1. Purpose: The purpose of the study was to evaluate the effects of the
rehabilitation program after simultaneously total knee replacement on
balance and functionality in patients with osteoarthritis and to compare
the results with the healthy individuals.
Methods: Two study groups of osteoarthritic patients between ages 50-
85 who underwent bilateral TKA and one control group (Group 3, n¼20)
were included in the study. Exercise program with frequent physical
therapist supervision (2 session/1 week) was performed in Group 1
(n¼20) and less supervised exercise program was performed Group 2
(n¼20) (1 session/2 week). Exercise program was continued during the
first month. All the patients were evaluated preoperatively as well as
1.and 2.month postoperatively. The patients were checked every week
by telephone interview. Functionality was evaluated with WOMAC (The
Western Ontario and McMaster Universities Arthritis) Index, dynamic
balance with Timed Up&Go (TUG) and stair climbing test (SCT), static
balance with single limb standing test (SLST).
Results: There was no statisitical significance between groups in terms
of age (p¼0,463). In preop evaluation there were significant differences
between groups in all tests in favor of the control group. WOMAC pain
scores in Groups 1 and 2 were similar in 1.month postop evaluation
while the decrease in the pain score in Group 1 was higher than the
scores of Group 2 and was close to the scores of healthy controls
(p¼0,00).
WOMAC stiffness scores were higher in postop period than preop
period in both Group 1 and 2. The stiffness scores of Group 1 were
significantly lower than Group 2 in 2. month postoperatively and were
close to healthy controls (p¼0,00).
WOMAC functionality scores of Group 1 and 2 were significantly higher
in 1.and 2.month postop compared to preop period. The functionality
scores of Group 1 were significantly higher in postop 2.month com-
pared to Group 2 and were closer to the scores of healthy controls
(p¼0,00).
The completion time of TUG test of the groups were shorter in 1.and
2.month scores and were similar to the 1.and 2.month scores of controls
(p¼0,33).
SCT time of the groups were shorter in both 1.and 2. month postop
period compared to preop period. Stair climbing time of Group 1 was
significantly shorter than Group 2 in 2.month postoperatively and was
closer to the scores of healthy controls (p¼0,00).
SLST for both legs of Group 1 were significantly higher than Group 2 in
1. and 2. month postoperatively and were closer to healthy controls
(p¼0,00).
Conclusions: The patients which have applied exercise program after
simultaneously bilateral TDR under the supervision of the physi-
otherapist 2 days a week for 1 month has obtained improvements in
balance and functional parameters and it has been shown that their
performances approach the healthy control group values.
662
INFLAMMATORY FACTORS DO NOT ENABLE BMP2-INDUCED
OSTEOPHYTE FORMATION
E.N. Blaney Davidson, M.I. Koenders, A.P. van Caam, E.L. Vitters,
M.B. Bennink, P.L. van Lent, F.A. van de Loo, P.M. van der Kraan. Radboud
Univ. Med. Ctr., Nijmegen, Netherlands
Purpose: The main hallmarks of osteoarthritis (OA) are cartilage
damage and osteophyte formation. The latter can clearly be observed in
X-rays of patients. What causes these osteophytes in OA is not entirely
clear. We have previously identified that in experimental models both
TGF-beta and BMP2 can induce de novo osteophyte formation. In
addition, when blocking TGF-beta signalling during experimental OA
we were able to prevent osteophyte formation. Recently we have shown
that chondrocyte-specific overexpression of BMP2 in a mouse model
does not induce osteophyte formation on its own, but that it greatly
enhances experimental OA-induced osteophyte formation. The latter
could be blocked by inhibition of TGF-beta activity. Apparently, TGF-
beta is involved in osteophyte formation but since there are inflam-
matory factors in (experimental) OA, it is still unclear whether they
might play a role in osteophyte formation. This led to our question: can
inflammatory factors enable BMP2 to induce osteophyte formation in a
mechanically stable joint?
Methods: We exposed Col2a1-rtTA-BMP2 mice (inducible BMP2 with
doxycycline) to a simple inflammatory trigger that has been implicated
in OA: interleukin-1, and to a more complex inflammatory trigger:
zymosan by intra-articular injection. Saline injections were used as
controls. In addition, since we have previously identified TGF-beta as a
major trigger for osteophyte formation in OA, we included a low-dose of
adenovirus overexpressing TGF-beta (ad-TGF-beta) as a trigger, which
on its own results in fibrosis but not yet osteophytes. We added ad-Luc
as a viral control. Mice were either fed standard diet (non-dox) or a diet
with doxycycline (dox) starting 7 days prior to i.a. injection for the
entire duration of the experiment and 14 days after intra-articular
injection of the adenovirus we isolated knee joints for histology.
