header advert
Bone & Joint Research Logo

Receive monthly Table of Contents alerts from Bone & Joint Research

Comprehensive article alerts can be set up and managed through your account settings

View my account settings

Visit Bone & Joint Research at:

Loading...

Loading...

Open Access

Arthritis

ATF3 as a potential diagnostic marker of early-stage osteoarthritis and its correlation with immune infiltration through bioinformatics analysis



Download PDF

Abstract

Aims

This study aimed, through bioinformatics analysis, to identify the potential diagnostic markers of osteoarthritis, and analyze the role of immune infiltration in synovial tissue.

Methods

The gene expression profiles were downloaded from the Gene Expression Omnibus (GEO) database. The differentially expressed genes (DEGs) were identified by R software. Functional enrichment analyses were performed and protein-protein interaction networks (PPI) were constructed. Then the hub genes were screened. Biomarkers with high value for the diagnosis of early osteoarthritis (OA) were validated by GEO datasets. Finally, the CIBERSORT algorithm was used to evaluate the immune infiltration between early-stage OA and end-stage OA, and the correlation between the diagnostic marker and infiltrating immune cells was analyzed.

Results

A total of 88 DEGs were identified. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses indicated that DEGs were significantly enriched in leucocyte migration and interleukin (IL)-17 signalling pathways. Disease ontology (DO) indicated that DEGs were mostly enriched in rheumatoid arthritis. Six hub genes including FosB proto-oncogene, AP-1 transcription factor subunit (FOSB); C-X-C motif chemokine ligand 2 (CXCL2); CXCL8; IL-6; Jun proto-oncogene, AP-1 transcription factor subunit (JUN); and Activating transcription factor 3 (ATF3) were identified and verified by GEO datasets. ATF3 (area under the curve = 0.975) turned out to be a potential biomarker for the diagnosis of early OA. Several infiltrating immune cells varied significantly between early-stage OA and end-stage OA, such as resting NK cells (p = 0.016), resting dendritic cells (p = 0.043), and plasma cells (p = 0.043). Additionally, ATF3 was significantly correlated with resting NK cells (p = 0.034), resting dendritic cells (p = 0.026), and regulatory T cells (Tregs, p = 0.018).

Conclusion

ATF3 may be a potential diagnostic marker for early diagnosis and treatment of OA, and immune cell infiltration provides new perspectives for understanding the mechanism during OA progression.

Cite this article: Bone Joint Res 2022;11(9):679–689.

Article focus

  • How can we diagnose and treat osteoarthritis (OA) at the early stage?

  • What is the relationship between OA and immune cell infiltration?

  • What is the significance of this relationship to the treatment of OA?

Key messages

  • Activating transcription factor 3 (ATF3) can be a potential diagnostic marker and therapeutic target for early-stage OA.

  • Immune cell infiltration is of great significance in the development of OA.

Strengths and limitations

  • The present study revealed that ATF3 could be a potential diagnostic marker and therapeutic target for early-stage OA, and we associated ATF3 with immune cell infiltration in OA to elucidate its role in the development of the disease.

  • Experimental verification and a large-scale sample size are still needed for better reliability of the results in the future.

Introduction

Osteoarthritis (OA) is the most common chronic degenerative joint disease which is mainly characterized by synovial inflammation and cartilage degeneration among older people.1,2 OA can cause serious limits on articular functions in major daily activities without an effective treatment.3 Over 200 million people worldwide suffer from OA, which is a great burden both for individuals and for society.4 However, early diagnosis and treatment are imperative to effectively prevent disease progression and joint damage.5 Currently, diagnosis is based on clinical manifestations and imaging, which can only detect quite advanced OA. Moreover, the severity of structural degeneration and clinical manifestations are not consistent.6 The lack of reliable methods to diagnose, stage, and monitor pathological changes in the joint has been one of the major impediments to advances in disease-modifying interventions.7 Therefore, it is essential to identify new and effective biomarkers for the early diagnosis and treatment of OA.

OA has now been recognized as a disease that affects whole-joint tissues, including the synovium (synovial tissue), which can produce synovial fluid and maintain the smoothness and motion of the joint. Some studies indicate that the pathological changes in synovium occur even before visible cartilage degeneration in OA.8 Multiple findings from MRI and ultrasound techniques have confirmed the existence of synovial pathological changes, such as thickening of the synovial lining layer and joint effusion, in early OA.9,10 With the onset of synovitis, proinflammatory cytokines produced by inflamed synovial tissue cause the cartilage breakdown which can, in turn, exacerbate synovitis, creating a vicious circle.11 However, only a few pieces of research focus on exploring the molecular mechanism between synovitis and the progression of OA. An increasing number of studies have demonstrated that cell infiltration plays a crucial role in the development of OA.12 Numerous immune cells have been discovered in the synovial tissue of OA patients. Among them, macrophages, T cells, and mast cells (MCs) are the most frequently identified infiltrating immune cells.13 In addition, a distinct infiltration pattern characterized by the increasing polarization of CD4+ T cells towards activated Th1 cells, and the increased secretion of immunomodulating cytokines, was recently discovered.14 Therefore, understanding infiltration of the immune cells is of great importance to elucidate the mechanisms of OA, while identifying the differences in components of infiltrating immune cells is beneficial to the development of new immunotherapeutic targets. However, no studies have yet analyzed the difference in immune cell infiltration between early-stage OA and end-stage OA.

Methods

Microarray datasets collection and preprocessing

We used the “GEOquery” package of R software (version 4.1.1, R Foundation for Statistical Computing, Austria) to download gene expression profiles of human synovial tissues, including three test sets GSE55235, GSE55457, and GSE55584 based on GPL96 platforms, and two validation sets including GSE12021 based on GPL96 platforms, and GSE32317 based on GPL570 platforms from the GEO database.15 Three test sets GSE55235, GSE55457, and GSE55584 contain 46 samples, including 20 normal samples and 26 OA samples. GSE12021 contains 19 samples, including nine normal samples and ten OA samples (Table I). GSE32317 contains 19 samples, including nine end-stage OA samples and ten early-stage OA samples. GSE55235, GSE55457, and GSE55584 gene expression matrixes were then combined, and the “sva” package was used to remove the inter-batch difference and other unwanted variations.16

Table I.

