Bioinformatics approach to explore the common pathogenic genes between lumbar disc degeneration and osteoarthritis

doi:10.21203/rs.3.rs-4860142/v1

Download PDF

Research Article

Bioinformatics approach to explore the common pathogenic genes between lumbar disc degeneration and osteoarthritis

https://doi.org/10.21203/rs.3.rs-4860142/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Osteoarthritis (OA) and lumbar disc degeneration disease (IDD) are both degenerative and chronic inflammatory diseases that directly and profoundly impact human quality of life and medical services.

Methods

To better understand how IDD is related to OA, we used transcriptomic and single-cell analysis to look for molecular and cellular overlaps between the two diseases in this work. Two gene expression profiles (GSE124272 and GSE55235) from the Gene Expression Omnibus database are downloaded to detect common differentially expressed genes (DEGs) for OA and IDD diseases to find similar pathways, possible medicines, and regulatory networks. The single-cell transcriptome of synovial tissues sequenced from OA patients was analyzed for the immune characteristics of the shared DEGs.

Results

Based on RNA-seq datasets, 184 DEGs shared by these two diseases were found. 114 hub genes were then identified using degree and Matthew's correlation coefficient algorithms based on creating the Protein-protein interactions network. Interactions between transcription factors and shared hub genes, coregulatory networks between miRNAs and these hub genes, and protein-drug interactions were found. Moreover, scRNA-seq of cells of synovial tissues from OA patients revealed that 30 of these hub genes were highly enriched in the inflammatory macrophages (IL1B⁺Mac). Finally, CSF1 and S100A12 were identified as potential biomarkers by machine learning, and their area under the curve in IDD and OA is greater than 0.7 by the receiver operating characteristic curve analysis.

Conclusions

IDD might contribute to the advancement of OA by inciting cytokine storms, thereby enhancing comprehension and treatment strategies for OA patients accompanied by IDD. Moreover, the identified candidate drugs, DEG-regulated networks, and potential diagnostic genes (CSF1 and S100A12) offer promising therapeutic avenues for addressing both conditions.

Osteoarthritis

lumbar disc degeneration disease

differentially expressed genes

hub genes

immune response

drug treatment

Osteoarthritis (OA), one of the most prevalent forms of arthritis in elderly adults, involves several joints and harms joint function, mainly characterized by synovial inflammation, cartilage degeneration, and reduced joint function[1]. Its clinical manifestations include local pain and limited joint movement[2]. Besides, patients with OA may experience symptoms such as swelling, pain, or stiffness, which can harm their quality of life[3]. Women over the age of 55 and males over the age of 65 have an increased risk of developing OA[4]. OA has been associated with an increased incidence rate in the past few decades due to the aging population and obesity[5]. According to a recent study, there will be approximately 400 million cases of OA in China by 2030[6]. However, the etiology and pathogenesis of the disease are still unknown. It is urgent to explore the potential mechanisms responsible for OA development, which is beneficial for developing novel and effective strategies.

Intervertebral disc degeneration (IDD) is a degenerative disease recognized as the primary cause of low back pain[7], and affects 70%-85% of the global population. This disease is distinguished by biochemical alterations in the nucleus pulposus, extracellular matrix degradation in the nucleus pulposus, and exterior damage to the fibrous ring accompanied by increasing fibrosis[8]. Research reports suggest that inflammation and signaling substances may play important roles in the occurrence and development of intervertebral disc degeneration, such as interleukin-1 (IL-1), tumor necrosis factor-alpha (TNF), nitric oxide, matrix metalloproteinases and metalloproteinases[9]. In addition, during the progression of IDD, macrophage infiltration leads to the release of many inflammatory factors, forming an inflammatory storm[10]. Meanwhile, OA is a chronic inflammatory disease, and the release of inflammatory factors has a memorable role in the process of OA. For example, enhanced IL-1 and TNF levels are related to the increased risk of OA[11, 12]. Besides, significantly increased levels of inflammatory cytokines, such as IL-1 and IL-6, were found in OA patients' synovial fluid, synovium, subchondral bone, and cartilage [13]. Inhibition of IL-1 and TNF can significantly reduce cartilage degeneration and inflammatory response in experimental OA[14, 15]. Meanwhile, clinical studies have shown a significant correlation between hip OA and endplate degeneration at L1, L3, and L5 levels[16], and degenerative diseases of the upper lumbar spine can induce the occurrence of primary unilateral knee osteoarthritis[17]. Therefore, IDD and OA may be interdependent and potentially have considerable pathological and physiological overlap. However, so far, the impact of IDD on the severity and progression of OA is still unclear. Given this, exploring the potential mechanisms underlying the impact of lumbar disc degeneration on the progression of osteoarthritis (OA) is crucial for improving the treatment of OA patients. By comprehensively deciphering the shared differentially expressed genes between these two diseases and identifying potential diagnostic biomarkers, a foundation can be established for more effective treatment strategies and ultimately improve clinical outcomes.

Here, we examined the overlapping molecular and cellular intersections of gene expression in IDDD and OA, utilizing transcriptome and single-cell sequencing data. Our investigation suggests that IDDD may exacerbate OA by an excessive cytokine storm. These findings have significant implications for advancing our comprehension and treatment strategies of OA patients with concurrent IDD. The entire workflow diagram in this study is shown in Fig. 1.

Dataset collection and processing

Synovial tissue microarray dataset (GSE55235), including 10 normal samples and 10 Osteoarthritis (OA) samples, was processed using the Affymetrix GPL96 platform, and GSE124272 [the database consists of 16 whole blood samples from 8 patients with Intervertebral disc degeneration (IDD) and 8 healthy controls] was analyzed using the platform of Agilent GPL21185. In addition, a validated dataset, GSE55457, was analyzed using the GPL96 Affymetrix U133A Array. GSE55457's gene expression profile included 10 OA patients and 10 healthy controls. GSE150408 contains the whole blood RNA-seq transcriptome data of 17 IDD patients and 17 healthy volunteers and was analyzed using the Agilent GPL21185 platform. GSE176205, containing 3 controls and 6 IDD tissue samples, was analyzed using GPL20301. All these datasets were obtained from the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo), and all four datasets of data's gene expression matrix files were created with the par pl package of Perl from raw RNA-seq data.

