Abstract
Introduction
Nonsteroidal anti-inflammatory drugs (NSAIDs) are the primary treatment for osteoarthritis (OA), but prolonged use has adverse effects and varying efficacy. Among NSAIDs, imrecoxib, a selective cyclooxygenase-2 (COX-2) inhibitor, reduces side effects yet remains ineffective for half of the patient population. This study aims to identify biomarkers for early evaluation of imrecoxib efficacy in OA for personalized therapy optimization.
Methods
From September 2021 to January 2022, imrecoxib was administered to patients with OA at Nanjing Drum Tower Hospital. Plasma samples from these patients underwent proteomic analysis through the four-dimensional data-independent acquisition (4D-DIA) method, followed by bioinformatics analysis. Potential differentially expressed proteins (DEPs) were validated using enzyme-linked immunosorbent assays (ELISA).
Results
Sixty-six patients with knee OA were included and divided into responders (n = 35) and non-responders (n = 31). Proteomic analysis was conducted on 15 patients from each group, with ELISA validation for every patient. We found 140 DEPs between the two groups after imrecoxib treatment, characterized by 29 proteins showing upregulation and 111 displaying downregulation (P < 0.05, fold change > ± 1.2). Galectin-1 (LGALS1), galectin-3 (LGALS3), and cluster of differentiation 44 (CD44) were identified as potential markers for evaluating clinical response to imrecoxib in OA following ELISA validation.
Conclusion
This study successfully identified biomarkers for evaluating imrecoxib’s clinical response in patients with OA using 4D-DIA technology. These biomarkers may play a vital role in future personalized OA treatment strategies, pending further confirmation.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Why carry out this study? |
There are individual differences in the efficacy of imrecoxib. Objective indicators for evaluating the early-stage effectiveness of imrecoxib in treating osteoarthritis (OA) are currently lacking. |
Can proteomic analysis, along with clinical studies, reveal potential indicators of imrecoxib efficacy? |
What was learned from this study? |
Three novel biomarkers, galectin-1, galectin-3, and cluster of differentiation 44, were identified, which may serve as potential indicators for assessing the early efficacy of imrecoxib in OA treatment. |
These findings may contribute to the development of personalized treatment strategies for patients with OA using nonsteroidal anti-inflammatory drugs. |
Introduction
Osteoarthritis (OA) is a degenerative joint condition that severely impairs patients’ quality of life, leading to joint pain and dysfunction [1]. Currently, the cumulative number of patients with OA is about 500 million worldwide, and this number continues to increase with the aging of the population [2]. OA has developed into a significant public health concern, imposing a considerable burden on both sufferers and healthcare resources.
The cornerstones of OA treatment today include nonpharmacologic, pharmacologic, and surgical approaches. Among these, pharmacotherapy occupies an important place in clinical practice, consisting of nonsteroidal anti-inflammatory drugs (NSAIDs), corticosteroid injections, and some adjuvant medications [3]. NSAIDs, as the first-choice treatment for OA, offer considerable therapeutic advantages to patients by blocking the activity of the enzyme cyclooxygenase (COX), which lowers the production of prostaglandins, an inflammatory mediator, and lessens inflammation and pain [4]. However, the effectiveness of NSAIDs varies significantly among patients, and some individuals do not respond to NSAID therapy. Additionally, prolonged NSAID use is linked to various adverse reactions, including gastrointestinal damage, cardiovascular risks, and renal issues [5, 6]. As a result, NSAID optimization is a crucial clinical concern.
Imrecoxib, a novel selective COX-2 inhibitor, was approved in China in 2011 for the treatment of OA and has comparable efficacy and reduced rates of adverse reactions compared to conventional NSAIDs [7,8,9,10]. However, clinical practice has revealed that individual efficacy differences still exist in treating OA, with a < 50% effective rate for a whole course of therapy [11, 12]. Therefore, it is anticipated that the discovery of biomarkers capable of precisely detecting the early efficacy of imrecoxib in treating OA will enable patients to get more customized treatment plans and improve disease management and efficacy assessment.
Mass spectrometry (MS)-based proteomics technologies are promising for a wide range of applications in biomarker discovery and disease mechanism studies [13,14,15]. Among them, data-independent acquisition (DIA) technology is an emerging MS analytical technique capable of thoroughly and extremely sensitively detecting and quantifying proteins in proteomics samples and has higher data acquisition efficiency, reproducibility, and quantitative accuracy compared to conventional data-dependent acquisition (DDA) [16]. Four-dimensional (4D) proteomics, which adds ion mobility separations to classical three-dimensional (3D) proteomics (consisting of retention time, m/z, and intensity), demonstrates greater scanning speed and detection sensitivity. The 4D-DIA combines the advantages of 4D proteomics and DIA technology, which significantly boosts data integrity and increases the sensitivity and depth of detection [17]. The application of 4D-DIA technology provides a powerful tool for parsing complex biologic samples and facilitates the discovery and analysis of drug-related biomarkers.