Results: Saline injections combined with BMP2 did not lead to osteo-
phyte formation. Moreover, neither did IL-1 injection combined with
BMP2. Zymosan injection resulted in clear synovial inflammation, which
was further enhanced by BMP2 overexpression. (fig 1A) Also, there were
minor fissures in medial tibial cartilage in zymosan-injected joints, which
showed a trend to becoming slightly more severe when combined with
BMP2, but this was not significant. (fig 1B) Despite the overt inflamma-
tion induced by injection of zymosan, the combination of zymosan with
BMP2 overexpression was insufficient to induce osteophytes. (fig 1C)
Injection of adenoviral TGF-beta resulted in massive fibrosis and minor
osteophyte formation. As expected, combining TGF-beta with BMP2
resulted in clearly enhanced osteophyte formation. (fig 1D)
Conclusions: Together these data show that neither a simple inflam-
matory trigger like IL-1, nor a more complex inflammatory trigger like
zymosan enable BMP2 to induce osteophytes. In contrast, BMP-2 in
combination with TGF-beta is a results in massive osteophyte for-
mation. The latter confirmed our prior data where we showed that by
blocking TGF-beta during experimental OA we could prevent BMP2
from inducing osteophytes.
663
CARTILAGE GENE EXPRESSION CHANGES IN EARLY MOUSE DMM
ARE DETECTED IN LATE HUMAN OA
J. Soul y, S.L. Dunn y, L.H. Kung y,
z, S. Anand x, J.-M. Schwartz y,
T.E. Hardingham y, R.P. Boot-Handford y. y
The Univ. of Manchester,
Manchester, United Kingdom; z
Murdoch Childrens Res. Inst., Melbourne,
Australia; x
Stockport NHS Fndn. Trust, Stockport, United Kingdom
Purpose: The DMM mouse has become the model of choice for much
experimental OA research because of its reproducible and relatively
rapid disease onset, coupled to the mouse’s genetic tractability. Nev-
ertheless, the DMM model represents a mechanically induced form of
OA developing quickly (2e8 weeks) in a young (10e12 week) animal
whereas most human OA is in an ageing population with idiopathic and
long term disease. We therefore compared the differentially regulated
gene profiles accompanying damage in human OA cartilage with sam-
ples of cartilage progressing to damage taken from DMM mice.
Methods: Global gene expression was determined by RNAseq analysis
on i) paired intact and damaged OA cartilage from 8 patients at TKR; and
ii) mouse tibial medial compartment cartilage laser dissected from
DMM and sham operated mice 2 weeks after operation. The expression
Abstracts / Osteoarthritis and Cartilage 24 (2016) S63eS534S396
Downloaded from ClinicalKey.com.au at University of Melbourne May 05, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.

2. data was batch effect corrected with RUVSeq, normalised and the fold
changes between the paired human samples and the DMM and sham
mice calculated with DESeq2. Benjamini-Hochberg multiple testing
correction was performed. Additional expression data obtained from
mouse cartilage isolated as described above from 2 and 6 week DMM
and sham wild type mice (GSE45793) was downloaded from GEO and
analysed with limma to identify differentially expressed genes.
Genes with an absolute fold change of !1.5 and a Benjamini-Hochberg
corrected p-value of 0.1 were classified as differentially expressed. For
comparison between mouse and human gene expression data, gene IDs
were mapped to ortholog gene groups from InParanoid. Hyper-
geometric statistics were used to calculate the statistical significance of
the gene overlaps.
Results: The human and mouse gene expression datasets shared 12600
orthologue gene groups. Comparison of RNAseq data from 8 paired
samples of damaged and intact OA cartilage from patients undergoing
TKR revealed 1237 genes were significantly dysregulated in the damaged
cartilage tissue. Data from mouse DMM experiments (2 datasets obtained
at 2 weeks and 1 data set at 6 weeks post operation) were analysed
against their respective shams generating a combined list of 1756 genes
that were significantly dysregulated by DMM. Comparison of the genes
dysregulated in damaged human OA and mouse DMM cartilage revealed
a highly significant (p¼ 4.403e-05) overlap of 220 genes (Figure 1).
Reactome pathway assessment of these 220 genes revealed the most
significant terms to involve extracellular matrix (ECM) including ECM
degradation / organisation, collagen formation / degradation and
integrin cell surface interactions. These terms were populated pre-
dominantly with collagen and other genes encoding structural ECM
proteins such as FN1, ACAN and LUM. Other significant terms included
peptide ligand binding receptors / G Protein-Coupled Receptor ligand
binding (which shared the AGTR2, F2R, EDNRA, JAK3, C3, CCBP2,
ANXA1, CXCR7 and ATGR1 genes) and signal transduction.
Of the 220 dysregulated genes common to human and one, or more of
the mouse DMM datasets, 21 genes were co-ordinately dysregulated in
all of the datasets (Table 1). These core set of genes were predominantly
involved with signalling and included chordin 2 (BMP signalling), GDF6
(joint development), PTGS2 (prostaglandin synthesis), Retinoic Acid
Receptor B and a limited number of ECM associated genes included
Lumican and TGFb induced protein In addition, the dysregulation of
three of these genes (NPR3, MOXD1, PTGS2) may be linked to their
reported sensitivity to mechanical stimulation and the increased stress
on chondrocytes associated with, or causing the tissue damage.
Conclusions: Around 18% of the genes found to be dysregulated in the
damaged cartilage taken from patients at TKR are also dysregulated in
cartilage from DMM mice at either 2 and/or 6 weeks post-operation.
