ad-
structural components that lead to homogeneity in the ministration (Suwon, Korea) and kindly donated for
secondary structure.11 Silk fibroin has been used in tissue this experiment (Fig. 1, A). Briefly, silk fibroin macro-
engineering as a scaffold for the repair of ligament, ten- molecules without sericin were dissolved in hydrochlo-
don, bone, and cartilage. Although silk may induce in- ric acid, and an electrodialysis system was used to
flammatory and immunogenic reactions, it can safely be remove salt. Then it was made into a powder.
used when it is prepared without glycosylated protein.12
As a scaffold, it delivers inductive cells to the healing site Surgical method
and provides cues to control the structure of newly formed General anesthesia was induced by intramuscular
tissue.13 Silk fibroin– based scaffold shows good biocom- injection of a combination of 0.4 mL ketamine (100
patibility,14 slow degradation,15 and excellent mechanical mg/mL; Ketara; Yuhan, Seoul, Korea) and 0.3 mL
properties.16 Silk is insoluble in organic solvent.15 Be- xylazine (10 mg/kg body weight; Rompun; Bayer Ko-
cause silk fibroin is an excellent scaffold, silk fibroin may rea, Seoul, Korea). Additionally, a 10-mL blood sample
be used as a scaffold for Choukroun PRF. was obtained through the ear vein of each rabbit (Fig. 1,
The objective of the present study was to determine the B). The sample was then centrifuged for 12 minutes at
capability of silk fibroin as a biomaterial template for bone a rate of 400g. After centrifugation, the blood was
formation when mixed with Choukroun PRF in vivo. separated into 3 layers (Fig. 1, C). Among these, the
middle layer, representing Choukroun PRF, was taken.
MATERIAL AND METHOD The cranium area was shaved and disinfected with
Animals and materials povidine-iodine. From the nasal bone to the occipital
Ten 3-month-old New Zealand white rabbits with an protuberance, a longitudinal incision was made in the
average weight of 2.3 kg (range 2.0-2.5 kg) were used in skull. Then a midline incision was created in the peri-
this experiment. This experiment was approved by the osteum. Sharp subperiosteal dissection reflected the
OOOOE
Volume 109, Number 5 Lee et al. e35

Table I. Microscopic computerized tomography analysis

6 weeks 12 weeks
Unfilled Silk fibroin ⫹ PRF P value Unfilled Silk fibroin ⫹ PRF P value
BMC (mg) 159.57 ⫾ 25.50 181.01 ⫾ 3.91 NS 271.25 ⫾ 20.69 279.37 ⫾ 14.82 NS
BMD (mg/mm3) 1,059.66 ⫾ 166.69 1,202.46 ⫾ 28.99 NS 1,467.61 ⫾ 480.28 1,550.46 ⫾ 415.06 NS
TMC 60.19 ⫾ 20.26 72.80 ⫾ 3.40 NS 126.42 ⫾ 6.62 132.09 ⫾ 4.41 .011
TMD 1,867.98 ⫾ 147.12 1,989.24 ⫾ 117.30 NS 2,029.72 ⫾ 668.22 2,088.88 ⫾ 648.34 .013
PRF, Platelet-rich-fibroin; BMC, bone mineral content; BMD, bone mineral density; TMC, tissue mineral content; TMD, tissue mineral density;
NS, not significant.

