Silk-based biomaterials - Biomedical...
Transcript of Silk-based biomaterials - Biomedical...
Biomaterials 24 (2003) 401–416
Silk-based biomaterials
Gregory H. Altman, Frank Diaz, Caroline Jakuba, Tara Calabro, Rebecca L. Horan,Jingsong Chen, Helen Lu, John Richmond, David L. Kaplan*
Department of Chemical and Biological Engineering, Bioengineering Center, Tufts University, 4 Colby Street, Medford, MA 02155, USA
Received 17 April 2002; accepted 19 June 2002
Abstract
Silk from the silkworm, Bombyx mori, has been used as biomedical suture material for centuries. The unique mechanical
properties of these fibers provided important clinical repair options for many applications. During the past 20 years, some
biocompatibility problems have been reported for silkworm silk; however, contamination from residual sericin (glue-like proteins)
was the likely cause. More recent studies with well-defined silkworm silk fibers and films suggest that the core silk fibroin fibers
exhibit comparable biocompatibility in vitro and in vivo with other commonly used biomaterials such as polylactic acid and
collagen. Furthermore, the unique mechanical properties of the silk fibers, the diversity of side chain chemistries for ‘decoration’
with growth and adhesion factors, and the ability to genetically tailor the protein provide additional rationale for the exploration of
this family of fibrous proteins for biomaterial applications. For example, in designing scaffolds for tissue engineering these
properties are particularly relevant and recent results with bone and ligament formation in vitro support the potential role for this
biomaterial in future applications. To date, studies with silks to address biomaterial and matrix scaffold needs have focused on
silkworm silk. With the diversity of silk-like fibrous proteins from spiders and insects, a range of native or bioengineered variants
can be expected for application to a diverse set of clinical needs.
r 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Silk; Sericin; Fibroin; Foreign body response; Suture; Tissue engineering; Biomaterial; Scaffold
1. Background
Silks are generally defined as protein polymers thatare spun into fibers by some lepidoptera larvae such assilkworms, spiders, scorpions, mites and flies [1–3]. Silkproteins are usually produced within specialized glandsafter biosynthesis in epithelial cells, followed by secre-tion into the lumen of these glands where the proteinsare stored prior to spinning into fibers. Silks differwidely in composition, structure and properties depend-ing on the specific source. The most extensivelycharacterized silks are from the domesticated silkworm,Bombyx mori, and from spiders (Nephila clavipes andAraneus diadematus). Many of the more evolutionarilyadvanced spiders synthesize different types of silks. Eachof these different silks has a different amino acidcomposition and exhibits mechanical properties tailored
to their specific functions: reproduction as cocooncapsular structures, lines for prey capture, lifelinesupport (dragline), web construction and adhesion.
Fibrous proteins, such as silks and collagens, arecharacterized by a highly repetitive primary sequencethat leads to significant homogeneity in secondarystructure, i.e., triple helices in the case of collagens andb-sheets in the case of many of the silks. These types ofproteins usually exhibit important mechanical proper-ties, in contrast to the catalytic and molecular recogni-tion functions of globular proteins. Because of theseimpressive mechanical properties, this family of proteinsprovides an important set of material options in thefields of controlled release, biomaterials and scaffoldsfor tissue engineering. The relative environmentalstability of these families of proteins, in comparison toglobular proteins, in combination with their biocompat-ibility, unique mechanical properties, and options forgenetic control to tailor sequence provides an importantbasis to exploit these natural proteins for biomedicalapplications.
*Corresponding author. Tel.: +1-617-627-3251; fax: +1-617-627-
3231.
E-mail address: [email protected] (D.L. Kaplan).
0142-9612/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 4 2 - 9 6 1 2 ( 0 2 ) 0 0 3 5 3 - 8
1.1. Silkworm silk
Silkworm silk has been used commercially as biome-dical sutures for decades, and in textile production forcenturies. The silk from the cocoon of B. mori containsat least two major fibroin proteins, light and heavychains, 25 and 325 kDa, respectively. These core fibersare encased in a sericin coat, a family of glue-likeproteins that holds two fibroin fibers together to formthe composite fibers of the cocoon case to protect thegrowing worm. This structural arrangement contrastswith spider silks where these glue-like proteins aregenerally absent. Silkworm cocoon silk production,known as sericulture, produces high yields since thelarvae can be maintained in high densities. The coresequence repeats in the fibroin heavy chain from B. mori
include alanine–glycine repeats with serine or tyrosine.
1.2. Spider silks
Spider silks have not been commercialized forbiomedical applications primarily due to the predatorynature of spiders and the relatively low levels ofproduction of these silks when compared to silkwormcocoon silk. The molecular weights of spider silkproteins vary depending on source but can range from70 to 700 kDa depending on the method of analysis. Thedragline silk from N. clavipes is characterized bypolyalanine and glycine–glycine-X regions, where X isoften tyrosine, glutamine or leucine. Genetic engineeringis being actively explored to construct, clone and expressnative and synthetic genes encoding recombinant spidersilk proteins to overcome limits to use of the nativeorganisms. This strategy has provided new opportunitiesin fundamental studies of spider silk genetics, silkprotein structure and function, and materials processing[2,4–6]. In general, interest in spider silk has increased inrecent years due to the differences in mechanicalproperties when compared to silkworm silk and thepresence of the multi-gene family encoding this group ofsilks as a basis for the study of protein structure-function relationships [4,6–11].
During the past 10 years a great deal of progress hasbeen made in understanding silk genetic and proteinstructures. Cloning and expression of native andsynthetic silks has been achieved in a variety of hostsystems. The sequences of cDNAs and genomic clonesencoding spider silks illustrate the highly repetitivestructures [4,6,8–11], which can be readily exploited toconstruct genetically engineered spider silk-like proteinsusing synthetic oligonucleotide versions of the consensusrepeats or variants of these repeats. This highlyrepetitive sequence was also recently fully described forthe fibroin heavy chain from the silkworm, B. mori [12].The unique organization of the family of silk genes inmore advanced spiders provides a fertile area for the
exploration of structure-function relationships in pro-tein design. For example, the dragline spider silk fromthe golden orb weaver N. clavipes displays impressivetoughness, and a balance of stiffness, strength andextensibility reflecting the native function of the silk orbweb construction [7,13,14]. Transgenic expression ofspider silks in plants (tobacco and potato) andmammalian epithelial cells has been reported [15,16]and may point the way toward more substantiveproduction of these proteins in the future.
Furthermore, since native and genetically engineeredversions of the silks tend to self-assemble into micro-fibrils, causing precipitation and leading to loss ofprotein during purification [3,17], alternative methods tomaintain solubility of the recombinant versions of theseproteins have been explored. For example, tweakingthe native or consensus sequences of spider silks toinclude encoded ‘triggers’ to regulate molecular-levelassembly of the proteins has been demonstrated [17–21].Chemical redox triggers, or a kinase recognition site forbiochemical phosphorylation, flanking the b-sheet form-ing regions of the proteins have been demonstrated.These modifications encoded in the synthetic genedesigns enabled improved control over solubility of thegenetically engineered silk protein.
1.3. Properties of silk
The enhanced environmental stability of silk fibers incomparison to globular proteins is due to the extensivehydrogen bonding, the hydrophobic nature of much ofthe protein, and the significant crystallinity. Silks areinsoluble in most solvents, including water, dilute acidand alkali. Detailed structural analysis of spider draglinesilk proteins has yielded information on the organiza-tion and orientation of the numerous but very smallb-sheet crystals in the fibers, and a high level oforganization of the protein even in the less crystallinedomains [2]. Liquid crystalline phases and conforma-tional polymorphism have been implicated in thebiological processing of these proteins to contribute tothe architectural features within the fibers [22–24]. Thesenanoscale features, factoring in the small, orientated andnumerous b-sheet crystals, a fuzzy interphase betweenthese crystals and the less crystalline domains, and theshear alignment of the chains, provides a basis for theorigin of the novel mechanical properties exhibited bysilk fibers. A comparison of mechanical properties(Table 1) suggests that they provide a remarkablecombination of strength and toughness. The distinguish-ing features of the spider silks are the very high strengthin combination with excellent elasticity in comparisonwith these other biomaterials. In addition, these fibersdisplay resistance to failure in compression thatdistinguishes them from other high performance fibers,such as Kevlar [13].
G.H. Altman et al. / Biomaterials 24 (2003) 401–416402
2. Utility in biomedical applications
B. mori silkworm silk fibers have been the primarysilk-like material used in biomedical applications parti-cularly as sutures. During decades of use, silk fibers haveproven to be effective in many clinical applications. Atthe same time, some biological responses to the proteinhave raised questions about biocompatibility. One of themajor difficulties in assessing the biological responsesreported to these silk fibers is the absence of detailedcharacterization of the fibers used including, extent ofextraction of the sericin, the chemical nature of wax-likecoatings sometimes used, and many related processingfactors. This variability in source material has resultedin confusion in the literature and in clinical settingsconcerning the benefits or potential concerns with thisclass of fibrous protein. For example, of greatestimportance is that based on many studies it is clearthat the sericin glue-like proteins are the major cause ofadverse problems with biocompatibility and hypersensi-tivity to silk (Table 2). If sericin is removed, the biolo-gical responses to the core fibroin fibers appear to becomparable to most other commonly used biomaterials(Table 3). An additional major misconception revolvesaround silk’s classification as non-degradable. Thereis clear evidence in the literature that silk, as a protein, issusceptible to proteolytic degradation in vivo and overlonger time periods in vivo will slowly be absorbed.