Identification of common differentially expressed genes (DEGs) between IDD and OA

First, use R software (version 4.2.1) to read the expression matrices of datasets GSE1919, GSE55235, and GSE51588 for batch correction[18] and normalization[19]. Subsequently, the "limma" R package was performed to identify DEGs between IDD patients and controls or OA patients and controls. Meanwhile, genes with |log2FC | > 0.585 and p-value < 0.05 were screened to be DEGs[20]. Volcano plots and DEG expression heatmaps were constructed using the "ggplot2" and "heatmap" packages. Besides, the common DEGs of GSE124272 and GSE55457 were acquired based on the Jvenn, a web-based Venn diagram tool.

Gene ontology (GO) and pathway enrichment analysis

The "clusterProfile" package[21], which is a widely used general enrichment tool for interpreting omics data, was next performed for analyzing GO and Disease Ontology (DO) to explore further the potential molecular mechanism of these shared DEGs and their link to disease. Disease ontology is a unified cross-species disease annotation[22]. Meanwhile, Enrichr, an online application (https://maayanlab.cloud/Enrichr/), was used to discover the major signaling pathways of these common DEGs, which included Kyoto Encyclopaedia of Genes and Genomes (KEGG) Pathways, WikiPathways, Reactome, and Hallmark. We used a p-value of 0.05 as the cut-off for significance[23].

Protein-protein interactions (PPI) network construction and hub gene identification

Cytoscape (v3.9.1), which is a software tool used for visualizing and analyzing biological molecular interaction networks, was used to generate a visual representation of the PPI networks based on gene interactions combined with a total score of > 0.4[24], which were derived from the STRING database (https://string-db.org/). The Cytoscape plugin CytoHubba, a crucial tool for analyzing and identifying the center of biological networks, was used to determine the degree of each node. Hub genes were identified as those with a degree score (highly connected nodes with a degree) of > 10[25, 26]. Then, two algorithms of the cytoHubba plugin, including degree and Matthew's correlation coefficient (MCC) algorithms, were employed to identify shared hub genes.

Recognition of TFs and miRNA interact with shared hub genes

Transcription factors (TFs) are proteins that attach to specific genes, exerting control over the rate of genetic information transcription. As a result, they play a pivotal role in facilitating a comprehensive understanding of molecular processes[27]. Using the 'NetworkAnalyst' function to find topologically plausible TFs derived from the JASPAR database that tend to bind to common DEGs opens up new avenues of investigation into the TFs regulating common hub genes. Six different taxonomic groupings are represented in the publicly available JASPAR database of TF binding profiles (https://jaspar.genereg.net/). Meta-analysis of gene expression data using the PPI network is performed in NetworkAnalyst to study biological mechanisms in greater depth[28].

Additionally, microRNAs (miRNAs), a class of small non-coding RNAs that bind to the transcripts of widely used DEG, thus altering protein expression, were identified by doing a target gene-miRNA interaction analysis. MirTarBase (https://mirtarbase.cuhk.edu.cn/~miRTarBase/) is the major database of miRNA-target gene interactions. Here, from the interaction of miRNAs-gene interaction based on the NetworkAnalyst platform, we extracted miRNAs that interact with shared hub genes focused on topological analysis from the mirTarBase database. NetworkAnalyst simplifies the exploration of complex datasets for researchers, helping to identify biological features and functions and formulating meaningful biological hypotheses[29]. The interaction network of TFs and miRNAs was visualized using Cytoscape.

Download and Analysis of scRNA-seq data

The single-cell raw matrix data of synovial tissues from GSE152805 were acquired from the GEO database. We then employed the Seurat package (version 4.1.2) to process the raw count matrix (UMI counts for each sample). The low-quality cells, defined as those with fewer than 200 genes per cell, less than 10 cells per gene, and more than 15% mitochondrial genes, were excluded. Following this, we performed normalization using the 'NormalizeData' function based on the corrected expression matrix. After data normalization, we scaled the data by applying the 'FindVariableGenes' function, focusing on the expression of 2000 highly variable genes (HVGs). Principal component analysis of these HVGs was executed using the 'RunPCA' function, and we eliminated batch effects between different samples using the 'harmony' function. The purpose of identifying cell types across all cell populations is to set the resolution to 1.0. Subsequently, we employed the "RunTSNE" function to project the combined cells into a two-dimensional space. Following this, we utilized the "Dimplot" function to visualize the cell clusters, concluding with their annotation based on known markers.

Assessment of Potential drugs

Understanding the features of diseases requires careful consideration of protein-drug interactions (PDIs) or drug molecule identification. As a result, we forecasted PDIs using the web application Enrichr. First, based on the shared hub genes, drug compounds were discovered utilizing the Drug Signatures database (DSigDB) via the Enrichr database based on the common hub genes of OA and IDD. Enrichr is a popular online platform that hosts a vast collection of diverse gene set libraries, enabling the exploration of gene set enrichment across the entire genome (https://maayanlab.cloud/Enrichr/). DSigDB (http://dsigdb.tanlab.org/DSigDBv1.0/) is a data repository for identifying targeted therapeutic compounds connected to these hub genes[30]. The database consists of 22,527 gene sets, and one accessible way to access the DSigDB database is through Enrichr under the Disease/Drug functionality category.

Immune Cell Infiltration and Single-sample Gene Set Enrichment Analysis (ssGSEA)

To explore the correlation between these common DEGs and immune cell infiltration characteristics in OA, infiltration levels of different immune cells in OA were evaluated using the "CIBERSORT" package, and the relative proportions of 22 immune cells were estimated via iterations (1000 times). The "CIBERSORT" package utilizes linear support vector regression on gene expression profiles, allowing for the estimation of immune cell quantities in samples using RNA sequencing data[31]. Subsequently, the Wilcoxon test was run to calculate the differences between the OA disease group and the normal control group in each immune cell population. Then, to further validate the activity of these DEGs during OA progression, the ssGSEA algorithm was implemented to elucidate the enrichment scores of differential expression genes quantitatively shared in each sample[32]. The Gene Set Variation Analysis (GSVA) R package (Version 1.26.0) was employed to calculate normalized enrichment scores for gene sets.