Many of these genes encode proteins that are part of, or associated with
the cartilage ECM. Of particular note is the core of 21 genes which are
involved in different aspects of signalling that are changed in expression
at the earliest stages of cartilage damage in the DMM mouse and also
changed in late stage human OA cartilage. These results provide sup-
porting evidence that the DMM model, although of joint destabilisation
in a young mouse, nevertheless reveals changes that suggest similar
processes are active in late human OA and these may represent
potential targets for future therapy. Supported by AR UK 19501 & 20414.
664
RAAV-MEDIATED COMBINED GENE TRANSFER AND
OVEREXPRESSION OF TGF-b AND SOX9 REMODELS HUMAN
OSTEOARTHRITIC ARTICULAR CARTILAGE
K. TAO, Sr. y,
z, A. Rey-Rico x, J. Frisch x, J. Venkatesan x, G. Schmitt x,
H. Madry x,
k, J. Lin ¶,
z, M. Cucchiarini x. y
Peking Univ. People's Hosp.,
Beijing, China; z
Peking Univ. Hlth. Sci. Ctr., Beijing, China; x
Ctr. of
Experimental Orthopaedics, Saarland Univ., Homburg, Saar, Germany;
k
Dept. of Orthopaedic Surgery, Saarland Univ., Homburg, Saar, Germany;
¶
Inst. of Arthritis, Peking Univ. People's Hosp., Beijing, China
Purpose: Direct administration of therapeutic candidate gene sequen-
ces using the safe and effective recombinant adeno-associated virus
(rAAV) vectors is a promising strategy to stimulate the biologic activities
of articular chondrocytes as an adapted tool to treat human osteo-
arthritic (OA) cartilage [1].
Methods: In the present study, we developed a combined gene transfer
approach based on the co-delivery of the pleiotropic transformation
growth factor beta (TGF-b) with the specific transcription factor SOX9
via rAAV to human normal and OA chondrocytes in vitro and cartilage
explants in situ in light of the mitogenic and pro-anabolic properties of
these factors [2, 3].
Results: Effective, durable co-overexpression of TGF-b and sox9 sig-
nificantly enhanced the levels of cell proliferation both in human nor-
mal and OA chondrocytes and cartilage explants over an extended
period of time (21 days) (Fig. 1A), while stimulating the biosynthesis of
key matrix components (proteoglycans, type-II collagen) compared
with control conditions (reporter lacZ gene transfer, absence of vector
treatment) (Fig. 1B, C, D). Of further note, expression of hypertrophic
type-X collagen significantly decreased following co-treatment by the
candidate vectors (Fig. 1B, E).
Conclusions: Here, we examined the feasibility of combining the
transfer of two crucial factors with chondroreparative activities (TGF-
b and SOX9) via direct, dual rAAV gene transfer as a means to
enhance the remodelling of human OA chondrocytes in vitro and OA
cartilage in situ. Our results first show that concomitant expression of
TGF-b and sox9 was successfully achieved via rAAV both in human
normal and OA chondrocytes in vitro and in human normal and OA
cartilage explants in situ for at least 21 days, probably due to a good
penetration and maintenance of the vectors in the targets as noted
when using individual gene treatments. The data further indicate
that high, durable levels of rAAV-co-delivered TGF-b/sox9 were
capable of activating the biological (proliferation, ECM synthesis)Figure 1. Genes dysregulated in damaged human OA and mouse DMM
cartilage and the overlap between the two populations.
Table 1. Genes dysregulated co-ordinately in damaged human OA and in all
mouse DMM cartilage datasets
Function Gene Human1
2 Wk DMM1
2 Wk DMM2
6 Wk DMM2
(log2FoldChange)
Extracellular COL8A1 1.18 1.28 2.64 2.80
Matrix LEPREL1 1.26 0.80 1.34 1.14
LUM 1.14 1.10 2.86 3.71
TGFBI 1.22 0.96 2.00 3.80
Transcription FHL2 1.14 1.95 2.35 1.61
Regulator ID2 0.74 1.19 2.37 1.65
SOX8 À0.58 À1.45 À2.41 À1.85
Signalling ARHGAP24 0.88 1.28 2.53 3.34
CHRDL2 À1.30 À1.72 À2.68 À2.75
EMP1 0.64 1.23 2.34 1.72
GDF6 0.82 2.58 2.74 2.22
MOXD1 1.06 1.55 1.62 3.08
NBL1 0.71 1.97 1.32 2.13
NFKBIZ 0.67 1.04 1.36 1.00
NPR3 1.49 1.97 3.20 2.24
NRP2 0.60 1.16 1.12 2.29
PTGS2 1.07 0.76 5.10 3.85
RARB 0.83 1.16 2.50 2.13
RNF152 0.78 0.82 1.60 1.83
TM4SF1 0.74 1.28 1.49 1.53
TSPAN2 0.79 0.79 1.53 1.13
1
¼ RNAseq data generated in-house (unpublished);
2
¼ microarray data from GSE45793
Abstracts / Osteoarthritis and Cartilage 24 (2016) S63eS534 S397
Downloaded from ClinicalKey.com.au at University of Melbourne May 05, 2016.
For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.