3. In vivo biocompatibility of virgin silk
Despite centuries of use as sutures [25], silk matricesare now being rediscovered and reconsidered as
potentially useful biomaterials for a range of applica-tions in clinical repairs and in vitro as scaffolds fortissue engineering [26]; some benefits and concerns ofsilk are detailed in Table 4. Silk has been usedmost extensively as sutures for wound ligation andbecame the most common natural suture surpassingcollagen (CatgutTM or Chromic CatgutTM, cross-linkedcollagen) used in the biomedical industry over the past100 years [25]. During the past 20 years, a variety ofdegradable synthetic materials including polyglycolicacid (PGA)(DexonTM), co-polymers of PGA (polygly-conate)(MaxonTM or gylcolide and trimethylene carbo-nate), and co-polymers of PGA and polylactic acid(VicrylTM or polyglactin 910, a co-polymer of 90%glycolide and 10% l-lactide), polydioxanone (PDS),and non-degradable synthetics including braided polye-ster (EthibondTM or MersileneTM), nylon (EthilonTM),Teflon coated polyester (TevdekTM) and polypropylene(ProleneTM or SurgileneTM) have dominated the suturemarket. In general, sutures should be strong, handleeasily, and form secure knots [25]. Sutures require thefollowing characteristics for general surgical applica-tions [27]:
1. Tensile strength—to match the clinical repair.2. Knot strength—the amount of force required to
cause a knot to slip.3. Elasticity—the ability to conform to the current stage
of wound repair.4. Memory—change in stiffness over time; the better the
suture, the less memory.5. Degradability—ability to be metabolized by the host
once its repair function has been completed.6. Tissue Reactivity—non-irritant.
Table 1
Comparison of mechanical properties of common silks (silkworm and spider dragline) to several types of biomaterial fibers and tissues commonly
used today
Material UTS (MPa) Modulus (GPa) % Strain at break Authors
B. mori silk (w/ sericin)a 500 5–12 19 Perez-Rigueiro et al. [68]
B. mori silk (w/o sericin)b 610–690 15–17 4–16 Perez-Rigueiro et al. [68]
B. mori silkc 740 10 20 Cunniff et al. [13]
Spider silkd 875–972 11–13 17–18 Cunniff et al. [13]
Collagene 0.9–7.4 0.0018–0.046 24–68 Pins et al. [69]
Collagen X-linkedf 47–72 0.4–0.8 12–16 Pins et al. [69]
PLAg 28–50 1.2–3.0 2–6 Engelberg and Kohn [70]
Tendon (comprised of mainly collagen) 150 1.5 12 Gosline et al. [71]
Bone 160 20 3 Gosline et al. [71]
Kevlar (49 fiber) 3600 130 2.7 Gosline et al. [71]
Synthetic Rubber 50 0.001 850 Gosline et al. [71]
a Bombyx mori silkworm silk—determined from bave (multithread fibers naturally produced from the silk worm coated in sericin).b Bombyx mori silkworm silk—determined from single brins (individual fibroin filaments following extraction of sericin).c Bombyx mori silkworm silk—average calculated from data in Ref. [13].d Nephila clavipes silk produced naturally and through controlled silking.e Rat-tail collagen Type I extruded fibers tested after stretching from 0% to 50%.f Rat-tail collagen dehydrothermally cross-linked and tested after stretching from 0% to 50%.g Polylactic acid with molecular weights ranging from 50,000 to 300,000.
G.H. Altman et al. / Biomaterials 24 (2003) 401–416 403
7. Free from infection—related to the material’s geo-metry, e.g., multifilament vs. monofilament.
Processing methods for virgin silk (fibroin containingsericin gum) were developed in industry to extractsericin from the inner silk fibroin fibers. As sutures, thesilk fibroin fibers are usually coated with waxes orsilicone to enhance material properties and reducefraying; these sutures are commonly referred to as blackbraided silk (e.g., Perma-HandTM). The transition fromvirgin silk to black braided sutures occurred in the late1970s and early 1980s; however, virgin silk is stillcommercially available today.
Virgin silk, like most proteins, is a potential allergen[28,29] causing a Type I allergic response in some cases[30]. Delayed allergic responses to silk (an average of 10months after initial exposure) induced some patientcomplications including asthma and specific upregula-tion in IgE levels [29,31]. Wen et al. [29] showed theextracted sericin was responsible for sensitization (thedevelopment of a T-cell mediated allergic response) by
skin testing of 64 children with asthma. Furtherbiochemical analysis by Dewair et al. [32] and Zaominget al. [33] concluded that upregulated IgEs wereproduced in response to the sericin, clarifying the roleof these fibroin contaminants as the main allergenicagent in silk and not the core fibroin fibers. Fibroindegradation products could also potentially be involvedin adverse biological responses. For example, one casewas reported where fibroin (black braided silk) mayhave induced a Type I allergic reaction [30]. However,the authors clarified later that there were no cases in theliterature that implicated black braided silk as inducer ofIgE hypersensitivity reactions. In addition, in thepreparation of this review no references were found inthe literature that implicated black braided silk ininducing hypersensitivity and allergic reactions. Kuro-saki et al. [30] concluded that the hypersensitivity intheir single case may have been the result of patientsensitization to silk from a prior surgical procedure 7years earlier. The type of silk used in this earlierexposure was unknown (e.g., virgin vs. black braided).
Table 2
Biological responses to virgin silk fibers
Type of silka Response In vivo or in vitro model/tissue site Authors
? Mucosal edema was greatest around silk and chromic
collagen compared to PGA
Dog/genitourinary tract Morrow et al. [38]
? Chronic inflammation including granulosous and fibrosis Rabbit/median or sciatic nerve
ligation
Nebel et al. [39]
Virgin (a) Acute and chronic inflammation initially including
conjunctival and episcleral hyperemia progressing to
chemosis and nodular episcleritis, peripheral corneal
ulceration and wound necrosis
Human/ocular cataract surgery Soong and Kenyon
[34]
(b) Sensitization to virgin silk during bilateral cataract
surgery following the use of virgin silk on the first eye
Black braided No comparable suture reactions Human/ocular cataract surgery Soong and Kenyon
[34]
Virgin IgE obtain from sera isolated from 8 of 9 patients with
allergy to silk bound specifically to sericin and was
negative towards fibroin
In vitro immunoblot of IgE and IgG
isolated from sera of 9 persons
allergic to silk
Dewair et al. [32]
? (a) Acute and chronic inflammation Rabbit/tracheal anastomosis Peleg et al. [40]
(b) However, the response was not significantly different
from VicrylTM, Chromic Catgut, TevdekTM and
polypropylene
? Severe delayed chronic inflammatory reaction to suture Three human pediatric patients/
neurosurgery
Rossitch et al. [31]
Virgin Delayed hypersensitivity to virgin silk resulting in asthma Children o15 yr old/skin tests Wen et al. [29]
Virgin (a) 90% of patient population developed an allergic
reaction to silk extract (sericin); none of the control
group was allergic
41 humans with asthma and a clinical
history of silk allergy; 4 control
patients with no allergy to silk/sera
and skin tests were used.
Zaoming et al. [33]
(b) IgE from 41% of the patient population sera bound
specifically to sericin in in vitro testing
? (a) Type I allergy response to black braided silk Single case in a human/trachea and
throat region
Kurosaki et al. [30]
(b) Patient sensitized to silk 7 years prior in an
independent surgical procedure
a ?=Type of silk used was not specified, e.g., virgin silk or black braided; Virgin=virgin silk containing sericin; Black braided=black braided
extracted silk fibroin (e.g., no sericin) coated in waxes or silicone.
G.H. Altman et al. / Biomaterials 24 (2003) 401–416404
Tab
le3
Co
mm
on
sutu
res
that
elic
ita
fore
ign
bo
dy
resp
on
sefo
llo
win
gim
pla
nta
tio
nin
viv
o
Su
ture
trad
en
am
e/m
ate
rial
Mo
del
/im
pla
nta
tio
nsi
teA
bso
rbab
lea
Fo
reig
nb
od
yre
spo
nse
Filam
ent
typ
ebA
uth
ors
Su
rgic
al
con
tro
lR
at/
—8.2
%(7
.3–10.6
)—
cF
osc
hi
etal.
[43]
Vic
rylT
M(p
oly
gla
ctin
910)
Ab
do
min
al
cavit
yY
es;o
60
days
8.7
%(7
.1–9.9
)M
ult
i
Catg
utT
M(C
ollagen
)Y
es;o
60
days
9.4
8%
(8.3
–15)
Mu
lti
Po
lygly
colic
aci
dsa
lt(P
GA
)Y
es;o
60
days
11.2
5%
(9.5
–12.7
)M
ult
i
Sil
k(a
ssu
med
bla
ckb
raid
ed)
Yes
;>
60
da
ys
14
.2%
(10
.9–2
2.3
)M
ult
i
Tit
an
ium
(met
al
clip
)N
o15.8
%(1
2.1
–19.1
)M
on
o
Extr
ud
edT
eflo
nT
MR
ab
bit
/N
oM
ild
(1)
Mo
no
dS
etze
net
al.
