Osteoarthritis (OA) is the most common degenerative joint disease, affecting over 500 million people worldwide and representing a leading cause of disability in the elderly population (1). The disease is characterized by progressive articular cartilage degradation, subchondral bone remodeling, osteophyte formation, and synovial inflammation, ultimately leading to joint pain, stiffness, and functional impairment (2). Despite its high prevalence and socioeconomic burden, there are currently no curative treatments, and diagnosis largely relies on radiographic evaluation and clinical symptoms, which often detect the disease only at advanced stages.

Inflammation plays a crucial role in the pathophysiology of OA. Pro-inflammatory cytokines, such as interleukin-1β (IL-1β), tumor necrosis factor-alpha (TNF-α), and interleukin-6 (IL-6), are elevated in OA-affected joints and synovial fluid (3). These cytokines not only stimulate chondrocyte catabolism but also activate synovial fibroblasts, thereby amplifying local inflammation. Importantly, they promote the expression and activation of matrix-degrading enzymes, notably matrix metalloproteinases (MMPs) and aggrecanases (ADAMTS family), which mediate the irreversible breakdown of the cartilage extracellular matrix (4, 5). Consequently, the simultaneous measurement of cytokine concentrations and enzyme activity levels provides valuable insights into both the inflammatory status and catabolic burden within the OA joint environment.

The identification of reliable, minimally invasive biomarkers that reflect early joint degeneration and correlate with disease severity is a major unmet need in OA research (6). Several serum and synovial fluid markers, such as cartilage oligomeric matrix protein (COMP), C-terminal cross-linked telopeptide of type II collagen (CTX-II), and MMPs, have been explored; however, their diagnostic and prognostic utility remains limited owing to their variable specificity and inconsistent association with disease progression (7, 8).

EphA2, a member of the ephrin receptor tyrosine kinase family, is widely recognized for its role in embryonic development, angiogenesis, and tumorigenesis (9). Emerging evidence suggests that EphA2 also plays a role in skeletal and joint biology, influencing bone remodeling by regulating osteoclast and osteoblast differentiation (10). Recent transcriptomic and experimental studies have revealed that EphA2 expression is markedly upregulated in osteoarthritic cartilage and synovium, promoting pro-inflammatory signaling, chondrocyte hypertrophy, and extracellular matrix degradation via MAPK and NF-κB pathways (11). Furthermore, pharmacologic inhibition of EphA2 in OA models has been shown to attenuate cartilage degeneration and reduce the expression of catabolic enzymes, such as MMP-13 (11).

EphA2 has been extensively studied in oncology, where it regulates tumor growth, angiogenesis, and metastasis (12, 13). Emerging evidence suggests that EphA2 may also contribute to OA pathology by promoting chondrocyte apoptosis, extracellular matrix degradation, and inflammatory responses (14). Although overexpression of EphA2 in joint tissues has been reported, its systemic expression and relationship with disease severity remain unclear, particularly in musculoskeletal disorders, such as knee OA. Preclinical studies further support targeting EphA2 in OA, as it is selectively upregulated in hypertrophic chondrocytes compared to other Eph receptors. In vitro inhibition attenuates the expression of inflammatory mediators and hypertrophic differentiation, whereas in vivo blockade with ALW-II-41-27 reduces synovial inflammation and cartilage damage (11), highlighting its potential as a promising biomarker. Therefore, investigating serum EphA2 levels may provide novel insights into its clinical significance, serving as a convenient biomarker for disease activity and progression. Accordingly, the present study aimed to evaluate serum EphA2 levels in patients with knee OA compared to those in healthy controls and examine their association with radiographic severity and clinical parameters.

In this study, we comprehensively evaluated the biochemical, inflammatory, oxidative stress, and ferroptosis-related profiles of patients with knee OA and compared them to those of healthy controls. Our findings confirmed that knee OA patients exhibited significantly elevated serum levels of cartilage degradation markers (COMP, HA, MMP-13, and CTX-2), pro-inflammatory cytokines (TNF-α, IL-1β, IL-6, and IL-17A), oxidative stress markers (MDA), ferroptosis-associated proteins (ACSL4), and notably, EphA2. Conversely, protective cartilage synthesis markers (PIICP and aggrecan) and antioxidant defenses (GSH and GPX4) were markedly reduced. These biochemical alterations were observed despite no significant differences in age or sex between the knee OA and control groups, consistent with prior reports that OA is largely driven by pathological rather than demographic factors (1).