Information of selected microarray datasets.

Accession numbers Platform Samples Mean age, yrs (SD) Sex, n (male/female) Source tissue Attribute
Normal OA Normal OA Normal OA
GSE55235 GPL96 10 10 N/A N/A N/A N/A Synovium Test set
GSE55457 GPL96 10 10 51 (18.7) 72.4 (5.6) 8/2 2/8 Synovium Test set
GSE55584 GPL96 0 6 N/A 73.2 (7.9) N/A 0/6 Synovium Test set
GSE12021 GPL96 9 10 50.2 (20.7) 71.9 (6.1) 7/2 2/8 Synovium Validation set
GSE32317 GPL570 0 19 (10 early-stage

and 9 end-stage)
N/A 62.3 (10.5) (early-stage) N/A 8/2 (early-stage) Synovium Validation set
69.1 (6.1) (end-stage) 5/4 (end-stage)
  1. Annotation: GPL96: [HG-U133A] Affymetrix Human Genome U133A Array; GPL570: [HG-U133_Plus_2] Affymetrix Human Genome U133 Plus 2.0 Array.

  1. N/A, not available; OA, osteoarthritis.

Identification of DEGs

The “limma” package was used to identify DEGs by comparing the expression value between normal and OA synovial tissues.17 The corresponding p-value of the gene symbols after the independent-samples t-test was defined as adjusted p < 0.05, and |log2FC| > 1.5 were considered as screening criteria. The “ComplexHeatmap” package18 and “ggplot2” package19 were then used to make the complex heatmap and volcano plot to better visualize these DEGs.

Gene ontology, KEGG, and disease ontology analyses of DEGs

We used the “clusterProfiler” package to perform the GO and KEGG pathway enrichment analyses of DEGs.20 A “GOplot” package was used to calculate the Z-score as well as visualize the results of these enrichment analyses.21 Disease ontology (DO) enrichment analysis was performed and visualized by the online tool Enrichr (Icahn School of Medicine at Mount Sinai, USA). A p-value < 0.05 was considered statistically significant.

PPI network construction and identification of the hub genes

The PPI network was constructed using the STRING database with a filter condition (confidence score 0.4).22 Then, Cytoscape (Institute for Systems Biology, USA) was used to better visualize the interaction information performed by STRING. Additionally, MCODE plugin was used to identify significant gene modules (degree cut-off = 2; node score cut-off = 0.2; k-score = 2; maximum depth = 100). The top eight genes were identified by the three algorithms including Degree, Maximal Clique Centrality (MCC), and Maximum Neighborhood Component (MNC) of CytoHubba.23 Next, the top eight genes were verified by the support vector machine-recursive feature elimination (SVM-RFE) algorithm.24 Finally, we intersected all these results to obtain the final hub genes. The GSE12021 dataset was analyzed to verify the hub genes of OA, while the GSE32317 dataset was analyzed to identify the significant differentially expressed genes between early-stage OA and end-stage OA.

Evaluation of immune cell infiltration

CIBERSORT (Stanford University, USA) was used to analyze the gene expression matrix data obtained previously, and the samples were filtered out according to p-value < 0.05. Next, a percentage diagram of each kind of immune cell in every sample was made. Then the “ggplot2” package was used to draw a correlation heatmap which can visualize the correlation of infiltrating immune cells.19 Next, the “ggplot2” package was used to visualize the different immune infiltration levels of each immune cell between early-stage OA and end-stage OA.

Correlation analysis between ATF3 and infiltrating immune cells

Finally, Spearman correlation analysis on ATF3 and infiltrating immune cells was performed using “ggstatsplot” package; “ggplot2” package was used to demonstrate the results.

Results

Identification of DEGs

To identify DEGs between OA and normal synovial tissue, the datasets GSE55235, GSE55457, and GSE55584 were selected. A total of 88 DEGs including 34 upregulated and 54 downregulated genes were detected. The results were visualized by a volcano map of all DEGs (Figure 1a), and a heatmap (Figure 1b) was made to show DEG expression.

Fig. 1 
            Differentially expressed genes (DEGs) of synovial tissue between osteoarthritis (OA) and normal controls. a) Heatmap of DEGs. Red rectangles represent high expression, and blue rectangles represent low expression. b) Volcano plot of DEGs. Red spots represent upregulated genes, and blue spots represent downregulated genes.

Fig. 1

Differentially expressed genes (DEGs) of synovial tissue between osteoarthritis (OA) and normal controls. a) Heatmap of DEGs. Red rectangles represent high expression, and blue rectangles represent low expression. b) Volcano plot of DEGs. Red spots represent upregulated genes, and blue spots represent downregulated genes.

Enrichment analyses of DEGs

GO enrichment analysis revealed that DEGs were mainly enriched in biological processes including leucocyte migration, regulation of inflammatory response, response to steroid hormone, and response to corticosteroid; cellular components including immunoglobulin complex and circulating immunoglobulin complex; and molecular functions including RNA polymerase II-specific DNA-binding transcription activator activity, cytokine activity, antigen binding, and chemokine activity (Figure 2a, Table II, Supplementary Figure a). KEGG enrichment analyses indicated that DEGs were mainly enriched in IL-17 signalling, cytokine-cytokine receptor interaction, osteoclast differentiation, and rheumatoid arthritis-related pathways (Figure 2b, Table II, Supplementary Figure b). DO analysis showed that diseases enriched by these DEGs mainly include rheumatoid arthritis, juvenile arthritis, arthritis, pulmonary hypertension, skin erosion, and OA (Figure 2c).

Fig. 2 
            Gene ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and disease ontology (DO) analyses of differentially expressed genes (DEGs) between osteoarthritis (OA) and normal controls. a) Bubble plot showing the top ten GO terms, including four biological processes, two cell components, and four molecular functions. b) KEGG pathway analysis of DEGs between OA and normal controls: the bubble plot shows the top ten enriched KEGG pathways of DEGs. c) Bar plot showing the DO enrichment analysis, where the horizontal axis represents the number of DEGs under the DO terms. IL-17, interleukin 17; NF-kappa B, nuclear factor kappa B; TNF, tumour necrosis factor.

