Content uploaded by Samuel Veres
Author content
All content in this area was uploaded by Samuel Veres on Oct 25, 2018
Content may be subject to copyright.
Full length article
Collagen fibrils in functionally distinct tendons have differing structural
responses to tendon rupture and fatigue loading
Tyler W. Herod
a
, Neil C. Chambers
b
, Samuel P. Veres
a,b,
⇑
a
School of Biomedical Engineering, Dalhousie University, Halifax, Nova Scotia, Canada
b
Division of Engineering, Saint Mary’s University, Halifax, Nova Scotia, Canada
article info
Article history:
Received 4 January 2016
Received in revised form 3 June 2016
Accepted 10 June 2016
Available online 14 June 2016
Keywords:
Collagen fibril
Structure-function
Mechanics
Tendon
Damage
abstract
In this study we investigate relationships between the nanoscale structure of collagen fibrils and the
macroscale functional response of collagenous tissues. To do so, we study two functionally distinct classes
of tendons, positional tendons and energy storing tendons, using a bovine forelimb model. Molecular-level
assessment using differential scanning calorimetry (DSC), functional crosslink assessment using
hydrothermal isometric tension (HIT) analysis, and ultrastructural assessment using scanning electron
microscopy (SEM) were used to study undamaged, ruptured, and cyclically loaded samples from the two
tendon types. HIT indicated differences in both crosslink type and crosslink density, with flexor tendons
having more thermally stable crosslinks than the extensor tendons (higher T
Fmax
of >90 vs. 75.1 ± 2.7 !C),
and greater total crosslink density than the extensor tendons (higher t
1/2
of 11.5 ± 1.9 vs. 3.5 ± 1.0 h after
NaBH
4
treatment). Despite having a lower crosslink density than flexor tendons, extensor tendons were
significantly stronger (37.6 ± 8.1 vs. 23.1 ± 7.7 MPa) and tougher (14.3 ± 3.6 vs. 6.8 ± 3.4 MJ/m
3
). SEM
showed that collagen fibrils in the tougher, stronger extensor tendons were able to undergo remarkable
levels of plastic deformation in the form of discrete plasticity, while those in the flexor tendons were
not able to plastically deform. When cyclically loaded, collagen fibrils in extensor tendons accumulated
fatigue damage rapidly in the form of kink bands, while those in flexor tendons did not accumulate signif-
icant fatigue damage. The results demonstrate that collagen fibrils in functionally distinct tendons respond
differently to mechanical loading, and suggests that fibrillar collagens may be subject to a strength vs.
fatigue resistance tradeoff.
Statement of Significance
Collagen fibrils—nanoscale biological cables—are the fundamental load-bearing elements of all structural
human tissues. While all collagen fibrils share common features, such as being composed of a precise
quarter-staggered polymeric arrangement of triple-helical collagen molecules, their structure can vary
significantly between tissue types, and even between different anatomical structures of the same tissue
type. To understand normal function, homeostasis, and disease of collagenous tissues requires detailed
knowledge of collagen fibril structure-function. Using anatomically proximate but structurally distinct
tendons, we show that collagen fibrils in functionally distinct tendons have differing susceptibilities to
damage under both tensile overload and cyclic fatigue loading. Our results suggest that the structure
of collagen fibrils may lead to a strength versus fatigue resistance tradeoff, where high strength is gained
at the expense of fatigue resistance, and vice versa.
"2016 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction
Collagen fibrils are the fundamental load bearing element of
structural tissues within the human body. They are only nanome-
ters in diameter (range !30–500 nm), but have lengths on the
order of millimeters [1], making them essentially nanoscale biolog-
ical cables. Each fibril is composed of a precise quarter-staggered
arrangement of triple-helical collagen molecules [2], which are
covalently crosslinked to one another [3], conferring strength to
the fibril [4,5], and stability to the molecules [6,7].
Because collagen fibrils are an integral part of so many tissues
within the body, understanding their response to functional
http://dx.doi.org/10.1016/j.actbio.2016.06.017
1742-7061/"2016 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
⇑
Corresponding author at: Division of Engineering, Saint Mary’s University, 923
Robie Street, Halifax, NS B3H 3C3, Canada.
E-mail address: sam.veres@smu.ca (S.P. Veres).
Acta Biomaterialia 42 (2016) 296–307
Contents lists available at ScienceDirect
Acta Biomaterialia
journal homepage: www.elsevier.com/locate/actabiomat
loading in both health and disease is of great importance. Linking
macroscale tissue response to nanoscale collagen fibril response
is necessary to understand the workings of collagenous tissues,
but is an extremely difficult process due to the disparate length
scales involved. To aid in understanding collagen structure-
function relations, tendons are commonly used as a study subject,
because they have the most basic architecture amongst the col-
lagenous tissues (though tendon architecture is by no means
simple).
Tendons are intermediate tissue structures that connect and
transmit tensile forces from muscle to bone. They are mostly
composed of water (50–60% of their wet weight [8,9]) and colla-
gen (70–85% of their dry weight [8–10]). While structural com-
plexities involving several levels of hierarchy exist including
fascicles and fibres [11], tendons are essentially a collection of
parallel, longitudinally aligned collagen fibrils. The forces trans-
mitted by tendons from muscle to bone are carried by their lon-
gitudinal collagen fibrils [12–14]. To date, most studies have
investigated tendon structure-function using a single tendon
model. Many studies have, and continue to use, the rat tail ten-
don [15–18]. Some research groups have adopted different mod-
els: the rat patellar tendon [19,20], the mouse supraspinatus
tendon [21,22], the bovine tail tendon [8,23]. All of these models
share the same basic structural features: the tendons are com-
posed of distinct fascicles, which are in turn composed of longitu-
dinally aligned collagen fibrils that are arranged in distinct fibres.
How generalizable are the results from any one model though?
How much does the structure of collagen fibrils differ between
tendons, and how large an impact can this have on functional
response?
Previous research has already demonstrated that important
structural and mechanical differences exist between different ten-
dons in the same individual. Two distinct classes of tendon are now
recognized: those that act to position limbs under little load (posi-
tional tendons), and those that drive locomotion by storing strain
energy under large stresses and subsequently releasing it (energy
storing tendons) [24]. Using equine and bovine models, it has been
shown that energy storing tendons have thinner collagen fibrils
[25], a higher concentration of sulfated glycosaminoglycans [24],
and greater cellularity [24] but slower collagen turnover [26] com-
pared to positional tendons from the same forelimb. During elon-
gation, energy storing tendons show greater levels of sliding
between fascicles [27], but reduced levels of fibre sliding within
fascicles [28]. Although energy storing tendons experience larger
tensile stresses in vivo [29], and demonstrate superior fatigue resis-
tance [30], positional tendons have greater stiffness and strength
[27].
While positional and energy storing tendons are clearly very
different in function, structure, and mechanics, there remains
much to be learned about how the differing mechanics of these
tendons relates to their structures, and how these structures
serve their tendons’ specialized functions. This is particularly
true at the nanoscale, where little is currently known about
how differences in collagen fibril structure may contribute to
the differences in macroscale mechanics observed between ten-
don types.
In this study, we use pair-matched positional and energy stor-
ing tendons from bovine forelimbs to show that significant nanos-
tructural differences can exist between anatomically proximate
tendons. We subsequently show that these structural differences
result in significant differences in mechanical properties. We then
go on to relate the mechanical differences observed on the macro-
scale to the nanoscale structural differences observed in the ten-
dons’ collagen fibrils. Through this process we show that
macroscale tendon strength may be related to collagen fibril plas-
ticity, and that collagen fibril fatigue resistance may be related to
crosslinking and molecular stability. Our results have interesting
potential implications for numerous collagenous tissues in the
form of a potential strength vs. fatigue resistance tradeoff that
may exist for fibrillar collagens.
2. Materials and methods
An overview of the experimental procedure is shown in Fig. 1
Remove tendons from bovine forelimbs (n = 7)
Each tendon cut into longitudinal subsamples:
Molecular
stability
and crosslink
assessment via
HIT testing
Assessment of
molecular stability
via calorimetry
Assessment of
nanostructure via
electron microscopy
Energy storing tendon:
supercial digital
exor
(n = 7)
Positional tendon:
common digital
extensor
(n = 7)
Precondition: 10 cycles to 10% strain
Rupture
500 cycles 1000 cycles
Rupture
Stress
30% of Rupture
Stress for:
Discard clamped ends
Subdivide
Fig. 1. Experimental procedure used to assess the molecular and nanostructural
consequences of rupture and fatigue loading in two functionally distinct classes of
tendon: positional tendons and energy storing tendons. In this study, the two
classes of tendon were represented by using common digital extensor tendons
(positional) and superficial digital flexor tendons (energy storing) from bovine
forelimbs.