To the best of our knowledge, there are no biomarkers to assess the early efficacy of imrecoxib in treating OA. In our pursuit of identifying biomarkers associated with variations in imrecoxib efficacy for OA treatment, we utilized 4D-DIA technology to conduct proteomic analysis on patients with knee OA undergoing imrecoxib treatment. We then quantitatively validated these biomarkers using enzyme-linked immunosorbent assays (ELISA). It is anticipated that a thorough investigation of these markers will uncover novel approaches to NSAID individualization and offer crucial direction and decision-making support to enhance therapeutic outcomes in OA.
Methods
Patients and Samples
This study was conducted at a single center, following the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [18], and was approved by the Ethics Committee of the Nanjing Drum Tower Hospital (2021-402-01). The entire trial was conducted under strict adherence to the Good Clinical Practice Guidelines of the International Council of Harmonization and the Declaration of Helsinki. Between September 2021 and January 2022, participants were recruited from the Joint Surgery Department’s outpatient clinic at Nanjing Drum Tower Hospital. Imrecoxib (manufactured by Jiangsu Hengrui Pharmaceuticals Co., Ltd.) was administered orally to patients at a standard dose of 100 mg twice daily for 4 weeks. Inclusion criteria were patients who consented to participate in the study, aged between 18 and 65 years. They had been diagnosed with knee OA following the American College of Rheumatology (ACR) knee OA guidelines [19]. They were classified as Kellgren-Lawrence (KL) grade I, II, or III based on x-ray changes, required to undergo NSAID treatment, and anticipated the continuous use of NSAIDs for a minimum duration of 4 weeks. Exclusion criteria were pregnant and lactating women; patients with active gastrointestinal ulcer/bleeding, or a history of recurrent ulcer/bleeding; those with a history of gastrointestinal bleeding or perforation, or asthma, urticaria, or allergic reactions following NSAID use; patients with moderate to severe dysfunction of the liver and kidneys; and patients who had used NSAIDs or other analgesics within one week prior to enrollment.
Clinical efficacy evaluation was conducted both before and after 4-week imrecoxib treatment [20]. According to the visual analog scale (VAS) and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) improvement rates, patients were categorized as responders or non-responders [21]. The VAS improvement rate was calculated as (VAS0 week–VAS4 week)/VAS0 week *100%. The WOMAC improvement rate was defined as (WOMAC0 week–WOMAC4 week)/WOMAC0 week *100%. Patients with VAS improvement rate and WOMAC improvement rate ≥ 25% were included as responders and patients with VAS improvement rate and WOMAC improvement rate < 25% were included as non-responders. Patients with ambiguous outcomes (with either VAS or WOMAC improvement rate < 25% or ≥ 25%) were excluded from the analysis.
Plasma samples were collected from patients before and after 4-week treatment, resulting in four distinct groups of plasma samples. Specifically, 1Y represents pre-treatment plasma samples from the responders, 1W represents pre-treatment plasma samples from the non-responders, 2Y represents post-treatment plasma samples from the responders, and 2W represents post-treatment plasma samples from the non-responders. After centrifugation at 3000 rpm for 5 min at 4 °C, the samples were kept at − 80 °C for subsequent proteomic analysis and ELISA confirmation.
Proteomic Analysis
We randomly selected 15 samples each from a pool of 35 responders and 31 non-responders for proteomic analysis.
Sample Preparation
The high-abundance proteins of plasma samples were processed with High Select™ Top14 Abundant Protein Depletion Mini Spin Columns (Thermo Fisher Scientific, USA). Subsequently, the low-abundance components were desalted and concentrated using a 5-kDa ultrafiltration membrane (Sartorius, Germany). Next, one volume of SDT buffer (4% sodium dodecyl sulfate and 100 mM Tris–HCl, pH 7.6) was added, and the mixture was boiled for 10 min. Afterward, the solution was centrifuged at 14,000 g for 15 min to remove the precipitate. Protein quantification was performed using a BCA protein assay kit (Beyotime, China). Each sample, containing 20 µg of proteins was added to 6X loading buffer (Beyotime, China) and boiled for an additional 5 min. Finally, the quality of protein separations was evaluated using 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE, Beyotime, China) and stained with Coomassie Blue R-250 (Beyotime, China).
Filter-Aided Sample Preparation (FASP) and Digestion
The FASP method was employed for protein digestion [22]. First, 200 µg of protein solution (from sample preparation) for each sample was mixed with dithiothreitol (DTT, Sigma, USA) to a final concentration of 100 mM. The mixture was then boiled for 5 min and allowed to cool to room temperature. Subsequently, 200 µl urea (UA) buffer (BIO-RAD, USA) was added, and the solution was transfered to 30-kDa ultrafiltration units (Sartorius, Germany), followed by centrifugation at 12,500 g for 25 min (this step was repeated twice). The samples were then mixed with 100 µl iodoacetamide (IAA, Sigma, USA) buffer (100 mM IAA in UA) and incubated in the dark for 30 min. Afterward, the samples underwent centrifugation (12,500 g, 25 min) and were washed twice with UA buffer. Next, 100 µl of 0.1 M triethylammonium bicarbonate (TEAB, Thermo Fisher Scientific, USA) was added, followed by two rounds of centrifugation (12,500 g, 15 min). The samples were digested with 40 µl trypsin buffer (4 µg trypsin in 40 µl 0.1 M TEAB solution) at 37 °C for 16–18 h. After digestion, we centrifuged (12,500 g, 15 min) the solution, and the peptides as filtrates were collected after adding 40 µl of 0.1 M TEAB solution and subsequent centrifugation (12,500 g, 15 min). The peptides were desalted with a C18 cartridge (Evosep, Denmark) and reconstituted in 40 µl of a 0.1% formic acid solution. Finally, the peptide content was estimated with UV light spectral density at 280 nm.
