Downloaded from ard.bmj.com on September 21, 2010 - Published by group.bmj.
com
Extended report
▶ Additional data are published
online only. To view these files
please visit the journal online
(http://ard.bmj.com).
1Rheumatology Service,
Hospital Clínico San Carlos,
Madrid, Spain
2Immunology Service, Hospital
Clínico San Carlos, Madrid,
Spain
3Neurology Service, Hospital
Clínico San Carlos, Madrid,
Spain
4Department of Cardiovascular
Epidemiology and Population
Genetics, Fundación Centro
Nacional de Investigaciones
Cardiovasculares Carlos III
(CNIC), Madrid, Spain
5Traumatology Service, Hospital
Clínico San Carlos, Madrid,
Spain
6Department of Proteomics,
Fundación Centro Nacional
de Investigaciones
Cardiovasculares Carlos III
(CNIC), Madrid, Spain
7Genomics Unit,
Fundación Centro Nacional
de Investigaciones
Cardiovasculares Carlos III
(CNIC), Madrid, Spain
Correspondence to
José Ramón Lamas,
Rheumatology Service, Hospital
Clínico San Carlos, C/ Profesor
Martín Lagos s/n, Madrid
28040, Spain;
jrlamas@gmail.com
Accepted 20 March 2010
1880
Large-scale gene expression in bone marrow
mesenchymal stem cells: a putative role for
COL10A1 in osteoarthritis
José Ramón Lamas,1 Luis Rodríguez-Rodríguez,1 Ana G Vigo,2 Roberto
Álvarez-Lafuente,3 Pedro López-Romero,4 Fernando Marco,5 Emilio Camafeita,6 Ana
Dopazo,7 Sergio Callejas,7 Esther Villafuertes,1 José Antonio Hoyas,1 María Pilar
Tornero-Esteban,1 Elena Urcelay,2 Benjamín Fernández-Gutiérrez1
ABSTRACT
Objectives To elucidate disease-specific molecular
changes occurring in osteoarthritis (OA) by analysing
the differential gene expression profiles of bone marrow
mesenchymal stem cells (BM-MSCs) from patients with
OA compared with those without OA.
Methods Expression profiles of BM-MSCs from eight
paired patients with OA and patients with hip fracture
without signs of OA were compared by DNA microarray
expression analysis and significant differences were
evaluated by computational Gene Set Enrichment
Analysis. To validate the involvement of COL10A1 as part
of the most downregulated gene set in OA, three tagging
single nucleotide polymorphisms were genotyped in 191
patients with OA and 283 controls. COL10A1 expression
was also assessed by quantitative RT-PCR in additional
subjects.
Results Expression levels in 9% (1967) of the overall
transcripts were significantly different (p<0.05) between
MSCs from patients with OA and controls (532 genes
reached twofold differences: 240 were upregulated
and 292 were downregulated). Cell development and
differentiation were the functional categories accounting
for most genes with altered expression. Interestingly,
several genes related to the Wnt/-catenin pathway
and collagen genes were downregulated in MSCs from
patients with OA. The collagen gene set was clearly
downregulated in OA. Furthermore, the expression of
COL10A1 was significantly reduced in patients with OA.
A genetic association between the COL10A1 rs11965969
polymorphism and OA was also found.
Conclusion COL10A1 downregulation seems to
have a role in the establishment of a defective and/or
unstable subchondral cartilage matrix in OA disease.
It is proposed that OA may be linked to the intrinsic
defective regenerative potential of BM-MSCs resulting
from its reduced expression of fate commitment-related
genes.
INTRODUCTION
Osteoarthritis (OA) is one of the most common
skeletal disorders clinically manifested by joint
pain, swelling and progressive loss of function. It is
characterised by cartilage degradation, hypertrophy
of the subchondral bone and osteophyte formation
at the joint margins. The severity and progression
of OA varies among patients, and fast progression
leads to disability and ultimately joint replacement
surgery.1
The aetiology and progression of OA has been
considered an articular cartilage disorder induced
by mechanical stress, articular injuries and involve-
ment of inflammatory mediators.2 Although the
underlying molecular mechanisms involved in the
disease remain unknown, disease establishment
and progression are likely to be independent pro-
cesses associated with different risk factors.3 4
Several studies suggest that cartilage and the
subchondral bone are likely to be the most impor-
tant structural elements in pain generation and dis-
ease progression. Their biology and the pathways
involved in cartilage and bone formation and turn-
over are therefore interesting targets for therapeu-
tic intervention. OA shows features of a systemic
musculoskeletal disease with a metabolic compo-
nent and a genetic predisposition leading to the for-
mation of a defective extracellular matrix (ECM).2 5
The collagen-rich ECM secreted by chondrocytes
provides the fundamental structure of cartilage and
plays an essential role in bone biology during both
embryo development and during endochondral
ossification and bone remodelling.
The cellular components of cartilage are chon-
drocytes, belonging to a mesenchymal stem cell
(MSC) chondrogenic progeny. In the growth plate
they enter in a programmed cell proliferation and
mature or differentiate to hypertrophic chondro-
cytes synthesising high levels of type X collagen.
Hypertrophic chondrocytes are subsequently
responsible for the deposition of mineralised ECM,
which serves as a template for bone formation.
This developmental process is under a fine-tuned
balance whose deregulation may be responsible for
defective matrix formation and cartilage degenera-
tion, as occurs in many skeletal dysplasias.6
MSCs show two well-known characteristics
of stemness: the self-renewal capacity and the
potential to differentiate towards several connec-
tive tissue cells.7 8 They are present in a variety of
tissues, but are primarily found in the bone mar-
row (BM).9 MSCs, as chondroprogenitor cells with
regenerative potential, have generated a great deal
of interest as a source for cell-based treatment in
rheumatic diseases.10 Their regenerative potential
has been shown in several tissues, including animal
models of OA,11 infracted myocardium,12 neural
tissue13 and in bone tissue regeneration.14 15
In this study we aimed to define a comparative
transcriptional profile to describe the gene signature
Ann Rheum Dis 2010;69:1880–1885. doi:10.1136/ard.2009.122564
 Downloaded from ard.bmj.com on September 21, 2010 - Published by group.bmj.com