pericranium from the outer table of the cranial valut, Histomorphometric evaluation
exposing the parietal bones. A dental-trephine bur was After radiodensitometry analysis, the calvaria were
used under copious saline irrigation to create a bilateral subjected to dehydration and embedding. The bones
full-thickness calvarial defect. Two 9-mm-diameter de- were dehydrated in ethanol and decalcified using 5%
fects were created, one on each side of the midline. The formic acid for 2 weeks. The right and left parietal
silk powder mixed with Choukroun PRF was placed on bones were separated through the midline sagittal su-
the right side, and the left side was left empty (Fig. 1, ture. Both segments were embedded to show the sag-
D). The experimental designs of previously published ittal sections in paraffin blocks. Then the sections were
studies were referenced in devising the present exper- sliced and stained with hematoxylin and eosin. The
iment.17,18 Then the pericranium and skin were closed section showing the widest defect area was selected and
in layers with 3-0 silk. After surgery, the rabbits re- the cuts from 50 ␮im before and after were also se-
ceived gentamicin 1 mg/kg (Kookje, Seoul, Korea) lected. Digital images of selected sections were taken
intramuscularly 3 times daily for 3 days. Each rabbit using a digital camera (DP-20; Olympus, Tokyo, Ja-
was individually caged and received food and water. pan). The images were analyzed by Sigma Scan Pro
Previously published studies were referenced to deter- (SPSS, Chicago, IL). Total new bone was calculated as
mine the observation point.19,20 Five animals were a percentage of the total region of the defect.
killed at 6 weeks and 5 at 12 weeks. The dimension of
the calvarial specimens was 25 ⫻ 12 ⫻ 3 mm at the Statistical analysis
largest including both defects. They were fixed in 10% A paired t test was used for comparison of the
formalin and underwent microscopic computerized to- samples from the same animal. The statistically signif-
mography (␮-CT). icant level was set as P ⬍ .05.

Microscopic computerized tomography RESULTS

The prepared specimens underwent ␮-CT using an Microscopic computerized tomography
Explored Locus SP ␮-CT scanner (GE Medical Sys- The results of the ␮-CT analysis are shown in Table
tems, London, Canada). After the calibration, the cal- I. The average value of all measured variables was
varial specimens were scanned in sections of 0.05 mm higher in the experimental group than in the control
thickness. The scanned images were reconstructed by group at 6 weeks after the operation (Fig. 2, A). How-
Microview software (GE Medical Systems). The cali- ever, there was no statistically significant difference
brated 3-dimensional images were shown in the gross (P ⬎ .05). The size of the remaining defect was larger
profiles of the specimens. Because the initial defect was in the control group than in the experimental group
round in shape with 9.0 mm diameter, the setting of the (Fig. 2, B). The TMC in the experimental group at 12
region of interest (ROI) was considered to be the initial weeks after operation was 132.09 ⫾ 4.41 and that in the
defect size and shape. A threshold level of 25% of the control group was 126.42 ⫾ 6.62 (Table I). This was a
bone standard was set as recommended by the manu- statistically significant difference (P ⫽ .011). The TMD
facturer. in the experimental group at 12 weeks after operation
The ROI of each specimen was analyzed for bone was 2,088.88 ⫾ 648.34 and that in the control group
mineral content (BMC) and bone mineral density was 2,029.72 ⫾ 668.22 (Table I). This was a statisti-
(BMD). The tissue mineral content (TMC) and tissue cally significant difference (P ⫽ .013). The other vari-
mineral density (TMD) were calculated using appropri- ables failed to show a statistically significant difference
ate software. (P ⬎ .05).
 OOOOE
e36 Lee et al. May 2010

Fig. 2. Microscopic computerized tomography. A, The defect size was smaller in the experimental group (asterisk) than in the
control group at 6 weeks. B, Most bony defects were filled by new bone at 12 weeks. The relative size of the defects was small
in the experimental group (asterisk).

Table II. Histomorphometric analysis

6 weeks 12 weeks
Unfilled Silk fibroin ⫹ PRF P value Unfilled Silk fibroin ⫹ PRF P value
Total new bone (%) 36.59 ⫾ 6.11 44.38 ⫾ 17.00 NS 49.86 ⫾ 7.49 59.83 ⫾ 10.92 .021
PRF, Platelet-rich-fibroin; NS, not significant.

Fig. 3. Histologic view at 6 weeks. The size of the unfilled defect was larger in the control group (A) than in the experimental
group (B). C, Magnified image of the control group (original magnification ⫻200). D, The experimental groups showed new bone
formation in the middle of the defect (original magnification ⫻200). Bar length ⫽ 2 mm.