Spectral Library Generation
Pooded peptides from all samples were separated using the Agilent 1260 Infinity II high-performance liquid chromatography (HPLC) system. Initially, the chromatographic column was equilibrated with buffer A (10 mM HCOONH4, 5% ACN, pH 10.0), and then the samples were loaded onto the chromatographic column (Waters, XBridge Peptide BEH C18 Column, 130 Å, 5 µm, 4.6 mm × 100 mm). Peptides were eluted at a flow rate of 1 ml/min with a gradient of buffer B (10 mM HCOONH4, 85% ACN, pH 10.0). The fractions were collected every 1 min for 26–62 min. Thirty-six fractions were connected, reconstituted in a 0.1% formic acid aqueous solution and combined into six fractions. The DDA method was employed to generate a spectral library. The six fractions were diluted to 10 ng/ul with 0.1% FA, mixed with iRT. For each sample, 1 μg of peptide segments was desalinated with Evotips. Then, these samples were separated using a nanoliter flow rate Evosep One system (Evosep, Denmark), which was coupled to a mass spectrometer Orbitrap Exploris 480 mass spectrometer (Thermo Fisher Scientific, OE 480) equipped with FAIMS PRO. Buffer solution A was 0.1% formic acid aqueous solution, and solution B was 0.1% formic acid acetonitrile solution (acetonitrile was 100%). Peptides were separated on a 15 cm × 150 μm analytical column with 1.9-μm C18 beads with a packed emitter tip (Evosep, Denmark). The column temperature was maintained at 55 °C. The 30 SPD (44 min gradient) method was used for separation, followed by MS analysis with the Orbitrap Exploris 480.
The MS settings were as follows: The field asymmetric ion mobility spectrometry (FAIMS) was set at − 45 V and − 65 V. The MS scan range (m/z) was 350–1500 at a resolution of 60,000. The auto gain control (AGC) target was 300% with a 20-ms maximum injection time. The included charge states were 2–7, and the filter dynamic exclusion duration was 45 s. The dd-MS2 isolation window was 1.6 m/z at a resolution of 15,000. The AGC target was 75%, with a maximum injection time of 22 ms and a normalized collision energy (NCE) of 30%.
Mass spectral raw data were processed using Spectronaut (Biognosys AG, Switzerland). This involved searching for the library and creating a spectral database using the uniprot_homo_20220308_20377_9606_swiss_prot database, assuming trypsin as the digestion enzyme. Carbamidomethyl was specified as the fixed modification. Oxidation (M) and acetyl (Protein N-term) were specified as the variable modifications. The Q-value cutoffs for both precursor and protein levels were set at 1%.
DIA Mass Spectrometry Analysis
The DIA method was conducted for identification and quantification of proteins. A 1 μg peptide sample mixed with iRT was analyzed on a Evosep One system (Evosep, Denmark) coupled to a Orbitrap Exploris 480 (Thermo Fisher Scientific) equipped with FAIMS pro. Buffer solution A was 0.1% formic acid aqueous solution, and solution B was 0.1% formic acid acetonitrile solution (acetonitrile was 100%). Peptides were separated on a 15 cm × 150 μm analytical column with 1.9-μm C18 beads with a packed emitter tip (Evosep, Denmark). The column temperature was maintained at 55 °C. The 30 SPD (44 min gradient) method was used for separation, followed by MS analysis with the Orbitrap Exploris 480. The MS parameters were the same in the DDA mode.
Regarding DIA MS data processing, Spectronaut was employed with default settings. The retention time prediction type was configured as dynamic iRT, with Spectronaut dynamically determining the optimal extraction window based on iRT calibration and gradient stability. Similarly, Q-value cutoffs for both precursor and protein levels remained at 1%. Following these analyses, differentially expressed proteins (DEPs) were selected using Student’s t-test, with proteins having P-values < 0.05 and fold changes of >1.2 being identified.
Bioinformatics Analysis
The molecular functions (MF), biological processes (BP), and cellular components (CC) of the DEPs were annotated using Blast2GO software (V6.0) [23]. Additionally, DEPs were annotated using the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database. Then, we used Interpro database for functional domain annotation analysis of DEPs. Fisher’s exact test was used to identify the significantly enriched Gene Ontology (GO) terms, KEGG pathways, and functional domains. To perform the cluster analysis, the target protein collection’s quantitative data were first normalized. Subsequently, both sample and protein expression dimensions were classified using matplotlib software (V3.7.2), followed by the generation of a hierarchical clustering heatmap. We utilized an online tool http://bioinformatics.psb.ugent.be/webtools/Venn/ to identify DEPs in 2Y vs. 2W while excluding those found in 1Y vs 1W. The results were presented in the form of a Venn diagram. These DEPs were then subjected to protein intersection analysis. Interaction networks of DEPs were generated using the String database and visualized with Cytoscape software (V3.9.1).
