Identification of Serum Biomarkers for Early-Stage Knee Osteoarthritis Using Proteomics in a Murine Model of Osteoarthritis

Abstract

Objective

To characterize the serum protein profiles of osteoarthritis model mice, particularly mice with early-stage osteoarthritis, using liquid chromatography-mass spectrometry (LC-MS/MS).

Design

Serum and knee samples were collected from 5 control mice and 15 osteoarthritis model mice that underwent destabilization of the medial meniscus (DMM). Osteoarthritic knee samples were collected 4, 8, and 12 weeks after DMM. Knee osteoarthritis severity was scored using the Osteoarthritis Research Society International (OARSI) scoring system. All serum samples were analyzed via LC-MS/MS after removing highly abundant proteins using a ProteoMiner kit (Bio-Rad).

Results

The average OARSI scores of the medial tibial plateau and medial femoral condyle at 8 weeks after DMM (3.40 ± 1.02 points, P = 0.03; 2.60 ± 1.71 points, P = 0.03) and 12 weeks after DMM (8.30 ± 3.12 points, P = 0.03; 4.80 ± 0.75 points, P = 0.03) were significantly higher than the corresponding values in the control group. Compared to those in the control model, 15, 35, and 56 proteins showed different expression levels at 4, 8, and 12 weeks after DMM, respectively. Differentially expressed proteins at 4 weeks after DMM included adenylate kinase, type IV intermediate filament protein, nestin, and insulin-like growth factor-binding proteins. The pathways activated at 4 weeks after DMM differed from those activated at 8 and 12 weeks after DMM.

Conclusion

Protein expression and activation pathways in the osteoarthritis model differed from those in the control model. Several proteins differentially expressed at 4 weeks postoperatively may be involved in the pathogenesis of osteoarthritis and serve as potential biomarkers of early-stage osteoarthritis.

Keywords

osteoarthritis diagnosis knee joint involved serum animal models proteomics

Introduction

Knee osteoarthritis (KOA) is a common joint disease associated with substantial pain, disability, and premature mortality.^1,2 KOA is clinically silent in most individuals during its initial stages, and patients show extensive cartilage deterioration by the time of diagnosis. Although surgical intervention is the treatment of choice for end-stage KOA,³ the early diagnosis can facilitate earlier nonsurgical interventions to modify the course of the disease, reducing patient morbidity and associated costs.⁴ However, early diagnosis and detection of minute changes are difficult using the conventional standard diagnosis and assessment methods based on Kellgren–Lawrence (KL) grading with radiographs. Thus, the identification of specific signs and symptoms, joint structure changes, and biomarkers are necessary to diagnose early KOA (EKOA).⁵

New criteria for detecting EKOA without assessment of radiographic abnormalities were recently proposed at the first international EKOA workshop.⁵ These criteria will allow the identification of patients with mild knee symptoms who have identical risk factors as those with definitive KOA. An increasing number of studies have evaluated the features of EKOA on the basis of these new criteria.^6,7 EKOA prevalence was estimated to be 9.5% and 15.0% in males and females, respectively, with the highest prevalence observed in middle-aged females in the Japanese population.⁸ Nevertheless, the pathology of EKOA remains unclear, and biomarkers that reflect the progression of EKOA are yet to be established.

To date, the identified KOA biomarkers include cartilage oligomeric matrix protein and type II collagen marker.⁹ These biomarkers exhibit changes that reflect the progression of KOA; however, no single clinically applicable biomarker characteristic of EKOA has been identified. More recently, different omics strategies have been employed to identify biomarkers and improve prognosis and diagnosis.¹⁰ In particular, proteomics strategies are being increasingly applied to osteoarthritis (OA) research and for the identification of OA biomarkers.¹¹ In contrast to metabolomics and genomics, proteomics can reflect the patient’s specific condition and allows more stable analyses than evaluations based on metabolites.¹² Proteomics using liquid chromatography–tandem mass spectrometry (LC-MS/MS) can identify and quantify thousands of proteins in a single analysis using a relatively small sample amount, making it an ideal approach for high-throughput analysis of a sample with a high dynamic range, such as serum.^13,14 In the current study, we aimed to identify serum biomarkers for diagnosing EKOA using LC-MS/MS with EKOA model mice. We hypothesized that the detected serum proteins would differ between the early and later stages of KOA.

Methods

Animal Models

Twenty male C57BL/6J mice were purchased from Clea Japan (Tokyo, Japan). The mice were maintained in temperature- and humidity-controlled rooms under a 12-h light-dark cycle (light from 8:00 to 20:00). All mouse experiments were approved by the Committee for the Ethics of Animal Experimentation of Hirosaki University and were conducted in accordance with the Principles of Laboratory Animal Care (National Institutes of Health, publication no. 85-23, revised 1985) and the Guidelines for Animal Experimentation of Hirosaki University. All mice were acclimatized to the laboratory environment for 2 weeks before the experiments. The mice were randomly divided into an OA group (n = 15) and a control group (n = 5). At 8 weeks of age, mice in the OA group underwent surgery to destabilize the medial meniscus (DMM).¹⁵ Briefly, the mice were administered anesthesia with 2.5% isoflurane. After a 3-mm longitudinal incision over the distal patella to the proximal tibial plateau of the right leg, the joint capsule immediately medial to the patellar tendon was incised using a #15 blade. The medial meniscus was then destabilized by sectioning the medial meniscotibial ligament using a #11 blade. The skin was closed using continuous 8-0 tapered Vicryl sutures. The mice exhibited excellent mobility within 2 h after surgery. The control group did not undergo any surgery.

Specimen Collection

The control mice were euthanized at 8 weeks of age for collection of serum and right knee samples. In contrast, 5 OA model mice each were euthanized at 4, 8, and 12 weeks after DMM. Serum protein composition may show post-traumatic changes immediately after DMM or sham surgery; therefore, this study did not evaluate mouse models at 0 weeks after DMM surgery. To collect serum samples under anesthesia with 2.5% isoflurane, 1,000 μL heart blood was obtained in a serum tube microtainer. Afterward, the tubes were gently inverted 6 times to activate clotting factors. After allowing 20 minutes to activate clotting factors, the serum samples were centrifuged at 1,200 × g at 4°C for 10 minutes to remove cells and clotting factors. The serum was stored at −80°C until processing. After collection of the right knee joint, each mouse was euthanized by cervical dislocation.

Histological Analysis

The knee joints were fixed in paraformaldehyde after excising the attached muscle, decalcified in ethylenediaminetetraacetic acid (EDTA) after washing with water, and dehydrated in a graded alcohol series. Finally, the specimens were embedded in paraffin and cut sagittally into serial sections. Each slide was stained with hematoxylin and eosin, safranin-O, and fast green for general histological evaluation.¹⁶ KOA severity at the medial tibial plateau (MTP), medial femoral condyle (MFC), lateral tibial plateau, and lateral femoral condyle was evaluated using the Osteoarthritis Research Society International (OARSI) scoring system.¹⁷ The OARSI scores are presented as mean ± standard deviation.