Fig. 2

Gene ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and disease ontology (DO) analyses of differentially expressed genes (DEGs) between osteoarthritis (OA) and normal controls. a) Bubble plot showing the top ten GO terms, including four biological processes, two cell components, and four molecular functions. b) KEGG pathway analysis of DEGs between OA and normal controls: the bubble plot shows the top ten enriched KEGG pathways of DEGs. c) Bar plot showing the DO enrichment analysis, where the horizontal axis represents the number of DEGs under the DO terms. IL-17, interleukin 17; NF-kappa B, nuclear factor kappa B; TNF, tumour necrosis factor.

Table II.

The Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analysis of differentially expressed genes.

Category ID Description Gene count Adjusted p-value Genes
BP GO:0050900 Leucocyte migration 17 4.88E-08 IGLV1-44, IGLC1, CXCL3, CXCL8, CXCL2, IL6, MMP1, CX3CR1, IGHM, CCL20, APOD, TNFSF11, TREM2, IGKC, SLC7A5, LRCH1, SELE
GO:0050727 Regulation of inflammatory response 14 1.43E-05 IGLV1-44, IGLC1, IL6, PTGS2, APOD, TNFSF11, TREM2, TAC1, IGKC, SOCS3, IL1R2, TLR7, IGHG1, SELE
GO:0031960 Response to corticosteroid 9 1.85E-05 FOSB, IL6, PTGS2, CSN1S1, KLF9, FOSL1, ZFP36, CDKN1A, DDIT4
GO:0048545 Response to steroid hormone 12 3.75E-05 FOSB, NR4A1, IL6, PTGS2, CSN1S1, NR4A2, KLF9, KDM5D, FOSL1, ZFP36, CDKN1A, DDIT4
GO:0032496 Response to lipopolysaccharide 11 4.60E-05 CXCL3, CXCL8, CXCL2, IL6, PTGS2, CX3CR1, TREM2, TAC1, ZFP36, JUN, SELE
CC GO:0042571 Immunoglobulin complex, circulating 4 2.90E-02 IGLC1, IGHM, IGKC, IGHG1
GO:0019814 Immunoglobulin complex 5 3.24E-02 IGLV1-44, IGLC1, IGHM, IGKC, IGHG1
MF GO:0005125 Cytokine activity 8 1.17E-03 CXCL3, CXCL8, CXCL2, IL6, CCL20, TNFSF11, NAMPT, TNFRSF11B
GO:0008009 Chemokine activity 4 5.48E-03 CXCL3, CXCL8, CXCL2, CCL20
GO:0003823 Antigen binding 6 5.48E-03 IGLV1-44, IGLC1, IGHM, IGKC, SLC7A5, IGHG1
GO:0001228 DNA-binding transcription activator activity, RNA polymerase II-specific 9 6.84E-03 FOSB, MAFF, NR4A1, ATF3, FOSL2, NR4A2, MYC, FOSL1, JUN
GO:0042379 Chemokine receptor binding 4 9.18E-03 CXCL3, CXCL8, CXCL2, CCL20
KEGG hsa04657 IL-17 signalling pathway 10 1.38E-08 FOSB, CXCL3, CXCL8, CXCL2, IL6, MMP1, PTGS2, CCL20, FOSL1, JUN
hsa05323 Rheumatoid arthritis 8 3.55E-06 CXCL3, CXCL8, CXCL2, IL6, MMP1, CCL20, TNFSF11, JUN
hsa04668 TNF signalling pathway 8 1.02E-05 CXCL3, CXCL2, IL6, PTGS2, CCL20, SOCS3, JUN, SELE
hsa04380 Osteoclast differentiation 8 2.15E-05 FOSB, TNFSF11, TREM2, FOSL2, SOCS3, FOSL1, JUN, TNFRSF11B
hsa05167 Kaposi sarcoma-associated herpesvirus infection 9 3.70E-05 CXCL3, CXCL8, CXCL2, IL6, PTGS2, MYC, ZFP36, JUN, CDKN1A
hsa04061 Viral protein interaction with cytokine and cytokine receptor 6 5.20E-04 CXCL3, CXCL8, CXCL2, IL6, CX3CR1, CCL20
hsa05146 Amoebiasis 6 5.20E-04 CXCL3, CXCL8, CXCL2, IL6, IL1R2, LAMA3
hsa04064 NF-kappa B signalling pathway 6 5.20E-04 CXCL3, CXCL8, CXCL2, PTGS2, GADD45B, TNFSF11
hsa05166 Human T-cell leukaemia virus 1 infection 8 5.25E-04 IL6, MYC, IL1R2, FOSL1, ZFP36, JUN, ADCY2, CDKN1A
hsa04060 Cytokine-cytokine receptor interaction 9 5.87E-04 CXCL3, CXCL8, CXCL2, IL6, CX3CR1, CCL20, TNFSF11, IL1R2, TNFRSF11B
  1. p < 0.05 was considered statistically significant.

  1. BP, biological process; CC, cellular component; DEGs, differentially expressed genes; IL-17, interleukin-17; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function; NF-kappa B, nuclear factor kappa beta; TNF, tumour necrosis factor.

PPI network construction, module analysis, and identification of the hub genes

The PPI network, which was composed of 48 nodes and 192 edges, was constructed by STRING and visualized by Cytoscape (Figure 3). MCODE plugin was used to do the module analysis, and two cluster modules were obtained according to the filter criteria (Supplementary Figure c). Cluster 1 had the highest score (score: 6, 9 nodes and 24 edges), followed by cluster 2 (score: 3.692, 14 nodes and 24 edges). Then, three algorithms including MCC, MNC, and degree plugins were used to identify the top eight hub genes. After verifying and intersecting these hub genes with the SVM-RFE algorithm, we finally obtained six hub genes which are the most important genes in the interaction network, and may play a pivotal role in the pathogenesis of OA, including CXCL2, FOSB, JUN, ATF3, IL6, and CXCL8. Six hub genes with detailed information are shown in Table III.

