Identification of crucial salivary proteins/genes and pathways involved in pathogenesis of temporomandibular disorders

2.1 Reagents

Dithiothreitol (DTT), iodoacetamide (IAA), and urea were purchased from Bio-Rad (Bio-Rad Laboratories, Hercules, CA, USA). Formic acid (FA) and ammonium bicarbonate (AB) were purchased from AppliChem (AppliChem GmbH, Darmstadt, Germany), while acetonitrile (ACN) was purchased from Fluka (Honeywell Fluka, Charlotte, North Carolina, USA). Sequence-grade quality-modified trypsin was obtained from Promega (Promega, Madison, Wisconsin, USA). Microcon 0.5 mL centrifugal filter units, Molecular weight cutoff (MWCO) 30 kDa, were purchased from Merck (Merck, Millipore, Billerica, MA, USA). To inhibit the activity of salivary enzymes, a protease inhibitor cocktail (Roche, Basel, Switzerland) was used. For Coomassie dye-binding assays, the BioRad Quick Start Bradford Protein Assay (Bio-Rad Laboratories, Hercules, CA, USA) was used.

2.2 Patients, sampling, and handling procedures

All saliva samples were collected between January 1, 2021 and May 30, 2022 at the Stomatology and Maxillofacial Surgery Clinic, Louis Pasteur University Hospital, Košice, Slovakia. The study included 20 TMD patients (5 men, 15 women) aged between 21 and 75 years. The control group consisted of 20 healthy individuals (7 men and 13 women) aged between 25 and 64 years. Only the newly diagnosed TMD patients and control individuals without any medication were enrolled. Participants were asked to refrain from eating and oral hygiene procedures 2 h prior to saliva collection. To avoid interference from the circadian cycle, saliva samples were collected between 9 a.m. and 11 a.m.

Collection of the whole unstimulated saliva was performed as previously described by spitting into sterile 50 mL Falcon tubes kept on ice [12]. Approximately 3 mL of saliva was accumulated over 5–10 min. To inhibit the activity of salivary enzymes, the saliva was immediately treated with a protease inhibitor cocktail (1:20, v/v). The samples were then centrifuged at 12,000×g for 30 min at 4°C to remove bacteria, cell fragments, and food debris. The supernatant of each sample was aspirated and stored at −80°C until further use.

2.3 Salivary sample processing for LC-MS/MS analysis

The saliva sample was processed following the procedure reported in our previous study [13].

Microcon 0.5 mL MWCO 30 kDa centrifugal filter units were used for filter-added sample preparation. The filter was loaded with a sample of 50 μg salivary proteins and then centrifuged at 14,000×g for 10 min at 20°C until the top of the filter was dried.

Thereafter, the filter was washed with 200 μL of 8 mol L⁻¹ urea in 25 mmol L⁻¹ AB to remove low-weight material and then centrifuged at 14,000×g for 20 min at 20°C.

To reduce proteins, 200 μL of 50 mmol L⁻¹ DTT in 8 mol L⁻¹ urea and 25 mmol L⁻¹ AB was added. The filter unit was incubated for 60 min at 37°C and then centrifuged at 14,000×g for 15 min at 20°C.

Alkylation of cysteine residues was performed by adding 100 μL of 50 mmol L⁻¹ IAA in 8 mol L⁻¹ urea and 25 mmol L⁻¹ AB solution. After incubation in the dark for 45 min at 37°C, the sample was centrifuged at 14,000×g for 20 min at 20°C.

Thereafter, the filter was washed twice with 200 μL of 25 mmol L⁻¹ AB followed by centrifugation at 14,000×g for 5 min each time at 20°C.

Proteolysis was performed by the addition of proteomic grade trypsin dissolved in 12.5 mmol L⁻¹ AB at an enzyme-to-protein ratio of 1:30 (w/w). Proteins were digested overnight at 37°C in a thermoshaker. The digested peptides were collected in a clean centrifuge tube by centrifugation at 14,000×g for 15 min at 20°C, followed by two additional 200 μL washes with deionized water at 14,000×g for 5 min and 16 min at 20°C, and then vacuum-dried at 55°C. The peptides were resuspended in 50 μL of 3% (v/v) ACN containing 0.1% (v/v) aqueous FA. The samples were then homogenized for 5 min on a vortex, followed by 10 min in an ultrasonic bath at 100% ultrasonic amplitude.