Targeted Validation by ELISA
Four selected candidate biomarkers, galectin-1 (LGALS1), galectin-3 (LGALS3), cluster determinant 44 (CD44), and insulin-like growth factor binding protein-1 (IGFBP-1), were examined with plasma samples from all recruited patients using commercially available ELISA kits from Aifang Bio. Co. Ltd. (Changsha, Hunan, China). The test was performed according to the instruction manual of the kit, and the optical density (OD) value of each well was acquired using a microtiter plate reader. The protein concentration in each sample was calculated by combining it with the standard curve. The target proteins were further analyzed by receiver operating characteristic (ROC) curve analysis based on the ELISA results, and the area under the curve (AUC) was used to assess the precision of the selected proteins for determining imrecoxib efficacy.
Statistical Analysis
The results of the current study were analyzed, and ROC curves were generated using GraphPad Prism (V9.5.0) and SPSS (V26.0) software. All data were subjected to normality testing. Parametric data were expressed as mean standard deviation (SD), nonparametric data as median (interquartile range), and count data as frequencies or constituent ratios (%). For comparisons between two groups of normally distributed measurement data, the t-test was used. The Mann-Whitney U test was used to assess non-normally distributed measurement data. To compare count data between the two groups, the χ2 test was used. The criterion for statistical significance was set at P < 0.05.
Results
Clinical Characteristics of Patients
An entire group of 142 patients with knee OA was included, and 7 cases were withdrawn because of loss to follow-up. A total of 35 patients with VAS and WOMAC improvement rates ≥ 25% were considered as responders, and a total of 31 patients with VAS and WOMAC improvement rates < 25% after medication were considered as non-responders. Figure 1 is the flowchart of the current study. Table 1 displays the general clinical data for both the responder and non-responder groups.
Quality Control and Repeatability Testing
Prior to analyzing DEPs, we initiated a quality control assessment with box plots (Supplementary Materials Fig. S1A, B). Post-normalization, the median protein abundance for each sample closely aligned with a horizontal line. We also performed principal component analysis (PCA) on identified DEPs (Supplementary Materials Fig. S1C, D). Group differences were not obvious, and there was an outlier sample in group 1Y, which was removed. Upon repeating PCA, differences among group samples remained small and indistinguishable, potentially because of high data variability. Subsequent analysis excluded this outlier sample. Pearson correlation coefficient analysis was employed for plasma samples (Fig. S1E), revealing strong overall correlation, except for the 2W14 sample, which exhibited weaker correlation. The relative standard deviation (RSD) values of the four groups of plasma samples were < 8, indicating good quantitative reproducibility (Supplementary Materials Fig. S1F).
Identification of DEPs
Proteomic analyses were carried out utilizing the 4D-DIA technique to examine and contrast the DEPs in the plasma of responders and non-responders before and after medication administration. We found a total of 1115 proteins in our investigation that belonged to 7720 different peptides. Between the 1Y and 1W groups, we identified 117 DEPs, consisting of 18 proteins showing upregulation and 99 displaying downregulation. Similarly, between the 2Y and 2W groups, we found 140 DEPs, with 29 proteins showing upregulation and 111 displaying downregulation (Supplementary Materials Table S1). Clustered heatmaps (Fig. 2A, B) and volcano graphs (Fig. 2C, D) showed DEPs between the responders and non-responders. Venn diagram showed that 88 proteins were statistically different only in 2Y vs. 2W but not in 1Y vs. 1W (Fig. 2E, Supplementary Materials Table S1). This indicates that these proteins were associated with the difference in imrecoxib treatment.
Identification of DEPs in plasma. A Hierarchical clustering heatmap showing the DEPs between the 1Y and 1W cohorts. B Hierarchical clustering heatmap showing the DEPs between the 2Y and 2W cohorts. C Volcano graphs of DEPs comparing the 1Y and 1W cohorts. D Volcano graphs of DEPs for the 2Y and 2W cohorts. E Venn diagram of DEPs. DEPs, differentially expressed proteins; 1Y, pre-treatment plasma samples from the responders; 1W, pre-treatment plasma samples from the non-responders; 2Y, post-treatment plasma samples from the responders; 2W, post-treatment plasma samples from the non-responders
Functional Classification of DEPs
One hundred forty DEPs from responders and non-responders after drug administration were analyzed for GO functional enrichment. In BP, DEPs were mainly enriched in leukocyte migration; Fc-gamma receptor signaling pathway was involved in phagocytosis, receptor-mediated endocytosis, and complement activation under the classical pathway. In CC, the main focus is on circulating immunoglobulin complexes, cytoplasmic membrane exterior, and immunoglobulin complexes. In MF, the main focus is on antigen binding and immunoglobulin binding (Fig. 3A, Supplementary Materials Table S2A). Enriched KEGG pathways were Fc epsilon RI signaling pathway, NF-κB signaling pathway, rheumatoid arthritis, B-cell receptor signaling pathway, Fc gamma R-mediated phagocytosis, and PI3K-Akt signaling pathway (Fig. 3B, Supplementary Materials Table S2B). The enriched structural domains include immunoglobulin-like structural domains, immunoglobulin V set structural domains, and the immunoglobulin-like structural domain superfamily (Fig. 3C, Supplementary Materials Table S2C).
