Osteoarthritis pain inversely correlates with histidine and glutamine following CSF and serum profiling

Osteoarthritis pain inversely correlates with histidine and glutamine following CSF and serum profiling | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Osteoarthritis pain inversely correlates with histidine and glutamine following CSF and serum profiling Isobel Kate Dunstan, Line J.H. Rasmussen, Fay Probert, Dorte A. Olsen, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8742887/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Background Osteoarthritis is a leading cause of pain and disability, yet the biological processes linking peripheral joint pathology with central pain mechanisms and wider symptom burden remain poorly defined. Methods We performed an integrated metabolomic and inflammatory analysis of cerebrospinal fluid and serum obtained from patients with osteoarthritis (n = 81) and pain-free controls (n = 70). Proton nuclear magnetic resonance spectroscopy was used for metabolomic profiling, alongside targeted protein assays for inflammatory mediators. Orthogonal partial least squares discriminant analysis was applied to assess separation between groups and to determine diagnostic accuracy. Associations between metabolites and clinical outcomes, including pain intensity, disability and sleep disturbance, were examined, with adjustment for age and BMI. Results Clear separation between osteoarthritis and control participants was observed in both biofluids, with classification accuracies of 87% for serum and 89% for cerebrospinal fluid. Reduced serum histidine, glutamine, albumin (lysyl) and lysine were key discriminators of osteoarthritis, while elevated lactate and glutamate and reduced glucose and glutamine characterised the cerebrospinal fluid profile. Combining metabolomic data with inflammatory proteins increased diagnostic accuracy to 90% and remained significant after matching for age and BMI. Reductions in serum histidine and glutamine were consistent across subgroups, including stratification by pain severity. These metabolites correlated inversely with pain intensity, disability, sleep disturbance and overall symptom impact, and were more markedly altered in women, who also reported greater symptom burden. Conclusions Osteoarthritis is associated with a distinct pattern of peripheral and central metabolic disturbance. Histidine and glutamine emerge as promising biomarkers related to pain and clinical severity, highlighting metabolic pathways as potential targets for improved stratification and intervention in osteoarthritis pain. Metabolomics Osteoarthritis Neuroinflammation Chronic Pain Biomarkers 1H NMR Spectroscopy cytokines CSF CRP Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Background Osteoarthritis (OA) is a complex joint disease characterized by cartilage degradation, synovial inflammation, and subchondral bone changes ( 1 – 3 ). It is the most prevalent joint disorder in older adults and a major cause of disability ( 2 , 4 ). OA pathogenesis involves multiple factors, including mechanical influences, aging, and genetic predisposition ( 1 ). Synovial inflammation and subchondral bone alterations are increasingly recognized as key contributors to OA onset and progression ( 5 , 6 ). These changes can precede visible cartilage degeneration and are associated with pain and functional impairment ( 6 ). The subchondral bone undergoes uncoupled remodelling, leading to osteosclerosis in advanced stages ( 5 ). Current treatments aim to relieve symptoms and improve joint function, but no curative therapy exists ( 7 ). Emerging strategies targeting cytokine inhibition and signalling pathways offer potential avenues for disease modification ( 8 ), but personalised approaches to prevention and treatment, tailoring interventions to distinct OA phenotypes are likely to be necessary ( 9 ). Biomarkers have emerged as a critical focus in OA research ( 10 ). Recent studies have focused on the role of biomarkers in assessing OA pain severity and subtypes. While serum and synovial fluid biomarkers, such as extracellular matrix molecules, have been extensively studied, the potential of cerebrospinal fluid (CSF) biomarkers is an emerging area of interest ( 11 , 12 ). CSF biomarkers are increasingly recognized as valuable tools in osteoarthritis (OA) research, despite the primary pathology occurring in joints. CSF biomarkers offer unique insights into central nervous system involvement in OA pain. Indeed, studies have identified elevated levels of inflammatory proteins in CSF of OA patients, indicating neuroinflammation ( 13 , 14 ). Interestingly, some CSF biomarkers, particularly those with neuroprotective effects, are associated with reduced pain intensity and milder symptoms ( 14 ). Specific proteins like monocyte chemoattractant protein 1 (MCP1) and interleukin (IL)-8 have been found elevated in CSF, suggesting their role in neuroimmune signalling ( 13 ). However, hitherto there has been no investigation of the metabolome profile of CSF from individuals with OA. Sex differences are a recognised feature of OA, with women experiencing higher prevalence, pain intensity, and poorer clinical outcomes compared to men. Women account for 60% of OA cases globally ( 15 ) and report more severe pain, functional limitations, and poorer surgical outcomes ( 15 , 16 ). These differences are influenced by hormonal factors, joint anatomy, central pain processing, and psychosocial aspects ( 17 , 18 ). The interplay between inflammatory markers and pain also varies by sex, suggesting distinct biological mechanisms underlying OA progression ( 17 ). Addressing these disparities is crucial for tailoring interventions and improving treatment outcomes across sexes ( 19 , 20 ). Metabolomics provides a comprehensive view of metabolic changes in OA, revealing alterations in amino acid, lipid, and energy metabolism ( 21 , 22 ). Nuclear magnetic resonance (NMR) spectroscopy of biofluids, particularly synovial fluid, has shown promise in identifying OA biomarkers ( 23 ). Here, we sought to test the hypothesis that osteoarthritis is accompanied by a measurable metabolic disturbance in both serum and CSF, and that this disturbance relates to the pain phenotype. To address this, we undertook a multiomics analysis integrating ¹H NMR metabolomics and targeted protein measurements from paired blood and CSF samples. The study had three objectives: first, to compare the metabolic and inflammatory profiles of OA and pain-free controls in serum and CSF; second, within OA, to determine whether the key discriminating metabolites and proteins were associated with clinical measures of pain and disability; and third, to explore whether these relationships differed by sex, given the higher symptomatic burden observed in women. Methods Aim The aim of this study was to determine whether osteoarthritis is associated with a metabolic and inflammatory signature in paired serum and cerebrospinal fluid, and whether key discriminating features relate to pain severity, symptom impact, and sex. Participants Participants aged 18–80 years who were able to understand and respond to the questionnaires were recruited as part of The Danish Pain Research Biobank (DANPAIN Biobank), as previously described ( 24 ). For the current study, samples were included from two groups: individuals with osteoarthritic pain scheduled for hip or knee arthroplasty surgery at a regional hospital (n = 81), and pain-free volunteers recruited locally (n = 70). Detailed recruitment procedures and cohort characteristics have been described previously ( 24 ). The study was conducted in accordance with the Declaration of Helsinki, with written informed consent obtained from all participants, and approval granted by the relevant regional ethics committees, as described previously ( 24 ). Sample Collection Blood (24 mL) and CSF (7 mL) samples were obtained following standard protocols. Serum was collected in plain tubes and stored at room temperature for 30 minutes before centrifugation. All samples were centrifuged at 2,000 G for 10 minutes at 4°C, aliquoted on ice blocks, and stored at − 80°C. CSF was collected via lumbar puncture using atraumatic needles and stored at − 80°C within 30 minutes of collection. Separate aliquots were analysed for neuroimmune biomarkers and for metabolomics to avoid freeze-thaw effects. CSF contaminated with blood was excluded, and samples were stored under identical conditions to blood. Clinical and Pain Assessments Functional and psychological impacts were assessed using validated instruments, including the Symptom Impact Questionnaire (SIQR) ( 25 ), Major Depression Inventory (MDI), Generalized Anxiety Disorder (GAD-7), Insomnia Severity Index (ISI), EuroQol visual analogue scale (EQ VAS), and the Oxford Hip or Knee score (OHS/OKS) ( 26 ). Pain intensity and interference were quantified using a Numeric Rating Scale (NRS) ( 27 ) and Pain Disability Index (PDI) ( 28 ), respectively. All clinical, pain, and questionnaire-based assessments were completed within 24 hours of blood and cerebrospinal fluid sampling, and in all cases prior to surgery for participants undergoing arthroplasty. Questionnaires were administered electronically using a standardised clinical registry platform, as previously described ( 24 ). Analysis of neuroimmune biomarkers A panel of biomarkers in CSF and blood was quantified using multiplex assays from Mesoscale Discovery as described elsewhere ( 29 ). High-sensitivity C-reactive protein (hsCRP, mg/L) was analyzed with the CRP high sensitive ELISA (TECAN, IBL International GmbH, Hamburg, Germany) according to the manufacturer’s instructions. Serum suPAR (suPAR, ng/mL) was analyzed with the suPARnostic® AUTO Flex ELISA (ViroGates A/S, Birkerød, Denmark) according to manufacturer’s instructions. Neurofilament light (NfL) and glial fibrillary acidic protein (GFAP) in CSF and serum were measured blinded to clinical data using a commercially available NfL and GFAP 2-plex assay on the Single Molecule Array (Simoa®) HD-X Analyzer (Quanterix, Billerica, MA, USA) ( 30 ). Further details on all assays, including limits of detection and assay ranges are listed in the Supplementary methods S1.1. Sample Processing and 1H NMR data collection Serum (150 µL) and CSF (100 µL) samples were mixed with 400 µL and 450 µL of NMR buffer (75 mM sodium phosphate buffer prepared in D 2 O, pH 7.4), respectively. All samples were then placed in a 5 mm NMR tube and measured using a 700-MHz Bruker AVII spectrometer operating at 16.4 T, equipped with a 1 H ( 13 C/ 15 N) TCI cryoprobe (Chemistry Research Laboratory, Department of Chemistry, University of Oxford), as described previously ( 31 , 32 ). Further details can be found in Supplementary methods S1.2. 1H NMR Data Processing Topspin 4.1.4 (Bruker) was used to phase, baseline correct, and reference all spectra to the lactate methyl resonance at δ = 1.33 ppm. ACD/Labs Spectrus Processor Academic Edition 12.01 (Advanced Chemistry Development, Inc.) was used to manually bin each resonance signal, excluding all noise from analysis. The integrals of these bins were sum normalised and exported to R version 4.3.0 ( 33 ). Metabolite assignment was conducted through a mix of in-house databases, literature reviews, 2D total correlation spectroscopy (TOCSY) experiments, and spiking reference compounds where needed ( 34 , 35 ). Multivariate Statistical Analysis Multivariate analyses were performed to identify key metabolite differences between groups, with rigorous cross-validation to ensure robustness. Specifically, the ropls package in R ( 36 ) was used to conduct orthogonal partial least squares discriminant analysis (OPLS-DA) between pain-free and OA predictor variables (i.e. all serum and CSF metabolites and proteins from the targeted panels, z-scaled). Model performance was evaluated against a null distribution generated by random permutation of class labels to assess the likelihood of spurious separation. Descriptions of the model building and cross validation scheme can be found in Supplementary methods S1.3. Univariate Statistical Analysis Univariate analyses such as Student’s T test and two-way ANOVA were carried out in R (v 4.3.0). The significance threshold was set to ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. All correlations were tested via Pearson correlational analyses, using the Benjamini-Hochberg method to correct for false discovery rate at 0.05. Univariate Receiver Operating Characteristic (ROC) analyses and multivariate ROC analyses on a combination of features using logistic regression were determined using the pROC package in R. Results Distinct serum and CSF metabolite profiles, and their integration with inflammatory markers, differentiate osteoarthritis from controls. We first examined whether untargeted ¹H NMR spectra from single biofluids could discriminate individuals with OA from pain-free controls. OPLS-DA models built on serum profiles showed a clear separation between groups (Fig. 1 A(i)), with a classification accuracy of 87%, sensitivity of 90% and specificity of 86%, all significantly better than the null permutation model (Fig. 1 A(iii–v)). The principal discriminatory serum signals were increased mobile, lipid-related resonances in the chylomicron/VLDL region, together with 3-hydroxybutyrate, and lower concentrations of amino acids including lysine, alanine and glutamine (Fig. 1 A(ii) and Supplementary Figure S2A(i-iv) for boxplots). CSF profiles gave a very similar picture, again separating OA from controls (Fig. 1 B(i)) and doing so with slightly higher diagnostic performance (accuracy 89%, sensitivity 94%, specificity 85%, Fig. 1 B(iii–v)). Here, the highest ranking variables were lactate, glutamine, glutamate and glucose (Fig. 1 B(ii) and Supplementary Figure S2B(i-iv) for boxplots), which indicates that metabolic disturbance is present in the central compartment as well as in the periphery. When serum and CSF metabolites were analysed together with inflammatory and neurodegeneration-related proteins, model performance improved further, with the integrated model achieving 90% accuracy, 90% sensitivity and 89% specificity (Fig. 1 C(iii–v)), driven chiefly by serum histidine and serum glutamine together with CSF CRP and CSF GFAP (Fig. 1 C(ii)). A random forest approach confirmed that OA and controls could also be separated using an orthogonal modelling strategy (Supplementary Figure S1 ), and the distribution of the top serum and CSF VIPs is shown as box plots in Supplementary Figure S2. These plots illustrate that, in serum, OA was characterised by higher very low density lipoprotein signals (mobile CH₃ in chylomicron/VLDL) and 3-hydroxybutyrate, together with lower lysine, alanine and glutamine. In CSF, lactate and glutamate were increased, while glutamine and glucose were reduced in OA compared with controls. These findings suggest distinct metabolic alterations in both systemic and, more surprisingly, central compartments in OA. Age- and BMI-matched analysis confirms a core OA metabolic-inflammatory signature independent of demographic differences. To exclude the possibility that the discrimination observed in the full cohort was driven by age and BMI, we rebuilt the model in an age and BMI matched subset (OA n = 31, controls n = 28). In this restricted sample, age and BMI no longer differed significantly between groups and only a modest imbalance in sex remained (which was not present in the full cohort, Fig. 2 A(ii)), yet the OPLS-DA model still separated OA from controls with an accuracy of 77% (Fig. 2 A(i)), significantly better than the null model (Supplementary Figure S3). The VIP ranking from this matched model again placed serum histidine at the top, followed by serum gp130, serum albumin (lysyl moiety), CSF glucose and serum glutamine (Fig. 2 A(iii)), several of which were also among the highest ranking variables in the full integrated model (Fig. 1 C), indicating that these signals are not explained by demographic confounders. We then took only those variables that remained significantly different in the small matched cohort and refitted a logistic model in the full dataset. This reduced panel still achieved a strong diagnostic performance, with an AUC of 0.906 (95% CI 0.858–0.953; Fig. 2 A(iv)), showing that a small, demographically robust signature is sufficient to distinguish OA from pain-free individuals. The direction and magnitude of these key variables in the whole cohort are illustrated in Fig. 2 B, where serum histidine, serum albumin (lysyl), serum glutamine and serum lysine are consistently lower in OA, while serum citrate, serum gp130, CSF acetate and CSF CRP are higher. Corresponding box plots for the matched subset are provided in Supplementary Figure S4 and a side by side comparison of VIP scores in the full and matched models is shown in Supplementary Figure S3B. Core metabolic signature associates with pain burden, symptom impact and sleep disturbance. To determine whether the demographically robust metabolic and inflammatory variables were clinically meaningful, we examined their correlations with pain and disability measures within the whole cohort. The restricted panel of serum metabolites (histidine, albumin lysyl resonance, glutamine, lysine and citrate) was strongly intercorrelated, forming a coherent serum metabolic module, and this module correlated inversely with several symptom scores, including the SIQR, total NRS, ISI and WPI (Fig. 3 ). In other words, lower serum histidine, albumin and glutamine were associated with greater symptom impact, higher pain ratings and poorer sleep. In contrast, central and inflammatory markers that survived matching, in particular CSF CRP, CSF acetate and CSF glutamate, showed positive correlations with these same clinical measures, consistent with a contribution of central immune activation to symptom severity. Anxiety and mood scores (GAD7, MDI) followed the same pattern, although the associations were weaker. A broader correlation matrix including the full inflammatory panel reproduced these relationships and showed additional links to CSF TNFRI/II and GFAP (Supplementary Figure S7), supporting the view that OA pain is accompanied by a coupled metabolic inflammatory disturbance across serum and CSF. Serum metabolites track intra-disease variation in pain intensity. We next asked whether the metabolic signature identified above also captured symptom severity within osteoarthritis itself. OA participants were stratified into low and high pain groups using the lower and upper quartiles of the total NRS (Fig. 4 A). Despite similar age, sex distribution, BMI and Kellgren–Lawrence scores between these two groups (Fig. 4 D), several of the key serum variables showed clear, directional differences. Serum albumin (lysyl resonance), histidine, lysine and glutamine were all lower in the high pain group, whereas the chylomicron/VLDL signal and serum GFAP, likely derived from chondrocytes ( 37 ), were higher (Fig. 4 B(i–vi)). Since structural burden and demographic factors did not differ, these findings indicate that this subset of metabolites reflects the pain phenotype rather than radiographic severity. A logistic model built only from the pain sensitive variables discriminated high from low pain OA with an AUC of 0.824 (95% CI 0.663 to 0.984; Fig. 4 C), suggesting that a small serum panel may have utility for phenotyping painful OA. Consistent with this, serum histidine also associated with broader measures of impact, showing lower values in participants with higher Pain Disability Index and SIQR scores and with poorer EQ-VAS ratings (Fig. 4 E). Together, these data show that the metabolic disturbance identified at disease level is graded by pain intensity and is therefore clinically meaningful. Female osteoarthritis shows a heavier symptom burden and a more pronounced metabolic–inflammatory disturbance. Because women are disproportionately affected by symptomatic OA, we examined whether the discriminatory signature differed by sex ( 21 ). When analysed separately, OPLS-DA models built from female participants showed robust separation of OA from controls (accuracy 88%), whereas the model in males, although still significant, was weaker (accuracy 76%) (Fig. 5 A,B; Supplementary Figure S5). The top VIPs in women were the same amino acid related serum variables that survived age/BMI matching, in particular serum albumin (lysyl), serum histidine and serum glutamine, whereas in men CSF CRP rose to the top of the ranking (Fig. 5 A(ii), 5B(ii)). Clinically, women with OA reported higher SIQR, ISI and, to a lesser extent, MDI scores than men, and for SIQR and ISI there were significant diagnosis by sex interactions, indicating that the disease effect is larger in females (Fig. 5 C(i–iii); Supplementary Table S6). Consistent with this, the metabolite changes that defined OA in the whole cohort, notably lower serum glutamine, histidine and albumin (lysyl), were clearly expressed in women but were weaker or absent in men (Fig. 5 C(iv–vi)). Interestingly, CSF glutamate was elevated in females with OA only, potentially related to their higher pain and symptom scores (Fig. 5 C(vii)). While CSF CRP was the top driver of the male OPLS-DA model, two-way ANOVA revealed it was significantly increased in both men and women with OA (Fig. 5 C(viii)). Taken together, these findings suggest that in women the OA phenotype is driven more by metabolic and excitatory changes that track symptom burden and men in this cohort display a similar, but less pronounced phenotype. Discussion This study asked whether OA has a reproducible biofluid signature, whether that signature relates to pain, and whether it differs by sex. Using untargeted ¹H NMR in serum and CSF, supported by targeted inflammatory markers, we showed that OA can be distinguished from pain-free controls with high accuracy, and that this signal persists after age and BMI matching. The same small set of variables, notably lower serum histidine, also tracked clinical severity, since OA cases with the greatest pain and symptom impact had the most divergent metabolite values. Finally, these effects were most pronounced in women, indicating that combined serum–CSF profiling can capture sex-specific pathways contributing to symptomatic OA. Metabolic Alterations in OA Our results reveal distinct metabolic signatures in OA, characterised by increases in VLDL and 3-hydroxybutyrate and decreases in lysine, alanine, histidine, and glutamine in serum, along with elevated lactate and glutamate and reduced glutamine and glucose in CSF. These results are consistent with growing evidence suggesting that metabolic dysregulation, particularly in amino acid and lipid metabolism, is linked to inflammation and joint degradation in OA ( 38 ). Chondrocyte metabolism is known to be affected by environmental stressors, which can lead to shifts between metabolic pathways and mitochondrial dysfunction ( 39 ). Furthermore, obesity and metabolic syndrome, often comorbid with OA, exacerbate these metabolic changes through complex interactions of biomechanical, inflammatory, and metabolic factors ( 40 ). These findings support the emerging concept of OA as a metabolic disease, with lipids and adipokines playing crucial roles in cartilage degradation ( 41 , 42 ). Similarly, reductions in amino acids such as lysine and glutamine may indicate their increased utilisation in pro-inflammatory and metabolic pathways ( 42 ). The differentiation between serum and CSF metabolite profiles provides further support that OA also involves systemic and central metabolic changes. Evidence suggests that elevated lactate and glutamate levels in CSF may reflect altered energy metabolism and excitotoxicity in OA, contributing to central pain sensitisation and neuroinflammatory responses. Studies have found increased levels of glutamate and other excitatory amino acids in synovial fluid of OA patients ( 43 , 44 ), however, this is the first report of an imbalance in the CNS. Metabolic changes in synovial fibroblasts and fluid, particularly in the glutamine-glutamate pathway, have been associated with inflammatory responses in OA ( 45 ). Exposure of rat spinal cord slices to CSF from OA patients has been shown to affect neuronal excitability, potentially altering nociceptive processing at the spinal level ( 46 ). Elucidating the altered composition of OA CSF here suggests that glutamate signalling may contribute to peripheral nociceptive transduction and inflammation in OA ( 47 ). Additionally, reactive astrocytes and their release of lactate have been implicated in chronic pain development and central sensitisation ( 48 – 50 ). These findings underscore the importance of considering systemic and central contributions to OA pathology and the potential utility of targeting metabolic pathways in therapeutic interventions. Diagnostic Potential of Biomarker Integration Combining serum and CSF metabolites with a targeted protein panel improved the diagnostic accuracy of OA models, reaching 90%. This integration highlights the value of a multi-modal biomarker approach in capturing the complex and multifactorial nature of OA. Previous research has shown the limitations of single biomarkers in diagnosing OA due to the heterogeneity of the disease ( 51 ). By incorporating markers of inflammation, neurodegeneration, and joint degradation, our model achieved a higher discriminatory power and highlighted unique relationships between metabolites and proteins in CSF and serum. The persistence of certain metabolites, such as reduced serum histidine and glutamine, in both the full and matched cohorts suggests their robustness as potential biomarkers of OA. Histidine is an essential amino acid with anti-inflammatory properties ( 52 ). Indeed, we show here that serum histidine levels are inversely correlated with cytokines both in the serum and CSF (Figure S7). Others have also highlighted histidine as a potential metabolomic biomarker for knee OA in which the ratio of branched-chain amino acids to histidine seems to be important ( 53 ). Pain-Associated Metabolic Changes One of the key findings of this study is the association of specific serum metabolites with OA pain intensity. Decreases in serum histidine, glutamine, lysine, and albumin (lysyl moiety) were observed in participants with high pain intensity, alongside an increase in serum GFAP. The observed reduction in histidine and glutamine in high-pain groups may reflect their roles in modulating oxidative stress and inflammatory responses. Histidine has been shown to scavenge reactive oxygen species (ROS) and inhibit inflammatory pathways in various studies ( 52 , 54 ). It can reduce fluid accumulation and protect intestinal tissue during inflammation ( 54 ), as well as inhibit IL-8 secretion in intestinal epithelial cells ( 55 ). Glutamine serves as a substrate for glutathione synthesis, which is crucial for antioxidant activities and immune cell function ( 56 , 57 ). Both amino acids have demonstrated potential in modulating respiratory burst in neutrophils ( 58 ). Their immunomodulatory effects make them promising candidates for therapeutic interventions in various conditions, including inflammation, infection, and oxidative stress-related disorders ( 52 , 59 ). The elevation of serum GFAP, potentially derived from chondrocytes, suggests a link between joint damage and systemic markers of neurodegeneration, further supporting the interconnectedness of peripheral and central mechanisms in OA pain ( 60 ). The correlation between these metabolic changes and clinical pain scores, such as the NRS and SIQR, provides a compelling case for the use of serum metabolites as objective measures of pain severity. This could address the current reliance on subjective pain assessments in clinical practice and trials, improving the precision of pain management strategies. Sex Differences in OA Female participants exhibited more severe OA symptoms and greater metabolic changes compared to men, consistent with broader findings of higher pain prevalence, worse functional outcomes, and distinct inflammatory profiles ( 15 ). The greater reductions in serum histidine and glutamine, as well as the notably higher CSF glutamate observed in females may reflect sex-specific differences in inflammatory and metabolic processes. Oestrogen influences cartilage metabolism, immune responses, and pain perception, particularly postmenopause, exacerbating OA pathology ( 61 ). Systematic reviews suggest oestrogen replacement therapy may lower OA prevalence by modulating cartilage turnover ( 62 , 63 ). Furthermore, metabolic imbalances influenced by oestrogen contribute to OA progression ( 64 ). Psychosocial factors, including anxiety and depression – affecting ~ 20% of OA patients ( 65 ) – correlate moderately with pain severity and are more common in women ( 66 ). While the causal link remains unclear, evidence suggests interventions such as cognitive-behavioural approaches may alleviate OA pain and disability ( 67 , 68 ). The pronounced metabolite changes in women highlight the need to consider sex differences in OA research and treatment. Tailored interventions that address the unique metabolic and inflammatory pathways in females could improve clinical outcomes and reduce the sex disparity in OA burden. Implications for Clinical Practice and Research This study provides a foundation for the development of biomarker-based diagnostic and therapeutic strategies in OA. Multi-modal biomarker panels integrating serum and CSF metabolites with targeted proteins could improve diagnostic accuracy, stratify patients by disease severity, and inform personalised OA management. The identification of robust biomarkers, such as histidine and glutamine, highlights potential therapeutic targets that warrant further investigation. Interventions aimed at restoring these metabolic pathways could mitigate inflammation and pain, offering a new avenue for disease-modifying treatments. Additionally, the sex-specific differences observed in this study call for the inclusion of sex-stratified analyses in future OA research to ensure equitable and effective treatment development. Limitations and Future Directions While this study advances our understanding of OA biomarkers, it has several limitations. The cross sectional design precludes causal inference, and the sample size, particularly in the matched analyses, may limit generalisability. We were not able to perform an independent external validation, largely because major population resources such as UK Biobank do not include CSF, which makes replication of a paired serum-CSF analysis challenging, but at the same time underscores the distinctiveness of the present cohort. Longitudinal studies are needed to define the temporal relationship between these metabolic alterations and OA progression and to test whether the serum variables that track pain can predict future symptom flares. Experimental work to probe the functional roles of histidine, glutamine and the CSF inflammatory markers in OA pathophysiology would also be valuable. Finally, combining this biofluid panel with imaging measures and a broader inflammatory/metabolic annotation could improve both diagnostic performance and biological interpretability. Conclusions In summary, this study identifies distinct metabolic and inflammatory profiles in OA. Specifically, reduced serum histidine and glutamine emerged as key discriminators, correlating inversely with pain intensity and disability scores and should be further explored for their role in OA pathophysiology and pain modulation as well as possible targets for personalised treatment strategies. Abbreviations AUC, area under the curve BMI, body mass index CRP, C reactive protein CSF, cerebrospinal fluid DANPAIN, Danish Pain Research Biobank D2O, deuterium oxide EQ VAS, EuroQol visual analogue scale GAD 7, Generalised Anxiety Disorder 7 item scale GFAP, glial fibrillary acidic protein hsCRP, high sensitivity C reactive protein IL, interleukin ISI, Insomnia Severity Index KL, Kellgren to Lawrence grade MCP 1, monocyte chemoattractant protein 1 MDI, Major Depression Inventory NfL, neurofilament light chain NMR, nuclear magnetic resonance NRS, numeric rating scale OA, osteoarthritis OHS, Oxford Hip Score OKS, Oxford Knee Score OPLS DA, orthogonal partial least squares discriminant analysis PDI, Pain Disability Index ROC, receiver operating characteristic SIQR, Symptom Impact Questionnaire Revised Simoa, Single Molecule Array suPAR, soluble urokinase plasminogen activator receptor TNF, tumour necrosis factor TNFR I, tumour necrosis factor receptor 1 TNFR II, tumour necrosis factor receptor 2 TOCSY, total correlation spectroscopy VLDL, very low density lipoprotein Declarations Ethics approval and consent to participate The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). Approval was obtained from the Regional Ethics Committee (references S-20160173 and S-20180003), and the study was duly notified to the Danish Data Protection Agency (reference 17/3391). All participants provided written informed consent prior to enrolment, including consent for sample collection, clinical data acquisition, and subsequent analysis for research purposes. No individual-identifying information is presented in this manuscript. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Funding Several private and public foundations funded the building of the DANPAIN Biobank but did not participate in any aspect of study design, data collection, data analysis, data interpretation, or writing of the manuscript. These funders include: The Danish Rheumatism Association (grants R155-A4866-B1363 and R175-A6088-B1363); Aase og Ejnar Danielsens Fond; Karen S. Jensens Legat; Oberstinde Kirsten Jensens la Cours Legat; Professor, Overlæge Sophus H. Johansens Fond of 23 August 1981; Foundation of the Danish Association for Anaesthesia and Intensive Care; Fonden til Lægevidenskabens Fremme; The Development Foundation, Lillebaelt Hospital; OUH Fund for Free Research; The Research Fund of the Danish Society of Anaesthesiology and Intensive Care Medicine. Author Contribution Isobel Kate Dunstan contributed to data curation, formal analysis, investigation, methodology, software, validation, visualisation, and writing – original draft. Line Jee Hartmann Rasmussen contributed to data curation, investigation, and writing – review & editing. Fay Probert contributed to software, methodology, supervision, and writing – review & editing. Dorte Aalund Olsen contributed to data curation, investigation, and writing – review & editing. Jonna Skov Madsen contributed to data curation, investigation, and writing – review & editing. Jesper Eugen-Olsen contributed to data curation, investigation, and writing – review & editing. Thomas Peter Enggaard contributed to conceptualisation and writing – review & editing. Claus Varnum contributed to data curation and writing – review & editing. Kate Lykke Lambertsen contributed to conceptualisation, investigation, methodology, data interpretation, and writing – review & editing. Daniel C Anthony contributed to conceptualisation, formal analysis, supervision, and writing – review & editing. Morten Rune Blichfeldt-Eckhardt contributed to conceptualisation, data curation, data interpretation, and writing – review & editing. Acknowledgement Data storage and analysis were supported by the OPEN research network (Odense Patient data Explorative Network), Region of Southern Denmark. All metabolomic and clinical data processing was conducted securely using the OPEN Analyse environment, which provides controlled and auditable access to pseudonymised data in compliance with data protection legislation. We thank the OPEN team for their technical support and for providing this secure analysis infrastructure. Data Availability The datasets generated and analysed during the current study are not publicly available as they are subject to ethical and governance restrictions associated with the DANPAIN Biobank, but are available from the corresponding author on reasonable request. References Goldring MB, Goldring SR. Osteoarthr J Cell Physiol. 2007;213(3):626–34. Kim J-R, Yoo J, Kim H. Therapeutics in osteoarthritis based on an understanding of its molecular pathogenesis. Int J Mol Sci. 2018;19(3):674. Wei Y, Bai L. Recent advances in the understanding of molecular mechanisms of cartilage degeneration, synovitis and subchondral bone changes in osteoarthritis. Connect Tissue Res. 2016;57(4):245–61. Li G, Yin J, Gao J, Cheng TS, Pavlos NJ, Zhang C, et al. Subchondral bone in osteoarthritis: insight into risk factors and microstructural changes. Arthritis Res Ther. 2013;15(6):223. Hügle T, Geurts J. What drives osteoarthritis?—synovial versus subchondral bone pathology. Rheumatology. 2016;56(9):1461–71. Mathiessen A, Conaghan PG. Synovitis in osteoarthritis: current understanding with therapeutic implications. Arthritis Res Ther. 2017;19(1):18. Altman RD. Overview of osteoarthritis. Am J Med. 1987;83(4):65–9. Malemud CJ, Islam N, Haqqi TM. Pathophysiological mechanisms in osteoarthritis lead to novel therapeutic strategies. Cells Tissues Organs. 2003;174(1–2):34–48. Mobasheri A, Batt M. An update on the pathophysiology of osteoarthritis. Ann Phys Rehabil Med. 2016;59(5–6):333–9. Attur M, Krasnokutsky-Samuels S, Samuels J, Abramson SB. Prognostic biomarkers in osteoarthritis. Curr Opin Rheumatol. 2013;25(1):136–44. Kalogera S, Jansen MP, Bay-Jensen A-C, Frederiksen P, Karsdal MA, Thudium CS, et al. Relevance of biomarkers in serum vs. synovial fluid in patients with knee osteoarthritis. Int J Mol Sci. 2023;24(11):9483. Mobasheri A. Osteoarthritis year 2012 in review: biomarkers. Osteoarthritis Cartilage. 2012;20(12):1451–64. Kosek E, Finn A, Ultenius C, Hugo A, Svensson C, Ahmed AS. Differences in neuroimmune signalling between male and female patients suffering from knee osteoarthritis. J Neuroimmunol. 2018;321:49–60. Palada V, Ahmed AS, Freyhult E, Hugo A, Kultima K, Svensson CI, et al. Elevated inflammatory proteins in cerebrospinal fluid from patients with painful knee osteoarthritis are associated with reduced symptom severity. J Neuroimmunol. 2020;349:577391. Segal NA, Nilges JM, Oo WM. Sex differences in osteoarthritis prevalence, pain perception, physical function and therapeutics. Osteoarthritis Cartilage. 2024;32(9):1045–53. Glass N, Segal NA, Sluka KA, Torner JC, Nevitt MC, Felson DT, et al. Examining sex differences in knee pain: the Multicenter Osteoarthritis Study. Osteoarthritis Cartilage. 2014;22(8):1100–6. Perruccio AV, Badley EM, Power JD, Canizares M, Kapoor M, Rockel J, et al. Sex differences in the relationship between individual systemic markers of inflammation and pain in knee osteoarthritis. Osteoarthr Cartil Open. 2019;1(1–2):100004. Sluka KA, Berkley KJ, O’Connor MI, Nicolella DP, Enoka RM, Boyan BD, et al. Neural and psychosocial contributions to sex differences in knee osteoarthritic pain. Biol Sex Differ. 2012;3(1):26. Tschon M, Contartese D, Pagani S, Borsari V, Fini M. Gender and sex are key determinants in osteoarthritis not only confounding variables. A systematic review of clinical data. JCM. 2021;10(14):3178. Valdrighi N, Vago JP, Blom AB, Loo, FAJvd. Blaney Davidson EN. Innate immunity at the core of sex differences in osteoarthritic pain? Front Pharmacol. 2022;13:881500. Adams SB, Setton LA, Nettles DL. The role of metabolomics in osteoarthritis research. J Am Acad Orthop Surg. 2012;21(1):63–4. Zhai G, Randell EW, Rahman P. Metabolomics of osteoarthritis: emerging novel markers and their potential clinical utility. Rheumatology. 2018;57(12):2087–95. Clarke EJ, Anderson JR, Peffers MJ. Nuclear magnetic resonance spectroscopy of biofluids for osteoarthritis. Br Med Bull. 2020;137(1):28–41. Blichfeldt-Eckhardt MR, Mortensen WCP, Varnum C, Bendix L, Lauridsen JT, Jensen HI, et al. The Danish Pain Research Biobank (DANPAIN-Biobank): a collection of blood, cerebrospinal fluid, and clinical data for the study of neuroimmune and glia-related biomarkers of chronic pain. Ann Transl Med. 2023;11(10):343. Friend R, Bennett RM. Evaluating disease severity in chronic pain patients with and without fibromyalgia: a comparison of the Symptom Impact Questionnaire and the Polysymptomatic Distress Scale. J Rheumatol. 2015;42(12):2404–11. Harris KK, Dawson J, Jones LD, Beard DJ, Price AJ. Extending the use of PROMs in the NHS-using the Oxford Knee Score in patients undergoing non-operative management for knee osteoarthritis: a validation study. BMJ Open. 2013;3(8):e003365. Ornetti P, Dougados M, Paternotte S, Logeart I, Gossec L. Validation of a numerical rating scale to assess functional impairment in hip and knee osteoarthritis: comparison with the WOMAC function scale. Ann Rheum Dis. 2011;70(5):740–6. Tait RC, Pollard CA, Margolis RB, Duckro PN, Krause SJ. The Pain Disability Index: psychometric and validity data. Arch Phys Med Rehabil. 1987;68(7):438–41. Martin NA, Bonner H, Elkjaer ML, D'Orsi B, Chen G, Konig HG, et al. BID mediates oxygen-glucose deprivation-induced neuronal injury in organotypic hippocampal slice cultures and modulates tissue inflammation in a transient focal cerebral ischemia model without changing lesion volume. Front Cell Neurosci. 2016;10:14. Nielsen HH, Soares CB, Hogedal SS, Madsen JS, Hansen RB, Christensen AA, et al. Acute Neurofilament Light Chain plasma levels correlate with stroke severity and clinical outcome in ischemic stroke patients. Front Neurol. 2020;11:448. Chan SY, Probert F, Radford-Smith DE, Hebert JC, Claridge TDW, Anthony DC, et al. Post-inflammatory behavioural despair in male mice is associated with reduced cortical glutamate-glutamine ratios, and circulating lipid and energy metabolites. Sci Rep. 2020;10(1):16857. Probert F, Yeo T, Zhou Y, Sealey M, Arora S, Palace J, et al. Determination of CSF GFAP, CCN5, and vWF levels enhances the diagnostic accuracy of clinically defined MS from non-MS Patients with CSF oligoclonal bands. Front Immunol. 2021;12:811351. R Core Team. R: a language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2024. Snytnikova OA, Khlichkina AA, Sagdeev RZ, Tsentalovich YP. Evaluation of sample preparation protocols for quantitative NMR-based metabolomics. Metabolomics. 2019;15(6):84. Lin HT, Cheng ML, Lo CJ, Hsu WC, Lin G, Liu FC. (1)H NMR metabolomic profiling of human cerebrospinal fluid in aging process. Am J Transl Res. 2021;13(11):12495–508. Thevenot EA, Roux A, Xu Y, Ezan E, Junot C. Analysis of the human adult urinary metabolome variations with age, Body Mass Index, and gender by implementing a comprehensive workflow for univariate and OPLS statistical analyses. J Proteome Res. 2015;14(8):3322–35. Kanazawa S, Nishizawa S, Takato T, Hoshi K. Biological roles of glial fibrillary acidic protein as a biomarker in cartilage regenerative medicine. J Cell Physiol. 2017;232(11):3182–93. Gkretsi V, Simopoulou T, Tsezou A. Lipid metabolism and osteoarthritis: Lessons from atherosclerosis. Prog Lipid Res. 2011;50(2):133–40. Zheng L, Zhang Z, Sheng P, Mobasheri A. The role of metabolism in chondrocyte dysfunction and the progression of osteoarthritis. Ageing Res Rev. 2021;66:101249. June RK, Liu-Bryan R, Long F, Griffin TM. Emerging role of metabolic signaling in synovial joint remodeling and osteoarthritis. J Orthop Res. 2016;34(12):2048–58. Masuko K, Murata M, Suematsu N, Okamoto K, Yudoh K, Nakamura H, et al. A metabolic aspect of osteoarthritis: lipid as a possible contributor to the pathogenesis of cartilage degradation. Clin Exp Rheumatol. 2009;27(2):347–53. Wu X, Fan X, Crawford R, Xiao Y, Prasadam I. The metabolic landscape in osteoarthritis. Aging Dis. 2022;13(4):1166. Damyanovich A, Staples J, Marshall K. 1H NMR investigation of changes in the metabolic profile of synovial fluid in bilateral canine osteoarthritis with unilateral joint denervation. Osteoarthritis Cartilage. 1999;7(2):165–72. McNearney T, Speegle D, Lawand N, Lisse J, Westlund K. Excitatory amino acid profiles of synovial fluid from patients with arthritis. J Rheumatol. 2000;27(3):739–43. Farah H, Wijesinghe SN, Nicholson T, Alnajjar F, Certo M, Alghamdi A, et al. Differential Metabotypes in Synovial Fibroblasts and Synovial Fluid in Hip Osteoarthritis Patients Support Inflammatory Responses. Int J Mol Sci. 2022;23(6):3266. Lu V, Peter AS. The excitability of dorsal horn neurons is affected by cerebrospinal fluid from humans with osteoarthritis. Can J Physiol Pharmacol. 2012;90(6):783–90. Wen Z, Chang Y-C, Jean Y. Excitatory amino acid glutamate: role in peripheral nociceptive transduction and inflammation in experimental and clinical osteoarthritis. Osteoarthritis Cartilage. 2015;23(11):2009–16. Lluch E, Torres R, Nijs J, Oosterwijck J. Evidence for central sensitization in patients with osteoarthritis pain: A systematic literature review. Eur J Pain. 2014;18(10):1367–75. James T, Mercedes B. Reactive astrocytes: critical players in the development of chronic pain. Front Psychiatry. 2021;12:682056. Liu-Bryan R. Inflammation and intracellular metabolism: new targets in OA. Osteoarthritis Cartilage. 2015;23(11):1835–42. Lotz M, Martel-Pelletier J, Christiansen C, Brandi ML, Bruyère O, Chapurlat R, et al. Value of biomarkers in osteoarthritis: current status and perspectives. Ann Rheum Dis. 2013;72(11):1756–63. Holeček M. Histidine in Health and Disease: Metabolism, Physiological Importance, and Use as a Supplement. Nutrients. 2020;12(3):848. Zhai G, Wang-Sattler R, Hart DJ, Arden NK, Hakim AJ, Illig T, et al. Serum branched-chain amino acid to histidine ratio: a novel metabolomic biomarker of knee osteoarthritis. Ann Rheum Dis. 2010;69(6):1227–31. Peterson J, Boldogh I, Popov V, Saini S, Chopra A. Anti-inflammatory and antisecretory potential of histidine in Salmonella-challenged mouse small intestine. Lab Invest. 1998;78(5):523–34. Son DO, Satsu H, Shimizu M. Histidine inhibits oxidative stress- and TNF‐α‐induced interleukin‐8 secretion in intestinal epithelial cells. FEBS Lett. 2005;579(21):4671–7. Amores-Sánchez MI, Medina MÁ. Glutamine, as a Precursor of Glutathione, and Oxidative Stress. Mol Genet Metab. 1999;67(2):100–5. Nagashima M, Soejima Y, Saito K. Glutamine and exercise. JPFSM. 2013;2(4):469–73. Moinard C, Caldefie-Chezet F, Walrand S, Vasson M-P, Cynober L. Evidence that glutamine modulates respiratory burst in stressed rat polymorphonuclear cells through its metabolism into arginine. Br J Nutr. 2002;88(6):689–95. Yoneda J, Andou A, Takehana K. Regulatory Roles of Amino Acids in Immune Response. CRR. 2009;5(4):252–8. Gil CM, Drylewicz J, Smits H, Wolterbeek N, van Dijk M, Lafeber FP, et al. Identification of biomarkers in cerebrospinal fluid associated with OA pain. Osteoarthritis Cartilage. 2023;31:S380. Gokhale JA, Frenkel SR, Dicesare PE. Estrogen and osteoarthritis. Am J Orthop (Belle Mead NJ). 2004;33(2):71–80. Felson DT, Nevitt MC. The effects of estrogen on osteoarthritis. Curr Opin Rheumatol. 