Identification and validation of hub genes of synovial tissue for patients with osteoarthritis and rheumatoid arthritis

doi:10.21203/rs.3.rs-564869/v1

Download PDF

Research Article

Identification and validation of hub genes of synovial tissue for patients with osteoarthritis and rheumatoid arthritis

https://doi.org/10.21203/rs.3.rs-564869/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Osteoarthritis (OA) and rheumatoid arthritis (RA) were two major joint diseases with partially common phenotypes and genotypes. This study aimed to determine the mechanistic similarities and differences between osteoarthritis and rheumatoid arthritis by analyzing the differentially expressed genes and signaling pathways. Microarray data of osteoarthritis and rheumatoid arthritis were obtained from the Gene Expression Omnibus. By integrating multiple gene data sets, specific differentially expressed genes (DEGs) were identified in synovial membrane samples from patients and healthy donations. Then, the Gene ontology significant functions annotation, Kyoto Encyclopedia of Genes and Genomes pathways and protein-protein interaction network analysis were conducted. Moreover, CIBERSORT was used to further distinguish OA and RA in immune infiltration. Finally, animal experimentation was conducted and the establishment of model, which was verified using PCR in the mouse. As an overlapping process, we identified 1116 DEGs between OA and RA. It was indicated that specific gene signatures differed significantly between OA and RA connected with the distinct pathways. Of identified DEGs, 9 immune cell types among 22 were identified to distinguish from each other. The qRT-PCR result showed that the eight-tenths expression levels of the hub genes were significantly increased in OA samples (P < 0.05). This large-scale gene expression study provided new insights for disease-associated genes and molecular mechanisms as well as their associated function in osteoarthritis and rheumatoid arthritis, which simultaneously offer a new direction for biomarker development and the distinguishment of gene-level mechanisms between osteoarthritis and rheumatoid arthritis.

Translational Medicine

Hepatobiliary & Transplant Surgery

osteoarthritis

rheumatoid

differentially expressed gene

bioinformatics analysis

Osteoarthritis (OA) is one of the most common arthritic diseases in elders, involving multiple joints with declining joint functions. As the main cause of disability disease, OA has gradually increased the source of societal costs in older adults [1]. Rheumatoid arthritis (RA) is an autoimmune disease relating to multifarious environmental as well as genetic factors, characterized by persistent synovitis, systemic inflammation, and the presence of autoantibodies [2]. RA typically involves small joints of the hands and feet, although the larger joints are also frequently affected [3]. Similar to OA, RA reduces productivity and increases disability in affected patients, and adds financial burden on these patients as well as society. Both OA and RA are related to intra-articular releasing of multiple cytokines and proteolytic enzymes, with symptoms including pain, swelling, and dysfunction around the joints [4]. Although these two diseases have many similarities of symptoms, there are still differences in pathological mechanisms and pathogenesis in terms of the synovial for both.

Despite the differences in the occurrence and development between OA and RA, the current therapies for these diseases are partial similarity, including oral medications (nonsteroidal anti-inflammatory drugs) [5, 6]; intra-articular injection (hormones and sodium hyaluronate), maintained functional capacities [7, 8]; and surgical intervention performed to rescue end-stage lesions [9, 10]. However, those interventions are not adequate to impede the occurrence and progression of OA and RA due to the unclear pathogenesis of both diseases. Evidence had demonstrated that degeneration and loss of articular cartilage, subchondral bone remodeling, and sclerosis, and inflammation of the synovium and synovia were the main pathomorphological changes of OA [11]. RA is characterized by an autoimmune-mediated attack of the involved joints synovial membranes, leading to a deterioration of cartilage and bone, and is one of the most common inflammatory arthritis among connective tissue disorders [12]. To distinguish OA from RA, a previous study reported that the positive rate and diagnostic accuracy can be enhanced by combining imaging procedures, such as the formation of bone spur and narrowing of joint space. Nevertheless, the pathogenesis and antidiastole of both OA and RA are still unclear so far. Besides, advanced RA progresses in the secondary are as same as OA, and then increase difficulty in the distinguishment.

In recent years, massive exploration had been made to probe the mechanisms of these two diseases at the level of genes and did a lot of work in bioinformatics analysis [13, 14]. Compared with individual gene expression data, integrated bioinformatics analysis of these profile data in multiple platforms will conducive to discover the potential mechanism of OA and RA with larger samples, more precisely. In this study, we performed functional annotation and constructed three protein-protein interaction (PPI) networks by data mining, to discover gene and molecular mechanisms underlying OA and RA and then further analyze the differences in mechanism between them. After that, immune infiltration was identified between OA and RA to discover the differences and relationships of 22 types immune cells. Finally, precisely overlapping genes of the DEGs in two diseases were detected and mice experimental were conducted using PCR to validate the hub genes.

Microarray data

The gene expression data was downloaded from the Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo) database, which is the biggest database of high throughput gene expression data in the world. Using the keywords “Rheumatoid arthritis” and “Osteoarthritis” to search the GEO database respectively, there were a total of 8046 results in the GEO database up to December 22, 2020. After that, Viewing these documents and then filtering through established inclusion criteria. The inclusion criteria of microarray data sets were as follows: (1) data sets should incorporate in messenger RNA transcriptome data; (2) samples were limited to synovium tissue.

Integration of microarray data and Screening DEGs

Probe ID extracted from the downloaded series matrix was converted into gene symbol using the Perl programming language (version 5.30.0). Aiming at the different comparisons, gene symbols were corresponding merged into three groups and saved in TXT files. Subsequently, the quantile normalization of multiple gene expression data sets was performed with the R software (version 3.6.1) to minimize the heterogeneity. Each gene was calculated by a t-test through its gene expression level and differentially expressed genes (DEGs) were screened employing “sva” and “limma” package in R software. The screening thresholds were log fold change (logFC) ༞ 1 and adjust-P value < 0.05. Hierarchical clustering analysis of the top 100 DEGs was conducted using “pheatmap” packages in R version 3.6.1.

Gene Ontology and pathway enrichment analysis

Biological function and DEG-associated biological pathways were discovered by Gene Ontology (GO) (http://geneontology.org/) [15] In this study, related enrichment analyses were enriched using R software in which they were involved. GO analysis included categories of 3 items, such as cellular component (CC), biological processes (BP), and molecular function (MF). Kyoto Encyclopedia of Genes and Genomes (KEGG) database (https://www.kegg.jp/) contains masses of pathway maps and their required encoding proteins, which do the favor of understanding high-level functions and utilities of the biological system. GO and KEGG-based pathway mapping of DEGs-associated were colored through “colorspace” and “stringi” packages. Besides, both GO and KEGG pathways were identified as significant under a correction at an FDR value cutoff of 0.05. For GO enrichment analysis, the other 3 packages (“enrichplot”, “DOSE” and “ggplot2”) were also used and the 3 GO terms included enriched biological processes (BP), cellular component (CC) and molecular function (MF) were also obtained.