Extended report

of MSCs from OA disease. Elucidation of the phenotypical alter- were digested with RNase-free DNase (Sigma-Aldrich; St Louis,
ations occurring in OA is important for the ascertainment of dis- Missouri, USA) at 70°C/15 min.
ease aetiology and for the development of effective treatments Transcriptor first-strand cDNA synthesis kit (Roche
for OA. For this purpose, DNA microarray analysis combined Diagnostics, S L Barcelona, Spain) was used to perform RT.
with bioinformatics and genetic validation were used as proven A 25°C/10 min incubation was followed by 50°C/60 min and
state of the art technologies16 for the study of gene expression. 85°C/5 min to inactivate the reaction and the cDNA was stored
at –80°C until used.
METHODS
Quantitative real-time PCR for cDNA of COL10A1
Patients and specimens Quantitative real-time PCR of COL10A1 was performed with
BM aspirates were obtained from the femoral channel of eight
a TaqMan Gene Expression Assay ID Hs00166657 (Applied
patients with OA (mean age 74.7 years, range 61–89) at the time
Biosystems). COL10A1 transcriptional expression results were
of surgery for total knee arthroplasty and from eight patients
normalised using the -actin (ACTB) gene (Eurogentec, Liege,
with hip fracture without signs of OA (mean age 66.6 years,
Belgium) as previously described.21 For the final relative quan-
range 44–90) during hip joint replacement surgery for the treat-
tification of expression, each RNA sample was assigned a
ment of subcapital fracture. The diagnosis of OA was based on
ΔCT value, CT (for COL10A1) minus CT (ACTB). The results
American College of Rheumatology criteria.17 18 The patients
were then normalised according to standard samples of RNA,
with subcapital fracture of the hip had no radiographic changes
extracted from the same number of healthy individuals run in
of OA, no diagnosis of osteoporosis in their clinical records,
parallel. These controls were assigned the normalisation ratio
they had not received treatment related to osteoporosis and had
(NR) of 1, and the NR of each experimental sample was calcu-
a densitometric T-score <2.5SD (Hologic QDR-4500C, Bedford,
Massachusetts, USA) performed after surgery. lated from the formula: NRexp=2CT, where –ΔΔCT was
ΔCT experimental minus ΔCT standard. ‘No RT’ controls were
used to ensure the absence of genomic DNA traces.
Cell culture
The isolation and in vitro expansion of MSCs were carried out
as previously described.19 At the end of primary culture, for a 10
serial passage, cells at approximately 50–60% confluence were n=240
detached with 0.25% trypsin-EDTA for 5 min at 37°C and
replated for continued passaging at 1:2. Confluent cells at the
third passage were used for the experiments.

Array data analysis

The chips were scanned on a Microarray Scanner (G2565BA;
Fold variation in gene expression

5
Agilent, Santa Clara, California, USA). Image analysis and data col-
lection were carried out using the Agilent Feature Extraction image
analysis software Version 9.1.3.1 (AFE). The data were analysed in R
(R Development Core Team) using packages of the Bioconductor
project20 as well as custom written R routines (see Methods sec- 2
tion in online supplement).