1998;10(3):269–72. Tanamas SK, Wijethilake P, Wluka AE, Davies-Tuck ML, Urquhart DM, Wang Y, et al. Sex hormones and structural changes in osteoarthritis: A systematic review. Maturitas. 2011;69(2):141–56. Bay-Jensen AC, Slagboom E, Chen-An P, Alexandersen P, Qvist P, Christiansen C, et al. Role of hormones in cartilage and joint metabolism. Menopause. 2013;20(5):578–86. Stubbs B, Aluko Y, Myint PK, Smith TO. Prevalence of depressive symptoms and anxiety in osteoarthritis: a systematic review and meta-analysis. Age Ageing. 2016;45(2):228–35. Fonseca-Rodrigues D, Rodrigues A, Martins T, Pinto J, Amorim D, Almeida A, et al. Correlation between pain severity and levels of anxiety and depression in osteoarthritis patients: a systematic review and meta-analysis. Rheumatology. 2021;61(1):53–75. Somers TJ, Keefe FJ, Godiwala N, Hoyler GH. Psychosocial factors and the pain experience of osteoarthritis patients: new findings and new directions. Curr Opin Rheumatol. 2009;21(5):501–6. Wang S-T, Ni G-X. Depression in osteoarthritis: current understanding. Neuropsychiatr Dis Treat. 2022;18:375–89. Additional Declarations No competing interests reported. Supplementary Files Dunstanetalsupplementarymethodsandfigures.pdf Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 11 Mar, 2026 Reviews received at journal 06 Mar, 2026 Reviewers agreed at journal 05 Mar, 2026 Reviewers agreed at journal 05 Mar, 2026 Reviews received at journal 26 Feb, 2026 Reviewers agreed at journal 17 Feb, 2026 Reviewers invited by journal 11 Feb, 2026 Editor assigned by journal 05 Feb, 2026 Submission checks completed at journal 04 Feb, 2026 First submitted to journal 30 Jan, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8742887","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":590410532,"identity":"994ca9e1-9c8d-48b8-9a2d-2f8994e43222","order_by":0,"name":"Isobel Kate Dunstan","email":"","orcid":"","institution":"University of Oxford","correspondingAuthor":false,"prefix":"","firstName":"Isobel","middleName":"Kate","lastName":"Dunstan","suffix":""},{"id":590410536,"identity":"83d75b89-14ae-4fa0-a0a2-791d71b821ef","order_by":1,"name":"Line J.H. Rasmussen","email":"","orcid":"","institution":"Copenhagen University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Line","middleName":"J.H.","lastName":"Rasmussen","suffix":""},{"id":590410538,"identity":"78d172a8-91f4-4a8c-83d8-b4998d5ac2c2","order_by":2,"name":"Fay Probert","email":"","orcid":"","institution":"University of Oxford","correspondingAuthor":false,"prefix":"","firstName":"Fay","middleName":"","lastName":"Probert","suffix":""},{"id":590410540,"identity":"966a5dbb-e6af-4bc2-a8d2-8b66eb571c26","order_by":3,"name":"Dorte A. Olsen","email":"","orcid":"","institution":"Lillebaelt Hospital","correspondingAuthor":false,"prefix":"","firstName":"Dorte","middleName":"A.","lastName":"Olsen","suffix":""},{"id":590410542,"identity":"018007ea-db93-4fd6-a42a-d4ed99ecbd6e","order_by":4,"name":"Jonna S. Madsen","email":"","orcid":"","institution":"Lillebaelt Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jonna","middleName":"S.","lastName":"Madsen","suffix":""},{"id":590410545,"identity":"90d3132f-21d2-4dcd-90c5-b73bd3c83a7c","order_by":5,"name":"Jesper Eugen-Olsen","email":"","orcid":"","institution":"Copenhagen University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jesper","middleName":"","lastName":"Eugen-Olsen","suffix":""},{"id":590410548,"identity":"a6e0700f-e761-4550-9117-442bf0ca697b","order_by":6,"name":"Thomas Peter Enggaard","email":"","orcid":"","institution":"Copenhagen University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"Peter","lastName":"Enggaard","suffix":""},{"id":590410554,"identity":"6ca602c4-f5b4-4d8e-9ac7-1a1a46607f20","order_by":7,"name":"Claus Varnum","email":"","orcid":"","institution":"Lillebaelt Hospital","correspondingAuthor":false,"prefix":"","firstName":"Claus","middleName":"","lastName":"Varnum","suffix":""},{"id":590410559,"identity":"dbc2c565-0f1b-4058-9da0-2265e3413bd0","order_by":8,"name":"Kate L. Lambertsen","email":"","orcid":"","institution":"Odense University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Kate","middleName":"L.","lastName":"Lambertsen","suffix":""},{"id":590410562,"identity":"af0b62b8-6955-495a-8403-91228cc8ddc1","order_by":9,"name":"Daniel C. Anthony","email":"data:image/png;base64,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","orcid":"","institution":"University of Oxford","correspondingAuthor":true,"prefix":"","firstName":"Daniel","middleName":"C.","lastName":"Anthony","suffix":""},{"id":590410566,"identity":"ad160f65-3acd-4a35-8d36-5b1959813051","order_by":10,"name":"Morten Rune Blichfeldt-Eckhardt","email":"","orcid":"","institution":"University of Southern Denmark","correspondingAuthor":false,"prefix":"","firstName":"Morten","middleName":"Rune","lastName":"Blichfeldt-Eckhardt","suffix":""}],"badges":[],"createdAt":"2026-01-30 15:53:05","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8742887/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8742887/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102828097,"identity":"db6fb076-25ab-4685-8b15-faf8703f4849","added_by":"auto","created_at":"2026-02-17 09:22:28","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":100613,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eMultivariate models of serum, CSF and combined biofluids distinguish osteoarthritis (OA) from pain-free controls. \u003c/strong\u003e\u003c/em\u003e\u003cem\u003e(A) Serum: (i) OPLS–DA scores plot of serum ¹H NMR data showing separation of OA (turquoise) from controls (black). (ii) Variable importance in projection (VIP) plot ranking discriminatory serum metabolites, with chylomicron/VLDL, lysine, 3-hydroxybutyrate, alanine and glutamine among the highest contributors. (iii) Accuracy, (iv) sensitivity and (v) specificity of the serum model compared with the permuted null model, showing Acc 87%, Sens 90% and Spec 86% (***p\u0026lt;0.001). (B) CSF: (i) OPLS–DA scores plot of CSF metabolites showing comparable separation. (ii) VIP plot indicating lactate, glutamine, glutamate and glucose as the main discriminatory CSF variables. (iii–v) Model metrics for CSF, giving Acc 89%, Sens 94% and Spec 85%, all superior to the null model (***p\u0026lt;0.001). (C) Serum + CSF + inflammatory markers: (i) OPLS–DA scores plot for the integrated dataset. (ii) VIP plot showing serum histidine and serum glutamine, together with CSF CRP and CSF GFAP, as the dominant variables when compartments are combined. (iii–v) Performance of the integrated model, yielding Acc 90%, Sens 90% and Spec 89%, again exceeding the null distribution (***p\u0026lt;0.001), indicating additive value of combining peripheral, central and inflammatory readouts. All models were built on z-scaled data and evaluated with repeated cross validation. Supplementary Figure S1 shows a concordant random forest analysis, and Supplementary Figure S2 presents box plots for the top serum and CSF VIPs from panels A and B.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8742887/v1/a8daf5b9abd318341d6d5181.png"},{"id":102828094,"identity":"fead0d56-0c29-4710-a35c-1cfd5fd73195","added_by":"auto","created_at":"2026-02-17 09:22:28","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":139232,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eAge and BMI matching confirm that key serum and CSF biomarkers remain altered in osteoarthritis.\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e (A) Matched multivariate model. (i) OPLS–DA scores plot for the age- and BMI-matched subset (OA n=31, controls n=28) showing that OA (turquoise) remains separable from controls (black), with an accuracy of 77%, indicating that group discrimination is not driven solely by demographic differences. (ii) Characteristics of the matched subset showing no significant differences in age or BMI between groups, with a modest imbalance in sex, which notably was not present in the full cohort. (iii) VIP plot from the matched model identifying serum histidine as the highest-ranking variable, followed by serum gp130, serum albumin (lysyl), CSF glucose and serum glutamine. (iv) ROC curve derived from refitting the matched-surviving variables in the full cohort, yielding an AUC of 0.906 (95% CI 0.858–0.953), which shows that a reduced, demographically robust panel is sufficient to distinguish OA from controls. (B) Distribution of matched-surviving variables in the full cohort. Box plots show lower concentrations of serum histidine (i), serum albumin (lysyl) (ii), serum glutamine (iii), serum lysine (iv) and serum citrate (v) in OA compared with controls, together with higher serum gp130 (vi), higher CSF acetate (vii) and higher CSF CRP (viii) in OA. These variables were selected because they were discriminatory in both the original and the matched models, so are unlikely to be confounded by age or BMI. Asterisks denote statistical significance: *p\u0026lt;0.05, **p\u0026lt;0.01, ***p\u0026lt;0.001.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8742887/v1/a643fab6b5db0f5212234f8d.png"},{"id":102828091,"identity":"d4cfb9d6-4cbb-465f-ba4a-afa57bab9dc6","added_by":"auto","created_at":"2026-02-17 09:22:28","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":133590,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eCorrelations between top discriminating metabolites and clinical outcomes in osteoarthritis.\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e Pearson correlation coefficients are shown for the serum and CSF variables that remained discriminatory after age/BMI matching (serum histidine, serum albumin (lysyl), serum glutamine, serum lysine, serum citrate, serum gp130, CSF acetate, CSF CRP, CSF glutamate and CSF glutamine) and for clinical and symptom scores. Circle colour and size indicate the direction and strength of the correlation. Correlations shown above the diagonal remain significant after false discovery rate correction. The horizontal and vertical dashed lines separate biochemical variables from clinical variables. Lower serum histidine, albumin and glutamine were associated with higher symptom burden, including SIQR, ISI and total NRS, while higher CSF CRP, CSF acetate and CSF glutamate tracked with worse clinical scores. Clinical measures: SIQR, Symptom Impact Questionnaire Revised; MDI, Major Depression Inventory; GAD7, Generalised Anxiety Disorder 7 item scale; ISI, Insomnia Severity Index; WPI, Widespread Pain Index; Total NRS, composite numeric rating scale totalled from average pain score, pain during exercise, pain at rest, and strongest pain; Oxford Score, Oxford hip/knee score; EQ VAS, EuroQol visual analogue scale.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8742887/v1/478ebe414e58d8873f7ee5bb.png"},{"id":102963087,"identity":"1caf6895-2f21-4e46-99b5-a9f7b04e9a17","added_by":"auto","created_at":"2026-02-19 04:13:22","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":120738,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eSerum metabolites distinguish between high and low pain intensity within osteoarthritis.\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e (A) Definition of pain groups. OA participants were stratified into low and high pain according to the lower and upper quartiles of the total Numeric Rating Scale (NRS), with higher values indicating greater pain intensity. (B) Metabolites associated with pain. Box plots show that serum albumin (lysyl) (i), serum histidine (ii), and serum lysine (iv) and serum glutamine (v) were significantly lower in the high pain group, whereas the chylomicron/VLDL signal (iii) and serum GFAP (vi) were higher. (C) Discriminative performance. A logistic model built from the pain sensitive metabolites in panel B separated high from low pain OA with an AUC of 0.824 (95% CI 0.663–0.984). (D) Group characteristics. Age, sex distribution, BMI and Kellgren–Lawrence (KL) grade did not differ significantly between high and low pain groups, indicating that the metabolic differences are related to pain state rather than structural severity. (E) Clinical relevance. Serum histidine remained significantly lower in participants with higher Pain Disability Index (PDI) (i) and SIQR scores (ii), and in those with lower EQ-VAS ratings (iii), confirming that this metabolite tracks broader symptom impact. Continuous variables were tested with unpaired t tests, categorical variables with Fisher’s exact test or the Monte Carlo extension for KL grade. Asterisks denote significance: *p\u0026lt;0.05, **p\u0026lt;0.01.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8742887/v1/96463124684eef839bb904e4.png"},{"id":102828096,"identity":"fc5c1fd3-c261-4843-9af3-201aecc683cd","added_by":"auto","created_at":"2026-02-17 09:22:28","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":101074,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eSex-specific differences in clinical and biofluid profiles in osteoarthritis\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e. (A) Female participants.\u003c/em\u003e\u003cem\u003e\u003cstrong\u003e \u003c/strong\u003e\u003c/em\u003e\u003cem\u003e(i) OPLS-DA scores plot of the integrated serum, CSF, and inflammatory dataset showing clear separation of osteoarthritis (OA, turquoise) from controls (black), with an accuracy of 88%. (ii) VIP plot identifying serum albumin (lysyl), serum histidine, and serum glutamine as the highest-ranking discriminatory variables in females. (B) Male participants. (i) OPLS-DA scores plot also separated OA from controls, although with lower accuracy (76%). (ii) VIP plot showing CSF CRP as the top-ranking variable in males. (C) Clinical outcomes by sex and diagnosis. Compared with sex-matched controls, women with OA reported higher Symptom Impact Questionnaire Revised (SIQR) scores (i), higher Insomnia Severity Index (ISI) scores (ii), and higher Major Depression Inventory (MDI) scores (iii). Two-way ANOVA showed a significant diagnosis × sex interaction for SIQR and a trend for ISI (interaction p values shown in pink). (C, iv–viii) Biofluid variables by sex and diagnosis. Women with OA showed larger reductions in serum albumin (lysyl) (iv), serum glutamine (v), and serum histidine (vi) than men. CSF glutamate was increased in females with OA only (vii), while CSF CRP was elevated in both men and women with OA (viii). Data are shown as box plots with overlaid points. Asterisks denote significance following two-way ANOVA with post hoc testing (*p\u0026lt;0.05, **p\u0026lt;0.01, ***p\u0026lt;0.001). Full two-way ANOVA results are provided in Supplementary Table S6.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-8742887/v1/505289f8e289b8b0e7ef71c4.png"},{"id":103056365,"identity":"b61f3c38-309b-4f63-8e1b-b9bacb5ef40f","added_by":"auto","created_at":"2026-02-20 09:08:15","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1847889,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8742887/v1/6b5bd04a-c56e-4be9-8b88-5aae2ae462b1.pdf"},{"id":102828095,"identity":"3389bf24-e748-4a80-b705-cfe5f1d974fd","added_by":"auto","created_at":"2026-02-17 09:22:28","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":931724,"visible":true,"origin":"","legend":"","description":"","filename":"Dunstanetalsupplementarymethodsandfigures.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8742887/v1/e336e23e74926a553bb32fa9.