Background

Dental pulp is loose connective tissue in the pulp cavity surrounded by rigid dentin and is necessary for tooth nutrition, innervation, and immunocompetency [1]. When the hard dental tissue is broken, various stimuli may induce pathological changes in dental pulp, which is difficult to heal due to the lack of collateral circulation [2]. Pulpitis is an inflammatory disease of the dental pulp, and bacterial infection is considered to be the most important trigger of pulpitis [3]. Whether a pathological change occurs in dental pulp and the degree of lesions are related not only to the virulence and amount of bacteria but also to the defensive capacity of the host [4]. Tertiary dentin is formed reactively when dental pulp is stimulated, and a balance exists between inflammation and reparative processes. If harmful stimuli cannot be removed in time, increasing numbers of immune cells are recruited to the pulp tissue [5] and the immune-inflammatory balance is disrupted. Thus, irreversible pulpitis can occur as a result of uncontrollable inflammation caused by invading bacteria [6]. Without appropriate treatment, pulpitis may result in pulp necrosis, periapical periodontitis and more severe conditions [7].

Different therapies for pulp disease can be selected according to the evaluation of pulp inflammation. Given the importance of tooth preservation, vital pulp therapies such as pulp cap**, pulpotomy, and stepwise excavation of caries may be more appropriate than root canal treatment under some circumstances [8]. Currently, guided by the criteria of the American Association of Endodontists (AAE), medical history and clinical examination are the main methods used to evaluate pulp inflammation severity, considering information such as pain quality and history and responses to pulp sensitivity tests [9]. Decisions regarding the preservation or removal of vital pulp depend on whether the pulpitis is considered reversible or irreversible [10]. However, histopathological examinations have revealed weak correlations between clinical features and pulp status [6]. Thus, tissues collected from cases diagnosed as irreversible pulpitis may not present severe inflammation in pathologic examination [11].

In clinical practice, incomplete or ambiguous clinical data make it difficult to determine whether a deep carious lesion has led to the widespread infection of dental pulp (irreversible pulpitis), the infection of only a small amount of tissue near the lesion, or no infection of pulp tissue (reversible pulpitis) [6]. Hence, clinical diagnosis has limitations for determining the degree of pulp inflammation [10]. As histopathological examinations of pulp tissue cannot be completed without tooth extraction, a new method of non-invasive pulp diagnostics is needed.

Dental pulp is not isolated in the oral cavity and releases many biological products to the external environment in response to external harmful stimuli [12,13,14]. At the cellular or molecular level, a wide range of molecules are released during pulpal and periapical inflammation, including cytokines, proteases, inflammatory mediators, growth factors, and antimicrobial peptides [15, 16]. Measurable levels of molecules can be found not only in pulp tissue but also in pulpal blood [17], dentinal fluid [18, 19], periapical fluid [1], and gingival crevicular fluid [20], which can be collected non-invasively and analysed without extirpating the pulpal tissue [19, 21, 22]. The factors that can be used to assess the level of pulp inflammation are called biomarkers of pulpitis [23]. Belonging to the matrix metalloproteinase (MMP) family, MMP9 can accelerate gelatinolytic activity in inflamed pulp tissue [24]. The levels of MMP9 have been found to be significantly increased in pulp blood collected from pulp exposure [17]. In addition, biomarkers in dental fluid allowing potential diagnosis of pulpitis have been identified [19]. An analysis of gingival crevicular fluid from teeth with irreversible pulpitis showed that the levels of interleukin-8 (CXCL8) were significantly higher than those in healthy contralateral teeth [20]. In addition, the upregulation of active TIMP-2 and myeloperoxidase (MPO) has been observed in inflamed pulp tissue [25]. Employed together with clinical examinations, biomarkers may serve as diagnostic tools to identify different stages of pulpitis [26].

In this study, we focused on gene expression in pulp tissue from pulpitis patients. We selected and analysed two microarray platform datasets in the GEO database, integrated the datasets and identified DEGs between pulpitis and normal pulp tissues. Then, GO enrichment analysis, KEGG pathway analysis and GSEA were used to analyse the major biological functions of the DEGs. Ten hub genes related to pulpitis were identified by constructing a PPI network with Cytoscape. The aim of the present study was to identify candidate biomarkers for pulpitis diagnosis and prognosis based on functional and molecular analyses by evaluating DEGs in pulpitis and normal tissue.

Methods

Two microarray datasets of pulpitis from GEO database were retrieved, with the keywords: “pulpitis”, “Homo sapiens” and “dental pulp”. The diagnostic criteria of normal pulp and pulpitis were in line with endodontics diagnoses system in accordance with the American Association of Endodontists (AAE) guidelines [27]. GSE77459 includes 6 samples of pulpitis and 6 samples of normal pulps, and GSE92681 includes 7 samples of pulpitis and 5 samples of normal pulps, respectively using the microarray platform GPL17692 and GPL16956. The details of both studies are shown in Table 1.

