Association of reduced Cu,Zn-superoxide dismutase with coronary microvascular dysfunction-An observational study of oxidative stress markers- Authors: Koichi Sato, Morihiko Takeda, Shohei Ikeda, Yoshihisa Kanaji, Masahiro Hada, Tsunekazu Kakuta, Takahiko Kiyooka, Yuichi Ozaki, Atsushi Tanaka, Yoshiaki Kawase, Hitoshi Matsuo, Yoshiaki Mitsutake, Yoshihiro Fukumoto, Satoshi Miyata, Hiroki Aoki, Hiroaki Shimokawa Published: 2026 ABSTRACT Highlights•Serum Cu,Zn-SOD level is reduced in patients with CMD.•Reduced Cu,Zn-SOD and Cav1 suggest impaired EDHF/H2O2 −mediated signaling in CMD.•Oxidative stress may be a key therapeutic target for CMD pathophysiology.Keywords: Coronary microvascular dysfunction, Cu.Zn-superoxide dismutase, Caveolin-1, Oxidative stress Introduction Angina with no obstructive coronary arteries (ANOCA) is increasingly recognized as a clinically important condition that affects a substantial proportion of patients presenting with symptoms of myocardial ischemia [1], [2]. Randomized trials including ORBITA [3] and ISCHEMIA [4] showed that percutaneous coronary intervention does not necessarily improve exercise capacity or long-term outcomes compared with optimal medical therapy, underscoring the importance of coronary microvascular dysfunction (CMD). Indeed, the Coronary Vasomotor Disorders International Study (COVADIS) group reported that patients with strictly diagnosed microvascular angina (MVA) experience a 7.7% annual rate of major adverse coronary events despite the absence of obstructive stenosis [5]. We further demonstrated that coronary spasm and CMD frequently coexist, conferring worse long-term outcomes [6]. Although traditionally considered a relatively benign entity, accumulating evidence indicates that patients with ANOCA experience impaired quality of life and are at increased risk of adverse cardiovascular outcomes [7]. CMD was reported to arise from multiple mechanisms, including vascular dysfunction, microvascular rarefaction, extravascular compression, and interstitial fibrosis [8], [9]. However, despite growing recognition of CMD, its underlying mechanisms remain incompletely understood. In particular, the role of systemic oxidative stress in the development of CMD has not been fully elucidated in clinical settings. Experimental studies have demonstrated that oxidative stress plays a critical role in endothelial dysfunction and microvascular regulation, including the modulation of endothelium-derived hyperpolarizing factor (EDHF)-mediated vasodilation [10], [11]. Among antioxidant systems, endothelial nitric oxide synthase (eNOS) generates both nitric oxide (NO) and superoxide (O2–), and Cu,Zn-superoxide dismutase (Cu,Zn-SOD) regulates redox balance by converting O2– into hydrogen peroxide (H2O2), which functions as EDHF, thereby preserving endothelial function [12], [13], [14]. In addition, Caveolin-1 (Cav1), a scaffolding and regulatory protein, is one of the important regulators of vascular homeostasis, suppressing eNOS, reducing NO and enhancing O2– production [15], [16]. Together, Cav1, eNOS, and Cu,Zn-SOD play crucial roles in microvascular regulation in experimental models (Fig. S1). However, the clinical relevance of these oxidative stress-related biomarkers in patients with ANOCA and CMD remains unclear. In particular, it is unknown whether circulating levels of Cu,Zn-SOD, Cav1, and eNOS are associated with distinct CMD phenotypes defined by invasive coronary function testing. Therefore, in the present multicenter observational study based on a nationwide registry, we aimed to investigate the association between circulating oxidative stress-related biomarkers and CMD in patients with ANOCA. By integrating systemic biomarker assessment with comprehensive invasive coronary phenotyping, we sought to provide novel insights into the pathophysiology of CMD in this clinically important population. Methods Japanese association of coronary microvascular dysfunction (J-CMD) The J-CMD study, established in 2022, is a nationwide, multi-center prospective cohort involving 30 hospitals in Japan (UMIN000047609). This study was conducted in accordance with the Declaration of Helsinki and approved by institutional review boards (No. 21-Nr-095). Informed consent was obtained from all participants. Study population We enrolled patients who had angina with non-obstructive coronary arteries (ANOCA), characterized by chest pain that limited daily activities, in whom coronary computed tomography or coronary angiography for suspected ischemic heart disease revealed no significant stenosis. Patients with clinically evident acute inflammatory or infectious conditions were excluded at enrollment (Supplement). All patients subsequently underwent interventional diagnostic procedures (IDP) to evaluate for vasospastic angina (VSA) and CMD, including assessment of coronary vasoreactivity testing (Coronary vasoreactivity testing refers to provocative test for coronary spasm with intracoronary injection of acetylcholine or ergonovine [17], [18].), coronary flow reserve (CFR), and index of microcirculatory resistance (IMR) [6] based on the guidelines by the Japanese Circulation Society [17]. Between August 2022 and December 2023, 295 ANOCA patients with serum samples were prospectively enrolled, irrespective of left ventricular ejection fraction, while those with FFR ≤ 0.80 or extensive old MI were excluded. Details of coronary vasoreactivity and coronary functional testing are provided in the Supplement. Measurement of serum concentrations of Cu,Zn-SOD, caveolin-1, eNOS and extracellular-SOD Venous blood samples were collected from all patients