Serum Concentrations of Minor Protein Fractions

A ProteoMiner kit (Bio-Rad) was used to enrich the minor component proteins in serum samples. In this method, serum proteins bind to beads bearing randomly assembled peptide ligands, removing the excess major proteins and enriching the minor protein fraction. A total of 100 µL of mouse serum was added to the beads equilibrated with phosphate-buffered saline and incubated at room temperature for 2 h with inverted mixing. The beads were washed 3 times with 40 µL of phosphate-buffered saline. The supernatant was removed by centrifugation at 1,000 × g for 1 minute, and 8 µL of eluent (8 M urea, 2% CHAPS) was added to the beads and incubated at room temperature for 15 minutes. The supernatant was collected via centrifugation. The elution procedure was repeated twice, and the eluates were pooled. The pooled eluate was added to 200 µL of acetone and maintained at −20°C for 2 h, following centrifugation at 15,000 × g for 10 minutes to remove the supernatant.

Proteomics Analysis

Label-free proteomics analyses of minor protein fractions in serum were performed using a NanoLC Eksigent 400 system coupled to a Triple TOF 6600 mass spectrometer (AB Sciex, Framingham, MA, USA). Briefly, acetone-precipitated proteins were denatured with 50 mM ammonium bicarbonate containing 8 M urea and reduced with 20 mM dithiothreitol. After incubation at 95°C for 15 minutes, iodoacetamide was added to achieve a final concentration of 40 mM to alkylate the free cysteines, followed by incubation for 30 minutes at room temperature in the dark. After 8-fold dilution with 50 mM ammonium bicarbonate, 1 µg of trypsin was added and incubated at 37°C for 24 h. Peptides were desalted and purified using ZipTip C18 (Merck Millipore). The tryptic peptides were analyzed using a mass spectrometer operated in information-dependent acquisition mode. The acquired spectra were searched against the UniProt database using a Paragon algorithm embedded in the ProteinPilot 5.0.1 software program (AB Sciex). Identification was considered positive when the identified proteins and peptides reached a 1% local false discovery rate. Each sample was assayed in a data-independent acquisition mode, and the proteins were quantified using information-dependent acquisition data as a library.

Statistical Analysis

The average OARSI scores were analyzed using the Steel–Dwass test. Statistical significance was set at P < 0.05. Statistical analysis of the proteomics data was performed using transformed normalized abundances in the t-test, with statistical significance set at P < 0.05. The expression levels between the control and OA groups were compared. Differentially expressed proteins were defined as those with a >2.0- or <0.5-fold change in expression, with P values <0.05. A protein expression level showing >2.0-fold change was defined as protein upregulation, and that showing <0.5-fold change was defined as protein downregulation.

Partial least squares discriminant analysis (PLS-DA) plots and heatmap analysis were performed using MetaboAnalyst 5.0 (http://www.Metaboanalyst.ca/; accessed on February 26, 2023). The heatmap was created for proteins with corrected P-values < 0.05 in a one-way analysis of variance among four groups. The statistical extraction of canonical pathways was performed using KeyMolnet (KM Data, Tokyo, Japan).¹⁸ The list of proteins was imported into KeyMolnet, and the “Interrelation search” algorithm was used to generate the network of molecular interactions between the related proteins. The top canonical pathways associated with the generated network were extracted from the KeyMolnet knowledge base, and P-values were calculated using a hypergeometric test. For pathway analysis, the protein list contained differentially expressed proteins with a >2.0-fold change or <0.5-fold change and P-values <0.05. IBM SPSS Statistical Software (version 25.0; IBM Corp., Armonk, NY, USA) was used for all analyses.

Results

Severity of OA in Histological Analysis

Compared with the control group, the OA groups showed time-dependent OA-related changes in the medial femorotibial joint (Fig. 1). The average OARSI score for both the MTP and MFC in the control model was 0. However, the average OARSI score for the MTP at 4, 8, and 12 weeks after DMM was 1.70 ± 1.89, 3.40 ± 1.02, and 8.30 ± 3.12, respectively. Compared with the control group, the average OARSI score for the MTP was significantly higher at 8 (P = 0.03) and 12 weeks after DMM (P = 0.03). The average OARSI score for the MFC at 4, 8, and 12 weeks after DMM was 1.40 ± 1.96, 2.60 ± 1.71, and 4.80 ± 0.75, respectively. Compared with the control group, the average OARSI score of the MFC was significantly higher at 8 weeks after DMM (P = 0.03) and 12 weeks after DMM (P = 0.03). No significant changes were observed in the OARSI scores of the lateral femorotibial joint (Fig. 1).

Figure 1.

Histological analysis of the medial tibial plateau (A–H). Regional vertical fissures (D: black arrow) of the cartilage surface were detected 12 weeks after DMM. Scale bar: 200 μm. OARSI scores of control mice and the 4-, 8-, and 12-week post-DMM surgery model mice (I–L). OARSI scores of the medial tibial plateau (I) and medial femoral condyle (J) are increase at all time points. No statistical difference in the scores can be observed on the lateral side of the joint (K and L). The average OARSI scores were analyzed using the Steel–Dwass test. Statistical significance was set at P < 0.05. DMM = destabilization of the medial meniscus; OARSI = Osteoarthritis Research Society International.

PLS-DA and Heatmap Analysis

PLS-DA modeling showed modest separations between the control group and the 4-week post-DMM model. In contrast, the 8- and 12-week post-DMM models showed marked group separations from the control group and 4-week post-DMM model (Fig. 2). The heatmap of clustered protein intensities revealed the patterns of protein upregulation and downregulation among the experimental groups. The significant differentially expressed proteins that correlated with the deviation between the control and OA groups showed different profiles. The 12-week post-DMM model showed more specific protein upregulation than the other groups (Fig. 3).

Figure 2.

PLS-DA score plots of the control group and the 4-, 8-, and 12-week post-DMM models. The shaded regions on the PLS-DA plots correspond to the 95% confidence interval region. Partial least squares discriminant analysis (PLS-DA) plots were performed using MetaboAnalyst 5.0 (http://www.Metaboanalyst.ca/; accessed on February 26, 2023). DMM = destabilization of the medial meniscus; PLS-DA = partial least squares discriminant analysis.

Figure 3.

Heat maps of differentially expressed proteins show different profiles based on time-dependent osteoarthritis changes. Each column indicates a different sample group. Each row indicates one protein. The red shading indicates increased expression, and the blue shading indicates decreased expression. Heatmap analysis was performed using MetaboAnalyst 5.0 (http://www.Metaboanalyst.ca/; accessed on May 24, 2024). The heatmap was plotted for proteins with corrected P-values < 0.05 in one-way analysis of variance among four groups.

Proteomic Analysis

LC-MS/MS analysis of the serum identified 605 proteins (Supplemental data S1). Highly abundant proteins in the serum, such as albumin and IgG, were removed using a ProteoMiner kit (Bio-Rad). Compared with that in the control group, 15 proteins were differentially expressed in the 4-week post-DMM group, with 8 proteins upregulated and 7 downregulated. In the 8-week post-DMM group, 35 proteins were differentially expressed, including 9 upregulated and 26 downregulated, when compared with the control group. In the 12-week post-DMM group, 56 proteins were differentially expressed compared with the control group, with 19 proteins upregulated and 37 downregulated (Table 1).