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Osteoarthritis pain inversely correlates with histidine and glutamine following CSF and serum profiling","fulltext":[{"header":"Background","content":"\u003cp\u003eOsteoarthritis (OA) is a complex joint disease characterized by cartilage degradation, synovial inflammation, and subchondral bone changes (\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e). It is the most prevalent joint disorder in older adults and a major cause of disability (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e). OA pathogenesis involves multiple factors, including mechanical influences, aging, and genetic predisposition (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). Synovial inflammation and subchondral bone alterations are increasingly recognized as key contributors to OA onset and progression (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). These changes can precede visible cartilage degeneration and are associated with pain and functional impairment (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). The subchondral bone undergoes uncoupled remodelling, leading to osteosclerosis in advanced stages (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). Current treatments aim to relieve symptoms and improve joint function, but no curative therapy exists (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Emerging strategies targeting cytokine inhibition and signalling pathways offer potential avenues for disease modification (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e), but personalised approaches to prevention and treatment, tailoring interventions to distinct OA phenotypes are likely to be necessary (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eBiomarkers have emerged as a critical focus in OA research (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). Recent studies have focused on the role of biomarkers in assessing OA pain severity and subtypes. While serum and synovial fluid biomarkers, such as extracellular matrix molecules, have been extensively studied, the potential of cerebrospinal fluid (CSF) biomarkers is an emerging area of interest (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). CSF biomarkers are increasingly recognized as valuable tools in osteoarthritis (OA) research, despite the primary pathology occurring in joints. CSF biomarkers offer unique insights into central nervous system involvement in OA pain. Indeed, studies have identified elevated levels of inflammatory proteins in CSF of OA patients, indicating neuroinflammation (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Interestingly, some CSF biomarkers, particularly those with neuroprotective effects, are associated with reduced pain intensity and milder symptoms (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Specific proteins like monocyte chemoattractant protein 1 (MCP1) and interleukin (IL)-8 have been found elevated in CSF, suggesting their role in neuroimmune signalling (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). However, hitherto there has been no investigation of the metabolome profile of CSF from individuals with OA.\u003c/p\u003e \u003cp\u003eSex differences are a recognised feature of OA, with women experiencing higher prevalence, pain intensity, and poorer clinical outcomes compared to men. Women account for 60% of OA cases globally (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e) and report more severe pain, functional limitations, and poorer surgical outcomes (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). These differences are influenced by hormonal factors, joint anatomy, central pain processing, and psychosocial aspects (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). The interplay between inflammatory markers and pain also varies by sex, suggesting distinct biological mechanisms underlying OA progression (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e). Addressing these disparities is crucial for tailoring interventions and improving treatment outcomes across sexes (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eMetabolomics provides a comprehensive view of metabolic changes in OA, revealing alterations in amino acid, lipid, and energy metabolism (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e). Nuclear magnetic resonance (NMR) spectroscopy of biofluids, particularly synovial fluid, has shown promise in identifying OA biomarkers (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). Here, we sought to test the hypothesis that osteoarthritis is accompanied by a measurable metabolic disturbance in both serum and CSF, and that this disturbance relates to the pain phenotype. To address this, we undertook a multiomics analysis integrating \u0026sup1;H NMR metabolomics and targeted protein measurements from paired blood and CSF samples. The study had three objectives: first, to compare the metabolic and inflammatory profiles of OA and pain-free controls in serum and CSF; second, within OA, to determine whether the key discriminating metabolites and proteins were associated with clinical measures of pain and disability; and third, to explore whether these relationships differed by sex, given the higher symptomatic burden observed in women.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eAim\u003c/h2\u003e \u003cp\u003eThe aim of this study was to determine whether osteoarthritis is associated with a metabolic and inflammatory signature in paired serum and cerebrospinal fluid, and whether key discriminating features relate to pain severity, symptom impact, and sex.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eParticipants\u003c/h3\u003e\n\u003cp\u003eParticipants aged 18\u0026ndash;80 years who were able to understand and respond to the questionnaires were recruited as part of The Danish Pain Research Biobank (DANPAIN Biobank), as previously described (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). For the current study, samples were included from two groups: individuals with osteoarthritic pain scheduled for hip or knee arthroplasty surgery at a regional hospital (n\u0026thinsp;=\u0026thinsp;81), and pain-free volunteers recruited locally (n\u0026thinsp;=\u0026thinsp;70). Detailed recruitment procedures and cohort characteristics have been described previously (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). The study was conducted in accordance with the Declaration of Helsinki, with written informed consent obtained from all participants, and approval granted by the relevant regional ethics committees, as described previously (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003eSample Collection\u003c/h3\u003e\n\u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eBlood (24 mL) and CSF (7 mL) samples were obtained following standard protocols. Serum was collected in plain tubes and stored at room temperature for 30 minutes before centrifugation. All samples were centrifuged at 2,000 G for 10 minutes at 4\u0026deg;C, aliquoted on ice blocks, and stored at \u0026minus;\u0026thinsp;80\u0026deg;C. CSF was collected via lumbar puncture using atraumatic needles and stored at \u0026minus;\u0026thinsp;80\u0026deg;C within 30 minutes of collection. Separate aliquots were analysed for neuroimmune biomarkers and for metabolomics to avoid freeze-thaw effects. CSF contaminated with blood was excluded, and samples were stored under identical conditions to blood.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eClinical and Pain Assessments\u003c/h3\u003e\n\u003cp\u003eFunctional and psychological impacts were assessed using validated instruments, including the Symptom Impact Questionnaire (SIQR) (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e), Major Depression Inventory (MDI), Generalized Anxiety Disorder (GAD-7), Insomnia Severity Index (ISI), EuroQol visual analogue scale (EQ VAS), and the Oxford Hip or Knee score (OHS/OKS) (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e). Pain intensity and interference were quantified using a Numeric Rating Scale (NRS) (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e) and Pain Disability Index (PDI) (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e), respectively. All clinical, pain, and questionnaire-based assessments were completed within 24 hours of blood and cerebrospinal fluid sampling, and in all cases prior to surgery for participants undergoing arthroplasty. Questionnaires were administered electronically using a standardised clinical registry platform, as previously described (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003eAnalysis of neuroimmune biomarkers\u003c/h3\u003e\n\u003cp\u003eA panel of biomarkers in CSF and blood was quantified using multiplex assays from Mesoscale Discovery as described elsewhere (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e). High-sensitivity C-reactive protein (hsCRP, mg/L) was analyzed with the CRP high sensitive ELISA (TECAN, IBL International GmbH, Hamburg, Germany) according to the manufacturer\u0026rsquo;s instructions. Serum suPAR (suPAR, ng/mL) was analyzed with the suPARnostic\u0026reg; AUTO Flex ELISA (ViroGates A/S, Birker\u0026oslash;d, Denmark) according to manufacturer\u0026rsquo;s instructions. Neurofilament light (NfL) and glial fibrillary acidic protein (GFAP) in CSF and serum were measured blinded to clinical data using a commercially available NfL and GFAP 2-plex assay on the Single Molecule Array (Simoa\u0026reg;) HD-X Analyzer (Quanterix, Billerica, MA, USA) (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e). Further details on all assays, including limits of detection and assay ranges are listed in the Supplementary methods S1.1.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eSample Processing and 1H NMR data collection\u003c/h2\u003e \u003cp\u003eSerum (150 \u0026micro;L) and CSF (100 \u0026micro;L) samples were mixed with 400 \u0026micro;L and 450 \u0026micro;L of NMR buffer (75 mM sodium phosphate buffer prepared in D\u003csub\u003e2\u003c/sub\u003eO, pH 7.4), respectively. All samples were then placed in a 5 mm NMR tube and measured using a 700-MHz Bruker AVII spectrometer operating at 16.4 T, equipped with a \u003csup\u003e1\u003c/sup\u003eH (\u003csup\u003e13\u003c/sup\u003eC/\u003csup\u003e15\u003c/sup\u003eN) TCI cryoprobe (Chemistry Research Laboratory, Department of Chemistry, University of Oxford), as described previously (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e). Further details can be found in Supplementary methods S1.2.\u003c/p\u003e \u003cp\u003e \u003cb\u003e1H NMR Data Processing\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTopspin 4.1.4 (Bruker) was used to phase, baseline correct, and reference all spectra to the lactate methyl resonance at δ\u0026thinsp;=\u0026thinsp;1.33 ppm. ACD/Labs Spectrus Processor Academic Edition 12.01 (Advanced Chemistry Development, Inc.) was used to manually bin each resonance signal, excluding all noise from analysis. The integrals of these bins were sum normalised and exported to R version 4.3.0 (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e). Metabolite assignment was conducted through a mix of in-house databases, literature reviews, 2D total correlation spectroscopy (TOCSY) experiments, and spiking reference compounds where needed (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eMultivariate Statistical Analysis\u003c/h3\u003e\n\u003cp\u003eMultivariate analyses were performed to identify key metabolite differences between groups, with rigorous cross-validation to ensure robustness. Specifically, the ropls package in R (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e) was used to conduct orthogonal partial least squares discriminant analysis (OPLS-DA) between pain-free and OA predictor variables (i.e. all serum and CSF metabolites and proteins from the targeted panels, z-scaled). Model performance was evaluated against a null distribution generated by random permutation of class labels to assess the likelihood of spurious separation. Descriptions of the model building and cross validation scheme can be found in Supplementary methods S1.3.\u003c/p\u003e\n\u003ch3\u003eUnivariate Statistical Analysis\u003c/h3\u003e\n\u003cp\u003eUnivariate analyses such as Student\u0026rsquo;s T test and two-way ANOVA were carried out in R (v 4.3.0). The significance threshold was set to \u0026lowast;p\u0026thinsp;\u0026lt;\u0026thinsp;0.05, \u0026lowast;\u0026lowast;p\u0026thinsp;\u0026lt;\u0026thinsp;0.01, \u0026lowast;\u0026lowast;\u0026lowast;p\u0026thinsp;\u0026lt;\u0026thinsp;0.001, \u0026lowast;\u0026lowast;\u0026lowast;\u0026lowast;p\u0026thinsp;\u0026lt;\u0026thinsp;0.0001. All correlations were tested via Pearson correlational analyses, using the Benjamini-Hochberg method to correct for false discovery rate at 0.05. Univariate Receiver Operating Characteristic (ROC) analyses and multivariate ROC analyses on a combination of features using logistic regression were determined using the pROC package in R.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e \u003cb\u003eDistinct serum and CSF metabolite profiles, and their integration with inflammatory markers, differentiate osteoarthritis from controls.\u003c/b\u003e \u003c/p\u003e \u003cp\u003eWe first examined whether untargeted \u0026sup1;H NMR spectra from single biofluids could discriminate individuals with OA from pain-free controls. OPLS-DA models built on serum profiles showed a clear separation between groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA(i)), with a classification accuracy of 87%, sensitivity of 90% and specificity of 86%, all significantly better than the null permutation model (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA(iii\u0026ndash;v)). The principal discriminatory serum signals were increased mobile, lipid-related resonances in the chylomicron/VLDL region, together with 3-hydroxybutyrate, and lower concentrations of amino acids including lysine, alanine and glutamine (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA(ii) and Supplementary Figure S2A(i-iv) for boxplots). CSF profiles gave a very similar picture, again separating OA from controls (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB(i)) and doing so with slightly higher diagnostic performance (accuracy 89%, sensitivity 94%, specificity 85%, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB(iii\u0026ndash;v)). Here, the highest ranking variables were lactate, glutamine, glutamate and glucose (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB(ii) and Supplementary Figure S2B(i-iv) for boxplots), which indicates that metabolic disturbance is present in the central compartment as well as in the periphery. When serum and CSF metabolites were analysed together with inflammatory and neurodegeneration-related proteins, model performance improved further, with the integrated model achieving 90% accuracy, 90% sensitivity and 89% specificity (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC(iii\u0026ndash;v)), driven chiefly by serum histidine and serum glutamine together with CSF CRP and CSF GFAP (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC(ii)). A random forest approach confirmed that OA and controls could also be separated using an orthogonal modelling strategy (Supplementary Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e), and the distribution of the top serum and CSF VIPs is shown as box plots in Supplementary Figure S2. These plots illustrate that, in serum, OA was characterised by higher very low density lipoprotein signals (mobile CH₃ in chylomicron/VLDL) and 3-hydroxybutyrate, together with lower lysine, alanine and glutamine. In CSF, lactate and glutamate were increased, while glutamine and glucose were reduced in OA compared with controls. These findings suggest distinct metabolic alterations in both systemic and, more surprisingly, central compartments in OA.