Bioinformatics analysis of DEPs. A GO term enrichment statistics bubble chart (top 10). B Enrichment statistical bubble map of the KEGG pathway (top 20). C Bubble plot of enrichment statistics for protein domain classification (top 20). D Diagram of protein–protein interaction networks. DEPs Differentially expressed proteins; GO gene ontology; KEGG Kyoto Encyclopedia of Genes and Genomes; IgA immunoglobulin A; NADPH nicotinamide adenine dinucleotide phosphate; Fc epsilon RI the high affinity receptor for immunoglobulin E
Protein–Protein Interaction Networks (PPI)
In this study, we selected 88 proteins for interaction network analysis based on Venn diagram results. These proteins exhibited significant variations only post-drug administration among responders and non-responders. No notable differences were observed pre-drug administration between the two groups. This strategy aims to concentrate on post-treatment effects, minimizing potential pre-treatment influences. Through the PPI network, among DEPs, LGALS3, haptoglobin (HP), apolipoprotein A-I (APOA1), CD44, and transferrin receptor protein 1 (TFRC) were in the center of the network module and highly connected to other proteins (Fig. 3D). These proteins may be the key proteins affecting the difference in efficacy of imrecoxib in patients with OA.
Targeted Validation by ELISA
LGALS1, LGALS3, CD44, and IGFBP-1 may serve as biomarkers to gauge imrecoxib’s effectiveness, according to a review of the PubMed database. They were chosen as potential biomarkers. The ELISA results of three proteins, LGALS1, LGALS3, and CD44, were in line with the proteomic analysis results. The ELISA results for IGFBP-1 were in contrast to the proteomic results (Fig. 4A, Supplementary Materials Table S3). The results suggested that LGALS1, LGALS3, and CD44 might be biomarkers for early prediction of the therapeutic efficacy of imrecoxib, and ROC curves were utilized to evaluate the predictive capability of the DEPs concerning therapeutic response. The predicted AUC values of LGALS1, LGALS3, and CD44 were 0.835 (95% CI = 0.738–0.932), 0.787 (95% CI = 0.676–0.898), and 0.770 (95% CI = 0.658–0.881), respectively (Fig. 4B). The above proteins’ binary logistic regression analyses were then run. The LGALS1 and LGALS3 combination's AUC was 0.857 (95% CI = 0.765–0.949), while the LGALS1 and CD44 combinations’ AUCs were 0.887 (95% CI = 0.807–0.966), and the LGALS3 and CD44 combination’s AUC was 0.889 (95% CI = 0.811–0.966). These combinations significantly outperformed plasma proteins alone. The combination of LGALS1, LGALS3, and CD44 had the highest evaluation certainty in separating non-responders from respondents, achieving an AUC of 0.918 (95% CI = 0.853–0.983) (Fig. 4C).
The ELISA results and ROC curve of selected DEPs. A ELISA results of LGALS1, LGALS3, CD44, and IGGBP-1 expression levels in responders and non-responders. B ROC curve of three DEPs (LGALS1, LGALS3, and CD44). C ROC curve of combination of three proteins. Student’s t-test was conducted (****P < 0.0001). LGALS1 galectin-1; LGALS3 galectin-3; CD44 the cluster of differentiation 44; IGFBP-1 insulin-like growth factor binding protein-1; ELISA enzyme-linked immunosorbent assays; ROC receiver operating characteristic; AUC area under the curve; DEPs differentially expressed proteins
Discussion
OA is a degenerative joint disease caused by a variety of complex causes that impacts over 500 million individuals globally and constitutes a substantial cause of both pain and functional disability [1]. OA is highly prevalent among middle-aged and elderly individuals, and epidemiologic surveys in several countries have shown that the prevalence of OA in the population is > 15% [24]. Guidelines and expert consensus recommend NSAIDs as the drug of choice for the treatment of OA [4, 25]. Imrecoxib is a novel NSAID originally developed in China. It has a lower risk of inducing cardiovascular and gastrointestinal events and is clinically safer than other drugs such as celecoxib and etoricoxib [26]. Clinical studies have shown that there are efficacy differences in the treatment of OA with imrecoxib [11, 12]. As MS advances and methods continually improve, mass spectrometry-based proteomics technologies have progressively emerged as a crucial and dependable tool for the identification of protein-level biomarkers [27]. To our current understanding, this is the first investigation utilizing 4D-DIA to explore biomarkers predictive of early imrecoxib treatment effectiveness in OA. We have uncovered three previously unreported biomarkers, LGALS1, LGALS3, and CD44, that may be useful in assessing the early efficacy of imrecoxib therapy for OA.