DNA isolation 0
DNA was extracted from blood of patients and controls using n=292
Qiagen columns (Qiagen, Hilden, Germany) according to the
manufacturer’s protocol. 2

Genotyping
COL10A1 is located in the chromosome region 6q22.1 and
extends ~7 kb. Three single nucleotide polymorphisms (SNPs) 5
were selected to cover most of the variability in a 17 kb region
surrounding this gene. The tagging was performed using the
‘aggressive tagging’ option in the Haploview software with r2
threshold set at 0.8 and minimum minor allele frequency at 0.1.
The SNPs (rs4945551, rs11965969, rs1931897) were genotyped
using assay from Applied Biosystems, Foster City, California,
USA and analysed in an ABI 7900HT system. All the polymor-
phisms conformed to Hardy–Weinberg proportions in the con- 10
trol sample.
Figure 1 Differential gene expression in osteoarthritis mesenchymal
stem cells (OA-MSCs). Fold change variation of gene expression.
RNA extraction and reverse transcription (RT) The y-axis represents the distribution of gene expression levels from
Total RNA was extracted from BM-MSCs using the QIAamp statistically significant upregulated to downregulated expression levels
RNA Kit (Qiagen). Total RNA was eluted in 60 μl RNase-free (p<0.05) in OA-MSCs. For graphical purposes, data points outside the
water. Blank reactions were interspersed within experimental axis limits are not represented. A complete list is available in table 1 in
samples. Purified RNA was quantified. Prior to RT, samples the online supplement.

Ann Rheum Dis 2010;69:1880–1885. doi:10.1136/ard.2009.122564 1881

 Downloaded from ard.bmj.com on September 21, 2010 - Published by group.bmj.com

Extended report

RESULTS A broad overview of results showed that signal transduction

Gene expression profile comparison for BM-MSCs
After the microarray data preprocessing steps were completed
(see Methods section in online supplement), 21 255 transcripts
were selected for statistical analysis. We found 1967 differ-
entially expressed transcripts between OA-MSCs and control
MSCs at the p<0.05 level of significance. A more than twofold
difference in expression was seen in 532 (~25%); 240 (~45%)
of these were upregulated and 292 (~55%) were downregu-
lated in OA-MSCs compared with control cells. Moreover,
40 downregulated and 21 upregulated transcripts showed an
expression level altered by at least fivefold (see file 1 in online
supplement and figure 1). The gene expression cut-off value
was fixed for those transcripts showing statistically signifi-
cant differences by more than twofold. With these data filtra-
tion criteria we attempted to minimise the number of genes
without missing relevant biological information for analy-
sis. The signature of altered genes fell mainly into the signal
transduction and development categories, including a series
of downregulated homeobox proteins (Hox-B3, –C12, –DLX2,
–B4, –C13, –SIX6, –Short stature) (see file 1 in online supple-
ment and figure 2B). Interestingly, other genes related to the
Wnt/-catenin pathway also showed downregulated expres-
sion in OA-MSCs. These genes included SFRP2, RUNX3,
BARX1, RSPO2, RSPO3, SPP1 and DACT1 (see file 1 in online
supplement).
Gene ontology analysis
To integrate expression profiles into functional categories, we
performed a gene ontology (GO)-based statistical analysis using
GeneCodis 2.0.22 23 Seventy-five genes from our list of 532 pro-
vided for comparison did not show annotations. The remain-
ing 457 genes were grouped into different GO categories based
on the subcellular location and functionality (see file 2 in online
supplement).
(45 genes), development (44 genes), differentiation (22 genes),
cell–cell signalling (15 genes) and cell proliferation (14 genes)
were among the five most enriched biological process catego-
ries. In addition, cytoplasm, extracellular region and integral to
membrane classes were among the first three significant GO cell
component categories. A total of 23 out of 45 genes pertaining
to the signal transduction category and 29 from the 44 related
to the development category were clearly downregulated.
Figure 2 shows the number of genes upregulated and down-
regulated in each annotation (figure 2A) as well as the variation
in expression (figure 2B). These results strongly suggest that the
altered genes—mainly signal transduction factors and molecules
involved in development—are to a large extent downregulated
in OA-MSCs. This finding most probably indicates the existence
of impaired mechanisms to restore damaged tissues in OA.
A significant part of the transcripts encoded proteins involved
in the protein binding molecular function category. When we
visually analysed the list of ‘specific’ genes whose expression
was significantly altered, we found that collagens belonging to
the extracellular region GO cell component category were clearly
downregulated in OA-MSCs for COL4A1, COL4A2, COL8A1I,
COL11A1 and, in particular, COL10A1.
Gene set enrichment analysis
Following gene set enrichment analysis (GSEA; see online
supplement), four gene sets from the BioCarta gene set col-
lection were significantly enriched in the control group at a
false discovery rate <25% (p<0.25). These were the CSK path-
way gene set, p=0.18 (composed of genes involved in modu-
lation of T cell receptor activation by c-src tyrosine kinase);
the AMI pathway gene set, p=0.21 (genes involved in acute
myocardial infarction); the NKT pathway gene set, p=0.05
(genes related to NKT cells during T cell polarisation) and the
collagen gene set, p=0.01. Bearing in mind that the collagen

