Spatiotemperal Dynamics of Osteoarthritis: Преобразование значений от стенда до Бедсайда.

Introduction

Osteoarthritis (OA) is the most prevalent form of degenerative joint disease. Patients with OA frequently experience joint pain and limited mobility in multiple joints [ 1 ]. Additionally, weight-bearing joints affected by OA are also at significantly increased risk of other cormobilities, including falling and fracturing [ 2 ]. The latest epidemiological studies have found that the prevalence of OA accounts for over 7% of the world's population, with an estimated 527.81 million cases of OA worldwide [ 3 ]. Of these cases, 40% are symptomatic in patients with moderate to severe imaging. In the United States, more than 400,000 joint replacements are performed annually, and the cost of OA treatment accounts for 0.25% to 2.5% of the gross national product, and these data increase yearly [ 4 ]. In addition, OA is a major contributor to disability [ 5 ], thereby imposing a considerable burden on both individuals and society. The current management of OA has failed to stop or alleviate the disease progression, leaving joint replacement as the last choice, primarily due to the lack of methods that prevent OA progression, owing to a limited understanding of OA aetiology.

The progression of different OA phenotypes varies from several years to decades; therefore, tracking the dynamic factors influencing OA progression is crucial for a comprehensive understanding. Nevertheless, existing research in OA predominantly focuses on evaluating limited instances of osteoarthritic progression at specific time points. However, OA exhibits complex phenotypes and endotypes across different levels, and part of the truth may hinder a complete understanding of OA’s pathology and aetiology [ 6 ]. Fortunately, spatiotemporal variation has been primarily introduced to the epidemiology of OA to account for variations in incidence, timing, and distribution influenced by various factors. It refers to the dynamic changes occurring in OA progression's spatial and temporal dimensions, which involve changes from two aspects: spatial and temporal changes that are holistically integrated. The spatial changes matter due to the highly organised structure of the osteochondral unit, which includes different layers of cartilage, calcified cartilage zone, subchondral bone plate and cancellous bone, each reacting differently during OA progression [ 7 ]. The variability in disease progression timelines and the susceptibility of OA to diverse risk factors contribute to the heterogeneity of OA phenotypes, posing significant challenges to developing precision medicine approaches. This complexity partly explains the high failure rate of clinical trials to treat OA. Insight into spatiotemporal variation helps understand OA's epidemiological patterns [ 8 , 9 ], disease burden, and healthcare needs across different times and places, guiding the optimal allocation of healthcare resources [ 10 ]. Advances in imaging, big data, artificial intelligence [ 11 - 13 ], and precision medicine have shifted the focus to OA's spatiotemporal dynamics, emphasising its complex, multi-stage progression influenced by genetic, metabolic, biomechanical, and environmental factors [ 14 , 15 ]. Recent advancements in imaging, tissue analysis, molecular biology, bioinformatics, and artificial intelligence (AI) analytics have enhanced our ability to study these variations in detail, offering early and precise intervention opportunities [ 16 ]. For advanced-stage OA patients, these technologies enable more accurate identification of disease subtypes and phenotypes, improving patient care by facilitating precision treatment plans based on the evolution of joint structure and function, thus improving treatment outcomes and quality of life [ 17 ].

Research on OA's aetiology and pathological transformations aims to uncover more effective treatments. Initially perceived as a wear-and-tear issue [ 18 ], recent research and advances reveal OA's complex genesis involving joint deterioration due to aging, stress, inflammation, and metabolic disorders. These lead to issues like synovitis and cartilage degeneration, accelerated by gene alterations and signalling pathway interactions [ 19 - 21 ]. As OA advances, cellular metabolism abnormalities in joint cartilage spark extensive inflammatory responses, exacerbating joint damage. Current clinical imaging techniques fail to detail these molecular and microstructural changes, highlighting the need for advanced spatial-temporal analysis for a comprehensive disease understanding [ 22 , 23 ]. Emerging spatially resolved technologies [ 24 , 25 ] have been validated to play pivotal roles in OA research, offering insights into disease phenotyping at various levels [ 8 , 26 ]. These advanced strategies enable detailed molecular information through analyses like transcriptomics and proteomics, enhancing our understanding of OA's pathogenesis [ 27 , 28 ]. Technologies like hyperspectral analysis allow non-invasive molecular detection in joint tissues, promising new diagnostic tools [ 8 ]. Although various studies have been published in this area, limited review papers have been collected to summarise the abovementioned advancements. This paper reviews the spatiotemporal developments in OA research from bench to bedside, aiming to inform precision treatment and prevention strategies ( Fig. 1 ).

Spatiotemporal Development Trajectory in OA at The Macro Level

OA Changes from Temporal Perspective

Developmental Risk Factors Induce OA Heterogeneity in Adults

The heritability of OA is influenced by a complex interaction of genetic, epigenetic, fetal, and environmental factors that contribute to its development and progression across the lifespan ( Fig. 1 ). Compared to other chronic musculoskeletal disorders, such as rheumatoid arthritis (RA), spinal arthritis (SpA) psoriatic arthritis (PsA), myalgia and osteoporosis, OA has a relatively more significant heritability [ 29 ]. Fetal risk factors are essential in the development and progression of OA. Genetic predispositions, including genetic and epigenetic factors, are critical in OA progression [ 30 ]. Solid familial aggregation has been previously observed in OA [ 31 ], and recent Musculoskeletal Pain in Ullensaker Study (MUST) cohort and Norweigian OA Twin study (Nor-Twin) suggested that maternal inheritance are contributor to OA than paternal inheritance [ 32 , 33 ]. Further, Magnusson et al. estimated the gender-specific genetic contributions to knee OA surgeries [ 34 ] using the largest twin registry worldwide, and it was discovered that knee OA surgery had a heritability of 53%. Additionally, the genetic contribution in males varied with body mass index (BMI) and age.

Genetic polymorphisms can predispose individuals to OA. Genetic variations related to cartilage structure, joint formation, and inflammation can be inherited, raising the risk of contracting OA later in life. The main genes associated with OA onset include Pattern Recognition Receptors, Interleukin 1 (IL-1), Interleukin 6 (IL-6), Matrix Metalloproteinases (MMPs), A Disintegrin and Metalloprotease with Thrombospondin Type I Motifs (ADAMTS), Intelectin 1 (ITLN1), Active Chemokine C-Motif Ligand 2 (XCL2) and DOT1 like histone lysine methyltransferase (DOT1L) [ 35 ].

Meanwhile, maternal exposure to certain environmental factors can lead to epigenetic modifications, which may affect gene expression related to cartilage health and inflammation, thereby increasing OA susceptibility. Epigenetic modifications are vital in regulating gene expression, including deoxyribonucleic acid (DNA) methylation, histone changes, and chromatin remodelling. Ribonucleic acid (RNA) epigenetics contributes significantly to this regulatory process [ 36 ]. Maternal health and exposure to environmental risk factors are also crucial for OA risk in later life. This process's main epigenetic regulatory factors include DNA methylation, histone modifications, and non-coding RNA [ 37 ]. Maternal smoking and alcohol use during pregnancy further exacerbate these risks. The study verified that maternal smoking could impair fetal growth, leading to lower bone mass and joint abnormalities, thereby increasing the risk of OA via the transforming growth factor-beta (TGF-β) signalling pathway [ 38 ]. Excessive alcohol intake during pregnancy can also affect fetal development, leading to skeletal abnormalities and a higher likelihood of OA via the IGF-1 signalling pathway [ 39 ]. Another study explored the long-term effects of early-life exposure to hyperglycaemia on OA susceptibility [ 40 ], and the mice model showed that early hyperglycaemic exposure leads to increased OA risk in adulthood primarily due to persistent low Sirtuin 3 (Sirt) expression, which disrupts mitochondrial function and mitophagy in chondrocytes. Enhancing Sirt3 expression or using Sirt3 agonists like Honokiol was shown to mitigate these effects and slow OA progression, which suggests that gestational diabetes may increase OA risk through epigenetic changes.

Up to date, evidence shows that poor or excessive maternal nutrition will lead to a newborn’s OA in later life. Poor maternal nutrition, such as insufficient nutrition during pregnancy, can have an impact on fetal growth and development, thereby increasing the likelihood of OA in a newborn's later life. By programmatic alterations of the hypothalamic-pituitary-adrenal (HPA) axis-related neuroendocrine metabolism, low birth weight can be induced, along with changes in glucose and lipid metabolism, and finally contribute to the poor-quality state of articular cartilage [ 41 ]. On the contrary, maternal obesity in pregnant women can also lead to fetal macrosomia (large birth weight), which is associated with increased mechanical stress on developing joints and a higher risk of OA. Gestational diabetes can also result in larger fetal size and altered fetal metabolism, increasing the risk of joint abnormalities and OA later in life. Meanwhile, a high-fat diet (HFD) in parents and grandparents affects metabolic and skeletal health, systemic inflammation, and the risk of OA in two generations of mice offspring [ 42 ]. Researchers examined the metabolic profiles and OA susceptibility in first- and second-generation mice whose parents were fed high-fat or low-fat diets, and they found that parental obesity and high-fat diets negatively impact the musculoskeletal health of subsequent generations.

Meanwhile, preterm birth, low birth weight and growth patterns (like childhood obesity) can significantly influence the risk of developing joint disorders and OA. It was believed that babies born with low birth weight may have underdeveloped cartilage and hip bone abnormalities, which increases their susceptibility to hip OA and other joint disorders [ 43 ]. Earlier studies used dogs as the model animals for the longitudinal study, and they observed that dog body weight could alter hip development, which may result in early-onset hip OA [ 44 ]. This was supported by another meta-analysis [ 43 ], which deduced that low birth weight and preterm birth are associated with abnormal hip morphology and an increased risk of hip OA. No association between birth weight and OA was found. However, Zengfeng et al. recently used Mendelian randomisation to assess birth weight, childhood obesity, and OA occurrence in adults using the Genetics of OA Consortium (n = 826,690). The results showed that childhood obesity has shown a causal relationship with OA in weight-bearing joints, suggesting a general link between obesity and OA [ 45 ]. Another study from the National Survey of Health and Development (NSHD) (n=2989) also discovered that the changes in Body Mass Index (BMI) during childhood among women and the changes in BMI during adolescence in men were likewise positively correlated with knee OA at the age of 53. [ 46 ]. The risk of high BMI accumulates with adulthood, which emphatically accentuates the crucial significance of weight control spanning an entire lifetime as a primary preventive measure for knee OA.

Other developmental factors are also worth careful consideration. Abnormal fetal positioning, such as breech presentation, can affect joint development and alignment, increasing the likelihood of joint abnormalities, such as developmental dysplasia of the hip and further inducing secondary OA [ 47 ]. In the UK Biobank, early menarche was also associated with a higher risk of OA [ 48 ]. In summary, developmental risk factors can lead to impaired joint development, altered biomechanics, and a pro-inflammatory state, increasing the risk of OA later in life. Understanding these mechanisms highlights the importance of maternal health and prenatal care in preventing OA and improving long-term joint health.

Longitudinal Study, Big Data, and Temporal Change of OA

Several databases have been established to track the OA progression temporally to monitor degenerative diseases, including OA. Famous databases include Osteoarthritis Initiative (OAI) [ 49 , 50 ], UK Biobank [ 51 ], Multicenter Osteoarthritis Study (MOST) [ 52 , 53 ], Framingham OA Study [ 54 ], Chingford General Population Survey [ 55 ], Rotterdam Study (ERGO) [ 56 ], Cohort Hip and Cohort Knee (CHECK) Study [ 57 ], Applied Public-Private Research enabling OA Clinical Headway funded by the Innovative Medicines Initiative (IMI-APPROACH) [ 58 ], Johnston County OA Project [ 59 ], Hertfordshire Cohort Study [ 60 , 61 ], Tasmanian Older Adult Cohort Study (TASOAC) [ 62 ] and Netherlands Epidemiology of Obesity (NEO) Study [ 63 ] ( Table 1 ). OAI identified genetic, biochemical, and imaging biomarkers that predict OA progression and established associated risk factors [ 64 ]. The UK Biobank has provided large-scale genomic data pinpointing genetic variants linked to OA and establishing connections between OA and cardiovascular diseases [ 49 ]. MOST explored risk factors such as age, obesity, and joint injury and assessed disease progression through imaging techniques [ 53 ]. The Framingham OA Study revealed a correlation between obesity and knee OA and identified both genetic predispositions and lifestyle factors influencing the disease. The Chingford General Population Survey examined the natural history and risk factors of OA, emphasising the impact of estrogen and menopause on the condition [ 65 ]. ERGO discovered associations between OA and metabolic syndrome, cardiovascular diseases, and genetic markers [ 56 ]. The CHECK study provided longitudinal data on early OA, identifying early symptoms and risk factors that predict disease progression [ 66 ]. IMI-APPROACH [ 58 ] identifies novel OA phenotypes, validates prediction models for OA progression and develops sensitive markers for OA progression [ 67 , 68 ]. The Johnston County OA Project highlighted racial and ethnic disparities in OA prevalence and progression and explored environmental influences on the disease [ 69 , 70 ]. The Hertfordshire Cohort Study linked early-life factors to later-life OA, underscoring the roles of nutrition and physical activity [ 60 , 71 ]. The TASOAC assessed genetic and environmental risk factors for OA, finding strong associations with bone health and physical activity [ 72 , 73 ]. Finally, the Netherlands' NEO Study examined the impact of obesity on OA, linking measures of adiposity with joint damage and pain [ 74 ]. The above longitudinal studies offered insights into the disease's progression. Magnetic Resonance Imaging (MRI) was particularly effective at detecting early cartilage changes and bone marrow edema, often correlated with other clinical manifestations [ 75 ].

OA Changes from Spatial Perspective

Clinical Imaging

Radiographic and clinical imaging techniques are indispensable in understanding the spatiotemporal dynamics of OA progression [ 76 ]. Traditional X-rays remain fundamental, offering valuable insights into joint space narrowing, subchondral bone sclerosis, and osteophyte formation, all hallmark features of OA. In contrast, advanced imaging modalities like MRI provide detailed visualisation of bone and soft tissue changes, enabling the early detection of cartilage degradation and synovitis, often preceding overt clinical symptoms [ 77 , 78 ]. These techniques capture OA's spatial variability and temporal progression, as evidenced by longitudinal imaging studies and clinical assessments [ 75 , 79 ]. For instance, radiographic assessments have been instrumental in monitoring disease progression, while MRI [ 77 , 78 ], ultrasound [ 79 ], nuclear medicine [ 77 ], and arthroscopy each offer unique perspectives on OA progression.

One notable study utilising the MOST database evaluated 3,026 individuals, revealing that even in "end-stage" OA (Kellgren-Lawrence grade 4), MRI detected ongoing cartilage loss, fluctuating bone marrow lesions, effusion, and synovitis over 30 months [ 80 ]. Such findings underscore the temporal utility of imaging in tracking the evolving pathology within the joint environment. Advances in machine learning have further enhanced imaging applications; for example, a classifier achieved 78% accuracy in predicting symptomatic OA progression three years before symptoms emerged [ 51 ]. Biochemical patterns of cartilage fissuring, identified through MRI, provide additional predictive markers, enriching the clinical understanding of disease onset and progression. Spatially, imaging tools delineate affected from unaffected areas within a joint, offering a comprehensive disease map.

