Key words Osteoarthritis - Type 2 Diabetes mellitus - Matrix metalloproteinase - Peripheral
Blood Mononuclear Cell
Introduction
Osteoarthritis (OA) is the most common form of chronic joint disorders, an ongoing
biological degenerative joint disease with erosion of articular cartilage, breakdown
of subchondral bone, and symptoms of inflammation in synovium, causing significant
functional impairment and disability in elderly [1 ]. OA may occur in different kinds of joints, while knee involvement is the most common
one. Female, increasing age, obesity, prior joint damage, genetic predisposition,
abnormal mechanical stress of joints, and endocrine factors are deemed to participate
in the occurrence and progression of OA [2 ]
[3 ]. Recently, growing evidence suggests that metabolic factors play a leading role
in both pathogenesis and progression of OA [3 ]
[4 ]. Proteolytic enzymes contribute to degrading of collagens and proteoglycans. Both
are components of the extracellular matrix (ECM), give rise to brittle cartilage that
is prone to mechanical damage. As a family of zinc-dependent proteolytic enzymes,
matrix metalloproteinases (MMPs) takes part in the degeneration of targeted proteins
through the cleavage of internal peptide bonds [5 ].
Normally, chondrocytes express multifarious proteolytic enzymes such as aggrecanases
and MMPs to regulate matric metabolism, which is responsible for cartilage remodeling
[6 ]. On the contrary, chondrocytes and synovial fibroblasts secrete more enzymes, resulting
in aberrant cartilage destruction in OA. Substantial researches have verified the
relationship between MMPs and OA [7 ]
[8 ]. MMP-1, -2, -8, -9 and -13 protein expression levels have been declared to be correlated
to the severity of the disease [9 ]
[10 ]. This phenomenon has attracted increased attention that MMPs as the promoter of
cartilage degradation, activating each other, producing “waterfall-like” amplification
action and conducting cartilage damage [11 ]
[12 ].
Associating with OA, type 2 diabetes mellitus (T2DM) could be an independent and prominent
risk factor [13 ]
[14 ], leading to concepts of diabetic OA phenotype or diabetes-accelerated OA [15 ]. However, the regulating mechanism of diabetic in OA remains elusive. Maintaining
tissue allostasis, MMPs are altered or increased in both T2DM and OA. However, few
studies evaluate the presence of MMPs in OA patients with diabetes. One purpose of
this study is to clarify the role of MMPs in OA patients, with or without diabetes,
to investigate correlations between MMPs and pathogenesis of diabetic OA.
Lipopolysaccharide (LPS) is one of the structural components of the outer membrane
of Gram-negative bacteria. LPS can induce inflammatory response in human peripheral
blood mononuclear cells (PBMCs), eventually leads to expression and production of
many pro-inflammatory cytokines, such as interleukin (IL) -1β, IL-6, tumor necrosis
factor-α (TNFα) and MMPs. Recent studies have shown that a low-grade systemic inflammatory
process such as diabetes may increase the risk of OA by two-fold [16 ]. Another purpose of this study is to research whether OA patients with diabetes
had a much higher inflammatory response than common OA patients.
Materials and Methods
Participants and study design
From January 2016 to January 2017, 30 knee OA patients without T2DM (OA group) and
20 knee OA patients with T2DM (DM-OA group) admitting at our hospital were enrolled.
Besides, 5 healthy volunteers without OA and T2DM were recruited as a healthy control
(HC) group. OA was diagnosed by clinical and radiological evaluation based on the
American College of Rheumatology criteria. T2DM was diagnosed according to the classic
American Diabetes Association criteria(fasting plasma glucose≥7.0 mmol/L and glycated
hemoglobin (HbA1c)≥6.5%) [17 ], or a clinical diagnosis of T2DM with ongoing antidiabetic treatment.
Patients with evident joint injury, previous knee surgery or with inflammatory arthritis
were excluded from the study. Also, those who received intra-articular corticosteroid
or hyaluronic acid injections within 3 months or with serious cardiovascular disease
(CVD), cerebrovascular conditions within 6 months before study enrollment were excluded
too. Diabetic patients were excluded if they had a history of ketoacidosis or suffering
from unstable or rapidly progressive diabetic retinopathy, nephropathy, or neuropathy.
This study conforms to the provisions of the Declaration of Helsinki and was approved
by the hospital medical ethics committees, and informed consent was obtained from
all participants before the study.
