A proteomic analysis of engineered tendon formation under dynamic mechanical loading in vitro.

Abstract

Previous studies have demonstrated the beneficial effect of mechanical loading on in vitro tendon engineering. To understand the mechanism, human tenocytes and polyglycolic acid long fibers were used for in vitro tendon engineering in a bioreactor system for 12 weeks with and without dynamic loading. The engineered neo-tendons were subjected to proteomic analysis using mass spectrometry along with shotgun strategy. As expected, mechanical loading resulted in a more mature tendon tissue characterized by a firmer tissue texture and densely deposited matrices which formed longitudinally aligned collagen fibers in a highly compact fashion. In contrast, non-loaded neo-tendon revealed loosely and less deposited matrices in a relatively less organized pattern. Proteins isolated from two groups of tissues exhibited similar distribution of isoeletric point and molecular weight indicating the similarity and comparability of the tissue specimens. Further, proteomic analysis showed that total 758 proteins were identified from both groups with 194 and 177 proteins uniquely presented in loaded and non-loaded tendons, respectively. Comparison of loaded and non-loaded tendons revealed 195 significantly up-regulated proteins and 189 significantly down-regulated proteins. The differentially expressed proteins could generally be classified into the categories of extracellular matrix, intra-cellular signaling, cytoskeleton and inflammatory response. Among them, significantly up-regulated collagens I and VI, MMP-14, WNT5A, microfilament molecules and some inflammatory factors suggest that the possible mechanism for this particular biological phenomenon may involve increased production of tendon specific matrices, enhanced cross-link of collagens and other matrix molecules, proper matrix remodeling for tissue maturation and mechanotransduction (including non-canonical Wnt signal pathway) mediated other biological processes.

Full text

Page 1

A proteomic analysis of engineered tendon formation under dynamic mechanical
loading in vitro
Yongkang Jianga,1, Hongwei Liub,1, Hong Lic,d, Fangjun Wangb, Kai Chengb, Guangdong Zhoua,c,
Wenjie Zhanga,c, Mingliang Yeb, Yinlin Caoa,c, Wei Liua,c,**, Hanfa Zoub,*
aDepartment of Plastic and Reconstructive Surgery, Shanghai 9th People’s Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai Key Laboratory of Tissue Engineering
Research, 639 Zhi Zao Ju Road, Shanghai 200011, PR China
bNational Chromatography R&A Centre, CAS Key Lab of Separation for Analytical Chemistry, Dalian Institute of Chemical Physics, CAS, Dalian 116023, PR China
cNational Tissue Engineering Center of China, Shanghai, PR China
dCollege of Life Information Science & Instrument Engineering, Hangzhou Dianzi University, Hangzhou 310018, PR China
a r t i c l e i n f o
Article history:
Received 9 January 2011
Accepted 15 February 2011
Available online 12 March 2011
Keywords:
Mechanical loading
Proteomics
Human tenocytes
Tendon engineering
In vitro
a b s t r a c t
Previous studies have demonstrated the beneficial effect of mechanical loading on in vitro tendon
engineering. To understand the mechanism, human tenocytes and polyglycolic acid long fibers were used
for in vitro tendon engineering in a bioreactor system for 12 weeks with and without dynamic loading.
The engineered neo-tendons were subjected to proteomic analysis using mass spectrometry along with
shotgun strategy. As expected, mechanical loading resulted in a more mature tendon tissue characterized
by a firmer tissue texture and densely deposited matrices which formed longitudinally aligned collagen
fibers in a highly compact fashion. In contrast, non-loaded neo-tendon revealed loosely and less
deposited matrices in a relatively less organized pattern. Proteins isolated from two groups of tissues
exhibited similar distribution of isoeletric point and molecular weight indicating the similarity and
comparability of the tissue specimens. Further, proteomic analysis showed that total 758 proteins were
identified from both groups with 194 and 177 proteins uniquely presented in loaded and non-loaded
tendons, respectively. Comparison of loaded and non-loaded tendons revealed 195 significantly up-
regulated proteins and 189 significantly down-regulated proteins. The differentially expressed proteins
could generally be classified into the categories of extracellular matrix, intra-cellular signaling, cyto-
skeleton and inflammatory response. Among them, significantly up-regulated collagens I and VI,
MMP-14, WNT5A, microfilament molecules and some inflammatory factors suggest that the possible
mechanism for this particular biological phenomenon may involve increased production of tendon
specific matrices, enhanced cross-link of collagens and other matrix molecules, proper matrix remod-
eling for tissue maturation and mechanotransduction (including non-canonical Wnt signal pathway)
mediated other biological processes.
? 2011 Elsevier Ltd. All rights reserved.
1. Introduction
With the progress of tissue engineering research, engineered
tendon may offer a promising alternative for tendon repair and
regeneration [1]. Toward its ultimate goal of practical application,
mechanism studies remain the prerequisite in order to master the
technology.
As one of the major research subjects, the effect of mechanical
stimulationontendonengineeringhas beenwidely investigated for
its various roles in engineered tendon/ligament formation [2e5], in
vitro tendon engineering [6,7], maturation of engineered tendon
after in vivo implantation under mechanical loading [8] and cell
phenotype switch [9] or stem cell tenogenic differentiation [10,11].
Otherstudieshaveinvestigatedtheeffectofmechanicalstimulation
on cellular activities including cell proliferation, cell alignment and
extracellular matrix (ECM) accumulation, etc. [12e15].
Although these studies have the merit of revealing the impor-
tance of mechanical loading, they may not be able to fully elucidate
* Corresponding author. Dalian Institute of Chemical Physics, CAS, Dalian 116023,
PR China. Tel.: þ86 411 84379610; fax: þ86 411 84379620.
** Corresponding author. Department of Plastic and Reconstructive Surgery,
Shanghai 9th People’s Hospital, 639 Zhi Zao Ju Road, Shanghai 200011, PR China.
Tel.: þ86 21 23271699x5601; fax: þ86 21 53078128.
E-mail addresses: liuwei_2000@yahoo.com (W. Liu), hanfazou@dicp.ac.cn
(H. Zou).
1Authors who contributed equally.
Contents lists available at ScienceDirect
Biomaterials
journal homepage: www.elsevier.com/locate/biomaterials
0142-9612/$ e see front matter ? 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biomaterials.2011.02.033
Biomaterials 32 (2011) 4085e4095