Gene Set Enrichment Analysis (GSEA)

The GSEA [16] software was downloaded from http://software.broadinstitute.org/gsea/index.jsp. To further identify the different function of the two disease hub genes, the GSEA analysis was conducted with merged data between OA-specific DEGs and RA-specific DEGs. Terms enriched in all real hub genes with false discovery rate (FDR) < 0.05 and adjusted-P value < 0.05 were set as the cut-off criteria.

PPI network analysis

In this study, we used the Search Tool for the Retrieval of Interacting Genes (STRING) database to build three gene-gene interaction networks [17, 18], which help us to discover the interactions of known and predicted proteins. Proteins included in the PPI networks were protein-encoding DEGs with a similar gene change in NC (Normal control) vs. OA, NC vs. RA, and OA vs. RA. In all three groups, the minimum required interaction score with a confidence score of 0.99 was applied to build the PPI networks. Using Cytoscape software (version 3.5), PPI networks of OA and RA-specific DEGs were constructed, respectively. Additionally, nodes represented proteins and edges indicated a connection between any two proteins. Moreover, the degree of each node was calculated and hub proteins were considered with a higher rank. The height of rectangles represents the connective degree with other genes.

Verification by ROC Analysis

We downloaded the gene expression data of GSE153015, including 4 OA patients and 10 RA patients. Using the “pROC” package [19] and evaluating the role of candidate genes, the expression values of the top overlapped 10 PPI hub genes in the network were used to acquire both receiver operating characteristic (ROC) curves and calculate area under curve (AUC) [20]. The AUC > 0.5, as well as adjusted-P value < 0.05, were considered as useful predictive values.

Immune infiltration

CIBERSORT [21] can be run online (http://cibersort.stanford.edu/) or downloaded for local use, and used as a method for estimating the cell proportion and cell types of multiple tissues. R software and Java implementations are essentially downloaded for the next data processing. In this study, we used CIBERSORT online tool to further explore the state of OA and RA synovial membrane with preestablished 22 types of immune cells. After that, the analysis result was filtered according to adjusted-P value < 0.05, and the immune cell composition of each sample was shown in the barplot. As a result, “corrplot” package [22] downloaded in R was used to visualize the correlation of 22 types of infiltrating immune cells (IICs); “ggplot2” script was run to draw violin diagrams to depict the differences in each immune cell types.

Animal Experiments

Male C57BL/6 (Grade SPF II) were provided by Shanghai Super B&K Laboratory Animal Co. Ltd. (Certificate number: SCXK (Shanghai) 2013-0016). The animal experiments were in accordance with the China legislation on the use and care of laboratory animals and approved by the Medical Norms and Ethics Committee of Zhejiang Chinese Medical University. C57BL/6 mice were randomly designated into 3 groups: destabilization of the medial meniscus (DMM) of OA-model, RA of RA-model and normal control (NC) group. In brief, after anesthetized with pentobarbital sodium (3%, 0.15ml/100g) intraperitoneally, the bilateral knee joint of mice (n = 10 per group; 8 weeks old; mean body weight = 25.3g) was exposed through a medial capsular incision. Then, the medial meniscotibial ligament was transected, and then the medial meniscus was visible displaced medially [23]. Finally, washed with 20ml saline and the incision was sutured. Two days after surgery, running wheels were put in the cage to encourage exercise. Other 10 mice (8 weeks old; mean body weight = 25.6g) were randomly chosen to inject Complete Freund’s Adjuvant [24] (CFA). Sham operation was done in parallel, and only the skin of the knee joints were resected. Mice were sacrificed at 8 weeks after surgery.

Evaluation of arthritis severity and tactile sensitivity testing

In RA group of 10 mice, for each of these limbs, a subjective scoring system [25] is applied to both lower limbs with the lowest grade of 0 and the highest grade of 4 (Table 1). Accordingly, the subjectivity of this score including ankylosis of the limb, arguably the severest form of arthritis, can be a concern for quantitative analysis model projects of RA.

Table 1

Scoring system for subjective evaluation of arthritis severity.
Severity	Degree of inflammation
0	No evidence of erythema and swelling
1	Erythema and mild swelling confined to the tarsals or ankle joint
2	Erythema and mild swelling extending from the ankle to the tarsals
3	Erythema and moderate swelling extending from the ankle to metatarsal joints
4	Erythema and severe swelling encompass the ankle, foot and digits, or ankylosis of the limb

In OA model group, 10 mice were acclimated for 30 minutes in closed chambers to von Frey testing (UGO, USA). The surface of bilateral hind paws were stimulated with ascending force to determine tactile sensitivity. Afterwards, a rapid withdrawal of the tested paw was defined as a positive reaction, and the number of positive responses was recorded digitally [26].

Real-time PCR (qPCR) assay

The 30 synovial samples were collected from mice after killed in a CO2 chamber. Total RNA was extracted with TRIzol reagent and quality controlled by NanoDrop2000 spectrophotometer (Thermo Scientific, USA) and agarose gel electrophoresis. Then, the reverse transcription was conducted to produce cDNA. The final total qPCR reaction system was 20µl, including 10µl SYBR® Premix Ex Taq II (Tli RnaseH Plus), 0.4µl PCR Forward Primer, 0.4µl PCR Reverse Primer, 1 µl template cDNA and 8.2 µl ddH2O, and the qPCR reaction conditions were as follows: 95°C for 5 min for initial denaturation, followed by 40 cycles of denaturation at 95°C for 10 sec, annealing and extension at 60°C for half a minute [27]. β-actin was used as an internal reference. Relative mRNA expression was calculated using the 2-ΔΔCt method (Table 2). One-way analysis of variance was used for the statistical analysis, and P < 0.05 indicated a significant difference.