Fig. 3 
            Protein-protein interaction network of differentially expressed genes (DEGs). The interaction network between proteins coded by DEGs was constructed based on STRING database and Cytoscape software. ARC, activity regulated cytoskeleton associated protein; APOLD1, apolipoprotein L domain containing 1; ATF3, activating transcription factor 3; CCL20, C-C motif chemokine ligand 20; CDKN1A, cyclin dependent kinase inhibitor 1A; CSN1S1, casein alpha s1; CXCL8, C-X-C motif chemokine ligand 8; CXCL2, C-X-C motif chemokine ligand 2; CXCL3, C-X-C motif chemokine ligand 3; CX3CR1, C-X3-C motif chemokine receptor 1; CYR61, as known as CCN1, cellular communication network factor 1; DDIT4, DNA damage inducible transcript 4; DUOX2, dual oxidase 2; DUSP2, dual specificity phosphatase 2; FOSB, FosB proto-oncogene, AP-1 transcription factor subunit; FOSL1, FOS like 1, AP-1 transcription factor subunit; FOSL2, FOS like 2, AP-1 transcription factor subunit; FKBP5, FKBP prolyl isomerase 5; GADD45B, growth arrest and DNA damage inducible beta; HTR2B, 5-hydroxytryptamine receptor 2B; IL1R2, interleukin 1 receptor type 2; IL6, interleukin 6; JUN, Jun proto-oncogene, AP-1 transcription factor subunit; KLF9, KLF transcription factor 9; MAFF, MAF bZIP transcription factor F; MMP1, matrix metallopeptidase 1; MYC, MYC proto-oncogene, bHLH transcription factor; NAMPT, nicotinamide phosphoribosyltransferase; NFIL3, nuclear factor, interleukin 3 regulated; NR4A1, nuclear receptor subfamily 4 group A member 1; NR4A2, nuclear receptor subfamily 4 group A member 2; PTGS2, prostaglandin-endoperoxide synthase 2; RGS1, regulator of G protein signaling 1; SELE, selectin E; SIK1, salt inducible kinase 1; SLC16A7, solute carrier family 16 member 7; SLC18A2, solute carrier family 18 member A2; SLC2A3, solute carrier family 2 member 3; SLC7A5, solute carrier family 7 member 5; SOCS3, suppressor of cytokine signaling 3; TAC1, tachykinin precursor 1; TLR7, toll like receptor 7; TNFSF11, TNF superfamily member 11; TNFRSF11B, TNF receptor superfamily member 11b; TREM2, triggering receptor expressed on myeloid cells 2; WIF1, WNT inhibitory factor 1; WNT5B, Wnt family member 5B; ZFP36, ZFP36 ring finger protein.

Fig. 3

Protein-protein interaction network of differentially expressed genes (DEGs). The interaction network between proteins coded by DEGs was constructed based on STRING database and Cytoscape software. ARC, activity regulated cytoskeleton associated protein; APOLD1, apolipoprotein L domain containing 1; ATF3, activating transcription factor 3; CCL20, C-C motif chemokine ligand 20; CDKN1A, cyclin dependent kinase inhibitor 1A; CSN1S1, casein alpha s1; CXCL8, C-X-C motif chemokine ligand 8; CXCL2, C-X-C motif chemokine ligand 2; CXCL3, C-X-C motif chemokine ligand 3; CX3CR1, C-X3-C motif chemokine receptor 1; CYR61, as known as CCN1, cellular communication network factor 1; DDIT4, DNA damage inducible transcript 4; DUOX2, dual oxidase 2; DUSP2, dual specificity phosphatase 2; FOSB, FosB proto-oncogene, AP-1 transcription factor subunit; FOSL1, FOS like 1, AP-1 transcription factor subunit; FOSL2, FOS like 2, AP-1 transcription factor subunit; FKBP5, FKBP prolyl isomerase 5; GADD45B, growth arrest and DNA damage inducible beta; HTR2B, 5-hydroxytryptamine receptor 2B; IL1R2, interleukin 1 receptor type 2; IL6, interleukin 6; JUN, Jun proto-oncogene, AP-1 transcription factor subunit; KLF9, KLF transcription factor 9; MAFF, MAF bZIP transcription factor F; MMP1, matrix metallopeptidase 1; MYC, MYC proto-oncogene, bHLH transcription factor; NAMPT, nicotinamide phosphoribosyltransferase; NFIL3, nuclear factor, interleukin 3 regulated; NR4A1, nuclear receptor subfamily 4 group A member 1; NR4A2, nuclear receptor subfamily 4 group A member 2; PTGS2, prostaglandin-endoperoxide synthase 2; RGS1, regulator of G protein signaling 1; SELE, selectin E; SIK1, salt inducible kinase 1; SLC16A7, solute carrier family 16 member 7; SLC18A2, solute carrier family 18 member A2; SLC2A3, solute carrier family 2 member 3; SLC7A5, solute carrier family 7 member 5; SOCS3, suppressor of cytokine signaling 3; TAC1, tachykinin precursor 1; TLR7, toll like receptor 7; TNFSF11, TNF superfamily member 11; TNFRSF11B, TNF receptor superfamily member 11b; TREM2, triggering receptor expressed on myeloid cells 2; WIF1, WNT inhibitory factor 1; WNT5B, Wnt family member 5B; ZFP36, ZFP36 ring finger protein.

Table III.

Six hub genes identified by three algorithms of cytoHubba and support vector machine-recursive feature elimination.

Gene symbol Description Fold-change Adjusted p-value Regulation
CXCL2 C-X-C motif chemokine ligand 2 -2.499 3.09E-05 Down
FOSB FosB proto-oncogene, AP-1 transcription factor subunit -3.581 2.42E-04 Down
JUN Jun proto-oncogene, AP-1 transcription factor subunit -1.696 3.67E-06 Down
ATF3 Activating transcription factor 3 -2.189 6.34E-06 Down
IL-6 Interleukin-6 -2.256 1.06E-03 Down
CXCL8 C-X-C motif chemokine ligand 8 -2.561 5.75E-04 Down
  1. SVM-RFE, support vector machine recursive feature elimination.