Sample collection
Synovial fluid harvesting was performed prior to total knee arthroplasty (TKA) or
utilizing joint cavity paracentesis, strictly avoiding hemarthrosis. Before collecting
synovial fluid samples of controls, 5 mL saline was injected into the articular cavity
and rinsed repeatedly. Then the flushing fluid was sequentially aspirated, sub-packaged
and stored in liquid nitrogen for later analysis. Fasting blood samples were collected
from a peripheral vein and drawn into vacuum blood tubes containing lithium heparin.
The samples were centrifuged or 20 min at 1,000×g at 4°C. The supernatant was collected
and stored at -80+°C until use.
ELISA assay
Serum and synovial fluid expression levels of MMP-1, -2, -3, -7, -8, -9, -10, -12,
-13 in three groups were detected by ELISA kit (Elabscience, Wuhan, China) according
to the manufacturer’s protocol, respectively. Briefly, the micro ELISA plate has been
pro-coated with an antibody specific to MMPs, and the appropriate diluent serum or
synovial fluid samples were added in each microplate well, followed with incubation
for 90 min at 37°C. After washing with wash buffer, the corresponding secondary antibody
was applied. After incubation for 60 min, the plates were washed again. The optical
density was measured spectrophotometrically at a wavelength of 450 nm. The concentration
of MMPs was calculated from a standard curve.
PBMC collection and stimulation
PBMCs were isolated by Ficoll gradient centrifugation from all participants, diluted
with equal amount of PBS, overlaid on Ficoll medium slowly, and centrifuged at 600 g
for 30 min at 20℃. The PBMCs band was aspirated, washed twice with PBS, and re-suspended
in cell freezing medium (90% FBS, 10% DMSO) before gradient cooling being stored in
liquid nitrogen for later use.
After all the PBMCs samples had been collected, the PBMCs were resuspended in medium
firstly, calculated the number to 2×105 /ml, then stimulated with LPS (a TLR4 ligand) at 100 ng/ml for 4 h, and the cell culture
supernatants were gathered before slit charging and being stored at −80℃. And the
MMPs in the supernatants were analyzed by Elisa Kit (Elabscience Biotechnology Co.,Ltd)
Statistical analysis
All statistical analyses were performed using IBM SPSS Statistic 19.0 (New York, NY).
Continuous variables were expressed as the mean±standard deviation (SD). Statistically
significant differences between multiple groups were determined analyzing variance
(ANOVA) in conjunction with Tukey's multiple comparison. A statistical significance
was defined when P<0.05.
Results
Descriptive statistics
The sample characteristics, including the general data and the laboratory results
of hospital are provided in [Table 1 ]. Our results showed that LDL concentration, fasting glucose and HbA1c in DM-OA group
were higher than in OA group (P<0.05). The age, c-reactive peptide and erythrocyte
sedimentation rate in DM-OA and OA group were higher than in control (P<0.05).
Table 1 Clinical and laboratory characteristics of study subjects.
control
OA
diabetic OA
n
5
20
30
Sex (M/F)
2/3
3/17
7/23
Age (years)
37.6±9.8
67.3±8.7*
70.4±8.3*
Body mass index
23.15±3.98
25.26±3.09
26.83±5.62
OA Duration (years)
-
7.14±3.88
8.43±4.70
Arthritic joints
-
8/12
11/19
Fasting glucose(mmol/L)
5.13±0.65
5.69±0.92
6.36±1.02#
HbA1c (%)
4.58±0.79
5.12±0.91
6.68±1.05# *
LDL cholesterol(mg/dL)
2.45±0.87
2.73±0.75
3.37±1.18#
HDL cholesterol(mg/dL)
1.07±0.08
1.29±0.29
1.41±0.38
Triglycerides(mg/dL)
0.97±0.34
1.36±0.50
1.52±0.76
Total cholesterol(mg/dL)
4.05±2.56
4.49±1.23
5.44±1.36
C-reactiv peptide(mg/L)
2.26±1.78
4.99±3.68*
5.67±4.97*
Erythrocyte sedimentation rate(mm/h)
13.1±6.48
28.45±17.26*
32.65±18.11*
Data were expressed as means ±SD. M, male; F, female; The P value was calculated by
One-way ANOVA tests. *P<0.05 vs. control. # P<0.05 vs. OA.
Basal expression of MMPs in healthy control, OA and DM-OA human synovial fluid
The levels of MMPs-1, -2, -3, -7, -8, -9, -10, -12, -13 in synovial fluid were measured
by ELISA. The expression of MMP-1 was significantly higher in OA and DM-OA group than
that in healthy control (both P<0.0001). Moreover, MMP-1 in DM-OA group was approximately
2.5 times higher than that in OA group (P<0.0001, [Fig. 1a ]).