One of the most notable results is the significant elevation of serum EphA2 in knee OA patients compared with healthy controls, with further increases in advanced disease stages (K-L III-IV). ROC analysis demonstrated excellent diagnostic accuracy (AUC = 0.924), exceeding that reported for many conventional OA biomarkers such as COMP (AUC ~ 0.80) (20) and CTX-2 (AUC ~ 0.88) (21). EphA2, a receptor tyrosine kinase, has been well-studied in cancer biology because of its role in cell proliferation, migration, angiogenesis, and resistance to therapy (22). More recently, evidence has implicated it in inflammatory joint disease. Recent transcriptomic and experimental work revealed that EphA2 is upregulated in osteoarthritic cartilage, and that pharmacologic inhibition of EphA2 reduces inflammation, hypertrophic marker expression, and MMP-mediated matrix breakdown in human OA chondrocytes and mouse OA models. The MAPK and NF-κB signaling pathways, both downstream of EphA2 activation, are well-established drivers of cartilage degradation (11, 23). Our results extend these findings to the clinical setting, suggesting that EphA2 may be both a marker and potential driver of OA pathophysiology.

We observed significant positive correlations between EphA2 and established cartilage breakdown markers, including COMP, HA, MMP-13, and CTX-2, and negative correlations with PIICP and aggrecan levels. COMP and CTX-2 are recognized indicators of cartilage degradation and predictors of OA progression (24). MMP-13, the primary collagenase responsible for type II collagen degradation in cartilage, is a critical mediator of OA pathogenesis (25). The association between EphA2 and MMP-13 is particularly noteworthy, as EphA2 has been shown to modulate MMP expression in chondrocyte cultures (26), supporting the idea that EphA2 may directly facilitate extracellular matrix breakdown.

We also found that serum EphA2 levels were positively correlated with inflammatory cytokines, including TNF-α, IL-1β, IL-6, and IL-17A, all of which are known to exacerbate cartilage catabolism and synovial inflammation (27). TNF-α and IL-1β stimulate MMP-13 production, whereas IL-6 and IL-17A synergistically enhance inflammatory cartilage destruction (28). The observed relationship between EphA2 and these cytokines is consistent with reports that EphA2 activation amplifies NF-κB– and MAPK-mediated inflammatory signaling in endothelial and immune cells (29), supporting the hypothesis that EphA2 contributes to the chronic inflammatory milieu in OA.

A novel and clinically relevant aspect of this study is the association between EphA2 and oxidative stress and ferroptosis-related markers. We found positive correlations between EphA2, MDA and ACSL4, along with negative correlations between EphA2, GSH and GPX4. Oxidative stress has long been implicated in OA, driving chondrocyte apoptosis and matrix degradation (30). Ferroptosis, an iron-dependent cell death process characterized by lipid peroxidation, has recently been recognized as a critical contributor to cartilage degeneration (31). ACSL4 promotes the esterification of polyunsaturated fatty acids, generating substrates for lipid peroxidation, whereas GPX4 and GSH serve as key inhibitors of ferroptosis by reducing the levels of lipid hydroperoxides. Our results suggest that EphA2 may be involved in ferroptotic signaling pathways in OA, a hypothesis that warrants further research.

Importantly, although we observed parallel associations between EphA2 and ferroptosis markers, the potential mechanistic link requires deeper consideration. EphA2 activates the MAPK/ERK and PI3K/Akt signaling cascades (22, 23), which are also implicated in the regulation of ferroptosis sensitivity (32, 33). For example, ERK activation enhances ACSL4 expression and lipid peroxidation, thereby promoting ferroptosis (34). Conversely, the PI3K/Akt/Nrf2 axis upregulates GPX4 and GSH biosynthesis, thereby suppressing ferroptosis (35). Dysregulated EphA2 signaling in OA chondrocytes may tilt the balance toward ERK-driven ACSL4 upregulation and impaired Nrf2-mediated antioxidant defense, favoring ferroptotic cell death. This hypothesis is consistent with the correlations observed in this study and provides a testable mechanistic framework for future studies.