Radiomics

Developing state-of-the-art techniques has propelled radiomics to the forefront of medical image analysis in OA research. Radiomics involves the high-throughput extraction of quantitative imaging features to assist in diagnosis, prognosis, and treatment planning [ 81 , 82 ]. A recent study introduced and validated the Joint Space Radiology Model (JS-RM) for predicting knee OA incidence. This model integrates radiological features of the meniscus and femorotibial cartilage derived from MRI and demonstrated high predictive accuracy, achieving an area under receiver operating characteristic curve (AUC) of 0.931, sensitivity of 84.4%, and specificity of 85.6% in a test cohort [ 83 ]. Additionally, the model improved the diagnostic performance of resident doctors in interpreting MRI scans, showcasing its potential as a clinical tool. Another recent study utilised radiomics analysis of patellofemoral joint imaging data from the MOST cohort, demonstrating that integrating radiomics features significantly improved knee replacement risk prediction models (AUC increased from 0.74 to 0.81), highlighting the potential of advanced imaging techniques for personalised OA management. Other radiomics studies leveraging X-rays [ 84 - 87 ], Computed Tomography (CT) [ 88 , 89 ], and MRI [ 90 - 92 ] have focused on disease classification [ 84 - 86 , 89 , 90 ], diagnosis [ 87 , 88 , 91 ], and progression prediction [ 91 , 92 ], further illustrating the versatility of this approach.

Functional Imaging

Functional imaging techniques are also gaining traction in OA research. Functional MRI [ 93 ], optical coherence tomography (OCT) [ 94 - 96 ], positron emission tomography-magnetic resonance imaging (PET-MRI) [ 97 ], and single-photon emission computed tomography (SPECT) [ 98 ] hold promise for dynamically tracing metabolic and inflammatory changes in OA. These techniques provide insights into the functional aspects of joint pathology, complementing structural assessments. However, additional research, standardisation of imaging feature definitions, and expanded databases are required to fully unlock their potential in early OA detection and clinical application [ 99 ].

Spatiotemporal Heterogeneity Provides the Basis for OA Classification

OA is a highly complex and heterogeneous disease, exhibiting significant variability across spatial and temporal dimensions. Multiple prognostic risk factors contribute to its progression, including education level, physical activity, pain severity, and bone edema observed through MRI [ 75 , 79 , 100 ]. Beyond these localised factors, OA’s interplay with systemic comorbidities has led some researchers to advocate for renaming it “systemic OA,” emphasising its broader systemic context and multifactorial etiology [ 101 ]. This complexity is visually represented in Figure 2 , which illustrates the interconnected nature of these factors.

The classification of OA into phenotypes has been an area of active exploration, though a universally accepted standard still needs to be discovered. Several foundational studies have laid the groundwork for such classifications, combining clinical imaging, systematic reviews, and expert consensus [ 9 , 102 , 103 ]. These efforts have divided OA into subtypes based on medical history, pathophysiological features, and imaging findings [ 104 - 108 ]. However, as our understanding of OA’s etiology deepens, research increasingly focuses on integrating cross-scale data and addressing endotype (molecular or cellular features) and phenotype (observable traits) [ 109 - 115 ]. Table 2 provides a detailed overview of these emerging insights.

One noteworthy example is applying the Subtype and Stage Inference (SuStaIn) model to OA classification [ 9 ]. Using data from the OA Initiative (OAI) database, this machine learning technique employed X-rays and MRI imaging to identify three distinct knee OA subtypes. These subtypes include early pain (15% of cases), structural damage with concurrent pain (61%), and late pain (24%). Each subtype reflects unique clinical symptoms and structural damage patterns at different stages of disease progression. This classification approach underscores the potential of advanced computational models to unravel the complexity of OA heterogeneity. Beyond the SuStaIn model, a systematic review has proposed additional phenotype categorisations based on OA’s leading risk factors [ 116 ]. These include six major phenotypes: (1) chronic pain dominated by central mechanisms like central sensitisation, (2) inflammatory phenotypes characterised by elevated inflammatory biomarkers, (3) metabolic syndrome-associated phenotypes linked to obesity, diabetes, and metabolic disturbances, (4) local bone and cartilage metabolic alterations, (5) mechanical overload-driven phenotypes predominantly associated with varus malalignment and medial compartment disease, and (6) minimal joint disease phenotypes marked by mild symptoms and slow progression over time. Similarly, Ali et al. have proposed an alternate phenotypic framework emphasising pathophysiological drivers. Their classification includes six subtypes: (1) aging-driven phenotype, (2) cartilage-driven phenotype, (3) metabolic phenotype, (4) synovitis-driven inflammatory phenotype, (5) traumatic injury-driven phenotype, and (6) subchondral bone phenotype [ 117 , 118 ]. This categorisation highlights the diverse mechanisms underlying OA progression and aligns with growing interest in personalised medicine. A recent addition to OA classifications is the accelerated knee OA subtype. This rapidly progressing form of OA is characterised by a transition from no radiographic evidence to advanced disease within four years [ 119 ]. Researchers have contextualised this subtype by comparing it to typical, slower-progressing OA cases and exploring potential factors such as aging, previous case reports, and relevant animal models. Despite these advancements, the field still lacks widely recognised and validated subcategories. Achieving this requires a multidisciplinary effort, integrating insights from diverse fields such as molecular biology, imaging, epidemiology, and clinical practice. A consensus-based approach, supported by input from field experts and robust validation, is essential for defining precise subcategories. This is critical for enhancing early diagnosis and accurate disease management and reducing the overall burden of OA on patients and healthcare systems.

Challenges in Integrating Cohort Data and Radiomics in OA

While promising, the above imaging strategies face several limitations that hinder their widespread adoption and effectiveness. A key issue is the lack of standardisation across feature extraction, data processing, and model evaluation, which complicates comparisons between studies and institutions. Additionally, the availability of high-quality, annotated datasets is limited, with small sample sizes and insufficient diversity reducing generalizability. Variability in imaging protocols, equipment, and settings between institutions introduces inconsistencies that challenge validation and clinical implementation. Computational complexity further adds to these challenges, as advanced models demand significant resources and expertise that are not universally accessible. Despite advancements, clinical integration still needs to be improved. Radiomics models often lack real-world validation and seamless incorporation into workflows, leading to hesitation among clinicians, mainly due to the "black box" nature of many machine learning models. Overfitting and reproducibility issues persist, especially in studies with small datasets, emphasising the need for more extensive and varied cohorts to ensure robust findings. Advanced imaging techniques, such as PET-MRI or OCT, come with their own set of challenges. These include high costs, specialised equipment requirements, and limited accessibility in routine healthcare settings. Furthermore, functional imaging techniques often lack extensive longitudinal data, particularly for OA research. Furthermore, radiomics features require further validation to confirm their clinical relevance and standardised imaging databases are essential for accurate modelling and inter-institutional validation. Addressing these limitations necessitates collaborative efforts focused on standardising protocols, enhancing dataset quality, improving interpretability, and validating results across diverse clinical environments. Such advancements will be critical for transitioning radiomics and clinical imaging into routine clinical practice, fulfilling their potential in early and precise OA diagnostics.

In summary, while substantial progress has been made in understanding OA’s heterogeneity and identifying phenotypes, many challenges remain. Future efforts should prioritise refining these classifications, validating them through cross-disciplinary collaboration, and leveraging advanced methodologies to integrate endotypes and phenotypes. Such advancements will pave the way for more targeted interventions, improving outcomes for individuals with OA.

Spatiotemporal Development Trajectory in OA at The Micro Level

Over recent decades, researchers have sought to delineate the development trajectories of OA at the histopathological level by employing various grading systems. These systems primarily fall into two categories: those following sequential stages of increasing OA severity [ 120 , 121 ] and those aggregating independent indicators of OA severity [ 122 , 123 ]. The Modified Mankin score, and the Osteoarthritis Research Society International (OARSI) grading system are the most renowned. Scientists have successfully built various OA grading systems in humans [ 121 ], horses [ 124 ], dogs [ 125 ], rabbits [ 126 ] and rats [ 127 ], and the grading systems have covered most of the tissues and cell types changed within the joint. The elements involved in the changes include cartilage, tidemark, proteoglycan, collagen integrity, subchondral bone plate, meniscus, synovium, and osteophytes, which can be seen via histopathological staining.

Cartilage

Normal: At the histopathological level, cartilage is the core tissue affected during OA progression. It is a highly differentiated tissue that maintains a high regeneration capacity during development and architecture after birth but loses its regenerative capacity during OA progression [ 128 - 130 ]. Therefore, genetic and transcriptomic pivots during development and growth are crucial to understanding OA changes [ 131 - 133 ]. Chondrogenesis is a well-coordinated process essential for embryonic development, adult cartilage maintenance, and vertebrate repair. Chondrocytes' fate determination and differentiation involve the differential expression of genes crucial at each stage of chondrogenesis [ 134 ]. During chondrogenesis, SRY-Box Transcription Factor 9 (SOX9) was believed to be crucial in endochondral ossification and bone elongation [ 135 ], owing to its key function in the initiation of chondrocyte differentiation [ 136 ], activation of cartilage matrix genes [ 137 , 138 ], maintenance of chondrocyte phenotype [ 136 ], inhibition of hypertrophy and osteogenic differentiation [ 138 , 139 ] and interaction with other factors [ 137 , 138 ]. A recent in situ study showed that SRY-Box Transcription Factor 8 (SOX8) is also expressed at levels comparable to SOX9 in reserve and early columnar chondrocytes [ 140 ]. The study showed that SOX8 expression stops as SOX9 peaks in late columnar and hypertrophic chondrocytes, and both SOX8 and SOX9 are essential for skeletal growth by promoting the conversion of growth plate chondrocytes to actively proliferating columnar cells [ 140 ].

Modern Technology, including single-cell RNA-sequencing (scRNA-seq) [ 141 ] and spatial transcriptomics [ 142 - 144 ], has improved our understanding of cartilage structure and its cell subpopulation at the spatial level. The distribution of chondrocyte subpopulations in normal cartilage reveals distinct regional preferences. Prehypertrophic chondrocytes (preHTC), inflammatory chondrocytes (InfC), pre-inflammatory chondrocytes (preInfC), and fibrocartilage chondrocytes (FC) predominantly occupy the surface and superficial layers [ 144 ]. In contrast, proliferative chondrocytes (ProC), hypertrophic chondrocytes (HTC), regulatory chondrocytes (RegC), and pre-inflammatory chondrocytes (preInfC) are significantly more abundant in these regions compared to the middle and deep layers. Spatial transcriptomics shows that preHTC, InfC, preInfC, and FC cells are abundant in weight-bearing (WB) and non-weight-bearing areas but less in normal controls. Hypertrophic chondrocytes (HTC), enriched with genes related to extracellular matrix organisation and ossification, are significantly more prevalent in the deep cartilage region. Similarly, RegC cells, associated with degradation processes such as amino sugar metabolism, are also more abundant in the deep area.

OA: In the early stage of OA, cartilage retains its surface integrity but may exhibit superficial fibrillation, such as microscopic cracks [ 121 ]. This phase also shows signs of matrix edema, leading to alterations in matrix density and staining properties with cationic dyes, reflecting early proteoglycan changes. Chondrocyte proliferation and apoptosis may be observed, particularly in superficial layers [ 145 , 146 ]. Progression to the next stage involves the focal loss of the superficial matrix, evidenced by the spallation of small matrix fragments and the beginning of surface layer discontinuity. This stage may also display deep fibrillation extending through the superficial zone, with matrix staining changes indicating altered biochemical composition [ 145 ]. As OA advances, fissures extend into the mid zone, forming vertical clefts that may branch. This results in more pronounced heterogeneity in matrix composition adjacent to fissures, with areas of proteoglycan depletion and increased staining. Chondrocyte death and complex chondron formations become more apparent [ 147 , 148 ]. These processes are often accompanied by increased chondrocyte death and matrix degradation. The most advanced stage is characterised by total loss of hyaline cartilage, exposing the underlying bone, which may exhibit increased density and thickness [ 121 ]. This stage may also involve fibrocartilage repair, which can fill the eroded cartilage volume. In the final stage, significant deformation of the joint surface occurs due to continued bone microfractures and reparative changes, including the formation of osteophytes. These osteophytes and fibrocartilaginous proliferation contribute to altering joint contour and function.

State-of-the-art techniques, including scRNA-seq [ 141 ] and multi-omics data integration [ 144 ], provided further evidence of chondrocyte cell population changes ( Table 3 ). In early OA, proliferative chondrocytes (ProCs) increase in the superficial zone as part of an initial reparative response. However, as OA progresses, the population of ProCs declines, resulting in reduced proliferative potential. Effector chondrocytes (ECCs), involved in cartilage matrix synthesis, exhibit a modest increase in the middle zone during early OA but remain relatively stable in advanced OA without significant functional changes. Homeostatic chondrocytes (HomCs), critical for maintaining protein folding and overall homeostasis, decrease in the deep zone during early OA and become absent in advanced OA, reflecting a loss of homeostatic capacity. Regulatory chondrocytes (RegCs), responsible for modulating catabolic processes, also decrease in the superficial zone during early OA and are minimally present in advanced OA as other cell types take over catabolic functions. Reparative chondrocytes (RepCs), which contribute to extracellular matrix (ECM) repair, show a mild increase in the SZ and MZ during early OA in response to initial cartilage damage. However, in advanced OA, RepCs are significantly reduced as they become overwhelmed by extensive matrix degradation. PreHTCs shift from the deep layer to the joint surface in the early stages of OA, exhibiting early ossification signals mediated by activation of TGF-β pathways. In advanced OA, preHTCs markedly increase in the articular surface and middle zone, promoting hypertrophic differentiation. HTCs, involved in ossification, show mild activation in the deep zone during early OA but become overabundant in both the deep zone and articular surface as OA progresses, contributing to significant matrix degradation and calcification. Fibrocartilage chondrocytes (FCs) emerge in the articular surface and middle zone in early OA, contributing to increased fibrocartilage-specific markers and eventually dominating the middle zone and deep zone in advanced OA, resulting in increased matrix stiffness and maladaptive repair processes. Prefibrocartilage chondrocytes (preFCs), which play a role in early ECM remodelling, are more prevalent in the AS during early OA but decline in advanced OA as hypertrophic or inflammatory chondrocytes replace them. Pre-inflammatory chondrocytes (preInfCs), absent in normal cartilage, emerge in the articular surface and superficial zone during early OA and are linked to apoptosis via IFI27 expression. In advanced OA, preInfCs differentiate into inflammatory chondrocytes (InfCs), which dominate the middle zone, contributing to inflammatory cascades, including macrophage migration inhibitory factor (MIF)-cluster of differentiation 74 (CD74) signalling [ 144 ].

Synovium

Normal : Under physiological conditions, the synovium consists primarily of two major synoviocyte types: Type A and Type B synoviocytes [ 149 ]. Type A synoviocytes, also called macrophage-like synoviocytes, are mainly involved in phagocytosis and debris clearance. In contrast, Type B synoviocytes, which are fibroblast-like, are responsible for synthesising synovial fluid and maintaining extracellular matrix (ECM) components, including collagen. These synoviocytes maintain synovial homeostasis, providing an optimally lubricated environment crucial for joint function.