T.W. Herod et al. / Acta Biomaterialia 42 (2016) 296–307 297
2.1. Tissue collection & sample preparation
Forelimbs from 24 to 36 month-old adult steers killed for food
were collected from a local abattoir. Only one forelimb per animal
was collected. The common digital extensor tendon and superficial
digital flexor tendon (hereinafter frequently referred to simply as
the ‘‘extensor tendon” and ‘‘flexor tendon”) were removed, laid flat
between sheets of gauze moistened with phosphate buffered saline
(PBS), double bagged, and stored at "86 !C.
In preparation for testing, tendons were removed from frozen
storage and allowed to thaw at room temperature in their sealed
bags. For this study, the extensor and flexor tendons from seven
forelimbs taken from seven different animals were used. Each ten-
don was cut to a length of 45 mm, and then cut longitudinally to
produce several parallel, matched-pair samples that ranged in
diameter from 3 to 5 mm. Samples were assigned to the following
five experimental groups: crosslink analysis, control (precondition-
ing only), rupture, 500 cycle fatigue, 1000 cycle fatigue.
2.2. Molecular stability & crosslink assessment
Samples from all seven extensor/flexor tendon pairs (n = 7)
were assessed for molecular stability and functional crosslinking
using hydrothermal isometric tension testing (HIT), as described
previously [31–33]. Each tendon sample was cut to a length of
15 mm and fixed using clamps to a load cell at one end and a fixed
support at the other end. Once mounted, tendon samples were
submerged in room temperature distilled, deionized water. A ten-
sile preload of 60 g was applied to each sample. Using a hotplate,
the temperature of the water bath containing the isometrically
constrained samples was increased at a constant rate to 90 !C. Dur-
ing the temperature ramp, time, temperature, and load data were
recorded at 0.2 Hz. Analyses of the resulting data were conducted
using Microsoft Excel (Office for Mac 2011), and each sample’s
denaturation temperature (T
d
) and temperature of maximum force
generation (T
Fmax
) were recorded (Fig. 2A).
Samples from an additional seven extensor/flexor tendon pairs
(n = 7), collected from seven different 24–36 month-old steers,
were used HIT thermal relaxation assessment, which provides an
indication of crosslink density. Four to six 15-mm-long longitudi-
nal samples were cut from the centre of each tendon. In order to
convert all crosslinks present to a heat-stable form, two samples
from each tendon underwent crosslink stabilization treatment
using sodium borohydride, as described previously [34]. Each sam-
ple underwent four 15-min rinses at 4 !C in 100 mL of a borate buf-
fer solution (pH 9.0) containing 0.1 mg/mL sodium borohydride,
under continuous agitation. Each sample was then rinsed twice
and stored in distilled water. The untreated and crosslink stabilized
samples were fixed in the HIT system and preloaded as described
above. The distilled, deionized water bath containing the isometri-
cally constrained samples was heated to 90 !C(Fig. 2B), and then
maintained at 90 ± 0.6 !C for three hours. During the isotherm,
time, temperature and load data were recorded at 0.2 Hz. For each
sample, a plot of ln(load/maximum load) vs. time was created, and
a linear line was fit to a 3000-s data interval starting at 1000 s
(Fig. 2C). The slope of the best fit line was used to calculate each
sample’s half-time of load decay (t
1/2
), as described previously
[33]. The half-times for the duplicate samples were then averaged
to produce a single value for each tendon.
2.3. Mechanical testing
Mechanical testing was conducted using a servo-hydraulic
materials testing system (MTS 458-series, MTS, Eden Prairie,
MN), controlled using custom software written in LabVIEW (ver-
sion 2010, National Instruments, Austin, TX). All testing was con-
ducted in a room-temperature PBS bath containing 1% antibiotic
antimycotic (anti-anti) solution (product #A5955, Sigma Aldrich,
St. Louis, MO). Prior to testing, each tendon sample was vertically
suspended by one end and photographed four times at 90!incre-
ments of axial rotation. The photos were then used to calculate
Force (g)
0
200
400
600
800
Temperature (°C)
50 55 60 65 70 75 80 85 90
Representative HIT Responses for Paired Bovine Forelimb
Extensor (positional) & Flexor (energy storing) Tendons:
Native, Untreated Tissue
Force (g)
0
100
200
300
400
500
Temperature (°C)
50 55 60 65 70 75 80 85 90
Paired HIT Responses after Crosslink Stabilization
via NaBH4 Treatment
Ln(Force/Max Force)
1.0
0.8
0.6
0.4
0.2
0
Time (s)
0 1000 2000 3000 4000 5000 6000
Load Decay during 90°C Isotherm for Paired Tendons
after NaBH4 Crosslink Stabilization
A
B
C
extensor tendon
extensor tendon
extensor tendon
Td
TFmax
Fig. 2. Representative load-temperature responses during hydrothermal isometric
tension (HIT) analysis for common digital extensor and superficial digital flexor
tendons taken from the same bovine forelimb. (A) Positional extensor tendons had
significantly lower denaturation temperatures (T
d
), indicating that they contained
collagen molecules with lower molecular stability compared to the energy storing
flexor tendons. The decline in load of extensor tendons well before 90 !C(T
Fmax
< 90 !C) indicated the absence of significant levels of thermally stable crosslinking.
(B) After borohydride treatment, extensor tendons no longer failed prior to 90 !C,
indicating the successful conversion of heat-labile crosslinks to a heat-stable form.
(C) For crosslink stabilized tissue, load decay was much slower during the 90 !C
isotherm for flexor tendons, indicating a much greater total crosslink density than
in the extensor tendons.
298 T.W. Herod et al. / Acta Biomaterialia 42 (2016) 296–307
the sample’s elliptical cross-sectional area. Samples were mounted
in the materials testing system using crush grips lined with 180
grit waterproof sandpaper. Samples were mounted with visible
slack, and then extended until a tensile load of 0.2 N was regis-
tered. The intergrip length of the samples at this position, which
varied from 9 to 11 mm, was taken as the initial, zero load length.
All samples that underwent mechanical testing were first precon-
ditioned by applying 10 loading cycles to a peak strain of 10%,
and all mechanical testing was conducted using a strain rate of
10%/s. Samples in the control group (n = 5 extensor/flexor pairs)
received preconditioning only. Samples in the rupture group
(n = 7 pairs) were preconditioned, and then pulled to rupture at a
strain rate of 10%/s. The ultimate tensile strength of each rupture
sample was then calculated. Samples in the 500 cycle fatigue group
(n = 5 pairs) were preconditioned, and then loaded at 10%/s to 30%
of the ultimate tensile strength of their matched-pair rupture sam-
ple 500 times. Samples in the 1000 cycle fatigue group (n = 4 pairs)
were tested the same way, but for 1000 cycles instead of 500.
After testing, the gripped ends of each sample were removed
using a razor blade. Samples were then prepared for differential
scanning calorimetry (DSC) and scanning electron microscopy
(SEM), as described below.
2.4. Differential scanning calorimetry
After mechanical testing, the structural state of collagen mole-
cules within the tendons was assessed using DSC. Immediately fol-
lowing mechanical testing a !2 mm length was removed from one
end of each tendon sample. These DSC samples were stored at 4 !C
in sealed tubes containing fresh PBS with anti-anti for 24–36 h
prior to DSC testing.
For both the extensor and flexor tendons, five control samples,
six ruptured samples, and eight fatigue samples (five 500 cycle
and three 1000 cycle) were assessed using DSC. For the DSC test,
each sample was blotted dry to remove excess surface liquid and
divided again, yielding subsamples with wet weights of
12.0 ± 3.0 mg. Each subsample was gently pressed into the bottom
of an aluminum sample pan to maximize pan/sample contact area.
The pans were then hermetically sealed. Each sample was run
against an empty pan using a Q200 differential scanning calorime-
ter (TA Instruments, New Castle, DE). Samples were equilibrated at
40 !C, and then ramped to 90 !C at 5 !C/min while data were
recorded at 5 Hz. After the DSC runs were complete, the sample
pans were pierced and dried in an oven at 110 !C for 24 h. Pans
were then reweighed, and the dry weight of each sample was
recorded.
