Osteoarthritis (OA) is a kind of joint, slow-onset, and highly disabling disease in clinical practice which causes irreversible joint lesions such as bone resorption, bone destruction, and bone fibrosis, eventually leading to disability (Duncan et al., 2009). In recent years, the incidence and disability rates of osteoarthritis have continued to rise, making it the second most common chronic disease among older patients, after hypertension, hyperglycemia, and hyperlipidemia. This trend poses an ongoing threat to public health systems worldwide, highlighting the urgent need for effective prevention and treatment strategies to address this growing health concern (Hawker, 2019). Osteoarthritis of the knee is the primary type of osteoarthritis, ranking 11th in the world in terms of disability and posing a substantial economic burden to patients, families, and society (Losina et al., 2015).

For many years, osteoarthritis was considered a “wear and tear” disease. However, with the development of clinical medicine, it has been recognized as a complex, multifaceted disease (Kraus et al., 2015). According to the International Osteoarthritis Research Society, osteoarthritis is initially manifested as a molecular disorder with joint tissue metabolism abnormalities, then followed by anatomical and/or physiological disturbances characterized by cartilage degradation, bone remodelling, bone redundancy formation, joint inflammation, and loss of normal joint function. There is growing evidence showing that inflammation plays a vital role in the pathogenesis of OA, and the link between inflammatory factors and OA risk has been increasingly reported (Molnar et al., 2021). Previous studies have shown that OA patients have higher circulating levels of several inflammatory cytokines than healthy controls. For example, Jiří Baloun et al. (2020) showed elevated levels of nine inflammatory mediators (e.g., eosinophil chemokine, monocyte chemotactic protein 1, interleukin-10(IL-10), and tumour necrosis factor) in patients with hip osteoarthritis (Baloun et al., 2020). In a meta-analysis designed for 682 individuals, elevated monocyte chemotactic protein-1 (MCP-1) concentrations were observed in patients with OA but not in healthy controls (Ni et al., 2020). It has also been shown that elevated CCL-5 contributes to OA progression (Monibi et al., 2015). However, the relationship between circulating levels of cytokines and the risk of osteoarthritis is questionable due to the limitations of observational studies, such as small sample sizes, follow-up time, and reverse causality, which can mislead the results (Davey Smith and Ebrahim, 2003).

Mendelian randomization approach is well suited to avoid the above problems. At this stage, Mendelian randomization (MR) is one of the most effective methods for making causal inferences. It is becoming increasingly popular as the results of many genome-wide association analyses are published and shared (Bowden et al., 2015). Mendelian randomization methods use genetic information (primarily single nucleotide polymorphisms (SNPs)) as instrumental variables for causal inference, and its studies are based on three hypotheses: 1).

Instrumental variables and exposure factors are strongly correlated; 2) instrumental variables and confounders are not correlated; and 3) instrumental variables are not directly correlated with the outcome, and its effect on the outcome can only be manifested through exposure (Miao et al., 2021). Therefore, MR can effectively avoid the confounding bias of traditional epidemiological studies.

Here, we applied a two-sample MR approach to reveal the causal impact of inflammatory factors on OA risk using pooled statistics of inflammatory cytokines and OA from a genome-wide association study (GWAS) conducted in a large cohort.

Osteoarthritis (OA), characterized by cartilage lesions, is a chronic degenerative disorder with a high prevalence and disability rate (Glyn-Jones et al., 2015). With the development of clinical medicine, OA has changed from a superficial cartilage “wear and tear” disease to a diverse disease with complex pathogenesis affecting all tissues in the joint (Primorac et al., 2020). OA has seriously affected the quality of life, especially for the elderly, and imposed a severe economic burden on families and the country. Still, the biological mechanisms underlying OA etiology are poorly understood. Due to these limitations, we performed a two-sample MR analysis; our study is the first MR analysis to determine whether inflammatory cytokine levels are associated with OA risk based on genetic data from publicly available databases.

We obtained evidence showing a causal association between high levels of MIP-1β/TNF-β and decreased risk of OA; increased circulating levels in Rantes had a suggestive association with increased risk OA.

Macrophage inflammatory protein-1beta (MIP-1β/CCL4) is the essential chemotactic cytokine in the immune response against infection and inflammation; A chemokine (chemotactic factors) is a small cytokine that attracts other cells to a local area to exert its biological effects. Chemokines, which can be secreted by leukocytes and some tissue cells, are a protein family with over 60 members, mostly 8–10 kDa in molecular weight (Allen et al., 2007). Few clinical observational studies or Meta-analyses have been performed correlation between MIP-1β and OA, but our current MR analysis determined that high levels of circulating MIP-1b were associated with a reduced risk of OA (OR = 0.998, 95% CI = 0.996–0.999, p = 9.61 × 10⁻⁵, per 1 Standard deviation (SD) increase).