OA : In OA progression, the synovium undergoes significant spatiotemporal changes, encompassing alterations in cellular subpopulations and their functional dynamics ( Table 4 ). As OA progresses, multiple macrophage subpopulations emerge within the synovium. In early OA, Transitional Macrophages (T-Mϕ) are important in initiating inflammatory responses. These macrophages express markers with both pro-inflammatory and anti-inflammatory properties, highlighting their transitional nature. In advanced OA, T-Mϕ exhibits increased activity, contributing to persistent and chronic inflammation. Another subpopulation, Fibrotic Immune-Regulated Macrophages (IR-Mϕ), is implicated in tissue repair and immune regulation. Their presence is limited in early OA but becomes more prominent in advanced stages, significantly contributing to synovial fibrosis. Interferon-stimulated macrophages (IFNS-Mϕ), are moderately abundant in early OA and increase in presence as OA progresses, exacerbating inflammatory responses. S100A8/9hi Macrophages (S100A8/9hi-Mϕ) are involved in auto debridement, facilitating the clearance of damaged tissue. Their levels increase significantly in advanced OA, corresponding with ongoing tissue damage and remodelling. Type B synoviocytes are localised in the synovial lining layer. During early OA, synovial fibroblasts activate, proliferate, and secrete ECM proteins that facilitate tissue repair. They also mediate inflammation by secreting pro-inflammatory cytokines. In advanced OA, fibroblast activity is marked by increased production of fibrotic ECM, leading to synovial fibrosis. These fibroblasts also contribute to chronic inflammation through enhanced secretion of inflammatory cytokines, such as IL-6 and Interleukin 8 (IL-8), which promote synovial hyperplasia and joint stiffness. Cluster of Differentiation 4 (CD4)+ T cells are key modulators of immune responses and differentiate into various helper subtypes, including Type 1 T-helper cell (Th1), Type 2 T-helper cell (Th2), Type 17 T-helper cell (Th17), and regulatory T cells (Tregs). In early OA, naive and memory CD4+ T cells are predominant. In contrast, in advanced OA, increased activation occurs, with regulatory subtypes (e.g., Tregs) playing a more significant role in immune regulation. CD8+ T cells also exhibit increased expression of cytotoxic and proliferative markers during early OA. They are predominantly found as effector memory T cells in advanced OA, suggesting an ongoing immune response. In early OA, B cells are present at relatively low levels, indicating a minimal contribution to the early inflammatory milieu—however, their abundance increases in advanced OA, suggesting an enhanced role in sustaining chronic inflammation. In summary, the synovium undergoes dynamic cellular alterations as OA progresses, with different synoviocyte populations contributing to inflammation, fibrosis, and tissue remodelling.

Subchondral Bone Plate

Normal : Under normal physiological conditions, subchondral bone exhibits a low turnover rate, which is essential for maintaining a stable microenvironment and supporting overall joint function. Osteocytes, residing within the bone matrix, are necessary for bone maintenance and mechanosensation. Regulating bone remodelling through signalling pathways involves markers such as sclerostin (SOST) and dentin matrix protein 1 (DMP1). The cellular constituents of subchondral bone, including osteoblasts (OBs), osteoclasts (OCs), mesenchymal stem cells (MSCs), endothelial cells (ECs), and immune cells, are typically maintained in a homeostatic state with minimal pathological activity [ 150 ].

The subpopulations of cells within the subchondral bone reside in specific niches, reflecting their specialised roles. Endothelial cells are located in the vascular regions of the subchondral bone, with two subpopulations identified: kinase insert domain receptor (KDR)-negative cells, which are associated with inflammation, and KDR-positive cells, which are involved in angiogenesis. Osteoblasts are present in different regions depending on their specific subtype. Endothelial osteoblasts (EnOBs) are found near blood vessels and contribute to vascularisation, stromal osteoblasts (StOBs) are located in areas of extracellular matrix production, and mineralising osteoblasts (MinOBs) are found in regions of active mineral deposition. Mesenchymal stem cells are typically located near areas of bone damage and early lesions, where they differentiate into osteoblasts or chondrocytes, contributing to bone formation and repair. These distinct subpopulations occupy different niches within the subchondral bone, ensuring proper bone function under normal physiological conditions.

OA : The progression of OA involves cellular and subpopulation changes within the subchondral bone, leading to significant alterations in bone structure and function [ 151 ] ( Table 5 ). In the initial stages of OA, osteoblasts have mild activation, contributing to early bone thickening and subchondral sclerosis. Osteoblasts at this stage are characterised by the upregulation of markers involved in bone formation and early osteophyte development [ 152 ]. Osteoclasts exhibit increased resorptive activity in response to microdamage [ 153 ]. This resorption is necessary to remove damaged bone and facilitate subsequent remodelling. MSCs are activated in early lesions, differentiating into osteoblasts to facilitate early repair processes and promote initial bone formation. The recruitment and differentiation of MSCs are driven by markers such as melanoma cell adhesion molecule (MCAM) and CD146, which are critical in osteoblast lineage commitment [ 154 ]. Endothelial cells (ECs) exhibit early angiogenic responses around areas of bone damage, enhancing nutrient supply to the affected regions through increased expression of angiogenesis markers [ 155 ]. Immune cells, including T cells, B cells, and macrophages, are present in limited numbers, contributing to localised cytokine production that remains relatively minor at this stage. However, this early immune response sets the stage for further immune-mediated processes exacerbating disease progression. In advanced OA, osteoblast activity becomes markedly increased, leading to the formation of osteophytes and significant subchondral bone sclerosis. Osteoblasts at this stage are highly active, contributing to extensive bone formation and mineralisation, often leading to joint deformities. Osteoclast activity persists, resulting in substantial bone resorption, increased microdamage, and extensive remodelling. Persistent osteoclast activity is marked by elevated levels of catabolic enzymes, particularly cathepsin K, which is essential for bone matrix degradation and contributes to the pathological remodelling of subchondral bone. Endothelial cells undergo pronounced angiogenesis, enhancing nutrient supply that supports inflammation and advanced bone remodelling. This angiogenesis is marked by upregulated vascular endothelial growth factor (VEGF) signalling and increased formation of vascular structures, which further contributes to subchondral bone pathology. There is heightened recruitment of MSCs, which accelerates osteoblast differentiation and fosters osteophyte formation, contributing to joint deformity. MSCs are recruited and differentiated at an increased rate, driven by an altered microenvironment characterised by inflammatory signals and mechanical stress.

Recent studies using single-cell RNA sequencing have revealed significant changes in the subpopulations of cells within the subchondral bone during OA progression [ 151 ]. Two distinct subpopulations in the endothelial cell compartment have been identified: kinase insert domain receptor (KDR)-negative endothelial cells characterised by an inflammatory response and KDR-positive endothelial cells associated with angiogenesis. The KDR-positive endothelial cells become increasingly prominent in advanced OA, indicating aberrant vascularisation. Osteoblasts can be further divided into endothelial osteoblasts (EnOBs), stromal osteoblasts (StOBs), and mineralising osteoblasts (MinOBs). Endothelial osteoblasts are involved in vascularisation, stromal osteoblasts contribute to extracellular matrix production, and mineralising osteoblasts are critical for mineral deposition. The relative abundance of these osteoblast subpopulations changes during OA progression, with endothelial osteoblasts predominating at early stages and mineralising osteoblasts becoming more prevalent in advanced stages as bone mineralisation and sclerosis progress. Mesenchymal stem cells also exhibit increased heterogeneity, with specific subpopulations differentiating more actively into osteoblasts or chondrocytes in response to inflammatory signals and mechanical stress. These changes in subpopulations reflect a dynamic remodelling process aimed at repairing damage, leading to pathological bone changes and joint deterioration. Additionally, there is extensive infiltration of immune cells, with significant production of pro-inflammatory cytokines, including tumour necrosis factor-alpha (TNF-α) and interleukin-1 beta (IL-1β), that exacerbate osteoclast-mediated bone resorption and remodelling. These cytokines further drive the inflammatory cascade, disrupting bone homeostasis and promoting chronic inflammation. In summary, targeting the distinct cellular and subpopulation changes within the subchondral bone may offer promising avenues for therapeutic interventions to prevent or slow the progression of OA.

Infrapatellar Fat Pad

Normal : The infrapatellar fat pad (IPFP) is a unique white adipose tissue in the knee joint, below the patella and behind the patellar ligament [ 156 ]. Under physiological conditions, the IPFP serves critical biomechanical and homeostatic functions, including acting as an impact dampener, facilitating the dispersion of synovial fluid, and contributing to joint stability. It is a highly vascularised tissue comprising various cellular populations, such as adipocytes, fibroblasts, immune cells, and endothelial cells. These diverse cell types are integral to maintaining knee joint integrity by modulating inflammatory responses, facilitating tissue repair, and providing structural support. The IPFP's close anatomical and functional association with the synovium forms an integrated unit that plays a vital role in joint lubrication and movement, underscoring its significance in overall knee health.

OA : During OA progression, the IPFP undergoes profound cellular and molecular changes contributing to the pathogenesis of OA [ 156 ]. OA induces a profibrotic and inflammatory phenotype in the IPFP, marked by increased fibrosis, chronic low-grade inflammation, and altered cellular heterogeneity. Intermediate fibroblasts emerge as key drivers of OA-associated fibrosis, while macrophages and other immune cells exacerbate inflammation by releasing pro-inflammatory cytokines. The transcriptomic landscape of the IPFP shifts markedly, with the upregulation of genes associated with fibrosis and inflammation, including those involved in the transforming growth factor-beta (TGF-β) signalling pathway [ 157 ]. These molecular alterations, combined with increased adipokines and other inflammatory mediators, exacerbate cartilage degradation and OA progression, highlighting the IPFP's pivotal role in OA pathogenesis.

Spatiotemporal changes at the micro level in the IPFP during OA progression further elucidate the pathological transformation of this tissue [ 157 , 158 ] ( Table 6 ). In early OA, precursor-like fibroblasts are distributed throughout the IPFP and contribute to tissue repair, shock absorption, and joint support. However, these fibroblasts exhibit increased expression of fibrosis-associated markers and inflammatory activation as OA advances. In late-stage OA, these cells reach an advanced differentiation state, displaying elevated profibrotic and inflammatory gene expression levels. Macrophages within the IPFP also undergo significant changes during OA progression. Homeostatic macrophages are involved in immune surveillance and maintaining internal homeostasis in normal conditions. In early OA, macrophages exhibit upregulation of pro-inflammatory cytokines, and as OA advances, markers are further upregulated and associated with pro-inflammatory and tissue remodelling functions. Adipocytes within the IPFP also display notable alterations during OA. In the early stages of OA, adipocytes show increased secretion of pro-inflammatory adipokines. In advanced OA, these cells demonstrate dysfunction, characterised by alterations in lipid metabolism that contribute to the inflammatory environment within the joint. Endothelial cells (EDCs) regulate vascular function and support tissue perfusion. During early OA, EDCs demonstrate increased expression of angiogenesis-related markers. In advanced OA, aberrant angiogenesis and elevated expression of inflammatory markers are evident, further exacerbating the inflammatory environment and contributing to pathological tissue remodelling within the IPFP.

Challenges in Combining scRNA-Seq and Spatial Transcriptomic Data in OA Progression

While scRNA-seq and spatial transcriptomics have revolutionised our understanding of cellular heterogeneity and spatial organisation in OA, integrating these two datasets presents significant challenges. Data integration requires sophisticated computational tools to align datasets with differing resolutions and scales. This complexity is exacerbated by sample preparation and processing protocol variations, which can introduce technical biases. Secondly, the high dimensionality of these datasets often necessitates advanced statistical frameworks and machine learning models, which require expertise and computational resources that may not be universally accessible. Lastly, the lack of standardised pipelines for data integration hinders reproducibility and cross-study comparisons. Addressing these challenges will require collaborative efforts to develop robust computational frameworks, standardised protocols, and accessible tools for integrating multi-omics data in OA research.

Spatiotemporal Development Trajectory in OA from the Molecular Perspective

Genetic and Epigenetic Polymorphism

Genetic and epigenetic factors play a crucial regulatory role in the OA spatiotemporal development trajectory. Genetic variations and epigenetic modifications collectively influence an individual's susceptibility to OA, the rate of disease progression, and the specific characteristics of affected joints. Studies on gene variations, such as collagen type II alpha 1 chain (COL2A1) [ 159 , 160 ] and growth differentiation factor 5 (GDF5) [ 161 - 163 ], have enabled scientists to identify individuals more prone to OA, determining the age of onset and the severity and progression patterns of the disease. Simultaneously, epigenetic mechanisms, such as DNA methylation, histone modifications, and interactions with non-coding RNA, offer a meticulous gene regulation capability of adjusting to environmental stimuli. These genetic and epigenetic factors significantly impact the pathophysiology of OA by modulating the expression of key catabolic enzymes and inflammatory mediators, driving cartilage degradation and synovial inflammation ( Table 7 ). The detailed review is summarised below:

Genetic polymorphisms are significant in the pathogenesis of OA. Polymorphisms in specific genes are linked to the risk and severity of OA. For instance, Toll-like receptor 9 (TLR9) gene rs187084 polymorphism is linked to a heightened risk of knee and hip OA [ 164 , 165 ]. In contrast, Toll-like receptor 10 (TLR10) gene rs11096957 polymorphism is linked to increased hip OA risk and severity [ 166 ]. The RAGE gene 82G/S polymorphism is also associated with an increased risk of knee OA [ 167 ]. Polymorphisms in cytokine and enzyme-related genes, such as interleukin-1 beta (IL-1β), Interleukin 1 Receptor Antagonist (IL-1RN), Interleukin 18 (IL-18), IL-6, MMP-1, Matrix Metallopeptidase 13 (MMP-13), and ADAM Metallopeptidase with Thrombospondin Type 1 Motif 14 (ADAMTS14), are also associated with increased OA risk and severity [ 30 ]. Other genes like B-cell Lymphoma 2 (BCL-2) [ 168 ], C-X-C Motif Chemokine Ligand 16 (CXCL16) [ 168 ], ITLN1/omentin [ 35 ], DOT1L [ 35 ], and matrix Gla protein [ 169 ] also show increased risk association with OA.

Epigenetic regulation, such as histone modifications, can influence gene expression, thereby regulating inflammatory responses and cartilage metabolism, contributing to the development of OA. For example, the enhancer of zeste 2 polycomb repressive complex 2 subunit (EZH2) promotes OA progression by methylating histone H3 lysine 27 (H3K27) to suppress miR-138. It can also regulate matrix metalloproteinases and inflammatory factors independently of its enzymatic activity [ 170 ]. The significant function of epigenetic mechanisms in the development of OA calls for additional research, such as exploring the interplays between genetic and epigenetic factors.

DNA methylation is a crucial epigenetic regulatory mechanism playing a significant role in OA pathogenesis. In OA patients, genes encoding collagen molecules undergo hypermethylation, inhibiting their transcription and promoting OA development. Conversely, demethylation of promoter regions of inflammatory mediator genes leads to their increased expression [ 171 ]. Differentially methylated enhancer regions between OA patients and normal individuals have also been identified. Therefore, DNA methylation regulates OA-related gene transcription, contributing significantly to OA pathogenesis.

Histone modifications also play essential roles in OA pathogenesis. Histone acetylation and deacetylation are critical epigenetic regulatory mechanisms. Histone deacetylase 2 (HDAC2) and Histone deacetylase 4 (HDAC4) regulate molecules related to inflammation and matrix degradation in OA [ 30 ]. Histone deacetylase inhibitors were shown to regulate immune response, stabilise mitochondria, and stabilise microtubules, which play a protective role in OA progression [ 172 , 173 ]. Histone methylation is also involved in OA pathogenesis. Histone H3 lysine 9 (H3K9) methylation is typically associated with gene expression suppression, while Histone H3 lysine 4 (H3K4) methylation promotes gene expression [ 174 ]. EZH2, a histone methyltransferase, catalyses the methylation of histone H3 on lysine 27 (H3K27), which is usually associated with transcriptional repression [ 170 ]. However, EZH2 may also influence gene expression through other mechanisms, such as DNA methylation and hydroxymethylation [ 175 ]. Furthermore, upregulation of various histone demethylases, such as lysine-specific demethylase 2A (KDM2A), lysine-specific demethylase 6A (KDM6A), and lysine-specific demethylase 7A (KDM7A), has been observed in OA patients. Inhibiting Histone Nε-methyl lysine demethylases (KDM2/7) can increase lysine 79 of histone H3 (H3K79) methylation, inhibiting cartilage damage [ 176 ].