Previous OA biomarker studies have primarily focused on individual catabolic or inflammatory markers, with limited integration of molecular pathways. For instance, CTX-2 and COMP predict OA progression but do not capture the full inflammatory or oxidative profile (20, 21). By integrating cartilage metabolism (MMP-13 and CTX-2), inflammation (TNF-α, IL-1β, IL-6, and IL-17A), oxidative stress (MDA), and ferroptosis markers (ACSL4 and GPX4) with EphA2, our study highlights EphA2 as a central node reflecting multiple aspects of OA pathology. This multi-pathway association distinguishes EphA2 from traditional biomarkers with a single mechanism.

Our cohort’s relatively young age and low BMI, despite a substantial proportion of patients with advanced radiographic OA-are consistent with patterns observed in populations engaged in high physical activity or labor-intensive occupations, where chronic mechanical joint loading may contribute to OA development independently of obesity-related risk factors (36). BMI is a crude measure of adiposity and does not account for body composition; in lean individuals with OA, reduced muscle mass and increased joint stress may play more important roles in disease progression (37, 38). Furthermore, inflammatory biomarkers such as cytokines and ESR in our cohort were generally low and within the reference ranges. While some between-group differences reached statistical significance, these changes are unlikely to have substantial clinical relevance in early or moderate OA, as low-grade inflammation in OA often does not produce marked alterations in systemic inflammatory markers (39, 40).

This study had several limitations. First, our study spanned a 2-year period (January 2023–December 2024) and employed a cross-sectional design, which limits the ability to infer causality; thus, our findings should be interpreted as associations rather than as causal relationships. Future longitudinal studies with extended follow-up are warranted to determine the predictive value of EphA2 and elucidate its underlying mechanisms in OA progression. Second, this study did not analyze OA subtypes, which may differ in molecular characteristics. Future studies with larger cohorts and detailed metabolic profiling are warranted to explore whether EphA2 expression varies across OA phenotypes. Third, the younger age distribution and absence of patients over 65 years in our study likely reflect sampling bias from recruiting only in specific outpatient clinics and community settings rather than the general population. Future studies should use population-based sampling, including multiple hospitals, primary care centers, and community sources, to ensure the inclusion of older adults and reduce age-related sampling bias. Fourth, analgesic use was not systematically recorded for statistical adjustment; thus, a small degree of bias in the WOMAC outcomes cannot be ruled out. Future studies should systematically document and adjust for all analgesic uses to more accurately isolate the relationship between biomarkers and clinical status. Fifth, although we adjusted for BMI in the multivariate analyses, residual confounding by unmeasured factors such as medication use, physical activity, and comorbidities cannot be excluded, underscoring the need for future studies with larger cohorts and more comprehensive clinical data to validate our findings. Finally, while our findings suggest a strong association between serum EphA2 levels and OA pathogenesis and disease severity, the underlying mechanisms remain unclear. Future functional studies, including in vitro chondrocyte models and in vivo OA models, are necessary to determine whether EphA2 acts as a causal driver or merely as a correlated “bystander” biomarker.

In summary, our study demonstrated that serum EphA2 levels were significantly elevated in patients with knee OA, particularly in those with advanced K-L grades, and exhibited strong diagnostic potential (AUC = 0.924). EphA2 levels were positively correlated with cartilage degradation markers, inflammatory cytokines, and oxidative stress indicators and negatively correlated with anabolic and antioxidant markers (Figure 6). These results suggest that EphA2 is a promising biomarker for the diagnosis of OA and assessment of disease severity. However, as this study was cross-sectional, causal relationships could not be inferred, and further mechanistic studies are required to explore its biological role in OA progression.