Both LGALS1 and LGALS3 belong to the galectin protein family. This family’s members participate in a variety of biologic functions, including angiogenesis, immune cell activity control, and cell adhesion [28,29,30]. The dysregulation of LGALS1 and LGALS3 is associated with inflammatory, neoplastic, and cardiovascular disorders. Studies have demonstrated that LGALS1 damages cartilage by stimulating NF-κB-mediated inflammation, leading to OA [31]. The inflamed synovial β-galactoside-binding lectin LGALS3 is highly expressed and secreted in patients with rheumatoid arthritis (RA) and OA. In addition, LGALS3 has also been demonstrated to play a significant role in arthritis by causing lesions that resemble OA and joint swelling [32]. CD44 is a transmembrane glycoprotein belonging to the family of adhesion molecules. It participates in biologic processes such as cell proliferation, differentiation, migration, and angiogenesis, mediates cell signaling, controlling tissue homeostasis, and performs other activities [33]. CD44 is a biomarker and therapeutic target for tumors and is involved in tumorigenesis, progression, and metastasis [34]. CD44 mRNA levels have been found to be higher in lymphoma, breast, colon, and endometrial malignancies [35]. Qadri et al. [36] discovered that CD44 regulates Toll-like receptor 2 (TLR2) responses in human macrophages. They observed that decreasing CD44 levels or engaging CD44 with its ligand (HA) or CD44-specific antibody reduces NF-κB translocation and the subsequent generation of pro-inflammatory cytokines. In addition, they demonstrated that a CD44-specific antibody can mitigate macrophage activation in OA synovial fluid. These findings indicate that CD44 presents a potential new therapeutic target for addressing OA treatment. Zhang et al. [37] used an immunohistochemical method to assess the CD44 levels in the articular cartilage of individuals with varying degrees of OA severity. Their research revealed a correlation between CD44 expression in articular cartilage and the extent of joint damage in progressive knee OA. In this study, the levels of LGALS1, LGALS3, and CD44 were similar between responders and non-responders before the administration of imrecoxib. However, after the use of imrecoxib, their levels in non-responders exceeded those in responders. These findings suggest that LGALS1, LGALS3, and CD44 may be negatively correlated with or involved in the therapeutic mechanism of the efficacy of imrecoxib.
Various arthritis diseases often result from an imbalance in the synthesis and degradation of extracellular matrix macromolecules. Insulin-like growth factor-I (IGF-I) is crucial for maintaining this balance [38]. Upon binding to its receptor, IGF-I exhibits tyrosine kinase activity, activating signaling pathways, thereby influencing cell proliferation, differentiation, survival, and apoptosis [39]. While IGFBPs are generally believed to inhibit the action of IGF, they paradoxically extend the half-life of IGF [40]. Studies indicate that levels of IGFBPs in joint cartilage and synovial fluid are elevated in patients with OA, thereby inhibiting the action of IGF [41]. In our study, MS data revealed lower IGFBP1 levels in the responders after imrecoxib use, contrasting with ELISA results. ELISA, influenced by sample heterogeneity, may introduce discrepancies due to unaccounted factors like biologic differences and sample storage conditions. Future research with a larger sample size may clarify the imrecoxib-IGFBP1 relationship.
Sixty-six proteins belong to immunoglobulin-like structural domains, and prior research has linked immune cells with clinical outcomes in people with psoriatic arthritis and RA [42, 43]. We found that proteins such as IGDCC4, FCGRT, and CEACAM5 were upregulated in responders, which predicted a favorable response, whereas IGLC7, IGLV2-11, and IGHV3-35 were downregulated in responders, which may indicate a poor response. These proteins were involved in pathways such as the Fc epsilon RI signaling pathway, NF-κB signaling pathway, and B-cell receptor signaling pathway by KEGG analysis (Supplementary Materials Table S2B). Therefore, we hypothesized that these immune-related proteins might influence the response to imrecoxib in patients with OA through these pathways.
It is important to mention several limitations in this study. First, this study was based on patients with OA; to determine whether LGALS1, LGALS3, and CD44 are viable biomarkers for assessing the efficacy of imrecoxib, it was necessary to include a healthy population for comparison. Second, only 4 of the 140 DEPs in responders and non-responders after administration of the drug were selected for ELISA validation in this study, whereas the remaining DEPs were not verified. Third, further validation of the four candidate biomarkers (LGALS1, LGALS3, CD44 and IGFBP-1) is needed in a larger cohort, as some of these biomarkers are abundant, and the current cohorts are relatively small and may not be sufficient for validation. Fourth, it is valuable and essential to investigate protein expression in plasma before initiating treatment, as this can serve as a predictive factor for patient treatment response, aiding in the selection of more suitable therapies. Fifth, based on the PCA results, we removed an outlier from the final dataset for DIA quantification. Our research results may be influenced by potential batch effects. Consequently, enrolling a substantial number of patients is imperative to enhance result specificity and sensitivity. Furthermore, further extensive research is warranted to delve into potential signaling pathways or networks implicated in the therapeutic mechanism of imrecoxib.
Conclusion
In summary, by combining the results of 4D-DIA and ELISA, we concluded that the plasma levels of LGALS1, LGALS3, and CD44 were higher in non-responders than in responders after imrecoxib administration, which suggested the possibility of these proteins to objectively assess the therapeutic response to imrecoxib and thus as biomarkers for early assessment of imrecoxib efficacy. It offers alternatives for investigations examining the mechanisms of imrecoxib treatment for OA and for enhancing NSAID efficacy techniques, and it has potential application value.