Table 1.

Differentially Expressed Proteins on Comparing the Control Group With the 4-, 8-, and 12-Week Post-DMM Group.

Protein	Fold Change	P-Value
a. Four weeks after DMM
Beta5-tublin	5.833	0.023
RELMbeta	4.755	0.02
Fiblin 1	3.037	0.029
Adk 1p	2.682	0.001
AK	2.682	0.001
b-tub	2.207	0.023
5'Ntdase	2.107	0.031
NT5C	2.107	0.031
AADISADH	0.41	0.027
IF-Ⅳ	0.28	0.001
Nestin	0.28	0.001
ALDO	0.173	0.032
AldlaseB	0.173	0.032
IGF-BP	0.171	0.006
IGF-BP5	0.171	0.006
b. Eight weeks after DMM
beta5-tublin	4.293	0.024
SAA	3.661	0.007
SAA4	3.661	0.007
ECM1	2.989	0.006
Lumican	2.721	0.029
PSMA3	2.64	0.047
TIMP	2.488	0.021
TIMP-3	2.488	0.021
PDXK	2.423	0.036
Arginase	0.464	0.008
Arginase1	0.464	0.008
Adipokine	0.448	0.021
AADISADH	0.414	0.038
Cyclophilin	0.339	0.008
PPIase	0.339	0.008
PPIC	0.339	0.008
IF-Ⅳ	0.23	0.001
Nestin	0.23	0.001
SLPI	0.205	0.032
WAP	0.205	0.032
Vitronectin	0.116	0.024
Gelsolin	0.101	0.038
Angiostatin	0.042	0.022
Plasmin	0.042	0.022
Plasminogen	0.042	0.022
Acylation-SP	0.01	0.021
C3c	0.01	0.021
C3dg	0.01	0.021
C3f	0.01	0.021
compl-C3	0.01	0.021
compl-C3a	0.01	0.021
compl-C3b	0.01	0.021
compl-C3d	0.01	0.021
iC3	0.01	0.021
iC3b	0.01	0.021
c. Twelve weeks after DMM
GRP94	5.029	0.045
C5 convertase	4.579	0.001
Factor B	4.579	0.001
Factor B-Ba	4.579	0.001
c. Twelve weeks after DMM
Factor B-Bb	4.579	0.001
PDXK	4.266	0.033
Adk 1p	3.413	0.012
AK	3.413	0.012
HSP90	2.87	0.045
ApoC-Ⅳ	2.524	0.048
PLA1	2.439	0.024
ADF	2.383	0.047
HSP	2.336	0.045
PLA	2.309	0.024
C3bBb	2.294	0.001
C3bBb3b	2.294	0.001
iC3Bb	2.294	0.001
Copine	2.001	0.041
Copine1	2.001	0.041
HDL	0.498	0.033
HDL2	0.498	0.033
HDL3	0.498	0.033
AADISADH	0.453	0.036
PGADH	0.431	0.005
PSP	0.429	0.011
Adipokine	0.406	0.021
LPL	0.395	0.024
C4b2a3b	0.335	0.021
PP2AA	0.323	0.043
PP2AAa	0.323	0.043
IGF-BP	0.306	0.013
IGF-BP5	0.306	0.013
SLP1	0.247	0.038
WAP	0.247	0.038
C1complex	0.199	0.035
compl-c1q	0.199	0.035
C1CC	0.125	0.035
ApoD	0.115	0.033
Angiostatin	0.107	0.026
Plasmin	0.107	0.026
Plasminogen	0.107	0.026
DENND2	0.095	0.038
DENND2A	0.095	0.038
LCAT	0.093	0.039
Vitronectin	0.044	0.021
Gensolin	0.04	0.032
Acylation-SP	0.01	0.021
C3c	0.01	0.021
C3dg	0.01	0.021
C3f	0.01	0.021
compl-C3	0.01	0.021
compl-C3a	0.01	0.021
compl-C3b	0.01	0.021
compl-C3d	0.01	0.021
iC3	0.01	0.021
iC3b	0.01	0.021

Statistical analysis of the proteomics data was performed using transformed normalized abundances in the t-test. Statistical significance was set at P < 0.05. The expression levels between the control and OA groups were compared. Differentially expressed proteins were defined as those with a >2.0- or <0.5-fold change in expression, with P-values < 0.05.

Compared with the control group, the 4-week post-DMM model shows significant upregulation or downregulation of 15 proteins (a). Compared with the control group, the 8-week post-DMM model shows significant upregulation or downregulation of 35 proteins (b). Compared with the control group, the 12-week post-DMM model shows significant upregulation or downregulation of 56 proteins (c). DMM = destabilization of the medial meniscus.

Pathway Analysis

To identify relationships between the molecular network of the serum proteins and the canonical pathway, an “interrelation” network search was performed using KeyMolnet. In the extracted molecular network, several pathways relevant to the differentially expressed proteins in the 4-, 8-, and 12-week post-DMM model scored > 20 and significantly contributed to the extracted network when compared with the control group (Table 2).

Table 2.

Pathways Relevant to the Differentially Expressed Proteins Between the Control Group and the 4-, 8-, and 12-Week Post-DMM Model.

Rank	Name	Score	P-Value
a. Four weeks after DMM
1	Intermediate filament signaling pathway	53.244	9.38E-17
2	p160 SRC signaling pathway	52.418	1.66E-16
3	Microtubule-associated protein signaling pathway	41.202	3.95E-13
4	GSK3 signaling pathway	41.158	4.08E-13
5	Mst(Hippo) signaling pathway	39.963	9.33E-13
6	Wnt signaling pathway	39.54	1.25E-12
7	Autophagy-related protein signaling pathway	39.329	1.45E-12
8	MAPK signaling pathway	38.142	3.30E-12
9	Transcriptional regulation by SMAD	32.659	1.48E-10
10	Arrestin signaling pathway	31.252	3.91E-10
b. Eight weeks after DMM
1	Complement pathway	108.621	2.00E-33
2	MMP signaling pathway	74.647	3.38E-23
3	Alternative complement pathway	63.445	7.96E-20
4	Fibrinolysis system	59.351	1.36E-18
5	Calpain signaling pathway	53.875	6.05E-17
6	SIBLING signaling pathway	52.855	1.23E-16
7	S100 family signaling pathway	45.455	2.07E-14
8	Transcriptional regulation by STAT	43.866	6.24E-14
9	ADAM signaling pathway	40.541	6.25E-13
10	PIN1 signaling pathway	39.831	1.02E-12
c. Twelve weeks after DMM
1	Complement pathway	186.506	7.18E-57
2	Classical complement pathway	94.511	3.54E-29
3	Lectin complement pathway	71.22	3.64E-22
4	Alternative complement pathway	66.695	8.37E-21
5	HSP90 signaling pathway	53.913	5.90E-17
6	MMP signaling pathway	45.727	1.72E-14
7	Fibrinolysis system	44.639	3.65E-14
8	AMPK signaling pathway	41.908	2.42E-13
9	p160 SRC signaling pathway	36.3	1.18E-11
10	Adipokine signaling pathway	36.161	1.30E-11