Endotherms were analyzed using Universal Analysis 2000 soft-
ware (version 4.5A, TA Instruments) for onset temperature (T
onset
),
peak temperature (T
peak
), full-width at half-maximum (FWHM),
and specific enthalpy of denaturation (
D
h) calculated based on
dry sample weight (Fig. 3).
2.5. Scanning electron microscopy
After mechanical testing, samples were also prepared for ultra-
structural assessment using SEM. From both the extensor and
flexor tendons, four control samples, four ruptured samples, and
five fatigue samples (three 500 cycle and two 1000 cycle) were
assessed. After mechanical testing, SEM samples were stored over-
night at 4 !C in fresh PBS with anti-anti. The following day, samples
underwent three, fifteen-minute rinses in distilled water, and were
then fixed for one hour in 2.5% SEM-grade glutaraldehyde. After
fixation, the samples were again triple-rinsed in distilled water.
Using a razor blade, each sample was bisected longitudinally to
expose its interior. The samples were then dehydrated in graded
ethanol, critical point dried, mounted on SEM stubs using carbon
tape with their internal, exposed surfaces facing upward, and
coated using gold-palladium. The samples were then examined
at magnifications of up to 90,000X using a Hitashi S-4700 scanning
electron microscope operating at 3 kV, 15
l
m.
For the control samples, both crimp wavelength and fibril diam-
eter were measured using the resulting SEM images. All measure-
ments were done using ImageJ software (version 1.44o, National
Institutes of Health). Crimp wavelength measurements were made
at least four different locations in each sample. At each location,
the total length of at least three consecutive waveforms was mea-
sured and used to calculate the average crimp wavelength at that
location. Fibril diameters were measured at least six locations in
each sample, with a minimum of 75 fibrils measured per sample.
The multiple measurements made on each sample were averaged,
producing a single measurement of crimp and fibril diameter for
each tendon.
2.6. Statistics
Statistical analyses were conducted using JMP software (version
11.0, SAS Institute Inc., Cary, NC). Data are presented as
mean ± standard deviation. For HIT data from the untreated exten-
sor and flexor tendons, samples were marked as having ‘‘survived”
the HIT test if they had a T
Fmax
P90 !C. The proportions of extensor
and flexor tendons that survived the HIT test were compared using
Fisher’s exact test. All other data were analyzed using repeated
measures ANOVA, followed by matched-pair Student’s t-tests. For
comparisons of HIT t
1/2
data, and DSC data between different tests
within each tendon type, the critical p-value required for signifi-
cance was reduced using a Bonferroni adjustment to p= 0.0167.
3. Results
3.1. Extensor tendons (positional) and flexor tendons (energy storing)
have significantly different ultrastructures
Four common digital extensor tendons and four superficial dig-
ital flexor tendons that had undergone preconditioning only were
assessed using SEM. Extensor and flexor tendons both contained
longitudinally aligned collagen fibrils organized into well-defined
bundles that had waveform crimp (Fig. S1). The wavelength of
Tonset
-4
-3
-2
-1
0
Heat Flow (W/g)
Tem p e ra t u re (° C )
Exo Up
30 40 50 60 70 80 90
FWHM
Representative DSC Responses for Paired
Flexor & Extensor Tendons
extensor tendon
Tpeak
Δh
Fig. 3. Representative power-temperature responses during differential scanning
calorimetry for digital flexor and extensor tendons taken from the same bovine
forelimb. The responses shown are from samples in the Control group, which
received mechanical preconditioning only. Compared to extensor tendons,
matched-pair analyses (Table 1) showed that flexor tendons were composed of
collagen molecules with a greater and more uniform thermal stability (higher T
onset
and T
peak
, and lower FWHM).
T.W. Herod et al. / Acta Biomaterialia 42 (2016) 296–307 299
the collagen crimp in extensor tendons was 124 ± 7
l
m, signifi-
cantly longer than the crimp wavelength in flexor tendons, which
measured 57 ± 10
l
m(p= 0.004). At the nanoscale, the crimp in
both extensor and flexor tendons was produced by smooth curves
in the collagen fibrils and not by fibril kinking (Fig. 4), important in
regard to the details of fatigue damage, presented below.
Collagen fibrils in both extensor and flexor tendons showed
clear
D
-banding (Fig. 4, bottom panel). The collagen fibrils in flexor
tendons were significantly thinner than those in the extensor ten-
dons. Average fibril diameter in the flexor tendons was 80 ± 7 nm,
while that in the extensor tendons was 134 ± 5 nm (p= 0.0004).
Histogram plots of all fibrils measured showed that extensor
tendons contained an interspersed mixture of small (70 nm) and
large (160 nm) collagen fibrils, with more of the large fibrils pre-
sent (Fig. S2, upper panel). Flexor tendons, on the other hand, con-
tained a large proportion of small diameter fibrils (35–75 nm)
mixed with relatively few larger diameter fibrils (Fig. S2, bottom
panel).
Other differences in the matrix of extensor and flexor tendons
were also apparent. The collagen fibrils of extensor tendons often
had little to no additional types of visible matrix covering or
attached to them (Fig. 4, bottom panel). In flexor tendons, on the
other hand, collagen fibrils were often covered in a thick webbing
of matrix composed of thin filaments measuring approximately
20 nm in diameter that lacked
D
-banding. Though poorly aligned,
the filaments generally ran across collagen fibrils, connecting them
laterally (Fig. 5).
3.2. Collagen in extensor and flexor tendons has significantly different
molecular stability and crosslinking
Representative HIT curves for extensor/flexor tendon pairs from
the same forelimbs are shown in Fig. 2.
HIT denaturation temperature of flexor tendons was signifi-
cantly higher than for the paired extensor tendons (Table 1), indi-
cating that the flexor tendons were composed of collagen
molecules with greater thermal stability. The dominant type of
crosslinking was also different between the two tendons: extensor
tendons generated maximum force during HIT at 75.1 ± 2.7 !C,
indicating that their collagen molecules were mostly joined by
immature aldimine crosslinks, which are heat labile [35,36]. The
force generated by all seven flexor tendons, on the other hand,
was still increasing at 90 !C, indicating that their collagen mole-
cules were mostly joined by either immature ketoamine crosslinks,
or mature trivalent crosslinks, both of which are heat stable [36].
During thermal relaxation in HIT, load decayed significantly
more slowly in crosslink stabilized flexor tendons
(t
1/2
= 11.5 ± 1.9 h) than in crosslink stabilized extensor tendons
(t
1/2
= 3.5 ± 1.0 h; p< 0.0001), indicating that total crosslinking
density was significantly higher in the flexor tendons (Table 1).
For the flexor tendons, NaBH
4
treatment significantly slowed the
rate of thermal relaxation compared to untreated flexor samples
(t
1/2
values of 11.5 ± 1.9 vs. 6.3 ± 2.4 h; p= 0.0074), indicating that
the flexor tendons contained a significant population of heat labile
crosslinks in addition to heat stable crosslinks.
Representative DSC curves of an extensor/flexor pair from the
same bovine forelimb are shown in Fig. 3. Endotherm onset and
peak temperatures for flexor tendons were significantly higher
than for extensor tendons (Table 1), again indicating greater
molecular stability in the flexor tendons. Endotherm full-width at
half-maximum was significantly smaller for flexor tendons com-
pared to extensor tendons, indicating greater uniformity in the col-
lagen structure. Flexor tendons also had a significantly lower
specific enthalpy of denaturation than extensor tendons.
3.3. Extensor tendons are significantly stronger and tougher than
flexor tendons
The tensile mechanical properties of matched-pair flexor and
extensor tendons were compared using the samples that had been
pulled to rupture. The positional extensor tendons, which HIT had
shown to be less crosslinked, were significantly stronger and
Fig. 4. In both control extensor and flexor tendon samples, crimp was produced by
smooth curves in the collagen fibrils, as shown in this series of micrographs taken at
progressively higher magnification. Compare to the sharp kink bands in extensor
tendons caused by fatigue loading (Fig. 9).
300 T.W. Herod et al. / Acta Biomaterialia 42 (2016) 296–307
tougher than matched-pair energy storing flexor tendons (Table 2).
Representative stress-strain curves from each tendon type are
shown in Supplementary Fig. S3.