Fig. 1 Synovial fluid expression level of MMPs in healthy controls (HC group), OA patients
without T2DM (OA group) and OA patients with T2DM (DM-OA group) were measured by ELISA.
a MMP-1; b MMP-2; c MMP-3; d MMP-7; e MMP-8; f MMP-9; g MMP-10; h MMP-12; i MMP-13. Each bar represents the mean±SD. *P<0.05 vs. healthy controls. #P<0.05 vs.
OA alone. p values result from a Tukey's multiple comparison test.
MMP-2, -3 and -13 protein levels were significantly higher in OA and DM-OA group than
healthy control (all P<0.0001). However, there was no significant difference between
OA and DM-OA group in MMP-2, -3 and -13 (P=0.177, 0.999, 0.974, respectively. [Fig. 1b, c and i ]).
The levels of MMP-7, -10 and -12 protein in OA and DM-OA group were substantially
higher than those in healthy control (all P<0.0001, except MMP-12 in OA group, P=0.011).
Surprisingly, the difference in MMP-7, -10 and -12 protein levels between OA and DM-OA
group was statistically significant (all P<0.0001, [Fig. 1d, g and h ]).
On average, MMP-8 and -9 protein levels were approximately 25 and 9 times higher in
DM-OA group than in OA group, respectively. Nevertheless, the difference between OA
group and healthy control did not reach statistical significance (P=0.958 and 0.984,
[Fig. 1e and f ]).
Basal expression of MMPs in healthy control, OA and DM-OA human serum
The levels of MMP-1, -3, -7, -12 and -13 protein in serum in OA and DM-OA group were
obviously higher than in control (P<0.05). Moreover, the levels of MMP-1 and -7 were
the highest in DM-OA group (P=0.001 and P<0.0001, [Fig. 2a and d ]). Nonetheless, MMP-3 in OA group was significantly higher than those in DM-OA group
(P<0.0001, [Fig. 2c ]), and there were no significance difference between OA and DM-OA group in MMP-12
and -13 (P=0.930 and 0.702, [Fig. 2h and i ]).
Fig. 2 MMPs serum expression level in HC, OA and DM-OA groups were measured by ELISA. a MMP-1; b MMP-2; c MMP-3; d MMP-7; e MMP-8; f MMP-9; g MMP-10; h MMP-12; i MMP-13. Each bar represents the mean±SD. *P<0.05 vs. healthy controls. #P<0.05 vs.
OA alone. p values result from a Tukey's multiple comparison test.
However, the levels of MMP-2 protein was the highest in OA group (P<0.0001, [Fig. 2b ]), and there were no differences between MMP-8, -9 and -10 in these three groups
(P=0.160, 0.996 and 0.990, respectively. [Fig. 2e, f and g ]).
Role of PBMC in modulating MMP expression profiles induced by LPS in healthy control,
OA and DM-OA human cell culture supernatant
PMBC failed to secrete MMP-2, -3, -7, -12, -13 before and after 4 h stimulation with
LPS, while analysis of MMPs production in the cell supernatant by ELISA revealed significant
augmentations of MMP-1, -8, -9, -10 after 4 h of stimulation with LPS compared to
those before stimulation, respectively (all P<0.0001, [Fig. 3 ]).
Fig. 3 Role of PBMC in modulating MMPs expression profiles induced by LPS in HC, OA and
DM-OA groups. Levels of MMP-1, -8, -9, -10 before and after 4 h stimulation with LPS
were measured by ELISA. a MMP-1; b MMP-8; c MMP-9; d MMP-10. Each bar represents the mean±SD. *P<0.05 vs. healthy control. #P<0.05 vs.
OA group. p values result from a Tukey's multiple comparison test.
As showed in [Fig. 3a ], in PMBC culture supernatant, secretion of MMP-1 was higher in OA and DM-OA group
than for healthy control before and after stimulation with LPS (P<0.05).
There was no difference in the production of MMP-8 among the three groups in the supernatant
of unstimulated cells (P=0.882, [Fig. 3b ]), and the levels of MMP-9 in OA and DM-OA groups were markedly and significantly
higher than those in healthy control (both P<0.0001, [Fig. 3c ]), whereas there was a significant different after 4 h stimulation with LPS in MMP-8
and -9 (both P<0.0001). The supernatant of PBMC in DM-OA group exhibited much higher
level of MMP-8 and -9 (both P<0.0001, [Fig. 3b and c ]).