(A) (B)
Organismal development 29/15

Signal transduction 23/22

Cell differentiation 15/7

Cell–cell signalling 7/8

Cell proliferation 9/5

Metabolic process 10/3

Carbohydrate Metabolic process 6/3

Anatomical structure morphogen 3/3

Cytoskeleton organisation 4/1

Response to stress 2/2

Response to external stimulus 1/1

Cellular component organisation 2/0

Growth 1/1

−10 −5 0 5 10 0 10 20 30 40
Fold decrease Fold increase Number of genes

Figure 2 Differential expression of genes in osteoarthritis mesenchymal stem cells (OA-MSCs) according to gene ontology (GO) categories. (A) The
y-axis represents individual genes classified according to the GO slim categories provided by the GeneCodis2 analysis. The x-axis represents the fold
variation in expression of OA-MSCs compared with control subjects (p<0.05). Only genes expressing at least twofold differences in expression were
considered. Data points outside the axis limits are not represented for graphical purposes. A complete list is given in file 2 in the online supplement.
(B) Bar graph showing the number of genes upregulated (white bars) and downregulated (black bars) in each category represented in (A).

1882 Ann Rheum Dis 2010;69:1880–1885. doi:10.1136/ard.2009.122564

 Downloaded from ard.bmj.com on September 21, 2010 - Published by group.bmj.com

Extended report

gene set was the most enriched in controls and therefore the
least expressed in OA-MSCs, we next focused our attention
on these genes. Another GSEA using all human collagen genes
was performed. Only collagen genes detected in at least one
experimental condition were used. Twenty-five out of 43 genes
showed annotations above our background level. The results
showed that 12 of the 25 collagen genes analysed, includ-
ing COL10A1, accounted for the enrichment of the collagen
gene set in control BM-MSCs and consequently they were
downregulated in OA-MSCs (figure 3). These results indicate
a consistently diminished transcriptional expression of colla-
gen genes in OA-MSCs; this decrease also affects COL10A1,
the only known hypertrophic chondrocyte-specific molecular
marker.24
Validation of results by quantitative PCR
The decreased expression previously observed by microar-
ray analysis was confirmed by the measurement of COL10A1
mRNA levels using quantitative RT-PCR. Thirteen of the
patients with OA analysed (seven different from those pre-
viously used) had an NR <1 compared with seven controls
(four additional samples), thus indicating reduced expression
(figure 4). However, the degree of decreased expression dif-
fered, probably as a result of the difference in the sensitivity of
the two assays. Nevertheless, the results validate the use and
reproducibility of our microarray protocol as an approach to
extrapolate biological significance.
OA-12
Genetic association of polymorphism in COL10A1 with OA
To further understand the role of COL10A1 in the pathogen-
esis of OA, three SNPs (rs4945551, rs1931897 and rs11965969)
were studied in a cohort including 191 patients with OA and
283 matched controls.
The SNPs rs4945551 and rs1931897 did not show statis-
tically significant differences in their genetic distribution
between patients with OA and controls. However, rs11965969
showed significant differences when comparing TT with
TG+GG between patients with OA and overall controls (OR
1.68; 95% CI 1.09 to 2.59; p=0.013, table 1). As the likelihood
for a diagnosis of OA increases with age, the same statisti-
cal analysis was performed using as the control group only
individuals aged >45 and >55 years. Whereas significant dif-
ferences were maintained, the ORs were increased (TT vs
TG+GG between patients with OA and controls aged >45
years (OR 1.70, 95% CI 0.98 to 2.95, p=0.043) and between
patients with OA and controls aged >55 years (OR=2.15, 95%
CI 1.09 to 4.27, p=0.0017)). These results provide evidence
for a genetic association between the COL10A1 rs11965969
polymorphism and OA.
DISCUSSION
BM-MSCs are in vivo chondroprogenitor and osteoprogenitor cells
with a critical role in the biology of articular cartilage and bone.25
26 We hypothesised that early alterations in BM-MSCs might
have a role in the pathogenesis of OA. Genetic and/or function-
ally defective MSCs are likely to give rise to altered chondrocytes
and, thus, to affect the correct formation of the main functional
element of joint cartilage, the ECM. In order to identify potential