Histomorphometry at 12 weeks after the operation (Fig. 4, B and D). The

The results of histomorphometry are shown in Table difference was statistically significant (P ⫽ .021).
II. The total new bone was 36.59 ⫾ 6.11% in the
control group at 6 weeks after the operation (Fig. 3, A DISCUSSION
and C). It was 44.38 ⫾ 17.00% in the experimental In this study, silk fibroin powder was used as a
group at 6 weeks after the operation (Fig. 3, B and D). combination template with Choukroun PRF for the
However, the difference was not statistically significant restoration of bony defect. Since fibrin without bone
(P ⬎ .05). The total new bone was 49.86 ⫾ 7.49% in the substitute material cannot increase bone regeneration
control group at 12 weeks after the operation (Fig. 4, A and even inhibits normal healing,7 the silk powder can
and C). It was 59.83 ⫾ 10.92% in the experimental group be considered a scaffold for Choukroun PRF.
OOOOE
Volume 109, Number 5 Lee et al. e37

Fig. 4. Histologic view at 12 weeks. The size of the unfilled defect was larger in the control group (A) than in the experimental
group (B). C, Magnified image of the control group showed dense fibrotic tissue (original magnification ⫻200). D, The
experimental groups showed narrow defects filled with loose fiber (original magnification ⫻200). Bar length ⫽ 2 mm.

Silk fibroin is produced by Bombyx mori and has control group (P ⫽ .011; Table I). The TMD was also
been widely studied as a tissue scaffold.12-14 The ad- significantly higher in the experimental group than in
vantages of the silk fibroin are that it is less expensive the control group (P ⫽ .013; Table I). The results were
and has good biocompatibility, slow degradation, and further confirmed by histomorphometric analysis (Ta-
plasticity.21,22 These advantages are highly important ble II). The total new bone was 49.86 ⫾ 7.49% in the
for the regeneration of connective tissue. However, control group at 12 weeks after the operation and
some features, such as slow degradation, may have an 59.83 ⫾ 10.92% in the experimental group (P ⫽ .021).
adverse effect on bone regeneration. It is also known to The ␮-CT analysis was highly correlated to the histo-
induce inflammatory reaction.23,24 For the successful morphometric analysis.29 Our study also showed a
regeneration of bone, degradation velocity and inflam- close correlation between the two analyses.
mation should be controlled. Silk fibroin is a macromolecule and can be degraded
The silk structure is modified in an attempt to over- by enzymes in vivo.30 Though the degradation time of
come the limitations of silk as a bone substitute. When silk is reported differently, it takes 2 years for the
silk fibroin film is chemically mixed with RGD peptide, complete loss of tensile strength.31 During the degra-
bone formation in vitro is significantly improved.25,26 dation, silk induces an immunogenic reaction mediated
The addition of bone morphogenetic protein 2 and by lymphocytes and multinucleated giant cells.23,24
nanohydroxyapatite to the silk fibroin scaffold can sig- Biodegradation is the breakdown of macromolecules
nificantly increase bone formation.13 Thin silk fibroin into small fragments.30 Because silk fibroin fragments
film scaffolds with stem cells presented new bone for- were used in the present study, the presence of an
mation.27 Although silk fibroin can induce the activity immunogenic reaction was hardly identified in the his-
of alkaline phosphatase, the silk fibroin does not have tologic exam. The silk fibroin fragment seemed to be
enough osteogenic power to induce new bone without degraded completely within 6 weeks. The inflammatory
any modification.28 reaction was minimal, unlike for the silk fibroin mac-
Choukroun PRF was produced from the patient. romolecule. Our silk powder increased the alkaline
Therefore, there were no immunologic or infection- phosphatase activity in proportion to the applied dose in
related concerns. The amount of Choukroun PRF may the osteoblast-like cells (data not shown). Therefore, in
have been limited because of the limited sampling size. terms of inflammatory potential and osteoinduction, our
When the defect size was relatively large, Choukroun silk powder was more appropriate as a scaffold for bone
PRF had to be used with other filling materials. Ac- regeneration than the silk macromolecule was. Because
cording to the ␮-CT results, silk fibroin powder was the presented product was a powder, its application
used as a combination template with Choukroun PRF might be limited to cavity-form deformities. Ideal bio-
(Table I). It was revealed that the TMC in the experi- degradability as a scaffold for bone regeneration is
mental group was significantly higher than that of the largely unknown. Degradation that is too rapid may not
 OOOOE
e38 Lee et al. May 2010