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eAge- and BMI-matched analysis confirms a core OA metabolic-inflammatory signature independent of demographic differences.\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTo exclude the possibility that the discrimination observed in the full cohort was driven by age and BMI, we rebuilt the model in an age and BMI matched subset (OA n\u0026thinsp;=\u0026thinsp;31, controls n\u0026thinsp;=\u0026thinsp;28). In this restricted sample, age and BMI no longer differed significantly between groups and only a modest imbalance in sex remained (which was not present in the full cohort, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA(ii)), yet the OPLS-DA model still separated OA from controls with an accuracy of 77% (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA(i)), significantly better than the null model (Supplementary Figure S3). The VIP ranking from this matched model again placed serum histidine at the top, followed by serum gp130, serum albumin (lysyl moiety), CSF glucose and serum glutamine (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA(iii)), several of which were also among the highest ranking variables in the full integrated model (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC), indicating that these signals are not explained by demographic confounders. We then took only those variables that remained significantly different in the small matched cohort and refitted a logistic model in the full dataset. This reduced panel still achieved a strong diagnostic performance, with an AUC of 0.906 (95% CI 0.858\u0026ndash;0.953; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA(iv)), showing that a small, demographically robust signature is sufficient to distinguish OA from pain-free individuals. The direction and magnitude of these key variables in the whole cohort are illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB, where serum histidine, serum albumin (lysyl), serum glutamine and serum lysine are consistently lower in OA, while serum citrate, serum gp130, CSF acetate and CSF CRP are higher. Corresponding box plots for the matched subset are provided in Supplementary Figure S4 and a side by side comparison of VIP scores in the full and matched models is shown in Supplementary Figure S3B.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eCore metabolic signature associates with pain burden, symptom impact and sleep disturbance.\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTo determine whether the demographically robust metabolic and inflammatory variables were clinically meaningful, we examined their correlations with pain and disability measures within the whole cohort. The restricted panel of serum metabolites (histidine, albumin lysyl resonance, glutamine, lysine and citrate) was strongly intercorrelated, forming a coherent serum metabolic module, and this module correlated inversely with several symptom scores, including the SIQR, total NRS, ISI and WPI (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). In other words, lower serum histidine, albumin and glutamine were associated with greater symptom impact, higher pain ratings and poorer sleep. In contrast, central and inflammatory markers that survived matching, in particular CSF CRP, CSF acetate and CSF glutamate, showed positive correlations with these same clinical measures, consistent with a contribution of central immune activation to symptom severity. Anxiety and mood scores (GAD7, MDI) followed the same pattern, although the associations were weaker. A broader correlation matrix including the full inflammatory panel reproduced these relationships and showed additional links to CSF TNFRI/II and GFAP (Supplementary Figure S7), supporting the view that OA pain is accompanied by a coupled metabolic inflammatory disturbance across serum and CSF.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eSerum metabolites track intra-disease variation in pain intensity.\u003c/b\u003e \u003c/p\u003e \u003cp\u003eWe next asked whether the metabolic signature identified above also captured symptom severity within osteoarthritis itself. OA participants were stratified into low and high pain groups using the lower and upper quartiles of the total NRS (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). Despite similar age, sex distribution, BMI and Kellgren\u0026ndash;Lawrence scores between these two groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD), several of the key serum variables showed clear, directional differences. Serum albumin (lysyl resonance), histidine, lysine and glutamine were all lower in the high pain group, whereas the chylomicron/VLDL signal and serum GFAP, likely derived from chondrocytes (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e), were higher (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB(i\u0026ndash;vi)). Since structural burden and demographic factors did not differ, these findings indicate that this subset of metabolites reflects the pain phenotype rather than radiographic severity. A logistic model built only from the pain sensitive variables discriminated high from low pain OA with an AUC of 0.824 (95% CI 0.663 to 0.984; Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC), suggesting that a small serum panel may have utility for phenotyping painful OA. Consistent with this, serum histidine also associated with broader measures of impact, showing lower values in participants with higher Pain Disability Index and SIQR scores and with poorer EQ-VAS ratings (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE). Together, these data show that the metabolic disturbance identified at disease level is graded by pain intensity and is therefore clinically meaningful.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eFemale osteoarthritis shows a heavier symptom burden and a more pronounced metabolic\u0026ndash;inflammatory disturbance.\u003c/b\u003e \u003c/p\u003e \u003cp\u003eBecause women are disproportionately affected by symptomatic OA, we examined whether the discriminatory signature differed by sex (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e). When analysed separately, OPLS-DA models built from female participants showed robust separation of OA from controls (accuracy 88%), whereas the model in males, although still significant, was weaker (accuracy 76%) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA,B; Supplementary Figure S5). The top VIPs in women were the same amino acid related serum variables that survived age/BMI matching, in particular serum albumin (lysyl), serum histidine and serum glutamine, whereas in men CSF CRP rose to the top of the ranking (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA(ii), 5B(ii)). Clinically, women with OA reported higher SIQR, ISI and, to a lesser extent, MDI scores than men, and for SIQR and ISI there were significant diagnosis by sex interactions, indicating that the disease effect is larger in females (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC(i\u0026ndash;iii); Supplementary Table S6). Consistent with this, the metabolite changes that defined OA in the whole cohort, notably lower serum glutamine, histidine and albumin (lysyl), were clearly expressed in women but were weaker or absent in men (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC(iv\u0026ndash;vi)). Interestingly, CSF glutamate was elevated in females with OA only, potentially related to their higher pain and symptom scores (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC(vii)). While CSF CRP was the top driver of the male OPLS-DA model, two-way ANOVA revealed it was significantly increased in both men and women with OA (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC(viii)). Taken together, these findings suggest that in women the OA phenotype is driven more by metabolic and excitatory changes that track symptom burden and men in this cohort display a similar, but less pronounced phenotype.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study asked whether OA has a reproducible biofluid signature, whether that signature relates to pain, and whether it differs by sex. Using untargeted \u0026sup1;H NMR in serum and CSF, supported by targeted inflammatory markers, we showed that OA can be distinguished from pain-free controls with high accuracy, and that this signal persists after age and BMI matching. The same small set of variables, notably lower serum histidine, also tracked clinical severity, since OA cases with the greatest pain and symptom impact had the most divergent metabolite values. Finally, these effects were most pronounced in women, indicating that combined serum\u0026ndash;CSF profiling can capture sex-specific pathways contributing to symptomatic OA.\u003c/p\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eMetabolic Alterations in OA\u003c/h2\u003e \u003cp\u003eOur results reveal distinct metabolic signatures in OA, characterised by increases in VLDL and 3-hydroxybutyrate and decreases in lysine, alanine, histidine, and glutamine in serum, along with elevated lactate and glutamate and reduced glutamine and glucose in CSF. These results are consistent with growing evidence suggesting that metabolic dysregulation, particularly in amino acid and lipid metabolism, is linked to inflammation and joint degradation in OA (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e). Chondrocyte metabolism is known to be affected by environmental stressors, which can lead to shifts between metabolic pathways and mitochondrial dysfunction (\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e). Furthermore, obesity and metabolic syndrome, often comorbid with OA, exacerbate these metabolic changes through complex interactions of biomechanical, inflammatory, and metabolic factors (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e). These findings support the emerging concept of OA as a metabolic disease, with lipids and adipokines playing crucial roles in cartilage degradation (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e). Similarly, reductions in amino acids such as lysine and glutamine may indicate their increased utilisation in pro-inflammatory and metabolic pathways (\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe differentiation between serum and CSF metabolite profiles provides further support that OA also involves systemic and central metabolic changes. Evidence suggests that elevated lactate and glutamate levels in CSF may reflect altered energy metabolism and excitotoxicity in OA, contributing to central pain sensitisation and neuroinflammatory responses. Studies have found increased levels of glutamate and other excitatory amino acids in synovial fluid of OA patients (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e), however, this is the first report of an imbalance in the CNS. Metabolic changes in synovial fibroblasts and fluid, particularly in the glutamine-glutamate pathway, have been associated with inflammatory responses in OA (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e). Exposure of rat spinal cord slices to CSF from OA patients has been shown to affect neuronal excitability, potentially altering nociceptive processing at the spinal level (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e). Elucidating the altered composition of OA CSF here suggests that glutamate signalling may contribute to peripheral nociceptive transduction and inflammation in OA (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e). Additionally, reactive astrocytes and their release of lactate have been implicated in chronic pain development and central sensitisation (\u003cspan additionalcitationids=\"CR49\" citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e). These findings underscore the importance of considering systemic and central contributions to OA pathology and the potential utility of targeting metabolic pathways in therapeutic interventions.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eDiagnostic Potential of Biomarker Integration\u003c/h2\u003e \u003cp\u003eCombining serum and CSF metabolites with a targeted protein panel improved the diagnostic accuracy of OA models, reaching 90%. This integration highlights the value of a multi-modal biomarker approach in capturing the complex and multifactorial nature of OA. Previous research has shown the limitations of single biomarkers in diagnosing OA due to the heterogeneity of the disease (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e). By incorporating markers of inflammation, neurodegeneration, and joint degradation, our model achieved a higher discriminatory power and highlighted unique relationships between metabolites and proteins in CSF and serum. The persistence of certain metabolites, such as reduced serum histidine and glutamine, in both the full and matched cohorts suggests their robustness as potential biomarkers of OA. Histidine is an essential amino acid with anti-inflammatory properties (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e). Indeed, we show here that serum histidine levels are inversely correlated with cytokines both in the serum and CSF (Figure S7). Others have also highlighted histidine as a potential metabolomic biomarker for knee OA in which the ratio of branched-chain amino acids to histidine seems to be important (\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003ePain-Associated Metabolic Changes\u003c/h2\u003e \u003cp\u003eOne of the key findings of this study is the association of specific serum metabolites with OA pain intensity. Decreases in serum histidine, glutamine, lysine, and albumin (lysyl moiety) were observed in participants with high pain intensity, alongside an increase in serum GFAP.\u003c/p\u003e \u003cp\u003eThe observed reduction in histidine and glutamine in high-pain groups may reflect their roles in modulating oxidative stress and inflammatory responses. Histidine has been shown to scavenge reactive oxygen species (ROS) and inhibit inflammatory pathways in various studies (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e). It can reduce fluid accumulation and protect intestinal tissue during inflammation (\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e), as well as inhibit IL-8 secretion in intestinal epithelial cells (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e). Glutamine serves as a substrate for glutathione synthesis, which is crucial for antioxidant activities and immune cell function (\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e, \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e). Both amino acids have demonstrated potential in modulating respiratory burst in neutrophils (\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e). Their immunomodulatory effects make them promising candidates for therapeutic interventions in various conditions, including inflammation, infection, and oxidative stress-related disorders (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e, \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e). The elevation of serum GFAP, potentially derived from chondrocytes, suggests a link between joint damage and systemic markers of neurodegeneration, further supporting the interconnectedness of peripheral and central mechanisms in OA pain (\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e). The correlation between these metabolic changes and clinical pain scores, such as the NRS and SIQR, provides a compelling case for the use of serum metabolites as objective measures of pain severity. This could address the current reliance on subjective pain assessments in clinical practice and trials, improving the precision of pain management strategies.