OA-20
OA-18
OA-14

SF-11
SF-12
OA-1
OA-8
OA-7

OA-3

SF-5
SF-8

SF-7
SF-2
SF-3
SF-4

COL11A2
COL8A2 1.0
COL16A1
NO
Enrichment, NO

COL6A3
Core enrichment,

COL14A1
Normalisation ratio (NR)

COL9A2
COL18A1
COL12A1
COL9A1
COL6A2
0.5
COL1A2
COL6A1
COL4A5
COL3A1
COL13A1
YES
Enrichment, YES

COL1A1
COL5A1
Core enrichment,

COL4A2
COL7A1 0.0
COL5A2
OA2
OA3
OA7
OA9
OA10
OA12
OA13
OA14
OA15
OA17
OA18
OA20
OA21
Median (IQR)

COL8A1
COL15A1
COL4A1
COL5A3 OA patients
COL10A1

Figure 3 Heat map showing clustered genes in the leading edge
subsets. The expression values are represented as colours, where the
range of colours (red, pink, light blue and dark blue) shows the range of
expression values (high, moderate, low and lowest). Genes with a ‘Yes’
value are those that contribute to the leading edge subset within the
gene set, the subset of genes that contributes most to the enrichment
result.
Figure 4 Validation by quantitative PCR of collagen gene expression.
COL10A1 mRNA in patients with osteoarthritis (OA). The results are
expressed as a normalisation ratio (NR). Healthy controls were assigned
an NR of 1 and all patients with OA had NRs <1. The mRNA expression
level of COL10A1 was lower in all patients with OA analysed (median
0.10 (IQR 0.03–0.16)). Patients with OA therefore expressed 10 times
less COL10A1 mRNA than healthy individuals.