induce new bone formation, whereas degradation that is 13. Hirano Y, Mooney DJ. Peptide and protein presenting materials
too slow may prevent the space occupied by the graft for tissue engineering. Adv Mater 2004;16:17-25.
14. Dal Pra I, Freddi G, Minic J, Chiarini A, Armato U. De novo
from being changed to bone. The biodegradability of engineering of reticular connective tissue in vivo by silk fibroin
silk fibroin might be controlled by the length of the silk nonwoven materials. Biomaterials 2005;26:1987-99.
fibroin fragment.30 However, the relationship between 15. Horan RL, Antle K, Collette AL, Wang Y, Huang J, Moreau
biodegradability and osteoinduction is still unclear. JE, et al. In vitro degradation of silk fibroin. Biomaterials
2005;26:3385-93.
Most of the osteogenic fragments of the digested silk
16. Gosline JM, Demont ME, Denny MW. The structure and prop-
fibroin will be studied in future research. erties of spider silk. Endeavour 1986;10:37-43.
In conclusion, a combined application of Choukroun 17. Mueller AA, Rahn BA, Gogolewski S, Leiggener CS. Early dural
PRF with the acid-digested silk fibroin showed more reaction to polylactide in cranial defects in rabbits. Pediatr Neu-
rapid bone healing than the unfilled control. Because rosurg 2005;41:285–21.
18. Nair MK, Nair UP, Seyedain A, Gassner R, Piesco N, Mooney
silk is a cheap and readily available material, silk M, et al. Correlation of tuned aperture computed tomography
fibroin and Choukroun PRF combination therapy could with conventional computed tomography for evaluation of osse-
provide a possible new bone substitute for the recon- ous healing in calvarial defects. Oral Surg Oral Med Oral Pathol
struction of various bony defects. However, results Oral Radiol Endod 2007;103:267-73.
from these studies on rabbits cannot be directly extrap- 19. Haddad AJ, Peel SA, Clokie CM, Sándor GK. Closure of rabbit
calvarial critical-sized defects using protective composite allogeneic
olated to human application, so further study is needed. and alloplastic bone substitutes. J Craniofac Surg 2006;17:926-34.
20. Moghadam HG, Sandor GK, Holmes HH, Clokie CM. Histomor-
phometric evaluation of bone regeneration using allogeneic and
REFERENCES
alloplastic bone substitutes. J Oral Maxillofac Surg 2004;62:202-13.
1. Kim SG, Kim WK, Park JC, Kim HJ. A comparative study of
21. Stitzel J, Liu J, Lee SJ, Komura M, Berrya J, Sokerc S, et al.
osseointegration of Avana implants in a demineralized freeze-
Controlled fabrication of a biological vascular substitute. Bio-
dried bone alone or with platelet-rich plasma. J Oral Maxillofac
materials 2006:27;1088-94.
Surg 2002;60:1018-25.
22. Murugan R, Ramakrishna S. Development of nanocomposites
2. Soffer E, Ouhayoun JP, Anagnostou F. Fibrin sealants and plate-
for bone grafting. Compos Sci Technol 2005;65:2385-406.
let preparations in bone and periodontal healing. Oral Surg Oral
23. Rossitch EJr, Bullard DE, Oakes WJ. Delayed foreign-body
Med Oral Pathol Oral Radio Endod 1998;85:638-46.
reaction to silk sutures in pediatric neurosurgical patients. Childs
3. Dohan MD, Choukroun J, Diss A, Dohan SL, Dohan AJ, Mouhyi Nerv Syst 1987;3:375-8.