Page 2

the mechanism due to the limitation of the used technologies. For
example, either RT-PCR or Western blot can only examine gene or
protein expression of some selected molecules; whereas the true
mechanism may involve the interactions among different mole-
cules that form a complex regulation network. Obviously, a high
throughput method, such as proteomics, is preferred to achieve
such an objective [16].
Previous studies demonstrated that tissue-based proteome is
able to decipher proteome complexities of tissue microenviron-
ment and deliver important information with appropriate patho-
logic and histological relevance [17e19]. Mass spectrometry (MS) is
a central analytical technique of the modern proteome analysis,
which allows identifying and characterizing thousands of proteins
in a complex mixture by combing with the chromatographic
separation prior to MS detection [20,21]. The “bottom up” proteo-
mics, which is also called as shotgun strategy [22,23] and is based
on digesting proteins into peptides and sequencing them using
tandem MS and automated database searching, has become the
method of choice for identifying proteins in most largescale studies
[24]. Shotgun proteome analysis is more compatible with liquid
chromatography and has much higher throughput in protein
identification and characterization, which makes it a popular
technology in current proteome analysis [25].
Ourpreviousstudiesdemonstratethatstaticmechanicalstrainis
important for in vitro engineered neo-tendon formation comparing
to non-strained tissue [6,7]. Further, our recent study revealed that
dynamic loading could engineer better quality tendon in vitro than
static strain. To decipher the proteome complexity of the protein
interactionnetworkrelatedto thisparticularbiological
phenomenon, the shotgun proteomics strategy was executed and
a semi-quantitative proteomics analysis was performed to compare
the protein expression profiles between dynamic loaded and non-
loaded neo-tendons in order to elucidate the potential mechanism.
2. Materials and methods
2.1. Isolation and culture of human tenocytes
Protocols for the handling of human tissue and cells were approved by the Ethics
Committee of Shanghai 9th People’s Hospital. Abandoned human tendon tissues
were donated for research purpose only from the patients who received amputation
operation. As previously described [7], tendons were minced into small fragments
followed by collagenase digestion to isolate human tenocytes. Cells were in vitro
cultured and expanded in Dulbecco’s Modified Eagle Medium (DMEM) containing
10% fetal bovine serum (FBS, Gibco), L-glutamine (292 mg/mL), penicillin (100 U/
mL), streptomycin (100 mg/mL), and ascorbic acid (50 mg/mL) [7].
2.2. Preparation of cell-scaffold constructs
Long polyglycolic acid (PGA) fibers provided by Shanghai Ju Rui Biomaterials
Company Inc (Shanghai, China) were used as the scaffold material. Briefly, 80 mg
PGA fibers (about 20 mm in diameter) were arranged into a cord with a length of
6 cm and a width of 1 cm, and then were secured on a custom-made stainless steel
frame (Fig.1B). The scaffold was sterilized by soaking in Povidone Iodine for 30 min
and 75% ethanol for 60 min. After washes in phosphate buffered saline (PBS) for 3
times, the PGA scaffold was pre-incubated in serum free DMEM overnight. Then
tenocytes were collected, suspended in DMEM culture medium at a density of
1.0 ? 107cells/mL followed by seeding onto the scaffold (1 mL per construct). The
cell-scaffold constructs were kept in an incubator up to 4 h to allow for complete
adhesion of the cells to the scaffold, and then culture medium was added followed
by static culture with medium change every other day as described [7].
Fig.1. Cell expansion and culture of cell-scaffold constructs. (A) Passage 2 human tenocytes exhibited a spindle shape and fibroblast-like morphology (bar ¼ 100 mm); (B) PGA long
fibers were secured on a custom-made stainless steel frame to serve as scaffolds; (C) A cell-scaffold construct was cultured in vitro for 1 week and examined under phase contrast
microscope (bar ¼ 100 mm); (D) SEM examination of a cell-scaffold construct cultured in vitro for 1 week (?500).
Y. Jiang et al. / Biomaterials 32 (2011) 4085e4095
4086