2.4 LC-MS/MS analysis and protein identification

For LC-MS analysis, an AmaZon Speed ETD ion trap mass spectrometer (Bruker Daltonik, Germany) coupled to an Ultimate 3000 RSLC NCP system (Thermo Scientific, USA) was used; 2 µL of each sample was injected into an Acclaim PepMap 100 (Dionex, Thermo Scientific, USA) trap column, 100 µm × 2 cm, C₁₈, 5 µm particles with water-ACN loading solvent at a ratio of 2:98, and v/v containing 0.1% FA at flow 8 µL min⁻¹. The peptides were eluted and separated on a home-made capillary column 75 µm × 30 cm, packed with reverse phase C₁₈, 3 µm particles (Magic C18 AQ, Michrom Bioresources, USA). Mobile phases consisted of 0.1% FA in water–ACN 98:2, v/v (A) and 0.1% FA in water-ACN 5:95, v/v (B) operated at a constant flow of 0.4 µL min⁻¹. The gradient is shown in Table 1. Samples were measured in auto MS/MS mode, 10 precursors for 1 MS scan, and only 2+ and 3+ precursors were taken for fragmentation with active exclusion set to 0.5 min. The ICC target was set to 400,000 for MS and MS/MS scans; the maximum accumulation time was 0.050 s for MS and 0.1 s for MS/MS. The isolation window was set to 2.2 Da. The scan range was set to 300–1,300 Da. The search engine Mascot 2.4 (Matrix Science, UK) against the Swiss-Prot database was used to identify proteins. The parameters were determined as follows: taxonomy – Homo sapiens (human), variable modification – oxidation of methionine, fixed modification – carbamidomethylation of cysteine, MS tolerance – 0.6 Da, MS/MS tolerance – 0.6 Da, and the false discovery rate (FDR) threshold was set to 1%. All samples were analyzed in triplicate.

2.5 Data preprocessing and identification screening of DEGs

The MS data were preprocessed in ProlineStudio 2.1.2 (Profi Proteomics, France) to determine weighted spectral counts of identified proteins. These data were subsequently analyzed using Reactome analysis tools using the Camera algorithm [14] to identify DEGs by comparing the expression level of salivary proteins between TMD patients and healthy controls. The criteria for screening DEGs were defined with a cut-off value of p < 0.05 to identify an upregulated group and a downregulated group of proteins, respectively.

2.6 Gene ontology enrichment analysis

The UniProt mapping service (https://www.uniprot.org) was used to convert protein identifiers/accession numbers into gene identifiers. Downloaded matched genes were used for further analysis.

DEGs of salivary proteins identified by LC-MS/MS were subjected to hierarchical gene ontology (GO) overrepresentation tests using the BiNGO plug-in 3.0.5 [15] for Cytoscape 3.9.0 [16]. GO enrichment analysis for three categories of biological process, cell component, and molecular function was determined at a significance level of 0.05 (p-value < 0.05) using the Benjamini & Hochberg correction for multiple testing.

2.7 Protein–protein interaction network and module analysis

The STRING (Search Tool for the Retrieval of Interacting Genes) database of known and predicted PPIs implemented in Cytoscape 3.9.0 was used to generate the DEG network. PPIs from DEGs were selected with confidence (score) cut-off of >0.4. PPIs that were not in the STRING database were manually added to the network using the database https://thebiogrid.org/. Then, the network generated by the StringApp 1.7.0 [17,18] was loaded into Cytoscape to detect hub genes and significant modules using the cytoHubba 0.1 [19] and the MCODE 2.0.0 (Molecular Complex Detection) [20] plug-ins.

The 11 methods using the cytoHubba plug-in algorithms were used. Local-based methods included Maximal Clique Centrality (MCC), Density of Maximum Neighborhood Component (DMNC), Maximum Neighborhood Component (MNC), and Degree method (Deg) algorithms. The network was also analyzed using Global-based methods − Edge Percolated Component (EPC), BottleNeck (BN), EcCentricity (EC), Closeness (Clo), Radiality (Rad), Betweenness (BC), and Stress (Str) algorithms [19].

The MCODE plug-in was applied to search densely intraconnected regions – clusters – and the hub genes with a higher degree of connectivity in a PPI network complex. The criteria for selecting a cluster were as follows: degree cut-off = 2, node score cut-off = 0.2, max depth = 100, and k-score = 2.