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eSex Differences in OA\u003c/h2\u003e \u003cp\u003eFemale participants exhibited more severe OA symptoms and greater metabolic changes compared to men, consistent with broader findings of higher pain prevalence, worse functional outcomes, and distinct inflammatory profiles (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e). The greater reductions in serum histidine and glutamine, as well as the notably higher CSF glutamate observed in females may reflect sex-specific differences in inflammatory and metabolic processes. Oestrogen influences cartilage metabolism, immune responses, and pain perception, particularly postmenopause, exacerbating OA pathology (\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e). Systematic reviews suggest oestrogen replacement therapy may lower OA prevalence by modulating cartilage turnover (\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e, \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e). Furthermore, metabolic imbalances influenced by oestrogen contribute to OA progression (\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e). Psychosocial factors, including anxiety and depression \u0026ndash; affecting\u0026thinsp;~\u0026thinsp;20% of OA patients (\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e) \u0026ndash; correlate moderately with pain severity and are more common in women (\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e). While the causal link remains unclear, evidence suggests interventions such as cognitive-behavioural approaches may alleviate OA pain and disability (\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e, \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e). The pronounced metabolite changes in women highlight the need to consider sex differences in OA research and treatment. Tailored interventions that address the unique metabolic and inflammatory pathways in females could improve clinical outcomes and reduce the sex disparity in OA burden.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eImplications for Clinical Practice and Research\u003c/h2\u003e \u003cp\u003eThis study provides a foundation for the development of biomarker-based diagnostic and therapeutic strategies in OA. Multi-modal biomarker panels integrating serum and CSF metabolites with targeted proteins could improve diagnostic accuracy, stratify patients by disease severity, and inform personalised OA management. The identification of robust biomarkers, such as histidine and glutamine, highlights potential therapeutic targets that warrant further investigation. Interventions aimed at restoring these metabolic pathways could mitigate inflammation and pain, offering a new avenue for disease-modifying treatments. Additionally, the sex-specific differences observed in this study call for the inclusion of sex-stratified analyses in future OA research to ensure equitable and effective treatment development.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eLimitations and Future Directions\u003c/h2\u003e \u003cp\u003eWhile this study advances our understanding of OA biomarkers, it has several limitations. The cross sectional design precludes causal inference, and the sample size, particularly in the matched analyses, may limit generalisability. We were not able to perform an independent external validation, largely because major population resources such as UK Biobank do not include CSF, which makes replication of a paired serum-CSF analysis challenging, but at the same time underscores the distinctiveness of the present cohort. Longitudinal studies are needed to define the temporal relationship between these metabolic alterations and OA progression and to test whether the serum variables that track pain can predict future symptom flares. Experimental work to probe the functional roles of histidine, glutamine and the CSF inflammatory markers in OA pathophysiology would also be valuable. Finally, combining this biofluid panel with imaging measures and a broader inflammatory/metabolic annotation could improve both diagnostic performance and biological interpretability.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eIn summary, this study identifies distinct metabolic and inflammatory profiles in OA. Specifically, reduced serum histidine and glutamine emerged as key discriminators, correlating inversely with pain intensity and disability scores and should be further explored for their role in OA pathophysiology and pain modulation as well as possible targets for personalised treatment strategies.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eAUC, area under the curve\u003c/p\u003e\u003cp\u003eBMI, body mass index\u003c/p\u003e\u003cp\u003eCRP, C reactive protein\u003c/p\u003e\u003cp\u003eCSF, cerebrospinal fluid\u003c/p\u003e\u003cp\u003eDANPAIN, Danish Pain Research Biobank\u003c/p\u003e\u003cp\u003eD2O, deuterium oxide\u003c/p\u003e\u003cp\u003eEQ VAS, EuroQol visual analogue scale\u003c/p\u003e\u003cp\u003eGAD 7, Generalised Anxiety Disorder 7 item scale\u003c/p\u003e\u003cp\u003eGFAP, glial fibrillary acidic protein\u003c/p\u003e\u003cp\u003ehsCRP, high sensitivity C reactive protein\u003c/p\u003e\u003cp\u003eIL, interleukin\u003c/p\u003e\u003cp\u003eISI, Insomnia Severity Index\u003c/p\u003e\u003cp\u003eKL, Kellgren to Lawrence grade\u003c/p\u003e\u003cp\u003eMCP 1, monocyte chemoattractant protein 1\u003c/p\u003e\u003cp\u003eMDI, Major Depression Inventory\u003c/p\u003e\u003cp\u003eNfL, neurofilament light chain\u003c/p\u003e\u003cp\u003eNMR, nuclear magnetic resonance\u003c/p\u003e\u003cp\u003eNRS, numeric rating scale\u003c/p\u003e\u003cp\u003eOA, osteoarthritis\u003c/p\u003e\u003cp\u003eOHS, Oxford Hip Score\u003c/p\u003e\u003cp\u003eOKS, Oxford Knee Score\u003c/p\u003e\u003cp\u003eOPLS DA, orthogonal partial least squares discriminant analysis\u003c/p\u003e\u003cp\u003ePDI, Pain Disability Index\u003c/p\u003e\u003cp\u003eROC, receiver operating characteristic\u003c/p\u003e\u003cp\u003eSIQR, Symptom Impact Questionnaire Revised\u003c/p\u003e\u003cp\u003eSimoa, Single Molecule Array\u003c/p\u003e\u003cp\u003esuPAR, soluble urokinase plasminogen activator receptor\u003c/p\u003e\u003cp\u003eTNF, tumour necrosis factor\u003c/p\u003e\u003cp\u003eTNFR I, tumour necrosis factor receptor 1\u003c/p\u003e\u003cp\u003eTNFR II, tumour necrosis factor receptor 2\u003c/p\u003e\u003cp\u003eTOCSY, total correlation spectroscopy\u003c/p\u003e\u003cp\u003eVLDL, very low density lipoprotein\u003c/p\u003e"},{"header":"Declarations","content":" \u003cp\u003e \u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e \u003c/p\u003e\u003cp\u003e The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). Approval was obtained from the Regional Ethics Committee (references S-20160173 and S-20180003), and the study was duly notified to the Danish Data Protection Agency (reference 17/3391). All participants provided written informed consent prior to enrolment, including consent for sample collection, clinical data acquisition, and subsequent analysis for research purposes. No individual-identifying information is presented in this manuscript.\u003c/p\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eConsent for publication\u003c/strong\u003e \u003c/p\u003e\u003cp\u003eNot applicable.\u003c/p\u003e \u003cp\u003e\u003c/p\u003e\u003cp\u003e \u003c/p\u003e\u003ch2\u003eCompeting interests\u003c/h2\u003e \u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e \u003cp\u003e\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eSeveral private and public foundations funded the building of the DANPAIN Biobank but did not participate in any aspect of study design, data collection, data analysis, data interpretation, or writing of the manuscript. These funders include: The Danish Rheumatism Association (grants R155-A4866-B1363 and R175-A6088-B1363); Aase og Ejnar Danielsens Fond; Karen S. Jensens Legat; Oberstinde Kirsten Jensens la Cours Legat; Professor, Overlæge Sophus H. Johansens Fond of 23 August 1981; Foundation of the Danish Association for Anaesthesia and Intensive Care; Fonden til Lægevidenskabens Fremme; The Development Foundation, Lillebaelt Hospital; OUH Fund for Free Research; The Research Fund of the Danish Society of Anaesthesiology and Intensive Care Medicine.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eIsobel Kate Dunstan contributed to data curation, formal analysis, investigation, methodology, software, validation, visualisation, and writing – original draft. Line Jee Hartmann Rasmussen contributed to data curation, investigation, and writing – review \u0026amp; editing. Fay Probert contributed to software, methodology, supervision, and writing – review \u0026amp; editing. Dorte Aalund Olsen contributed to data curation, investigation, and writing – review \u0026amp; editing. Jonna Skov Madsen contributed to data curation, investigation, and writing – review \u0026amp; editing. Jesper Eugen-Olsen contributed to data curation, investigation, and writing – review \u0026amp; editing. Thomas Peter Enggaard contributed to conceptualisation and writing – review \u0026amp; editing. Claus Varnum contributed to data curation and writing – review \u0026amp; editing. Kate Lykke Lambertsen contributed to conceptualisation, investigation, methodology, data interpretation, and writing – review \u0026amp; editing. Daniel C Anthony contributed to conceptualisation, formal analysis, supervision, and writing – review \u0026amp; editing. Morten Rune Blichfeldt-Eckhardt contributed to conceptualisation, data curation, data interpretation, and writing – review \u0026amp; editing.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eData storage and analysis were supported by the OPEN research network (Odense Patient data Explorative Network), Region of Southern Denmark. All metabolomic and clinical data processing was conducted securely using the OPEN Analyse environment, which provides controlled and auditable access to pseudonymised data in compliance with data protection legislation. We thank the OPEN team for their technical support and for providing this secure analysis infrastructure.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe datasets generated and analysed during the current study are not publicly available as they are subject to ethical and governance restrictions associated with the DANPAIN Biobank, but\u0026nbsp;are available from the corresponding author on reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eGoldring MB, Goldring SR. Osteoarthr J Cell Physiol. 2007;213(3):626\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim J-R, Yoo J, Kim H. Therapeutics in osteoarthritis based on an understanding of its molecular pathogenesis. Int J Mol Sci. 2018;19(3):674.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWei Y, Bai L. Recent advances in the understanding of molecular mechanisms of cartilage degeneration, synovitis and subchondral bone changes in osteoarthritis. Connect Tissue Res. 2016;57(4):245\u0026ndash;61.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi G, Yin J, Gao J, Cheng TS, Pavlos NJ, Zhang C, et al. Subchondral bone in osteoarthritis: insight into risk factors and microstructural changes. Arthritis Res Ther. 2013;15(6):223.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH\u0026uuml;gle T, Geurts J. What drives osteoarthritis?\u0026mdash;synovial versus subchondral bone pathology. Rheumatology. 2016;56(9):1461\u0026ndash;71.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMathiessen A, Conaghan PG. Synovitis in osteoarthritis: current understanding with therapeutic implications. Arthritis Res Ther. 2017;19(1):18.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAltman RD. Overview of osteoarthritis. Am J Med. 1987;83(4):65\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMalemud CJ, Islam N, Haqqi TM. Pathophysiological mechanisms in osteoarthritis lead to novel therapeutic strategies. Cells Tissues Organs. 2003;174(1\u0026ndash;2):34\u0026ndash;48.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMobasheri A, Batt M. An update on the pathophysiology of osteoarthritis. Ann Phys Rehabil Med. 2016;59(5\u0026ndash;6):333\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAttur M, Krasnokutsky-Samuels S, Samuels J, Abramson SB. Prognostic biomarkers in osteoarthritis. Curr Opin Rheumatol. 2013;25(1):136\u0026ndash;44.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKalogera S, Jansen MP, Bay-Jensen A-C, Frederiksen P, Karsdal MA, Thudium CS, et al. Relevance of biomarkers in serum vs. synovial fluid in patients with knee osteoarthritis. Int J Mol Sci. 2023;24(11):9483.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMobasheri A. Osteoarthritis year 2012 in review: biomarkers. Osteoarthritis Cartilage. 2012;20(12):1451\u0026ndash;64.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKosek E, Finn A, Ultenius C, Hugo A, Svensson C, Ahmed AS. Differences in neuroimmune signalling between male and female patients suffering from knee osteoarthritis. J Neuroimmunol. 2018;321:49\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePalada V, Ahmed AS, Freyhult E, Hugo A, Kultima K, Svensson CI, et al. Elevated inflammatory proteins in cerebrospinal fluid from patients with painful knee osteoarthritis are associated with reduced symptom severity. J Neuroimmunol. 2020;349:577391.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSegal NA, Nilges JM, Oo WM. Sex differences in osteoarthritis prevalence, pain perception, physical function and therapeutics. Osteoarthritis Cartilage. 2024;32(9):1045\u0026ndash;53.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGlass N, Segal NA, Sluka KA, Torner JC, Nevitt MC, Felson DT, et al. Examining sex differences in knee pain: the Multicenter Osteoarthritis Study. Osteoarthritis Cartilage. 2014;22(8):1100\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePerruccio AV, Badley EM, Power JD, Canizares M, Kapoor M, Rockel J, et al. Sex differences in the relationship between individual systemic markers of inflammation and pain in knee osteoarthritis. Osteoarthr Cartil Open. 2019;1(1\u0026ndash;2):100004.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSluka KA, Berkley KJ, O\u0026rsquo;Connor MI, Nicolella DP, Enoka RM, Boyan BD, et al. Neural and psychosocial contributions to sex differences in knee osteoarthritic pain. Biol Sex Differ. 2012;3(1):26.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTschon M, Contartese D, Pagani S, Borsari V, Fini M. Gender and sex are key determinants in osteoarthritis not only confounding variables. A systematic review of clinical data. JCM. 2021;10(14):3178.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eValdrighi N, Vago JP, Blom AB, Loo, FAJvd. Blaney Davidson EN. Innate immunity at the core of sex differences in osteoarthritic pain? Front Pharmacol. 2022;13:881500.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAdams SB, Setton LA, Nettles DL. The role of metabolomics in osteoarthritis research. J Am Acad Orthop Surg. 2012;21(1):63\u0026ndash;4.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhai G, Randell EW, Rahman P. Metabolomics of osteoarthritis: emerging novel markers and their potential clinical utility. Rheumatology. 2018;57(12):2087\u0026ndash;95.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eClarke EJ, Anderson JR, Peffers MJ. Nuclear magnetic resonance spectroscopy of biofluids for osteoarthritis. Br Med Bull. 2020;137(1):28\u0026ndash;41.