Non-coding RNAs, including long non-coding RNAs (lncRNA), circular RNAs (circRNA), and small non-coding RNAs (sncRNA) like microRNAs (miRNA), small interfering RNAs (siRNA), and piwi-interacting RNAs (piRNA), are crucial genetic regulatory factors in OA [ 177 ]. These non-coding RNAs regulate gene expression through various mechanisms [ 178 ]. For instance, miRNAs can post-transcriptionally inhibit gene expression, while lncRNAs and circRNAs can act as competing endogenous RNAs or sponges. Research has shown that non-coding RNAs play significant roles in the progression or inhibition of OA. For example, miR-17 can inhibit the expression of Nitric oxide synthase 2 (Nos2), ADAM Metallopeptidase With Thrombospondin Type 1 Motif 5 (ADAMTS5), MMP-3, and MMP-13, protecting cartilage [ 179 ]. miR-149 can inhibit inflammation by suppressing Vascular cell adhesion protein 1 (VCAM-1) [ 180 ], while miR-34a-5p can reduce Collagen Type II Alpha 1 Chain (COL2A1) and aggrecan expression and increase MMP-13 and ADAMTS5 expression, promoting OA progression [ 181 ]. Non-coding RNAs interact with histone-modifying enzymes; for instance, miR-222 targets HDAC4, inhibiting MMP-13 expression [ 182 ]. Polymorphisms and methylation of non-coding RNA genes can also affect their expression and function, contributing to OA pathogenesis. Specific miRNAs, such as miR-23a-3p, miR-146a-5p, and miR-652-3p, show elevated expression levels in OA patients, potentially serving as diagnostic markers for OA [ 183 ]. Consequently, non-coding RNAs are essential genetic regulatory elements that participate in the pathogenesis of OA via diverse mechanisms. The above genetic and epigenetic factors could serve as clinical biomarkers to distinguish high-risk individuals for OA. However, the above spatiotemporal changes of the genetic and epigenetic change remain to be discovered, especially for the tissue-specific change. Conflicting study results might be attributed to sample size or population differences. Some genetic variants might be less common in specific populations, necessitating large-scale studies across diverse groups to evaluate their precise impact. A comprehensive understanding of these factors' spatiotemporal variations across different populations and individuals offers new perspectives for precision medicine, aiding in developing personalised therapeutic strategies tailored to specific genetic and epigenetic profiles, thus enhancing the efficacy of OA treatments.

Biochemical Pathways

The biochemical landscape of OA is dominated by a cascade of molecular pathways that contribute significantly to the cartilage degradation and inflammation characteristic of the disease even before the occurrence of visible radiographic changes. In a cohort of 200 women, 24 serum protein biomarkers were identified to predict knee OA before radiographic abnormalities. The predictions based on these blood biomarkers outperform traditional predictions based on age and BMI (AUC 51%) or knee pain (AUC 57%). Among them, the majority (58%) also predict the progression of knee OA. This discovery implies that the molecular pathophysiology beneath incident OA and progressive knee OA is similar, thus uncovering the presence of a pathophysiological "OA continuum". [ 64 ].

Traditional biochemical players in these pathways include cytokines such as Interleukin-1 beta (IL-1β) and tumour necrosis factor-alpha (TNF-α), which initiate and perpetuate inflammatory processes within joint tissues [ 184 , 185 ]. Alongside these cytokines, MMPs are crucial for degenerating extracellular matrix components, directly leading to cartilage erosion [ 186 ]. These biochemical mediators do not act uniformly across all affected joints or throughout all stages of OA. Instead, their expression and activity can vary significantly, contributing to the temporal dynamics of disease progression—where early stages may show mild symptoms that escalate as the biochemical activity intensifies. Spatial variations are also evident, as different joints may exhibit distinct biochemical signatures depending on the mechanical stresses they endure and their biological responses. Understanding these spatiotemporal changes in biochemical pathways is essential for developing targeted therapeutic interventions that can potentially halt or reverse the progression of OA at various stages and in specific joints.

With the emergence of more research cohorts in OA research, researchers started to utilise multiple biomarkers capable of tracking the temporal changes of OA to classify the endotypes. Recently, a study employed machine learning techniques to analyse the biochemical marker data from the IMI-APPROACH cohort to identify different subtypes of OA [ 67 ]. This innovative and forward-looking research method provides new ideas and means for the precise classification of OA. The IMI-APPROACH is a prospective cohort study that includes 297 knee OA patients who meet the American College of Rheumatology classification criteria. These patients were pre-selected from existing cohorts through a machine learning model that indicated a higher possibility of joint space width reduction and/or the progression of knee pain. Among the 433 OA patients screened from 5 centres, 297 were finally included because they were most likely to experience pain and/or structural progression during the 2-year follow-up period. This screening process is rigorous and scientific, ensuring the pertinence and representativeness of the research subjects. Three main subtypes of OA were identified by analysing 16 biochemical markers related to joint tissue metabolism in the cohort. Low tissue metabolism, low repair ability, and cartilage/bone metabolism levels characterise the C1 subtype. The C2 subtype is associated with structural damage, featuring high bone formation/resorption and cartilage degradation. The C3 subtype involves systemic inflammation, including joint tissue degradation, inflammation, and cartilage degradation. This classification method based on biomarkers provides important clues for a deeper understanding of the pathogenesis of OA [ 67 ]. Significant differences exist among these three subtypes regarding clinical characteristics and disease progression. The C1 subtype has the highest proportion of patients without disease progression, indicating the least degree of disease progression. The C2 subtype is mainly related to the longitudinal structural changes of joints over time. The C3 subtype is linked to persistent or worsening pain symptoms. This finding is significant for clinical treatment and helps doctors formulate personalised treatment plans for different subtypes. The biomarker-based, data-driven subtype identification method may significantly value patient stratification in OA clinical trials and contribute to implementing precision medicine strategies. It can be foreseen that with the continuous progress of technology and the deepening of research, this method will play an increasingly important role in diagnosing and treating OA [ 67 , 187 ].

Meanwhile, data-driven techniques for OA predictions have surged in recent years, which indicates that multiple molecular changes happen before the radiographic changes. Due to the construction of various databases containing multi-omic and radiographic data, some omics biomarkers have replaced traditional clinical risk markers, with several proving to be more predictive of OA risk than BMI. Another latest study incorporated clinical, lifestyle and biomarker data using the UK Biobank database [ 49 ]. They divided OA into 14 subgroups using interpretable machine learning [ 49 ], and they found out that key gene and signalling pathways, including growth differentiation factor 5 (GDF5) and transforming growth factor beta (TFG-β) signalling pathways, play an essential role in OA progression. Cartilage acidic protein 1 (CRTAC1) emerged as the most predictive protein, previously identified as an OA risk biomarker [ 49 , 188 - 190 ]. CRTAC1 is linked to OA diagnosis in various joints and correlates with the severity of OA [ 188 - 190 ]. Pro-inflammatory cytokines are believed to upregulate CRTAC1 in OA-affected joints [ 191 ]. Another significant biomarker, collagen type IX alpha 1 chain (COL9A1), predicts OA diagnosis and has genetic and epigenetic associations with OA [ 191 - 195 ]. Mutations in COL9A1 are also linked to multiple epiphyseal dysplasia, a hereditary condition marked by early-onset OA [ 191 ]. Additionally, actin alpha 2 (ACTA2) has been associated with specific OA subgroups and smooth muscle cell clusters in the OA synovium [ 196 ]. Finally, the ectodysplasin A2 receptor (EDA2R), another predictive protein, has been linked to TNF-mediated inflammation in RA [ 197 ]. The above studies provide a novel perspective for OA's precision prevention and treatment development.

Metabolic Changes

Metabolic changes play a crucial role in the spatiotemporal development trajectories of OA from a molecular perspective. During the early stages of OA, there is a notable shift in chondrocyte metabolism, characterised by increased glycolysis [ 198 ] and a corresponding decrease in oxidative phosphorylation [ 199 ]. This metabolic reprogramming is driven by hypoxic conditions within the cartilage, leading to upregulation of hypoxia-inducible factor 1-alpha (HIF-1α). As OA progresses, the altered metabolic state of chondrocytes results in the accumulation of lactate and other metabolic by-products, contributing to an acidic microenvironment that exacerbates cartilage degradation [ 200 ]. Furthermore, inflammatory cytokines such as IL-1β and TNF-α amplify these metabolic disruptions by inducing nitric oxide synthase (NOS) activity, further impairing mitochondrial function and energy production [ 201 , 202 ]. The dysregulation of anabolic and catabolic gene expression also influences the catabolic processes, including the suppression of collagen type II and aggrecan synthesis and the upregulation of MMPs and aggrecanases. In advanced stages of OA, synovial inflammation and subchondral bone sclerosis add complexity to the metabolic landscape, promoting a feedback loop that perpetuates joint degeneration [ 203 ]. Understanding these metabolic changes at various stages of OA development is essential for identifying potential therapeutic targets to restore metabolic balance and halt disease progression.

Normal : A novel study used matrix-assisted laser desorption/ionisation Fourier transform ion cyclotron resonance mass spectrometry (MALDI-FT-ICR-MS) imaging experiments on human knee OA cartilage for technical validation [ 204 ]. They detected spatial differences in the distribution of metabolites such as adenosine triphosphate (ATP), adenosine diphosphate (ADP), and adenosine monophosphate (AMP). These nucleotide metabolites were mainly concentrated in the superficial zone of the cartilage rather than in the deeper layers. The authors hypothesised that a particular situation could be explained by the fact that the superficial part of the cartilage has direct contact with the synovial fluid and that it is the only region where cartilage progenitor cells have been identified.

OA : Recent advances in the molecular changes of OA from a spatiotemporal perspective mainly focus on spatial metabolomics and lipidomics. A pivotal study examined the femoral heads of patients in advanced OA (grades III-IV), revealing that metabolic phenotypes vary according to the stage of the disease and the specific joint regions affected. Advanced analysis of cartilage and subchondral bone samples from quadrants of the femoral head identified critical metabolic pathways driving these variations. These pathways include lipid metabolism, amino acid metabolism, nucleotide metabolism, and glucose metabolism. Notably, glycosaminoglycan degradation and amino acid metabolism were predominantly concentrated in the upper regions of the femoral head, suggesting that metabolic disruptions in advanced OA are anatomically and pathologically heterogeneous [ 205 ]. OA synovium's lipidomic profile has also exhibited distinctive characteristics compared to healthy synovium and other inflammatory arthropathies such as RA and PsA. In OA synovium, elevated levels of phosphatidylcholine, fatty acids, and lysophosphatidic acid contrasted with reduced levels of lysophosphatidylcholine. Furthermore, spatial distribution analysis revealed that specific glycerol-phospholipid profiles were associated with hypertrophy, inflammation, and vascularisation of the synovium. Interestingly, OA tissues exhibited lower levels of phosphatidylethanolamine plasmalogens compared to RA and PsA counterparts. These findings emphasise the unique lipidomic signature of OA synovium and its potential utility in diagnostic and therapeutic applications [ 206 ]. These investigations provide invaluable insights into OA's complex metabolic and lipidomic adaptations. The findings may shed light on developing precise diagnostic tools and targeted interventions, paving the way for enhanced patient care and improved disease management strategies.

Challenges in Combining Existing Spatial-Omics Data in OA Progression

Combining spatial-omics data in OA progression, including spatial transcriptomics, proteomics, metabolomics, elemental mapping, and mechanical mapping, presents several challenges. Each omics layer provides unique insights—transcriptomics reveals gene expression, proteomics shows protein localisation and metabolomics tracks biochemical processes. Integrating these diverse datasets while maintaining spatial accuracy is challenging. The overlap between molecular data from different layers is not always perfect, making it challenging to correlate findings across platforms. The volume of spatial transcriptomics and proteomics data can be overwhelming, requiring advanced computational tools. Developing algorithms to reduce dimensionality while preserving essential biological information is crucial but complex. Managing these large-scale datasets demands efficient storage, processing, and analysis methods. Different spatial-omics techniques have varying levels of spatial resolution and sensitivity. While spatial transcriptomics offers high-resolution gene expression data, proteomics and metabolomics may have lower spatial resolution, making capturing fine-scale changes in molecular interactions difficult. Similarly, mechanical mapping, which assesses tissue stiffness, lacks the sensitivity for early-stage OA detection. Integrating molecular data with mechanical properties (e.g., cartilage stiffness) requires sophisticated multivariate analysis. OA exhibits patient-specific progression patterns, so developing machine learning models to correlate these data types and predict disease trajectories should be further developed. Existing methods often fall short in accounting for OA’s complex biological and mechanical features. Integrating these diverse data types requires collaboration across multiple disciplines, including biology, chemistry, physics, and computational science. This interdisciplinary effort is necessary for designing experiments, interpreting complex data, and developing tools that can handle multi-omics data effectively. In summary, the main challenges in combining spatial-omics data for OA research involve data integration, scaling, resolution, sensitivity, and variability. Overcoming these barriers requires advanced computational methods, interdisciplinary teamwork, and ongoing technological improvements.

Integrative Insights and Therapeutic Implications

Novel Technologies and Methodology in OA Spatiotemporal Changes

As the understanding of OA evolves, emerging tools and technologies significantly enhance the study of OA’s spatiotemperal dynamics. Among those, three crucial techniques are needed to understand OA progression: high-resolution imaging techniques, quantitative biomarker surveillance and artificial intelligence/machine learning.

High-resolution imaging techniques, featured by spatial transcriptomics and spatial proteomics, are complementary fields that study the spatial organisation of molecular components within biological systems [ 24 , 207 ]. Spatial transcriptomics maps gene expression by localising RNA molecules within tissue architecture, using methods like in situ hybridisation and spatial barcoding to uncover patterns of gene activity and tissue heterogeneity. Spatial proteomics focuses on protein localisation and interactions, employing advanced imaging and mass spectrometry techniques to investigate cellular functions and disease mechanisms. Together, these fields integrate molecular and spatial data, providing a comprehensive understanding of how the spatial organisation of transcripts and proteins influences biological processes and disease. With spatial proteomics and spatial transcriptomics won as Nature Method of the Year 2024 and 2020, respectively [ 24 , 207 ], these up-to-date techniques, alongside spatial metabolomics, provide a probability for researchers to understand OA progression. Technical advances have been advancing in two directions: to reach ultra-high-resolution to reach a subcellular spatial resolution and to correlate with multi-omics data in situ. Higher resolutions are mainly achieved by coupling fluorophore labelling using optical scanning microscopy or ToF-SIM, a mass spectrometry imaging-based system. Advancements in technology have opened new avenues for understanding the progression of OA by linking spatial proteomics with transcriptomics. In contrast to traditional clinical imaging techniques, newly developed multi-omics imaging platforms now offer comprehensive solutions for analysing spatial transcriptomics and spatial proteomics and famous instruments include CODEX (Akoya Bioscience), Stereo-seq (Beijing Genomics Institute), Visium (10× Genomics), GeoMx DSP (Nanostring), CosMx SMI (Nanostring), MERFISH (Vizgen), MultiOmyx (NeoGenomics), Discovery Ultra/Opal multiplex IHC (Roche/Akoya Bioscience), Hyperion (Fluidigm), and MIBIscope (IONpath) [ 207 ]. Other imaging techniques, like hyperspectral imaging techniques, elemental mapping, and function imaging, can also help reveal the spatiotemporal changes of OA [ 8 ]. To conclude, novel methodologies, models and instruments offer unparalleled insights into OA's spatiotemperal dynamics, driving precision diagnostics and innovative therapies for improved patient outcomes.

Although consecutive histopathological tracing of the human knee joint in vivo is still tricky, the dynamic tracing for OA progression in an animal model has shed light on OA progression. The practice hasn’t been utilised in OA changes yet, but the mouse embryo development as the spatiotemporal transcriptomic modification patterns have been verified using various techniques, including Slide-seq and the three-dimensional reconstruction and visualisation tool (sc3D) [ 208 ], Stereo-seq using DNA nanoball [ 209 ] and sequential fluorescence in situ hybridisation (seqFISH) [ 210 ].

Artificial Intelligence (AI) holds transformative potential in OA research, particularly in three key areas: integrating data across different scales, automating the annotation of genes or proteins, and uncovering risk factors or biomarkers through data-driven approaches. Machine learning (ML) algorithms can combine spatial multi-omics data with clinical and biochemical indicators, bridging the gap between laboratory research and clinical application. By leveraging ML in spatial proteomics, researchers can improve image-based analyses, enabling precise and efficient identification of proteomic signatures within joint tissues [ 211 ]. However, ML heavily relies on annotation libraries derived from in-silico experimental data, meaning its effectiveness depends on the availability of highly accurate, tissue-specific proteomic data. Unfortunately, current efforts to construct and standardise proteomics databases for cartilage remain inadequate and require further development. While these advancements in technology and methodology offer promising opportunities to enhance OA diagnosis and treatment, their successful implementation in clinical practice will depend on overcoming significant logistical and technical challenges.