Page 3

2.3. Scanning electron microscopic (SEM) examination
As previously described [7], after in vitro culture for 7 days, part of the cell-
scaffold construct was prepared for SEM examination by pre-fixation with 2% glu-
taraldyhyde and post-fixation with 1% osmic acid for 2 h at 4?C respectively. After 2
washes in PBS, the samples were subjected to dehydration in gradient ethanol and
critical point drying (HCP-2, Hitachi, Japan). The samples were then mounted,
sputter-coated with gold (BAL-TEC, Philips, Eindhoven, Netherlands), and examined
with a scanning electron microscope (Philips-XL-30) to observe cell attachment and
matrix production on PGA scaffold.
2.4. In vitro tendon engineering with dynamic loading in a bioreactor
In order to provide a dynamic stretch to the cell-scaffold constructs, a custom-
made bioreactor system was designed and fabricated. The bioreactor system was
composed of power machine, a culture chamber and a control system. Generally,
through dynamic gas compression, the power machine could provide a force that
was transmitted via a plastic tube to the culture chamber to periodically drive
a sliding rod back and forth, and therefore to provide a periodical mechanical
stretch.
After 2 weeks of static culture, the cell-scaffold constructs were transferred from
the frames to the culture chamber of the bioreactor and secured by fixing to the
clamps connecting to a fixed plate at one end and to the sliding rod at the other end.
The mechanical loading was applied for 60 min a day with an amplitude of 1/10
length of the constructs, and with a frequency of 6 times per minute, each lasting for
5 s with an interval of 5 s.
Total 12 cell-scaffold constructs were included and divided into two groups: (1)
Experimental group: 6 constructs were subjected to a dynamic culture for 10 weeks
in a bioreactor after 2 weeks of static culture; (2) Control group: 6 constructs were
subject to static culture for 12 weeks constantly.
2.5. Macrographic and histological examinations
All specimens were harvested after 12 weeks of culture and examined for their
gross appearance and tissue texture and parts of engineered tendons from each
group were randomly selected for tissue section and H&E staining as previously
described [7].
2.6. Protein extraction and proteolysis
Three batches of duplicated engineered tendon tissues were collected and
combined, the proteins were extracted by the lysis buffer containing 8 M urea, 4%
CHAPS (w/v), 40 mM TriseCl, 20 mM dithiothreitol (DTT),1 mM phenylmethylsulfonyl
fluoride (PMSF),1 mM EDTA,10 mM NaF and 1 mM NaVO4in pH 7.4 at 4?C, and then
digested by trypsin in the same procedure as reported previously [26].
2.7. Online multidimensional mHPLC/MS/MS analysis
Online multidimensional mHPLC/MS/MS analysis (also termed as MudPIT) was
performed to analyze the tryptic digestions according to a reported method with
minor modifications as we previously reported [27] and both samples were
analyzedin triplicateinparallel. Generally, 40mg trypticdigestions were loaded onto
the phosphate monolith SCX trap column. Then, the SCX column was directly con-
nected to the reversed phase (RP) separation column. The peptides were gradually
eluted from SCX segment onto the RP separation column using a serial stepwise
elution, and then a 122 min binary RP separation gradient was explored to separate
the peptides retained onto the RP separation column. Once the RP gradient sepa-
ration started, the data acquisition of MS began immediately.
2.8. Data analysis and statistics
The acquired MS/MS spectra were searched using the Turbo SEQUEST (version
2.7) in the Bioworks 3.3 suite (Thermo, San Jose, CA) against a composite database
including both original and reversed human protein database of International
Protein Index (ipi. human 3.17.fasta, including 60234 entries, http://www.ebi.ac.uk/
IPI/IPIhuman.html).
Initial searching results were filtered with the following parameters as reported
previously [28]: the minimum Xcorr of 1.8 for a singly charged peptide; 2.5 for
a doubly charged peptide; and 3.5 for a triply charged peptide; the minimum delta