Table 2

Primer sequences of target genes.
Gene	Forward primer	Reverse primer
β-actin	5′-AGGGGCCGGACTCGTCATACT-3′	5′-GGCGGCACCACCATGTACCCT-3′
RPS6	5′-AGCTCCGCACCTTCTATGAGA-3′	5′-GGGAAAACCTTGCTTGTCATTC-3′
RPS14	5′-TGCCACATCTTTGCATCCTTC-3′	5′-ACTCATCTCGGTCAGCCTTCA-3′
RPS25	5′-CCCAGTAAATAAATCTGGTGGCA-3′	5′-CGGAACCTCCTTACAGAGCTT-3′
RPL11	5′-ATGGCGCAAGATCAAGGGG-3′	5′-GACTGTGCAGTGAACAGCAAT-3′
RPL27	5′-AAAGCCGTCATCGTGAAGAAC-3′	5′-GCTGTCACTTTCCGGGGATAG-3′
RPS29	5′-GTCTGATCCGCAAATACGGG-3′	5′-AGCCTATGTCCTTCGCGTACT-3′
SNRPE	5′-CAGGGCCAAAAGGTGCAGAA-3′	5′-ATTCACTTGTTCATACAGCCACA-3′
EEF2	5′-TGTCAGTCATCGCCCATGTG-3′	5′-CATCCTTGCGAGTGTCAGTGA-3′
RPL10A	5′-ATGAGCAGCAAAGTCTCACG-3′	5′-GGTCGTAGTTCTTCAGGCTGAT-3′
RPL19	5′-ATGAGTATGCTCAGGCTACAGA-3′	5′-GCATTGGCGATTTCATTGGTC-3′

Statistical analysis

β-actin was used as the reference gene and the relative mRNA expression data from 3 different groups were compared using one-way ANOVA followed by Fisher’s least significant difference (LSD) comparison. The mRNA data were processed by Statistical Product and Service Solutions (SPSS) 21.0 software. A p-value < 0.05 was considered a significant difference and p-value < 0.01 considered to indicate a very significant difference.

Basic datasets of OA and RA

In this study, we performed an integrative analysis of gene expression and the entire workflow was shown in Fig. 1. A total of 11 expression data sets of OA, RA, and NC were downloaded with the GEO accession number of GSE1919, GSE12021, GSE29746, GSE36700, GSE39340, GSE55235, GSE55457, GSE55584, GSE77298, and GSE82107. The latest GSE153015 was solely used on performing ROC analysis as well as hub genes verification. In brief, 265 persons were enrolled in this study and data were merged, among that, the synovial membrane of 88 patients with OA, 114 patients with RA, and 63 NC (Normal control). Figure 2 depicts the relevant details of the selection process. All relevant information of these eleven GEO datasets was showed in Table 3. Venn diagram of shared DEGs between OA, RA, and NC was shown in Fig. 3A. The volcano plot representing all up-regulated genes and down-regulated genes identified in NC vs. OA and NC vs. RA was shown in Fig. 3(B-C). 1577 DEGs (44 up-regulated and 1533 down-regulated DEGs) were identified between RA and OA. The volcano plot was shown in Fig. 3D.

Table 3

Summary information of studies included in analysis.
GEO accession	Author	Public date	Platform	OA: RA: NC
GSE1919	Ungethuem U	Nov 04, 2004	GPL91 [HG_U95A] Affymetrix Human Genome U95A Array	5:5:5
GSE12021 (GPL96)	Huber R	Sep 02, 2008	GPL96 [HG-U133A] Affymetrix Human Genome U133A Array	10:12:9
GSE12021 (GPL97)	Huber R	Sep 02, 2008	GPL97 [HG-U133B] Affymetrix Human Genome U133B Array	10:12:4
GSE29746	Del Rey MJ	Oct 25, 2011	GPL4133 Agilent-014850 Whole Human Genome Microarray 4x44K G4112F (Feature Number version)	11:9:11
GSE36700	Lauwerys BR	Mar 27, 2012	GPL570 [HG-U133_Plus_2] Affymetrix Human Genome U133 Plus 2.0 Array	5:7:0
GSE39340	Chang X	Oct 22, 2012	GPL10558 Illumina HumanHT-12 V4.0 expression beadchip	7:10:0
GSE55235	Woetzel D	Feb 21, 2014	GPL96 [HG-U133A] Affymetrix Human Genome U133A Array	10:10:10
GSE55457	Woetzel D	Mar 05, 2014	GPL96 [HG-U133A] Affymetrix Human Genome U134A Array	10:13:10
GSE55584	Woetzel D	Mar 05, 2014	GPL96 [HG-U133A] Affymetrix Human Genome U135A Array	6:10:0
GSE77298	Broeren MG	Jan 27, 2016	GPL570 [HG-U133_Plus_2] Affymetrix Human Genome U133 Plus 2.0 Array	0:16:7
GSE82107	de Vries M	Jun 02, 2016	GPL570 [HG-U133_Plus_2] Affymetrix Human Genome U133 Plus 2.0 Array	10:0:7
GSE153015	Triaille C	Jun 22, 2020	GPL570[HG-U133_Plus_2] Affymetrix Human Genome U133 Plus 2.0 Array	4:10:0
GEO: Gene Expression Omnibus; OA: osteoarthritis; RA: rheumatoid arthritis; NC: normal control.

DEGs screening

Compared with the healthy synovial membrane samples, a total of 1723 and 1460 DEGs were identified in OA and RA samples, respectively. From the results, we found that a total of 1690 genes were up-regulated in both NC and OA samples. While a total of 33 same down-regulated genes were identified in these two groups. The number of up-regulated DEGs was 1278 for NC versus OA and 183 for down-regulated. After overlapped all DEGs, we obtained 1116 DEGs that were differentially expressed in OA and RA compared with NC samples. Three heatmap diagrams of DEGs were represented in Fig. 4(A-C), visually showing that these similar DEGs could significantly distinguish from each other. The top 5 up- and down-regulated DEGs of the three groups were shown in Table 4. Next, the probe expression matrix files downloaded from the GEO database were normalized and the three sets of data were shown in Fig. 5. And then, we merged the expression matrix of 10 items into one, and the order of samples in dataset was rearranged.

Table 4

Top 5 up- and downregulated DEGs between OA, RA and NC.
	NC vs OA				NC vs RA				OA vs RA
Up-regulated	ID	Symbol	logFC	Adj.P.Val	ID	Symbol	logFC	Adj.P.Val	ID	Symbol	logFC	Adj.P.Val
	4613	MYCN	9.180080878	8.02E-28	8935	SKAP2	9.553082085	1.26E-35	23094	SIPA1L3	7.79814072	5.46E-25
	6598	SMARCB1	9.03478793	8.02E-28	10745	PHTF1	9.39043259	1.29E-34	1404	HAPLN1	7.87947808	3.69E-22
	23365	ARHGEF12	9.51514144	3.28E-27	22925	PLA2R1	7.801574317	1.29E-34	5074	PAWR	6.908165575	1.22E-21
	5137	PDE1C	9.117030454	4.96E-27	9488	PIGB	9.97554221	1.04E-33	10100	TSPAN2	5.2303578	4.83E-18
	8697	CDC23	9.841678124	6.56E-27	9294	S1PR2	9.470029664	1.04E-33	22999	RIMS1	4.725490721	6.50E-18
Down-regulated	5340	PLG	-5.068734955	2.10E-17	5208	PFKFB2	-7.906263354	5.11E-25	954	ENTPD2	-10.57701223	1.29E-43
Down-regulated	7368	UGT8	-7.56536003	7.07E-17	9024	BRSK2	-6.728489168	2.71E-24	4846	NOS3	-9.366652871	3.92E-41
	3575	IL7R	-6.966596378	3.51E-16	23308	ICOSLG	-7.439436875	8.41E-24	6875	TAF4B	-9.588028779	2.08E-40
	1404	HAPLN1	-8.500030865	2.05E-15	2676	GFRA3	-7.481477156	2.29E-23	2395	FXN	-9.989764717	6.81E-40
	22999	RIMS1	-5.348208851	4.24E-14	6875	TAF4B	-8.729849528	2.50E-23	777	CACNA1E	-9.95687779	7.39E-39
DEGs: differentially expressed genes; OA: osteoarthritis; RA: rheumatoid arthritis; NC: normal control; logFC: log fold change.