Emerging Research Direction and Theoretical Frame for OA Aetiology

Regarding research models, four key aspects of OA progression are crucial: organelle-organelle, cell-cell, tissue-tissue communication, and cross-scale interactions [ 212 ]. Multi-scale models integrating data across molecular, cellular, tissue, and organ levels are essential for understanding OA progression. Despite significant advances, cross-scale investigations remain underdeveloped. Key areas for exploration include (1) correlating local histological changes with systemic biomarker shifts, (2) developing robust post-processing techniques for large-scale imaging and phenotype data, (3) linking spatial endotype changes in local tissues with clinical phenotypes and imaging, and (4) associating developmental gene polymorphisms with late-stage OA ( Fig. 3 ).

A critical challenge in OA research is translating histopathological knowledge into clinical practice. One promising pathway involves integrating big and multi-omics data, such as the UK Biobank. For instance, MILTON's machine learning tool predicts 3,213 diseases using biomarkers and longitudinal health records, outperforming polygenic risk scores and identifying novel gene-disease associations [ 213 ]. Similarly, "sc-linker" connects GWAS findings with specific cell types and processes by integrating scRNA-seq data, epigenomic SNP-to-gene maps, and GWAS summary statistics [ 214 ]. These advances may inspire future investigation of data-driven approaches to enhance genetic research and personalised OA care.

Another challenge in understanding OA's heterogeneity requires distinguishing phenotypes—observable traits from clinical assessments—and endotypes, which reflect underlying pathobiological mechanisms like inflammation or biomechanical stress [ 67 , 187 , 215 - 218 ]. This distinction highlights that different mechanistic pathways may result in similar clinical presentations but require tailored therapeutic strategies. Although separately investigated, the combination of spatiotemporal analysis in the OA field remains to be investigated, which paved the way for future personalised interventions [ 219 ].

Despite identifying novel cell subtypes through scRNA-seq and multi-omics imaging, their roles in OA progression remain unclear. For example, while infants can regenerate cartilage, this ability diminishes with growth. Whether changes or manipulation to specific cell subtypes contribute to loss of regeneration remains to be determined. OA's biomechanical nature also suggests that cell interactions in different environments may vary, presenting opportunities to identify vulnerable subtypes and develop targeted therapies. These unresolved questions highlight the complexity of modelling OA's molecular and cellular mechanisms. Advancing our understanding of these interactions will require collaborative efforts across disciplines to address the temporal and spatial dynamics of tissue degeneration.

Precision Medicine Targeting Spatiotemporal Risk Factors of OA

Emerging therapies in OA are increasingly informed by our evolving understanding of the disease's spatiotemporal dynamics. Recent advances in pharmacological treatments focus on targeting specific subtypes of OA based on spatiotemporal subtypes implicated in the progressive degeneration of joint tissues.

For example, metabolic regulative drugs have been proven to have life-changing benefits in OA [ 220 ]. Semaglutide is a Glucagon-like peptide-1 (GLP-1) agonist approved by the FDA for treating obesity. In the latest study involving a 68-week randomised trial, once-weekly injectable semaglutide significantly reduced body weight and knee OA pain compared to placebo in obese participants with moderate-to-severe pain. At the same time, adverse events were similar between groups, proving the efficacy of semaglutide on obese patients with knee OA [ 221 ]. NAD + is another hot topic in metabolic improvement. The Sirtuin (SIRT) family consists of nicotinamide adenine dinucleotide (NAD + )-dependent protein deacetylases that play an active role in regulating chondrocyte functions during the pathophysiology of OA. This family is crucial for preventing cartilage aging and degradation by maintaining extracellular matrix (ECM) homeostasis, regulating chondrocyte metabolism, and inhibiting chondrocyte apoptosis and autophagy through their deacetylation activity. [ 222 , 223 ]. Mouse models have demonstrated that inhibition of CD38, an enzyme of NAD + , may serve as a promising therapeutic strategy for OA treatment by reducing cartilage degradation, synovial inflammation, and associated pain following joint injury [ 224 ]. Other FDA-approved metabolic manipulation medicines and therapies, like statins [ 225 , 226 ], metformin [ 227 ], caloric restriction [ 228 ], minimally invasive arthroscopy and autologous micro-fragmented adipose tissue transplantation [ 229 ], can also be seen as future strategies.

Meanwhile, multiple studies have proved that anti-aging strategies benefit OA [ 230 , 231 ]. Key regulators in aging, including stem cell therapy [ 232 ], senescent cell elimination [ 233 ], autophagy enhancement [ 234 ], physical exercise [ 235 ], cellular reprogramming [ 236 ] and telomere reactivation [ 237 ], Parabiosis (Blood Exchange) [ 238 ] and other techniques have also raised researchers' interest and proven to be promising in OA. Despite advancements in therapy, the foundational approach of spatiotemporal phenotyping and clustering remains unexplored. Existing precision medicine strategies predominantly involve systemic drug treatments, often resulting in unnecessary side effects. Future research could illuminate methods for the targeted enrichment of medicinal compounds at specific sites, potentially halting or reversing the progression of OA.

Conclusion

In summary, incorporating spatiotemporal dynamics offers a holistic view for exploring the interactions between tissues in OA research. This interdisciplinary strategy could improve our knowledge of this suffering disease that impacts millions globally. A deeper understanding of the fundamental mechanisms underlying OA in a spatiotemporal manner provides essential evidence for developing effective DMOADs. This advancement can improve treatment options and, subsequently, the quality of life for individuals affected by this debilitating condition.

Acknowledgments

XM acknowledges the Hunan Provincial Science and Technology Department, China (2018JJ2579, 2015JC 3034). XF acknowledges the Science and Technology Innovation Program of Hunan Province (2024RC3053). IP acknowledges the NHMRC Investigator grant fellowship (APP1176298). AS acknowledges the CBT ECR/MCR scheme (324910-0028/07). XW acknowledges the National Natural Science Foundation of China (32300652).

Funding Statement

XM acknowledges the Hunan Provincial Science and Technology Department, China (2018JJ2579, 2015JC 3034). XF acknowledges the Science and Technology Innovation Program of Hunan Province (2024RC3053). IP acknowledges the NHMRC Investigator grant fellowship (APP1176298). AS acknowledges the CBT ECR/MCR scheme (324910-0028/07). XW acknowledges the National Natural Science Foundation of China (32300652).

Conflicts of interest

The authors declare no conflict of interest.

References

[1] Hattori T, Ohga S, Shimo K, Matsubara T (2024). Pathology of knee osteoarthritis pain: contribution of joint structural changes and pain sensitization to movement-evoked pain in knee osteoarthritis. Pain Rep, 9:e1124.

[2] Wang P, Tang Z, Lin X, Shao W, Wang B, Li Z, et al. (2024). Associations of osteoarthritis with risk of future fracture events: prospective study in UK biobank. Int [J] Surg.

[3] Wong AY, Samartzis D, Maher C (2023). The global burden of osteoarthritis: past and future perspectives. Lancet Rheumatol, 5:e496-e497.

[4] Scheuing WJ, Reginato AM, Deeb M, Acer Kasman S (2023). The burden of osteoarthritis: Is it a rising problem? Best Pract Res Clin Rheumatol, 37:101836.

[5] Pandey P, Singh R, Srivastava S, Mishra MK (2024). A Review on Osteoarthritis and its Eradication Approaches: An Update. Curr Drug Res Rev.

[6] Minnig MCC, Golightly YM, Nelson AE (2024). Epidemiology of osteoarthritis: literature update 2022-2023. Curr Opin Rheumatol, 36:108-112.

[7] Fan X, Wu X, Crawford R, Xiao Y, Prasadam I (2021). Macro, Micro, and Molecular. Changes of the Osteochondral Interface in Osteoarthritis Development. Front Cell Dev Biol, 9:659654.

[8] Fan X, Sun AR, Young RSE, Afara IO, Hamilton BR, Ong LJY, et al. (2024). Spatial analysis of the osteoarthritis microenvironment: techniques, insights, and applications. Bone Res, 12:7.

[9] Li M, Lan L, Luo J, Peng L, Li X, Zhou X (2021). Identifying the Phenotypic and Temporal Heterogeneity of Knee Osteoarthritis: Data From the Osteoarthritis Initiative. Front Public Health, 9:726140.

[10] Riad MH, Cohnstaedt LW, Scoglio CM (2021). Risk Assessment of Dengue Transmission in Bangladesh Using a Spatiotemporal Network Model and Climate Data. Am J Trop Med Hyg, 104:1444-1455.

[11] Dai Y, O'Brien TD, Pei G, Zhao Z, Jia P (2020). Characterization of genome-wide association study data reveals spatiotemporal heterogeneity of mental disorders. BMC Med Genomics, 13:192.

[12] Turhan G, Kucuk H, Isik EO (2022). Spatio-temporal convolution for classification of alzheimer disease and mild cognitive impairment. Comput Methods Programs Biomed, 221:106825.

[13] Romo-Bucheli D, Erfurth US, Bogunovic H (2020). End-to-End Deep Learning Model for Predicting Treatment Requirements in Neovascular AMD From Longitudinal Retinal OCT Imaging. IEEE J Biomed Health Inform, 24:3456-3465.

[14] Jimenez-Sanchez J, Martinez-Rubio A, Popov A, Perez-Beteta J, Azimzade Y, Molina-Garcia D, et al. (2021). A mesoscopic simulator to uncover heterogeneity and evolutionary dynamics in tumors. PLoS Comput Biol, 17:e1008266.

[15] Golightly YM, Renner JB, Helmick CG, Jordan JM, Nelson AE (2024). Looking back on 30+ years of the Johnston County Osteoarthritis Project while looking forward with the Johnston County Health Study: A narrative review. Osteoarthritis Cartilage, 32:430-438.

[16] Geusens PP, van den Bergh JP (2016). Osteoporosis and osteoarthritis: shared mechanisms and epidemiology. Curr Opin Rheumatol, 28:97-103.

[17] Hiranaka T, Furumatsu T, Yokoyama Y, Higashihara N, Tamura M, Kawada K, et al. (2024). Intercondylar notch width and osteophyte width impact meniscal healing and clinical outcomes following transtibial pullout repair of medial meniscus posterior root tears. Knee Surg Sports Traumatol Arthrosc, 32:116-123.

[18] Liu Y, Wu C, Chen C, Zhang L, Xing G, Wu K, et al. (2022). Impact of soft tissue around the knee on the efficacy of extracorporeal shockwave therapy in knee osteoarthritis. Medicine (Baltimore), 101:e32334.

[19] Bakinowska E, Kielbowski K, Pawlik A (2023). The Role of Extracellular Vesicles in the Pathogenesis and Treatment of Rheumatoid Arthritis and Osteoarthritis. Cells, 12:2716.

[20] Naselli F, Bellavia D, Costa V, De Luca A, Raimondi L, Giavaresi G, et al. (2023). Osteoarthritis in the Elderly Population: Preclinical Evidence of Nutrigenomic Activities of Flavonoids. Nutrients, 16:112.

[21] Wei G, Lu K, Umar M, Zhu Z, Lu WW, Speakman JR, et al. (2023). Risk of metabolic abnormalities in osteoarthritis: a new perspective to understand its pathological mechanisms. Bone Res, 11:63.

[22] Katz JN, Arant KR, Loeser RF (2021). Diagnosis and Treatment of Hip and Knee Osteoarthritis: A Review. JAMA, 325:568-578.

[23] Grol MW (2024). The evolving landscape of gene therapy strategies for the treatment of osteoarthritis. Osteoarthritis Cartilage, 32:372-384.

[24] Editorials NM (2021). Method of the Year 2020: spatially resolved transcriptomics. Nat Methods, 18:1.

[25] Zhuang X (2021). Spatially resolved single-cell genomics and transcriptomics by imaging. Nat Methods, 18:18-22.

[26] Rai MF, Collins KH, Lang A, Maerz T, Geurts J, Ruiz-Romero C, et al. (2024). Three decades of advancements in osteoarthritis research: insights from transcriptomic, proteomic, and metabolomic studies. Osteoarthritis Cartilage, 32:385-397.

[27] Wu H, Dixon EE, Xuanyuan Q, Guo J, Yoshimura Y, Debashish C, et al. (2024). High resolution spatial profiling of kidney injury and repair using RNA hybridization-based in situ sequencing. Nat Commun, 15:1396.

[28] Kumar R, Zemaitis KJ, Fulcher JM, Pasa-Tolic L (2024). Advances in mass spectrometry-enabled at single-cell resolution. Curr Opin Biotechnol, 87:103096.

[29] Magnusson K, Turkiewicz A, Rydén M, Englund M (2024). Genetic Influence on Osteoarthritis Versus Other Rheumatic Diseases. Arthritis Rheumatol, 76:206-215.

[30] Kiełbowski K, Herian M, Bakinowska E, Banach B, Sroczyński T, Pawlik A (2023). The Role of Genetics and Epigenetic Regulation in the Pathogenesis of Osteoarthritis. Int J Mol Sci, 24.

[31] Riyazi N, Meulenbelt I, Kroon HM, Ronday KH, Hellio le Graverand MP, Rosendaal FR, et al. (2005). Evidence for familial aggregation of hand, hip, and spine but not knee osteoarthritis in siblings with multiple joint involvement: the GARP study. Ann Rheum Dis, 64:438-443.

[32] Weldingh E, Johnsen MB, Hagen K, Østerås N, Slatkowsky-Christensen B, Nordsletten L, et al. (2019). The maternal and paternal effects on clinical and surgical definitions of osteoarthritis. Osteoarthritis and Cartilage, 27:S47.

[33] Collison J (2019). Mother-daughter inheritance patterns in OA. Nat Rev Rheumatol, 15:450.

[34] Magnusson K, Turkiewicz A, Englund M (2019). Nature vs nurture in knee osteoarthritis - the importance of age, sex and body mass index. Osteoarthritis Cartilage, 27:586-592.

[35] Wang C, Zhang R (2023). The effect of ITLN1, XCL2 and DOT1L variants on knee osteoarthritis risk in the Han population. Arch Orthop Trauma Surg, 143:4821-4831.

[36] Zhao LY, Song J, Liu Y, Song CX, Yi C (2020). Mapping the epigenetic modifications of DNA and RNA. Protein Cell, 11:792-808.

[37] Kiełbowski K, Herian M, Bakinowska E, Banach B, Sroczyński T, Pawlik A (2023). The Role of Genetics and Epigenetic Regulation in the Pathogenesis of Osteoarthritis. Int J Mol Sci, 24:11655.

[38] Chen B, Lu KH, Ni QB, Li QX, Gao H, Wang H, et al. (2019). Prenatal nicotine exposure increases osteoarthritis susceptibility in male elderly offspring rats via low-function programming of the TGFβ signaling pathway. Toxicol Lett, 314:18-26.

[39] Ni Q, Tan Y, Zhang X, Luo H, Deng Y, Magdalou J, et al. (2015). Prenatal ethanol exposure increases osteoarthritis susceptibility in female rat offspring by programming a low-functioning IGF-1 signaling pathway. Sci Rep, 5:14711.

[40] Li X, Zhu W, Cheng Y, Ren Z, Liu X, Yang H, et al. (2024). Intrauterine hyperglycemia induces SIRT3-mediated mitochondrial dysfunction: the fetal origin pathogenesis of precocious osteoarthritis. Osteoarthritis Cartilage, 32:950-962.

[41] Tan Y, Wu Y, Ni Q, Deng Y, Li J, Wang L, et al. (2016). Prenatal food restriction induces poor-quality articular cartilage in female rat offspring fed a post-weaning high-fat diet and its intra-uterine programming mechanisms. Br J Nutr, 116:1346-1355.