Verification of six hub genes from the GEO dataset GSE12021

GSE12021, which included nine normal synovial samples and ten OA synovial samples, was selected to verify the six hub genes. The results are shown in Figure 4a, which indicate that the messenger RNA (mRNA) expression levels of the six hub genes in OA samples were significantly decreased compared with those in the normal samples (p < 0.05).

Fig. 4 
            Verification of the six hub genes by two datasets of the Gene Ontology database. a) Verification by GSE12021 (GPL96). Compared with normal controls, all hub genes were downregulated in osteoarthritis (OA) with significance. b) Verification by GSE32317 (GPL570). Compared with early OA, activating transcription factor 3 (ATF3) was downregulated in end-stage OA with significance, while others had no significance. ***p < 0.001; *p < 0.05; ns, no significant difference. CXCL2, C-X-C motif ligand 2; FOSB, FosB proto-oncogene, AP-1 transcription factor subunit; IL6, interleukin-6; JUN, Jun proto-oncogene; AP-1 transcription factor subunit.

Fig. 4

Verification of the six hub genes by two datasets of the Gene Ontology database. a) Verification by GSE12021 (GPL96). Compared with normal controls, all hub genes were downregulated in osteoarthritis (OA) with significance. b) Verification by GSE32317 (GPL570). Compared with early OA, activating transcription factor 3 (ATF3) was downregulated in end-stage OA with significance, while others had no significance. ***p < 0.001; *p < 0.05; ns, no significant difference. CXCL2, C-X-C motif ligand 2; FOSB, FosB proto-oncogene, AP-1 transcription factor subunit; IL6, interleukin-6; JUN, Jun proto-oncogene; AP-1 transcription factor subunit.

Identification of the effective diagnostic markers of early OA

GSE32317, which included ten early-stage OA and nine end-stage OA samples, was used to identify the effective diagnostic markers of early OA. We observed that the mRNA expression level of ATF3 was significantly decreased in the end-stage OA samples compared with the early ones (p < 0.001, Figure 4b).

Receiver operating characteristic curve of six hub genes

We created the receiver operating characteristic (ROC) curves using the hub genes expression profiles of normal and OA samples in GSE12021, as well as the ROC curves of hub genes based on expression profiles of early-stage OA and end-stage OA samples in GSE32317 (Supplementary Figure d). Among the six hub genes, JUN has the highest diagnostic value (area under the curve (AUC) 0.988) in the OA samples, while ATF3 has the highest diagnostic value (AUC 0.975) in the end-stage OA samples. In order to identify better diagnostic markers, we combined the results of ROC curves with their expression levels; compared with end-stage OA, ATF3 was significantly upregulated in early-stage OA (p < 0.001). Therefore, ATF3 may have a high diagnostic value for early diagnosis of OA and could be a potential therapeutic target.

Immune cell infiltration analyses results

The landscape of immune cell infiltration in different stages of OA has not been fully elucidated. We first investigated the difference in immune cell infiltration between early-stage OA and end-stage OA tissues. The correlation heatmap revealed that M2 macrophages had a significant negative correlation with memory B cells (p = 0.012), plasma cells (p = 0.017), follicular helper T cells (p = 0.038), and activated NK cells (p = 0.009). Regulatory T cells (Tregs) had a significant negative correlation with follicular helper T cells (p = 0.014) and activated NK cells had a significant negative correlation with resting NK cells (p < 0.001), while resting mast cells also had a significant negative correlation with monocytes (p = 0.011). The plasma cells had a significant positive correlation with memory B cells (p = 0.047). M0 Macrophages had a significant positive correlation with activated NK cells (p = 0.019), while gamma delta T cells had a significant positive correlation with activated dendritic cells and resting mast cells (p = 0.003, Supplementary Figure e). The violin plot showed that several infiltrating immune cells varied significantly between early-stage OA and end-stage OA. Compared with early-stage OA tissue, end-stage OA tissue contained a lower proportion of resting NK cells (p = 0.016) and resting dendritic cells (p = 0.043), while plasma cells (p = 0.043) infiltrated more (Figure 5).

Fig. 5 
            Evaluation and visualization of the landscape of immune infiltration between early-stage osteoarthritis (OA) and end-stage OA. The difference of immune cell infiltration between early-stage OA and end-stage OA. The red marks represent the significant difference in infiltration between the two groups of samples. p < 0.05 was considered statistically significant. NK cells, natural killer cells. ‘Fraction’ refers to the proportion of each immune cell.

Fig. 5

Evaluation and visualization of the landscape of immune infiltration between early-stage osteoarthritis (OA) and end-stage OA. The difference of immune cell infiltration between early-stage OA and end-stage OA. The red marks represent the significant difference in infiltration between the two groups of samples. p < 0.05 was considered statistically significant. NK cells, natural killer cells. ‘Fraction’ refers to the proportion of each immune cell.

Correlation analysis between ATF3 and immune infiltrating cells

Correlation analysis (Figure 6) revealed that ATF3 was positively correlated with resting NK cells (p = 0.034, |Cor| = 0.488), resting dendritic cells (p = 0.026, |Cor| = 0.510), and regulatory T cells (Tregs, p = 0.018, |Cor| = 0.535).

Fig. 6 
            Correlation between activating transcription factor 3 (ATF3) and immune infiltrating cells. The size of the dots represents the strength of the correlation between ATF3 and immune cells: the larger the dots, the stronger the correlation. The colour of the dots represents the p-value: the brighter shade of blue indicates a lower p-value. p < 0.05 was considered statistically significant.

Fig. 6

Correlation between activating transcription factor 3 (ATF3) and immune infiltrating cells. The size of the dots represents the strength of the correlation between ATF3 and immune cells: the larger the dots, the stronger the correlation. The colour of the dots represents the p-value: the brighter shade of blue indicates a lower p-value. p < 0.05 was considered statistically significant.