2.8 Pathway enrichment analysis of common DEGs

The Reactome pathway database (http://www.reactome.org, release 79) [21] was used to perform enrichment analysis of DEGs to link information from large molecular data sets. The Reactome Camera workflow was adopted for quantitative pathway analysis. The MetScape plug-in (version 3.1.3, [22]) in the Cytoscape software was used for gene enrichment and cluster analysis of metabolic pathways.

4.1 Identification of common up- and downregulated genes/proteins

Based on the in silico analysis by the Reactome analysis tool, a total of 140 genes were identified as DEGs, 91 of which were upregulated and 49 downregulated (Figure 1), and further analyzed. Table 2 shows the top ten upregulated genes/proteins and downregulated genes/proteins in the saliva of individuals with TMD compared to those in the control group.

Figure 1

Volcano plot of differentially expressed proteins from samples from TMD patients and a healthy control group, visualizing the fold-change (x-axis) and statistical significance (y-axis). The red line corresponds to p = 0.05 and the blue line to p = 0.01.

4.2 GO identification by gene/protein set enrichment analysis

To understand the representation of gene and gene product attributes in three non-overlapping areas of molecular biology, a gene ontology (GO) analysis of DEGs in TMD patients and the healthy control group was performed. A series of enrichment analyzes were performed to obtain detailed information on biological process, molecular function, and cell components mapping GO terms within specific gene/protein sets using the BiNGO plug-in for Cytoscape. GO enrichment revealed that the downregulated genes were involved in 102 GO terms, including 54 terms related to biological processes, 46 terms related to molecular functions, and 2 terms related to cell components. The upregulated genes were involved in 204 GO terms, including 137 terms associated with biological processes, 48 terms associated with molecular functions, and 59 terms associated with cell components. The top terms of GO enrichment related to DEGs are presented in Figure 2 and Table 3. The upregulated DEGs were mainly associated with the extracellular space and cytoplasm (Figure 2 and Table 3), and the central molecular function of these genes was found to bind to proteins and regulate enzyme activity. Among the upregulated DEGs, numerous genes were implicated in catabolic processes involving simple carbohydrates. The downregulated DEGs, which are also located in the extracellular area, seemed to be particularly involved in the defense and immune response of the organism.

Figure 2

Gene ontology analysis of DEGs of three ontologies: biological process, molecular function, and cell component. The x-axis represents the number of up/downregulated genes/proteins. The y-axis represents the GO term.

4.3 Protein–protein interaction network and network cluster analysis

A PPI was constructed based on 140 identified DEGs with combined scores greater than 0.4. The network formed from the corresponding 142 genes and 2333 edges is shown in Figure 3. Topological metrics of the network is summarized in Table 4.

Figure 3

Identification of DEGs and construction of the TMD-related PPI. According to the node fill color mapping, the red color denotes the upregulated genes, and the green color stands for the downregulated genes; edges represented by lines are gray. The color intensity corresponds to the fold-change.

The complex network topology was examined with the cytoHubba plug-in integrated into Cytoscape3.9.0. The highly interconnected hub genes have been filtered out of the main network of DEGs. The network was analyzed with each of the 11 algorithms embodied in cytoHubba. Then, 22 genes found in the intersection of at least four methods were sequentially ordered as follows: ACTB, ALB, APOA1, B2M, CAT, CLU, CTSD, ENO1, GAPDH, GSN, HBB, HP, HSPA8, LYZ, MMP9, S100A9, SERPINA1, TF, TXN, C3, LTF, and TPI1 (Table 5).

Figure 4 shows the network of key genes with high binding degree identified by the MNC algorithm of the cytoHubba plug-in in the set of DEGs.

Figure 4

Module analysis of the protein-protein interaction by the cytoHubba plug-in. The PPI hub gene module of the top 20 genes ranked by MNC algorithm.

In addition to the previous clustering method, the MCODE plug-in in Cytoscape 3.9.0 was used to identify the most densely connected regions. The MCODE application enabled the identification of 10 intraconnected regions/clusters, and the nine hub genes – seeds, with high-level connectivity in a PPI network complex of DEGs (Figure 5 and Table 6).

Figure 5

Top two cluster networks generated from the complex PPI network by the Molecular Complex Detection (MCODE) plug-in in Cytoscape. The nodes are represented by circles (green – unclustered, red – cluster 1, and pink – cluster 2) and edges as lines (gray).