For pathway analysis, the protein list contained differentially expressed proteins with a >2.0- or <0.5-fold changes and P-values < 0.05. The statistical extraction of canonical pathways was performed using KeyMolnet (KM Data, Tokyo, Japan).¹⁸ The list of proteins was imported into KeyMolnet, and the “Interrelation search” algorithm was used to generate the network of molecular interactions between the related proteins. The top canonical pathways associated with the generated network were extracted from the KeyMolnet knowledge base, and P-values were calculated using a hypergeometric test. (a) Pathways relevant to the differentially expressed proteins between the control group and the 4-week post-DMM model. (b) Pathways relevant to the differentially expressed proteins between the control group and the 8-week post-DMM model. (c) Pathways relevant to the differentially expressed proteins between the control group and the 12-week post-DMM model. DMM = destabilization of the medial meniscus.

Comparing the control and 4-week post-DMM model, the top three relevant pathways were the intermediate filament signaling pathway (score 53.244), p160 SRC (Steroid receptor coactivators) signaling pathway (score 52.418), and microtubule-associated protein signaling pathway (score 41.202). Upon comparing the control and 8-week post-DMM model, the top three relevant pathways were the complement pathway (score 108.621), matrix metalloprotease (MMP) signaling pathway (score 74.647), and alternative complement pathway (score 63.445). Comparing the control and 12-week post-DMM model, the top three relevant pathways were the complement pathway (score 186.506), classical complement pathway (score 94.511), and lectin complement pathway (score 71.22). Comparing the control and 4-week post-DMM model, the relevant pathways were distinct from the pathways identified in the 8- and 12-week post-DMM model.

Molecular Network Analysis

We conducted a precise network analysis focused on the differentially expressed proteins to extract abnormal signaling patterns using KeyMolnet.¹⁸ Network analysis showed three different molecular networks in the 4-, 8-, and 12-week post-DMM models (Figs. 4 –6). The 4-week post-DMM model exhibited fewer correlated proteins than the other models. In addition, the 4-week post-DMM model showed no correlations with complements, including C3, C5, and factor B.

Figure 4.

Molecular network analysis by KeyMolnet between the control group and the 4-week post-DMM model. The statistical extraction of canonical pathways was performed using KeyMolnet (KM Data, Tokyo, Japan). The list of proteins was imported into KeyMolnet, and the “Interrelation search” algorithm was used to generate the network of molecular interactions between the related proteins. Oval, rectangular, capsule-like, hexagonal, and asymmetrical hexagonal shapes represent proteins, metabolites, complexes, exogenous molecules, and micro-RNA, respectively. The molecular relationships are indicated by a solid line with an arrow (direct binding or activation), a solid line with an arrow and stop (direct inactivation), a solid line without an arrow (complex formation), a dashed line with an arrow (transcriptional activation), and a dashed line with an arrow and stop (transcriptional repression). The black line indicates correlations of the expressed molecules. The gray lines indicate correlations of unexpressed molecules. Red, violet, pink, light blue, blue, and dark blue coloration indicate expression levels over 2-fold higher, within 2-fold higher, within 1.5-fold lower, within 1-fold lower, within 0.75-fold lower, and under 0.5-fold, respectively. DMM = destabilization of the medial meniscus.

Figure 5.

Molecular network analysis by KeyMolnet between the control group and the 8-week post-DMM OA model. The statistical extraction of canonical pathways was performed using KeyMolnet (KM Data, Tokyo, Japan). The list of proteins was imported into KeyMolnet, and the “Interrelation search” algorithm was used to generate the network of molecular interactions between the related proteins. Oval, rectangular, capsule-like, hexagonal, and asymmetrical hexagonal shapes represent proteins, metabolites, complexes, exogenous molecules, and micro-RNA, respectively. The molecular relationships are indicated by a solid line with an arrow (direct binding or activation), a solid line with an arrow and stop (direct inactivation), a solid line without an arrow (complex formation), a dashed line with an arrow (transcriptional activation), and a dashed line with an arrow and stop (transcriptional repression). The black line indicates correlations of the expressed molecules. The gray lines indicate the correlations of unexpressed molecules. Red, violet, pink, light blue, blue, and dark blue coloration indicate expression levels over 2-fold higher, within 2-fold higher, within 1.5-fold lower, within 1-fold lower, within 0.75-fold lower, and under 0.5-fold, respectively. DMM = destabilization of the medial meniscus; OA = osteoarthritis.

Figure 6.

Molecular network analysis by KeyMolnet between the control group and the 12-week post-DMM OA model. The statistical extraction of canonical pathways was performed using KeyMolnet (KM Data, Tokyo, Japan). The list of proteins was imported into KeyMolnet, and the “Interrelation search” algorithm was used to generate the network of molecular interactions between the related proteins. Oval, rectangular, capsule-like, hexagonal, and asymmetrical hexagonal shapes represent proteins, metabolites, complexes, exogenous molecules, and micro-RNA, respectively. The molecular relationships are indicated by a solid line with an arrow (direct binding or activation), a solid line with an arrow and stop (direct inactivation), a solid line without an arrow (complex formation), a dashed line with an arrow (transcriptional activation), and a dashed line with an arrow and stop (transcriptional repression). The black line indicates correlations of the expressed molecules. The gray lines indicate the correlations of unexpressed molecules. Red, violet, pink, light blue, blue, and dark blue coloration indicate expression levels over 2-fold higher, within 2-fold higher, within 1.5-fold lower, within 1-fold lower, within 0.75-fold lower, and under 0.5-fold, respectively. DMM = destabilization of the medial meniscus; OA = osteoarthritis.

Discussion

The objective of the current study was to investigate the proteins expressed during the early stages of KOA. The most important finding of this study was the detection of serum changes in mice over time after joint trauma using proteomics. Differentially expressed proteins were detected in the control and OA groups. PLS-DA and heat maps revealed that protein properties and expressions differed at 4, 8, and 12 weeks post-DMM surgery. Moreover, the activated pathways and molecular networks in the 4-week post-DMM model were distinct from those observed in the 8- and 12-week post-DMM models. These results may help characterize the pathology of EKOA and facilitate the identification of EKOA biomarkers.

Luyten et al. stated that EKOA is defined by the presence of some knee symptoms with KL grade¹⁹ 0-I radiographs showing the absence of osteophytes in the knee.⁵ In their histological assessment of EKOA in humans, Madry et al.²⁰ defined EKOA as OARSI grade 1–3 (depth progression into cartilage) and OARSI score 1–12. In contrast, it is unclear how many weeks post-DMM mice can be deemed equivalent to EKOA in humans. Huang et al.²¹ focused on OA changes in the DMM model mice and reported increased OARSI scores in the MTP and MFC over time when compared with the sham group. The authors further reported that the period beyond 5 weeks post-DMM represents late stages of OA, exhibiting subchondral bone with bone sclerosis, substantially increased bone mineral density in medial compartments, and severe osteophyte formation around the joint when compared with the sham group. In the present study, all mice were within OARSI grade 1–3, with an OARSI score of ≤12. Based on previous reports, the 4-week post-DMM model was considered to be the closest to the EKOA stage in humans.