[44]
Pro
len
eTM
(po
lyp
rop
yle
ne)
Su
bcu
tan
eou
sN
oM
ild
(1)
Mo
no
Su
rgilen
eTM
(po
lyp
rop
yle
ne)
No
Mild
(1)
Mo
no
PD
ST
M(p
oly
dio
xan
on
e)Y
es;o
60
days
Mo
der
ate
(2)
Mu
lti
Vic
rylT
MY
es;o
60
days
Mo
der
ate
(2)
Mu
lti
Sil
k(b
lack
bra
ided
)Y
es;
>6
0d
ay
sM
od
era
te(2
)M
ult
i
Nylo
nN
oM
od
erate
(2)
Mo
no
Eth
ibo
nd
TM
(bra
ided
po
lyes
ter)
No
Mo
der
ate
(2)
Mu
lti
Mer
silin
eTM
(bra
ided
po
lyes
ter)
No
Mo
der
ate
(2)
Mu
lti
Tev
dek
TM
(Tefl
on
coate
d
po
lyes
ter)
No
Mo
der
ate
(2)
Mu
lti
Ch
rom
icC
atg
utT
M(c
ross
lin
ked
collagen
)
Yes
;o
60
days
Exte
nsi
ve
(3)
Mu
lti
Nylo
nR
at/
No
Min
imal
Mo
no
eB
uck
nall
etal.
[45]
Nylo
nA
bd
om
inal
inci
sio
nN
oM
inim
al
Mu
lti
PG
AY
es;o
60
days
Min
imal
Mu
lti
Sil
k(b
lack
bra
ided
)Y
es;
>6
0d
ay
sM
od
era
teM
ult
i
aA
bso
rbab
leis
defi
ned
as
the
loss
of
ten
sile
stre
ngth
wit
hin
60
days
inviv
o,
bu
td
oes
no
tre
flec
tth
ep
ersi
sten
ceo
fa
fore
ign
bo
dy
resp
on
se(F
BR
)fo
rgre
ate
rth
an
60
days.
bM
on
ofi
lam
ent
sutu
res
con
tain
asi
ngle
fib
eran
dm
ult
ifila
men
tsu
ture
sco
nta
inm
ult
iple
fib
ers
incr
easi
ng
mate
rial
surf
ace
are
ain
viv
oan
dth
ep
ote
nti
al
risk
of
infe
ctio
n(i
.e.,
mo
rep
lace
sfo
r
con
tam
inati
on
toh
ide
inm
ult
ifila
men
tsu
ture
s).
cF
RB
was
chara
cter
ized
by
the
per
cen
tgro
wth
of
neo
ves
sels
(an
gio
gen
esis
)aro
un
dth
esu
ture
7d
ays
po
st-i
mp
lan
tati
on
.dF
BR
was
chara
cter
ized
on
a3
level
scale
30
days
po
st-i
mp
lan
tati
on
:M
ild
(1)
incl
ud
edth
ep
rese
nce
of
his
tio
cyte
san
dfo
reig
n-b
od
ygia
nt
cells,
Mo
der
ate
(2)
ind
icate
sh
isti
ocy
tes
an
dfo
reig
n-b
od
y
gia
nt
cells,
an
dcl
ust
ero
fly
mp
ho
cyte
san
da
few
pla
smacy
tes,
an
dE
xte
nsi
ve
(3)
incl
ud
esh
isti
ocy
tes,
fore
ign
-bo
dy
gia
nt
cells,
an
dd
iffu
se,
scatt
ered
lym
ph
ocy
tes
an
dp
lasm
acy
tes.
eF
BR
was
chara
cter
ized
his
tolo
gic
ally
10
days
inviv
o;
ho
wev
er,
the
au
tho
rsd
on
ot
pro
vid
ea
defi
nit
ion
of
‘min
imal’
or
‘mo
der
ate
’in
flam
mati
on
an
dp
rovid
en
om
ean
sb
yw
hic
hth
ey
syst
emati
cally
chara
cter
ize
the
resp
on
seh
isto
logic
ally.
G.H. Altman et al. / Biomaterials 24 (2003) 401–416 405
In other cases where the type of silk suture wasunknown [31], silk induced a delayed allergic reactionthat necessitated the removal of the suture. The authorsdid not distinguish between the use of virgin silkcontaining sericin versus black braided silk. However,based on the time frame of the surgical procedures(1970s) and the severity of the reaction, sericin can beconsidered as the most likely cause of the response.Soong and Kenyon [34] detailed 12 patients from 1980to 1983 that reported severe reactions to virgin silksutures following cataract surgery. No comparablereactions were noted when black-braided silk sutureswere used. Further, they provided evidence of patientsensitization to virgin silk in bilateral surgical proce-dures; an allergic reaction developed against virgin silkfollowing its use in the first eye procedure.
The use of virgin silk during the 1960s to the early1980s negatively impacted the general acceptance of thisbiomaterial from the surgical practitioner perspective[35–37]. For example, Morrow et al. [38] observed silkand collagen (catgut) to be moderately more reactivethan polyglycolic acid (PGA) in the genitourinary tractfollowing 3 weeks of implantation in vivo, althoughagain the type of silk used in the study was notidentified. Nebel et al. [39] observed cellular infiltration,granulosous and fibrosis for silk sutures that persistedfor more than 4 weeks compared to autologous collagensutures used for peripheral nerve ligation in the rabbit.Again the specific type of silk was not described. Of notein the study was that inflammation was comparable tochromic collagen sutures. Peleg et al. [40] drew a similarconclusion regarding the inflammatory potential ofsilk used in tracheal anastomosis in the rabbit with-out identifying the type of silk used. Of interest was asimilar acute and chronic inflammatory response toVicrylTM and TeflonTM coated polyester sutures in thestudy.
4. Foreign body response to silk
All biomaterials derived from a non-autologoussource will elicit some level of foreign body response(FBR) following implantation in vivo. Absorbable andnon-degradable biomaterials generally do not induce aT-cell mediated hypersensitivity immune response (aswas observed with sericin) [30]. However, while rare,sensitization to fibroin due to pre-exposure or a failedphagocytic response can result in the formation of agranuloma years after the implantation of black braidedsilk [30]. Some common materials that have been shownto induce a FBR are detailed in Table 3.
Biomaterial characteristics, including implantationsite, size, geometry [41], and surface topography caninfluence the level of the foreign body response [42]. Thisresponse can be predicted in part based on the surface tovolume ratio of the biomaterial [42]. This relationship isfrequently not addressed in the relative comparisonbetween different suture types in vivo. Several studieshave shown black braided silk to elicit a greaterFBR when compared to other commonly used sutures[43–45]. However, a single fiber of native silk iscomposed of several individual fibroin filaments and asingle black braided suture comprises several extractedsilk fibers. No studies have attempted to normalize theFBR to the suture surface area, rather equating suturesbased on their final diameter. Thus studies that reportsilk as inducing a heightened FBR may not be reflectiveof the inflammatory potential of the silk but rather amanifestation of the overall suture structure andgeometry.
Silk sutures induced angiogenesis, a limiting step inmounting a foreign body response, in the rat’smesenteric window 7 days post-implantation to anlesser extent as titanium and to a greater extent thancollagen, polyglactin, and PGA; however the differences
Table 4
Benefits and concerns with the use of silks for biomedical applications
Benefits
Novel mechanical properties of some silks that are superior to any other natural fiber and rival many high performance fibers
Natural fiber with a long standing history of use in clinical applications
The ability to process silks in aqueous solutions for subsequent formation of films and other material formats, with relatively simple
insolubilization via exposure to alcohols and other environmental factors
Easily chemically modified with surface decorations, such as adhesion sites or cytokines, due to the availability of amine and acid side chains on
some of the amino acids
Genetically tailorable composition and sequence to moderate specific features, such as molecular weight, crystallinity, solubility
Slow rates of degradation in vitro and in vivo, this is particularly useful in biodegradable scaffolds in which slow tissue ingrowth is desirable
No known risk of bioburden
Concerns
Adequate removal of contaminating sericin from silkworm silk to avoid biocompatibility problems
Slow degradation of crystalline (b-sheet) regions
Aborted proteolytic attack by macrophages and giant cells leading to encapsulation and the formation of a granuloma
Potential sensitization to silk fibroin resulting in an allergic response upon exposure to the biomaterial
G.H. Altman et al. / Biomaterials 24 (2003) 401–416406
were not statistically significant ðp > 0:05Þ (Table 3) [43].The type of silk suture (virgin versus black braided) wasnot described but it is assumed to be black braidedbased on the year of study. A curiosity about the studywas the choice to close all the wounds of all studygroups with silk suture, potentially sensitizing theanimals to the silk suture under investigation. Bucknallet al. [45] examined four sutures in abdominal woundclosure in a rat model 10, 30, and 70 days post-implantation. Histological examination revealed silkelicited a ‘moderate’ zone of inflammation comparedto a ‘minimal response’ induced by multi- and mono-filament nylon and multifilament PGA (Table 3). Nohistological criteria or ranking system was provided.However, the authors noted that even monofilamentnylon, thought to be the most biocompatible of thematerials evaluated, induced fibrous encapsulation 10days post-implantation. Of interest was the finding thatPGA, despite a 92% decrease in tensile strength 30 dayspost-implantation continued to elicit a FBR over 70days in vivo, the final time point of the study.