Discussion

OA has become a common disease that brings substantial burdens not only to society but also to the healthcare system. However, because of the lack of early diagnostic biomarkers, once OA is diagnosed, it is always at the end phase of the disease, resulting in a poor prognosis. Moreover, in previous studies, cartilage was considered as the central component bearing the full brunt of the OA. Recently, more evidence indicated that OA is a whole-joint disease affecting entire articular tissues including cartilage, synovium, and subchondral bone.25 Furthermore, during the development of OA, synovitis is always concomitant with OA from the early stage to the end stage.26 It is reported that synovitis can lead to secondary OA by initiating cartilage degeneration and bone reconstruction.27 In addition, researches also reveal that immune cell infiltration plays a vital role in the pathogenesis of OA.12,14 Therefore, it is of great significance and urgency to clarify the underlying mechanism of OA and identify effective biomarkers for the early diagnosis of OA. In this study, we aimed to identify the effective markers for the early diagnosis of OA and explore the role of immune cell infiltration in OA.

Through functional analyses, the results indicate that inflammatory response, immune response, and synovial membrane signal transduction play an important role in OA, resulting in arthritis and pain, which are known as the main clinical manifestations of OA.28 Additionally, the immune response may be related to infiltration of the immune cells, which needs to be further explored.

In this study, we found that ATF3 was significantly upregulated in early-stage OA with a high diagnostic value compared with end-stage OA. Therefore, we hypothesize that ATF3 may be an effective biomarker for the early diagnosis of OA based on our analysis. ATF3 is a member of the ATF/cyclic AMP response element-binding (CREB) protein family of transcription factors.29 A couple of studies have indicated that ATF3 may play an important role in regulating the cell biological processes of the joint. Chan et al30 reported that ATF3 can directly affect the transcription of matrix metalloproteinase (MMP)13. However, the increased expression of MMP13 would lead to the degradation of type II collagen, which could result in the loss of the extracellular matrix (ECM) and the imbalance of the cartilage homeostasis. Furthermore, IL-1β, which is considered to be related to the catabolic effects including growth inhibition and apoptosis induction of the chondrocyte, is reported to be mediated by ATF3.31 All of these effects on cartilage would lead to bone remodelling in OA. Hence, we have reasons to believe that ATF3 is of great significance in maintaining the cells and joint function, as well as in the course of OA development. In addition, one study showed that the JNK/SAPK pathway is involved in the induction of ATF3, which indicated its potential role in stress responses.32 Another study reported that ATF3 functions as a ‘hub’ of the cellular adaptive-response network, especially in the role of ATF3 in modulating inflammatory responses.33 Moreover, a study showed that ATF3 is related to the regulation of cell growth, apoptosis, invasion, and collagen synthesis in keloid fibroblast through the transforming growth factor β (TGF-beta)/SMAD signalling pathway.34,35 The roles of ATF3 in immunity, which may be related to Toll-like receptors, have also been reported.36 In addition, during the literature search, we noticed that there may be some interactions between ATF3 and NF-kappa B1,37 while manifestations of inflammation of synovial tissues in early-stage OA were associated with increased expression of NF-kappa B1.38 Jiang et al39 and Li et al40 also found that ATF3 may be of great help to diagnose early-stage OA, and can be an effective therapeutic target. Therefore, we believe that ATF3 is likely to be involved in the pathogenic, immunological, and inflamed processes of OA. Furthermore, ATF3 is associated with some classic inflammation-related signalling pathways which play important roles in the progression of OA. However, more studies are still needed to fully reveal the roles of ATF3 in osteoarthritic synovial tissues.

As for immune cell infiltration in the development of OA, we found that plasma cells infiltrated significantly more while resting NK cells and resting dendritic cells infiltrated significantly less in end-stage OA, which may be related to OA progression. Previous studies have shown that OA patients often exhibit inflammatory infiltration of synovial membranes by plasma cells, NK cells, and dendritic cells.41,42 One study revealed that biopsy samples from intensely inflamed synovium contained plasma cells.43 Another study confirmed that NK cells play an important role in OA.44 Further, a significant increase of dendritic cells has been observed in the synovial fluid of OA patients.45 The literature above, combined with our analysis, indicates that the infiltration of plasma cells, NK cells, and dendritic cells may play crucial roles in OA development. However, further experimental evidence is needed to reveal the underlying mechanism. Our analysis also discovered some details of immune cell infiltration of OA, such as the finding that activated NK cells are significantly related to resting NK cells and M2 macrophages, while activated dendritic cells are significantly related to gamma delta T cells. As for the results of the correlation between ATF3 and immune cells, we hypothesize that ATF3 raises the number of NK cells, dendritic cells, and regulatory T cells, which are related to the occurrence and progression of OA. Our study found that ATF3 might correlate with immune cell infiltration and change the course of the development of the disease eventually, however all of these results require validation by cell and animal experiments; this can be the focus of further studies.

Despite some of the novel data-analyzing methods like SVM-RFE and CIBERSORT used, there are some limitations to our research. First of all, the sample size for analysis and verification is relatively small, which could result in the deviation of the hub genes and CIBERSORT analysis. In addition, it is reported that age and sex are considered to relate to OA prevalence.46 However, some of the demographic information such as sex and age was missing from the selected datasets. The possible differences in the age and the sex distribution of the normal compared to OA samples might lead to unreliability in the results. Lastly, our analysis is based on previously published datasets. Although some previous studies are consistent with our results, the credibility of our results needs to be validated in further experiments.

To summarize, in our study we found that ATF3 may be a potential biomarker for the early diagnosis and treatment of OA. We also found that the differences in immune cell infiltration may be related to the differentially expressed ATF3 between early-stage OA and end-stage OA synovial tissue samples. Our study provides a theoretical basis and research direction for the early diagnosis, monitoring, and potential therapeutic intervention of OA.