Table 1 Summary of two individual datasets of pulpitis
Full size table

Data processing

After removing the probes of lncRNA in GSE92681, mRNAs of two datasets were merged into one file, and then ComBat normalization in SVA package (https://bioconductor.org/packages/sva/) was used to remove batch effects based on the standard protocol [28, 29]. Then, the raw data was converted into the form of an expression matrix and handled with the Linear Models for Microarray data (limma, https://bioconductor.org/packages/limma/) package in Bioconductor. Up- or downregulated DEGs between samples of pulpitis and normal pulps were identified with the cut-off criteria of adjusted p-value (adj. P. val) < 0.05 and |fold change (FC)| > 2.

Functional analysis of DEGs

GO enrichment and KEGG pathway analyses were used to investigate the functional progression of pulpitis. Biological process (BP), molecular function (MF), cellular component (CC) in GO analysis and potential pathways in KEGG analysis were performed in the Database for Annotation Visualization and Integrated Discovery (DAVID) (https://david.ncifcrf.gov/). P < 0.05 and false discovery rate (FDR) < 5% were used as the cut-off criteria.

The gene set enrichment analysis

To further explore the function of DEGs in inflammatory progression, GSEA was performed using h.all.v7.0.symbols.gmt (http://software.broadinstitute.org/gsea/ downloads.jsp) as a reference gene set [30]. The GSEA software (version 4.0) is available on the GSEA website (http://software.broadinstitute.org/gsea/index.jsp). Gene set permutations were performed 1000 times, and the pathway set list was sorted by the Normalized Enrichment Score (NES). P < 0.05 and FDR < 0.25 were considered statistically significant.

PPI network analysis and hub gene identification

The PPI network of DEGs was constructed using the Search Tool for THE Retrieval of Interacting Genes (STRING) database (Version 11.0, http://string-db.org/). PPI pairs and PPI network were visualized in the Cytoscape software (Version 3.7.1), and cytohHubba and MCODE plugin in Cytoscape were used to calculate the degrees of protein nodes and to select the significant modules. Top 10 genes were identified as hub genes.

Results

Dataset integration and identification of DEGs

According to the principal component analysis (PCA), the data from two samples, GSM2434473 and GSM2434475, were excluded (Fig. 1). Then, an expression matrix of normal pulp and pulpitis was obtained, representing 12,813 mRNAs and 22 groups (11 normal samples and 11 pulpitis samples, Fig. 2). A total of 470 DEGs were screened from the two merged microarray platform datasets, including 394 upregulated genes and 75 downregulated genes in pulpitis relative to normal tissue, shown in a volcano plot and heatmap (Fig. 3). Details of the expression matrix and DEGs are provided in Additional files 1 and 2.

Fig. 1
figure 1

PCA results before and after batch effect removal

Full size image
Fig. 2
figure 2

Box plot of data from 22 samples in the pooled dataset that were normalized. The X-axis presents samples from the dataset, and the Y-axis presents normalized intensity values. a Box plot showing the expression profiles of the pooled dataset before normalization; b Box plot showing the expression profiles of the pooled dataset after normalization using the SVA package

Full size image
Fig. 3
figure 3

Volcano plot and heatmap of DEGs. a Volcano plot for the DEGs in the merged dataset. The X-axis shows the -log10 adj.P, and the Y-axis shows the log FC. The DEGs were identified according to the criteria |FC| > 1.0 and adj. P. val < 0.05. The red and green circles denote upregulated genes and downregulated genes in pulpitis, respectively. The grey circles denote genes with no significant difference in expression between pulpitis and healthy tissue. b Heatmap of the DEGs in the merged dataset. Blue and white represent upregulation and downregulation, respectively, of the mRNAs

Full size image

Functional analysis of DEGs

GO BP, CC and MF and KEGG pathway analyses were performed for functional analysis of the DEGs. Regarding BP terms, the DEGs were enriched in inflammatory response, signal transduction, immune response and cell adhesion. Regarding CC terms, the DEGs were enriched in integral component of membrane, plasma membrane, and extracellular exosome. Regarding the MF terms, the DEGs were enriched in protein binding, calcium ion binding, and protein homodimerization activity (Fig. 4, Table 2). The significantly enriched KEGG pathways of the DEGs included cytokine-cytokine receptor interaction, osteoclast differentiation, chemokine signalling pathway, NF-κB signalling pathway, T cell receptor signalling pathway and other pathways related to pulp inflammation (Fig. 5). The detailed results of the GO enrichment and KEGG pathway analyses are provided in Additional files 3 and 4.