Setzen et al. [44] in a comprehensive blind study,statistically analyzed tissue responses to 11 types ofsutures subcutaneously implanted in the rabbit (Table 3).Responses were based on the number of giant cellspresent in the fibrous capsule and within the individualfilaments of the sutures, and the thickness of the fibrouscapsule surrounding the suture. Evaluations wereperformed 30, 60 and 120 days post-implantation. Theinflammatory response surrounding the encapsulationwas graded on a scale of 1–3 based on the type of cellsand the number of vessels found. Silk elicited anequivalent response to absorbable sutures and multi-filament non-absorbable sutures when evaluating thenumber of foreign body cells surrounding each suture 30days post-implantation. Non-degradable EthibondTM
(braided polyester) and TevdekTM(Teflon coated polye-ster) induced a significantly greater accumulation of cellsat the wound site compared to the group of suturesunder investigation. However, when assessing thethickness of the fibrous capsule, black braided silk aswell as EthibondTM induced the formation of signifi-cantly thicker capsules compared to the mean responseof the group. Histological analysis 30 days afterimplantation showed the tissue response to the silkwas equivalent to VicrylTM and PDSTM (two commonsynthetic absorbables), nylon (monofilament), andEthibondTM, MersileneTM and TevdekTM (three com-mon multifilament non-degradables), and less thancross-linked collagen (Chromic CatgutTM). The foreignbody response to VicrylTM lasted in excess of 120 days invivo, long after the loss of tensile integrity. Inconclusion, Setzen et al. [44] confirmed that theinflammatory response induced by silk was no greaterthan that in response to common absorbable sutures 30days post-implantation in a subcutaneous rabbit in vivo
model (Table 3). Furthermore, they found a significantlyhigher response to multifilament sutures than monofila-ment, confirming the influence of higher surface area tovolume ratio as indicative of induction of a greaterinflammatory response.
As a suture, silk is still popular in ocular, neural andcardiovascular surgery, but has also been used in avariety of other tissues in the body. Silk’s knot strength,handling characteristics and ability to lay low to thetissue surface make it a popular suture in cardiovascularapplications where bland tissue reactions are desirablefor the coherence of the sutured structures [36].However, black braided silk is thrombotic followingimplantation within ateriovessels [46,47]. Using a ratmodel, Dahike et al. [47] showed black braided silkbinds fibrin and platelets within 3 days in vivo followingimplantation into arteriovessels. Within 7 days, silk wasencapsulated by a dense thrombus consisting of plateletsand lymphocytes, erythrocytes, and granulocytes. From14 to 28 days, the gradual decline of the thrombus andthe development of a new endothelium layer wasobserved. By 28 days in vivo, silk was completelyendothelialized, indicating the presence of regeneratedendothelial tissue. While silk was initially thrombotic(e.g, 3 and 7 days post-implantation) when compared toProleneTM, EthilonTM, VicrylTM and MersileneTM, by28 days in vivo, silk was less so when compared to bothEthilonTM and VicrylTM. At 56 days post-implantationonly traces of thrombogenicity were detected with blackbraided silk by the presence of loosely bound platelets tothe neoendothelium.
The ability of black braided silk to induce thrombosisimmediately following implantation in blood flow maybe due to the surface properties and ability to bindproteins in the clotting cascade such as fibrin. Alteringthe surface properties of extracted fibroin silk (i.e.,eliminating the wax coating) significantly diminished theinitially high thrombotic response to silk [48]. Sakabeet al. [49] coated polyester suture with solubilized fibroinand demonstrated that the reconstituted fibroin proteindid not induce significant thrombogenicity in vivo. Invitro studies performed by Santin et al. [50] confirmedthe ability of silk fibroin films (extracted sericin-freefibroin dissolved in LiBr salt, dialyzed in dIH2O, castand made insoluble with methanol) to bind fibrinogen asa mechanism by which the alternative complementpathway is activated in vivo. However, compared totwo reference materials, polystyrene and poly(2-hydro-xyethyl methacrylate), fibroin films bound fibrinogen toa lesser extent [50]. C3 and IgG, components of thehuman plasma complement system bound similarlywhen the silk was compared to the reference materials.The authors concluded that C3 binding may be theresult of specific binding to hydrophobic patches (e.g., b-sheets) of fibroin. When compared to the referencematerials after 24 h in vitro, diminished interleukin-1b
G.H. Altman et al. / Biomaterials 24 (2003) 401–416 407
production from mononuclear cells isolated from hu-man plasma incubated with fibroin was observed [50].
Uff et al. [51] examined the effects of soluble factorsfrom sutures on macrophage activation (e.g., attach-ment, phagacytosis and the production of lysozyme andproinflammatory tumor necrosis factor (TNF)) in vitro.Soluble factors were acquired by subjecting sterilesutures to ultrasound for 1 h in detergent. Solublefactors from silk were found to elicit the highestphagocytic response from macrophages when comparedto catgut, PDSTM, steel, nylon and VicrylTM. However,VicrylTM induced the highest response when character-izing macrophage adherence and TNF production.PDSTM induced the highest production of lysozyme ofthe group of sutures extracted. The authors failed tocharacterize the soluble factors produced by sonificationof the suture materials. In the case of silk, the resultingresponse from macrophages could be due to a variety offactors including residual sericin, waxes or silicones usedin manufacture of the sutures, or the specific size offibroin particles or crystals potentially generated. Thus,the response may be completely independent of actualsilk fibroin material properties and rather a function ofthe coating and/or geometry and size of particlesproduced.
5. Silk degradation
According to the US Pharmacopeia an absorbable(suture) biomaterial is defined as one that ‘loses most ofits tensile strength within 60 days’ post-implantation invivo. Within this definition, silk is correctly classified asnon-degradable. However, according to the literature,silk is degradable but over longer time periods due toproteolytic degradation usually mediated by a foreignbody response [31,34,41,52]. Several studies detailvariable rates of silk absorption in vivo dependent onthe animal model and tissue implantation site (Table 5).In general, silk fibers lose the majority of their tensilestrength within 1 year in vivo, and fail to be recognizedat the site within 2 years [53].
Silk studies in vitro have demonstrated that proteasessuch as chymotrypsin will cleave the less-crystallineregions of the protein to peptides which are then capableof being phagocytosed for further metabolism by thecell. Furthermore, protease cocktails [54] and chymo-trypsin (known to be produced by macrophages) arecapable of enzymatically degrading silk [55]. Of interest,the silkworm B. mori produces a protease inhibitorin the silk gland embedding it within the silk cocoonfor protection against premature proteolytic degrada-tion [56].
In a comparative study of six absorbable and fournon-degradable sutures implanted circumferentiallyunder the skin of rats, silk lost 55% of its tensile T
ab
le5
Evid
ence
of
silk
deg
rad
ati
on
invit
roan
din
viv
o
Typ
eo
fsi
lka
Inviv
o/i
nvit
roM
ech
an
ism
Deg
ree
an
dm
easu
reo
fd
egra
dati
on
Au
tho
r
Extr
act
edfi
bro
infi
lmIn
vit
roP
rote
oly
tic
deg
rad
ati
on
B10%
wei
gh
tlo
ss5
days
follo
win
g
enzy
mati
cd
iges
tio
n
Min
ou
raet
al.
[54]
Un
kn
ow
n/a
ssu
med
bla
ckb
raid
edR
at/
sub
cuta
neo
us
Un
kn
ow
n/a
ssu
med
fore
ign
bo
dy
resp
on
se
55%
loss
inte
nsi
lest
ren
gth
6w
eek
s
inviv
o
Gre
enw
ald
etal.
[57]
Bla
ckb
raid
edR
at/
sub
cuta
neo
us
Un
kn
ow
n/a
ssu
med
fore
ign
bo
dy
resp
on
se
83%
loss
inte
nsi
lest
ren
gth
10
wee
ks
inviv
o
Bu
ckn
all
etal.
[45]
Un
kn
ow
n/a
ssu
med
bla
ckb
raid
edR
at/
ab
do
min
al
wall
mu
scle
Fo
reig
nb
od
yre
spo
nse
(pro
teo
lyti
c
deg
rad
ati
on
)
Fra
gm
enta
tio
nat
6w
eek
s;n
ot
det
ecte
d
at
24
wee
ks
Lam
etal.
[41]
Bla
ckb
raid
edR
ab
bit
/co
rnea
,sc
lera
an
do
cula
rm
usc
leF
ore
ign
bo
dy
resp
on
se(p
rote
oly
tic
deg
rad
ati
on
)
Red
uce
dn
um
ber
of
fila
men
tsan
d
dia
met
erat
42
days;
ab
sorp
tio
nat
90
days
inviv
o
Salt
ho
use
etal.