GO pathway of OA-specific and RA-specific DEGs

GO annotations analysis showed that OA-specific (both in NC vs OA and OA vs RA) DEGs (Table 5) were predominantly enriched in 3 items including BP, CC and MF. In the BP group, the DEGs were mainly enriched in regulation of neuron projection development, regulation of cell morphogenesis involved in differentiation, positive regulation of neuron differentiation, etc. In the CC group, the DEGs were mainly enriched in nuclear speck, RNA polymerase II transcription factor complex, adherens junction, etc. In the MF group, the DEGs were mainly enriched in protein serine/threonine kinase activity, proximal promoter sequence-specific DNA binding, RNA polymerase II proximal promoter sequence-specific DNA binding, etc. RA-specific (both in NC vs RA and OA vs RA) DEGs (Table 5) were primarily enriched in BP group of cell morphogenesis involved in neuron differentiation, regulation of cell morphogenesis involved in differentiation, positive regulation of neuron differentiation, etc. In the CC group, the DEGs were mainly enriched in nuclear speck, RNA polymerase II transcription factor complex, transcription factor complex, etc. In the MF group, the DEGs were mainly enriched in protein serine/threonine kinase activity, SMAD binding, proximal promoter sequence-specific DNA binding, etc. The GO distribution of OA and RA including up- and down-regulated DEGs were presented in Fig. 6A and 6B, respectively.

Table 5

GO pathway analysis of differentially expressed genes.
Term		ID	Description	Adj.P.Val	Count
OA-specific	BP	GO:0010975	regulation of neuron projection development	3.97E-10	92
		GO:0010769	regulation of cell morphogenesis involved in differentiation	3.97E-10	63
		GO:0045666	positive regulation of neuron differentiation	6.79E-09	74
	CC	GO:0016607	nuclear speck	4.51E-10	78
		GO:0090575	RNA polymerase II transcription factor complex	2.64E-09	37
		GO:0005912	adherens junction	2.12E-08	87
	MF	GO:0004674	protein serine/threonine kinase activity	6.24385E-10	80
		GO:0000987	proximal promoter sequence-specific DNA binding	6.24385E-10	91
		GO:0000978	RNA polymerase II proximal promoter sequence-specific DNA binding	3.26907E-09	86
RA-specific	BP	GO:0048667	cell morphogenesis involved in neuron differentiation	8.82E-10	87
		GO:0010769	regulation of cell morphogenesis involved in differentiation	1.52E-09	56
		GO:0045666	positive regulation of neuron differentiation	1.52E-09	68
	CC	GO:0016607	nuclear speck	1.53E-08	67
		GO:0090575	RNA polymerase II transcription factor complex	1.17E-06	30
		GO:0005667	transcription factor complex	1.15E-05	46
	MF	GO:0004674	protein serine/threonine kinase activity	2.69356E-11	75
		GO:0046332	SMAD binding	4.45873E-08	25
		GO:0000987	proximal promoter sequence-specific DNA binding	2.44254E-06	72
GO: Gene Ontology; OA: osteoarthritis; RA: rheumatoid arthritis.

KEGG pathway of OA-specific and RA-specific DEGs

As shown in Fig. 7 and Table 6, the signaling pathways of OA-specific (both in NC vs OA and OA vs RA) DEGs were mainly enriched in the EGFR tyrosine kinase inhibitor resistance, Ubiquitin mediated proteolysis, FoxO signaling pathway, etc. Significant results of the KEGG pathway enrichment of DEGs in RA were FoxO signaling pathway, TGF-beta signaling pathway, EGFR tyrosine kinase inhibitor resistance, etc.

Table 6

KEGG analysis of differentially expressed genes.
Term	ID	Description	Adj.P.Val	Count
OA-specific	hsa01521	EGFR tyrosine kinase inhibitor resistance	7.94E-06	27
	hsa04120	Ubiquitin mediated proteolysis	7.94E-06	38
	hsa04068	FoxO signaling pathway	1.88E-05	36
	hsa04012	ErbB signaling pathway	2.08E-05	27
	hsa05215	Prostate cancer	2.59E-05	29
	hsa04910	Insulin signaling pathway	9.18E-05	35
	hsa04350	TGF-beta signaling pathway	0.000109585	27
	hsa04931	Insulin resistance	0.000153499	29
	hsa05205	Proteoglycans in cancer	0.000153499	45
RA-specific	hsa04068	FoxO signaling pathway	2.39E-05	33
	hsa04350	TGF-beta signaling pathway	3.84E-05	26
	hsa01521	EGFR tyrosine kinase inhibitor resistance	3.84E-05	23
	hsa04012	ErbB signaling pathway	3.84E-05	24
	hsa04120	Ubiquitin mediated proteolysis	0.000112284	31
	hsa04510	Focal adhesion	0.000112284	40
	hsa04931	Insulin resistance	0.000204948	26
	hsa04151	PI3K-Akt signaling pathway	0.00022268	59
	hsa05215	Prostate cancer	0.00022268	24
KEGG: Kyoto Encyclopedia of Genes and Genomes; OA: osteoarthritis; RA: rheumatoid arthritis.

GSEA of involved signaling pathways

GSEA analysis unveiled that DEGs between OA and RA were significantly enriched in several pathways (Fig. 8). We found that the RA was significantly associated with the high expression of cell cycle (P = 0.0143), type II diabetes mellitus (P = 0.0126), oocyte meiosis (P = 0.0201), spliceosome (P = 0.0432), p53 signaling pathway (P = 0.0452). OA process, in contrast, had lower expression in the activities.