Upon comparing the control and 4-week post-DMM groups, 15 proteins showing different expression levels were identified: adenylate kinase 1 protein (Adk 1p), adenylate kinase (AK), type IV intermediate filament protein (IF-IV), nestin, insulin-like growth factor-binding proteins (IGFBPs), insulin-like growth factor-binding protein 5 (IGFBP-5), tubulin (beta5-tublin and b-tub), resistin-like molecules beta (RELMbeta), fibulin-1, 5', 3'-nucleotidase, cytosolic (5'Ntdase, NT5C), aminoadipate-semialdehyde dehydrogenase (AADISADH), and fructose biphosphatealdolase (ALDO, aldolase B).

Among the detected proteins, Adk 1p, AK, IF-IV, nestin, IGFBPs, and IGFBP-5 showed lower P-values (P = 0.001). We searched for literature that reported a relationship between OA and the detected proteins.

Nestin is a cytoskeletal protein classified as a type IV intermediate filament, and its protein sequence is similar to that of type IV intermediate filaments.²² Although this protein was initially detected in neural stem cells of the brain,²³ it was subsequently found to be expressed in various tissues and stem or progenitor cells. Nestin+ cells from the adult brain can generate differentiated cells of neuronal and astrocytic lineages,²³ whereas nestin+ cells from the adult bone marrow can differentiate into osteoblastic, adipocytic, and chondrocytic lineages.²⁴ Aberrant mechanical stress plays a role in the alterations observed in the subchondral bone during OA development.²⁵ Zhen et al. demonstrated that transforming growth factor b1 (TGF-b1) was activated in the subchondral bone of the knee in response to abnormal mechanical loading in an OA mouse model. Higher concentrations of TGF-b1 induced the formation of nestin+ mesenchymal stem cell clusters and led to the formation of marrow osteoid islets along with high levels of angiogenesis.²⁵ The association between nestin-positive cells and OA is becoming more apparent, and the correlation between serum nestin and EKOA warrants future examination.

IGFBP-5 is one of the IGFBPs regulating IGF bioavailability by binding to IGF ligands.²⁶ Clemmons et al.²⁷ investigated the effect of IGFBP-5 on improvements in joint architecture in canine OA models. Increased IGF-1 expression reportedly inhibits MMPs by enhancing the cartilage synthesis of tissue inhibitors of metalloproteases.²⁸ An increase in the concentration of IGFBP-5 in synovial fluid leads to a rise in IGF-1 levels. Thus, an increase in IGFBP-5 contributes to an increase in IGF-1 levels in the synovial fluid and reduces joint destruction. Although the relationship and mechanism between the expression levels of IGFBP-5 in serum and joint fluid remain elusive, a decrease in serum IGFBP-5 expression could indicate the onset of EKOA.

AKs are phosphotransferases that regulate nucleotide homeostasis and energetic metabolism by catalyzing reversible phosphoryl transfer.²⁹ A search of databases, such as PubMed, failed to identify any relevant literature on the association between AKs and OA. Thus, further studies are required to identify their potential as EKOA biomarkers.

Fibulin-1 is one of the proteins that was only detected in the 4-week post-DMM model. Tardif et al.¹² performed a proteome analysis of human serum proteins between patients with OA and controls and found that revealed that fibulin-1 was markedly upregulated in patients with OA. Fibulin-1 is an extracellular matrix (ECM) protein that assembles into microfibrils to form the template for elastic fiber formation. In cartilage metabolism and OA pathogenesis, fibulin-1 can activate the TGF-b pathway and promote chondrocyte proliferation via the phosphorylation of Smad2 in the signaling pathway.³⁰ Moreover, fibulin-1 is a new regulator of ADAMTS-1-mediated proteoglycan proteolysis and may play an important role in proteoglycan turnover in tissues.³¹ Fibulin-1 would be an interesting protein for further analysis as a potential EKOA biomarker.

Herein, we compared pathways activated at 4, 8, and 12 weeks after DMM and detected differences in the pathways activated before and after 4 weeks after DMM. Although the complement pathway was the most activated at 8 and 12 weeks after DMM, it was not activated at 4 weeks post-DMM. The complement system is an important component of the innate immune system and consists of the alternative, classical, and lectin complement pathways. Activation of the complement system is evident in patients with OA. The alternative, classical, and lectin complement pathways are activated in the synovial fluid of patients with OA.³² Quantitative proteomics of serum from patients with OA revealed that the levels of complement proteins (C3, 4A, 6, and 9) were substantially altered in late OA (KL-Ⅳ) when compared with levels detected in moderate OA (KL-II).¹⁴ The absence of activation of the complement pathway at 4 weeks post-DMM may reflect the pathology of EKOA and its differences from the later stages of OA.

The MMP pathway was identified as an activated pathway in the 8- and 12-week post-DMM models. MMPs are enzymes that can degrade all ECM components. Within each synovial joint, the ECM undergoes degradation to produce several inflammatory mediators,³³ which stimulate the production of MMPs. MMP-1, produced by synovial cells, and MMP-13, produced by chondrocytes,³⁴ play notable roles in OA because they function as rate-limiting factors in the collagen-degradation process. In addition, the expression of other MMPs, such as MMP-2, MMP-3, and MMP-9, is known to be elevated in arthritis, and these enzymes degrade the non-collagenous substrate components of the joint.³⁵ The results of the molecular network analysis revealed that the only relevant MMP in the 4-week-post-DMM model was MMP2, which differed from that in the 8- and 12-week DMM models.

Conversely, pathways activated in the 4-week post-DMM model were the intermediate filament, p160 SRC, microtubule-associated protein, GSK3, Wnt, and mitogen-activated protein kinase (MAPK) signaling pathways. Members of the p160 SRC family, SRC-1, SRC-2, and SRC-3, interact with nuclear receptors and other transcription factors to drive target gene expression by assembling transcriptional coactivator complexes to increase the transcription of growth-related genes.³⁶ In a mouse model of OA, loss of the nuclear receptor coactivator 3 (NCOA3), which is one of the SRC-3, promoted post-traumatic OA by enhancing NF-κB activation.³⁷ The Wnt signaling pathway is involved in the homeostasis and degradation of the cartilage matrix by regulating the expression of anabolic or catabolic genes, such as increased expression of MMP–2, MMP–3, MMP–9, MMP–13, ADAMTS4, and ADAMTS5, which promote the degradation of the ECM and trigger KOA.³⁸ Gsk3b is critical for maintaining the chondrocytic phenotype by controlling the canonical Wnt signaling pathway. Inhibition of Gsk3b in chondrocytes ex vivo leads to loss of cartilage markers expression, induces matrix degradation by stimulating the expression of MMPs, and inhibits chondrocyte proliferation.³⁹ DDR-mediated p38 MAPK activation can induce MMP-13 activity.⁴⁰ The intermediate filaments relay mechanical signals into downstream biochemical responses. At the tissue level, the ECM, composed mostly of collagen II and proteoglycans, provides mechanical integrity; at the cellular level, the cytoskeleton fulfills this role, with vimentin intermediate filaments exerting a major function.⁴¹ Early OA changes in developmental dysplasia of the hip (DDH) cartilage have been shown to originate at the chondrocyte level and could be linked to intermediate filaments.⁴² Although the undamaged DDH cartilage exhibits minimal ECM changes, a pronounced reorganization of vimentin intermediate filaments occurs. The activated pathways detected in the 4-week post-DMM model were associated with the pathology of KOA. Further research is necessary to comprehensively clarify the pathogenesis of EKOA.