By the conjunction of the results of the two top-scoring modules of the MCODE clustering (cluster 1 and cluster 2) with the results of the cytoHubba analysis, a total of 20 hub genes including ALB, APOA1, B2M, C3, CAT, CLU, CTSD, ENO1, GSN, HBB, HP, HSPA8, LTF, LYZ, MMP9, S100A9, SERPINA1, TF, TPI1, and TXN were selected.

4.4 Pathway enrichment analysis

To map the signaling pathways of TMD, the associated genes of all down- and upregulated proteins were uploaded to the Reactome pathways database. As seen in Table 7 and Figure 6, the downregulated proteins were significantly enriched in the top five different signaling pathways including the R-HSA-6803157 antimicrobial peptides pathway, while the upregulated proteins were enriched in several different signaling pathways, including R-HSA-70171 glycolysis, R-HSA-70263 gluconeogenesis, and R-HSA-71387 metabolism of carbohydrates.

Figure 6

Volcano plot of differentially expressed Reactome pathways from TMD patient samples and a healthy control group, visualizing the fold-change (x-axis) and statistical significance (y-axis). The red line corresponds to p = 0.05, and the blue line p = 0.01.

From cluster analysis, we identified a set of 20 hub genes that were significantly modulated in TMD (Figure 4). These genes were further analyzed by MetScape. Two enriched metabolic pathways, glycolysis, and gluconeogenesis, and tryptophan metabolism pathways were identified, including the upregulated hub genes CAT, ENO1, and TPI1.

The rest of the hub genes were mainly enriched in the innate immune system and antimicrobial peptides taking into account the number of entities found. Among the signaling pathways, the p-value of the neutrophil degranulation pathway was the lowest mapped on 13 Reactome entities including upregulated B2M, C3, CTSD, GSN, HP, HSPA8, SERPINA1, TF and downregulated LTF, MMP9, S100A9, HBB, and LYZ.

5.1 Upregulated hub proteins/genes involved in the hub protein network

Beta-2-microglobulin (B2M − protein-encoding gene) is an LMW protein (17 kDa) synthesized in all nucleated cells including immunocompetent cells such as macrophages, active T and B lymphocytes, and salivary gland epithelial cells [53]. Beta-2-microglobulin associates with the major histocompatibility complex I-related gene protein and forms the light chain subunit of the class l human leukocyte antigen on the surface of all nucleated cells [54].

It is well documented that the pathologically high serum concentration of beta-2-microglobulin, which leads to protein aggregation, has the amyloidogenic potential [55,56,57,58]. Amyloid deposition in skeletal joints, bones, and muscles leads to microarchitectural deterioration of bone tissue and movement impairment [57]. In addition, along with transferrin, beta-2-microglobulin is a known regulator of cartilage regeneration and differentiation. There is evidence that beta-2-microglobulin is highly expressed in cartilage and synovial fluid of patients with OA compared to control [59]. Since OA may be localized in the TMJ, elevated beta-2-microglobulin levels may also impair condylar chondrocyte function, which could contribute to TMD pathogenesis [60]. At the same time, beta-2-microglobulin is a recognized indicator of inflammatory disease activity [61]. The author of the study points to the association of immune system, inflammation, and metabolic deregulation with the onset symptoms of OA with elevated beta-2-microglobulin in chondrocytes. The positive correlations between serum and synovial fluid protein levels in RA patients have also been reported [62,63].

In addition, beta-2-microglobulin, including modified beta-2-microglobulin, is of significant importance in innate defenses with antimicrobial activity against a variety of microorganisms, including Streptococci [64]. Therefore, in the future, it would be interesting to investigate the connection between the occurrence of TMD and possible changes that have occurred in the microbial composition of the oral cavity or the microbiome as such.

Complement C3 (C3 − protein-encoding gene) is an essential protein of the immune system. This protein affects innate and adaptive immunity and plays a crucial role in immunosurveillance and homeostasis in the human body. It can cause the system to malfunction if it becomes dysregulated or over-activated, such as in several inflammatory diseases.

The complement system is believed to be actively involved in the pathogenesis of OA. It is produced and activated in the OA joint with the involvement of cartilage, bone, and synovium [65,66]. The complement system promotes inflammation of the joint tissue, which manifests itself in the degradation of the cartilage and the release of components into the synovial fluid [67]. Activation of complement C3 as part of the complement system can be triggered in cartilage cultured by interleukin-1 alpha, as the study authors have shown [68]. In addition, a mass spectrometry study of the synovial fluid showed that several other pro-inflammatory cytokines, namely, interleukin-6, interleukin-8, and interleukin-18, play a significant role in the etiology and progression of OA [69], which is present in 10–17% of patients with TMJ pain [70].