Secretion of MMP-10 was higher in OA and DM-OA groups than healthy control in unstimulated
state (both P<0.0001). After 4 h of LPS stimulation, MMP-10 protein levels increased
to 543.44±15.38 pg/mL (HC), 559.16±42.41 pg/mL (OA), and 550.72±58.16 pg/mL (DM-OA)
([Fig. 3d ]), respectively. This increase in the expression of MMP-10 was not significant for
the three groups (P=0.711).
The present study demonstrate that there is significantly more MMP-8, -9 secreted
by PBMC after LPS activation in DM-OA group. Synchronous finding was observed in synovial
fluid.
Discussion
Articular cartilage, a specialized connective tissue, consists of chondrocytes and
ECM. Chondrocytes are unique cellular components embedded in ECM that maintain homeostasis
between synthesis and catabolism under normal physiological environment [18 ]. Cartilage integrity crucially facilitates distributing loads and ensures a near-frictionless
motion of joints.
OA, one of age-related skeletal diseases, results in the progressive damage of the
articular cartilage owing to inflammation of synovium and the consequent imbalance
between catabolism and anabolism of synoviocytes and chondrocytes. MMPs take part
in the process of tissue remodeling, either physiological or pathological [19 ]. Plenty of studies revealed that after articular injury, MMPs, especially MMP-2
and MMP-9, have been participated in the disease progression of OA [20 ]. Catterall et al. [21 ] reported that MMP-3 concentration distinctly elevates in synovial fluid within 3
weeks after ACL injury whereas Tchetverikov [22 ] demonstrated that both MMP-1 and MMP-3 synchronously rose 4 days after collateral
ligaments of knee injury. Several MMPs and degraded aggrecan concentrations in synovial
fluid were increased after an acute intraarticular fracture compared with a uninjured
joint in the same patient, and maintained in the high level at a secondary time [23 ].
A few researches indicated that there was a distinct positive correlation between
plasma glucose levels and radiological evidence of OA [24 ]. Several studies support that OA is more prevalent and severe in T2DM patients [13 ]
[25 ]. OA is increasingly regarded as a “metabolic disorder disease” associated with obesity,
diabetes [26 ].
The cardinal aim of this study was to ascertain the expression differences of MMPs
in synovial fluid and plasma as well as release of MMPs by mononuclear cells after
stimulation with LPS in OA patients with or without diabetes. We found that OA patients
with diabetes have high levels of MMP-1, -7, -8, -9, -10, and -12 in synovial fluid.
These MMPs directly impair cartilage integrity by upsetting modulating balance of
anabolism and catabolism.
The results are presented in [Figs. 1i ] and [2i ] are consistent with previous studies [27 ] that MMP-13 expression is higher in OA than in normal synovial fluid and serum.
In addition, our data revealed that MMP-13 expression in the synovial fluid of OA
patients with or without diabetes did not differ significantly. Rosa et al. [28 ] reported that the expression of MMP-13 and collagen I was not affected by high glucose
in normal chondrocytes. This suggests that MMP-13 is not a major factor of diabetes
susceptibility to OA. Indeed, as a key collagenase, MMP-13 expression has been shown
to be augmented in OA cartilage [29 ], suggesting a significant role in cartilage degradation as it preferentially cleaves
type II collagen fibers [30 ].
Generally, MMP-2 proteins are gelatinases of the MMP family [20 ]. MMP-2 are secreted extracellularly, and these gelatinases cleave other targets,
including growth factors, ECM, and cytokines, which results in the release of these
ligands, which in turn activate major signaling pathways that are involved in cell
growth, invasion, inflammation, and migration [31 ]. Previous study [32 ] showed that, MMP-3 levels in OA patient’s blood and synovial fluid were higher than
that in healthy people, and the increment was consistent with the extent of cartilage
damage. MMP-2 expression in the synovial fluid of OA and DM-OA patients was substantially
greater than that of healthy control, while the difference in serum did not differ
significantly between DM-OA patients and healthy control. Moreover, MMP-2 and MMP-3
in serum were higher in OA patients than in DM-OA patients. We anticipated that upregulated
MMP-2, -3 secretion might have unfavorable effects on cartilage integrity, but it
is not the cause of diabetes susceptibility to OA. There may be a factor of diabetes
that limits the MMP-2, -3 secretion in serum, potentially a protective mechanism exists.
If MMP-2, -3 secretion is to be inhibited from chondrocytes and synovial fibroblasts,
it may be possible to avoid the occurrence of OA.