In recent years, several studies have explored potential human serum proteomic biomarkers for KOA.^43-46 In a large study that examined 4792 proteins in human plasma, cartilage acidic protein 1 (CRTAC1) was found to be associated with OA pain.¹ In the Rotterdam cohort, serum CRTAC1 was identified as a biomarker for the severity and progression of radiographic KOA after adjusting for age, sex, and body mass index.⁴⁴ Multiple-reaction-monitoring mass spectrometry of human serum peptides in the New York University cohort also revealed that CRTAC1 had the highest odds of predicting OA progression.⁴⁶ In the Foundation for the National Institutes of Health cohort, 3 proteins, VTDB, CRTAC1, and C1R, were selected as essential biomarkers, demonstrating their importance as prognostic indicators of KOA progression.⁴⁵ CRTAC1 is a glycosylated ECM protein that is found in the interterritorial matrix of articular deep zone cartilage.⁴⁷ Patients with OA reportedly exhibit higher concentrations of CRTAC1 in the knee synovial fluid than controls, along with higher levels in OA cartilage.⁴⁸ Recently, CRTAC1 has attracted attention as a candidate novel biomarker for KOA. However, its function in the EKOA is yet to be comprehensively established, and CRTAC1 was not detected as a serum biomarker in the current study. Examining whether CRTAC1 can be detected in the knee articular cartilage of mice 4 weeks after DMM surgery as an EKOA model may provide more detailed information on the mechanism underlying EKOA.

Although not a serum protein, the ECM protein periostin has been identified as a new therapeutic target for preventing the progression of OA in human knee joints. Periostin is a matricellular secretory matrix protein expressed by mesenchymal stem cells and the periosteum.⁴⁹ Chijimatsu et al.⁵⁰ revealed that periostin was positively stained in the erosive surface layer of human knee cartilage. The authors concluded that periostin may be upregulated in chondrocytes in response to mechanical stress. Attur et al.⁵¹ reported that the expression of periostin mRNA was substantially upregulated in osteoarthritic cartilage. The authors suggested that periostin promotes cartilage degeneration in KOA by upregulating MMP-13 via canonical Wnt signaling.⁵¹ In the current study, periostin was not detected as a differentially expressed protein. Activation of the Wnt signaling pathway in the 4-week post-DMM model may be associated with cartilage degeneration by upregulation of periostin expression. To date, there have been no reports regarding periostin expression in human or mouse EKOA cartilage. Immunostaining of periostin in the cartilage of the 4-week post-DMM mouse may provide further insights into the pathogenesis of EKOA.

This study had some limitations. First, this study did not incorporate the use of sham surgery models. The composition of serum protein may exhibit post-traumatic alterations following DMM or sham surgery. Ideally, this study should encompass sham surgery models at 4, 8, and 12 weeks post-sham surgery to validate the impact of post-traumatic changes resulting from the DMM surgery. Consequently, our findings do not allow us to dismiss the influence of post-traumatic changes on serum proteins, as this proteome analysis solely compared the control and DMM models. Although mouse joints are similar to human joints, mice and humans show inherent differences (e.g., mouse physes remain open in mature animals compared with the adolescent closure in humans, quadruped vs. bipedal gait).⁵² These differences may limit the clinical generalizability of these data. Finally, our results only support the associations between cartilage damage and specific serum proteins in early post-traumatic OA. Further functional studies should be conducted to explore the mechanisms of EKOA and determine the potential use of the suggested serum proteins or activated pathways as biomarkers for EKOA onset or progression.

Collectively, this study used proteomics analysis to detect differentially expressed proteins by investigating serum changes over time following the onset of OA in mice. The histological evaluations indicated that the 4-week post-DMM model was the closest to EKOA in humans. The differentially expressed proteins in the 4-week post-DMM model included Adk 1p, AK, IF-IV, nestin, IGF-BP, and IGF-BP5. The activated pathways and molecular networks in the 4-week post-DMM model differed from those in the 8- and 12-week post-DMM models. These results may help elucidate the pathology of EKOA and facilitate the identification of EKOA biomarkers.

Supplemental Material

sj-xlsx-1-car-10.1177_19476035251363443 – Supplemental material for Identification of Serum Biomarkers for Early-Stage Knee Osteoarthritis Using Proteomics in a Murine Model of Osteoarthritis

Supplemental material, sj-xlsx-1-car-10.1177_19476035251363443 for Identification of Serum Biomarkers for Early-Stage Knee Osteoarthritis Using Proteomics in a Murine Model of Osteoarthritis by Shohei Yamauchi, Eiji Sasaki, Yota Tatara, Kyota Ishibashi, Takahiro Tsushima, Yuka Kimura, Eiichi Tsuda and Yasuyuki Ishibashi in CARTILAGE

Footnotes

Acknowledgements

The authors thank the Shared Facility Center for Science and Technology, Hirosaki University (SFCST) for helping with proteomic analysis. The authors also thank Editage () for the English editing.

ORCID iD

Shohei Yamauchi

Ethical Considerations

All animal experiments were approved by the Committee for the Ethics of Animal Experimentation of Hirosaki University and were conducted in accordance with the Guidelines for Animal Experimentation of Hirosaki University.

Author Contributions

S.Y., E.S., K.I., T.T., Y.K., E.T., and Y.I. conceived and designed this study. S.Y., E.S., and Y.T. contributed to data acquisition. S.Y., E.S., and Y.T. contributed to data analysis and/or interpretation. S.Y., E.S., and Y.T. drafted the manuscript. Y.K., E.T., and Y.I. critically revised the manuscript for important intellectual content. All authors have approved the final version of the manuscript for publication. Y.I. is the guarantor. The corresponding author attests that all listed authors meet the authorship criteria and that no other authors meeting the criteria have been omitted.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The datasets generated and/or analyzed in the current study are available from the corresponding author upon reasonable request.

Supplemental Material

Supplementary material for this article is available on the Cartilage website at .

References

Liu

Niu

Huang

Tang

, et al. Knee osteoarthritis and all-cause mortality: the Wuchuan osteoarthritis study. Osteoarthritis Cartilage. 2015;23:1154-7. doi:10.1016/j.joca.2015.03.021.

Cleveland

Nelson

Callahan

. Knee and hip osteoarthritis as predictors of premature death: a review of the evidence. Clin Exp Rheumatol. 2019;37 Suppl 120(5):24-30.