Candidate diagnosis marker identification

Then, to identify the diagnostic markers in OA patients, we used two machine learning algorithms (LASSO logistic regression and the SVM algorithm). As a dimensionality reduction method, compared to regression analysis, LASSO regression demonstrates superior performance when evaluating high-dimensional data. It employs regularization to enhance prediction accuracy. Here, the "Glmnet" software package for LASSO regression analysis is used to identify genes significantly associated with the discrimination of OA and healthy samples. Meanwhile, as a supervised machine learning method for support vectors, support vector machines (SVM) rank features recursively to prevent overfitting and find the optimal variable by deleting the feature vectors generated by the support vector machine[33]. The SVM classifier from the "e1071" R package was adopted for screened biomarkers in diagnosing OA patients. Receiver Operating Characteristic (ROC) analysis of these feature genes was conducted in RStudio using the pROC package. Area under the curve (AUC), with a range of 0.7 to 1, typically has good predictive performance[34]. Therefore, feature genes with an AUC value of > 0.7 are considered diagnostic for OA.

GSEA function enrichment and immune cell infiltration analysis of CSF1 and S100A12 in OA

The "limma" and "cluster Profiler" packages in R were used to carry out gene set enrichment analysis (GSEA) at a threshold of p < 0.05 to identify differences in signaling pathways of CSF1 and S100A12 (grouped by median expression level). GSEA is for enrichment analysis of pathways based on whole-genome expression profiles[35]. The gene sets of c2.cp.kegg.v7.4.symbols were derived from the MSigDB database for running the GSEA analysis. To determine the interaction relationship among two effective biomarkers (CSF1 and S100A12) and differential immune cells and inflammatory factors, the "limma" and "vioplot" packages in R were used to compare the relative proportions of the immune cells between control and OA patients, and the "correlation analysis" was run to perform Pearson correlation analysis for these two genes (CSF1and S100A12) and immune cells and inflammatory factors.

184 shared DEGs were identified between IDD and OA

To examine the potential influences of IDD on OA, we compared the differential gene expression in the whole blood samples of 8 intervertebral disc degeneration patients and 8 healthy intervertebral disc patients from dataset GSE124272. Meanwhile, we analyzed the differential gene expression between 10 osteoarthritides (OA) patients' pathological synovial tissues and 10 healthy control tissues using GEO data GSE55235. Subsequently, 2381 DEGs (986 upregulated and 1395 down-regulated) in the IDD dataset and 2361 DEGs (938 upregulated and 1423 down-regulated) in the OA dataset were identified using "limma" R package with a threshold of p-value < 0.05 and |log2FC| ≥ 0.585. Volcano plots and heatmaps are commonly used to visualize DEGs for IDD and OA (Supplementary Fig. 1A-D). Next, 184 common DEGs between IDD and OA datasets were identified using the Venn diagram (Fig. 2). Hence, these findings can be inferred that there exists a mutual linkage between these two diseases, as they exhibit a shared presence of multiple common genes (184 common DEGs); these shared DEGs were used in further analysis to delve deeper into exploring the potential connections between the two diseases.

GO and pathway enrichment analysis

To explore the potential biological function and diseases related to these common DEGs, the "clusterProfile" package was performed to carry out GO and DO enrichment analysis. The top 10 enriched GO terms for these shared DEGs are shown in Supplementary Fig. 2A. The GO analysis showed that shared DEGs were remarkably enriched in neutrophil activation involved in immune response, neutrophil-mediated immunity, and negative regulation of inflammatory response; nuclear chromosome, phagocytic vesicle lumen and phagocytic vesicle; transition metal ion binding, protein phosphatase binding and zinc ion binding (Supplementary Fig. 2A and Supplementary Table 1). Meanwhile, the top 40 enriched DO terms for these shared DEGs are shown in the Fig. 3B. The findings indicated that these DEGs exhibited enrichment in diseases such as hip osteoarthritis, autoimmune diseases, systemic lupus erythematosus, breast cancer, rheumatoid arthritis, and other related illnesses (Supplementary Fig. 2B and Supplementary Table 2). Subsequently, the four global databases were used further to investigate the possible pathways of these common DEGs. The results showed that these common DEGs were significantly related to the MAPK signaling pathway, T cell receptor signaling pathway, Rap1 signaling pathway, immune system, innate immune system, and VEGFA-VEGFR2 signaling pathway (Supplementary Fig. 3A-B and Supplementary Table 3, 4). These results indicate that these DEGs may influence the occurrence of OA through pathways associated with inflammation and immune response.

Hub gene classification and construction of regulatory networks

The PPI network of these shared DEGs was created by using the STRING database (https://string-db.org/), which is a database for predicting functional associations between proteins, and Cytoscape (v3.7.2) visualized these DEGs with a combined score ≥ 0.4 as the cut-off criteria. Figure 3A depicts the PPI network of common DEGs consisting of 184 nodes and 912 edges. The degree and MCC algorithms of the cytoHubba plugin in Cytoscape were utilized to calculate the protein-protein interaction (PPI) network of 181 DEGs. Among 184 common DEGs, 131 hub genes were identified with a threshold degree of > 10 based on Degree algorithms of the cytoHubba plugin (Fig. 3B and Supplementary Table 5). A Venn diagram was then constructed using the intersection of hub genes obtained from these two algorithms, and 114 shared hub genes were identified (Fig. 3C). Additionally, to ascertain the primary alterations at the transcriptional level and delve further into key protein regulatory molecules or common DEGs, we employed a network-based approach to decode the regulatory transcription factors (TFs) and miRNAs. Figure 4A depicts the network of TF regulators of shared DEGs, and Fig. 4B depicts the interaction of miRNAs with shared DEGs. We identified 36 transcription factors (TFs) and 29 post-transcriptional regulators (miRNAs) regulated with more than one shared DEGs by analyzing the interaction network between transcription factor genes and miRNA genes. These results essentially indicate a strong interplay between them.

scRNA-seq reveals the accumulation of common 22 IRGs in inflammatory macrophages of OA patients

To clarify the impact of IDD on the process of OA, we subsequently explored the immune characteristics of these common hub genes in OA patients. The scRNA-seq dataset of OA patients (including three OA synovial samples) was obtained from the dataset GSE152805. A total of 9822 cells were acquired, and these cells were divided into 17 clusters based on unsupervised graph-based clustering (Fig. 5A). These clusters were then defined as synovial subintimal fibroblasts (SSF), synovial intimal fibroblasts (SIF), macrophage, DC, endothelial cell (EC), smooth muscle cell (SMC), Mast Cell, proliferating immune cell (ProIC), T cell, and B cell based on the known cell markers (Fig. 5B-E). Subsequently, we analyzed the expression of these shared hub genes in different cell types within OA patients. Among their hub genes were 30 genes highly expressed in macrophages, compared with other cell types (Fig. 5F). To delve deeper into the potential roles of these 29 genes in macrophages of OA patients, we conducted a subgroup analysis by extracting macrophages. Firstly, we performed unsupervised graph-based clustering of macrophages, identifying a total of 10 clusters (Fig. 6A), and then these clusters were defined as FOLR2⁺macrophages (FOLR2⁺Mac) and IL1B⁺macrophages (IL1B⁺Mac) according to specific gene signatures (Fig. 6B-E). Subsequently, we further explored the distribution of these 30 genes in the subgroups of macrophages.

Interestingly, we noticed that almost all these genes were highly expressed in IL1B⁺Mac compared with the FOLR2⁺Mac except HCRT (Fig. 6F). Simultaneously, we further observed that these genes were highly expressed in IDD patients (blood and tissue samples) compared to normal samples and also found to be highly expressed in OA patients (Supplementary Fig. 4A-B). These genes were highly expressed in IL1B + Mac and were defined as inflammation-related DEGs (IRGs). Subsequently, to clarify the characteristic changes of these IRGs in OA immune cell infiltration. We then utilized single-sample Gene Set Enrichment Analysis (ssGSEA) to calculate each sample's scores based on these IRGs' expressions (Supplementary Fig. 5A). The results indicated that the ssGSEA score (DEGs signature) is significantly higher in the OA disease samples than in the control samples (Supplementary Fig. 5B). Besides, this score also shows a significant positive correlation with M1 macrophages, follicular helper T cells, Eosinophils and Neutrophils (Supplementary Fig. 5C-G). Simultaneously, we further assessed the associations between several pro-inflammatory cytokines (IL1B, IL2, IL6, IL12B and TNF) in OA and DEGs signature. The results reveal a significant positive correlation between these pro-inflammatory cytokines and the DEG signature (Supplementary Fig. 6). To conclude, these results indicated that the IRGs, which were highly expressed by IL1B⁺Mac, may be the main contributors to the cytokine storm in OA disease, as well as IDD may also promote the progression of OA through cytokine storms.

Determination and verification of diagnostic markers in OA patients by machine learning

Finally, to explore potential biomarkers for OA, we further screened diagnostic genes based on these IRGs of the OA datasets using LASSO regression. As shown in Fig. 7A, LASSO regression identified 13 variables that were determined among these common hub genes as characteristic genes for OA. Meanwhile, other arithmetics (SVM-RFE method) were utilized to explore OA's diagnostic biomarkers. The data showed that 22 genes were identified as signature genes by the SVM-RFE algorithm (Fig. 7B). Then, regarding the results of the two algorithms, after taking the intersection. In total, 2 overlapping genes (Colony stimulating factor 1, CSF1 and S100 calcium-binding protein A12, S100A12) were selected between those two arithmetics via Venn diagrams (Fig. 7C).

Additionally, Supplementary Table S6 displays the shared hub genes discovered by the LASSO regression and SVM-RFE algorithm. Subsequently, we further validated the expression level of these two genes between OA patients and controls. The results showed that these two genes were significantly upregulated in the OA patients compared with control samples (Fig. 7D). Besides, ROC analysis confirmed that these two genes exhibited strong diagnosis stability, and the AUC value was 0.820 and 0.870, respectively (Fig. 7E). Thus, these two genes may be critical diagnostic markers involved in OA progression.

Meanwhile, these two genes were also significantly more expressed in the whole blood and tissue samples of IDD patients than in control samples (Fig. 7F), and the AUC value of these two genes also revealed an excellent diagnostic value for LDDD patients (Fig. 7G). Moreover, the expression and AUC value of these two potential genes were also confirmed in other OA datasets (GSE55457) and IDD external independent datasets (GSE150408) (Supplementary Fig. 7A-D). In conclusion, these results demonstrated that these two genes may be potential targets for treating and diagnosing OA and IDD.

Correlation between CSF1 and S100A12 expression and immune cell infiltration across OA

To further investigate the potential role of these two genes (CSF1 and S100A12) in OA, GSEA was used to screen the gene sets enriched in different expression levels of these two genes. As shown in Supplementary Fig. 8A-B, compared with the group of CSF1 low expression, the Calcium signaling pathway, Focal-adhesion, and ECM-receptor interaction were significantly enriched in the CSF1 high expression group. Chemokine signaling pathway, Focal-adhesion and Cytokine-cytokine receptor interaction were more involved in the S100A12 high expression group than in the S100A12 low expression group. Therefore, these findings showed that these two genes may influence the occurrence of OA through pathways associated with chemokine and cytokine-cytokine receptor interaction. The study mentioned above reveals a significant enrichment of inflammation-related infiltrating cells (M1 macrophages and neutrophils) in patients with OA. Therefore, to further investigate whether these two prognostic remarkers are associated with immune cell infiltration during the development of OA, Pearson correlation analysis was performed on these two genes (CSF1 and S100A12) and immune cells, and the results showed that the expression of these two genes was appreciably positively correlated with the MO and M1 macrophages (Fig. 8A-B).

Meanwhile, we further explored the relationship between pro-inflammatory cytokines and these two genes in OA. We observed a significant positive correlation between the expression of these pro-inflammatory cytokines and CSF1 and S100A12 (Fig. 8C). Thus, the above results suggested that these two genes may promote the progression of OA through inflammatory-related responses.

Identification of candidate drugs

Assessment of PDIs is essential to understand the structural features recommended for receptor sensitivity. In the aspects of common DEGs as potential drug targets in OA and IDD, the top 15 candidate therapeutic compounds were identified by Enrichr based on the transcriptome signatures from the DSigDB database. These chemical compounds are extracted based on their p-value. These potential drugs are suggested for the common hub genes. These drugs could be chemical compounds that are used to treat both diseases. The DSigDB database's list of effective medications for common hub genes is presented in Table 1.

Table 1

List of the top 15 suggested drugs for OA.
Term	p-value	Counts	Chemical Formula
VALPROIC ACID	0.001155287	64	C₈H₁₆O₂
Retinoic acid	2.99E-06	46	C₂₀H₂₈O₂
estradiol	1.24E-05	45	C₁₈H₂₄O₂
benzo[a]pyrene	2.43E-04	42	C₂₀H₁₂
Tetradioxin	2.97E-05	40	C₁₂H₄C_l4O₂
trichostatin A	5.37E-05	38	C₁₇H₂₂N₂O₃
AFLATOXIN B1	3.89E-04	32	C₁₇H₁₂O₆
Triclocarban	5.92E-04	29	C₁₃H₉C_l3N₂O
COPPER	1.87E-07	28	Cu
resveratrol	2.35E-07	27	C₁₄H₁₂O₃
calcitriol	1.11E-05	27	C₂₇H₄₄O₃
hydrogen peroxide	0.001913058	27	H₂O₂
Bisphenol A	7.66E-09	26	C₁₅H₁₆O₂
progesterone	2.21E-05	26	C₂₁H₃₀O₂
Decitabine	4.76E-04	22	C₈H₁₂N₄O₄

OA, a widespread chronic joint disease, affects many individuals worldwide[36]. OA has gradually increased healthcare and social expenditures as the primary cause of impairment in older persons [37]. Although multiple diagnostic and therapeutic approaches are available for OA, the clinical outcomes are still unsatisfactory. IDD is a degenerative spine disease that causes chronic lower back pain and motor disturbance[38]. In more severe stages, loss of joint space, subchondral sclerosis, and osteophytes, similar to OA in the articular joint, accompany intervertebral disc degeneration[39]. Thus, IDD may accelerate the progression of OA. Nevertheless, there are few reports illustrating the interactions between the two diseases. Here, we employed various bioinformatics analysis methods to reveal the potential molecular mechanism of IDD in the pathogenesis of OA and provide a possible basis for future targeted and preventive treatment.

Here, 184 common DEGs were identified between OA and IDD. After constructing the PPI network of shared DEGs, 114 common hub genes were identified based on the degree and MCC algorithm of the cytoHubba plugin in Cytoscape. Subsequently, 114 identified shared hub genes were evaluated by functional enrichment and pathway analysis. The enrichment analysis results showed that these common hub genes were significantly associated with the MAPK signaling pathway, T cell receptor signaling pathway, Rap1 signaling pathway, immune response, neutrophil-mediated immunity, and negative regulation of inflammatory response. Existing research indicates an inherent connection between the MAPK signaling pathway and osteoarthritis (OA) advancement. An imbalance in the MAPK pathway expedites inflammatory reactions, releasing a multitude of cartilage matrix-degrading enzymes and intensifying cartilage degeneration[40]. The Rap1 signaling pathway is an essential component of pathological osteoclast resorption, contributing to the progression of arthritis[41]. Furthermore, research has reported that T-cell signaling mediates the onset of various inflammatory diseases, such as inflammatory bowel disease and autoimmune arthritis[42]. Therefore, these data indicated that these common DEG-related pathways are closely associated with the inflammatory response.

Previous research has shown that OA, as an inflammatory disease, can be exacerbated by a dysregulated inflammatory storm[43], and large amounts of pro-inflammatory cytokines can be detected in patients with IDD[44]. To investigate the distribution of these DEGs across various immune cell types, we delved deeper into the characteristics of single-cell atlases derived from OA tissue samples. Our data proved that 30 of these hub genes were broadly enriched in macrophages compared to other cell types, and subgroup analysis of macrophages also revealed that the IL1B⁺ macrophage subtype was a prominent source of expression for these 30 genes. Meanwhile, we further observed that these genes were highly expressed in IDD and OA patients compared to control samples. Subsequently, based on the expression of these 30 genes, the score derived from ssGSEA calculations for these gene sets is to be constructed to explore further the characteristic changes of these genes in OA immune cell infiltration. The ssGSEA score of these gene sets was significantly higher in the OA disease group than in the control group. These scores also positively correlated with M1 macrophages, eosinophils, and neutrophils. Meanwhile, we further assessed the associations between several pro-inflammatory cytokines in OA and these scores. The results showed a significant positive correlation between these pro-inflammatory cytokines and these scores. Polarization of synovial macrophages towards M1 exacerbates the development of OA[45].

Additionally, research has indicated that neutrophil elastase (NE) contributes to tissue degeneration, osteophyte formation, and the release of inflammatory cytokines and chemokines, thereby participating in the progression of OA[46]. Therefore, based on these findings, it can be inferred that the highly expressed IRGs in IL1B⁺ macrophages could be the primary drivers of the cytokine storm in OA. IDD might also contribute to OA progression through cytokine storms.

Subsequently, we conducted feature selection utilizing two distinct algorithms to delve deeper into potential biomarkers for osteoarthritis (OA). The results showed that two hub genes (CSF1 and S100A12) were identified as feature genes using LASSO regression analysis and the SVM-RFE method. CSF1, the protein encoded by this gene, is a cytokine that controls macrophage production, differentiation, and function. CSF regulates the development, proliferation, differentiation and activation of myeloid cells in health and disease[47]. In addition, it has been reported that CSF-1 can promote the survival, proliferation and differentiation of bone marrow cells (including monocytes, macrophages and osteoclast) by binding to the tyrosine kinase receptor CSF1R[48]. Exogenous administration of CSF-1 exacerbates the incidence and severity of OA[49]. The protein encoded by is a member of the S100 protein family containing two EF-hand calcium-binding motifs. S100 protein has also been proven to have important extracellular pro-inflammatory effects and cytokine-like activity[51]. Clinical research shows serum S100A12 expression is elevated in patients with OA [52]. The change in S100A12 expression level in synovial fluid can reflect the clinical severity of patients with primary knee osteoarthritis[53]. Here, our findings revealed significant overexpression of CSF1 and S100A12 in both the test and validation sets among IDD and OA patients and the OA patients with high CSF1 and S100A12 levels had positively correlated with M1 macrophages.

Meanwhile, the results of the correlation analysis indicate a significant positive correlation between these two genes and pro-inflammatory cytokines in the progression of OA. Experimental studies also demonstrated that blocking the CSF1R can reduce inflammation in human and mouse models of rheumatoid arthritis[50]. Furthermore, clinical case study results indicate that S100A12 in the serum and synovial fluid (SF) of patients with OA may be an independent factor for predicting the risk of this condition[53]. Therefore, these two genes may serve as potential diagnostic biomarkers for OA patients and could be involved in the pathogenesis of OA through inflammatory responses.

Transcription factors (TFs) govern the proportion of transcriptional processes, while miRNAs play a crucial role in gene regulation at the post-transcriptional level and in RNA silencing. Both TFs and miRNAs are of significant importance in understanding disease progression. Then, a study of the interactions between TFs-genes and miRNAs-genes was done to determine which regulators control the transcription and expression of common hub genes. 36 TFs and 29 miRNAs were found to have substantial interactions with common hub genes. For example, YY1[54] and NFKB1[55] are associated with the regulation and progression of OA diseases. Furthermore, some miRNAs, such as miR-335-5p[56] and miR-218-5p[57], participated in the progression of OA.

In recent years, several chemical agents and drugs have been utilized as potential therapeutic approaches to OA. Local nonsteroidal anti-inflammatory drugs (NSAIDs) such as lurbiprofen cataplasms and loxoprofen sodium cataplasm have effectively treated OA[58]. However, the incidence rate of OA disease has increased significantly, so it is urgent to develop new drugs to treat OA. Based on shared hub genes, some potential drugs and compounds targeting OA and IDD have been identified using database, such as valproic acid (VPA), estradiol, and trichostatin A. VPA is a flavonoid compound extracted from herbaceous plants and is a first-line drug for treating epilepsy and mania[59]. VPA has been proven to have multiple therapeutic effects, including anticancer, neuroprotective, differentiation, and neurodegenerative activity[60]. Besides, studies found that VPA was identified as a putative agent associated with OA progression and protected chondrocytes from damage caused by OA stimulation[61]. Estradiol is a well-known homeostatic regulator of the bone and immune system and can also control exercise ability and pain sensitivity[62]. In addition, experimental studies have confirmed that the physiological level of estradiol limits the progression of OA in mice[63]. Trichostatin A (TSA) is a non-selective histone deacetylase inhibitor that inhibits OA and rheumatoid arthritis progression by inhibiting the expression of MMP[64]. So, these chemical compounds might be good choices for drugs to treat OA. However, there are still some limitations to this study. First, the sample size of this study is small, which may affect the reliability of the conclusions. In addition, the results of this study are all based on the analysis of database data. The necessary support of experimental evidence is lacking. CSF1 and S100A12 need further validation through cell experiments, animal models and clinical samples. Additionally, these drugs need to undergo chemical experiments for further confirmation. Despite these limitations, the results of the current study may provide potential directions for future research.

Our study elucidates the potential molecular connections between IDD and OA at the cellular and molecular levels through comprehensive transcriptome and single-cell sequencing data analysis. Our research has identified that IDDD exacerbates the severity of OA through an excessive cytokine storm. Additionally, we identified two potential diagnostic biomarkers for these diseases. By expanding our understanding of these conditions, our research provides potential directions for further investigating the molecular mechanisms of OA. It may lead to more effective therapeutic strategies and improved clinical outcomes for OA patients.

Osteoarthritis (OA)

lumbar disc degeneration disease (IDD)

differentially expressed genes (DEGs)

interleukin-1 (IL-1)

tumor necrosis factor-alpha (TNF)

L1, L3, and L5 levels

Gene Expression Omnibus (GEO)

Gene ontology (GO)

Disease Ontology (DO)

Kyoto Encyclopaedia of Genes and Genomes (KEGG)

Protein-protein interactions (PPI)

Matthew's correlation coefficient (MCC)

Transcription factors (TFs)

microRNAs (miRNAs)

protein-drug interactions (PDIs)

Drug Signatures database (DSigDB)

Single-sample Gene Set Enrichment Analysis (ssGSEA)

Gene Set Variation Analysis (GSVA)

support vector machines (SVM)

Receiver Operating Characteristic (ROC)

Area under the curve (AUC)

gene set enrichment analysis (GSEA)

transcription factors (TFs)

protein-protein interaction (PPI)

subintimal fibroblasts (SSF)

synovial intimal fibroblasts (SIF)

endothelial cell (EC)

smooth muscle cell (SMC)

proliferating immune cell (ProIC)

FOLR2⁺macrophages (FOLR2⁺Mac)

IL1B⁺macrophages (IL1B⁺Mac)

inflammation-related DEGs (IRGs)

ingle-sample Gene Set Enrichment Analysis (ssGSEA)

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Funding

This work was supported by the National Natural Science Foundation of China (82371751).

Data availability statement

All data needed to evaluate the conclusions in the paper are presented in the main manuscript and the Supplementary Information.

Acknowledgments

Not applicable.

Authors' contributions

WSC and ZMY designed the study. HGL conducted and interpreted the bioinformatics analysis. WGZ and SYY prepared the draft of the manuscript. All authors revised and approved the final manuscript

Competing interests

The authors declare no competing interests.

Kraus VB, Blanco FJ, Englund M, Karsdal MA, Lohmander LS. Call for standardized definitions of osteoarthritis and risk stratification for clinical trials and clinical use. Osteoarthritis Cartilage. 2015;23(8):1233–41.
Winter AR, Collins JE, Katz JN. The likelihood of total knee arthroplasty following arthroscopic surgery for osteoarthritis: a systematic review. BMC Musculoskelet Disord. 2017;18(1):408.
Roseti L, Desando G, Cavallo C, Petretta M, Grigolo B. Articular Cartilage Regeneration in Osteoarthritis. Cells. 2019; 8(11).
Jimenez G, Cobo-Molinos J, Antich C, Lopez-Ruiz E. Osteoarthritis: Trauma vs Disease. Adv Exp Med Biol. 2018;1059:63–83.
Nelson AE. Osteoarthritis year in review 2017: clinical. Osteoarthritis Cartilage. 2018;26(3):319–25.
Zhang Z, Huang C, Jiang Q, et al. Guidelines for the diagnosis and treatment of osteoarthritis in China (2019 edition). Ann Transl Med. 2020;8(19):1213.
Dagenais S, Caro J, Haldeman S. A systematic review of low back pain cost of illness studies in the United States and internationally. Spine J. 2008;8(1):8–20.
Nishida K, Gilbertson LG, Robbins PD, Evans CH, Kang JD. Potential applications of gene therapy to the treatment of intervertebral disc disorders. Clin Orthop Relat Res. 2000; (379 Suppl):S234–41.
Wang L, He T, Liu J, et al. Revealing the Immune Infiltration Landscape and Identifying Diagnostic Biomarkers for Lumbar Disc Herniation. Front Immunol. 2021;12:666355.
Risbud MV, Shapiro IM. Role of cytokines in intervertebral disc degeneration: pain and disc content. Nat Rev Rheumatol. 2014;10(1):44–56.
Cai H, Sun HJ, Wang YH, Zhang Z. Relationships of common polymorphisms in IL-6, IL-1A, and IL-1B genes with susceptibility to osteoarthritis: a meta-analysis. Clin Rheumatol. 2015;34(8):1443–53.
Wojdasiewicz P, Poniatowski LA, Szukiewicz D. The role of inflammatory and anti-inflammatory cytokines in the pathogenesis of osteoarthritis. Mediators Inflamm. 2014; 2014:561459.
Wang T, He C. Pro-inflammatory cytokines: The link between obesity and osteoarthritis. Cytokine Growth Factor Rev. 2018;44:38–50.
Chien SY, Tsai CH, Liu SC et al. Noggin Inhibits IL-1beta and BMP-2 Expression, and Attenuates Cartilage Degeneration and Subchondral Bone Destruction in Experimental Osteoarthritis. Cells 2020; 9(4).
Chen H, Shao X, Li L, et al. Electroacupuncture serum inhibits TNF–alpha–mediated chondrocyte inflammation via the Ras–Raf–MEK1/2–ERK1/2 signaling pathway. Mol Med Rep. 2017;16(5):5807–14.
Bajwa NS, Toy JO, Young EY, Cooperman DR, Ahn NU. Disk degeneration in lumbar spine precedes osteoarthritic changes in hip. Am J Orthop (Belle Mead NJ). 2013;42(7):309–12.
Kim YC, Kim KT, Ha KY, et al. Radiographic and clinical evidence: osteoarthritic knee can change surgical result for lumbar degenerative disease patient undergone surgery for 3-year follow-up: a retrospective comparative clinical study. BMC Musculoskelet Disord. 2020;21(1):740.
Yang M, Zheng H, Su Y, et al. Bioinformatics Analysis Identified the Hub Genes, mRNA-miRNA-lncRNA Axis, and Signaling Pathways Involved in Rheumatoid Arthritis Pathogenesis. Int J Gen Med. 2022;15:3879–93.
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.
Chen Y, Li ZY, Zhou GQ, Sun Y. An Immune-Related Gene Prognostic Index for Head and Neck Squamous Cell Carcinoma. Clin Cancer Res. 2021;27(1):330–41.
Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS. 2012;16(5):284–7.
Bello SM, Shimoyama M, Mitraka E et al. Disease Ontology: improving and unifying disease annotations across species. Dis Model Mech. 2018; 11(3).
Zhang C, Ma Z, Nan X, et al. Comprehensive analysis to identify the influences of SARS-CoV-2 infections to inflammatory bowel disease. Front Immunol. 2023;14:1024041.
Song X, Du R, Gui H, et al. Identification of potential hub genes related to the progression and prognosis of hepatocellular carcinoma through integrated bioinformatics analysis. Oncol Rep. 2020;43(1):133–46.
Zhu X, Li T, Niu X, Chen L, Ge C. Identification of UBE2T as an independent prognostic biomarker for gallbladder cancer. Oncol Lett. 2020;20(4):44.
Teng P, Xu X, Ni C, et al. Identification of key genes in calcific aortic valve disease by integrated bioinformatics analysis. Med (Baltim). 2020;99(29):e21286.
Ye Z, Wang F, Yan F, et al. Bioinformatic identification of candidate biomarkers and related transcription factors in nasopharyngeal carcinoma. World J Surg Oncol. 2019;17(1):60.
Zhou G, Soufan O, Ewald J, Hancock REW, Basu N, Xia J. NetworkAnalyst 3.0: a visual analytics platform for comprehensive gene expression profiling and meta-analysis. Nucleic Acids Res. 2019;47(W1):W234–41.
Xia J, Gill EE, Hancock RE. NetworkAnalyst for statistical, visual and network-based meta-analysis of gene expression data. Nat Protoc. 2015;10(6):823–44.
Yoo M, Shin J, Kim J, et al. DSigDB: drug signatures database for gene set analysis. Bioinformatics. 2015;31(18):3069–71.
Newman AM, Steen CB, Liu CL, et al. Determining cell type abundance and expression from bulk tissues with digital cytometry. Nat Biotechnol. 2019;37(7):773–82.
Barbie DA, Tamayo P, Boehm JS, et al. Systematic RNA interference reveals that oncogenic KRAS-driven cancers require TBK1. Nature. 2009;462(7269):108–12.
Zhao E, Xie H, Zhang Y. Predicting Diagnostic Gene Biomarkers Associated With Immune Infiltration in Patients With Acute Myocardial Infarction. Front Cardiovasc Med. 2020;7:586871.
Mandrekar JN. Receiver operating characteristic curve in diagnostic test assessment. J Thorac Oncol. 2010;5(9):1315–6.
Subramanian A, Tamayo P, Mootha VK, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545–50.
Rabago D, Nourani B. Prolotherapy for Osteoarthritis and Tendinopathy: a Descriptive Review. Curr Rheumatol Rep. 2017;19(6):34.
Zhou J, Xiao Y, Li C, Liu T, Wang W. Association between use of nonsteroidal anti-inflammatory drugs and risk of myocardial infarction in patients with spondyloarthritis and osteoarthritis. Ann Rheum Dis. 2019;78(8):e78.
Geurts JW, Willems PC, Kallewaard JW, van Kleef M, Dirksen C. The Impact of Chronic Discogenic Low Back Pain: Costs and Patients' Burden. Pain Res Manag. 2018; 2018:4696180.
Rustenburg CME, Emanuel KS, Peeters M, Lems WF, Vergroesen PA, Smit TH. Osteoarthritis and intervertebral disc degeneration: Quite different, quite similar. JOR Spine. 2018;1(4):e1033.
Shi J, Zhang C, Yi Z, Lan C. Explore the variation of MMP3, JNK, p38 MAPKs, and autophagy at the early stage of osteoarthritis. IUBMB Life. 2016;68(4):293–302.
Zou W, Izawa T, Zhu T, et al. Talin1 and Rap1 are critical for osteoclast function. Mol Cell Biol. 2013;33(4):830–44.
Yan J, Pandey SP, Barnes BJ, Turner JR, Abraham CT, Cell-Intrinsic. IRF5 Regulates T Cell Signaling, Migration, and Differentiation and Promotes Intestinal Inflammation. Cell Rep. 2020;31(13):107820.
Manchikanti L, Soin A, Benyamin RM, et al. An Update of the Systematic Appraisal of the Accuracy and Utility of Discography in Chronic Spinal Pain. Pain Physician. 2018;21(2):91–110.
Siebuhr AS, Bay-Jensen AC, Jordan JM, et al. Inflammation (or synovitis)-driven osteoarthritis: an opportunity for personalizing prognosis and treatment? Scand J Rheumatol. 2016;45(2):87–98.
Zhang H, Cai D, Bai X. Macrophages regulate the progression of osteoarthritis. Osteoarthritis Cartilage. 2020;28(5):555–61.
Chaney S, Vergara R, Qiryaqoz Z, Suggs K, Akkouch A. The Involvement of Neutrophils in the Pathophysiology and Treatment of Osteoarthritis. Biomedicines. 2022; 10(7).
Hamilton JA, Achuthan A. Colony stimulating factors and myeloid cell biology in health and disease. Trends Immunol. 2013;34(2):81–9.
Hume DA, MacDonald KP. Therapeutic applications of macrophage colony-stimulating factor-1 (CSF-1) and antagonists of CSF-1 receptor (CSF-1R) signaling. Blood. 2012;119(8):1810–20.
Bischof RJ, Zafiropoulos D, Hamilton JA, Campbell IK. Exacerbation of acute inflammatory arthritis by the colony-stimulating factors CSF-1 and granulocyte macrophage (GM)-CSF: evidence of macrophage infiltration and local proliferation. Clin Exp Immunol. 2000;119(2):361–7.
Garcia S, Hartkamp LM, Malvar-Fernandez B, et al. Colony-stimulating factor (CSF) 1 receptor blockade reduces inflammation in human and murine models of rheumatoid arthritis. Arthritis Res Ther. 2016;18:75.
Pietzsch J, Hoppmann S. Human S100A12: a novel key player in inflammation? Amino Acids. 2009;36(3):381–9.
Myles A, Viswanath V, Singh YP, Aggarwal A. Soluble receptor for advanced glycation endproducts is decreased in patients with juvenile idiopathic arthritis (ERA category) and inversely correlates with disease activity and S100A12 levels. J Rheumatol. 2011;38(9):1994–9.
Wang LC, Zhang HY, Shao L, et al. S100A12 levels in synovial fluid may reflect clinical severity in patients with primary knee osteoarthritis. Biomarkers. 2013;18(3):216–20.
Yue S, Su X, Teng J, Wang J, Guo M. Cryptotanshinone interferes with chondrocyte apoptosis in osteoarthritis by inhibiting the expression of miR–574–5p. Mol Med Rep. 2021; 23(6).
Tang S, Nie X, Ruan J, Cao Y, Kang J, Ding C. Circular RNA circNFKB1 promotes osteoarthritis progression through interacting with ENO1 and sustaining NF-kappaB signaling. Cell Death Dis. 2022;13(8):695.
Lu X, Li Y, Chen H, Pan Y, Lin R, Chen S. miR-335-5P contributes to human osteoarthritis by targeting HBP1. Exp Ther Med. 2021;21(2):109.
Liu Y, Zhao D, Wang X, Dong Y, Ding F. LncRNA KCNQ1OT1 attenuates osteoarthritic chondrocyte dysfunction via the miR-218-5p/PIK3C2A axis. Cell Tissue Res. 2021;385(1):115–26.
Li D, Cheng Y, Yuan P et al. Efficacy and safety of flurbiprofen cataplasms versus loxoprofen sodium cataplasms in knee osteoarthritis: a randomized controlled trial. Chin Med J (Engl). 2023.
Henry TR. The history of valproate in clinical neuroscience. Psychopharmacol Bull. 2003;37(Suppl 2):5–16.
Abematsu M, Tsujimura K, Yamano M, et al. Neurons derived from transplanted neural stem cells restore disrupted neuronal circuitry in a mouse model of spinal cord injury. J Clin Invest. 2010;120(9):3255–66.
Sun L, Zheng W, Liu QD, Ge L. Valproic Acid Protects Chondrocytes from LPS-Stimulated Damage via Regulating miR-302d-3p/ITGB4 Axis and Mediating the PI3K-AKT Signaling Pathway. Front Mol Biosci. 2021;8:633315.
Straub RH. The complex role of estrogens in inflammation. Endocr Rev. 2007;28(5):521–74.
Corciulo C, Scheffler JM, Humeniuk P, et al. Physiological levels of estradiol limit murine osteoarthritis progression. J Endocrinol. 2022;255(2):39–51.
Chen WP, Bao JP, Tang JL, Hu PF, Wu LD. Trichostatin A inhibits expression of cathepsins in experimental osteoarthritis. Rheumatol Int. 2011;31(10):1325–31.

No competing interests reported.

SupplementaryFigures.docx

Download PDF

Editorial decision: Revision requested
01 Sep, 2024
Reviews received at journal
01 Sep, 2024
Reviews received at journal
27 Aug, 2024
Reviewers agreed at journal
23 Aug, 2024
Reviewers agreed at journal
17 Aug, 2024
Reviewers invited by journal
09 Aug, 2024
Editor assigned by journal
09 Aug, 2024
Submission checks completed at journal
09 Aug, 2024
First submitted to journal
05 Aug, 2024

You are reading this latest preprint version

Bioinformatics approach to explore the common pathogenic genes between lumbar disc degeneration and osteoarthritis

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Introduction

Material and Methods

Dataset collection and processing

Identification of common differentially expressed genes (DEGs) between IDD and OA

Gene ontology (GO) and pathway enrichment analysis

Protein-protein interactions (PPI) network construction and hub gene identification

Recognition of TFs and miRNA interact with shared hub genes

Download and Analysis of scRNA-seq data

Assessment of Potential drugs

Immune Cell Infiltration and Single-sample Gene Set Enrichment Analysis (ssGSEA)

Candidate diagnosis marker identification

GSEA function enrichment and immune cell infiltration analysis of CSF1 and S100A12 in OA

Results

184 shared DEGs were identified between IDD and OA

GO and pathway enrichment analysis

Hub gene classification and construction of regulatory networks

scRNA-seq reveals the accumulation of common 22 IRGs in inflammatory macrophages of OA patients

Determination and verification of diagnostic markers in OA patients by machine learning

Correlation between CSF1 and S100A12 expression and immune cell infiltration across OA

Identification of candidate drugs

Discussion

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1