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlichfeldt-Eckhardt MR, Mortensen WCP, Varnum C, Bendix L, Lauridsen JT, Jensen HI, et al. The Danish Pain Research Biobank (DANPAIN-Biobank): a collection of blood, cerebrospinal fluid, and clinical data for the study of neuroimmune and glia-related biomarkers of chronic pain. Ann Transl Med. 2023;11(10):343.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFriend R, Bennett RM. Evaluating disease severity in chronic pain patients with and without fibromyalgia: a comparison of the Symptom Impact Questionnaire and the Polysymptomatic Distress Scale. J Rheumatol. 2015;42(12):2404\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHarris KK, Dawson J, Jones LD, Beard DJ, Price AJ. Extending the use of PROMs in the NHS-using the Oxford Knee Score in patients undergoing non-operative management for knee osteoarthritis: a validation study. BMJ Open. 2013;3(8):e003365.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOrnetti P, Dougados M, Paternotte S, Logeart I, Gossec L. Validation of a numerical rating scale to assess functional impairment in hip and knee osteoarthritis: comparison with the WOMAC function scale. Ann Rheum Dis. 2011;70(5):740\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTait RC, Pollard CA, Margolis RB, Duckro PN, Krause SJ. The Pain Disability Index: psychometric and validity data. Arch Phys Med Rehabil. 1987;68(7):438\u0026ndash;41.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMartin NA, Bonner H, Elkjaer ML, D'Orsi B, Chen G, Konig HG, et al. BID mediates oxygen-glucose deprivation-induced neuronal injury in organotypic hippocampal slice cultures and modulates tissue inflammation in a transient focal cerebral ischemia model without changing lesion volume. Front Cell Neurosci. 2016;10:14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNielsen HH, Soares CB, Hogedal SS, Madsen JS, Hansen RB, Christensen AA, et al. Acute Neurofilament Light Chain plasma levels correlate with stroke severity and clinical outcome in ischemic stroke patients. Front Neurol. 2020;11:448.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChan SY, Probert F, Radford-Smith DE, Hebert JC, Claridge TDW, Anthony DC, et al. Post-inflammatory behavioural despair in male mice is associated with reduced cortical glutamate-glutamine ratios, and circulating lipid and energy metabolites. Sci Rep. 2020;10(1):16857.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eProbert F, Yeo T, Zhou Y, Sealey M, Arora S, Palace J, et al. Determination of CSF GFAP, CCN5, and vWF levels enhances the diagnostic accuracy of clinically defined MS from non-MS Patients with CSF oligoclonal bands. Front Immunol. 2021;12:811351.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR Core Team. R: a language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2024.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSnytnikova OA, Khlichkina AA, Sagdeev RZ, Tsentalovich YP. Evaluation of sample preparation protocols for quantitative NMR-based metabolomics. Metabolomics. 2019;15(6):84.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin HT, Cheng ML, Lo CJ, Hsu WC, Lin G, Liu FC. (1)H NMR metabolomic profiling of human cerebrospinal fluid in aging process. Am J Transl Res. 2021;13(11):12495\u0026ndash;508.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThevenot EA, Roux A, Xu Y, Ezan E, Junot C. Analysis of the human adult urinary metabolome variations with age, Body Mass Index, and gender by implementing a comprehensive workflow for univariate and OPLS statistical analyses. J Proteome Res. 2015;14(8):3322\u0026ndash;35.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKanazawa S, Nishizawa S, Takato T, Hoshi K. Biological roles of glial fibrillary acidic protein as a biomarker in cartilage regenerative medicine. J Cell Physiol. 2017;232(11):3182\u0026ndash;93.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGkretsi V, Simopoulou T, Tsezou A. Lipid metabolism and osteoarthritis: Lessons from atherosclerosis. Prog Lipid Res. 2011;50(2):133\u0026ndash;40.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng L, Zhang Z, Sheng P, Mobasheri A. The role of metabolism in chondrocyte dysfunction and the progression of osteoarthritis. Ageing Res Rev. 2021;66:101249.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJune RK, Liu-Bryan R, Long F, Griffin TM. Emerging role of metabolic signaling in synovial joint remodeling and osteoarthritis. J Orthop Res. 2016;34(12):2048\u0026ndash;58.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMasuko K, Murata M, Suematsu N, Okamoto K, Yudoh K, Nakamura H, et al. A metabolic aspect of osteoarthritis: lipid as a possible contributor to the pathogenesis of cartilage degradation. Clin Exp Rheumatol. 2009;27(2):347\u0026ndash;53.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu X, Fan X, Crawford R, Xiao Y, Prasadam I. The metabolic landscape in osteoarthritis. Aging Dis. 2022;13(4):1166.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDamyanovich A, Staples J, Marshall K. 1H NMR investigation of changes in the metabolic profile of synovial fluid in bilateral canine osteoarthritis with unilateral joint denervation. Osteoarthritis Cartilage. 1999;7(2):165\u0026ndash;72.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcNearney T, Speegle D, Lawand N, Lisse J, Westlund K. Excitatory amino acid profiles of synovial fluid from patients with arthritis. J Rheumatol. 2000;27(3):739\u0026ndash;43.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFarah H, Wijesinghe SN, Nicholson T, Alnajjar F, Certo M, Alghamdi A, et al. Differential Metabotypes in Synovial Fibroblasts and Synovial Fluid in Hip Osteoarthritis Patients Support Inflammatory Responses. Int J Mol Sci. 2022;23(6):3266.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu V, Peter AS. The excitability of dorsal horn neurons is affected by cerebrospinal fluid from humans with osteoarthritis. Can J Physiol Pharmacol. 2012;90(6):783\u0026ndash;90.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWen Z, Chang Y-C, Jean Y. Excitatory amino acid glutamate: role in peripheral nociceptive transduction and inflammation in experimental and clinical osteoarthritis. Osteoarthritis Cartilage. 2015;23(11):2009\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLluch E, Torres R, Nijs J, Oosterwijck J. Evidence for central sensitization in patients with osteoarthritis pain: A systematic literature review. Eur J Pain. 2014;18(10):1367\u0026ndash;75.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJames T, Mercedes B. Reactive astrocytes: critical players in the development of chronic pain. Front Psychiatry. 2021;12:682056.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu-Bryan R. Inflammation and intracellular metabolism: new targets in OA. Osteoarthritis Cartilage. 2015;23(11):1835\u0026ndash;42.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLotz M, Martel-Pelletier J, Christiansen C, Brandi ML, Bruy\u0026egrave;re O, Chapurlat R, et al. Value of biomarkers in osteoarthritis: current status and perspectives. Ann Rheum Dis. 2013;72(11):1756\u0026ndash;63.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHoleček M. Histidine in Health and Disease: Metabolism, Physiological Importance, and Use as a Supplement. Nutrients. 2020;12(3):848.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhai G, Wang-Sattler R, Hart DJ, Arden NK, Hakim AJ, Illig T, et al. Serum branched-chain amino acid to histidine ratio: a novel metabolomic biomarker of knee osteoarthritis. Ann Rheum Dis. 2010;69(6):1227\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePeterson J, Boldogh I, Popov V, Saini S, Chopra A. Anti-inflammatory and antisecretory potential of histidine in Salmonella-challenged mouse small intestine. Lab Invest. 1998;78(5):523\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSon DO, Satsu H, Shimizu M. Histidine inhibits oxidative stress- and TNF‐α‐induced interleukin‐8 secretion in intestinal epithelial cells. FEBS Lett. 2005;579(21):4671\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmores-S\u0026aacute;nchez MI, Medina M\u0026Aacute;. Glutamine, as a Precursor of Glutathione, and Oxidative Stress. Mol Genet Metab. 1999;67(2):100\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNagashima M, Soejima Y, Saito K. Glutamine and exercise. JPFSM. 2013;2(4):469\u0026ndash;73.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMoinard C, Caldefie-Chezet F, Walrand S, Vasson M-P, Cynober L. Evidence that glutamine modulates respiratory burst in stressed rat polymorphonuclear cells through its metabolism into arginine. Br J Nutr. 2002;88(6):689\u0026ndash;95.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoneda J, Andou A, Takehana K. Regulatory Roles of Amino Acids in Immune Response. CRR. 2009;5(4):252\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGil CM, Drylewicz J, Smits H, Wolterbeek N, van Dijk M, Lafeber FP, et al. Identification of biomarkers in cerebrospinal fluid associated with OA pain. Osteoarthritis Cartilage. 2023;31:S380.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGokhale JA, Frenkel SR, Dicesare PE. Estrogen and osteoarthritis. Am J Orthop (Belle Mead NJ). 2004;33(2):71\u0026ndash;80.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFelson DT, Nevitt MC. The effects of estrogen on osteoarthritis. Curr Opin Rheumatol. 1998;10(3):269\u0026ndash;72.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTanamas SK, Wijethilake P, Wluka AE, Davies-Tuck ML, Urquhart DM, Wang Y, et al. Sex hormones and structural changes in osteoarthritis: A systematic review. Maturitas. 2011;69(2):141\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBay-Jensen AC, Slagboom E, Chen-An P, Alexandersen P, Qvist P, Christiansen C, et al. Role of hormones in cartilage and joint metabolism. Menopause. 2013;20(5):578\u0026ndash;86.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStubbs B, Aluko Y, Myint PK, Smith TO. Prevalence of depressive symptoms and anxiety in osteoarthritis: a systematic review and meta-analysis. Age Ageing. 2016;45(2):228\u0026ndash;35.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFonseca-Rodrigues D, Rodrigues A, Martins T, Pinto J, Amorim D, Almeida A, et al. Correlation between pain severity and levels of anxiety and depression in osteoarthritis patients: a systematic review and meta-analysis. Rheumatology. 2021;61(1):53\u0026ndash;75.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSomers TJ, Keefe FJ, Godiwala N, Hoyler GH. Psychosocial factors and the pain experience of osteoarthritis patients: new findings and new directions. Curr Opin Rheumatol. 2009;21(5):501\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang S-T, Ni G-X. Depression in osteoarthritis: current understanding. Neuropsychiatr Dis Treat. 2022;18:375\u0026ndash;89.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"arthritis-research-and-therapy","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"arrt","sideBox":"Learn more about [Arthritis Research \u0026 Therapy](http://arthritis-research.biomedcentral.com/)","snPcode":"13075","submissionUrl":"https://submission.nature.com/new-submission/13075/3","title":"Arthritis Research \u0026 Therapy","twitterHandle":"@ArthritisRes","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Metabolomics, Osteoarthritis, Neuroinflammation, Chronic Pain Biomarkers, 1H NMR Spectroscopy, cytokines, CSF, CRP","lastPublishedDoi":"10.21203/rs.3.rs-8742887/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8742887/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eOsteoarthritis is a leading cause of pain and disability, yet the biological processes linking peripheral joint pathology with central pain mechanisms and wider symptom burden remain poorly defined.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eWe performed an integrated metabolomic and inflammatory analysis of cerebrospinal fluid and serum obtained from patients with osteoarthritis (n\u0026thinsp;=\u0026thinsp;81) and pain-free controls (n\u0026thinsp;=\u0026thinsp;70). Proton nuclear magnetic resonance spectroscopy was used for metabolomic profiling, alongside targeted protein assays for inflammatory mediators. Orthogonal partial least squares discriminant analysis was applied to assess separation between groups and to determine diagnostic accuracy. Associations between metabolites and clinical outcomes, including pain intensity, disability and sleep disturbance, were examined, with adjustment for age and BMI.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eClear separation between osteoarthritis and control participants was observed in both biofluids, with classification accuracies of 87% for serum and 89% for cerebrospinal fluid. Reduced serum histidine, glutamine, albumin (lysyl) and lysine were key discriminators of osteoarthritis, while elevated lactate and glutamate and reduced glucose and glutamine characterised the cerebrospinal fluid profile. Combining metabolomic data with inflammatory proteins increased diagnostic accuracy to 90% and remained significant after matching for age and BMI. Reductions in serum histidine and glutamine were consistent across subgroups, including stratification by pain severity. These metabolites correlated inversely with pain intensity, disability, sleep disturbance and overall symptom impact, and were more markedly altered in women, who also reported greater symptom burden.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eOsteoarthritis is associated with a distinct pattern of peripheral and central metabolic disturbance. Histidine and glutamine emerge as promising biomarkers related to pain and clinical severity, highlighting metabolic pathways as potential targets for improved stratification and intervention in osteoarthritis pain.\u003c/p\u003e","manuscriptTitle":"Osteoarthritis pain inversely correlates with histidine and glutamine following CSF and serum profiling","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-17 09:22:23","doi":"10.21203/rs.3.rs-8742887/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-11T14:40:19+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-06T16:19:44+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"333820917651961827527546850556391721007","date":"2026-03-05T15:41:46+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"175530665930549441212100029716887648790","date":"2026-03-05T15:18:02+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-27T01:16:42+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"96822537303690255106277748529884777799","date":"2026-02-17T20:09:06+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-11T07:52:12+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-05T07:20:04+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-05T01:42:04+00:00","index":"","fulltext":""},{"type":"submitted","content":"Arthritis Research \u0026 Therapy","date":"2026-01-30T15:21:26+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"arthritis-research-and-therapy","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"arrt","sideBox":"Learn more about [Arthritis Research \u0026 Therapy](http://arthritis-research.biomedcentral.com/)","snPcode":"13075","submissionUrl":"https://submission.nature.com/new-submission/13075/3","title":"Arthritis Research \u0026 Therapy","twitterHandle":"@ArthritisRes","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"69157162-8af2-4207-a769-47703d4d0460","owner":[],"postedDate":"February 17th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-21T10:53:21+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-17 09:22:23","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8742887","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8742887","identity":"rs-8742887","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}