PPI network and modules

The OA-specific PPI network consisted of 391 nodes and 483 edges (Fig. 9A). The top five hub nodes were RPS6 (ribosomal protein S6, degree = 19), RPS14 (ribosomal protein S14, degree = 16), RPS19 (ribosomal protein S19, degree = 15), RPL11 (ribosomal protein L11, degree = 13) and RPS27 (ribosomal protein S27, degree = 13). The RA-specific PPI network consisted of 273 nodes and 259 edges (Fig. 9B). The top five hub nodes were CTNNB1 (catenin beta 1, degree = 8), SNRPA1 (small nuclear ribonucleoprotein polypeptide A', degree = 8), CDC42 (cell division cycle 42, degree = 7), PRPF8 (pre-mRNA processing factor 8, degree = 7), MAD2L1 (mitotic arrest deficient 2 like 1, degree = 6). PPI constructed by DEGs between OA synovium and RA synovium was shown in Fig. 9C. In the top 20 connections of these three groups, the overlap associated genes were shown in Fig. 9D, as follows.

Verification of Hub Genes by ROC Analysis

To further validate the expression of the top of 10 PPI hub genes in the synovium of other patients, we selected GSE153015 as a testing dataset and performed ROC analysis in RStudio software. After calculation, all 10 genes (RPS6, RPS14, RPS25, RPL11, RPL27, RPS29, SNRPE, EEF2, RPL10A and RPL19) with AUC more than 0.50 were considered as hub genes, indicating that they can distinguish two diseases and testify the availability of the merged microarray data. The analysis results were available in Fig. 10.

Immune infiltration

In this study, the IICs in 22 subpopulations of immune cells were identified. As shown in Fig. 11A, the proportions and percentage of immune cells were depicted. The heatmap showed that the level of NK cells activated, dendritic cells resting, T cells regulatory (Tregs) and macrophages M2 is relatively high of the 22 immune cells. The violin plot (Fig. 11B) indicated that the proportion of B cells memory (P = 0.005), T cells regulatory (Tregs) (P = 0.003) and NK cells activated (P < 0.001) was relatively high in OA tissues compared to RA synovial, while B cells naive (P < 0.001), plasma cells (P = 0.004), T cells CD4 naive (P = 0.018), T cells CD4 memory activated(P = 0.001), dendritic cells activated (P = 0.003) and eosinophils (P < 0.001) was relatively low in OA. Finally, the quantified contrast of the distribution of IICs subsets between OA and RA synovial tissues was shown in Fig. 11C. The result showed that there was a positive correlation existed between Monocytes and B cells memory. On the contrary, a negative correlation was detected between Dendritic cells activated and T cells CD8.

Validation of the key genes by Real-time PCR

After the model establishment and verification (Fig. 12), a real-time PCR (qPCR) assay on an ABI QuantStudio™ 7 Flex Real-Time PCR System (Applied Biosystems; Thermo Scientific, USA) was used to compare the expression of the identified genes in mice synovial tissue in the OA model, RA model and normal control. Based on the qPCR results, the expression of RPS6, RPS14, RPS25, RPL11, RPL27, SNRPE and EEF2 (Fig. 13) were significantly down-regulated in RA mice compared with RA synovial. All validations except RPS29 and RPL10A were consistent with the microarray data and analytical results in this study.

Recently, with an increase of aging populations and chronic conditions, both OA and RA become the most common causes of musculoskeletal-related chronic joint disorders in elders [28, 29]. There were many common methods for the diagnosis and treatment of OA and RA. However, the clinical outcomes were still not satisfactory [30]. Genomic studies had recently been widely used to enhance the diagnosis and treatment of many diseases [13]. As a result, a large number of fragmented research and analysis had been done to obtain genomic information related to OA and RA. To further investigate the genomic profiles of OA and RA, we systematically analyzed these multiple microarray data sets using an integrated method with the biggest sample sizes so far.

A total of 1723 OA-specific DEGs, including 1690 up- and 33 down-regulated were screened, respectively. Ubiquitin was significantly enriched in OA-related GO terms and pathways of DEGs. PPI network highlighted some genes such as RPS6 and RPS14. Among that, 1683 overlapping genes were obtained, and 71 DEGs were acquired such as TPM2, NCAM2, MFHAS1 after eliminating RA-related genes. The study found 1460 RA-specific DEGs in total (1278 up- and 182 down-regulated DEGs). Some genes were identified through PPI networks such as CTNNB1 and SNRPA1. A total of 40 RA-specific genes were acquired such as CUX1, KANK1, MBTPS2 after calculation. These DEGs may be of important meaning for the occurrence and progression of OA and RA.

Autophagy was related to cellular homeostasis mechanism and responsible for metabolic regulation, largely via the removal and regression of biologically damaged intracellular products, which were in consequence of improving cartilage aging and development of structural changes [31, 32]. The recent finding supported that autophagy response machineries were regulated by some genes, such as the mammalian target of rapamycin (mTOR). Mammalian TORC-1 was adjusted by PI3K/Akt signaling pathways, as well as further regulated multiple protein synthesis and modification by phosphorylating ribosomal S6 protein kinase (RPS6) [33, 34]. RPS6 participated in the mTOR signal pathway as well as dependent growth program and was triggered by anabolic signals [35]. There was evidence suggested that expression of mTOR increased in OA joint cartilage [36], thus, genetic or pharmaceutical inhibition of mTOR signaling pathway can lead to autophagy activation and further protection from OA [36, 37].

Recent study data suggested that the multi aspects contribution of the sphingolipid machinery concerning osteoclast-osteoblast crosstalk represented one of the significant interactions underlying bone and cartilage homeostasis [38]. Under this situation, imbalances in the interrelated between osteoblasts and osteoclasts might cause bone-related illnesses such as osteoporosis, Paget’s disease, RA, and bone metastases [39, 40]. Quint P found that activation of both S1PR1 and S1PR2 can synergistically stimulate the migration of osteoblast precursors [41]. Far more telling was that the expression of S1PR2 preserved sensitivity to S1P at the preosteoblast process as well as osteoblast stage [42]. Broadly speaking, the recent data showed that the expression of S1PR2 was a pivotal gene for the progression of osteoblasts to S1P ensuring the scenario under which the osteoblast precursors were conserved in the marrow during osteoblastogenesis [38].

Fibroblast-like synoviocytes (FLSs) participated in the pathogenesis and occurrence of RA. Experiments demonstrated that MALAT1 was decreased expressed in the FLSs of RA illness. The current study revealed that MALAT1 reduced the secretion of multiple inflammatory cytokines, while its purpose was to promote the apoptosis of FLSs in RA [43] A previous report suggested that MALAT1 could promote the methylation on the CTNNB1 promoter region and consequently inhibit CTNNB1, resulting in the inactivation of the Wnt signaling pathway [44]. Li et al. found that the CTNNB1 promoter methylation and Wnt pathway inactivation inhibited the secretion of MALAT1-related inflammatory cytokine production of FLSs such as IL-1β, TNF-α, IL-6, and IL-8 in RA [45]. Together with the above studies, the understanding of the MALAT1-CTNNB1-Wnt pathway in the progression of RA may be of benefit to the understanding of cellular and molecular mechanisms of RA, with underlying of acting as a target point for treatments of RA.