Yu Fan. E-mail:

References

1. Murphy CA , Garg AK , Silva-Correia J , Reis RL , Oliveira JM , Collins MN . The meniscus in normal and osteoarthritic tissues: facing the structure property challenges and current treatment trends . Annu Rev Biomed Eng . 2019 ; 21 : 495 521 . Crossref PubMed Google Scholar

2. Sanghani-Kerai A , Black C , Cheng SO , et al. Clinical outcomes following intra-articular injection of autologous adipose-derived mesenchymal stem cells for the treatment of osteoarthritis in dogs characterized by weight-bearing asymmetry . Bone Joint Res . 2021 ; 10 ( 10 ): 650 658 . Crossref PubMed Google Scholar

3. He Z , Nie P , Lu J , et al. Less mechanical loading attenuates osteoarthritis by reducing cartilage degeneration, subchondral bone remodelling, secondary inflammation, and activation of NLRP3 inflammasome . Bone Joint Res . 2020 ; 9 ( 10 ): 731 741 . Crossref PubMed Google Scholar

4. Hunter DJ , Bierma-Zeinstra S . Osteoarthritis . Lancet . 2019 ; 393 ( 10182 ): 1745 1759 . Crossref PubMed Google Scholar

5. Chu CR , Williams AA , Coyle CH , Bowers ME . Early diagnosis to enable early treatment of pre-osteoarthritis . Arthritis Res Ther . 2012 ; 14 ( 3 ): 212 . Crossref PubMed Google Scholar

6. Dieppe PA , Lohmander LS . Pathogenesis and management of pain in osteoarthritis . Lancet . 2005 ; 365 ( 9463 ): 965 973 . Crossref PubMed Google Scholar

7. Jayadev C , Hulley P , Swales C , et al. Synovial fluid fingerprinting in end-stage knee osteoarthritis: a novel biomarker concept . Bone Joint Res . 2020 ; 9 ( 9 ): 623 632 . Crossref PubMed Google Scholar

8. Mathiessen A , Conaghan PG . Synovitis in osteoarthritis: current understanding with therapeutic implications . Arthritis Res Ther . 2017 ; 19 ( 1 ): 18 . Crossref PubMed Google Scholar

9. Guermazi A , Roemer FW , Hayashi D , et al. Assessment of synovitis with contrast-enhanced MRI using a whole-joint semiquantitative scoring system in people with, or at high risk of, knee osteoarthritis: the MOST study . Ann Rheum Dis . 2011 ; 70 ( 5 ): 805 811 . Crossref PubMed Google Scholar

10. Roemer FW , Guermazi A , Felson DT , et al. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study . Ann Rheum Dis . 2011 ; 70 ( 10 ): 1804 1809 . Crossref PubMed Google Scholar

11. Sellam J , Berenbaum F . The role of synovitis in pathophysiology and clinical symptoms of osteoarthritis . Nat Rev Rheumatol . 2010 ; 6 ( 11 ): 625 635 . Crossref PubMed Google Scholar

12. Ponchel F , Burska AN , Hensor EMA , et al. Changes in peripheral blood immune cell composition in osteoarthritis . Osteoarthr Cartil . 2015 ; 23 ( 11 ): 1870 1878 . Crossref Google Scholar

13. de Lange-Brokaar BJE , Ioan-Facsinay A , van Osch GJVM , et al. Synovial inflammation, immune cells and their cytokines in osteoarthritis: a review . Osteoarthr Cartil . 2012 ; 20 ( 12 ): 1484 1499 . Crossref PubMed Google Scholar

14. Rosshirt N , Hagmann S , Tripel E , et al. A predominant Th1 polarization is present in synovial fluid of end-stage osteoarthritic knee joints: analysis of peripheral blood, synovial fluid and synovial membrane . Clin Exp Immunol . 2019 ; 195 ( 3 ): 395 406 . Crossref PubMed Google Scholar

15. Davis S , Meltzer PS . GEOquery: a bridge between the Gene Expression Omnibus (GEO) and BioConductor . Bioinformatics . 2007 ; 23 ( 14 ): 1846 1847 . Crossref PubMed Google Scholar

16. Parker HS , Leek JT , Favorov AV , et al. Preserving biological heterogeneity with a permuted surrogate variable analysis for genomics batch correction . Bioinformatics . 2014 ; 30 ( 19 ): 2757 2763 . Crossref PubMed Google Scholar

17. Ritchie ME , Phipson B , Wu D , et al. limma powers differential expression analyses for RNA-sequencing and microarray studies . Nucleic Acids Res . 2015 ; 43 ( 7 ): e47 . Crossref PubMed Google Scholar

18. Gu Z , Eils R , Schlesner M . Complex heatmaps reveal patterns and correlations in multidimensional genomic data . Bioinformatics . 2016 ; 32 ( 18 ): 2847 2849 . Crossref PubMed Google Scholar

19. Ginestet C . ggplot2: elegant graphics for data analysis . Journal of the Royal Statistical Society . 2011 ; 174 ( 1 ): 245 246 . Crossref Google Scholar

20. Yu G , Wang LG , Han Y , He QY . clusterProfiler: an R package for comparing biological themes among gene clusters . OMICS . 2012 ; 16 ( 5 ): 284 287 . Crossref PubMed Google Scholar

21. Walter W , Sánchez-Cabo F , Ricote M . GOplot: an R package for visually combining expression data with functional analysis . Bioinformatics . 2015 ; 31 ( 17 ): 2912 2914 . Crossref PubMed Google Scholar

22. No authors listed . STRING Database . STRING Consortium . 2022 . https://string-db.org/cgi/about ( date last accessed 8 August 2022 ). Google Scholar

23. Chin CH , Chen SH , Wu HH , Ho CW , Ko MT , Lin CY . cytoHubba: identifying hub objects and sub-networks from complex interactome . BMC Syst Biol . 2014 ; 8 Suppl 4 : S11 . Crossref PubMed Google Scholar

24. Suykens JAK , Vandewalle J . Least squares support vector machine classifiers . Neural Processing Letters . 1999 ; 9 ( 3 ): 293 300 . Crossref Google Scholar

25. Scanzello CR , Goldring SR . The role of synovitis in osteoarthritis pathogenesis . Bone . 2012 ; 51 ( 2 ): 249 257 . Crossref PubMed Google Scholar

26. Wang H , Wang Q , Yang M , et al. Histomorphology and innate immunity during the progression of osteoarthritis: Does synovitis affect cartilage degradation? J Cell Physiol . 2018 ; 233 ( 2 ): 1342 1358 . Crossref PubMed Google Scholar

27. Ayral X , Pickering EH , Woodworth TG , Mackillop N , Dougados M . Synovitis: a potential predictive factor of structural progression of medial tibiofemoral knee osteoarthritis -- results of a 1 year longitudinal arthroscopic study in 422 patients . Osteoarthr Cartil . 2005 ; 13 ( 5 ): 361 367 . Crossref PubMed Google Scholar