Ann Rheum Dis 2010;69:1880–1885. doi:10.1136/ard.2009.122564 1883

 Downloaded from ard.bmj.com on September 21, 2010 - Published by group.bmj.com
Extended report
Table 1 Genotypic and allelic frequencies of COL10A1 rs11965969 polymorphism in patients with
osteoarthritis and controls
Osteoarthritis
rs11965969 n=191 % n=283
Genotype TT 62 32
TG 74 39
GG 55 29
Allele T 198 52
G 184 48
*TT vs (TG+GG): OR 1.68 (95% CI 1.09 to 2.59); p=0.013.
†TT vs (TG+GG): OR 1.70 (95% CI 0.98 to 2.95); p=0.043.
‡TT vs (TG+GG): OR 2.15 (95% CI 1.09 to 4.27); p=0.00017.
transcriptional alterations present in BM-MSCs, we studied dif-
ferences in cells from patients with OA and non-OA individuals.
A possible genetic component was also studied.
Functional categories showing a major number of genes with
downregulated expression in OA-MSCs were signal transduction,
development and cell differentiation, which in turn are key func-
tions in pluripotential cells. This probably indicates that MSCs from
patients with OA have a more limited potential to generate a func-
tional lineage of cells involved in musculoskeletal tissue homeosta-
sis. Besides, other genes coding for structural and regulatory proteins
were also underexpressed in OA-MSCs. This is the case for SPP1 (or
osteopontin), a secreted glycoprotein that binds to components of
the structural ECM and is involved in the regulation of bone min-
eralisation. It is expressed early in mesenchymal cell differentiation
and has been implicated in cell migration as well as osteogenesis.27
Collagen and proteoglycans are the two main components of
the ECM. Collagens II, IX and XI are maximally synthesised by
resting, proliferating and maturing chondrocytes, whereas type
X collagen is synthesised by hypertrophic chondrocytes before
mineralisation of the ECM occurs. The importance of collagens
in matrix formation has been reported in many skeletal dyspla-
sias.6 In humans, type X collagen deficiency occurs in Schmid
metaphyseal chondrodysplasia. A diminished and/or defective
collagen X production, such as that observed in our study, could
be hypothetically responsible (at least in part) for an altered and/
or unstable subchondral bone formation. Subchondral bone pro-
vides support for the articular cartilage and a damaged ECM,
and therefore can also affect joint conformation, inducing defor-
mation of articular cartilage and bone surface and creating areas
under stress in the surrounding cartilage.
The Wnt pathway includes signalling factors that regulate
developmental processes critical for growth plate organisa-
tion, cartilage boundary definition and endochondral ossifica-
tion.28 In addition, Wnt proteins have been reported to inhibit
chondrogenesis,29 adipogenesis30 and to have a dual effect on
osteogenesis.31 Wnt/-catenin signalling regulates chondrocyte
phenotype, maturation and function in a developmentally regu-
lated manner via binding to Frizzled and LDL receptor-related
protein receptors.31 This pathway therefore has a role in articu-
lar chondrocyte function and OA pathogenesis.32 The Wnt path-
way is negatively regulated by several modulators, including
secreted Frizzled-related proteins (SFRPs).33 In our study, SFRP2
was downregulated and thus probably prevents OA-MSCs from
entering the terminal differentiation process to take on other
musculoskeletal lineages.
RUNX3 forms a ternary complex with -catenin, attenuat-
ing Wnt signalling activity. The RUNX family is comprised
of three members expressed during chondrogenesis34 that are
1884
Controls
Overall >45 years >55 years
% n=127 % n=82 %
63 22 28 22 15 18
155 55 71 56 45 55
65 23* 28 22† 22 27‡
281 50 127 50 75 46
285 50 127 50 89 54
components of signalling cascades mediated by both trans-
forming growth factor- and bone morphogenetic proteins.35
RUNX1 is involved in haematopoiesis, RUNX2 has a number of
roles in osteogenesis and RUNX3 functions in the development.
In addition, in mice RUNX3 also mediates chondrocyte matu-
ration.36 Osterix, a transcription factor required for osteogen-
esis, is downregulated by RUNX3. Overexpression of RUNX3
also leads to enhanced expression of integrins and thus could
potentially contribute to the enhancement of migration.37 The
decreased expression of RUNX3 observed in our study therefore
probably contributes to defective chondrocyte differentiation.
BARX homeobox 1 has been described as an attenuator of
the Wnt receptor signalling pathway essential for cellular prolif-
eration, aggregation and the condensation of prechondrocytes.38
Members of the R-Spondin (RSPO) family of secreted proteins
were reported to be involved in vertebrate development, where
they act as ligands for Frizzled8 and LRP6 (LDL-receptor-related
protein), promoting the activation of the Wnt signalling.39 40 In
agreement with these functions, it is not surprising that OA-MSCs
showed a downregulated expression of both RSPO2 and RSPO3
in our study. DACT1 belongs to the Frodo/dapper protein family
which contains signalling adaptors that interact and regulate the
Wnt and other signalling pathways during development and can-
cer.41 In our study, based on gene array data alone, it is difficult
to evaluate whether the changes observed are present initially or
are a consequence of the disease process itself. Further validation
of these results is therefore needed.