[52]
Un
kn
ow
n/a
ssu
med
vir
gin
silk
Rab
bit
/ab
do
min
al
wall
mu
scle
Fo
reig
nb
od
yre
spo
nse
(pro
teo
lyti
c
deg
rad
ati
on
)
80%
dec
rease
inte
nsi
lest
ren
gth
at
12
wee
ks;
0%
stre
ngth
at
2yea
rs;
dec
rease
inth
en
um
ber
of
fib
ers
ob
serv
ed
his
tolo
gic
ally;
fragm
enta
tio
nfo
llo
win
g
4w
eek
sin
viv
o
Pro
stle
thw
ait
[36]
aT
yp
eo
fsi
lk:
vir
gin
=ra
wsi
lkco
nta
inin
gse
rici
n;
bla
ckb
raid
ed=
extr
act
edsi
lkfi
bro
infi
ber
s(e
.g.,
no
seri
cin
)co
ate
din
waxes
or
silico
ne;
fib
roin
film
=ex
tract
edan
dso
lub
iliz
edfi
bro
inca
stan
d
inso
lub
iliz
edw
ith
met
han
ol.
G.H. Altman et al. / Biomaterials 24 (2003) 401–416408
strength and 16% of its elastic modulus 6 weeks post-implantation [57] (Table 6). In a rat model withsubcutaneous implantation, silk fibers lost 29% of theirtensile strength at 10 days, 73% at 30 days and 83% 70days post-implantation (Table 7) [45]. In abdominalwound closures in the rat, silk promoted a moderateforeign body response compared to mono- and multi-filament nylon and polyglycolic acid (PGA) sutures [45].However, common to all sutures, inflammation wasrequired as the main mechanism for degradation.
Lam et al. [41] described silk as biodegradable due toits susceptibility to proteolytic enzymes. The studycompared dry-spun hot-drawn poly(l-lactic acid)(PLLA) fibers to several absorbable and non-degradablesutures in the muscle layer surrounding the abdomen ofrats. PLLA, PDSTM, VicrylTM, and black braided silkshowed signs of degradation after 2 weeks in vivo asdetermined by scanning electron microscopy. Allsutures, in addition to monofilament nylon, provokeda chronic inflammatory response at 2 weeks. At 6 and 12weeks in vivo, the authors were unable to retrieve silkdue to fragmentation. At 24 weeks, neither silk norinflammation was observed in vivo most likely due therapid degradation of silk by proteolytic enzymaticdigestion. Inflammation subsided by 6 weeks in vivofor Vicryl; a chronic inflammatory response was presentsurrounding PLLA at 80 weeks in vivo, the final timepoint of the study.
In general, silk is slowly absorbed in vivo. The rate ofabsorption is dependent upon the implantation site,mechanical environment, and variables related to thehealth and physiological status of the patient, the type(virgin silk versus extracted black braided fibroin), andthe diameter of the silk fiber [31,34,52]. Furthermore,alterations in silk processing may cause conformationalchanges in the protein structure potentially increasing ordecreasing susceptibility to degradation. None ofthese variables has been studied in detail; therefore,it is difficult to gain a clear understanding of therelationships between structure, processing and degrad-ability. Regardless, silk protein fibers will degrade invivo; rates are variable depending on the factors listedabove.T
ab
le6
Mate
rial
pro
per
ties
of
10
com
mo
nsu
ture
sfo
llo
win
g6
wee
ks
of
sub
cuta
neo
us
imp
lan
tati
on
ina
rat
mo
del
.V
alu
esat
tim
e0–6
wee
ks
tak
enfr
om
Tab
le1
of
Ref
.[5
7]
Su
ture
nam
eS
tren
gth
(N/m
2)a
Tan
gen
tel
ast
icm
od
ulu
s(N
/m2)b
0W
eek
s6
Wee
ks
%D
ecre
ase
0W
eek
s6
Wee
ks
%D
ecre
ase
Vic
rylT
M(p
oly
gla
ctin
910)
1.4
2E
14
Fra
gm
ente
dB
100
3.0
0E
8—
100
Dex
on
TM
(po
lygly
colic
aci
d)
3.4
6E
14
Fra
gm
ente
dB
100
2.6
2E
8—
100
Catg
utT
M(c
ollagen
)2.3
0E
14
Fra
gm
ente
dB
100
1.4
1E
8—
100
Ch
rom
icC
atg
utT
M(c
ross
-lin
ked
collagen
)2.2
8E
14
Fra
gm
ente
dB
100
1.5
6E
8—
100
PD
ST
M(p
oly
dio
xan
on
e)7.2
0E
14
9.7
2E
13
86.5
9.8
7E
77.1
3E
785.8
Maxo
nT
M(p
oly
gly
con
ate
)5.9
6E
14
8.2
9E
13
86.1
1.0
0E
87.4
E7
26
TM
Sil
kc
8.4
8E
13
3.8
0E
13
55
.22
.42
E8
2.0
3E
81
6.1
Eth
ilo
nT
M(n
ylo
n)
4.5
7E
14
2.5
4E
14
44.4
1.0
4E
81.0
7E
80
Pro
len
eTM
(po
lyp
rop
yle
ne)
4.0
2E
14
2.9
9E
14
25.6
1.0
0E
81.0
1E
80
Eth
ibo
nd
TM
(po
lyes
ter)
2.4
9E
14
2.3
2E
14
6.8
2.4
7E
82.3
0E
86.9
aS
tren
gth
of
init
ial
sutu
rep
rio
rto
imp
lan
tati
on
(0w
eek
s)an
dat
6w
eek
sfo
llo
win
gin
viv
oci
rcu
mfe
ren
tial
sub
cuta
neo
us
imp
lan
tati
on
ina
rat
mo
del
.F
ragm
ente
dsu
ture
sth
at
wer
eu
nab
leto
be
harv
este
dfo
llo
win
gim
pla
nta
tio
nat
6w
eek
sw
ere
ass
um
edto
have
lost
100%
of
thei
rte
nsi
lein
tegri
ty.
bE
last
icm
od
ulu
so
fsu
ture
s(N
/m2)
pri
or
toim
pla
nta
tio
nan
dfo
llo
win
g6
wee
ks
inviv
o.
cT
yp
eo
fsi
lk(e
.g.,
vir
gin
or
bla
ckb
raid
ed)
was
no
td
escr
ibed
.It
isass
um
edto
be
bla
ck-b
raid
edsi
lk.
Table 7
Percent loss in suture mechanical tensile strength over time sub-
cutaneously implanted in a rat model. Durability data was taken from
Table 3 of Ref. [45]
Suture Time post-implantation
10 days 30 days 70 days
Polyglycolic acid 65% 92% 99%
Silk, black braided 29% 73% 83%
Multifilament nylon 16% 19% 22%
Monofilament nylon 3% 7% 16%
G.H. Altman et al. / Biomaterials 24 (2003) 401–416 409
It should be recognized that amyloid-like deposits inAlzheimer’s and related diseases are often associatedwith b-sheet structures, particularly cross-beta crystals[58]. Similar structures are also implicated in membrane-associated cytotoxicity [58]. It is unknown if silks, suchas those from silkworm, spiders or genetic variantsthereof, will negatively influence biological function in asimilar mode as these amyloid systems. This issue shouldbe addressed over time as tissue-engineering studiesmove forward. The outcomes of these studies wouldinfluence the types of tissue applications that should beaddressed with silk. It is encouraging that no amyloid-related problems have been reported throughout theliterature to date associated with silk fibers, despite theirextensive use in the clinical setting for many centuries.As the use of silk-based biomaterials broadens, pre-sumably the ASTM would be involved to establishstandard guidelines for this type of assessment.
6. Scaffolds for tissue engineering
The emergence of tissue engineering has increased thedemand for a diverse portfolio of biomaterials tosupport the development of tissues in vitro prior toimplantation in vivo. The biomaterial or matrix plays akey role in communicating or transducing environmen-tal cues to cells seeded within or on the matrix. Inessence, the matrix acts as the translator between thelocal (e.g., mechanical) environment (either in vitro or invivo) and the developing tissue, aiding in the develop-ment of biologically viable functional tissue. The matrixmust support cell attachment, spreading, growth anddifferentiation. In most instances it is advantageous ifthe matrix degrades into biocompatible fragments ormonomers capable of being metabolized by host cells;however, the rate of degradation must match or be lessthan the rate of tissue ingrowth and development. This
balance assures appropriate mechanical and physiologi-cal compatibility during integration of the host and/orimplanted tissue in vivo. A mismatch in these rates canlead to premature failure of the tissue. Silk fibroin offersversatility in matrix scaffold design for a number oftissue engineering needs (Table 8) in which mechanicalperformance and biological interactions are majorfactors for success, including bone, ligaments, tendons,blood vessels and cartilage. Silk fibroin can be processedinto foams, films, fibers and meshes.
Inouye et al. [59] demonstrated the utility of a silkfibroin film in culturing animal cells (SE1116, humancolon adenocarcinoma; KB, human mouth epidermoidcarcinoma; Colo201, human colon adenocarcinoma;QG56, human lung carcinoma) in comparison to acollagen matrix. Films of both fibroin and collagensupported equivalent cell growth after 5 days in cellculture conditions. Minoura et al. [26] compared theability of fibroin and sericin films, as well as a control,collagen (cast bovine acid soluble) film, in supportingattachment, spreading and growth of the L-929 fibro-blast cell line. The results indicated that fibroin andcollagen films were equivalent in their ability to supportcell attachment, physiological morphology and growthwhen compared to the sericin film. These results supportthe concept that extracted fibroin free of sericin is asuitable matrix for cell and tissue culture.