MMP-7 protein was detected in synovial fluid and serum, we found that the expression
of MMP-7 protein in DM-OA patients was higher than that in OA patients and healthy
control. According to previous studies, high glucose concentrations favor catabolism
in primary human chondrocytes and facilitate the development and progression of OA
[28 ]. Elevated MMP-7 has been found to increase cartilage degradation and play a role
in inflammatory joint disease [33 ]. Therefore, our results indicated that increasing secretion of MMP-7 in serum and
synovial fluid due to diabetes mediates inflammatory OA-like effects in articular
cartilage.
Further detailed studies are warranted to understand the roles of the MMP-1 and -9
proteins in the pathogenesis of OA as a prelude to clinical application of the knowledge
gained [34 ]. Our results indicated that MMP-1,-8, -9, -10 and -12 expressions were highest in
DM-OA patients in synovial fluid. Interestingly, we found unexpected results that
serum did not exhibit such statistically significant differences in MMP-8, -9 and
-10 protein levels among three groups. Hyperglycemia, a participant that led to excessive
secretion of MMP-1 in serum and synovial fluid, is considered one of the reasons for
facilitating the development and/or progression of OA.
Epidemiologic studies have established an association between inflammatory biomarkers
and the occurrence of T2DM and complications. Low-grade inflammation is a common feature
in subjects with T2DM [35 ]
[36 ]. Synovitis has been considered as a secondary phenomenon not associated with systemic
inflammation, but it is responsible for symptoms of OA. Although inflammation is unlikely
to play a primary role in initiating synovitis, growing evidences show that inflammation
accelerates articular degeneration [37 ].
PBMCs consist of lymphocytes (T cells, B cells, NK cells) and monocytes, which can
be extracted from whole blood using ficoll. There are also mononuclear cells in synovial
fluid, called synovial fluid mononuclear cells (SFMCs). Limit to the amount of synovial
fluid, we did not extract mononuclear cells from synovial fluid. Stimulation experiments
with LPS are widely used to study the secretory function of PBMC. In response to LPS,
the releases of MMP-1, -8, -9 and -10 were significantly increased at 4 h after stimulation
of PMBC in three groups. In agreement with other studies [38 ], our presented results revealed no differences in the release of MMP-8 and -9 by
PBMC between OA and DM-OA groups under unstimulated conditions. After 4 h stimulation
with LPS however, PBMC released much more MMP-8 and -9 in DM-OA group than in OA group
or healthy controls. The extracellular glucose concentration, either increased or
decreased, has been shown to directly affect some chondrocyte functions [39 ]
[40 ].
MMPs inside synovial fluid are produced by chondrocytes and synoviocytes especially
when stimulated by cytokines [41 ]. Our results also demonstrate that mononuclear cells have the ability to secrete
MMP-1, -8, -9 and -10. The level of MMPs in synovial fluid may be influenced by function
of synovial fluid mononuclear cells to a certain degree. Although there are few mononuclear
cells inside synovial fluid, the hyperfunction of mononuclear cells caused by inflammation
might be one of the mechanisms of diabetes prone to OA.
There were no significant differences of MMP-10 in three groups after LPS stimulation.
It indicated that the function of PBMC secretion of MMP-10 did not change. MMP-10
in synovial fluid in DM-OA patients may rely on other inflammatory cytokines and MMPs
for upregulation in chondrocytes and synovial fibroblasts, and may not be a malfunction
of PBMC.
Our study had some limitations. First, our sample size was relatively small, especially
the healthy control group, limiting its statistical power to detect existing associations.
Moreover, the age of the healthy control group was much younger than OA and DM-OA
groups. Whether age, blood glucose control and diabetes mellitus complication affect
the level of MMPs in plasma or synovial fluid is unclear, it needs to be investigated
in further study. Further research should evaluate whether diabetes promotes articular
cartilage degradation, facilitating the development and/or progression of OA via of
these MMPs in vivo and in vitro experiment.
In summary, high levels of MMP-1, -7, -8, -9, -10, and -12 in the synovial fluid might
be one of important reasons that diabetes patients are more frequently suffered from
OA. Inflammation-induced malfunction of mononuclear cells would stimulate MMP-8 and
-9 secretion to various extents. Abnormal intrinsic regulatory mechanisms may activate
other MMPs and produce amplification effect. Thus it can significantly increase MMPs
activity and content. When the increment exceeds the degree, MMPs will lose the balance
and further compromising cartilage integrity [42 ], thus promoting the degradation of articular cartilage matrix and inducing OA.