[42] Harasymowicz NS, Choi YR, Wu CL, Iannucci L, Tang R, Guilak F (2020). Intergenerational Transmission of Diet-Induced Obesity, Metabolic Imbalance, and Osteoarthritis in Mice. Arthritis Rheumatol, 72:632-644.

[43] Hussain SM, Ackerman IN, Wang Y, Zomer E, Cicuttini FM (2018). Could low birth weight and preterm birth be associated with significant burden of hip osteoarthritis? A systematic review. Arthritis Res Ther, 20:121.

[44] Vanden Berg-Foels WS, Todhunter RJ, Schwager SJ, Reeves AP (2006). Effect of early postnatal body weight on femoral head ossification onset and hip osteoarthritis in a canine model of developmental dysplasia of the hip. Pediatr Res, 60:549-554.

[45] Xin Z, Xu L, Sun L (2024). Assessing the causal relationship of birth weight and childhood obesity on osteoarthritis: a Mendelian randomization study. J Dev Orig Health Dis, 15:e12.

[46] Wills AK, Black S, Cooper R, Coppack RJ, Hardy R, Martin KR, et al. (2012). Life course body mass index and risk of knee osteoarthritis at the age of 53 years: evidence from the 1946 British birth cohort study. Ann Rheum Dis, 71:655-660.

[47] Khabiah MM, A MAH, Al Mudawi EA, Alkhalifah KM, Asiri NA, Alqahtani RF, et al. (2023). Prevalence of Breech Presentation and Other Gestational/Delivery Characteristics Among Patients Born With Developmental Dysplasia of the Hip. Cureus, 15:e42750.

[48] Day FR, Elks CE, Murray A, Ong KK, Perry JR (2015). Puberty timing associated with diabetes, cardiovascular disease and also diverse health outcomes in men and women: the UK Biobank study. Sci Rep, 5:11208.

[49] Nielsen RL, Monfeuga T, Kitchen RR, Egerod L, Leal LG, Schreyer ATH, et al. (2024). Data-driven identification of predictive risk biomarkers for subgroups of osteoarthritis using interpretable machine learning. Nat Commun, 15:2817.

[50] Chang AH, Almagor O, Lee JJ, Song J, Muhammad LN, Chmiel JS, et al. (2023). The Natural History of Knee Osteoarthritis Pain Experience and Risk Profiles. J Pain, 24:2175-2185.

[51] Kundu S, Ashinsky BG, Bouhrara M, Dam EB, Demehri S, Shifat ERM, et al. (2020). Enabling early detection of osteoarthritis from presymptomatic cartilage texture maps via transport-based learning. Proc Natl Acad Sci U S A, 117:24709-24719.

[52] Roemer FW, Zhang Y, Niu J, Lynch JA, Crema MD, Marra MD, et al. (2009). Tibiofemoral joint osteoarthritis: risk factors for MR-depicted fast cartilage loss over a 30-month period in the multicenter osteoarthritis study. Radiology, 252:772-780.

[53] Segal NA, Nevitt MC, Gross KD, Hietpas J, Glass NA, Lewis CE, et al. (2013). The Multicenter Osteoarthritis Study: opportunities for rehabilitation research. PM&R, 5:647-654.

[54] Guermazi A, Niu J, Hayashi D, Roemer FW, Englund M, Neogi T, et al. (2012). Prevalence of abnormalities in knees detected by MRI in adults without knee osteoarthritis: population based observational study (Framingham Osteoarthritis Study). Bmj, 345:e5339.

[55] Perry TA, Wang X, Gates L, Parsons CM, Sanchez-Santos MT, Garriga C, et al. (2020). Occupation and risk of knee osteoarthritis and knee replacement: A longitudinal, multiple-cohort study. Semin Arthritis Rheum, 50:1006-1014.

[56] Ikram MA, Brusselle GGO, Murad SD, van Duijn CM, Franco OH, Goedegebure A, et al. (2017). The Rotterdam Study: 2018 update on objectives, design and main results. Eur J Epidemiol, 32:807-850.

[57] Wesseling J, Boers M, Viergever MA, Hilberdink WK, Lafeber FP, Dekker J, et al. (2016). Cohort Profile: Cohort Hip and Cohort Knee (CHECK) study. Int J Epidemiol, 45:36-44.

[58] van Helvoort EM, van Spil WE, Jansen MP, Welsing PMJ, Kloppenburg M, Loef M, et al. (2020). Cohort profile: The Applied Public-Private Research enabling OsteoArthritis Clinical Headway (IMI-APPROACH) study: a 2-year, European, cohort study to describe, validate and predict phenotypes of osteoarthritis using clinical, imaging and biochemical markers. BMJ Open, 10:e035101.

[59] Nelson AE, Hu D, Arbeeva L, Alvarez C, Cleveland RJ, Schwartz TA, et al. (2021). The Prevalence of Knee Symptoms, Radiographic, and Symptomatic Osteoarthritis at Four Time Points: The Johnston County Osteoarthritis Project, 1999-2018. ACR Open Rheumatol, 3:558-565.

[60] Syddall HE, Simmonds SJ, Carter SA, Robinson SM, Dennison EM, Cooper C (2019). The Hertfordshire Cohort Study: an overview. F1000Res, 8:82.

[61] Syddall HE, Aihie Sayer A, Dennison EM, Martin HJ, Barker DJ, Cooper C (2005). Cohort profile: the Hertfordshire cohort study. Int J Epidemiol, 34:1234-1242.

[62] Ding C, Parameswaran V, Cicuttini F, Burgess J, Zhai G, Quinn S, et al. (2008). Association between leptin, body composition, sex and knee cartilage morphology in older adults: the Tasmanian older adult cohort (TASOAC) study. Ann Rheum Dis, 67:1256-1261.

[63] de Mutsert R, den Heijer M, Rabelink TJ, Smit JW, Romijn JA, Jukema JW, et al. (2013). The Netherlands Epidemiology of Obesity (NEO) study: study design and data collection. Eur J Epidemiol, 28:513-523.

[64] Kraus VB, Sun S, Reed A, Soderblom EJ, Moseley MA, Zhou K, et al. (2024). An osteoarthritis pathophysiological continuum revealed by molecular biomarkers. Sci Adv, 10:eadj6814.

[65] Palazzo C, Nguyen C, Lefevre-Colau MM, Rannou F, Poiraudeau S (2016). Risk factors and burden of osteoarthritis. Ann Phys Rehabil Med, 59:134-138.

[66] Sedaghati-Khayat B, Boer CG, Runhaar J, Bierma-Zeinstra SMA, Broer L, Ikram MA, et al. (2022). Risk Assessment for Hip and Knee Osteoarthritis Using Polygenic Risk Scores. Arthritis Rheumatol, 74:1488-1496.

[67] Angelini F, Widera P, Mobasheri A, Blair J, Struglics A, Uebelhoer M, et al. (2022). Osteoarthritis endotype discovery via clustering of biochemical marker data. Ann Rheum Dis, 81:666-675.

[68] Roemer FW, Jansen M, Marijnissen ACA, Guermazi A, Heiss R, Maschek S, et al. (2022). Structural tissue damage and 24-month progression of semi-quantitative MRI biomarkers of knee osteoarthritis in the IMI-APPROACH cohort. BMC Musculoskelet Disord, 23:988.

[69] Callahan LF, Cleveland RJ, Allen KD, Golightly Y (2021). Racial/Ethnic, Socioeconomic, and Geographic Disparities in the Epidemiology of Knee and Hip Osteoarthritis. Rheum Dis Clin North Am, 47:1-20.

[70] Nelson A (2023). Epidemiology of Hip Osteoarthritis: The Johnston County Osteoarthritis Project. Hss j, 19:413-417.

[71] Clynes MA, Parsons C, Edwards MH, Jameson KA, Harvey NC, Sayer AA, et al. (2014). Further evidence of the developmental origins of osteoarthritis: results from the Hertfordshire Cohort Study. J Dev Orig Health Dis, 5:453-458.

[72] McMillan LB, Aitken D, Ebeling P, Jones G, Scott D (2018). The relationship between objectively assessed physical activity and bone health in older adults differs by sex and is mediated by lean mass. Osteoporos Int, 29:1379-1388.

[73] Shen Z, Wang Y, Xing X, Jones G, Cai G (2023). Association of smoking with cartilage loss of knee osteoarthritis: data from two longitudinal cohorts. BMC Musculoskelet Disord, 24:812.

[74] Visser AW, Ioan-Facsinay A, de Mutsert R, Widya RL, Loef M, de Roos A, et al. (2014). Adiposity and hand osteoarthritis: the Netherlands Epidemiology of Obesity study. Arthritis Res Ther, 16:R19.

[75] Newman S, Ahmed H, Rehmatullah N (2022). Radiographic vs. MRI vs. arthroscopic assessment and grading of knee osteoarthritis - are we using appropriate imaging? J Exp Orthop, 9:2.

[76] Hunter DJ, Bierma-Zeinstra S (2019). Osteoarthritis. Lancet, 393:1745-1759.

[77] Braun HJ, Gold GE (2012). Diagnosis of osteoarthritis: imaging. Bone, 51:278-288.

[78] Sukerkar PA, Doyle Z (2022). Imaging of Osteoarthritis of the Knee. Radiol Clin North Am, 60:605-616.

[79] Wang X, Oo WM, Linklater JM (2018). What is the role of imaging in the clinical diagnosis of osteoarthritis and disease management? Rheumatology (Oxford), 57:iv51-iv60.

[80] Guermazi A, Hayashi D, Roemer F, Felson DT, Wang K, Lynch J, et al. (2015). Severe radiographic knee osteoarthritis--does Kellgren and Lawrence grade 4 represent end stage disease?--the MOST study. Osteoarthritis Cartilage, 23:1499-1505.

[81] Pan Y, Lei X, Zhang Y (2022). Association predictions of genomics, proteinomics, transcriptomics, microbiome, metabolomics, pathomics, radiomics, drug, symptoms, environment factor, and disease networks: A comprehensive approach. Med Res Rev, 42:441-461.

[82] Zhang X, Zhang Y, Zhang G, Qiu X, Tan W, Yin X, et al. (2022). Deep Learning With Radiomics for Disease Diagnosis and Treatment: Challenges and Potential. Front Oncol, 12:773840.

[83] Li S, Cao P, Li J, Chen T, Luo P, Ruan G, et al. (2024). Integrating Radiomics and Neural Networks for Knee Osteoarthritis Incidence Prediction. Arthritis Rheumatol, 76:1377-1386.

[84] Minciullo L, Bromiley PA, Felson DT, Cootes TF.2017. Indecisive trees for classification and prediction of knee osteoarthritis. In Machine Learning in Medical Imaging: 8th International Workshop, MLMI 2017, Held in Conjunction with MICCAI 2017, Quebec City, QC, Canada, September 10, 2017, Proceedings 8: Springer. 283-290.

[85] Bayramoglu N, Nieminen MT, Saarakkala S (2022). Machine learning based texture analysis of patella from X-rays for detecting patellofemoral osteoarthritis. Int J Med Inform, 157:104627.

[86] Navale DI, Hegadi RS, Mendgudli N.2015. Block based texture analysis approach for knee osteoarthritis identification using SVM. In 2015 IEEE International WIE Conference on Electrical and Computer Engineering (WIECON-ECE): IEEE. 338-341.

[87] Thomson J, O’Neill T, Felson D, Cootes T.2015. Automated shape and texture analysis for detection of osteoarthritis from radiographs of the knee. In Medical Image Computing and Computer-Assisted Intervention--MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part II 18: Springer. 127-134.

[88] Halilaj E, Moore DC, Laidlaw DH, Got CJ, Weiss AP, Ladd AL, et al. (2014). The morphology of the thumb carpometacarpal joint does not differ between men and women, but changes with aging and early osteoarthritis. J Biomech, 47:2709-2714.

[89] Bianchi J, Gonçalves JR, de Oliveira Ruellas AC, Ashman LM, Vimort JB, Yatabe M, et al. (2021). Quantitative bone imaging biomarkers to diagnose temporomandibular joint osteoarthritis. Int J Oral Maxillofac Surg, 50:227-235.

[90] Xie Y, Dan Y, Tao H, Wang C, Zhang C, Wang Y, et al. (2021). Radiomics Feature Analysis of Cartilage and Subchondral Bone in Differentiating Knees Predisposed to Posttraumatic Osteoarthritis after Anterior Cruciate Ligament Reconstruction from Healthy Knees. Biomed Res Int, 2021:4351499.

[91] Marques J, Genant HK, Lillholm M, Dam EB (2013). Diagnosis of osteoarthritis and prognosis of tibial cartilage loss by quantification of tibia trabecular bone from MRI. Magn Reson Med, 70:568-575.

[92] Lin T, Peng S, Lu S, Fu S, Zeng D, Li J, et al. (2023). Prediction of knee pain improvement over two years for knee osteoarthritis using a dynamic nomogram based on MRI-derived radiomics: a proof-of-concept study. Osteoarthritis Cartilage, 31:267-278.

[93] Gwilym SE, Keltner JR, Warnaby CE, Carr AJ, Chizh B, Chessell I, et al. (2009). Psychophysical and functional imaging evidence supporting the presence of central sensitization in a cohort of osteoarthritis patients. Arthritis Rheum, 61:1226-1234.

[94] Chu CR, Lin D, Geisler JL, Chu CT, Fu FH, Pan Y (2004). Arthroscopic microscopy of articular cartilage using optical coherence tomography. Am J Sports Med, 32:699-709.

[95] Chu CR, Izzo NJ, Irrgang JJ, Ferretti M, Studer RK (2007). Clinical diagnosis of potentially treatable early articular cartilage degeneration using optical coherence tomography. J Biomed Opt, 12:051703.

[96] Chu CR, Williams A, Tolliver D, Kwoh CK, Bruno S 3rd, Irrgang JJ (2010). Clinical optical coherence tomography of early articular cartilage degeneration in patients with degenerative meniscal tears. Arthritis Rheum, 62:1412-1420.

[97] Kogan F, Watkins LE, Goyal A (2025). PET-MRI: The promise of multi-tissue imaging of early disease mechanisms in osteoarthritis. Osteoarthritis Cartilage, 33:5-8.

[98] Zarringam D, Saris DBF, Bekkers JEJ (2021). The Value of SPECT/CT for Knee Osteoarthritis: A Systematic Review. Cartilage, 12:431-437.

[99] Jiang T, Lau SH, Zhang J, Chan LC, Wang W, Chan PK, et al. (2024). Radiomics signature of osteoarthritis: Current status and perspective. J Orthop Translat, 45:100-106.

[100] Bastick AN, Runhaar J, Belo JN, Bierma-Zeinstra SM (2015). Prognostic factors for progression of clinical osteoarthritis of the knee: a systematic review of observational studies. Arthritis Res Ther, 17:152.

[101] Herrero-Beaumont G, Castro-Dominguez F, Migliore A, Naredo E, Largo R, Reginster JY (2024). Systemic osteoarthritis: the difficulty of categorically naming a continuous condition. Aging Clin Exp Res, 36:45.

[102] Bierma-Zeinstra SM, van Middelkoop M (2017). Osteoarthritis: In search of phenotypes. Nat Rev Rheumatol, 13:705-706.

[103] Deveza LA, Melo L, Yamato TP, Mills K, Ravi V, Hunter DJ (2017). Knee osteoarthritis phenotypes and their relevance for outcomes: a systematic review. Osteoarthritis Cartilage, 25:1926-1941.

[104] Petersen ET, Rytter S, Koppens D, Dalsgaard J, Hansen TB, Larsen NE, et al. (2022). Patients with knee osteoarthritis can be divided into subgroups based on tibiofemoral joint kinematics of gait - an exploratory and dynamic radiostereometric study. Osteoarthritis Cartilage, 30:249-259.