Cn cut-off value of 0.08; Furthermore, only the proteins identified in all three
parallel experiments for each sample with a relative standard deviation (RSD) of
peptide spectral counts less than 50% and more than two unique peptides measured
per protein were saved. For semi-quantitative comparison of the proteins identified
in Experimental group and Control group, spectral counts for each identified protein
from each of the groups were extracted, averaged, normalized and compared as
described previously and as noted in the footnotes for Supplementary Table 1
[28,29]. The biological process, cellular location and molecular function involved
for these identified proteins were assigned using Gominer, a web-based application
based on GeneOntology annotation (www.discovery.nci.nih.Gov/gominer),
a resource for biological interpretation of genomic and proteomic data [30,31].
Differentially expressed proteins with at least 5 fold differences in spectral count
ratio of loaded vs non-loaded samples were respectivelyassigned to up-regulated or
down-regulated categories. (For additional details of 2.6, 2.7 and 2.8, see
Supplemental Experimental Procedures.)
3. Result
3.1. Cell culture
Primary cultured human tenocytes exhibited a spindle and
fibroblast-like morphology (Fig. 1A) and proliferated rapidly by
reaching 80% confluence within 7 days. After subculture, the
passaged cells grew faster and could reach 80% confluence within 5
days. At passage 2, these cells were observed able to maintain good
proliferation and matrix production, and thus were used for tendon
engineering by seeding them on PGA fibers (Fig. 1B).
3.2. Phase contrast microscopic and SEM examination of cell-
scaffold constructs
The human tenocytes adhered to and spread along PGA fibers
within 48 h post-seeding when observed with phase contrast
microscopy. After 1 week of in vitro culture, microscopic observa-
tion also showed that rich matrices were produced by the seeded
tenocytes (Fig.1C). SEM examination further demonstrated that the
cells bound tightly to PGA fibers and produced rich extracelluar
matrices (Fig. 1D). All of these indicated that the PGA material had
good biocompatibility with seeded human tenocytes.
3.3. Macrographic examination and histology of engineered
tendons
There were clear differences in gross appearance and histolog-
ical structure between two groups. In the Experimental group with
dynamic loading, the effect of mechanical loading resulted in
thinner engineered tendons relative to those of the Control group
without loading. In addition, the engineered tendons became
firmer in their tissue texture with apparently enhanced mutuality
(Fig. 2A). In contrast, a relatively thicker neo-tendon was observed
with much soft tissue texture in the non-loading group (Fig. 2B).
This difference was also supported by the histological finding. As
shown, the engineered tendons with dynamic loading exhibited
a relatively more mature tissue structure with densely deposited
collagen matrices. Furthermore, the formed collagen fibers were
longitudinally aligned in a highly compact fashion, resembling the
tissue structure of a native tendon tissue (Fig. 2C). In contrast,
without an effective loading, the neo-tissue formed in the control
group revealed an immature tissue structure characterized by
loosely deposited collagens which were partially randomly
arranged and apparently disorganized (Fig. 2D). This distinctive
structural difference between two groups due to the mechanical
loading effect provided the solid tissue basis for further proteomic
analysis.
3.4. Protein extraction
In this study, three batches of duplicated cultured tissue were
used for the analysis. Total tissue weight collected from 6 samples
of the Experimental and Control groups was 68.0 mg and 61.3 mg
respectively. After the extraction bylysis bufferand precipitation by
acidic acetone, the corresponding quantity of protein extracted
from loaded and non-loaded groups was 0.68 mg and 1.2 mg
respectively, which were sufficient for triplicate proteomic tests for
each group.
Y. Jiang et al. / Biomaterials 32 (2011) 4085e4095
4087