28. Haviv B , Bronak S , Thein R . The complexity of pain around the knee in patients with osteoarthritis . Isr Med Assoc J . 2013 ; 15 ( 4 ): 178 181 . PubMed Google Scholar

29. Hai T . The ATF Transcription Factors in Cellular Adaptive Responses. InMaJ. ed. Gene Expression and Regulation. New York, NY: Springer, 2006: 329340. Google Scholar

30. Chan CM , Macdonald CD , Litherland GJ , et al. Cytokine-induced MMP13 expression in human chondrocytes is dependent on activating Transcription Factor 3 (ATF3) regulation . J Biol Chem . 2017 ; 292 ( 5 ): 1625 1636 . Crossref PubMed Google Scholar

31. Li X , Li Y , Yang X , et al. PR11-364P22.2/ATF3 protein interaction mediates IL-1β-induced catabolic effects in cartilage tissue and chondrocytes . J Cell Mol Med . 2021 ; 25 ( 13 ): 6188 6202 . Crossref PubMed Google Scholar

32. Hai T , Wolfgang CD , Marsee DK , Allen AE , Sivaprasad U . ATF3 and stress responses . Gene Expr . 1999 ; 7 ( 4–6 ): 321 335 . PubMed Google Scholar

33. Hai T , Wolford CC , Chang YS . ATF3, a hub of the cellular adaptive-response network, in the pathogenesis of diseases: is modulation of inflammation a unifying component? Gene Expr . 2010 ; 15 ( 1 ): 1 11 . Crossref PubMed Google Scholar

34. Wang XM , Liu XM , Wang Y , Chen ZY . Activating transcription factor 3 (ATF3) regulates cell growth, apoptosis, invasion and collagen synthesis in keloid fibroblast through transforming growth factor beta (TGF-beta)/SMAD signaling pathway . Bioengineered . 2021 ; 12 ( 1 ): 117 126 . Crossref PubMed Google Scholar

35. Duan M , Wang Q , Liu Y , Xie J . The role of TGF-β2 in cartilage development and diseases . Bone Joint Res . 2021 ; 10 ( 8 ): 474 487 . Crossref PubMed Google Scholar

36. Thompson MR , Xu D , Williams BRG . ATF3 transcription factor and its emerging roles in immunity and cancer . J Mol Med (Berl) . 2009 ; 87 ( 11 ): 1053 1060 . Crossref PubMed Google Scholar

37. Zheng S , Abraham C . NF-κB1 inhibits NOD2-induced cytokine secretion through ATF3-dependent mechanisms . Mol Cell Biol . 2013 ; 33 ( 24 ): 4857 4871 . Crossref PubMed Google Scholar

38. Benito MJ , Veale DJ , FitzGerald O , van den Berg WB , Bresnihan B . Synovial tissue inflammation in early and late osteoarthritis . Ann Rheum Dis . 2005 ; 64 ( 9 ): 1263 1267 . Crossref PubMed Google Scholar

39. Jiang A , Xu P , Zhao Z , et al. Identification of candidate genetic markers and a novel 4-genes diagnostic model in osteoarthritis through integrating multiple microarray data . Comb Chem High Throughput Screen . 2020 ; 23 ( 8 ): 805 813 . Crossref PubMed Google Scholar

40. Li H , Yang HH , Sun ZG , Tang HB , Min JK . Whole-transcriptome sequencing of knee joint cartilage from osteoarthritis patients . Bone Joint Res . 2019 ; 8 ( 7 ): 290 303 . Crossref PubMed Google Scholar

41. Li YS , Luo W , Zhu SA , Lei GH . T cells in osteoarthritis: alterations and beyond . Front Immunol . 2017 ; 8 : 356 . Crossref PubMed Google Scholar

42. Pettit AR , Ahern MJ , Zehntner S , Smith MD , Thomas R . Comparison of differentiated dendritic cell infiltration of autoimmune and osteoarthritis synovial tissue . Arthritis Rheum . 2001 ; 44 ( 1 ): 105 110 . Crossref PubMed Google Scholar

43. Lindblad S , Hedfors E . Arthroscopic and immunohistologic characterization of knee joint synovitis in osteoarthritis . Arthritis Rheum . 1987 ; 30 ( 10 ): 1081 1088 . Crossref PubMed Google Scholar

44. Benigni G , Dimitrova P , Antonangeli F , et al. CXCR3/CXCL10 axis regulates neutrophil-NK cell cross-talk determining the severity of experimental osteoarthritis . J Immunol . 2017 ; 198 ( 5 ): 2115 2124 . Crossref PubMed Google Scholar

45. Alahdal M , Zhang H , Huang R , et al. Potential efficacy of dendritic cell immunomodulation in the treatment of osteoarthritis . Rheumatology (Oxford) . 2021 ; 60 ( 2 ): 507 517 . Crossref PubMed Google Scholar

46. Liu Z , Wang H , Wang S , Gao J , Niu L . PARP-1 inhibition attenuates the inflammatory response in the cartilage of a rat model of osteoarthritis . Bone Joint Res . 2021 ; 10 ( 7 ): 401 410 . Crossref PubMed Google Scholar

Author contributions

J. Yang: Software, Visualization, Formal analysis, Methodology, Writing – original draft.

Y. Fan: Conceptualization, Supervision, Methodology, Writing – review & editing.

S. Liu: Validation, Visualization, Writing – review & editing.

Funding statement

The authors received no financial or material support for the research, authorship, and/or publication of this article.

Acknowledgements

This study was the re-analysis based on the published data from GEO database. We would like to thank GEO database for the sharing of data. We also thank www.xiantao.love for the figure technology support.

Open access funding

The open access funding for this manuscript was self-funded.

Supplementary material

Figures showing more details about the osteogenesis of osteoarthritis (OA) and the role that activating transcription factor 3 plays in the development of OA, as well as its correlation with immune infiltration.

© 2022 Author(s) et al. This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial No Derivatives (CC BY-NC-ND 4.0) licence, which permits the copying and redistribution of the work only, and provided the original author and source are credited. See https://creativecommons.org/licenses/by-nc-nd/4.0/