[105] Knoop J, de Joode JW, Brandt H, Dekker J, Ostelo R (2022). Patients' and clinicians' experiences with stratified exercise therapy in knee osteoarthritis: a qualitative study. BMC Musculoskelet Disord, 23:559.

[106] Vongsirinavarat M, Nilmart P, Somprasong S, Apinonkul B (2020). Identification of knee osteoarthritis disability phenotypes regarding activity limitation: a cluster analysis. BMC Musculoskelet Disord, 21:237.

[107] Radojčić MR, Arden NK, Yang X, Strauss VY, Birrell F, Cooper C, et al. (2020). Pain trajectory defines knee osteoarthritis subgroups: a prospective observational study. Pain, 161:2841-2851.

[108] Wyatt LA, Moreton BJ, Mapp PI, Wilson D, Hill R, Ferguson E, et al. (2017). Histopathological subgroups in knee osteoarthritis. Osteoarthritis Cartilage, 25:14-22.

[109] Calvet J, Berenguer-Llergo A, Orellana C, García-Manrique M, Rusiñol M, Garcia-Cirera S, et al. (2024). Specific-cytokine associations with outcomes in knee osteoarthritis subgroups: breaking down disease heterogeneity with phenotyping. Arthritis Res Ther, 26:19.

[110] Xue Y, Zhou L, Wang J (2023). Classification of distinct osteoarthritis subtypes with different knee joint tissues by gene expression profiles. Bone Joint Res, 12:702-711.

[111] Cao X, Cui Z, Ding Z, Chen Y, Wu S, Wang X, et al. (2022). An osteoarthritis subtype characterized by synovial lipid metabolism disorder and fibroblast-like synoviocyte dysfunction. J Orthop Translat, 33:142-152.

[112] Werdyani S, Liu M, Zhang H, Sun G, Furey A, Randell EW, et al. (2021). Endotypes of primary osteoarthritis identified by plasma metabolomics analysis. Rheumatology (Oxford), 60:2735-2744.

[113] Yuan C, Pan Z, Zhao K, Li J, Sheng Z, Yao X, et al. (2020). Classification of four distinct osteoarthritis subtypes with a knee joint tissue transcriptome atlas. Bone Res, 8:38.

[114] Soul J, Dunn SL, Anand S, Serracino-Inglott F, Schwartz JM, Boot-Handford RP, et al. (2018). Stratification of knee osteoarthritis: two major patient subgroups identified by genome-wide expression analysis of articular cartilage. Ann Rheum Dis, 77:423.

[115] Hannani MT, Thudium CS, Gellhorn AC, Larkin J, Karsdal MA, Lisowska-Petersen Z, et al. (2025). Longitudinal stability of molecular endotypes of knee osteoarthritis patients. Osteoarthritis Cartilage, 33:166-175.

[116] Dell'Isola A, Allan R, Smith SL, Marreiros SS, Steultjens M (2016). Identification of clinical phenotypes in knee osteoarthritis: a systematic review of the literature. BMC Musculoskelet Disord, 17:425.

[117] Mobasheri A, Rayman MP, Gualillo O, Sellam J, van der Kraan P, Fearon U (2017). The role of metabolism in the pathogenesis of osteoarthritis. Nat Rev Rheumatol, 13:302-311.

[118] Mobasheri A, van Spil WE, Budd E, Uzieliene I, Bernotiene E, Bay-Jensen AC, et al. (2019). Molecular taxonomy of osteoarthritis for patient stratification, disease management and drug development: biochemical markers associated with emerging clinical phenotypes and molecular endotypes. Curr Opin Rheumatol, 31:80-89.

[119] Driban JB, Harkey MS, Barbe MF, Ward RJ, MacKay JW, Davis JE, et al. (2020). Risk factors and the natural history of accelerated knee osteoarthritis: a narrative review. BMC Musculoskelet Disord, 21:332.

[120] Collins DH, Mc ET (1960). Sulphate (35SO4) uptake by chondrocytes in relation to histological changes in osteoarthritic human articular cartilage. Ann Rheum Dis, 19:318-330.

[121] Pritzker KP, Gay S, Jimenez SA, Ostergaard K, Pelletier JP, Revell PA, et al. (2006). Osteoarthritis cartilage histopathology: grading and staging. Osteoarthritis Cartilage, 14:13-29.

[122] Yoshimi T, Kikuchi T, Obara T, Yamaguchi T, Sakakibara Y, Itoh H, et al. (1994). Effects of high-molecular-weight sodium hyaluronate on experimental osteoarthrosis induced by the resection of rabbit anterior cruciate ligament. Clin Orthop Relat Res: 296-304.

[123] Leroux MA, Cheung HS, Bau JL, Wang JY, Howell DS, Setton LA (2001). Altered mechanics and histomorphometry of canine tibial cartilage following joint immobilization. Osteoarthritis Cartilage, 9:633-640.

[124] McIlwraith CW, Frisbie DD, Kawcak CE, Fuller CJ, Hurtig M, Cruz A (2010). The OARSI histopathology initiative - recommendations for histological assessments of osteoarthritis in the horse. Osteoarthritis Cartilage, 18 Suppl 3:S93-105.

[125] Cook JL, Kuroki K, Visco D, Pelletier JP, Schulz L, Lafeber FP (2010). The OARSI histopathology initiative - recommendations for histological assessments of osteoarthritis in the dog. Osteoarthritis Cartilage, 18 Suppl 3:S66-79.

[126] Laverty S, Girard CA, Williams JM, Hunziker EB, Pritzker KP (2010). The OARSI histopathology initiative - recommendations for histological assessments of osteoarthritis in the rabbit. Osteoarthritis Cartilage, 18 Suppl 3:S53-65.

[127] Gerwin N, Bendele AM, Glasson S, Carlson CS (2010). The OARSI histopathology initiative - recommendations for histological assessments of osteoarthritis in the rat. Osteoarthritis Cartilage, 18 Suppl 3:S24-34.

[128] Roseti L, Desando G, Cavallo C, Petretta M, Grigolo B (2019). Articular Cartilage Regeneration in Osteoarthritis. Cells, 8:1305.

[129] Fujii Y, Liu L, Yagasaki L, Inotsume M, Chiba T, Asahara H (2022). Cartilage Homeostasis and Osteoarthritis. Int J Mol Sci, 23:6316.

[130] Tong L, Yu H, Huang X, Shen J, Xiao G, Chen L, et al. (2022). Current understanding of osteoarthritis pathogenesis and relevant new approaches. Bone Res, 10:60.

[131] Zhao D, Zeng LF, Liang GH, Luo MH, Pan JK, Dou YX, et al. (2024). Transcriptomic analyses and machine-learning methods reveal dysregulated key genes and potential pathogenesis in human osteoarthritic cartilage. Bone Joint Res, 13:66-82.

[132] Li C, Wei P, Wang L, Wang Q, Wang H, Zhang Y (2023). Integrated Analysis of Transcriptome Changes in Osteoarthritis: Gene Expression, Pathways and Alternative Splicing. Cartilage, 14:235-246.

[133] van der Kraan PM (2013). Understanding developmental mechanisms in the context of osteoarthritis. Curr Rheumatol Rep, 15:333.

[134] Song H, Park KH (2020). Regulation and function of SOX9 during cartilage development and regeneration. Semin Cancer Biol, 67:12-23.

[135] Lefebvre V, Angelozzi M, Haseeb A (2019). SOX9 in cartilage development and disease. Curr Opin Cell Biol, 61:39-47.

[136] Zuscik MJ, Hilton MJ, Zhang X, Chen D, O'Keefe RJ (2008). Regulation of chondrogenesis and chondrocyte differentiation by stress. J Clin Invest, 118:429-438.

[137] Li G, Fang Y, Xu N, Ding Y, Liu D (2024). Fibroblast-like synoviocytes-derived exosomal circFTO deteriorates rheumatoid arthritis by enhancing N6-methyladenosine modification of SOX9 in chondrocytes. Arthritis Res Ther, 26:56.

[138] Liao J, Hu N, Zhou N, Lin L, Zhao C, Yi S, et al. (2014). Sox9 potentiates BMP2-induced chondrogenic differentiation and inhibits BMP2-induced osteogenic differentiation. PLoS One, 9:e89025.

[139] Leung VY, Gao B, Leung KK, Melhado IG, Wynn SL, Au TY, et al. (2011). SOX9 governs differentiation stage-specific gene expression in growth plate chondrocytes via direct concomitant transactivation and repression. PLoS Genet, 7:e1002356.

[140] Molin AN, Contentin R, Angelozzi M, Karvande A, Kc R, Haseeb A, et al. (2024). Skeletal growth is enhanced by a shared role for SOX8 and SOX9 in promoting reserve chondrocyte commitment to columnar proliferation. Proc Natl Acad Sci U S A, 121:e2316969121.

[141] Ji Q, Zheng Y, Zhang G, Hu Y, Fan X, Hou Y, et al. (2019). Single-cell RNA-seq analysis reveals the progression of human osteoarthritis. Ann Rheum Dis, 78:100-110.

[142] Park J, Kim J, Lewy T, Rice CM, Elemento O, Rendeiro AF, et al. (2022). Spatial omics technologies at multimodal and single cell/subcellular level. Genome Biol, 23:256.

[143] Zhang L, Chen D, Song D, Liu X, Zhang Y, Xu X, et al. (2022). Clinical and translational values of spatial transcriptomics. Signal Transduct Target Ther, 7:111.

[144] Fan Y, Bian X, Meng X, Li L, Fu L, Zhang Y, et al. (2024). Unveiling inflammatory and prehypertrophic cell populations as key contributors to knee cartilage degeneration in osteoarthritis using multi-omics data integration. Ann Rheum Dis, 83:926-944.

[145] Sandell LJ, Aigner T (2001). Articular cartilage and changes in arthritis. An introduction: cell biology of osteoarthritis. Arthritis Res, 3:107-113.

[146] Hu W, Chen Y, Dou C, Dong S (2021). Microenvironment in subchondral bone: predominant regulator for the treatment of osteoarthritis. Ann Rheum Dis, 80:413-422.

[147] Kurz B, Lange T, Voelker M, Hart ML, Rolauffs B (2023). Articular Cartilage-From Basic Science Structural Imaging to Non-Invasive Clinical Quantitative Molecular Functional Information for AI Classification and Prediction. Int J Mol Sci, 24:14974.

[148] Maldonado M, Nam J (2013). The role of changes in extracellular matrix of cartilage in the presence of inflammation on the pathology of osteoarthritis. Biomed Res Int, 2013:284873.

[149] Sanchez-Lopez E, Coras R, Torres A, Lane NE, Guma M (2022). Synovial inflammation in osteoarthritis progression. Nat Rev Rheumatol, 18:258-275.

[150] Xu H, Wang W, Liu X, Huang W, Zhu C, Xu Y, et al. (2023). Targeting strategies for bone diseases: signaling pathways and clinical studies. Signal Transduct Target Ther, 8:202.

[151] Hu Y, Cui J, Liu H, Wang S, Zhou Q, Zhang H, et al. (2022). Single-cell RNA-sequencing analysis reveals the molecular mechanism of subchondral bone cell heterogeneity in the development of osteoarthritis. RMD Open, 8:e002314.

[152] Maruotti N, Corrado A, Cantatore FP (2017). Osteoblast role in osteoarthritis pathogenesis. J Cell Physiol, 232:2957-2963.

[153] Wang H, Yuan T, Wang Y, Liu C, Li D, Li Z, et al. (2024). Osteoclasts and osteoarthritis: Novel intervention targets and therapeutic potentials during aging. Aging Cell, 23:e14092.

[154] Wangler S, Menzel U, Li Z, Ma J, Hoppe S, Benneker LM, et al. (2019). CD146/MCAM distinguishes stem cell subpopulations with distinct migration and regenerative potential in degenerative intervertebral discs. Osteoarthritis Cartilage, 27:1094-1105.

[155] Bonnet CS, Walsh DA (2004). Osteoarthritis, angiogenesis and inflammation. Rheumatology, 44:7-16.

[156] Wang MG, Seale P, Furman D (2024). The infrapatellar fat pad in inflammaging, knee joint health, and osteoarthritis. NPJ Aging, 10:34.

[157] Peters H, Potla P, Rockel JS, Tockovska T, Pastrello C, Jurisica I, et al. (2024). Cell and transcriptomic diversity of infrapatellar fat pad during knee osteoarthritis. Ann Rheum Dis:ard--2024-225928.

[158] Tang S, Yao L, Ruan J, Kang J, Cao Y, Nie X, et al. (2024). Single-cell atlas of human infrapatellar fat pad and synovium implicates APOE signaling in osteoarthritis pathology. Sci Transl Med, 16:eadf4590.

[159] Li P, Wang A, Li J, Li X, Sun W, Liu Q (2021). COL2A1 Mutation (c.611G>C) Leads to Early-Onset Osteoarthritis in a Chinese Family. Int J Gen Med, 14:2569-2574.

[160] Hämäläinen S, Solovieva S, Hirvonen A, Vehmas T, Takala EP, Riihimäki H, et al. (2009). COL2A1 gene polymorphisms and susceptibility to osteoarthritis of the hand in Finnish women. Ann Rheum Dis, 68:1633-1637.

[161] Kania K, Colella F, Riemen AHK, Wang H, Howard KA, Aigner T, et al. (2020). Regulation of Gdf5 expression in joint remodelling, repair and osteoarthritis. Sci Rep, 10:157.

[162] Lettre G (2017). The osteoarthritis and height GDF5 locus yields its secrets. Nat Genet, 49:1165-1166.

[163] Kania K, Colella F, Riemen AH, Wang H, Howard KA, Aigner T, et al. (2020). Expression of growth and differentiation factor 5 during joint repair and osteoarthritis. Osteoarthritis and Cartilage, 28:S34.

[164] Balbaloglu O, Sabah Ozcan S, Korkmaz M, Yılmaz N (2017). Promoter polymorphism (T-1486C) of TLR-9 gene is associated with knee osteoarthritis in a Turkish population. J Orthop Res, 35:2484-2489.

[165] Zheng M, Shi S, Zheng Q, Wang Y, Ying X, Jin Y (2017). Association between TLR-9 gene rs187084 polymorphism and knee osteoarthritis in a Chinese population. Biosci Rep, 37:BSR20170844.

[166] Vrgoc G, Vrbanec J, Eftedal RK, Dembic PL, Balen S, Dembic Z, et al. (2018). Interleukin-17 and Toll-like Receptor 10 genetic polymorphisms and susceptibility to large joint osteoarthritis. J Orthop Res, 36:1684-1693.

[167] Han Z, Liu Q, Sun C, Li Y (2012). The interaction between obesity and RAGE polymorphisms on the risk of knee osteoarthritis in Chinese population. Cell Physiol Biochem, 30:898-904.

[168] Alimoradi N, Tahami M, Firouzabadi N, Haem E, Ramezani A (2023). Metformin attenuates symptoms of osteoarthritis: role of genetic diversity of Bcl2 and CXCL16 in OA. Arthritis Res Ther, 25:35.

[169] Hui W, Cao Z, Wang X, Zhu J (2019). Association of matrix Gla protein polymorphism and knee osteoarthritis in a chinese population. Biosci Rep, 39:BSR20182228.

[170] Wang J, Wang X, Ding X, Huang T, Song D, Tao H (2021). EZH2 is associated with cartilage degeneration in osteoarthritis by promoting SDC1 expression via histone methylation of the microRNA-138 promoter. Lab Invest, 101:600-611.

[171] Liu W, Jiang T, Zheng W, Zhang J, Li A, Lu C, et al. (2023). FTO-mediated m6A demethylation of pri-miR-3591 alleviates osteoarthritis progression. Arthritis Res Ther, 25:53.

[172] Ye J, Deng R, Wang X, Song S, Xu X, Zhang JY, et al. (2023). Intra-articular Histone Deacetylase Inhibitor Microcarrier Delivery to Reduce Osteoarthritis. Nano Lett, 23:10832-10840.