Cathepsin D (CTSD − protein-coding gene) is a lysosomal proteinase. Cathepsin D released from lysosomes is involved in the digestion of surrounding tissues and also plays a role in the processing of protein antigens for presentation to the immune system [71].

The mechanism of the degenerative processes in the joint triggered by catabolic enzymes has not yet been clarified. However, there is some work confirming the involvement of cathepsin D concomitantly with interleukin-1 in the modulation of terminal complement complex formation in cultured human disc tissue [72] or local changes within the osteoarthritic joint [73].

Gelsolin (GSN − protein-coding gene) is a calcium-regulated, actin-modulating protein that can promote assembly and disassembly of actin filaments. Gelsolin is involved in cell formation, metabolism, and wound healing processes. The protein has been found to be associated with joint homeostasis in various types of arthritis and has positive anti-inflammatory and chondroprotective effects [74,75].

In the proteomic analysis of mouse interleukin-1 alpha and retinoic acid, which stimulated cartilage degradation and protein release, the authors of the study identified gelsolin as one of the potential biomarkers for cartilage deterioration. A particularly novel finding was that the modified expression of gelsolin in cartilage leads to reduced concentrations of gelsolin in explant media [76].

Haptoglobin (HP − protein-coding gene) belongs to the family of acute-phase plasma proteins that play a role in modulating many aspects of the acute-phase response and also exhibit antibacterial activity. An increase in haptoglobin plasma levels has been found to be associated with carcinogenesis, infection, and chronic inflammation or autoimmune inflammatory rheumatic disease [77,78].

However, there is not much data correlating the risk of joint defects with haptoglobin expression in cartilage, synovial fluid, or plasma. One of the few studies states that porcine articular cartilage homogenates with increased haptoglobin concentrations have higher (p < 0.05) zymographic activities of the active form of MMP-2, but there is no association with degenerative processes in the cartilage [79].

HSPA8 protein (HSPA8 − protein-coding gene) is a cognate protein of the HSP70 family that is upregulated during cellular stress states such as inflammation, infection, and cancer [80]. HSPA8 plays a key role in the presentation of peptide antigens to CD4⁺ T cells, which regulates T- and B-cell activation and antibody secretion by plasma cells [81]. According to the researchers [82], HSPA8 is persistently expressed in the nucleus pulposus tissues of the human intervertebral disc and its expression correlates with the degree of disc degeneration. On the other hand, the upregulation of heat shock protein 70 may also contribute to the robustness and active matrix production of articular cartilage chondrons [83].

Using in-gel-based label-free ESI-LC-MS/MS analysis, bacteria-responsive host-defense proteins/pathways leading to inflammation were identified and compared between non-degenerated and degenerated discs. Stress-responsive and antioxidant proteins − HSPA8 protein, peroxiredoxins-1,2,6, catalase representing the host response to inflammaging – have been specifically expressed in degenerated discs [84]. HSPA8 has also been identified as one of the key hub genes correlating with RA onset and progression [85].

Alpha-1-antitrypsin (SERPINA1 − protein-coding gene) is a member of the serine protease inhibitor superfamily. A high transcript-level expression of SERPINA1 gene has been identified predominantly in superficial cartilage zones, as well as high levels of the alpha‐1‐antitrypsin protein have been found in synovial fluid [86,87].

Alpha-1-antitrypsin is an acute phase reactant that protects tissues from enzymes released during inflammation with highest affinity for neutrophil elastase [88,89]. Authors of the studies [74,90] also presented the properties of alpha‐1‐antitrypsin in the context of cartilage protection, joint inflammation, and manifestations of associated pain. In ex vivo analyzes of arthritic joints, they showed that alpha‐1‐antitrypsin promoted transcription of collagen alpha-1(II) chain, aggrecan core protein, and transcription factor SOX-9, as well as downregulated MMP-13, disintegrin, and metalloproteinase with thrombospondin motifs 5.

On the other hand, the concentration of alpha-1-antitrypsin activity inversely correlates with the levels of the residual neutrophil elastase − proteoglycan-degrading enzyme that has historically been associated with inflammatory arthritis [91]. Bioinformatic analyzes showed that OA cartilage exhibited intense alpha-1-antitrypsin staining in the most damaged areas of the cartilage due to the infiltration of greater amounts of alpha-1-antitrypsin from synovial fluid after degradation of the proteoglycan matrix [92].