In addition, our MR analysis revealed a potential association between the chemokine CCL5, also known as Rantes, and an increased risk of osteoarthritis (OA). CCL5 is a member of the CC chemokine family and has been shown to regulate the expression and secretion of normal or activated T Cells. Our findings suggest that increased circulating levels of CCL5 may contribute to the development of OA (Suffee et al., 2011). Several observational studies have previously demonstrated a link between CCL5 and OA; Beekhuizen et al. reported that the most significantly elevated mediator in synovial fluid in the articular cavity of patients with OA compared with controls was CCL5. A further study confirmed this, documenting significantly higher CCL5 levels in 18 patients compared with the control group (Beekhuizen et al., 2013). Nevertheless, due to the limitations of studies, the results are suspicious and ambiguous; for the following reasons: small sample size and potential confounders. Our MR analysis, therefore, yielded a more reliable suggestive association; an increase of one SD was associated with a 1.3% increased risk of OA (OR = 1.013, 95% CI = 1.002–1.024, p = 0.016).

Biologically, CLL4 is the blocker of infection and inflammation, leading to a better amelioration of various inflammatory conditions through the corresponding immune response. CCL4 and CCL5 are CCR5 receptor-binding ligands, but CCL4 plays a minimal role for CCR5 compared to CCL5 (Molnar et al., 2021). Elevated CCL4 activates the NF-kappaB signalling pathway, which leads to the development of OA inflammation is effectively mitigated. Several studies have elucidated the relationship between NF-kappaB and OA. NF-kappaB is strongly associated with chondrocyte survival and apoptosis in osteoarthritis (Barreto et al., 2020). In a study by Tianlong Pan et al., it has been shown that activation of the NF-kappaB signalling pathway inhibited chondrocyte apoptosis induced by inflammatory factors (especially interleukin-1β), thereby delaying the progression of OA (Pan et al., 2017). CCL5 and CCL4 have opposite effects, with CCL5 binding to CCR5 leading to cartilage breakdown in the joint (Borzì et al., 2000). The main reason for this is that CCL5 and CCR5 binding to each other regulates the expression of MMP-3, which leads to the inhibition of proteoglycan synthesis and increases the release of proteoglycan, thereby degrading cartilage (Masuko-Hongo and Yudoh, 2005). CCL5/CCR5 not only degrades cartilage but also increases the production of the remaining inflammatory mediators (such as IL6) by human synovial fibroblasts through the PKCδ/c-Src/c-Jun signal pathway and AP-1 signaling pathways, allowing further development of OA (Tang et al., 2010).

Tumour necrosis factor beta (TNF-β), also called lymphotoxin A, is a close homolog of tumour necrosis factor alpha (TNF-α) (Aggarwal et al., 2012). In the present MR analysis, we found that elevated circulating levels of TNF-β were associated with a decreased risk of developing OA (OR = 0.996,95% CI = 0.994–0.999, p = 0.002, per 1 Standard deviation (sd) increase). TNF-β, as well as TNF-α, could bind to TNF receptors (TNFR 1 and TNFR2) (Doss C G, 2014). Prior studies have suggested that the sequences of TNF-β and TNF-α are similar, so that TNF-β increases the risk of inflammatory diseases as does the TNF-α. However, the affinity of TNF-β for receptors is much lower than that of TNF-α, and TNF-β is 1,000 times less effective in receptor activation. Therefore, higher levels of TNF-β can compete with TNF-α for the TNFR1 and TNFR2 receptor sites, and act as an antagonist of TNF-α, reducing the risk of OA.

Our study has several strengths. (1) This is the first MR study to elucidate the relationship between inflammatory cytokines and OA risk. (2) Unlike observational studies, our current study avoids confounders and reverse causality to the greatest extent possible, providing a reliable pair of cause-and-effect relationships. (3) Our study data were obtained from the publicly available GWASS database with a large amount of original study data, which provided a strong guarantee for this study. (4) Unlike the time-consuming randomized controlled studies (RCTs), the cost of time and money invested in this study is highly cost-effective for the results we obtained. However, this study also has some limitations. First, because the data in the database were sourced from Europe, the study was limited to European participants, and the usefulness of the results for other populations remains to be seen. (润色) Second, this study ignored the diversity of OA diseases and did not analyze site-specific OA, and cytokines may have a causal effect on knee OA/hip OA. Third, for the 20 cytokines that did not yield significance in the current study, a relationship between them and OA cannot be excluded, possibly due to the small number of SNPs involved in the study and the lack of MR computational power. Fourth, cytokines are a dynamic indicator, and unlike other indicators such as weight, MR does not address the dynamics of cytokine levels.

Our study finally identified a causal association between circulating levels of TNF-β, MIP-1β and OA risk and a suggestive link between Rantes and OA. Our study may provide a deeper understanding of the pathogenesis of OA, as well as the development of effective management strategies for the clinic. We suggest that TNF-β, MIP-1β, and Rantes may serve as potential therapeutic targets for OA development.