[173] Zheng Y, Chen Y, Lu X, Weng Q, Dai G, Yu Y, et al. (2020). Inhibition of Histone Deacetylase 6 by Tubastatin A Attenuates the Progress of Osteoarthritis via Improving Mitochondrial Function. Am J Pathol, 190:2376-2386.

[174] Orouji E, Utikal J (2018). Tackling malignant melanoma epigenetically: histone lysine methylation. Clin Epigenetics, 10:145.

[175] Viré E, Brenner C, Deplus R, Blanchon L, Fraga M, Didelot C, et al. (2006). The Polycomb group protein EZH2 directly controls DNA methylation. Nature, 439:871-874.

[176] Assi R, Cherifi C, Cornelis FMF, Zhou Q, Storms L, Pazmino S, et al. (2023). Inhibition of KDM7A/B histone demethylases restores H3K79 methylation and protects against osteoarthritis. Ann Rheum Dis, 82:963-973.

[177] Zhang P, Wu W, Chen Q, Chen M (2019). Non-Coding RNAs and their Integrated Networks. J Integr Bioinform, 16:20190027.

[178] Lu TX, Rothenberg ME (2018). MicroRNA. J Allergy Clin Immunol, 141:1202-1207.

[179] Zhang Y, Li S, Jin P, Shang T, Sun R, Lu L, et al. (2022). Dual functions of microRNA-17 in maintaining cartilage homeostasis and protection against osteoarthritis. Nat Commun, 13:2447.

[180] Jiang Y, Zhang L, Tian H (2023). MicroRNA-149 improves osteoarthritis via repression of VCAM-1 and inactivation of PI3K/AKT pathway. Exp Gerontol, 174:112103.

[181] Endisha H, Datta P, Sharma A, Nakamura S, Rossomacha E, Younan C, et al. (2021). MicroRNA-34a-5p Promotes Joint Destruction During Osteoarthritis. Arthritis Rheumatol, 73:426-439.

[182] Song J, Jin EH, Kim D, Kim KY, Chun CH, Jin EJ (2015). MicroRNA-222 regulates MMP-13 via targeting HDAC-4 during osteoarthritis pathogenesis. BBA Clin, 3:79-89.

[183] Baloun J, Pekáčová A, Švec X, Kropáčková T, Horvathová V, Hulejová H, et al. (2023). Circulating miRNAs in hand osteoarthritis. Osteoarthritis Cartilage, 31:228-237.

[184] Jenei-Lanzl Z, Meurer A, Zaucke F (2019). Interleukin-1β signaling in osteoarthritis - chondrocytes in focus. Cell Signal, 53:212-223.

[185] Chisari E, Yaghmour KM, Khan WS (2020). The effects of TNF-alpha inhibition on cartilage: a systematic review of preclinical studies. Osteoarthritis Cartilage, 28:708-718.

[186] Mehana EE, Khafaga AF, El-Blehi SS (2019). The role of matrix metalloproteinases in osteoarthritis pathogenesis: An updated review. Life Sci, 234:116786.

[187] Hannani MT, Thudium CS, Karsdal MA, Ladel C, Mobasheri A, Uebelhoer M, et al. (2024). From biochemical markers to molecular endotypes of osteoarthritis: a review on validated biomarkers. Expert Rev Mol Diagn, 24:23-38.

[188] Styrkarsdottir U, Lund SH, Saevarsdottir S, Magnusson MI, Gunnarsdottir K, Norddahl GL, et al. (2021). The CRTAC1 Protein in Plasma Is Associated With Osteoarthritis and Predicts Progression to Joint Replacement: A Large-Scale Proteomics Scan in Iceland. Arthritis Rheumatol, 73:2025-2034.

[189] Styrkarsdottir U, Lund SH, Thorleifsson G, Saevarsdottir S, Gudbjartsson DF, Thorsteinsdottir U, et al. (2023). Cartilage Acidic Protein 1 in Plasma Associates With Prevalent Osteoarthritis and Predicts Future Risk as Well as Progression to Joint Replacements: Results From the UK Biobank Resource. Arthritis Rheumatol, 75:544-552.

[190] Szilagyi IA, Vallerga CL, Boer CG, Schiphof D, Ikram MA, Bierma-Zeinstra SMA, et al. (2023). Plasma proteomics identifies CRTAC1 as a biomarker for osteoarthritis severity and progression. Rheumatology (Oxford), 62:1286-1295.

[191] Li Y, Xu L, Olsen BR (2007). Lessons from genetic forms of osteoarthritis for the pathogenesis of the disease. Osteoarthritis Cartilage, 15:1101-1105.

[192] Imagawa K, de Andrés MC, Hashimoto K, Itoi E, Otero M, Roach HI, et al. (2014). Association of reduced type IX collagen gene expression in human osteoarthritic chondrocytes with epigenetic silencing by DNA hypermethylation. Arthritis Rheumatol, 66:3040-3051.

[193] Mustafa Z, Chapman K, Irven C, Carr AJ, Clipsham K, Chitnavis J, et al. (2000). Linkage analysis of candidate genes as susceptibility loci for osteoarthritis-suggestive linkage of COL9A1 to female hip osteoarthritis. Rheumatology (Oxford), 39:299-306.

[194] Hu K, Xu L, Cao L, Flahiff CM, Brussiau J, Ho K, et al. (2006). Pathogenesis of osteoarthritis-like changes in the joints of mice deficient in type IX collagen. Arthritis Rheum, 54:2891-2900.

[195] Alizadeh BZ, Njajou OT, Bijkerk C, Meulenbelt I, Wildt SC, Hofman A, et al. (2005). Evidence for a role of the genomic region of the gene encoding for the alpha1 chain of type IX collagen (COL9A1) in hip osteoarthritis: A population-based study. Arthritis Rheum, 52:1437-1442.

[196] Zhang W, McWilliams DF, Ingham SL, Doherty SA, Muthuri S, Muir KR, et al. (2011). Nottingham knee osteoarthritis risk prediction models. Ann Rheum Dis, 70:1599-1604.

[197] Chou CH, Jain V, Gibson J, Attarian DE, Haraden CA, Yohn CB, et al. (2020). Synovial cell cross-talk with cartilage plays a major role in the pathogenesis of osteoarthritis. Sci Rep, 10:10868.

[198] Tan C, Li L, Han J, Xu K, Liu X (2022). A new strategy for osteoarthritis therapy: Inhibition of glycolysis. Front Pharmacol, 13:1057229.

[199] Zheng L, Zhang Z, Sheng P, Mobasheri A (2021). The role of metabolism in chondrocyte dysfunction and the progression of osteoarthritis. Ageing Res Rev, 66:101249.

[200] Zeng CY, Wang XF, Hua FZ (2022). HIF-1α in Osteoarthritis: From Pathogenesis to Therapeutic Implications. Front Pharmacol, 13:927126.

[201] Zahan OM, Serban O, Gherman C, Fodor D (2020). The evaluation of oxidative stress in osteoarthritis. Med Pharm Rep, 93:12-22.

[202] Molnar V, Matišić V, Kodvanj I, Bjelica R, Jeleč Ž, Hudetz D, et al. (2021). Cytokines and Chemokines Involved in Osteoarthritis Pathogenesis. Int J Mol Sci, 22:9208.

[203] Mukherjee A, Das B (2024). The role of inflammatory mediators and matrix metalloproteinases (MMPs) in the progression of osteoarthritis. Biomater Biosyst, 13:100090.

[204] Eveque-Mourroux MR, Emans PJ, Zautsen RRM, Boonen A, Heeren RMA, Cillero-Pastor B (2019). Spatially resolved endogenous improved metabolite detection in human osteoarthritis cartilage by matrix assisted laser desorption ionization mass spectrometry imaging. Analyst, 144:5953-5958.

[205] Welhaven HD, Viles E, Starke J, Wallace C, Bothner B, June RK, et al. (2024). Metabolomic profiles of cartilage and bone reflect tissue type, radiography-confirmed osteoarthritis, and spatial location within the joint. Biochem Biophys Res Commun, 703:149683.

[206] Rocha B, Cillero-Pastor B, Ruiz-Romero C, Paine MRL, Cañete JD, Heeren RMA, et al. (2021). Identification of a distinct lipidomic profile in the osteoarthritic synovial membrane by mass spectrometry imaging. Osteoarthritis Cartilage, 29:750-761.

[207] Editorials NM (2024). Method of the Year 2024: spatial proteomics. Nat Methods, 21:2195-2196.

[208] Sampath Kumar A, Tian L, Bolondi A, Hernández AA, Stickels R, Kretzmer H, et al. (2023). Spatiotemporal transcriptomic maps of whole mouse embryos at the onset of organogenesis. Nat Genet, 55:1176-1185.

[209] Chen A, Liao S, Cheng M, Ma K, Wu L, Lai Y, et al. (2022). Spatiotemporal transcriptomic atlas of mouse organogenesis using DNA nanoball-patterned arrays. Cell, 185:1777-1792.e1721.

[210] Lohoff T, Ghazanfar S, Missarova A, Koulena N, Pierson N, Griffiths JA, et al. (2022). Integration of spatial and single-cell transcriptomic data elucidates mouse organogenesis. Nat Biotechnol, 40:74-85.

[211] Guo G, Papanicolaou M, Demarais NJ, Wang Z, Schey KL, Timpson P, et al. (2021). Automated annotation and visualisation of high-resolution spatial proteomic mass spectrometry imaging data using HIT-MAP. Nat Commun, 12:3241.

[212] Rainbow R, Ren W, Zeng L (2012). Inflammation and Joint Tissue Interactions in OA: Implications for Potential Therapeutic Approaches. Arthritis, 2012:741582.

[213] Garg M, Karpinski M, Matelska D, Middleton L, Burren OS, Hu F, et al. (2024). Disease prediction with multi-omics and biomarkers empowers case-control genetic discoveries in the UK Biobank. Nat Genet, 56:1821-1831.

[214] Jagadeesh KA, Dey KK, Montoro DT, Mohan R, Gazal S, Engreitz JM, et al. (2022). Identifying disease-critical cell types and cellular processes by integrating single-cell RNA-sequencing and human genetics. Nat Genet, 54:1479-1492.

[215] Mobasheri A, Saarakkala S, Finnilä M, Karsdal MA, Bay-Jensen AC, van Spil WE (2019). Recent advances in understanding the phenotypes of osteoarthritis. F1000Res, 8.

[216] Deveza LA, Nelson AE, Loeser RF (2019). Phenotypes of osteoarthritis: current state and future implications. Clin Exp Rheumatol, 37 Suppl 120:64-72.

[217] Roemer FW, Jarraya M, Collins JE, Kwoh CK, Hayashi D, Hunter DJ, et al. (2023). Structural phenotypes of knee osteoarthritis: potential clinical and research relevance. Skeletal Radiol, 52:2021-2030.

[218] Felson DT (2010). Identifying different osteoarthritis phenotypes through epidemiology. Osteoarthritis Cartilage, 18:601-604.

[219] Roos EM, Arden NK (2016). Strategies for the prevention of knee osteoarthritis. Nat Rev Rheumatol, 12:92-101.

[220] Watson T (2024). Anti-obesity drug has life-changing benefits for arthritis. Nature.

[221] Bliddal H, Bays H, Czernichow S, Uddén Hemmingsson J, Hjelmesæth J, Hoffmann Morville T, et al. (2024). Once-Weekly Semaglutide in Persons with Obesity and Knee Osteoarthritis. N Engl J Med, 391:1573-1583.

[222] Sun K, Wu Y, Zeng Y, Xu J, Wu L, Li M, et al. (2022). The role of the sirtuin family in cartilage and osteoarthritis: molecular mechanisms and therapeutic targets. Arthritis Res Ther, 24:286.

[223] Liao Z, Cai X, Zheng Y, Lin J, Yang X, Lin W, et al. (2024). Sirtuin 1 in osteoarthritis: Perspectives on regulating glucose metabolism. Pharmacol Res, 202:107141.

[224] Gil Alabarse P, Chen LY, Oliveira P, Qin H, Liu-Bryan R (2023). Targeting CD38 to Suppress Osteoarthritis Development and Associated Pain After Joint Injury in Mice. Arthritis Rheumatol, 75:364-374.

[225] Chen W, Sun Z, Xiong X, Tan H, Hu J, Liu C, et al. (2024). Exploring the causal link among statin drugs and the osteoarthritis risk based on Mendelian randomization research. Front Genet, 15:1390387.

[226] Farnaghi S, Prasadam I, Cai G, Friis T, Du Z, Crawford R, et al. (2017). Protective effects of mitochondria-targeted antioxidants and statins on cholesterol-induced osteoarthritis. Faseb j, 31:356-367.

[227] Zou Z, Hu W, Kang F, Xu Z, Li Y, Zhang J, et al. (2024). Interplay between lipid dysregulation and ferroptosis in chondrocytes and the targeted therapy effect of metformin on osteoarthritis. J Adv Res:S2090-1232(2024)00155-00153.

[228] Ciaffi J, Mancarella L, Pederzani G, Lisi L, Brusi V, Pignatti F, et al. (2024). Efficacy, Safety, and Tolerability of a Very Low-Calorie Ketogenic Diet in Women with Obesity and Symptomatic Knee Osteoarthritis: A Pilot Interventional Study. Nutrients, 16:3236.

[229] Li W, Guo H, Wang C, Zhang Y, Wang J (2023). Autologous micro-fragmented adipose tissue in the treatment of atherosclerosis patients with knee osteoarthritis in geriatric population: A systematic review and meta-analysis. PLoS One, 18:e0289610.

[230] Wakale S, Wu X, Sonar Y, Sun A, Fan X, Crawford R, et al. (2023). How are Aging and Osteoarthritis Related? Aging Dis, 14:592-604.

[231] Tenchov R, Sasso JM, Wang X, Zhou QA (2024). Antiaging Strategies and Remedies: A Landscape of Research Progress and Promise. ACS Chem Neurosci, 15:408-446.

[232] Thoene M, Bejer-Olenska E, Wojtkiewicz J (2023). The Current State of Osteoarthritis Treatment Options Using Stem Cells for Regenerative Therapy: A Review. Int J Mol Sci, 24:8925.

[233] Fujita M, Sasada M, Iyoda T, Fukai F (2024). Involvement of Matricellular Proteins in Cellular Senescence: Potential Therapeutic Targets for Age-Related Diseases. Int J Mol Sci, 25:6591.

[234] Rastaldo R, Vitale E, Giachino C (2020). Dual Role of Autophagy in Regulation of Mesenchymal Stem Cell Senescence. Front Cell Dev Biol, 8:276.

[235] Leong DJ, Sun HB (2014). Osteoarthritis - Why Exercise? J Exerc Sports Orthop, 1:4.

[236] Liu Y, Li W, Tang H, Yang Z, Wei M, Zhou W, et al. (2024). Ruscogenin attenuates osteoarthritis by modulating oxidative stress-mediated macrophage reprogramming via directly targeting Sirt3. Int Immunopharmacol, 143:113336.

[237] Zhang Y, He X, Ge Z, Wang B, Ni M, Cai G (2024). Investigating the differential therapeutic efficacy and mechanisms of human umbilical cord mesenchymal stem cells at various passages in osteoarthritis treatment. Tissue Cell, 90:102499.

[238] Zhang B (2024). Extended heterochronic parabiosis as an approach to study rejuvenation. Nat Rev Mol Cell Biol, 25:83.

[239] Spector TD, Nandra D, Hart DJ, Doyle DV (1997). Is hormone replacement therapy protective for hand and knee osteoarthritis in women?: The Chingford Study. Ann Rheum Dis, 56:432-434.

[240] Huang ZY, Luo ZY, Cai YR, Chou CH, Yao ML, Pei FX, et al. (2022). Single cell transcriptomics in human osteoarthritis synovium and in silico deconvoluted bulk RNA sequencing. Osteoarthritis Cartilage, 30:475-480.