Catalase (CAT − protein-coding gene) is a key antioxidant enzyme in the metabolism of H₂O₂ and reactive nitrogen species. The subcellular localization of catalase is mainly peroxisomal, but a cytosolic catalase can bind cytosolic proteins or be localized to the cytoplasmic membrane involved in the protection of important cellular elements (i.e., proteins and chromosomes) from oxidative damage [93].

Oxidative stress due to an imbalance between pro-oxidants and antioxidants can activate antioxidant defenses, leading to differential expression of some genes involved in inflammatory pathways [94]. The hypothesis that oxidative stress may be involved in the pathogenesis of joint diseases and lead to TMD was the subject of a recent authors’ study examining oxidative stress biomarkers in the saliva of TMD patients. Although they had significantly higher malondialdehyde levels, there were no significant differences in TMD salivary catalase levels compared to healthy controls [95]. Likewise, no correlation was observed between oxidative stress markers, including catalase, in the TMD subjects and the control group [96]. On the other hand, there was a decrease in catalase activity in the oral fluid of children [97].

Enolase 1 (ENO1 − protein-coding gene) is a multifunctional enzyme expressed in the cytoplasm of prokaryotic and eukaryotic cells and on the surface of several cell types, where it acts as a plasminogen receptor [98,99]. Enolase 1 mediates activation of plasmin, degradation of the extracellular matrix, and induces a specific humoral and cellular immune response [100]. The enzyme has been detected in the synovial fluid of inflamed joints [101] and in the serum [102] of RA patients. Enolase 1 stimulates the production of pro-inflammatory mediators from monocytes and macrophages isolated from RA patients via the p38-MAPK and NF-kappa-B signaling pathways [103,104].

Triosephosphate isomerase 1 (TPI1 − protein-coding gene) is a key enzyme in the glycolysis and gluconeogenesis signaling pathway that is found in almost all cell types. Furthermore, it has been shown that triosephosphate isomerase 1 is an RA-related metabolic regulator capable of catalyzing the interconversion of dihydroxyacetone phosphate and D-glyceraldehyde-3-phosphate and balancing glycolysis and gluconeogenesis. Overexpression of triosephosphate isomerase in macrophages from inflamed joints facilitates macrophage polarization to the pro-inflammatory M1 phenotype by promoting glycolysis [105].

The enzyme was also identified by SILAC and nLC-MS/MS as one of the chondrogenesis-regulated proteins in mesenchymal stromal cells of osteoarthritic human bone marrow [106].

The intercellular communication between chondrocytes and osteoblasts plays a new and crucial role in the metabolic homeostasis of the bone-cartilage unit. The experiments on a non-contact co-culture model show that osteoblasts trigger increased ATP generation in chondrocytes via an energetic shift represented by enhanced glycolysis and an impaired mitochondrial tricarboxylic acid cycle. Intensified glycolysis is also reflected in the upregulation of glycolytic enzymes, including triosephosphate isomerase 1 [107].

Thioredoxin (TXN − protein-coding gene) is a component of the thioredoxin system, which forms redox-dependent signaling pathways critical to fundamental cellular processes, including metabolism, the redox processes associated with oxidative stress, and the regulation of nitrosative stress [108].

Salivary oxidative stress profiles between patients with myogenic TMD and healthy subjects were analyzed to reveal associations between total antioxidant capacity and total oxidative status with clinical features of the disease. While total antioxidant capacity scores were significantly higher, total oxidative status scores decreased in myogenic TMD patients compared to those in controls [109]. The results confirmed data from the authors’ previous proteomic studies of the whole stimulated saliva from TMD myalgia patients. Statistical analysis of the quantitative proteomics data revealed 20 differentially regulated proteins, including the antioxidant proteins thioredoxin and S100-A8 [110].

The truncated form of thioredoxin is a potent mitogenic cytokine. Pro-inflammatory cytokines interleukin-1 beta and tumor necrosis factor or H₂O₂-induced increase in the truncated form of thioredoxin released from synoviocytes in RA patients may implicate a link between inflammation and the immune system in RA [111].