3.4. Tendon strength and toughness may be related to collagen fibril
plasticity
Tendon samples that had been pulled to rupture were assessed
using scanning electron microscopy. In the stronger, tougher
extensor tendons, but not in the energy storing flexor tendons,
extensive regions of collagen fibrils that had undergone discrete
plasticity were found (Fig. 6, top and middle panels). These
strain-damaged fibrils, not present in the control samples, were
characterized by the presence of longitudinally repeating kinks.
Regions containing fibrils with discrete plasticity often had little
to no visible microscale crimp remaining, despite the samples hav-
ing been allowed to recover unloaded in PBS for a minimum of 15 h
between rupture and glutaraldehyde fixation the next day.
Unlike the positional extensor tendons, rupturing energy stor-
ing flexor tendons at 10%/s did not produce collagen fibrils with
discrete plasticity. Some regions with what appeared to be
strain-damaged fibrils were found (Fig. 6, bottom panel), however,
they were rare and the fibrils lacked the systematic longitudinal
kinking characteristic of discrete plasticity.
The ruptured flexor and extensor samples were also analyzed
using DSC, and compared to control samples (preconditioned only)
from the same tendons. Consistent with our SEM observations,
rupture had a detrimental effect on the thermal stability of colla-
gen molecules within extensor tendons, but not flexor tendons.
For extensor tendons only, rupture decreased the stability of the
least stable collagen molecules within the tendons, significantly
decreasing endotherm onset temperature (Fig. 7). For flexor ten-
dons, which did not show discrete plasticity under SEM, the ther-
mal stability of collagen molecules remained unchanged
following rupture.
3.5. Collagen fibrils in flexor tendons show greater resistance to fatigue
loading than those in extensor tendons
Mechanical data for the four pairs of extensor and flexor sam-
ples subjected to 1000 cycles of fatigue loading are shown in
Fig. 8. Extensor tendons had significantly greater modulus of elas-
ticity, peak strain, and hysteresis than the flexor tendons. Com-
pared to their responses during the first loading cycle, the two
tendon types underwent statistically similar fractional changes in
modulus and peak strain (Fig. 8, right column). Hysteresis, which
was greater in the extensor tendons, underwent a greater frac-
tional decrease in the extensor tendons compared to the flexor
tendons.
Despite undergoing similar relative changes in mechanical
properties during cyclic loading, SEM analysis showed an accumu-
lation of fatigue damage to the collagen fibrils in extensor tendons,
but not in flexor tendons. In extensor tendons that underwent both
500 and 1000 cycles of fatigue loading, kink bands were found at
locations throughout the tendons (Figs. 9–11,S4). Kink bands were
formed by fibrils that had been laterally kinked in register with one
another, sometimes along a single plane (Fig. 9), and other times
along multiple planes (Fig. 11). At high magnification, individual
fibril kinks making up the kink band often had a twisted or folded
appearance (Fig. 9C). Occasionally, more highly disrupted (Fig. S4)
and even broken fibrils (Fig. 10) were found within kink bands.
The collagen fibrils in flexor tendons appeared to be highly
resistant to fatigue damage: kink bands were rarely observed in
the 1000 cycle samples, and never seen in the 500 cycle samples.
Evidence of matrix shearing was the only form of fatigue damage
frequently observed in the energy storing flexor tendons (Fig. 12).
In addition to kink bands, the extensor tendons that had under-
gone 1000 cycles of fatigue loading also contained fibrils with dis-
crete plasticity damage (Fig. 11), consistent with those seen in the
ruptured extensor tendons (Fig. 6). The presence of discrete plas-
Fig. 5. Collagen fibrils in the energy storing flexor tendons were frequently coated
in a webbing of filamentous matrix that connected the fibrils laterally to one
another. Little of this lateral webbing was observed in positional extensor tendons,
where collagen fibrils were generally free of significant quantities of other forms of
matrix (Fig. 4, bottom panel).
Table 1
Hydrothermal isometric tension (HIT) data and differential scanning calorimetry
(DSC) data indicate that flexor and extensor tendons have significantly different
molecular stabilities and intermolecular collagen crosslinking.
Common digital
extensor tendon
Superficial digital
flexor tendon
HIT data
Native, untreated tissue:
T
d
(!C) 62.7 ± 0.4 65.4 ± 0.7 p< 0.0001
T
Fmax
(!C) 75.1 ± 2.7 >90 p= 0.0006
t
1/2
(hrs) – 6.3 ± 2.4
NaBH
4
crosslink stabilized tissue:
t
1/2
(hrs) 3.5 ± 1.0 11.5 ± 1.9 p< 0.0001
DSC data
T
onset
(!C) 63.1 ± 1.0 64.4 ± 0.7 p= 0.0041
T
peak
(!C) 65.4 ± 0.7 65.8 ± 0.8 p= 0.0421
FWHM (!C) 2.7 ± 0.5 1.7 ± 0.1 p= 0.0071
D
h(J/g dry
mass)
58.0 ± 13.7 40.7 ± 10.1 p= 0.0478
Samples used for HIT crosslink assessment were not mechanically tested. For this
table, DSC data is from control samples that underwent mechanical preconditioning
only.
Table 2
When pulled to rupture at 10%/s, positional common digital extensor tendons were
significantly stronger and tougher than energy storing superficial digital flexor
tendons from the same bovine forelimb.
Common digital
extensor tendon
Superficial digital
flexor tendon
Young’s modulus, E
(MPa)
150.7 ± 46.3 102.0 ± 33.9 p> 0.1
Yield strain (%) 32.5 ± 12.4 38.2 ± 18.2 p> 0.5
Yield strength,
r
y
(MPa)
29.7 ± 8.9 19.4 ± 6.5 p= 0.0148
Ultimate strength,
r
u
(MPa)
37.6 ± 8.1 23.1 ± 7.7 p= 0.0038
Modulus of
toughness, U
(MJ/m
3
)
14.3 ± 3.6 6.8 ± 3.4 p= 0.0108
T.W. Herod et al. / Acta Biomaterialia 42 (2016) 296–307 301
ticity suggests that these fibrils had reached their normal limit of
elongation, and that the tendons were perhaps close to entering
the tertiary region of creep leading to failure. Compared to the
Fig. 6. Top & Middle: In the stronger, tougher extensor tendons, rupture at 10%/s
created extensive regions of fibrils with discrete plasticity damage, characterized by
kinks that repeated along the length of individual fibrils with nanoscale spatial
frequency. The ability of fibrils to undergo discrete plasticity, as seen here, may
increase tendon load bearing capacity: discrete plasticity was not observed in the
weaker energy storing flexor tendons after rupture. Bottom: Rupture at 10%/s did
not alter the structure of fibrils in the flexor tendons notably. In rare instances,
regions with fibrils that appeared to have been plastically damaged were found, as
shown here. Unlike the ruptured extensor tendons, though, regions of fibrils with
discrete plasticity were not found.
Onset Temperature (°C)
60
62
64
66
Extensor Flexor
Peak Temperature (°C)
64
65
66
67
Extensor Flexor
FWHM (°C)
1
2
3
4
5
Extensor Flexor
Enthalpy (J/g dry weight)
0
20
40
60
80
100
Extensor Flexor
control
rupture
fatigue
*
Fig. 7. DSC-based structural assessment of collagen molecules within digital
extensor and flexor tendons after rupture and fatigue loading compared to
preconditioned controls. Only rupture affected enough collagen molecules to
produce a signal change in DSC, and only in the positional extensor tendons. For the
extensor tendons, the decrease in endotherm onset temperature indicates that
rupture increased the lateral spacing between collagen molecules. This change is
likely due to the formation of discrete plasticity damage during rupture (Fig. 6),
which occurred in extensor tendons but not in flexor tendons.
*
p60.0167.
302 T.W. Herod et al. / Acta Biomaterialia 42 (2016) 296–307
ruptured extensor tendons, far fewer regions of fibrils with discrete
plasticity were observed, and when found, tended to be less expan-
sive in the lateral direction.
The cyclically loaded extensor and flexor tendons were also
assessed using DSC. No significant change to the structure of colla-
gen molecules within either tendon type occurred as the result of
cyclic loading (Fig. 7), consistent with SEM observations of little
to no fibril damage in the flexor tendons, and highly localized fibril
damage within the extensor tendons.