Previous studies suggested that the biological process of ubiquitination and the proteasome led to OA by regulating the expression of inflammatory cytokines, such as MMP13 [46]. Frank et al. [47] found that ubiquitin was an integral part of ubiquitin-like proteins, which covered the Small Ubiquitin-related Modifiers (SUMO). Through the histological analysis of synovial tissue obtained from OA and RA patients, SUMO had definitely been related to MMP13 expression level. In the meantime, via the NF-κB pathway, SUMO-2/3 was differentially controlled by TNF-α and selectively regulated corresponding MMP expression [48].

Jia et al. suggested that the EGFR signaling pathway was needed for acquiring an adequate number of joint superficial chondrocytes via regulating proliferation ability, improving survival, and keeping superficial properties [49]. Previous work by the same institution found that EGFR signaling pathway played a pivotal role in growth plate development and secondary ossification center formation through building a chondrocyte-specific EDFR knockout mouse model [50, 51]. Sun et al. [52] put forward a concept of heterogeneous OA subgroup, especially, the pEGFR^high expression OA subgroup. The research pointed out that the character of EGFR in deeply exploring OA made EGFR a promising hot topic of therapy. Findings from the Daneshmand study also suggested that modulation of the EGFR pathway can promote mesochondrium synthesis and suppress OA cartilage degradation [53].

Genome-Wide Association Studies (GWAS) had identified hundreds of associations of autoimmune disorders. Utilizing this data, Lee et al. detected FOXO3 activity as the severity of rheumatoid arthritis, acting via the modulation of cytokine production in monocytes[54]. In the meantime, a Foxo3a haplotype was related to erosion scores in adult RA as the previous report. Some reports said that in RA joint synovial tissue, modulation of known transcriptional FoxO-related factors played a role in FoxO proteins of integrating inflammatory stimuli to regulate cell survival and apoptosis [55]. Besides, inhibition of both PI3K/AKT and MEK/ERK signaling pathways induced FOXO transcriptional activity, and the result was significantly inhibited cell migration as well as capillary tube formation [56]. What’s more, angiogenesis was the pathological characteristic of the RA synovial membrane [57].

Three PPI networks of multigroup DEGs were established using STRING and overlapped connected-nodes were constructed to further acquire the deep relationship of OA and RA. Moreover, we validated the top 10 genes by performing ROC analysis with the testing latest dataset GSE153015. As a result, all 10 genes with p value < 0.05 as well as AUC > 0.50 showed satisfactory diagnostic value for the pooled data. Then, 10 of distinguished hub genes including RPS6, RPS14, RPS25, RPL11, RPL27, SNRPE, EEF2, RPS29 and RPL10A were conducted for experimental verification. After OA and RA model establishment, 30 mice were sacrificed and synovial in bilateral knees were acquired. Further qPCR was performed to study the accuracy of the top 10 genes, and as a result, except RPS29 and RPL10A, other 8 genes were showed consistency with this bioinformatic analysis. Finally, our team chose 30 mice to further verify the veracity of the top 10 hub PPI genes.

Our study also had many limitations that needed to be improved. First, this bioinformatics analysis conducted by synovial tissues may not represent the best way for differential diagnosis of OA and RA. Second, despite collecting maximal samples to the best of our ability, the GEO database in this study was the only approach to acquire related data and perform this data mining. Thus, a multicenter, larger-scale clinical survey was an indispensable thing for further research. In this study, multiple similar genes and signaling pathways were detected to synergistically regulate in both OA and RA. The connection between these genes offered us new evidence for the next research and help us to discover potential biomarkers.

The findings from this multiple bioinformatics networks analysis study may contribute to the distinguished diagnosis of RA and RA using synovial membrane. Though the promising results, the small samples, and single database in this study remind us of acquiring more experimental research. Thus, further statistical collection with larger samples and sophisticated algorithms may provide a more persuasive result for therapeutic strategy.

OA: osteoarthritis; RA:rheumatoid arthritis; DEGs:differentially expressed genes; PPI:protein-protein interaction; GEO:Gene Expression Omnibus:logFC:log fold change; GO:Gene Ontology; KEGG:Kyoto Encyclopedia of Genes and Genomes; BP:biological processes; CC:cellular component; MF:molecular function; GSEA:Gene Set Enrichment Analysis; FDR:false discovery rate; STRING:Search Tool for the Retrieval of Interacting Genes; ROC:receiver operating characteristic; IICs:infiltrating immune cells; DMM:destabilization of the medial meniscus; CFA:Complete Freund’s Adjuvant.

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Funding

This work was supported by Zhejiang Education Commission [grant number Y202044541 to author Yanzhi Ge], National Natural Science Foundation of China [grant number 81774331 and 82074464 to author Letian Shan], Zhejiang Natural Science Foundation Young Scholars [grant number LQ20H270009 to author Li Zhou], Zhejiang Traditional Chinese Medical Science Foundation [grant number 2020ZA039 to author Li Zhou] and Natural Science Foundation of Zhejiang Chinese Medical University [grant number 020ZG42 to author Li Zhou].

Authors' contributions

Yanzhi Ge, Zuxiang Chen and Haipeng Xu analyzed and extracted the data, contributed analysis tools. Yanzhi Ge, Yanbin Fu and Li Zhou prepared figures and tables. Yanzhi Ge and Letian Shan wrote the main protocol and prepared the manuscript. Peijian Tong conceived and designed the study, approved the final draft.

Acknowledgements

Not applicable.