Page 4

3.5. Protein identification and characterization
Online multidimensionalmHPLC/MS/MS analysiswas performed
in this study. As shown in Table 1, 564 and 581 unique proteins
were identified from non-loaded and loaded groups respectively.
Due to the same proteins that were identified in both groups, 758
proteins in total were actually found based on the cumulative
proteins identified from the two types of engineered tendons as
listed in Supplementary Table 1. Among them, 387 (51.0%) were
found in both tendons, while only 194 (25.6%) and 177 (23.4%) were
uniquely observed in loaded and non-loaded groups, respectively
(Table 1). The high overlap of the proteins identified in these
samples indicated the similarity and comparability of the samples.
This was also supported by the similar distribution of isoelectric
point and molecular weight of the proteins identified from each of
two groups (Supplementary Fig. 1). In addition, the accumulative
numbers of the identified unique peptides in total from loaded and
non-loaded groups were 3602 and 3456, respectively, which meant
that every protein was identified by averaged 6 unique peptides
with the range from 2 to more than 6 peptides. The large numbers
of spectral counts shown inTable 1 and multiple unique peptides to
identify a protein made the dataset much more reliable.
The spectral count has good correlation with the protein
abundance and it has been successfully applied for quantitative
analysis of many proteome samples [32]. By comparing the spectral
counts of the identified proteins between two groups,195 proteins
were found up-regulated and 189 proteins down-regulated, which
are listed in Supplementary Table 2. To better understand their
biological roles, these identified proteins were further classified
using Gene Ontology annotation [30,31] and were summarized
according to their molecular functions (Fig. 3A), biological
processes (Fig. 3B) and cellular locations (Fig. 3C). As shown in
Fig. 3, the binding and catalytic activities were the two major
molecular functions of both the up- and down-regulated proteins,
whereas physiological process and cellular process were the two
major biological processes; the major locations included both intra-
cellular and extracellular regions. In this study, proteomic analysis
was focused on four areas that are closely related to cellular
response to mechanical stimulation, including extracellular matrix,
intra-cellular signaling, cytoskeleton and inflammatory response.
3.6. Differential expression of extracellular matrix molecules
Collagens, proteoglycans and other molecules constitute the
extracellular matrices of tendon tissue and thus became the target
molecules for investigation and the results were summarized in
Fig. 4 (A: up-regulated; B: down-regulated; C: no significant
change) and in Supplementary Table 3-1. This study revealed that
dynamic loading could lead to up-regulated expression of collagen
molecules. For examples, significantly up-regulated expressions of
COL1A2 (Type I collagen alpha-2) and COL6A3i3 (Type VI collagen
alpha-3 isoform 3) were observed. Although only 2e3 folds
increase in spectral count ratio, the expressions of COL1A1 (Type I
collagen alpha-1), COl6A1 (Type VI collagen alpha 1), COL6A2i2c2
(Collagen VI alpha-2 isoform 2C2) and COL6A3i1 (Type VI collagen
alpha-3 isoform 1) were also up-regulated. There was no obvious
change of the expression of COL5A1 (Type V collagen alpha-1) and
COL14i1 (Type XIV collagen alpha-1 isoform 1).
Proteoglycans and some small molecules play important roles in
collagen fibrillogenesis and thus became the target molecules for
investigation. As shown in Fig. 4 and Supplementary Table 3-1,
significantly up-regulated expressions of other extracellular matrix
proteins wereobserved, and theyincluded DERM (Dermatopontin),
SPARC (secreted protein acidic and rich in cysteine), FBN2 (fibrillin-
2) and FMOD (fibromodulin). Proteins with more than 5 folds of
decreased expression included ADAMTS1 (A disintegrin and met-
alloproteinase with thrombospondin motifs 1), LAMA4 (Splice
isoform 1 of laminin alpha-4 chain precursor), FBLN2 (Fibulin-2
Table 1
Total spectral counts and proteins identified from the non-loaded tendons and the
loaded tendonsa.
Sample nameTotal spectral
counts
Unique peptidesTotal identified
proteins
564 (387bþ 177c)
Non-loaded tendon 1
Non-loaded tendon 2
Non-loaded tendon 3
Loaded tendon 1
Loaded tendon 2
Loaded tendon 3
9884
9627
9325
10,648
10,297
10,885
3456
3602 581 (387bþ 194c)
aThe criteria for filtering protein identifications with the false discovery rates
(FDRs) less than 1%: (1) the minimum DCn cut-off value was 0.08, the minimum
Xcorr for singly, doubly and triply charged peptides were 1.8, 2.5 and 3.5, respec-
tively; (2) proteins were identified in all triplicated experiments and their RSD
(relative standard deviation) for spectral count less than 50%; (3) at least two unique
peptides were identified per protein.
bThe number of identified proteins in both groups.
cThe number of identified proteins only in each group.
Fig. 2. Gross view (A, B) and histology (C, D, bar ¼ 100 mm) of the in vitro engineered tendons with dynamic loading for 10 weeks in the bioreactor after 2 weeks of static culture
(A, C), and engineered tendon without loading for 12 weeks (B, D).
Y. Jiang et al. / Biomaterials 32 (2011) 4085e4095
4088