4. Discussion
In this study we have shown that anatomically proximate ten-
dons can have dramatically different collagen structures, leading
to significant differences in their response to tensile loading, rup-
ture, and fatigue. The common digital extensor tendon from bovine
forelimb has high strength and toughness, which may result from
the ability of its collagen fibrils to plastically deform when over-
loaded. However, this tendon rapidly accumulates fatigue damage
when cyclically loaded. The situation is reversed for the superficial
digital flexor tendon from bovine forelimb: its collagen fibrils do
not readily incur fatigue damage, possibly from a higher level of
intermolecular crosslinking and a network of filamentous webbing
that joins fibrils laterally, but is significantly weaker and less tough
than the extensor tendon. These results pose the intriguing ques-
tion of whether the structure of collagen fibrils may be subject to
a strength vs. fatigue resistance tradeoff, where higher strength
can only be gained at the detriment of fatigue resistance, and vice
versa.
Both bovine and equine common digital extensor and superfi-
cial digital flexor forelimb tendons have been used previously as
examples of positional and energy storing tendons. The equine
flexor tendon is indeed a true energy storing tendon, stretching
by 10–15% during galloping [27,37,38] and storing and releasing
considerable energy [39]. While similar data is not available for
the bovine forelimb model, measurements of muscle cross-
sectional area indicate that bovine flexor tendons experience
significantly more stress than extensor tendons: 69 MPa vs.
8–11 MPa, respectively [29]. These values are very close to those
reported for the equine forelimb analogues [29].
Despite sustaining significantly lower tensile stresses in vivo
[29], in this study we found that common digital extensor tendons
from bovine forelimbs were significantly stronger and tougher
than superficial digital flexor tendons (Table 2). Similar results
have been reported previously, for both individual fascicles from
bovine forelimb tendons [30] and whole equine forelimb tendons
[27]. The significant differences in strength and toughness between
positional and energy storing tendons are most likely caused by
the structural differences in their collagen fibrils, particularly dif-
ferences in intermolecular crosslinking. While we did not perform
chemical crosslink analysis in the current study, which we
acknowledge as a limitation of the present work, our hydrothermal
isometric tension (HIT) analyses clearly demonstrate that signifi-
cant differences in collagen crosslinking exist between these two
Extensor Tendon Flexor Tendon
Elastic Modulus (MPa)
50
100
150
200
250
Loading Cycle
1 500 1000 1 500 1000
Peak Strain (mm/mm)
0.15
0.20
0.25
0.30
0.35
1 500 1000 1 500 1000
Hysteresis (%)
0
10
20
30
40
1 500 1000 1 500 1000
Extensor Tendon Flexor Tendon
Fraction of Initial Modulus
1.0
1.2
1.4
1.6
1.8
Loading Cycle
1 500 1000 1 500 1000
Fraction of Initial Peak Strain
1.0
1.1
1.2
1.3
1 500 1000 1 500 1000
Fraction of Initial Hysteresis
0
0.2
0.4
0.6
0.8
1.0
1 500 1000 1 500 1000
Fig. 8. Mechanical data for flexor and extensor tendon samples subject to cyclic creep loading to 30% of their ultimate stresses for 1000 cycles. Extensor tendons generally had
a greater modulus of elasticity, achieved larger peak strains, and dissipated a greater proportion of the input strain energy during cyclic loading compared to the flexor
tendons. While peak strain data indicate that both extensor and flexor samples were within the secondary creep region after 1000 loading cycles, SEM observation suggests
that extensor samples were probably close to entering the tertiary creep region (Fig. 11).
T.W. Herod et al. / Acta Biomaterialia 42 (2016) 296–307 303
tendons. First, our T
Fmax
data show that the dominant species of
crosslinking in flexor and extensor tendons are chemically distinct.
After the denaturation temperature, the linear increase in tension
with temperature for untreated flexor tendons demonstrates that
the intermolecular crosslinks joining dissociated
a
-chains are not
heat labile, meaning that they are either mature, trivalent cross-
links, or divalent keto-amine crosslinks [36]. Previous work
Fig. 9. In the positional extensor tendons, cyclic loading caused the development of
kink bands (arrows in A point to three kink bands). Kink bands were readily found
in extensor tendons that had undergone both 1000 cycles (A) and 500 cycles (B, C)
of loading. The kink bands were caused by a laterally aligned folding of individual
collagen fibrils (C). Kink bands were not found in control samples, and were rarely
seen in the cyclically loaded flexor tendons.
Fig. 10. The fate of fibrils within kink bands: fatigue rupture. In occasional cases,
the broken ends of fibrils were seen within the kink bands that formed in extensor
tendons (red arrows in lower image). (For interpretation of the references to colour
in this figure legend, the reader is referred to the web version of this article.)
Fig. 11. In addition to kink bands (yellow arrows), extensor tendons that
underwent 1000 cycles also contained fibrils that had undergone discrete plasticity
(note the repeating kinks present along the length of individual fibrils—a phenom-
ena quite distinct from the kink band). The presence of fibrils with discrete
plasticity indicates that these tendons were likely in or nearing the tertiary phase of
creep, leading to rupture. (For interpretation of the references to colour in this
figure legend, the reader is referred to the web version of this article.)
304 T.W. Herod et al. / Acta Biomaterialia 42 (2016) 296–307
indicates that most are probably the mature, trivalent hydroxylysyl
pyridinoline crosslink [40]. In contrast, the rate of tension genera-
tion with increasing temperature declines quickly in untreated
extensor tendons (Fig. 2A), indicating an overwhelming presence
of immature aldimine crosslinks, which are easily broken via heat-
ing [35,36].
Second, our t
1/2
data indicate that the energy storing digital
flexor tendons had significantly greater total crosslinking density
than the anatomically proximate, positional extensor tendons.
When held under isometric constraint at 90 !C using HIT, the ten-
sion accumulated in a collagenous tissue during the proceeding
temperature ramp declines following a Maxwell decay as hydroly-
sis of the collagen molecule polypeptide
a
-chains occur [41]. The
rate of load decay is dependent on the concentration of heat stable
crosslinks: tissues with a higher concentration of heat stable cross-
links will experience slower load decay. For example, maturation
and aging of many collagenous tissues is accompanied by an
increase in heat stable crosslinking as divalent crosslinks are con-
verted into mature, trivalent crosslinks [40,42]. Accordingly, the
rate of load decay in HIT decreases with increasing tissue age
[34,41,43,44]. Exogenous crosslinking has also been shown to slow
isothermal load decay [45]. The relationship between crosslinking
and load decay in HIT can be directly shown by employing borohy-
dride crosslink stabilization. Treating collagen with borohydride
chemically reduces heat labile, immature, aldimine crosslinks to
a heat stable form [35,46]. For tissues with these crosslinks, boro-
hydride treatment increases the total number of crosslinks present
during the isotherm, thereby reducing the rate of load decay
[34,44]. The utility of borohydride treatment is best shown in the
present study by the extensor tendons, where following stabiliza-
tion they no longer failed during the temperature ramp to 90 !C
(Fig. 2A vs. B). Comparing the rate of load decay for crosslink
stabilized tissues thus provides a relative measure of the total con-
centration of crosslinking present—immature, mature, and non-
enzymatic, combined. For the crosslink stabilized tissue tested in
the present study, load decay occurred more than three times
slower for the flexor tendons than it did for extensor tendons
(Fig 2C; Table 1), indicating greater overall levels of crosslinking
in the flexor tendons.
The inhibition of enzymatic crosslinking leads to a dramatic
reduction in the strength of collagenous tissues. This has been well
shown in studies that have used experimental lathyrism. Intake of
b-aminopropionitrile inhibits lysyl oxidase [47], preventing the
formation of the lysine and allysine aldehydes required to form
intermolecular crosslinks. The tail tendons of rats raised on diets
supplemented with b-aminopropionitrile show marked reductions
in ultimate strength, despite having normal collagen content and
normally sized fibrils [4,48]. Similar results have been reported
for other tissues, including the aorta [49], and cortical bone [50].
Given that inhibiting enzymatic crosslinking reduces tendon
strength [4,48], one would expect that increasing crosslinking den-
sity would strengthen a tendon. Yet, the more highly crosslinked
superficial digital flexor tendons tested in the current study were
weaker and less tough than the less crosslinked common digital
extensor tendons. Hansen et al. [51] reported similar findings for
fascicles from human patellar tendons: posterior fascicles were
significantly weaker than anterior fascicles despite having higher
levels of crosslinking. Previous research has shown that during
extension, greater levels of fibre sliding occur within the fascicles
of equine extensor tendons compared to flexor tendons [28].