Availability of data and material

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Hunter DJ, Bierma-Zeinstra S, Osteoarthritis. Lancet. 2019;393:1745–59.
Scott DL, Wolfe F, Huizinga TW. Rheumatoid arthritis. Lancet. 2010;376:1094–108.
Bartok B, Firestein GS. Fibroblast-like synoviocytes: key effector cells in rheumatoid arthritis. Immunol Rev. 2010;233:233–55.
Ma VY, Chan L, Carruthers KJ. Incidence, prevalence, costs, and impact on disability of common conditions requiring rehabilitation in the United States: stroke, spinal cord injury, traumatic brain injury, multiple sclerosis, osteoarthritis, rheumatoid arthritis, limb loss, and back pain. Archives of physical medicine and rehabilitation. 2014; 95.
Wasserman AM. Diagnosis and management of rheumatoid arthritis. Am Family Phys. 2011;84:1245–52.
Rees HW. Management of Osteoarthritis of the Hip. J Am Acad Orthop Surg. 2019.
Reum Son A, Kim DY, Hun Park S, Yong Jang J, Kim K, Ju Kim B, Yun Yin X, Ho Kim J, Hyun Min B, Keun Han D, Suk Kim M. Direct chemotherapeutic dual drug delivery through intra-articular injection for synergistic enhancement of rheumatoid arthritis treatment. Scientific reports. 2015;5:14713.
DeRogatis M, Anis HK, Sodhi N, Ehiorobo JO, Chughtai M, Bhave A, Mont MA. Non-operative treatment options for knee osteoarthritis. Annals of translational medicine. 2019;7:245.
Maderbacher G, Greimel F, Schaumburger J, Grifka J, Baier C. [The knee joint in rheumatoid arthritis-current orthopaedic surgical treatment options]. Z Rheumatol. 2018;77:882–8.
King LK, Marshall DA, Faris P, Woodhouse L, Jones CA, Noseworthy T, Bohm E, Dunbar M, Hawker GA. Use of recommended non-surgical knee osteoarthritis management in patients prior to total knee arthroplasty: a cross-sectional study. The Journal of rheumatology. 2019.
Heidari B. Knee osteoarthritis prevalence, risk factors, pathogenesis and features: Part I. Caspian journal of internal medicine. 2011;2:205–12.
Townsend MJ. Molecular and cellular heterogeneity in the Rheumatoid Arthritis synovium: clinical correlates of synovitis. Best practice & research Clinical rheumatology. 2014; 28: 539 – 49.
Lu Q-Y, Han Q-H, Li X, Li Z-C, Pan Y-T, Liu L, Fu Q-G. Analysis of differentially expressed genes between rheumatoid arthritis and osteoarthritis based on the gene co-expression network. Mol Med Rep. 2014;10:119–24.
Liu F-Q. Analysis of differentially expressed genes in rheumatoid arthritis and osteoarthritis by integrated microarray analysis. Journal of cellular biochemistry. 2019;120:12653–64.
Huang DW, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic acids research. 2009;37:1–13.
Reimand J, Isserlin R, Voisin V, Kucera M, Tannus-Lopes C, Rostamianfar A, Wadi L, Meyer M, Wong J, Xu C, Merico D, Bader GD. Pathway enrichment analysis and visualization of omics data using g:Profiler, GSEA, Cytoscape and EnrichmentMap. Nat Protoc. 2019;14:482–517.
Szklarczyk D, Franceschini A, Kuhn M, Simonovic M, Roth A, Minguez P, Doerks T, Stark M, Muller J, Bork P, Jensen LJ, von Mering C. The STRING database in 2011: functional interaction networks of proteins, globally integrated and scored. Nucleic acids research. 2011;39:D561-D8.
Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J, Simonovic M, Doncheva NT, Morris JH, Bork P, Jensen LJ, Mering CV. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019;47:D607-d13.
Robin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez JC, Müller M. pROC: an open-source package for R and S + to analyze and compare ROC curves. BMC Bioinformatics. 2011;12:77.
Muschelli J. ROC and AUC with a Binary Predictor: a Potentially Misleading Metric. J Classif. 2020;37:696–708.
Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, Hoang CD, Diehn M, Alizadeh AA. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015;12:453–7.
Zhang H, Liu R, Sun L, Guo W, Ji X, Hu X. Comprehensive Analysis of Gene Expression Changes and Validation in Hepatocellular Carcinoma. Onco Targets Ther. 2021;14:1021–31.
Glasson SS, Blanchet TJ, Morris EA. The surgical destabilization of the medial meniscus (DMM) model of osteoarthritis in the 129/SvEv mouse. Osteoarthritis Cartilage. 2007;15:1061–9.
Torres-Guzman AM, Morado-Urbina CE, Alvarado-Vazquez PA, Acosta-Gonzalez RI, Chávez-Piña AE, Montiel-Ruiz RM, Jimenez-Andrade JM. Chronic oral or intraarticular administration of docosahexaenoic acid reduces nociception and knee edema and improves functional outcomes in a mouse model of Complete Freund's Adjuvant-induced knee arthritis. Arthritis Res Ther. 2014;16:R64.
Brand DD, Latham KA, Rosloniec EF. Collagen-induced arthritis. Nat Protoc. 2007;2:1269–75.
Hwang HS, Park IY, Hong JI, Kim JR, Kim HA. Comparison of joint degeneration and pain in male and female mice in DMM model of osteoarthritis. Osteoarthritis Cartilage. 2021.
Yan L, Zhou L, Xie D, Du W, Chen F, Yuan Q, Tong P, Shan L, Efferth T. Chondroprotective effects of platelet lysate towards monoiodoacetate-induced arthritis by suppression of TNF-α-induced activation of NF-ĸB pathway in chondrocytes. Aging. 2019;11:2797–811.
Woolf AD, Pfleger B. Burden of major musculoskeletal conditions. Bull World Health Organ. 2003;81:646–56.
Neogi T. The epidemiology and impact of pain in osteoarthritis. Osteoarthritis cartilage. 2013;21:1145–53.
Long NP, Park S, Anh NH, Min JE, Yoon SJ, Kim HM, Nghi TD, Lim DK, Park JH, Lim J, Kwon SW. Efficacy of Integrating a Novel 16-Gene Biomarker Panel and Intelligence Classifiers for Differential Diagnosis of Rheumatoid Arthritis and Osteoarthritis. Journal of clinical medicine. 2019; 8.
Matsuzaki T, Alvarez-Garcia O, Mokuda S, Nagira K, Olmer M, Gamini R, Miyata K, Akasaki Y, Su AI, Asahara H, Lotz MK. FoxO transcription factors modulate autophagy and proteoglycan 4 in cartilage homeostasis and osteoarthritis. Science translational medicine. 2018;10:eaan0746.
Ribeiro M, López de Figueroa P, Blanco FJ, Mendes AF, Caramés B. Insulin decreases autophagy and leads to cartilage degradation. Osteoarthritis cartilage. 2016;24:731–9.
Fenton TR, Gout IT. Functions and regulation of the 70kDa ribosomal S6 kinases. Int J Biochem Cell Biol. 2011;43:47–59.
Conn CS, Qian S-B. mTOR signaling in protein homeostasis: less is more? Cell cycle (Georgetown. Tex). 2011;10:1940–7.