Bagherifard

Jabalameli

Mirzaei

Khodabandeh

Abedi

Yahyazadeh

. Retaining the medial collateral ligament in high tibial medial open-wedge osteotomy mostly results in post-operative intra-articular gap reduction. Knee Surg Sports Traumatol Arthrosc. 2020;28(5):1388-93. doi:10.1007/s00167-019-05473-8.

Chu

Williams

Coyle

Bowers

. Early diagnosis to enable early treatment of pre-osteoarthritis. Arthritis Res Ther. 2012;14(3):212. doi:10.1186/ar3845.

Luyten

Bierma-Zeinstra

Dell’Accio

Kraus

Nakata

Sekiya

, et al. Toward classification criteria for early osteoarthritis of the knee. Semin Arthritis Rheum. 2018;47(4):457-63. doi:10.1016/j.semarthrit.2017.08.006.

Ishibashi

Sasaki

Ota

Chiba

Yamamoto

Tsuda

, et al. Detection of synovitis in early knee osteoarthritis by MRI and serum biomarkers in Japanese general population. Sci Rep. 2020;10:12310. doi:10.1038/s41598-020-69328-w.

Ota

Sasaki

Chiba

Kimura

Yamamoto

, et al. Relationship between abnormalities detected by magnetic resonance imaging and knee symptoms in early knee osteoarthritis. Sci Rep. 2021;11:15179. doi:10.1038/s41598-021-94382-3.

Sasaki

Ota

Chiba

Kimura

Sasaki

Yamamoto

, et al. Early knee osteoarthritis prevalence is highest among middle-aged adult females with obesity based on new set of diagnostic criteria from a large sample cohort study in the Japanese general population. Knee Surg Sports Traumatol Arthrosc. 2020;28(3):984-94. doi:10.1007/s00167-019-05614-z.

Kumavat

Kumar

Malhotra

Pandit

Jones

Ponchel

, et al. Biomarkers of joint damage in osteoarthritis: current status and future directions. Mediators Inflamm. 2021;2021:5574582. doi:10.1155/2021/5574582.

10.

Underwood

Gerber

Nguyen

Delitto

Han

Thomas

, et al. Protein signatures and tissue diagnosis of pancreatic cancer. J Am Coll Surg. 2020;230(1):26-36. doi:10.7754/Clin.Lab.2019.190340.

11.

Hulme

Wilson

Peffers

Roberts

Simpson

Richardson

, et al. Autologous chondrocyte implantation-derived synovial fluids display distinct responder and non-responder proteomic profiles. Arthritis Res Ther. 2017;19:150. doi:10.1186/s13075-017-1336-7.

12.

Tardif

Paré

Gotti

Roux-Dalvai

Droit

Zhai

, et al. Mass spectrometry-based proteomics identify novel serum osteoarthritis biomarkers. Arthritis Res Ther. 2022;24:120. doi:10.1186/s13075-022-02801-1.

13.

Folkesson

Turkiewicz

Englund

Önnerfjord

. Differential protein expression in human knee articular cartilage and medial meniscus using two different proteomic methods: a pilot analysis. BMC Musculoskelet Disord. 2018;19(1):416. doi:10.1186/s12891-018-2346-6.

14.

Fernández-Puente

Mateos

Fernández-Costa

Oreiro

Fernández-López

Ruiz-Romero

, et al. Identification of a panel of novel serum osteoarthritis biomarkers. J Proteome Res. 2011;10:5095-101. doi:10.1021/pr200695p.

15.

Glasson

Blanchet

Morris

. The surgical destabilization of the medial meniscus (DMM) model of osteoarthritis in the 129/SvEv mouse. Osteoarthritis Cartilage. 2007;15(9):1061-9. doi:10.1016/j.joca.2007.03.006.

16.

Schmitz

Laverty

Kraus

Aigner

. Basic methods in histopathology of joint tissues. Osteoarthritis Cartilage. 2010;18 Suppl 3:S113-S116. doi:10.1016/j.joca.2010.05.026.

17.

Pritzker

Gay

Jimenez

Pelletier

Revell

Salter

, et al. Osteoarthritis cartilage histopathology: grading and staging. Osteoarthritis Cartilage. 2006;14:13-29. doi:10.1016/j.joca.2005.07.014.

18.

Hirayama

Kobayashi

Mizuguchi

Morikawa

Nagayama

Midorikawa

, et al. Integrated proteomics identified novel activation of dynein IC2-GR-COX-1 signaling in neurofibromatosis type I (NF1) disease model cells. Mol Cell Proteomics. 2013;12(5):1377-94. doi:10.1074/mcp.M112.024802.

19.

Kellgren

Lawrence

. Radiological assessment of osteo-arthrosis. Ann Rheum Dis. 1957;16(4):494-502. doi:10.1136/ard.16.4.494.

20.

Madry

Luyten

Facchini

. Biological aspects of early osteoarthritis. Knee Surg Sports Traumatol Arthrosc. 2012;20(3):407-22. doi:10.1007/s00167-011-1705-8.

21.

Fang

Huang

Welch

Norley

Holdsworth

Beier

, et al. Early changes of articular cartilage and subchondral bone in the DMM mouse model of osteoarthritis. Sci Rep. 2018;8:2855. doi:10.1038/s41598-018-21184-5.

22.

Steinert

Chou

Prahlad

Parry

Marekov

, et al. A high molecular weight intermediate filament-associated protein in BHK-21 cells is nestin, a type VI intermediate filament protein. Limited co-assembly in vitro to form heteropolymers with type III vimentin and type IV alpha-internexin. J Biol Chem. 1999;274:9881-90. doi:10.1074/jbc.274.14.9881.

23.

Mignone

Kukekov

Chiang

Steindler

Enikolopov

. Neural stem and progenitor cells in nestin-GFP transgenic mice. J Comp Neurol. 2004;469:311-24. doi:10.1002/cne.10964.

24.

Méndez-Ferrer

Michurina

Ferraro

Mazloom

Macarthur

Lira

, et al. Mesenchymal and haematopoietic stem cells form a unique bone marrow niche. Nature. 2010;466:829-34. doi:10.1038/nature09262.

25.

Zhen

Wen

Jia

Crane

Mears

, et al. Inhibition of TGF-β signaling in mesenchymal stem cells of subchondral bone attenuates osteoarthritis. Nat Med. 2013;19(6):704-12. doi:10.1038/nm.3143.

26.

Firth

Baxter

. Cellular actions of the insulin-like growth factor binding proteins. Endocr Rev. 2002;23(6):824-54. doi:10.1210/er.2001-0033.

27.

Clemmons

Busby

Jr Garmong

Schultz

Howell

Altman

, et al. Inhibition of insulin-like growth factor binding protein 5 proteolysis in articular cartilage and joint fluid results in enhanced concentrations of insulin-like growth factor 1 and is associated with improved osteoarthritis. Arthritis Rheum. 2002;46(3):694-703. doi:10.1002/art.10222.

28.