Whereas human genetic association studies suggest that
genetic factors contribute to the pathogenesis of OA,42–46 the
direct genetic evidence and molecular mechanism for the patho-
genesis of OA are still unknown. In this study we describe for
the first time an association between the COL10A1 gene and
OA susceptibility. The associated polymorphism (rs11965969) is
most likely a genetic marker in linkage disequilibrium with some
other variant that is responsible for the observed association.
Studies using fine mapping throughout the gene are warranted
to determine the presence of the causative(s) polymorphism(s).
Mutations in the COL10A1 gene have been reported in Schmid
metaphyseal chondrodysplasia,6 47 and it would be possible
for the presence of a common polymorphism within this gene
to cause less severe but more prevalent diseases such as OA.
Interestingly, several mutations in the GDF5 gene, one of the
most robust genes associated with OA, have been described
to lead to a number of skeletal malformation syndromes
(OMIM*601146), indicating that bone formation and OA are in
some way connected.
Many factors such as mechanical stress, ageing, genetic back-
ground, inflammation and phenotypic changes influence the
Ann Rheum Dis 2010;69:1880–1885. doi:10.1136/ard.2009.122564
 Downloaded from ard.bmj.com on September 21, 2010 - Published by group.bmj.com
pathogenesis of OA and suggest the involvement of many regu-
latory mechanisms in gene expression. Overall, our findings pro-
vide a reference dataset of genes related to essential functions for
the normal biology of MSCs that become altered in OA. Based
on the function of the proteins encoded by these genes, our
results suggest both a diminished differentiation and regenera-
tive potential in patients with OA. Our results therefore suggest
that the underlying biological changes which occur during OA
disease might be related, at least in part, to defects in the ECM
and the formation of subchondral bone, an essential structure
for providing joint stability. The involvement of COL10A1 in the
development of OA evidenced by the gene expression analysis
has been validated through an association study of tagging SNPs
for the chromosomal region.
Acknowledgements The authors thank the orthopaedic surgeons at the Hospital
Clínico San Carlos for providing human samples.
Funding This work was supported by grants from Fundacion Mutua Madrileña and
FIS 04/1698.
Competing interests None.
Ethics approval This study was conducted with the approval of the ethics
committee of the Hospital Clínico San Carlos and all samples were obtained after
patients’ written informed consent.
Provenance and peer review Not commissioned; externally peer reviewed.
REFERENCES
1. Driban JB, Sitler MR, Barbe MF, et al. Is osteoarthritis a heterogeneous disease that
can be stratified into subsets? Clin Rheumatol 2010;29:123–31.
2. Brandt KD, Radin EL, Dieppe PA, et al. Yet more evidence that osteoarthritis is not a
cartilage disease. Ann Rheum Dis 2006;65:1261–4.
3. Dieppe P. Subchondral bone should be the main target for the treatment of pain and
disease progression in osteoarthritis. Osteoarthr Cartil 1999;7:325–6.
4. Valdes AM, Spector TD. The genetic epidemiology of osteoarthritis. Curr Opin
Rheumatol 2010;22:139–43.
5. Aspden RM. Osteoarthritis: a problem of growth not decay? Rheumatology (Oxford)
2008;47:1452–60.
6. Newman B, Wallis GA. Skeletal dysplasias caused by a disruption of skeletal
patterning and endochondral ossification. Clin Genet 2003;63:241–51.
7. Bruder SP, Jaiswal N, Haynesworth SE. Growth kinetics, self-renewal, and the
osteogenic potential of purified human mesenchymal stem cells during extensive
subcultivation and following cryopreservation. J Cell Biochem 1997;64:278–94.
8. Pittenger MF, Mackay AM, Beck SC, et al. Multilineage potential of adult human
mesenchymal stem cells. Science 1999;284:143–7.
9. Caplan AI. Review: mesenchymal stem cells: cell-based reconstructive therapy in
orthopedics. Tissue Eng 2005;11:1198–211.
10. Djouad F, Bouffi C, Ghannam S, et al. Mesenchymal stem cells: innovative
therapeutic tools for rheumatic diseases. Nat Rev Rheumatol 2009;5:392–9.
11. Murphy JM, Fink DJ, Hunziker EB, et al. Stem cell therapy in a caprine model of
osteoarthritis. Arthritis Rheum 2003;48:3464–74.
12. Barbash IM, Chouraqui P, Baron J, et al. Systemic delivery of bone marrow-derived
mesenchymal stem cells to the infarcted myocardium: feasibility, cell migration, and
body distribution. Circulation 2003;108:863–8.
13. Hofstetter CP, Schwarz EJ, Hess D, et al. Marrow stromal cells form guiding
strands in the injured spinal cord and promote recovery. Proc Natl Acad Sci USA
2002;99:2199–204.
14. Liu G, Zhao L, Zhang W, et al. Repair of goat tibial defects with bone marrow stromal
cells and beta-tricalcium phosphate. J Mater Sci Mater Med 2008;19:2367–76.
15. Petite H, Viateau V, Bensaïd W, et al. Tissue-engineered bone regeneration. Nat
Biotechnol 2000;18:959–63.
16. Licatalosi DD, Darnell RB. RNA processing and its regulation: global insights into
biological networks. Nat Rev Genet 2010;11:75–87.
17. Altman R, Asch E, Bloch D, et al. Development of criteria for the classification and
reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and
Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis
Rheum 1986;29:1039–49.
18. Arnett FC, Edworthy SM, Bloch DA, et al. The American Rheumatism Association
1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum
1988;31:315–24.
Ann Rheum Dis 2010;69:1880–1885. doi:10.1136/ard.2009.122564
Extended report
19. Rollín R, Alvarez-Lafuente R, Marco F, et al. Human parvovirus B19, varicella
zoster virus, and human herpesvirus-6 in mesenchymal stem cells of patients
with osteoarthritis: analysis with quantitative real-time polymerase chain reaction.
Osteoarthr Cartil 2007;15:475–8.
20. Gentleman RC, Carey VJ, Bates DM, et al. Bioconductor: open software
development for computational biology and bioinformatics. Genome Biol 2004;5:R80.
21. Pachner A, Narayan K, Price N, et al. MxA gene expression analysis as an interferon-
beta bioactivity measurement in patients with multiple sclerosis and the identification
of antibody-mediated decreased bioactivity. Mol Diagn 2003;7:17–25.
22. Carmona-Saez P, Chagoyen M, Tirado F, et al. GENECODIS: a web-based tool for
finding significant concurrent annotations in gene lists. Genome Biol 2007;8:R3.
23. Nogales-Cadenas R, Carmona-Saez P, Vazquez M, et al. GeneCodis: interpreting
gene lists through enrichment analysis and integration of diverse biological
information. Nucl Acids Res 2009;37:W317–22.
24. Zheng Q, Zhou G, Morello R, et al. Type X collagen gene regulation by Runx2
contributes directly to its hypertrophic chondrocyte-specific expression in vivo. J Cell
Biol 2003;162:833–42.
25. Aigner T, Söder S, Gebhard PM, et al. Mechanisms of disease: role of chondrocytes
in the pathogenesis of osteoarthritis–structure, chaos and senescence. Nat Clin Pract
Rheumatol 2007;3:391–9.
26. Tchetina EV, Squires G, Poole AR. Increased type II collagen degradation and very early
focal cartilage degeneration is associated with upregulation of chondrocyte differentiation
related genes in early human articular cartilage lesions. J Rheumatol 2005;32:876–86.
27. Giachelli CM, Steitz S. Osteopontin: a versatile regulator of inflammation and
biomineralization. Matrix Biol 2000;19:615–22.
28. Tamamura Y, Otani T, Kanatani N, et al. Developmental regulation of Wnt/
beta-catenin signals is required for growth plate assembly, cartilage integrity, and
endochondral ossification. J Biol Chem 2005;280:19185–95.
29. Day TF, Guo X, Garrett-Beal L, et al. Wnt/beta-catenin signaling in mesenchymal
progenitors controls osteoblast and chondrocyte differentiation during vertebrate
skeletogenesis. Dev Cell 2005;8:739–50.
30. Christodoulides C, Lagathu C, Sethi JK, et al. Adipogenesis and WNT signalling.
Trends Endocrinol Metab 2009;20:16–24.
31. Baksh D, Boland GM, Tuan RS. Cross-talk between Wnt signaling pathways in
human mesenchymal stem cells leads to functional antagonism during osteogenic
differentiation. J Cell Biochem 2007;101:1109–24.
32. Zhu M, Tang D, Wu Q, et al. Activation of beta-catenin signaling in articular
chondrocytes leads to osteoarthritis-like phenotype in adult beta-catenin conditional
activation mice. J Bone Miner Res 2009;24:12–21.
33. Alfaro MP, Pagni M, Vincent A, et al. The Wnt modulator sFRP2 enhances
mesenchymal stem cell engraftment, granulation tissue formation and myocardial
repair. Proc Natl Acad Sci USA 2008;105:18366–71.
34. Yoshida CA, Komori T. Role of Runx proteins in chondrogenesis. Crit Rev Eukaryot
Gene Expr 2005;15:243–54.
35. Ito Y, Miyazono K. RUNX transcription factors as key targets of TGF-beta superfamily
signaling. Curr Opin Genet Dev 2003;13:43–7.
36. Soung do Y, Dong Y, Wang Y, et al. Runx3/AML2/Cbfa3 regulates early and late
chondrocyte differentiation. J Bone Miner Res 2007;22:1260–70.
37. Domínguez-Soto A, Relloso M, Vega MA, et al. RUNX3 regulates the activity of the
CD11a and CD49d integrin gene promoters. Immunobiology 2005;210:133–9.
38. Sperber SM, Dawid IB. barx1 is necessary for ectomesenchyme proliferation and
osteochondroprogenitor condensation in the zebrafish pharyngeal arches. Dev Biol
2008;321:101–10.
39. Friedman MS, Oyserman SM, Hankenson KD. Wnt11 promotes
osteoblast maturation and mineralization through R-spondin 2. J Biol Chem
2009;284:14117–25.
40. Kim KA, Wagle M, Tran K, et al. R-Spondin family members regulate the Wnt
pathway by a common mechanism. Mol Biol Cell 2008;19:2588–96.
41. Brott BK, Sokol SY. Frodo proteins: modulators of Wnt signaling in vertebrate
development. Differentiation 2005;73:323–9.
42. Lane NE, Lian K, Nevitt MC, et al. Frizzled-related protein variants are risk factors for
hip osteoarthritis. Arthritis Rheum 2006;54:1246–54.
43. Loughlin J. The genetic epidemiology of human primary osteoarthritis: current status.
Expert Rev Mol Med 2005;7:1–12.
44. Loughlin J, Dowling B, Chapman K, et al. Functional variants within the secreted
frizzled-related protein 3 gene are associated with hip osteoarthritis in females. Proc
Natl Acad Sci USA 2004;101:9757–62.
45. Min JL, Meulenbelt I, Riyazi N, et al. Association of the Frizzled-related
protein gene with symptomatic osteoarthritis at multiple sites. Arthritis Rheum
2005;52:1077–80.
46. Ralston SH. New developments in the search for genetic determinants of
osteoarthritis. Osteoarthr Cartil 2007;15:117–19.
47. Bateman JF, Wilson R, Freddi S, et al. Mutations of COL10A1 in Schmid
metaphyseal chondrodysplasia. Hum Mutat 2005;25:525–34.
1885