Fig. 4
figure 4

Top 10 terms from the GO enrichment analysis, including BP, MF and CC. Red and dots represent the upregulated genes and downregulated genes respectively. The size and colour of the sectors indicate the adjusted adj. P. val and stand score (z-score), respectively, of each GO term

Full size image
Table 2 Supplementary information of GO analysis
Full size table
Fig. 5
figure 5

Top 20 pathways from the KEGG pathways analysis. The X-axis presents the ratio of enriched genes to all of the DEGs, and the Y-axis presents the names of the KEGG pathway terms. The size and colour of the circles indicate the number of genes and -log10(P-Value), respectively, for each enriched term

Full size image

Gene set enrichment analysis

The GSEA results revealed that the enriched biological processes mainly involved the IL2/STAT5 signalling pathway (NES = 1.51; P = 0.008), the IL6/JAK/STAT3 signalling pathway (NES = 1.49; P = 0.006) and inflammatory response pathways (NES = 1.49; P = 0.001), as shown in Fig. 6.

Fig. 6
figure 6

Gene Set Enrichment Analysis results. The red, blue and green lines represent the IL2/STAT5 signalling pathway, IL6/JAK/STAT3 signalling pathway and inflammatory response pathways, respectively. NES: Normalized Enrichment Score

Full size image

PPI network and hub genes

The PPI network constructed with the online STRING program consisted of 3873 edges and 465 nodes (Additional files 5 and 6). Used the MCODE plugin, two significant modules were obtained, one containing 22 nodes and 105 edges (module 1) and the other containing 56 nodes and 645 edges (module 2), as shown in Fig. 7. The hub genes of pulpitis from the PPI network are listed in Table 3.

Fig. 7
figure 7

The hub genes and significant modules identified from the PPI network. a Top 10 hub mRNAs of pulpitis; b Module 1, with 22 nodes and 105 edges; c Module 2, with 56 nodes and 645 edges. The polygonal nodes represent the hub genes, and the round nodes are the other genes in the module network. The size and colour saturation of a node represent the degree of the node in each network

Full size image
Table 3 Top 10 hub genes with higher degrees of PPI network of pulpitis
Full size table

Discussion

It is consistently challenging for clinicians to make accurate diagnoses regarding the level of dental pulp inflammation [4]. As new biological materials are developed, vital pulp therapies are increasingly used, which can increase the fracture resistance and long-time survival of teeth [31]. However, there is currently no accurate diagnostic tool for guiding dental pulp treatment; this shortcoming is an important cause of failure in vital pulp therapy [32]. Since histological examinations show that the extent of inflammation is not correlated with clinical manifestation, traditional diagnostic approaches based on symptoms or electrical/thermal pulp tests, require improvement [9, 33].

During inflammation, cells in human dental pulp, including odontoblasts, macrophages, vascular endothelial cells, precursor cells and other cells that can activate the immune system, secrete large amounts of cytokines, chemokines and neuropeptides, which have been reported to play critical roles in inflammation [34, 35]. Sivakami et al. found that the levels of IL-6 and the cytokine interleukin-1β (IL-1β) were clearly increased in saliva during pulpal and periapical inflammation [36]. An analysis of cytokines in pupal blood revealed that the levels of IL-8 were significantly higher in irreversible pulpitis than in caries-exposed pulp [37]. Several cytokines have been reported to be candidate diagnostic markers of pulpal inflammation [23, 38]. However, to the best of our knowledge, no biomarkers with high degrees of accuracy have yet been used in clinical examination to diagnose early pulpitis.

Research involving microarray analysis is rapidly expanding due to the rapid development of transcriptomic studies, resulting in an increasing understanding of the biological mechanisms underlying oral diseases [39]. Through the integration of several similar datasets, we can generate comprehensive expression profiles to identify key genes that can serve as biomarkers of pulpitis diagnosis and prognosis.

In the current study, expression data of pulpitis tissue from the GEO database were integrated and then used to identify the underlying characteristics of DEGs and candidate biomarkers for diagnosing the inflammation level of dental pulp. Li et al. reported DEGs and enrichment results obtained by analysing data in the GSE77459 dataset [

Conclusions

In summary, the present study analysed merged datasets of pulpitis tissue, aiming to gain insight into methods for the diagnosis or treatment of different inflammatory levels of pulpitis. In addition to investigating the possible regulatory mechanisms of DEGs, we screened key genes as biomarker candidates for the diagnosis of pulpitis, including PTPRC, CD86, CCL2, IL6, TLR8, MMP9, CXCL8 and ICAM1. The use of bioinformatic methods to analyse merged datasets is effective for identifying biomarkers of diseases. The present findings require further validation in future studies.