Data Availability
The MS proteomics data have been deposited in the ProteomeXchange Consortium via the PRIDE partner repository [44] with the dataset identifier PXD047236. The release date was 2023-11-26.
References
Glyn-Jones S, Palmer AJ, Agricola R, et al. Osteoarthritis. Lancet. 2015;386(9991):376–87.
Leifer VP, Katz JN, Losina E. The burden of OA-health services and economics. Osteoarthr Cartil. 2022;30(1):10–6.
Katz JN, Arant KR, Loeser RF. Diagnosis and treatment of hip and knee osteoarthritis: a review. JAMA. 2021;325(6):568–78.
Geenen R, Overman CL, Christensen R, et al. EULAR recommendations for the health professional’s approach to pain management in inflammatory arthritis and osteoarthritis. Ann Rheum Dis. 2018;77(6):797–807.
Wittenberg RH, Schell E, Krehan G, et al. First-dose analgesic effect of the cyclo-oxygenase-2 selective inhibitor lumiracoxib in osteoarthritis of the knee: a randomized, double-blind, placebo-controlled comparison with celecoxib [NCT00267215]. Arthritis Res Ther. 2006;8(2):R35.
Lanas A. A review of the gastrointestinal safety data–a gastroenterologist’s perspective. Rheumatology (Oxford). 2010;49(Suppl 2):ii3-10.
Chen XH, Bai JY, Shen F, et al. Imrecoxib: a novel and selective cyclooxygenase 2 inhibitor with anti-inflammatory effect. Acta Pharmacol Sin. 2004;25(7):927–31.
Sun X, Zhen X, Hu X, et al. Cost-utility analysis of imrecoxib compared with celecoxib for patients with osteoarthritis. Ann Transl Med. 2021;9(7):575.
Guo W, Liu Y, Li J. Imrecoxib versus celecoxib as postoperative analgesia for patients receiving arthroscopic knee surgery: a randomized, controlled, non-inferiority study. Inflammopharmacology. 2022;30(3):875–81.
Jiang Y, Qi L, Liu J, et al. Analgesic efficacy of imrecoxib for postoperative pain following oral surgery: a prospective randomized, active-controlled, non-inferiority trial. Ann Transl Med. 2021;9(6):469.
Huang JL, Gu JR, Pan YF, et al. A multicenter, double-blind and randomized controlled phase II trial of imrecoxib in treatment of knee osteoarthritis. Chin Pharm J. 2011;46:1740–5.
Xu D, Zhang FC, Gu JR, et al. A multicenter double blind and randomized controlled phase III trial of imrecoxib in the treatment of knee osteoarthritis. Chinese J Allergy Clin Immunol. 2014;8(3):205–10.
Mun S, Lee J, Park A, et al. Proteomics approach for the discovery of rheumatoid arthritis biomarkers using mass spectrometry. Int J Mol Sci. 2019;20(18):4368.
Tardif G, Paré F, Gotti C, et al. Mass spectrometry-based proteomics identify novel serum osteoarthritis biomarkers. Arthritis Res Ther. 2022;24(1):120.
Braga FHG, Gómez-Mendoza DP, Lemos RP, et al. Proteomic analysis reveals stage-specific reprogramed metabolism for the primary breast cancer cell lines MGSO-3 and MACL-1. Proteomics. 2022;22(17): e2200095.
Guo T, Aebersold R. Recent advances of data-independent acquisition mass spectrometry-based proteomics. Proteomics. 2023;23(7–8): e2200011.
Meier F, Brunner AD, Frank M, et al. diaPASEF: parallel accumulation-serial fragmentation combined with data-independent acquisition. Nat Methods. 2020;17(12):1229–36.
Cuschieri S. The STROBE guidelines. Saudi J Anaesth. 2019;13:S31–4.
Altman R, Asch E, Bloch D, et al. Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and therapeutic criteria committee of the American rheumatism association. Arthritis Rheum. 1986;29(8):1039–49.
Schnitzer TJ, Tesser JR, Cooper KM, Altman RD. A 4-week randomized study of acetaminophen extended-release vs rofecoxib in knee osteoarthritis. Osteoarthr Cartil. 2009;17(1):1–7.
Pham T, Van Der Heijde D, Lassere M, et al. Outcome variables for osteoarthritis clinical trials: the OMERACT-OARSI set of responder criteria. J Rheumatol. 2003;30(7):1648–54.
Wiśniewski JR, Zougman A, Nagaraj N, et al. Universal sample preparation method for proteome analysis. Nat Methods. 2009;6(5):359–62.
Götz S, García-Gómez JM, Terol J, et al. High-throughput functional annotation and data mining with the Blast2GO suite. Nucleic Acids Res. 2008;36(10):3420–35.
Li Y, Xie W, Xiao W, et al. Progress in osteoarthritis research by the National Natural Science Foundation of China. Bone Res. 2022;10(1):41.
Bannuru RR, Osani MC, Vaysbrot EE, et al. OARSI guidelines for the non-surgical management of knee, hip, and polyarticular osteoarthritis. Osteoarthr Cartil. 2019;27(11):1578–89.
Meng Q, Zhang Z, Li F, et al. The prescription patterns and safety profiles of oral non-steroidal anti-inflammatory drugs in China: an 8-year real-life analysis. Ann Palliat Med. 2021;10(2):2224–37.