Reduced fibre sliding may prevent even load distribution between
fibrils, potentially lowering the strength of flexor tendons despite a
high level of intermolecular crosslinking within fibrils.
In addition to the important differences in fascicle and fibre
sliding characteristics of energy storing and positional tendons
[27,28], differences in fibril extensibility may also contribute to
the observed differences in strength and toughness. As opposed
to simply rupturing when overloaded, collagen fibrils that are able
to deform plastically could help to evenly distribute force within a
tendon. Our SEM analyses support this idea. Following rupture, the
stronger, tougher extensor tendons contained large regions of col-
lagen fibrils that had undergone discrete plasticity (Fig. 6, top and
middle panels), as observed previously in other positional tendons
[23,52,53]. In contrast, rupture of the weaker flexor tendons did
not produce significant levels of plastic damage within collagen
fibrils. While a high density of enzymatic crosslinking probably
increases the strength of individual collagen fibrils, it may also
reduce fibril plasticity, therefore compromising load sharing abil-
ity, and hence tendon strength.
Despite being strong and tough, positional extensor tendons
rapidly accumulated fatigue damage during cyclic loading in the
form of kink bands (Figs. 9–11,S4). Kink band formation has been
well studied in numerous fields ranging from geology to fibre-
reinforced composites, and can occur in a variety of anisotropic
materials due to compressive stress [54–57]. Rather than the
repeated application of tension, the presence of kink bands sug-
gests that fatigue damage in the extensor tendons was created dur-
ing the unloading phase of each loading cycle. With the peak strain
achieved by the samples continuously increasing during testing
Fig. 12. Within the energy storing flexor tendons, the only form of fatigue damage
frequently found after 1000 cycles of loading to 30% of ultimate stress was matrix
shearing, as shown here. Note the sharp change in fibril direction in the centre of
the lower image, indicating that neighboring fibrils are well adhered to one another.
Also note the filamentous webbing covering the fibrils, a feature frequently seen in
the flexor tendons but not in the extensor tendons.
T.W. Herod et al. / Acta Biomaterialia 42 (2016) 296–307 305
(Fig. 8), but each loading cycle ending at the original zero displace-
ment mark, compressive forces resulting in matrix shearing and
fibre buckling could certainly occur. At high magnification, the kink
bands were produced by individual, laterally aligned folds in the
collagen fibrils (Fig. 9C), indicating significant disruption to molec-
ular packing at these locations. Had loading continued beyond
1000 cycles, further disruption (Fig. S4) and eventual breakage of
fibrils within the kink bands (Fig. 10) would likely have reduced
the load bearing capacity of the tendon during elongation, leading
to tensile overload and the development of discrete plasticity
(Fig. 11).
In contrast to the rapid development of kink bands in the posi-
tional extensor tendons, 1000 cycles of loading caused minimal
damage to fibrils in the energy storing flexor tendons, despite
flexor tendons being significantly weaker in tension. It is interest-
ing to note that in vivo, collagen in positional tendons appears to be
turned over significantly faster than in energy storing tendons
[26,58], possibly indicating a faster accumulation of fibril damage
and subsequent remodeling of the tendon structure. The difference
in fatigue response between the two tendons found in this study is
consistent with previous results showing that fascicles from bovine
digital flexor tendons have a greater fatigue life than those from
digital extensor tendons [30]. Considering the mode of fatigue fail-
ure in extensor tendons and the structural differences between
flexor and extensor tendons, the fatigue resistance of energy stor-
ing tendons may arise from two sources. First, collagen fibrils in
the superficial digital flexor tendons examined in this study
appeared to be bundled together laterally by filamentous webbing
that lacked
D
-banding (Fig. 5). While the nature of this material
was not identified, it appears similar to reported images of type
VI collagen [59], which is known to exist in tendon [60]. The pres-
ence of a filamentous meshwork connecting fibrils laterally could
significantly increase the bending stiffness of collagen fibres,
thereby increasing the resistance of individual fibrils to buckling.
Second, in a similar fashion but within individual fibrils, the higher
intermolecular crosslinking levels present in flexor tendon fibrils
would increase fibril bending stiffness by increasing the resistance
to molecular shearing. These two mechanisms, acting at different
length scales, may account for the reduced susceptibility of energy
storing flexor tendons to fatigue damage.
Tendinopathy is a common cause of disability in developed
countries [61–64]. There is good evidence to support that
tendinopathy develops as the result of repetitive mechanical stress
[65–67], as has been anecdotally discussed for many years, hence
terms like ‘‘jumpers knee” and ‘‘tennis elbow”. While mechanically
triggered, predisposition to tendinopathy may be genetically
determined [68–71]. It is possible that energy storing tendons in
certain individuals do not possess the fibril-level structural archi-
tectures necessary to adequately sustain repeated loading at high
stress. Interestingly, features similar to the kink bands observed
in this study have recently been reported in tendinopathic human
Achilles tendons [72]. Although several studies have shown that
physical training can alter the structure of tendons, with an aver-
age collagen half-life of nearly 200 years in energy storing tendons
any remodeling that occurs would appear to be very gradual [26].
Further, mechanical loading alone may not be capable of stimulat-
ing the conversion to a more fatigue-appropriate structure. Prelim-
inary results from our laboratory indicate that the structures of
energy storing and positional tendons begin to diverge, at least
in cattle, during fetal development in the absence of differences
in mechanical loading [73]. Understanding the development of
these two classes of tendon in greater detail could aid in our under-
standing of why certain tissues in certain individuals are sensitive
to mechanical injury. It would also be of great benefit to the field of
tissue engineering, the optimization of surgical graft selection, and
the design of injury prevention therapies.
In conclusion, positional and energy storing tendons have sig-
nificantly different collagen fibril architectures. These differences
lead to significant differences in tensile strength and toughness,
and susceptibility to fibril-level fatigue damage during cyclic load-
ing. It is possible that optimization for strength and fatigue resis-
tance in fibrillar collagen may be mutually exclusive, where high
strength can only be gained at the expense of fatigue resistance,
and vice versa. Understanding the development of these structural
differences, and how they vary between individuals may help in
understanding the development of repetitive stress injuries, choos-
ing replacement tissues, and designing novel therapies.
Acknowledgements
This work was supported by a grant to SPV from the Natural
Sciences and Engineering Research Council of Canada (NSERC).
NCC is grateful for the support from NSERC in the form of an
Undergraduate Student Research Award. TWH acknowledges fund-
ing support from the Government of Nova Scotia. For our use of
SEM and DSC, we acknowledge the support of the Canada Founda-
tion for Innovation, the Atlantic Innovation Fund, and other part-
ners which fund the Facilities for Materials Characterization,
managed by the Institute for Research in Materials, Dalhousie
University.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.actbio.2016.06.
017.
References
[1] N.S. Kalson, Y. Lu, S.H. Taylor, T. Starborg, D.F. Holmes, K.E. Kadler, A structure-
based extracellular matrix expansion mechanism of fibrous tissue growth,
eLife 4 (2015) e05958.
[2] J.P.R.O. Orgel, T.C. Irving, A. Miller, T.J. Wess, Microfibrillar structure of type I
collagen in situ, Proc. Natl. Acad. Sci. U.S.A. 103 (2006) 9001–9005.
[3] D.R. Eyre, J.-J. Wu, Collagen cross-links, Top. Curr. Chem. 247 (2005) 207–229.
[4] R.C. Haut, The effect of a lathyritic diet on the sensitivity of tendon to strain
rate, J. Biomech. Eng. 107 (1985) 166.
[5] P. Hansen, T. Hassenkam, R.B. Svensson, P. Aagaard, T. Trappe, B.T. Haraldsson,
M. Kjaer, P. Magnusson, Glutaraldehyde cross-linking of tendon—mechanical
effects at the level of the tendon fascicle and fibril, Connect Tissue Res. 50
(2009) 211–222.
[6] C.A. Miles, M. Ghelashvili, Polymer-in-a-box mechanism for the thermal
stabilization of collagen molecules in fibers, Biophys. J. 76 (1999) 3243–3252.
[7] C.A. Miles, N.C. Avery, V.V. Rodin, A.J. Bailey, The increase in denaturation
temperature following cross-linking of collagen is caused by dehydration of
the fibres, J. Mol. Biol. 346 (2005) 551–556.