Albumin (ALB − protein-coding gene) Bovine serum albumin is widely used as an agent capable of inducing TMJ inflammation in model systems [112,113]. By comparing essential proteins in the serum of TMD patients and healthy controls, a standard blood sample analysis revealed that TMD patients had, among other things, significantly higher serum levels of albumin [114]. Similarly, a remarkable association of serum albumin with TMD was found in patients on chronic hemodialysis, in whom increased risk factors for the development of TMD were observed [115]. Furthermore, the analysis of the redox states of albumin in the serum and in the synovial fluid of patients with TMD showed that the oxidized fraction of albumin in the synovial fluid is expressed significantly higher than that of healthy subjects. On this basis, the authors of the study hypothesized that synovial albumin might play a scavenging role against intra-articular oxidative stress [116].

Apolipoprotein A-I (APOA1−protein-coding gene) is the major protein component of high-density lipoprotein in plasma [117]. Apolipoprotein A-I, possibly derived from adipose tissue, may be involved in the pathologic process within the joint [118]. Pro-inflammatory properties of apolipoprotein A-I have been confirmed in vitro in human joint cells isolated from cartilage and synovial membrane of OA patients. A dysregulated lipid profile in the synovial fluid of OA patients has been correlated with apolipoprotein A-I-triggered release of inflammatory parameters such as interleukin-6, MMP-1, and MMP-3 [119].

5.2 Downregulated proteins/genes involved in hub protein network

Clusterin (CLU − protein-coding gene) is present in all fluids of the organism except the intracellular matrix; it is also produced in the articular cartilage and synovium [120]. Numerous functions of clusterin include neuroprotection, cardioprotection, and pain modulation. The protein is also associated with inflammation [121]. While the intracellular form of clusterin can suppress stress-induced apoptosis, the secreted form of the protein has cytoprotective functions in osteoarticular tissues [122]. Suitable sources of the secreted form of clusterin are synovial and systemic fluids [123].

Serotransferrin (TF − protein-coding gene) is an iron-binding protein that primarily transports iron from storage sites to regions of iron metabolism. The protein belongs to the acute phase proteins with antimicrobial activity, capable of preventing the adhesion of gram-positive and gram-negative bacteria to surfaces. Transferrin levels in serum drop during, e.g., infection and inflammation [124].

As a traditional acute-phase protein, differentially expressed serotransferrin can indicate an elevated inflammatory status. Serum protein profiles in RA examined using two-dimensional differential gel electrophoresis and mass spectrometry identified several differentially expressed serum proteins between patients with RA and healthy controls. Among them, serotransferrin was significantly downregulated in RA patients [125]. Similarly, serotransferrin was downregulated in early OA sets compared to the control, as identified by a quantitative proteomic ¹⁸O labeling approach [126].

Lactotransferrin (LTF − protein-coding gene) is an iron-binding multifunctional cationic glycoprotein. As a key element of host defense, lactotransferrin maintains various biological functions as it is an anti-inflammatory, antibacterial, antiviral, antioxidant, antitumor, and immunomodulatory agent [127,128]. Numerous research data related to such functions of lactotransferrin as the modulation of regenerative processes in bone or cartilage [128]. Lactotransferrin has the potential to inhibit chondrocyte apoptosis, which is responsible for cartilage degeneration [129], and to regulate chondrocyte metabolism [130]. When released into the synovial fluid, lactotransferrin allegedly facilitates cartilage remodeling by promoting the degradation of cartilage material [131].

Matrix metalloproteinase-9 (MMP-9 − protein-coding gene)

MMPs form a family of calcium-dependent and zinc-dependent endopeptidases. Increased proteolytic activity of MMP, particularly MMP-9, induces degradation of organic matrix components of bone and erosion of cartilage and is closely associated with the inflammatory response [131]. Overexpressed MMP-7 and MMP-9, which are involved in extracellular matrix homeostasis and articular disc remodeling, were found in the synovial tissue of patients with anterior disc displacement of the TMJ without reduction using an immunohistochemical approach [131]. Recent advances in understanding tissue-degrading MMP-9 levels indicated that the upregulated enzyme can be found in the serum, synovial fluid, cartilage, and synovial and subchondral bone tissues of patients with OA or RA, in contrast to controls [7,132,133,134,135,136].

TMDs are defined as musculoskeletal conditions involving acute or chronic orofacial pain that can be coupled with inflammatory processes in the TMJ. The authors of the current study investigated how the production and activity of MMPs, including MMP-9, are regulated by the limbic system during persistent orofacial pain-induced temporomandibular inflammation [137]. Normally, MMPs are downregulated in the tissue by the four specific metalloproteinase inhibitors [131]. Interestingly, in our experiments, the MMP-9 expression found in the saliva of TMD patients was slightly reduced compared to healthy controls, but the level of metalloproteinase inhibitor 1 expression in the TMD saliva samples was also reduced.