Haartmans MJJ, Emanuel KS, Tuijthof GJM, et al. Mass spectrometry-based biomarkers for knee osteoarthritis: a systematic review. Expert Rev Proteom. 2021;18(8):693–706.
Thijssen VL. Galectins in endothelial cell biology and angiogenesis: the basics. Biomolecules. 2021;11(9):1386.
Liu FT, Stowell SR. The role of galectins in immunity and infection. Nat Rev Immunol. 2023;23(8):479–94.
Elola MT, Wolfenstein-Todel C, Troncoso MF, et al. Galectins: matricellular glycan-binding proteins linking cell adhesion, migration, and survival. Cell Mol Life Sci. 2007;64(13):1679–700.
Osório J. Osteoarthritis: galectin-1 damages cartilage via inflammation. Nat Rev Rheumatol. 2016;12(3):132.
Hu Y, Yéléhé-Okouma M, Ea HK, et al. Galectin-3: a key player in arthritis. Joint Bone Spine. 2017;84(1):15–20.
Dzwonek J, Wilczynski GM. CD44: molecular interactions, signaling and functions in the nervous system. Front Cell Neurosci. 2015;9:175.
Xu H, Niu M, Yuan X, et al. CD44 as a tumor biomarker and therapeutic target. Exp Hematol Oncol. 2020;9(1):36.
Fujita N, Yaegashi N, Ide Y, et al. Expression of CD44 in normal human versus tumor endometrial tissues: possible implication of reduced expression of CD44 in lymph-vascular space involvement of cancer cells. Cancer Res. 1994;54(14):3922–8.
Qadri M, Almadani S, Jay GD, et al. Role of CD44 in regulating TLR2 activation of human macrophages and downstream expression of proinflammatory cytokines. J Immunol. 2018;200(2):758–67.
Zhang FJ, Luo W, Gao SG, et al. Expression of CD44 in articular cartilage is associated with disease severity in knee osteoarthritis. Mod Rheumatol. 2013;23(6):1186–91.
Tavera C, Abribat T, Reboul P, et al. IGF and IGF-binding protein system in the synovial fluid of osteoarthritic and rheumatoid arthritic patients. Osteoarthritis Cartilage. 1996;4(4):263–74.
Mathew R, Pal Bhadra M, Bhadra U. Insulin/insulin-like growth factor-1 signalling (IIS) based regulation of lifespan across species. Biogerontology. 2017;18(1):35–53.
Bach LA. Insulin-like growth factor binding proteins–an update. Pediatr Endocrinol Rev. 2015;13(2):521–30.
Galasso O, De Gori M, Nocera A, Brunetti A, Gasparini G. Regulatory functions of insulin-like growth factor binding proteins in osteoarthritis. Int J Immunopathol Pharmacol. 2011;24(1 Suppl 2):55–9.
Chen J, Tang MS, Xu LC, et al. Proteomic analysis of biomarkers predicting the response to triple therapy in patients with rheumatoid arthritis. Biomed Pharmacother. 2019;116: 109026.
van Kuijk AW, Gerlag DM, Vos K, et al. A prospective, randomised, placebo-controlled study to identify biomarkers associated with active treatment in psoriatic arthritis: effects of adalimumab treatment on synovial tissue. Ann Rheum Dis. 2009;68(8):1303–9.
Perez-Rivero Y, Bai J, Bandla C, et al. The PRIDE database resources in 2022: a hub for mass spectrometry-based proteomics evidences. Nucleic Acids Res. 2022;50(D1):D543–52.
Acknowledgements
We sincerely thank Shanghai Genechem for assisting in proteomic analysis and all the patients who participated in this study.
Funding
This work was supported by the National Natural Science Foundation of China (no.72104105), the Special Fund for Clinical Research of Nanjing Drum Tower Hospital (no. 2021-LCYJ-PY-07), and the Clinical Comprehensive Drugs Evaluation in Jiangsu Province. The Rapid Service Fee was funded by Nanjing Drum Tower Hospital.
Author information
Authors and Affiliations
Contributions
Han Xie and Weihong Ge contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Yuan Zhang, Zunyi Zhu, Jingxuan Wei, and Gulinigeer Ainiwaer. The first draft of the manuscript was written by Yuan Zhang, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Conflict of Interest
Han Xie, Yuan Zhang, Zunyi Zhu, Jingxuan Wei, Gulinigeer Ainiwaer, and Weihong Ge report no conflicts of interest.
Ethical Approval
This study was approved by the Ethics Committee of Nanjing Drum Tower Hospital (2021-402-02), and the entire trial was conducted under strict adherence to the Good Clinical Practice Guidelines of the International Council of Harmonization and the Declaration of Helsinki. All subjects gave written informed consent.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License, which permits any non-commercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc/4.0/.
About this article
Cite this article
Xie, H., Zhang, Y., Zhu, Z. et al. Plasma Proteomic Analysis Based on 4D-DIA Evaluates the Clinical Response to Imrecoxib in the Early Treatment of Osteoarthritis. Rheumatol Ther 11, 269–283 (2024). https://doi.org/10.1007/s40744-023-00636-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40744-023-00636-z