Chauvin C, Koka V, Nouschi A, Mieulet V, Hoareau-Aveilla C, Dreazen A, Cagnard N, Carpentier W, Kiss T, Meyuhas O, Pende M. Ribosomal protein S6 kinase activity controls the ribosome biogenesis transcriptional program. Oncogene. 2014;33:474–83.
Zhang Y, Vasheghani F, Li Y-H, Blati M, Simeone K, Fahmi H, Lussier B, Roughley P, Lagares D, Pelletier J-P, Martel-Pelletier J, Kapoor M. Cartilage-specific deletion of mTOR upregulates autophagy and protects mice from osteoarthritis. Ann Rheum Dis. 2015;74:1432–40.
Caramés B, Hasegawa A, Taniguchi N, Miyaki S, Blanco FJ, Lotz M. Autophagy activation by rapamycin reduces severity of experimental osteoarthritis. Ann Rheum Dis. 2012;71:575–81.
Meshcheryakova A, Mechtcheriakova D, Pietschmann P. Sphingosine 1-phosphate signaling in bone remodeling: multifaceted roles and therapeutic potential. Expert Opin Ther Targets. 2017;21:725–37.
Hofbauer LC, Schoppet M. Clinical implications of the osteoprotegerin/RANKL/RANK system for bone and vascular diseases. JAMA. 2004;292:490–5.
Rachner TD, Khosla S, Hofbauer LC. Osteoporosis: now and the future. Lancet. 2011;377:1276–87.
Quint P, Ruan M, Pederson L, Kassem M, Westendorf JJ, Khosla S, Oursler MJ. Sphingosine 1-phosphate (S1P) receptors 1 and 2 coordinately induce mesenchymal cell migration through S1P activation of complementary kinase pathways. J Biol Chem. 2013;288:5398–406.
Roelofsen T, Akkers R, Beumer W, Apotheker M, Steeghs I, van de Ven J, Gelderblom C, Garritsen A, Dechering K. Sphingosine-1-phosphate acts as a developmental stage specific inhibitor of platelet-derived growth factor-induced chemotaxis of osteoblasts. Journal of cellular biochemistry. 2008;105:1128–38.
Gutschner T, Hämmerle M, Diederichs S. MALAT1 -- a paradigm for long noncoding RNA function in cancer. J Mol Med. 2013;91:791–801.
Enzo MV, Rastrelli M, Rossi CR, Hladnik U, Segat D. The Wnt/β-catenin pathway in human fibrotic-like diseases and its eligibility as a therapeutic target. Molecular cellular therapies. 2015;3:1.
Li G-Q, Fang Y-X, Liu Y, Meng F-R, Wu X, Zhang C-W, Zhang Y, Liu D, Gao B. MALAT1-Driven Inhibition of Wnt Signal Impedes Proliferation and Inflammation in Fibroblast-Like Synoviocytes Through CTNNB1 Promoter Methylation in Rheumatoid Arthritis. Human gene therapy. 2019;30:1008–22.
Radwan M, Wilkinson DJ, Hui W, Destrument APM, Charlton SH, Barter MJ, Gibson B, Coulombe J, Gray DA, Rowan AD, Young DA. Protection against murine osteoarthritis by inhibition of the 26S proteasome and lysine-48 linked ubiquitination. Ann Rheum Dis. 2015;74:1580–7.
Frank S, Peters MA, Wehmeyer C, Strietholt S, Koers-Wunrau C, Bertrand J, Heitzmann M, Hillmann A, Sherwood J, Seyfert C, Gay S, Pap T. Regulation of matrixmetalloproteinase-3 and matrixmetalloproteinase-13 by SUMO-2/3 through the transcription factor NF-κB. Ann Rheum Dis. 2013;72:1874–81.
Huang TT, Wuerzberger-Davis SM, Wu Z-H, Miyamoto S. Sequential modification of NEMO/IKKgamma by SUMO-1 and ubiquitin mediates NF-kappaB activation by genotoxic stress. Cell. 2003;115:565–76.
Jia H, Ma X, Tong W, Doyran B, Sun Z, Wang L, Zhang X, Zhou Y, Badar F, Chandra A, Lu XL, Xia Y, Han L, et al. EGFR signaling is critical for maintaining the superficial layer of articular cartilage and preventing osteoarthritis initiation. Proc Natl Acad Sci USA. 2016;113:14360–5.
Zhang X, Siclari VA, Lan S, Zhu J, Koyama E, Dupuis HL, Enomoto-Iwamoto M, Beier F, Qin L. The critical role of the epidermal growth factor receptor in endochondral ossification. Journal of bone mineral research: the official journal of the American Society for Bone Mineral Research. 2011;26:2622–33.
Zhang X, Zhu J, Li Y, Lin T, Siclari VA, Chandra A, Candela EM, Koyama E, Enomoto-Iwamoto M, Qin L. Epidermal growth factor receptor (EGFR) signaling regulates epiphyseal cartilage development through β-catenin-dependent and -independent pathways. J Biol Chem. 2013;288:32229–40.
Sun H, Wu Y, Pan Z, Yu D, Chen P, Zhang X, Wu H, Zhang X, An C, Chen Y, Qin T, Lei X, Yuan C, et al. Gefitinib for Epidermal Growth Factor Receptor Activated Osteoarthritis Subpopulation Treatment. EBioMedicine. 2018;32:223–33.
Daneshmand M, Parolin DAE, Hirte HW, Major P, Goss G, Stewart D, Batist G, Miller WH, Matthews S, Seymour L, Lorimer IAJ. A pharmacodynamic study of the epidermal growth factor receptor tyrosine kinase inhibitor ZD1839 in metastatic colorectal cancer patients. Clinical cancer research: an official journal of the American Association for Cancer Research. 2003;9:2457–64.
Schaid DJ, Chen W, Larson NB. From genome-wide associations to candidate causal variants by statistical fine-mapping. Nature reviews Genetics. 2018;19:491–504.
Reedquist KA, Ludikhuize J, Tak PP. Phosphoinositide 3-kinase signalling and FoxO transcription factors in rheumatoid arthritis. Biochemical Society transactions. 2006;34:727–30.
Srivastava RK, Unterman TG, Shankar S. FOXO transcription factors and VEGF neutralizing antibody enhance antiangiogenic effects of resveratrol. Molecular cellular biochemistry. 2010;337:201–12.
McInnes IB, Schett G. The pathogenesis of rheumatoid arthritis. N Engl J Med. 2011;365:2205–19.

Download PDF

Version 1

posted

You are reading this latest preprint version

Identification and validation of hub genes of synovial tissue for patients with osteoarthritis and rheumatoid arthritis

Status:

Version 1

Abstract

Figures

Introduction

Materials And Methods

Microarray data

Integration of microarray data and Screening DEGs

Gene Ontology and pathway enrichment analysis

Gene Set Enrichment Analysis (GSEA)

PPI network analysis

Verification by ROC Analysis

Immune infiltration

Animal Experiments

Evaluation of arthritis severity and tactile sensitivity testing

Real-time PCR (qPCR) assay

Statistical analysis

Results

Basic datasets of OA and RA

DEGs screening

GO pathway of OA-specific and RA-specific DEGs

KEGG pathway of OA-specific and RA-specific DEGs

GSEA of involved signaling pathways

PPI network and modules

Verification of Hub Genes by ROC Analysis

Immune infiltration

Validation of the key genes by Real-time PCR

Discussion

Conclusions

Abbreviations

Declarations

References

Status:

Version 1