[8] T.L. Willett, R.S. Labow, N.C. Avery, J.M. Lee, Increased proteolysis of collagen in
an in vitro tensile overload tendon model, Ann. Biomed. Eng. 35 (2007) 1961–
1972.
[9] A.P. Rumian, A.L. Wallace, H.L. Birch, Tendons and ligaments are anatomically
distinct but overlap in molecular and morphological features—a comparative
study in an ovine model, J. Orthop. Res. 25 (2007) 458–464.
[10] D. Amiel, C. Frank, F. Harwood, J. Fronek, W. Akeson, Tendons and ligaments: a
morphological and biochemical comparison, J. Orthop. Res. 1 (1984) 257–265.
[11] J. Kastelic, A. Galeski, E. Baer, The multicomposite structure of tendon, Connect
Tissue Res. 6 (1978) 11–23.
[12] R.B. Svensson, T. Hassenkam, P. Hansen, M. Kjaer, S.P. Magnusson, Tensile force
transmission in human patellar tendon fascicles is not mediated by
glycosaminoglycans, Connect Tissue Res. 52 (2011) 415–421.
[13] H.R.C. Screen, J.C. Shelton, V.H. Chhaya, M.V. Kayser, D.L. Bader, D.A. Lee, The
influence of noncollagenous matrix components on the micromechanical
environment of tendon fascicles, Ann. Biomed. Eng. 33 (2005) 1090–1099.
[14] G. Fessel, J.G. Snedeker, Evidence against proteoglycan mediated collagen fibril
load transmission and dynamic viscoelasticity in tendon, Matrix Biol. 28
(2009) 503–510.
[15] S. Torp, E. Baer, B. Friedman, Effects of age and of mechanical deformation on
the ultrastructure of tendon, Colston Papers 26 (1975) 223–250.
[16] T.M. Grant, C. Yapp, Q. Chen, J.T. Czernuszka, M.S. Thompson, The mechanical,
structural, and compositional changes of tendon exposed to elastase, Ann.
Biomed. Eng. 43 (2015) 2477–2486.
[17] J. Kastelic, E. Baer, Deformation in tendon collagen, Symp. Soc. Exp. Biol. 34
(1980) 397.
306 T.W. Herod et al. / Acta Biomaterialia 42 (2016) 296–307
[18] G. Fessel, Y. Li, V. Diederich, M. Guizar-Sicairos, P. Schneider, D.R. Sell, V.M.
Monnier, J.G. Snedeker, Advanced glycation end-products reduce collagen
molecular sliding to affect collagen fibril damage mechanisms but not
stiffness, PLoS One 9 (2014) e110948.
[19] H.B. Sun, N. Andarawis-Puri, Y. Li, D.T. Fung, J.Y. Lee, V.M. Wang, J. Basta-
Pljakic, D.J. Leong, J.B. Sereysky, S.J. Ros, R.A. Klug, J. Braman, M.B. Schaffler, K.J.
Jepsen, E.L. Flatow, Cycle-dependent matrix remodeling gene expression
response in fatigue-loaded rat patellar tendons, J. Orthop. Res. 28 (2010)
1380–1386.
[20] D.T. Fung, J.B. Sereysky, J. Basta-Pljakic, D.M. Laudier, R. Huq, K.J. Jepsen, M.B.
Schaffler, E.L. Flatow, Second harmonic generation imaging and fourier
transform spectral analysis reveal damage in fatigue-loaded tendons, Ann.
Biomed. Eng. 38 (2010) 1741–1751.
[21] B.K. Connizzo, J.J. Sarver, L. Han, L.J. Soslowsky, In situ fibril stretch and sliding
is location-dependent in mouse supraspinatus tendons, J. Biomech. 47 (2014)
3794–3798.
[22] K.S. Miller, B.K. Connizzo, E. Feeney, L.J. Soslowsky, Characterizing local
collagen fiber re-alignment and crimp behavior throughout mechanical testing
in a mature mouse supraspinatus tendon model, J. Biomech. 45 (2012) 2061–
2065.
[23] S.P. Veres, J.M. Harrison, J.M. Lee, Repeated subrupture overload causes
progression of nanoscaled discrete plasticity damage in tendon collagen fibrils,
J. Orthop. Res. 31 (2012) 731–737.
[24] E.L. Batson, R.J. Paramour, T.J. Smith, H.L. Birch, J.C. Patterson-Kane, A.E.
Goodship, Are the material properties and matrix composition of equine flexor
and extensor tendons determined by their functions?, Equine Vet J. 35 (2003)
314–318.
[25] H.L. Birch, Tendon matrix composition and turnover in relation to functional
requirements, Int. J. Exp. Pathol. 88 (2007) 241–248.
[26] C.T. Thorpe, I. Streeter, G.L. Pinchbeck, A.E. Goodship, P.D. Clegg, H.L. Birch,
Aspartic acid racemization and collagen degradation markers reveal an
accumulation of damage in tendon collagen that is enhanced with aging, J.
Biol. Chem. 285 (2010) 15674–15681.
[27] C.T. Thorpe, C.P. Udeze, H.L. Birch, P.D. Clegg, H.R.C. Screen, Specialization of
tendon mechanical properties results from interfascicular differences, J. R. Soc.
Interface 9 (2012) 3108–3117.
[28] C.T. Thorpe, C. Klemt, G.P. Riley, H.L. Birch, P.D. Clegg, H.R.C. Screen, Helical
sub-structures in energy-storing tendons provide a possible mechanism for
efficient energy storage and return, Acta Biomater. 9 (2013) 7948–7956.
[29] R.F. Ker, R. Alexander, M.B. Bennett, Why are mammalian tendons so thick?, J
Zool. 216 (1988) 309–324.
[30] J.H. Shepherd, K. Legerlotz, T. Demirci, C. Klemt, G.P. Riley, H.R.C. Screen,
Functionally distinct tendon fascicles exhibit different creep and stress
relaxation behaviour, Proc. Inst. Mech. Eng. H 228 (2014) 49–59.
[31] J.M. Lee, C.A. Pereira, D. Abdulla, W.A. Naimark, I. Crawford, A multi-sample
denaturation temperature tester for collagenous biomaterials, Med. Eng. Phys.
17 (1995) 115–121.
[32] S.P. Veres, J.M. Harrison, J.M. Lee, Cross-link stabilization does not affect the
response of collagen molecules, fibrils, or tendons to tensile overload, J.
Orthop. Res. 31 (2013) 1907–1913.
[33] I.G. Aldous, S.P. Veres, A. Jahangir, J.M. Lee, Differences in collagen cross-
linking between the four valves of the bovine heart: a possible role in
adaptation to mechanical fatigue, Am. J. Physiol. Heart Circ. Physiol. 296
(2009) H1898–H1906.
[34] S.M. Wells, S.L. Adamson, B.L. Langille, J.M. Lee, Thermomechanical analysis of
collagen crosslinking in the developing ovine thoracic aorta, Biorheology 35
(1998) 399–414.
[35] A.J. Bailey, Intermediate labile intermolecular crosslinks in collagen fibres,
Biochim. Biophys. Acta 160 (1968) 447–453.
[36] N.C. Avery, A.J. Bailey, Restraining cross-links responsible for the mechanical
properties of collagen fibers: natural and artificial, Collagen (2008) 81–110.
[37] P.R. Stephens, D.M. Nunamaker, D.M. Butterweck, Application of a Hall-effect
transducer for measurement of tendon strains in horses, Am. J. Vet. Res. 50
(1989) 1089–1095.
[38] N.J. Dimery, R. Alexander, R.F. Ker, Elastic extension of leg tendons in the
locomotion of horses (Equus caballus), J. Zool. 210 (1986) 415–425.
[39] A.A. Biewener, Muscle-tendon stresses and elastic energy storage during
locomotion in the horse, Comp. Biochem. Physiol. B: Biochem. Mol. Biol. 120
(1998) 73–87.
[40] J.C. Patterson-Kane, D.A. Parry, H.L. Birch, A.E. Goodship, E.C. Firth, An age-
related study of morphology and cross-link composition of collagen fibrils in
the digital flexor tendons of young thoroughbred horses, Connect Tissue Res.
36 (1997) 253–260.
[41] M. Le Lous, J.C. Allain, L. Cohen-Solal, P. Maroteaux, The rate of collagen
maturation in rat and human skin, Connect Tissue Res. 9 (1982) 253–262.