J, et al. Platelet-rich fibrin (PRF): a second-generation platelet 24. Soong HK, Kenyon KR. Adverse reactions to virgin silk sutures
concentrate. Part I: technological concepts and evolution. Oral in cataract surgery. Ophthalmology 1984;91:479-83.
Surg Oral Med Oral Pathol Oral Radiol Endod 2006;101:e37-44. 25. Zhao J, Zhang Z, Wang S, Sun X, Zhang X, Chen J, et al.
4. Mannaioni PF, Di Bello MG, Masini E. Platelets and inflamma- Apatite-coated silk fibroin scaffolds to healing mandibular bor-
tion: role of platelet-derived growth factor, adhesion molecules der defects in canines. Bone 2009;45:517-27.
and histamine. Inflamm Res 1997;46:4-18. 26. Li C. Electrospun silk-BMP-2 scaffolds for bone tissue engineer-
5. Kania RE, Meunier A, Hamadouche M, Sedel L, Petite H. ing. Biomaterials 2006;27:3115-24.
Addition of fibrin sealant to ceramic promotes bone repair: 27. Wang X, Kim HJ, Xu P, Matsumoto A, Kaplan DL. Biomaterial
long-term study in rabbit femoral defect model. J Biomed Mater coating by stepwise deposition of silk fibroin. Langmuir 2005;21:
Res 1998;43:38-45. 11335-41.
6. Tayapongsak P, O’Brien DA, Monteiro CB, Arceo-Diaz LY. 28. Cai K, Yao K, Lin S, Yang Z, Li X, Xie H, et al. Poly(D, L-lactic
Autologus fibrin adhesive in mandibular reconstruction with acid) surfaces modified by silk fibroin: effects on the culture of
particulate cancellous bone and marrow. J Oral Maxillofac Sug osteoblast in vitro. Biomaterials 2002;23:1153-60.
1994;52:161-5. 29. Buchman SR, Sherick DG, Goulet RW, Goldstein SA. Use of
7. Jung RE, Schmoekel HG. Platelet-rich plasma and fibrin as delivery microcomputed tomography scanning as a new technique for the
system for rhBMP-2. Clin Oral Implants Res 2005;16:676-82. evaluation of membranous bone. J Craniofac Surg 1998;9:48-54.
8. Carmagnola D, Berglundh T, Lindhe J. The effect of a fibrin glue 30. Cao Y, Wang B. Biodegradation of silk biomaterials. Int J Mol
on the integration of Bio-Oss with bone tissue. An experimental Sci 2009;10:1514-24.
study in labrador dogs. J Clin Periodontol 2002;29:377-83. 31. Postlethwait RW. Tissue reaction to surgical sutures. In: Dum-
9. Kaplan DL, Fossey S, Viney C, Muller W. Self-organization (as- phy JE, van Winkle W, editors. Repair and regeneration. New
sembly) in biosynthesis of silk fibers—a hierarchical problem. In: York: McGraw-Hill; 1969. p. 263-85.
Aksay IA, Baer E, Sarikaya M, Tirrell DA, editors. Hierarchically
structured materials. Mater Res Symp Proc 1992;255:19-29. Reprint requests:
10. Bai J, Ma T, Chu W, Wang R, Silva L, Michal C, et al. Dr. S. G. Kim
Regenerated spider silk as a new biomaterial for MEMS. Biomed Department of Oral and Maxillofacial Surgery
Microdevices 2006;8:317-23. College of Dentistry
11. Craig CL, Riekel C. Comparative architecture of silks, fibrous Gangneung-Wonju National University
proteins, and their encoding genes in insects and spiders. Comp Jibyun-Dong, Gangneung
Biochem Physiol B Biochem Mol Biol 2003;135:721-2. Gangwondo, 210-702
12. Lorena M, Robert F. Silk implants for the healing of critical size Korea
bone defect. Bone 2005;37:688-98. [email protected]