In our own studies, fibroin films induced bone tissuegrowth in vitro when seeded with osteoblasts [60]. Whenthe films were chemically decorated with the peptideRGD to promote integrin interactions for adhesion, theinduction of bone formation in vitro was significantlyenhanced. The response was determined based onincreased alkaline phosphatase levels, upregulation ofbone-specific transcripts, and calcification levels over 4weeks. Similar responses were not observed whenparathyroid hormone was immobilized and also assayedwith osteoblasts. The utility of a protein matrix offering
Table 8
Examples of silk’s utility as a matrix material in tissue engineering
Silk form Supported cell type in vitro Comments Authors
Fibroin film L-929 mouse fibroblast Comparable growth rates to collagen films Minoura et al. [26]
Fibroin film SE1116 (human colon adenocarcinoma); KB
(human mouth epidermoid carcinoma);
Colo201 (human colon adenocarcinoma);
QG56 (human lung carcinoma)
Comparable growth rates to collagen films as
well as rates of protein production of
carcinoembryonic antigen (CEA)
Inouye et al. [59]
Fibroin film Saos-2 (human osteoblast-like cells) Bone formation was evident on fibroin films,
but was enhanced on RGD-coupled matrices
Sofia et al. [60]
Fibroin film hBMSC (human bone marrow stromal cells) Supports bone nodule formation from adult
stem cells
(Unpublished data)
Fibroin fibers hBMSC; human adult anterior cruciate
ligament fibroblasts
Supports ligament specific development
in vitro
Altman et al. [61]
G.H. Altman et al. / Biomaterials 24 (2003) 401–416410
a diversity of sites (e.g., amino acids) for selectivechemical couplings for tissue engineering is a benefit ofthe silk system. Bone tissue formation was a logicaltarget tissue for silk due to the unique mechanicalproperties of the protein in fiber and film forms.
Recent research with silk has focused on the devel-opment of a wire rope matrix for the development ofautologous tissue engineered anterior cruciate ligaments(ACL) using a patient’s own adult stem cells [61].Biologically compatible and mechanically robust bio-materials are critical based on the stringent requirementsof ligaments and the dynamic and demanding mechan-ical environment of the knee. Human bone marrowstromal cells (BMSCs) cultured in collagen gels weregrown in a mechanically dynamic environment relevantto that present in vivo leading to the formation ofligament fibroblasts (from the BMSCs) and maturetissue development in vitro [62]. However, the poorintegrity of the collagen gels and the demanding anddynamic intra-articular mechanical and biochemicalenvironment of the knee have prompted renewedinterest in silk as a long term absorbable material withgood mechanical integrity and biocompatibility.
Collagen fibers have been used as a matrix biomater-ial to support de novo ligament formation in vivo withlimited success. In the early 1990s Dunn et al. [63]reported initial work on the development of a ligamentprosthesis involving a collagenous ACL prosthesis [63].Results showed inconsistent neoligament formation andsignificant weakening of the prosthesis in a rabbitmodel. Enhanced scaffolds were examined including acollagen fiber-PLA composite to further maintainmechanical integrity allowing for neoligament tissueingrowth [64]. In both studies, only half of the structuresremained intact 4 weeks post-reconstruction suggestingthat this composition (collagen and PLA) was inade-quate for the rigorous in vivo environment of the ACL.
Silk’s unique mechanical properties, coupled with theability to weave the fibers into a wire-rope geometryprovides control over the matrix’s final mechanicalproperties (i.e., matrix stiffness can be altered byadjusting the pitch angle of a wire rope) to mimic themechanical properties of the native ACL and supporthost tissue ingrowth offering new options in ACL tissueengineering. Pilot scale manufacturing equipment hasbeen developed for the fabrication of ACL wire-rope
Fig. 1. (A) Pilot-scale manufacturing equipment for the fabrication of silk wire-rope matrices showing (i) two computer controlled twisting
machines, (ii) a 160 l stainless steel water bath with immersion heaters, and (iii) a custom designed fiber extraction rack holding 94 individual groups
of fibers to be extracted free of sericin; (B) a close up of the extraction rack containing 94 individually anchored groups of fibers; (C) spring-loaded
clamps used to keep fibers or groups of fibers in equal tension during extraction, rinsing and drying.
G.H. Altman et al. / Biomaterials 24 (2003) 401–416 411
fibroin matrices (Fig. 1). A 160 l stainless steel waterbath (Fig. 1A) and custom designed fiber extractionracks to keep individual fibers or groups of fibers inequal tension were developed. The extraction rack wasdesigned to anchor up to 94 individual fibers or groupsof fibers (Fig. 1B) with spring-loaded clamps (Fig. 1C).The water bath (214 cm� 61 cm� 15 cm) combined withtwo immersion heaters and an external pump allowrecirculating flow of the extraction broth at 901Cover the rack containing the silk. Following sericinextraction, two computer controlled twisting machines(Figs. 1A and 2) were developed to fabricate wire-ropematrices with desired geometries (e.g., material proper-ties). One of many fabricated matrices is shown inFig. 3. Confirming silks utility as a biocompatiblematerial for in vitro ligament tissue engineering, BMSCswere shown to attach, spread and proliferate on the silkfiber matrices [61]. Scanning electron microscopy (SEM)images of virgin silk and extracted silk (Fig. 4) as well asBMSC seeded fibroin wire-rope matrices are shown(Fig. 5). BMSC are able to attach within 5 min(Fig. 5A), initially spread at 1 h (Fig. 5B), and formconfluent cell sheets with possible extracellular matrixformation 7 days and 14 days post-seeding (Figs. 5Cand D).
It is generally accepted in the literature that silk willslowly degrade in vivo as a function of proteolyticattack. However, in vitro in immune system deficienttissue culture conditions, sericin-extracted fibroin silkfibers retained their initial tensile integrity over 21 days
[61]. In vitro studies of the extracted silk fibroinexhibited a negligible response from macrophages asassessed by cytokine release in vitro (unpublished data).As a result, appropriate, directed and rigorous mechan-ical stimuli can be communicated to adhered stem cellsin vitro without concern for premature graft mechanicalfailure; however, the effects of mechanical fatigue ongraft integrity will need to be characterized for eachculture condition and variant matrix geometry.
We have shown that the application of mechanicalstimuli will induce BMSC ligament specific differentia-tion and matrix formation in vitro [62]. The utility ofthis tissue engineering approach in combination with amechanical robust silk matrix is the potential inhibitionof a FBR following implantation in vivo due to the
Fig. 2. A close-up view of the twisting machines showing the motor controlled spring-loaded clamps. The clamps can anchor from 2 to 6 fibers or
groups of fibers for twisting.
Fig. 3. One of many silk cords manufactured by the twisting
equipment. The silk cord shown contains 5 levels of twisting hierarchy
and 540 individual fibers twisted to approximate the stiffness of the
human ACL.
G.H. Altman et al. / Biomaterials 24 (2003) 401–416412
Fig. 4. (A) SEM of raw virgin B. mori silk fibers prior to extraction showing the gum-like sericin proteins coating the core fibroin and (B) following
extraction at 901C for 60 min in conditions previously described [59] showing the individual smooth fibroin filaments following the removal of the
covering sericin.
Fig. 5. (A) Initially attached BMSCs on the fibroin cord shown in Fig. 3, 5 min following seeding of the fibers with 2 million cells/ml (1 ml total per
3 cm of silk cord). (B) Initial cell spreading on the fibroin 1 h after seeding. (C) Seven days post-seeding with a cell and ECM sheet evident. (D)
Fourteen days following seeding, thick encapsulation of the matrix by cells and ECM.
G.H. Altman et al. / Biomaterials 24 (2003) 401–416 413
presentation of autologous ligament tissue (derived fromthe BMSC) to the host. This approach should allowboth infiltration of host tissue and the in vitro developedligament tissue incorporated into and surrounding thematrix to form mature functional, biologically viableautologous ligament tissue. Studies are currently on-going to identify an optimal matrix geometry andstiffness to best support tissue ingrowth both in vitroand in vivo. Furthermore, silk’s high ultimate tensilestress (N/mm2) provides significant void volume in vivooccupying only B12% of the space of a human ACL(given a human ACL 27 mm in length and 8 mm indiameter). We suspect, following angiogenesis andrevascularization of the tissue engineered ligamentwithin 8–12 weeks in vivo, the core original silk matrixwill be proteolytically degraded and metabolized in vivo.The effects of the rigorous mechanical knee jointenvironment on matrix fatigue life over the long termremain to be determined but will clearly guarantee theeventual failure of the silk matrix in vivo, a desired goalto ensure biologically viable, host ligamentous tissueeventually sustains the functional roles of the ACL overthe life of the patient.
In terms of matrices for tissue engineering, the novelmechanical properties of silk in film, fiber (Table 1) orsponge forms, coupled with the facile chemical decora-tion, the potential to form complex rope designs tomatch functional requirements for specific tissues, andthe slow rate of degradation in vivo to provide adequaterobust support, suggest that silks offer many benefitswhen compared to other types of natural or syntheticdegradable fibers, films and foams. For example, in thecase of bone, biomaterials that can provide robustsupport during compression would be advantageous,and silks are known to function well under compressionwith little evidence for cracking or crazing underconditions in which synthetic high performance fibersfail [13]. Furthermore, in the case of ligaments, the hightensile strength of silks provides a strong advantage inmatching ligament function to immediately restore kneefunction, a feature difficult to achieve with collagen [61].