Amiri

Christians

. PAPP-A2 expression by osteoblasts is required for normal postnatal growth in mice. Growth Horm IGF Res. 2015;25(6):274-80. doi:10.1016/j.ghir.2015.09.003.

29.

Wujak

Kozakiewicz

Ciarkowska

Loch

Barwiolek

Sokolowska

, et al. Assessing the interactions of statins with human adenylate kinase isoenzyme 1: fluorescence and enzyme kinetic studies. Int J Mol Sci. 2021;22:5541. doi:10.3390/ijms22115541.

30.

Geng

Yang

Zhang

. FBLN1 promotes chondrocyte proliferation by increasing phosphorylation of Smad2. J Orthop Sci. 2022;27(1):242-8. doi:10.1016/j.jos.2020.12.006.

31.

Lee

Rodriguez-Manzaneque

Thai

Twal

Luque

Lyons

, et al. Fibulin-1 acts as a cofactor for the matrix metalloprotease ADAMTS-1. J Biol Chem. 2005;280:34796-804. doi:10.1074/jbc.M506980200.

32.

Struglics

Okroj

Swärd

Frobell

Saxne

Lohmander

, et al. The complement system is activated in synovial fluid from subjects with knee injury and from patients with osteoarthritis. Arthritis Res Ther. 2016;18:223. doi:10.1186/s13075-016-1123-x.

33.

Chan

Macdonald

Litherland

Wilkinson

Skelton

Europe-Finner

, et al. Cytokine-induced MMP13 expression in human chondrocytes is dependent on activating transcription factor 3 (ATF3) regulation. J Biol Chem. 2017;292:1625-36. doi:10.1074/jbc.M116.756601.

34.

Johnson

Pavlovsky

Ortwine

Prior

Man

Bornemeier

, et al. Discovery and characterization of a novel inhibitor of matrix metalloprotease-13 that reduces cartilage damage in vivo without joint fibroplasia side effects. J Biol Chem. 2007;282:27781-91. doi:10.1074/jbc.M703286200.

35.

Mehana

Khafaga

El- Blehi

. The role of matrix metalloproteinases in osteoarthritis pathogenesis: an updated review. Life Sci. 2019;234:116786. doi:10.1016/j.lfs.2019.116786.

36.

Mullany

Lonard

O’Malley

. Wound healing-related functions of the p160 steroid receptor coactivator family. Endocrinology. 2021;162(3):bqaa232. doi:10.1210/endocr/bqaa232.

37.

Zhang

Wang

Guo

Kang

Liu

, et al. NCOA3 loss disrupts molecular signature of chondrocytes and promotes posttraumatic osteoarthritis progression. Cell Physiol Biochem. 2018;49(6):2396-413. doi:10.1159/000493839.

38.

Shu

Miao

Tang

Zhan

Zeng

Jiang

. The GSK3β/βcatenin signaling pathway is involved in HMGB1induced chondrocyte apoptosis and cartilage matrix degradation. Int J Mol Med. 2020;45(3):769-78. doi:10.3892/ijmm.2020.4460.

39.

Miclea

Siebelt

Finos

Goeman

Löwik

Oostdijk

, et al. Inhibition of Gsk3β in cartilage induces osteoarthritic features through activation of the canonical Wnt signaling pathway. Osteoarthritis Cartilage. 2011;19(11):1363-72. doi:10.1016/j.joca.2011.07.014.

40.

Olivotto

Vitellozzi

Fernandez

Falcieri

Battistelli

Burattini

, et al. Chondrocyte hypertrophy and apoptosis induced by GROalpha require three-dimensional interaction with the extracellular matrix and a co-receptor role of chondroitin sulfate and are associated with the mitochondrial splicing variant of cathepsin B. J Cell Physiol. 2007;210(2):417-27. doi:10.1002/jcp.20864.

41.

Haudenschild

Chen

Pang

Steklov

Grogan

Lotz

, et al. Vimentin contributes to changes in chondrocyte stiffness in osteoarthritis. J Orthop Res. 2011;29(1):20-5. doi:10.1002/jor.21198.

42.

Hernandez

Wells

Usheva

Nakonezny

Barati

Gonzalez

, et al. Early-onset osteoarthritis originates at the chondrocyte level in hip dysplasia. Sci Rep. 2020;10:627. doi:10.1038/s41598-020-57431-x.

43.

Styrkarsdottir

Lund

Saevarsdottir

Magnusson

Gunnarsdottir

Norddahl

, et al. The CRTAC1 protein in plasma is associated with osteoarthritis and predicts progression to joint replacement: a large-scale proteomics scan in Iceland. Arthritis Rheumatol. 2021;73(11):2025-34. doi:10.1002/art.41793.

44.

Szilagyi

Vallerga

Boer

Schiphof

Ikram

Bierma-Zeinstra

SMA

, et al. Plasma proteomics identifies CRTAC1 as a biomarker for osteoarthritis severity and progression. Rheumatology (Oxford). 2023;62:1286-95.

45.

Zhou

Soderblom

Reed

Jain

Sun

, et al. A “best-in-class” systemic biomarker predictor of clinically relevant knee osteoarthritis structural and pain progression. Sci Adv. 2023;9(4):eabq5095. doi:10.1126/sciadv.abq5095.

46.

Kraus

Reed

Soderblom

Moseley

Hsueh

Attur

, et al. Serum proteomic panel validated for prediction of knee osteoarthritis progression. Osteoarthr Cartil Open. 2024;6(1):100425. doi:10.1016/j.ocarto.2023.100425.

47.

Steck

Bräun

Pelttari

Kadel

Kalbacher

Richter

. Chondrocyte secreted CRTAC1: a glycosylated extracellular matrix molecule of human articular cartilage. Matrix Biol. 2007;26(1):30-41. doi:10.1016/j.matbio.2006.09.006.

48.

Ritter

Tsang

Shi

Takei

Aliprantis

. Sex-specific protection of osteoarthritis by deleting cartilage acid protein 1. PLoS ONE. 2016;11(7):e0159157.

49.

Coutu

Monette

Rivard

Blostein

Galipeau

. Periostin, a member of a novel family of vitamin K-dependent proteins, is expressed by mesenchymal stromal cells. J Biol Chem. 2008;283(26):17991-8001. doi:10.1074/jbc.M708029200.

50.

Chijimatsu

Kunugiza

Taniyama

Nakamura

Tomita

Yoshikawa

. Expression and pathological effects of periostin in human osteoarthritis cartilage. BMC Musculoskelet Disord. 2015;16215. doi:10.1186/s12891-015-0682-3.

51.

Attur

Yang

Shimada

Tachida

Nagase

Mignatti

, et al. Elevated expression of periostin in human osteoarthritic cartilage and its potential role in matrix degradation via matrix metalloproteinase-13. FASEB J. 2015;29(10):4107-21. doi:10.1096/fj.15-272427.

52.

Wallace

Hislop

Hahn

Erdogan

Brahmachary

June

. Correlations between metabolites in the synovial fluid and serum: a mouse injury study. J Orthop Res. 2022;40(12):2792-802. doi:10.1002/jor.25310.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB