Bioinformatic analysis to find small molecules related to rheumatoid arthritis

Authors


Abstract

Background

Rheumatoid arthritis (RA) is a chronic, systemic inflammatory disorder that may affect many tissues and organs, but principally attacks flexible (synovial) joints.

Aims

Our aim is to explore the change of gene expression profile in patients with RA, and investigate the underlying mechanism of the pathogenesis and progression of RA.

Methods

We downloaded the dataset GSE2053 from Gene Expression Omnibus database and screened the differentially expressed genes by analyzing the profiles between RA and normal cells with bioinformatics methods. Furthermore, Gene Ontology (GO) function analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were used to screen GO and the significantly changed signaling pathways in RA cells with the Database for Annotation, Visualization and Integrated Discovery (DAVID).

Results

By bioinformatics methods, we obtained the metabolic pathway changed in the cells of patients with RA, and explored small molecule drugs that can restore these changes.

Conclusions

These results may provide a new approach for explore the pathogenesis of RA and a new breakthrough in the medical treatment of patients with RA.

Introduction

Rheumatoid arthritis (RA) is a chronic, systemic inflammatory disorder that may affect many tissues and organs, but principally attacks flexible (synovial) joints.[1] The pathology of the disease process often leads to the destruction of articular cartilage and ankylosis (fusion) of the joints.[2] About 1% of the world's population is afflicted by RA, women three times more often than men.[3]

Rheumatoid arthritis is a form of autoimmunity, the causes of which are still incompletely known. Previous studies have suggested that the disease involves abnormal B cell–T cell interaction and inflammation is then driven either by B cell or T cell products stimulating release of tumor necrosis factor (TNF) and other cytokines.[4, 5] Epidemiological studies have confirmed a potential association between RA and two herpesvirus infections.[6] The various pathology of RA brings much difficulty for us in curing it. The goals of treatment include minimizing clinical symptoms such as pain and swelling, as well as preventing bone deformity and radiographic damage (e.g., bone erosions visible in X-rays), and maintaining quality of life in terms of day-to-day activities.[7] Pharmacological treatment of RA can be divided into disease-modifying antirheumatic drugs (DMARDs), anti-inflammatory agents and analgesics.[8, 9]

In our study, we extracted and clustered the differentially expressed genes (DEGs) with Gene Ontology (GO) enrichment from intracellular composition, molecular function and biological processes. In addition, we mad enrichment analysis for all pathways with the Database for Annotation, Visualization and Integrated Discovery (DAVID) to obtain the changed signaling pathways in RA cells. Finally, we divided the DEGs into up-regulation and down-regulation, and compared these with the DEGs worked by small molecules using Gene Set Enrichment Analysis (GSEA) to obtain the small molecules.

Methods

Gene expression profiles of RA cells and normal cells

In order to explore the difference between RA cells and normal cells, we detected the DEGs on the gene level and investigated the mechanism of RA. The RA and normal cells were collected and microarray analysis was performed to obtain their gene profiles. The specimen GSE2053 which was based on GPL1740 (HUMAN UNIGENE SetI Part 1) was from Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/). We downloaded the original CEL files and the platform probe annotation information files of five RA cells and five normal cells for further research.

Extraction of DEGs

After obtaining the original data, they were divided into a RA group and control group, and then analyzed by R language.[10] First, we used the RMA (Robust Multichip Averaging) method[11] to normalize the different microarrays, and compared the microarray of different groups with Limma package,[12] aiming to detect the DEGs.

GO functional analysis of DEGs

In order to find the changes and DEGs on a cellular level and their functional cluster, GO database[13] was used to cluster genes according to their function and location. We clustered the DEGs with GO enrichment[14] from intracellular composition, molecular function and biological processes.

Pathway analysis

To reveal the changes of development in RA on a molecular level, we focused on the pathway related to RA. We screened all metabolic pathways and non-metabolic pathways from the current authoritative pathway database Kyoto Encyclopedia of Genes and Genomes (KEGG), and then made enrichment analyses[15] for them with DAVID[16] to obtain the signaling pathways changed in RA cells.

Small molecules profile analysis

CMAP (Connectivity Map) database stores the human cell gene transcriptional expression profile under the active small molecule intervention, including 6100 groups of small molecule interventional experiments (small interfering groups and the normal control groups) and a total of 7056 expression profiles.[17] We analyzed the expression difference between normal cells and RA cells, then compared them with the DEGs interfered by small molecules, hoping to find the small molecule that could simulate or reverse gene changes in RA cells or normal cells. We divided the DEGs into up-regulation and down-regulation, selected the 500 most significant probes respectively, and compared them with the DEGs worked by small molecules using GSEA to obtain the enrichment value. The value was between −1 and 1, which reflected the similarity between the gene expression profiles and those caused by the phenotype of interest (1 represents the state of normal cells, −1 represents the state of RA cells).

Results

Recognizing the DEGs

We used the classic t-test to examine the profiles of normal or RA cells respectively, and obtained the DEGs. Selecting P < 0.05 as the significant threshold, the expression of a total 1301 probes changed, relating to 844 genes.

GO functional analysis of DEGs

We performed GO function enrichment analysis for DEGs, which could be categorized into three main categories (cellular component, biological process and molecular function). The DEGs enriched in cellular component are shown in Figure 1. The DEGs enriched in biological processes are shown in Figure 2. The DEGs enriched in molecular function are shown in Figure 3.

Figure 1.

The differentially expressed genes (DEGs) enriched in cellular components. Circles represent the removed entries after processing by the local redundancy algorithm, squares represent removed entries after processing by global redundancy algorithm, and rectangles represent the final remaining entries.

Figure 2.

The differentially expressed genes (DEGs) enriched in biological processes. Circles represent the removed entries after processing by the local redundancy algorithm, squares represent removed entries after processing by global redundancy algorithm, and rectangles represent the final remaining entries.

Figure 3.

The differentially expressed genes (DEGs) enriched in molecular functions. Circles represent the removed entries after processing by the local redundancy algorithm, squares represent removed entries after processing by global redundancy algorithm, and rectangles represent the final remaining entries.

From the cellular component, the DEGs of RA were focused on the extracellular region, the proteinaceous extracellular matrix, extracellular space, external side of the plasma membrane, cell surface, extracellular matrix, T cell receptor complex and the extracellular region. All the locations reflected that the signaling pathway had changed. Among them, it was worth mentioning that the changes of T cell receptor complex could be the direct cause of RA. From the biological process, the DEGs of RA were focused on osteoblast differentiation, the leukotriene metabolic process, defense response, negative regulation of cell proliferation, response to lipids, the cellular alkene metabolic process, small molecule biosynthetic process, response to stimulus and cellular response to stimulus. As we can see from these, cell differentiation, proliferation, immune and signal responses of the RA cells can be changed. From the molecular function, receptor activity, receptor binding, extracellular matrix structural constituent, metalloenzyme regulator activity and growth factor binding have changed, which are focused on signal transduction.

Significantly changed pathway in RA cells

When RA occurs, the cellular expression profile changes significantly, wherein some of the genes under pathological conditions compared with normal cells have a significant change. Here we performed a sub-KEGG pathway enrichment analysis to find the RA signaling pathway. We selected P < 0.1 and at least two genes in this pathway as limiting conditions to obtain a significant change in biological pathways, which are shown in Table 1.

Table 1. Significantly changed pathway in Rheumatoid arthritis (RA) cells
TermP-value
hsa04520: Adherens junction0.012
hsa03320: PPAR signaling pathway0.019
hsa04512: ECM-receptor interaction0.021
hsa00480: Glutathione metabolism0.034
hsa04115: p53 signaling pathway0.047
hsa05216: Thyroid cancer0.051
hsa00590: Arachidonic acid metabolism0.054
hsa00020: Citrate cycle (TCA cycle)0.063
hsa04310:Wnt signaling pathway0.065
hsa04110: Cell cycle0.085
hsa05340: Primary immunodeficiency0.09
hsa00290: Valine, leucine and isoleucine biosynthesis0.098

Exploring the related small molecules

We analyzed the expression difference between normal cells and RA cells, then compared them with the DEGs interfered by small molecules, hoping to find the small molecule that could simulate or reverse gene changes in RA cells or normal cells. Twenty strongest correlation (minimum P-value) small molecules are shown in Table 2.

Table 2. Twenty strongest correlation small molecules
Connectivity map nameEnrichmentP-value
H-70.9420
Trichostatin A0.2570
0175029–00000.8350.00002
Camptothecin0.9810.00004
GW-85100.9270.00004
8-azaguanine0.9250.00004
Thioguanosine0.9140.00004
Ethotoin0.8120.00008
Podophyllotoxin−0.9130.0001
Sulconazole0.9030.0001
Etiocholanolone−0.8220.0001
Dipyridamole0.7970.00012
Fludrocortisone−0.7120.00012
Sanguinarine0.990.00014
Vincamine−0.7560.00044
Felbinac−0.8810.00048
Naringenin−0.8660.00062
Prestwick-692−0.8420.00107
11-deoxy-16,16-dimethylprostaglandin E20.8320.00119
Chlorpromazine0.430.00122
Alsterpaullone0.9140.00132

As we can see from the Table 1, small molecules sanguinarine (enrichment = 0.99) and camptothecin (enrichment = 0.981) can simulate the state of normal cells; that is, two small molecules can repair the abnormal state of RA cells, which are good potential therapeutic drugs for treatment of RA. Meanwhile, small molecule podophyllotoxin (enrichment = −0.913) can simulate the state of RA cells, which means that analogs of this small molecule likely are strongly induced RAs drugs. In addition, researching the mechanism of these small molecules may provide a new way to explore the pathogenesis of RA.

Conclusion

Rheumatoid arthritis is a chronic inflammatory disease characterized by joint swelling, joint tenderness and destruction of synovial joints, leading to severe disability and premature mortality.[18, 19] Its occurrence causes much pain. Therefore, the study of RA has an important role in human health.

There are many DEGs between RA and normal cells; in this microarray data 844 differentially expressed genes were obtained, meaning that the mechanism of development in RA, and even the key genes to the treatment of RA also are present in these genes.

Rheumatoid arthritis is a synovitis characterized by chronic systemic autoimmune disease.[20] It can be predicted from GO clustering and pathway analysis that causes of RA are changes in the extracellular signaling pathway, eventually leading to metabolic disorders, immune change and change of cell proliferation and apoptosis. Therefore, to explore the means of medical treatment of RA, we might try to find a breakthrough using these aspects.

The pathway after clustering focused on signal transduction, immune-related and amino acid synthesis pathways; also included are the metabolism and disease-related pathway. The important change in the cells of patients with RA lies in signal transduction. First, adherens junction and extracellular matrix (ECM)-receptor interaction pathways have changed. Adherens junctions are protein complexes that occur at cell–cell junctions in epithelial tissues, usually more basal than tight junctions.[21] An adherens junction is defined as a cell junction in which the cytoplasmic face is linked to the actin cytoskeleton. The ECM is the extracellular part of animal tissue that usually provides structural support to the animal cells in addition to performing various other important functions.[22] These two pathways are important ways to receive extracellular signals, and their changes trigger a series of changes in intracellular signaling. Another cellular pathway in patients with RA is the metabolic aspect, involving the citric acid cycle citrate cycle (TCA cycle), glutathione metabolism, arachidonic acid metabolism, valine and leucine isoleucine biosynthesis. In addition, peroxisome proliferator-activated receptor (PPAR) signaling pathway is closely related to change in lipid metabolism. Apart from the obvious symptoms of RA, and immune-related and cancer-related signal pathways also change. The change of primary immunodeficiency may be a major factor causing RA pain.

Disease-modifying anti-rheumatic drug (DMARD) originally are drugs that affect biological measures, but are now usually used to reduce the rate of damage to bone and cartilage.[23] DMARDs have been used in the treatment of RA for a long time now, and this type of drug usually is of the small molecule type, such as azathioprine, cyclosporine A, D-penicillamine, gold salts, hydroxychlorquine, leflunomide, methotrexate, minocycline and so on. Unfortunately, some drugs are not very potent, and are usually insufficient to control symptoms. Methotrexate is considered to be the most important DMARD; however, it has shown to have toxic gastrointestinal, hematologic, pulmonary and hepatic side-effects.[24] Sulphasalazine may cause side effects that can range in severity from mild to serious; even leukopenia has also been reported in therapies with sulphasalazine, but in very rare cases. In our study, we analyzed the expression difference between normal cells and RA cells and found small molecules similar or opposite to genes in RA cells or normal cells. Sanguinarine (enrichment = 0.99) and camptothecin (enrichment = 0.981) can simulate the state of normal cells (Table 2), which are good potential therapeutic drugs for treatment of RA. Sanguinarine is a quaternary ammonium salt from the group of benzylisoquinolinealkaloids, which usually is extracted from plants. Camptothecin is a cytotoxicquinolinealkaloid which inhibits the DNA enzymetopoisomerase, which is isolated from the bark and stem of Camptotheca acuminata. Camptothecin showed remarkable anticancer activity in preliminary clinical trials, the analogues of which have been used in cancer chemotherapy, such topotecan[25] and irinotecan.[26] Podophyllotoxin (enrichment = −0.913) can simulate state of RA cells, which means it can induce RAs. Podophyllotoxin arrests the cell cycle in the metaphase through the inhibition of tubulin polymerization.[27]

In our study, we further discussed metabolic pathways changed in the cells of patients with RA, and explored small molecule drugs that can respond to these changes, which could provide a new breakthrough in the medical treatment of patients with RA.

Ancillary