Protein S100-A9 (S100A9 − protein-coding gene) is a calcium- and zinc-binding protein of the S100 protein family involved in various biological processes. Its differential expression is associated with both intra- and extracellular environments [138].

Protein S100-A9 is predominantly expressed by neutrophils, monocytes, and activated macrophages involved in pathological processes including inflammation. The actively released protein S100-A9 modulates the inflammatory response by stimulating leukocyte recruitment and inducing cytokine secretion [139,140]. Protein S100-A9 molecules are supposed to act as damage-associated molecular patterns; however, the mechanism of their action has not yet been fully elucidated [141].

Protein S100-A9 is known to affect joint integrity. Stimulation of osteoarthritic chondrocytes with protein S100-A9 (and S100-A8) induced significant upregulation of catabolic markers (MMP-1, MMP-3, MMP-9 and MMP-13, interleukin-6, interleukin-8, and monocyte chemoattractant protein-1) and downregulation of anabolic markers (aggrecan core protein and collagen alpha 1 type II), which can thus stimulate cartilage degradation [142]. Upregulated proteins of most S100 family proteins are closely related to RA pathogenesis [143] and are a marker of disease activity in RA and systemic juvenile idiopathic arthritis [144,145].

Nevertheless, in our experiments, all proteins from the S100 protein family including S100-A9 found in the saliva of TMD patients were downregulated compared to healthy control.

Hemoglobin subunit beta (HBB − protein-coding gene) is an essential component of the hemoglobin α₂β₂ heterotetramer, which is particularly known as the iron-containing protein essential for O₂ transport in mammals [146]. Furthermore, the antimicrobial activity of a group of peptides produced by proteolytic cleavage of hemoglobin and acting as a second line of defense in enhancing the innate immune system of organisms has been observed [147]. As established by in silico analysis, the abundance of peptides derived from antimicrobial regions of hemoglobin subunit beta can contribute to an antimicrobial environment of infected wounds [148]. Human hemoglobin-derived peptides inhibit the growth of microbial invaders and reduce the biological activity of endotoxins through lipopolysaccharide binding [149].

A transcriptomic profile of the masticatory muscles of TMD patients with migraine compared to controls revealed a predominant downregulation of genes involved in the myogenetic process. Among them, a downregulation of hemoglobin subunit beta was registered [150]. The researchers concluded that the muscle regeneration process may be slowed down in TMD patients with migraine due to a decrease in muscle strength that occurs in the presence of anemia [151].

Lysozyme C (LYZ − protein-coding gene) is a 14 kDa protein expressed in mucosal secretions (tears, saliva, and mucus) and tissues. Lysozyme, whose natural substrate is the bacterial cell wall peptidoglycan, is an important component of innate immunity. Simultaneously with its direct antimicrobial role, lysozyme modulates the host’s immune response to infection [152]. At the same time, compared to the control group, a reduced activity of lysozyme and antimicrobial protection, as well as an increase in the degree of contamination with pathogenic and conditionally pathogenic microflora, was observed [97,153]. The downregulation of lysozyme in oral fluid suggests the possibility of a host immune response to infection rather than a direct effect on the joint [154].

Candidate biomarkers for the diagnosis of TMD were discovered and analyzed by quantitative proteomic analysis. A computational bioinformatic approach enabled the determination of multiple hub proteins/genes ALB, APOA1, B2M, C3, CAT, CLU, CTSD, ENO1, GSN, HBB, HP, HSPA8, LTF, LYZ, MMP9, S100A9, SERPINA1, TF, TPI1, and TXN from the set of differentially expressed proteins in TMD compared to healthy control. In addition, the construction of PPIs enabled a deeper understanding of the underlying molecular mechanisms of TMD and offered the opportunity to improve therapeutic strategies in TMD. Our analysis provided important information on signaling pathway changes involved in TMD progression, including upregulated glycolysis, gluconeogenesis, and metabolism of carbohydrates, as well as downregulated innate immune system and antimicrobial peptide signaling pathways. The results suggest that the identified hub proteins could be biomarkers for TMD after experiments were completed to determine the full extent of metabolic networks and colonization of oral bacteria in salivary secretion and synovial fluid in TMD.