[42] A.J. Bailey, R.G. Paul, L. Knott, Mechanisms of maturation and ageing of
collagen, Mech. Ageing Dev. 106 (1998) 1–56.
[43] M. Le Lous, L. Cohen-Solal, J.C. Allain, J. Bonaventure, P. Maroteaux, Age related
evolution of stable collagen reticulation in human skin, Connect Tissue Res. 13
(1985) 145–155.
[44] W.A. Naimark, S.D. Waldman, R.J. Anderson, B. Suzuki, C.A. Pereira, J.M. Lee,
Thermomechanical analysis of collagen crosslinking in the developing lamb
pericardium, Biorheology 35 (1998) 1–16.
[45] E.P. Brennan-Pierce, I. MacAskill, R.B. Price, J.M. Lee, Riboflavin-sensitized
photo-crosslinking of collagen using a dental curing light, Biomed. Mater. Eng.
24 (2014) 1659–1671.
[46] A.J. Bailey, D. Lister, Thermally labile cross-links in native collagen, Nature 220
(1968) 280–281.
[47] S.R. Pinnell, G.R. Martin, The cross-linking of collagen and elastin: enzymatic
conversion of lysine in peptide linkage to alpha-aminoadipic-delta-
semialdehyde (allysine) by an extract from bone, Proc. Natl. Acad. Sci. U.S.A.
61 (1968) 708–716.
[48] R. Puxkandl, I. Zizak, O. Paris, J. Keckes, W. Tesch, S. Bernstorff, P. Purslow, P.
Fratzl, Viscoelastic properties of collagen: synchrotron radiation investigations
and structural model, Philos. Trans. R. Soc. London, Ser. B 357 (2002) 191–197.
[49] A. Brüel, G. Ortoft, H. Oxlund, Inhibition of cross-links in collagen is associated
with reduced stiffness of the aorta in young rats, Atherosclerosis 140 (1998)
135–145.
[50] H. Oxlund, M. Barckman, G. Ortoft, T.T. Andreassen, Reduced concentrations of
collagen cross-links are associated with reduced strength of bone, Bone 17
(1995) 365S–371S.
[51] P. Hansen, B.T. Haraldsson, P. Aagaard, V. Kovanen, N.C. Avery, K. Qvortrup, J.O.
Larsen, M. Krogsgaard, M. Kjaer, S.P. Magnusson, Lower strength of the human
posterior patellar tendon seems unrelated to mature collagen cross-linking
and fibril morphology, J. Appl. Physiol. 108 (2010) 47–52.
[52] S.P. Veres, J.M. Harrison, J.M. Lee, Mechanically overloading collagen fibrils
uncoils collagen molecules, placing them in a stable, denatured state, Matrix
Biol. 33 (2014) 54–59.
[53] S.P. Veres, J.M. Lee, Designed to fail: a novel mode of collagen fibril disruption
and its relevance to tissue toughness, Biophys. J. 102 (2012) 2876–2884.
[54] M.A. Wadee, G.W. Hunt, M.A. Peletier, Kink band instability in layered
structures, J. Mech. Phys. Solids 52 (2004) 1071–1091.
[55] R.T. Faill, Kink-band folding, Valley and Ridge Province, Pennsylvania, Geol.
Soc. Am. Bull. 84 (1973) 1289–1314.
[56] T.J. Vogler, S. Kyriakides, On the initiation and growth of kink bands in fiber
composites: Part I. experiments, Int. J. Solids Struct. 38 (2001) 2639–2651.
[57] P.M. Moran, X.H. Liu, C.F. Shih, Kink band formation and band broadening in
fiber composites under compressive loading, Acta Metall. Mater. 43 (1995)
2943–2958.
[58] H.L. Birch, S. Worboys, S. Eissa, B. Jackson, S. Strassburg, P.D. Clegg, Matrix
metabolism rate differs in functionally distinct tendons, Matrix Biol. 27 (2008)
182–189.
[59] D.R. Keene, E. Engvall, R.W. Glanville, Ultrastructure of type VI collagen in
human skin and cartilage suggests an anchoring function for this filamentous
network, J. Cell Biol. 107 (1988) 1995–2006.
[60] A.D. Waggett, J.R. Ralphs, A.P. Kwan, D. Woodnutt, M. Benjamin,
Characterization of collagens and proteoglycans at the insertion of the
human Achilles tendon, Matrix Biol. 16 (1998) 457–470.
[61] C. Littlewood, S. May, S. Walters, Epidemiology of rotator cuff tendinopathy: a
systematic review, Shoulder Elbow 5 (2013) 256–265.
[62] T.L. Sanders, H. Maradit Kremers, A.J. Bryan, J.E. Ransom, J. Smith, B.F. Morrey,
The epidemiology and health care burden of tennis elbow: a population-based
study, Am. J. Sports Med. 43 (2015) 1066–1071.
[63] J. Zwerver, S.W. Bredeweg, I. van den Akker-Scheek, Prevalence of Jumper’s
knee among nonelite athletes from different sports: a cross-sectional survey,
Am. J. Sports Med. 39 (2011) 1984–1988.
[64] S. de Jonge, C. van den Berg, R.J. de Vos, H.J.L. van der Heide, A. Weir, J.A.N.
Verhaar, S.M.A. Bierma-Zeinstra, J.L. Tol, Incidence of midportion Achilles
tendinopathy in the general population, Br. J. Sports Med. 45 (2011) 1026–
1028.
[65] U.M. Kujala, S. Sarna, J. Kaprio, Cumulative incidence of achilles tendon
rupture and tendinopathy in male former elite athletes, Clin. J. Sport Med. 15
(2005) 133–135.
[66] O.B. Lian, L. Engebretsen, R. Bahr, Prevalence of jumper’s knee among elite
athletes from different sports: a cross-sectional study, Am. J. Sports Med. 33
(2005) 561–567.
[67] J. Toppi, J. Fairley, F.M. Cicuttini, J. Cook, S.R. Davis, R.J. Bell, F. Hanna, Y. Wang,
Factors associated with magnetic resonance imaging defined patellar
tendinopathy in community-based middle-aged women: a prospective
cohort study, BMC Musculoskelet. Disord. 16 (2015) 184.
[68] R.Z. Tashjian, J.M. Farnham, F.S. Albright, Evidence for an inherited
predisposition contributing to the risk for rotator cuff disease, J. Bone Joint
Surg. Am. 91 (2009) 1136–1142.
[69] P. Harvie, S.J. Ostlere, J. Teh, Genetic influences in the aetiology of tears of the
rotator cuff, J. Bone Joint Surg. Bm. 86-B (2004) 696–700.
[70] G.G. Mokone, M.P. Schwellnus, T.D. Noakes, M. Collins, The COL5A1 gene and
Achilles tendon pathology, Scand. J. Med. Sci. Sports 16 (2006) 19–26.
[71] G.G. Mokone, M. Gajjar, A.V. September, The guanine-thymine dinucleotide
repeat polymorphism within the tenascin-C gene is associated with Achilles
tendon injuries, Am. J. Sports Med. 33 (2005) 1016–1021.
[72] J. Pingel, Y. Lu, T. Starborg, U. Fredberg, H. Langberg, A. Nedergaard, M. Weis, D.
Eyre, M. Kjaer, K.E. Kadler, 3-D ultrastructure and collagen composition of
healthy and overloaded human tendon: evidence of tenocyte and matrix
buckling, J. Anat. 224 (2014) 548–555.
[73] S. Sparavalo, C.A.M. Bray, T.M. Brock-Fisher, N.M. Easton, C.A. Guinard, S.M.
Wells, J.M. Lee, S.P. Veres, Structural differences between distinct tendon types
arise during fetal development, in: Biomedical Engineering Society, Library of
Annual Meeting Abstracts, 2015.
T.W. Herod et al. / Acta Biomaterialia 42 (2016) 296–307 307
Flexor Tendon Fibrils
393 fibrils
Number of Fibrils
0
5
10
15
20
25
30
35
Fibril Diameter (nm)
050 100 150 200 250
Extensor Tendon Fibrils
416 fibrils
Number of Fibrils
0
5
10
15
20
25
Fibril Diameter (nm)
50 100 150 200 250
extensor tendon
flexor tendon
Stress (MPa)
0
10
20
30
40
Strain (m/m)
00.2 0.4 0.6
Representative Stress-Strain Responses for Samples of
Extensor and Flexor Tendons from Bovine Forelimb