Silks fibers and films are often prepared fromresolubilized protein after dissolution in high concentra-tions of lithium salts or calcium nitrate. Once soluble,the proteins can be dialyzed, lyophilized and resolubi-lized in organic solvents such as hexafluoroisopropanolor in water for limited periods of time depending onconcentration. In the case of fiber formation, silks[65,66] and collagens [67] have been electrospun to formnanometer-scale diameter fibers. Thus, modulation infiber diameters from the native 10’s of microns in thecase of B. mori to submicron by electrospinning offers arange of morphologies for cell interactions. It is alsointeresting to note that unrestrained dragline silk fromN. clavipes will supercontract. However, when em-bedded in a material or restrained, this unusual feature
is lost. Silkworm silk fibers from B. mori do notsupercontract.
7. Conclusions
B. mori silk fibers are composed primarily of twotypes of proteins: (1) sericin, the antigenic gum-likeprotein surrounding the fibers and (2) fibroin, the corefilaments of silk comprised of highly organized b-sheetcrystal regions and semi-crystalline regions responsiblefor silk’s elasticity compared to fibers of similar tensileintegrity. Silk has been used in biomedical applicationsfor centuries primarily for the ligation of wounds. Virginsilk suture (containing sericin) induces hypersensitivityin patients, causing a Type I allergic reaction. Exposureto silk debris (e.g., broken virgin silk fibers used inbedding and fabrics) may sensitize patients to silkcausing adverse allergic reactions when silk is used asa suture material. Sericin, identified as the antigenicagent of silk, is removed and replaced with a wax orsilicone coating in commercial black braided silksutures. As a result, T-cell mediated hypersensitivityhas not been observed in response to black braided silkin the literature. Unfortunately, poor documentation ofthe exact type of silk suture used when describing hostresponses to silk in the literature has led to confusionregarding the utility of silk in biomedical applications.As a result, the use of silk over the past 20 years hasdeclined since the arrival of commonly used absorbableand non-degradable synthetics such as VicrylTM andEthibondTM.
Another misconception about silk revolves around itsability to degrade in vivo. Defined as non-degradable bythe USP because it retains the majority of its tensilestrength beyond 60 days in vivo, silk is commonlythought of as a permanent suture once implanted intothe body. As a protein, silk is susceptible to proteolyticdegradation, but over a longer time periods. Based onthe literature, it is apparent that silk fibroin will lose themajority of its tensile strength within 1 year in vivo. Insome cases, the silk or the features associated withinflammation cannot be found at the implantation sitewithin months following surgery. Therefore, silk is along-term absorbable suture.
Silk fibroin elicits a foreign body response followingimplantation in vivo. Yet the response is comparable tothe most popular synthetic materials in use today asbiomaterials, and is dependent upon the implantationsite and model used for investigation. In rare cases agranuloma may form as a result of an abandonedphagocytic response to silk by macrophages and giantbody cells. Again, the response is dependent on theimplantation site of the silk. It is known that withinblood, silk is highly thrombic. Clearly, this results froma unique feature of the silk fibroin protein to bind
G.H. Altman et al. / Biomaterials 24 (2003) 401–416414
components of the clotting cascade such as fibrin andfibrinogen. However, in most cases, the response ismoderate and subsides with time.
To examine the material properties of silk fibroin inrelation to other common degradable and non-degrad-able biomaterials in vivo, comparisons should be madeby equating surface area of the materials underinvestigation. Individual silk fibroin filaments, due totheir small diameter (B5 mm) increase the surface areato volume ratio of silk sutures when compared to otherbiomaterials where the equating factor is typically finalsuture diameter. As a result, in cases that demonstrate agreater foreign body response to silk in vivo, thecomparison may not be equitable to the inherentmaterial properties of silk fibroin. Therefore, silk fibroinwhen utilized as films, foams and fibers, may offer a‘new’ alternative biomaterial for use as matrices in tissueengineering where mechanically robust, long-term de-gradable materials are needed.
Acknowledgements
We thank the NIH (R01 DE13405-01 and R01AR46563-01), the NSF (DMR-0090384) and TissueRegeneration, Inc., for support of this research.
References
[1] Kaplan DL, Mello SM, Arcidiacono S, Fossey S, Senecal K,
Muller W. Silk. In: McGrath K, Kaplan DL, editors. Protein
based materials. Boston: Birkhauser, 1998. p. 103–31.
[2] Kaplan DL, Adams WW, Farmer B, Viney C. Silk: biology,
structure, properties and genetics. In: Kaplan DL, Adams WW,
Farmer B, Viney C, editors. Silk polymers: materials science and
biotechnology. ACS Symp Ser 1994;544:2–16.
[3] Kaplan DL, Fossey S, Viney C, Muller W. Self-organization
(assembly) in biosynthesis of silk fibers—a hierarchical problem.
In: Aksay IA, Baer E, Sarikaya M, Tirrell DA, editors.
Hierarchically structured materials. Materials Res Symp Proc
1992;255:19–29.
[4] Guerette PA, Ginzinger DG, Weber BHF, Gosline JM. Silk
properties determined by gland-specific expression of a spider
fibroin gene family. Science 1996;272:112–5.
[5] Cappello J. The biological production of protein polymers and
their use. Trends Biotechnol 1990;8:309–11.
[6] Xu M, Lewis RV. Structure of a protein superfiber: spider
dragline silk. Proc Natl Acad Sci USA 1990;87:7120–4.
[7] Gosline JM, Denny MW, DeMont ME. Spider silk as rubber.
Nature 1984;309:551–2.
[8] Hayashi CY, Lewis RV. Molecular architecture and evolution of
a modular spider silk protein gene. Science 2000;287:1477–9.
[9] Arcidiacono S, Mello C, Kaplan DL, Cheley S, Bayley H.
Purification and characterization of recombinant spider silk
expressed in Escherichia coli. Appl Microbiol Biotechnol
1998;49:31–8.
[10] Hayashi CY, Lewis RV. Evidence from flagelliform silk cDNA
for the structural basis of elasticity and modular nature of spider
silks. J Mol Biol 1998;275:773–84.
[11] Hinman MB, Lewis RV. Isolation of a clone encoding a second
dragline silk fibroin. J Biol Chem 1992;267:19320–4.
[12] Zhou CZ, Confalonieri F, Medina N, Zivanovic Y, Esnault C,
Yang T, Jacquet M, Janin J, Duguet M, Perasso R, Li ZG. Fine
organization of Bombyx mori fibroin heavy chain gene. Nucleic
Acids Res 2000;28:2413–9.
[13] Cunniff PM, Fossey SA, Auerbach MA, Song JW, Kaplan DL,
Adams WW, Eby RK, Mahoney D, Vezie DL. Mechanical and
thermal properties of dragline silk from the spider, Nephila
clavipes. Polym Adv Technol 1994;5:401–10.
[14] Gosline JM, DeMont ME, Denny MW. The structure and
properties of spider silk. Endeavour 1986;10:37–43.
[15] Lazaris A, Arcidiacono S, Huang Y, Zhou JF, Duguay F,
Chretien N, Welsh EA, Soares JW, Karatzas CN. Spider silk
fibers spun from soluble recombinant silk produced in mamma-
lian cells. Science 2002;295:472–6.
[16] Hood EE, Jilka JM. Plant-based production of xenogenic
proteins. Curr Opin Biotechnol 2000;10:382–6.
[17] Wilson D, Valluzzi R, Kaplan DL. Conformational transitions in
model silk peptides. Biophys J 2000;78:2690–701.
[18] Szela S, Avtges P, Valluzzi R, Winkler S, Wilson D, Kirschner D,
Kaplan DL. Reduction-oxidation control of b-sheet assembly
in genetically engineered silk. Biomacromolecules 2000;1:
534–42.
[19] Winkler S, Wilson D, Kaplan DL. Controlling b-sheet assembly
in genetically engineered silk by enzymatic phosphorylation/
dephosphorylation. Biochemistry 2000;39:12739–46.
[20] Valluzzi R, Szela S, Avtges P, Kirschner D, Kaplan DL.
Methionine redox controlled crystallization of biosynthetic silk
spidroin. J Phys Chem B 1999;103:11382–92.
[21] Winkler S, Szela S, Avtges P, Valluzzi R, Kirschner DA, Kaplan
DL. Designing recombinant spider silk proteins to control
assembly. Int J Biol Macro 1999;24:265–70.
[22] Winkler S, Kaplan DL. Molecular biology of spider silk. Rev Mol
Biotech 2000;74:85–93.
[23] Heslot H. Artificial fibrous proteins: a review. Biochimie
1998;80:19–31.
[24] Kadler KE. Fibril forming collagens. In: Sheterline P, editor.
Protein profiles: extracellular matrix, vol. 1. London, UK:
Academic Press, 1995.
[25] Moy RL, Lee A, Zalka A. Commonly used suture materials in
skin surgery. Am Fam Physician 1991;44:2123–8.
[26] Minoura N, Aiba S, Gotoh Y, Tsukada M, Imai Y. Attachment
and growth of cultured fibroblast cells on silk protein matrices.