Page 5

precursor), FBLN4 (EGF-containing fibulin-like extracellular matrix
protein 2 precursor) and TSP3 (Thrombospondin-3 precursor).
Those with no significant change in their expression included DCN
(Splice isoform a of decorin precursor), BGN (Biglycan precursor),
LUM (Lumican precursor), TNC (Splice isoform 1 of tenascin
precursor), FBN1 (Fibrillin-1 precursor), FBLN1 (Splice isoform D of
fibulin-1 precursor), TSP2 (Thrombospondin-2 precursor) and TSP4
(Thrombospondin-4 precursor).
Matrix metalloproteinase (MMP) and its regulator, metal-
loproteinase inhibitor (TIMP) are both important for matrix depo-
sition and degradation and tissue remodeling during tendon
formation. As shown in Fig. 4 and Supplementary Table 3-1, MMP-
14 expression was significantly increased, whereas MMP3 expres-
sion was significantly decreased. However, there was no significant
change for MMP-1, MMP-2 and TIMP-1 in protein expression. In
addition, TIMP-3 was up-regulated with 3 folds in spectral count
ratio although not significantly.
3.7. Differential expression of intra-cellular signaling molecules
Mechanical stimulation exerts its biological effect on cells via
various signaling pathways. This study also revealed differentially
expressed signaling molecules when comparing loaded with non-
loaded samples. As shown in Fig. 5 (A: up-regulated; B: down-
regulated; C: no significant change) and Supplementary Table 3-2,
G-protein signaling related molecules, such as GNA13 (Guanine
nucleotide-binding protein alpha-13 subunit), GNA14 (Guanine
nucleotide-binding protein alpha-14 subunit), PI4KA (Splice Iso-
form 1 of phosphatidylinositol 4-kinase alpha), PKC-delta (Protein
kinase C delta binding protein) were significantly up-regulated in
their expressions. However, there was no apparent change of the
expression of CaM (Calmodulin) and GNB1 (Guanine nucleotide-
binding protein G(I)/G(S)/G(T) beta subunit 1). Additionally, RhoA
and RhoC expressions were found significantly down-regulated.
Non-canonical Wnt signal pathway may also participate. As
shown in Fig. 5 and Supplementary Table 3-2, the expression of
WNT5A (Protein Wnt-5a precursor) was significantly up-regulated,
whereas the expressions of DKK2 (Dickkopf-related protein 2
precursor)andSFRP4(Secretedfrizzled-relatedprotein4precursor)
were significantly down-regulated. The expression of SFRP2
(Secreted frizzled-related protein 2 precursor) was not significantly
altered.
Ras family molecules have been reported to play their roles in
mechano-signaling, which agrees with the findings of this study
(Fig. 5 and Supplementary Table 3-2). RAB5B (Ras-related protein
Rab-5B) was significantly up-regulated, whereas the expression of
RAB1B (Ras-related protein Rab-1B) was significantly down-regu-
lated. However, there was no significant change of the expression
for RAB2A (Ras-related protein Rab-2A), IQGAP1 (Ras GTPase-
activating-like protein IQGAP1).
Fig. 3. The up-regulated proteins (loaded tendon/non-loaded tendon spectral count ratio ?5) and down-regulated proteins (loaded tendon/non-loaded tendon spectral count ratio
?0.2) were classified using Gene Ontology annotation into molecular function (A), biological process (B) and cellular locations (C).
Y. Jiang et al. / Biomaterials 32 (2011) 4085e4095
4089