J Biomed Mater Res 1995;29:1215–21.
[27] Moynihan BGA. The ritual of a surgical operation. Br J Surg
1920;8:27–35.
[28] Hollander DH. Interstitial cystitis and silk allergy. Med Hypoth-
eses 1994;43:155–6.
[29] Wen CM, Ye ST, Zhou LX, Yu Y. Silk-induced asthma in
children: a report of 64 cases. Ann Allergy 1990;65:375–8.
[30] Kurosaki S, Otsuka H, Kunitomo M, Koyama M, Pawankar R,
Matumoto K. Fibroin allergy: IgE mediated hypersensitivity
to silk suture materials. Nippon Ika Daigaku Zasshi 1999;66:
41–4.
[31] Rossitch Jr. E, Bullard DE, Oakes WJ. Delayed foreign-body
reaction to silk sutures in pediatric neurosurgical patients. Childs
Nerv Syst 1987;3:375–8.
[32] Dewair M, Baur X, Ziegler K. Use of immunoblot technique for
detection of human IgE and IgG antibodies to individual silk
proteins. J Allergy Clin Immunol 1985;76:537–42.
[33] Zaoming W, Codina R, Fernandez-Caldas E, Lockey RF. Partial
characterization of the silk allergens in mulberry silk extract. J Inv
Allergy Clin Immunol 1996;6:237–41.
[34] Soong HK, Kenyon KR. Adverse reactions to virgin silk sutures
in cataract surgery. Ophthalmology 1984;91:479–83.
G.H. Altman et al. / Biomaterials 24 (2003) 401–416 415
[35] Postlethwait RW. Long-term comparative study of nonabsorb-
able sutures. Ann Surg 1970;171:892–8.
[36] Postlethwait RW. Tissue reaction to surgical sutures. In: Dumphy
JE, Van Winkle W, editors. Repair and regeneration. New York:
McGraw-Hill, 1969. p. 263–85.
[37] Postlethwait RW, Dillon MD, Reeves JW. Experimental study of
silk suture. Arch Surg 1962;84:698–702.
[38] Morrow FA, Kogan SJ, Freed SZ, Laufman H. In vivo
comparison of polyglycolic acid, chromic catgut and silk in tissue
of the genitourinary tract: an experimental study of tissue retrieval
and calculogenesis. J Urol 1974;112:655–8.
[39] Nebel L, Rosenberg G, Tobias B, Nathan H. Autograft suture in
peripheral nerves. Eur Surg Res 1977;9:224–34.
[40] Peleg H, Rao UN, Emrich LJ. An experimental comparison of
suture materials for tracheal and bronchial anastomoses. J Thorac
Cardiovasc Surg 1986;34:384–8.
[41] Lam KH, Nijenhuis AJ, Bartels H, Postema AR, Jonkman MF,
Pennings AJ, Nieuwenhuis P. Reinforced poly(l-Lactic Acid)
fibers as suture material. J Appl Biomater 1995;6:191–7.
[42] Anderson JM. Inflammation, wound healing, and the foreign
body response. In: Ratner BD, Hoffman AS, Schoen FJ, Lemons
JE, editors. Biomaterials science: an introduction to materials in
medicine. Boston: Academic Press, 1996. p. 165–72.
[43] Foschi D, Corsi F, Cellerino P, Rizzi A, Morandi E, Trabucchi E.
Angiogenic effects of suture biomaterials. An experimental study
in rats. Eur Surg Res 2001;33:16–20.
[44] Setzen G, Williams EF. Tissue response to suture materials
implanted subcutaneously in a rabbit model. Plast Reconstr Surg
1997;100:1788–95.
[45] Bucknall TE, Teare L, Ellis H. The choice of a suture to close
abdominal incisions. Eur Surg Res 1983;15:59–66.
[46] Wong PC, Crain EJ, Nguan O, Watson CA, Racanelli A.
Antithrombotic actions of selective inhibitors of blood coagula-
tion factor Xa in rat models of thrombosis. Thromb Res
1996;83:117–26.
[47] Dahlke H, Dociu N, Thurau K. Thrombogenicity of different
suture materials as revealed by scanning electron microscopy.
J Biomed Mater Res 1980;14:251–68.
[48] Lee KY, Kong SJ, Park WH, Ha WS, Kwon IC. Effect of surface
properties on the antithrombogenicity of silk fibroin/S-carbox-
ymethyl kerateine blend films. J Biomater Sci Polym Ed
1998;9:905–14.
[49] Sakabe H, Ito H, Miyamoto T, Noishiki Y, Ha WS. In vivo blood
compatibility of regenerated silk fibroin. Sen-I Gakkaishi
1989;45:487–90.
[50] Santin M, Motta A, Freddi G, Cannas M. In vitro evaluation of
the inflammatory potential of the silk fibroin. J Biomed Mater
Res 1999;46:382–9.
[51] Uff CR, Scott AD, Pockley AG, Phillips RK. Influence of soluble
suture factors on in vitro macrophage function. Biomaterials
1995;16:355–60.
[52] Salthouse TN, Matlaga BF, Wykoff MH. Comparative tissue
response to six suture materials in rabbit cornea, sclera, and
ocular muscle. Am J Ophthalmol 1977;84:224–33.
[53] Ethicon, Inc., Wound closure manual. The suture: specific
suturing materials. Non-absorbable sutures. Somerville, NJ:
Ethicon, Inc. [copyright 2000, Chapter 2].
[54] Minoura N, Tsukada M, Nagura M. Physico-chemical properties
of silk fibroin membrane as a biomaterial. Biomaterials
1990;11:430–4.
[55] Asakura T, Demura M, Date T, Miyashita N, Ogawa K,
Williamson MP. NMR study of silk I structure of Bombyx mori
silk fibroin with N-15- and C-13-NMR chemical shift contour
plots. Biopolymers 1997;41:193–203.
[56] Kurioka A, Yamazaki M, Hirano H. Primary structure and
possible functions of a trypsin inhibitor of Bombyx mori. Eur J
Biochem 1999;259:120–6.
[57] Greenwald D, Shumway S, Albear P, Gottlieb L. Mechanical
comparison of 10 suture materials before and after in vivo
incubation. J Surg Res 1994;56:372–7.
[58] Kourie JI, Shorthouse AA. Properties of cytotoxic peptide-
formed ion channels. Am J Physiol Cell Physiol 2000;278:
C1063–87.
[59] Inouye K, Kurokawa M, Nishikawa S, Tsukada M. Use of
Bombyx mori silk fibroin as a substratum for cultivation of animal
cells. J Biochem Biophys Meth 1998;37:159–64.
[60] Sofia S, McCarthy MB, Gronowicz G, Kaplan DL. Functiona-
lized silk-based biomaterials for bone formation. J Biomed Mater
Res 2001;54:139–48.
[61] Altman GH, Horan RL, Lu H, Moreau J, Martin I, Richmond
JC, Kaplan DL. Silk matrix for tissue engineered anterior cruciate
ligaments. Biomaterials 2002;23:4131–41.
[62] Altman GH, Horan RL, Martin I, Farhadi J, Stark PRH, Volloch
V, Richmond JC, Vunjak-Novakovic, Kaplan DL. Cell differ-
entiation by mechanical stress. FASEB J 2001;16:270–2 (print)
and 10:1096 (electronic).
[63] Dunn MG, Tria AJ, Kato P, Bechler JR, Ochner RS, Zawadski
JP, Silver FH. Anterior cruciate ligament reconstruction using a
composite collagenous prosthesis. Am J Sports Med 1992;20:
507–15.
[64] Dunn MG, Bellincampi LD, Tria AJ, Zawadski JP. Preliminary
development of a collagen-PLA composite for ACL reconstruc-
tion. J Appl Polym Sci 1997;63:1423–8.
[65] Zarkoob S, Reneker DH, Eby RK, Hudson SD, Ertley D, Adams
WW. Generation of synthetic elastin-mimetic small diameter
fibers and fiber networks. Polym Preprints Am Chem Soc
1998;39:244–5.
[66] Jin H-J, Fridrikh SV, Rutledge GC, Kaplan DL. Electrospinning
Bombyx mori silk with poly(ethylene oxide). Biomacromolecules,
in press.
[67] Matthews JA, Wnek GE, Simpson DG, Bowlin GL. Electrospin-
ning of collagen nanofibers. Biomacromolecules 2002;3:232–8.
[68] Perez-Rigueiro J, Viney C, Llorca J, Elices M. Mechanical
properties of single-brin silkworm silk. J Appl Polym Sci
2000;75:1270–7.
[69] Pins GD, Christiansen DL, Patel R, Silver FH. Self-assembly of
collagen fibers: influence of fibrillar alignment and decorin on
mechanical properties. Biophys J 1997;73:2164–72.
[70] Engelberg I, Kohn J. Physicomechanical properties of degradable
polymers used in medical applications: a comparative study.
Biomaterials 1991;12:292–304.
[71] Gosline JM, Guerette PA, Ortlepp CS, Savage KN. The
mechanical design of spider silks: from fibroin sequence to
mechanical function. J Exp Biol 1999;202(Pt 23):3295–303.
G.H. Altman et al. / Biomaterials 24 (2003) 401–416416