at the time of cardiac catheterization. After centrifugation, the samples were collected and stored at −20°C until biochemical analysis. All samples were transported in a frozen state to the Cardiovascular Research Institute of Kurume University (Kurume, Japan), where central biochemical analysis was performed. Serum concentrations of Cu,Zn-SOD, Cav1 and eNOS were determined using a double-antibody sandwich human enzyme-linked immunosorbent assay (ELISA) kit. Additionally, we also measured serum concentrations of extracellular-SOD (ecSOD), another type of SOD isoform [14] in 128 patients with leftover serum samples. Regarding the reproducibility of ELISA, triplicate measurements were performed for all samples across all kits to confirm reproducibility. The intra-assay coefficients of variation were below 10% across the measurement range for all kits. The validity of the ELISA method was confirmed based on the use of commercially available ELISA kits and the fact that the recombinant proteins provided with the kits were measured with the expected sensitivity. Regarding the dilution ratio, preliminary experiments were conducted to determine the appropriate sample dilution ratio that allows measurement within the linear range of the standard concentration-absorbance relationship. For a small number of samples that exceeded the detection limit, the dilution ratio was appropriately increased, and the samples were remeasured to ensure they fell within the measurable range. The product names of the ELISA kits used are listed in the Supplement. Statistical analysis Continuous variables are presented as means and standard deviations, or medians and interquartile ranges (IQRs). Categorical variables are presented as numerals and percentages. Comparisons between two groups were made with Mann-Whitney test for continuous variables and Pearson’s chi-square test for categorical variables. Comparisons among 3 or more groups were performed with Kruskal-Wallis test followed by Steel Dwass test for multiple comparison. Classification and regression tree (CART) analysis was performed to identify optimal cutoff values for serum Cu,Zn-SOD, Cav1, and eNOS. CART is a statistical method that recursively partitions data to identify variables and cutoff points that best discriminate the outcome. This approach enables data-driven determination of cutoff values for continuous variables, providing potentially more clinically relevant thresholds than arbitrary categorizations such as quartiles [19]. The CART model was pruned so that the maximum depth of the final tree was set to two to avoid overfitting. To examine the correlation between Cav1 and CFR/IMR, as well as Cu,Zn-SOD and CFR/IMR, the natural logarithm of each parameter was taken, and linear regression analysis was performed. Logistic regression analysis was used to analyze the predictor for the presence of CMD by stepwise backward elimination with the following covariates; age, male sex, hypertension, dyslipidemia, diabetes mellitus, NT-proBNP level, serum Cav1 concentration below the cut-off value (6.40 ng/ml), and serum Cu,Zn-SOD concentration below the cut-off value (97.2 ng/ml). Statistical analysis was performed with SPSS statistics 28.0.1.0 (IBM Corp, Armonk, NY, USA) and R version 3.1.1 (https://www.r-project.org, 21 September 2016). A P-value of < 0.05 was considered to be statistically significant. Results Clinical characteristics of ANOCA patients Patient characteristics are summarized in Table 1. A total of 295 ANOCA patients (134 men, 161 women; mean age 63.7 ± 13.9 years) were enrolled, of whom 127 (43%) were diagnosed with CMD. Chest pain at rest was the predominant symptom in both groups (73%). Compared to the non-CMD group, CMD patients were older and had a higher prevalence of hypertension, diabetes mellitus, former smoking, ischemic heart disease, and chronic kidney disease. The CMD group also had higher prescription rates of ACE inhibitors/ARBs and beta-blockers. Laboratory findings showed higher levels of LDL cholesterol, HbA1c, and NT-proBNP in the CMD group. Echocardiography revealed elevated E/e’ values in CMD patients, suggesting diastolic dysfunction. VSA was present in 53% of non-CMD and 47% of CMD patients, with 20% (n = 59) showing overlap of CMD and VSA. Fig. 1 shows patient distribution based on IMR and CFR. Patients were categorized into four groups: normal (IMR < 25, CFR > 2.0, n = 168); abnormal IMR (IMR > 25, CFR > 2.0, n = 61); abnormal CFR (IMR < 25, CFR < 2.0, n = 28); and abnormal IMR/CFR (IMR > 25, CFR < 2.0, n = 38). Table 1Baseline clinical characteristics of patients with ANOCA (n = 295).CharacteristicsNon-CMDCMDP-value(n = 168)(n = 127)Patient characteristicsAge (yrs)63(53,75)69(58,76)0.03Male sex, n (%)71(42%)63(50%)0.24BW (kg)60(53,72)61(54,71)0.52Height (cm)161(155,169)160(154,168)0.26BMI (kg/m2)24(20,26)24(22,27)0.17 Symptom, n (%)Rest angina122(73)93(73)0.91Effort angina14(8)12(9)0.74Rest/effort angina26(16)18(14)0.76Dyspnea6(4)4(3)0.84 Coronary risk factors, n (%)Hypertension83(49)89(70)<0.001Dyslipidemia82(49)68(54)0.42Diabetes mellitus31(19)36(28)0.045Current smoking22(13)17(13)0.94Former smoking43(26)46(36)0.049Never smoking103(61)64(50)0.06Family history of CAD28(17)19(15)0.69 Medical history, n (%)Ischemic heart disease20(12)27(21)0.03Chronic kidney disease20(12)28(22)0.02Hemodialysis3(1.8)4(3.1)0.45ASO0(0)1(0.8)0.25 EchocardiographyLVEF (%)64(60,69)65(60,70)0.18rEF patients (n, %)2(1.3)1(0.9)0.74LAD (mm)34(30,38)36(32,40)0.01E/A0.9(0.7,1.3)0.8(0.7,1.1)0.09E/e’9(7,11)10(8,12)0.004 Blood testHb (mg/dL)13.5(12.6,15.0)13.7(12.6,14.5)0.87Creatinine (mg/dL)0.7(0.6,0.9)0.8(0.6,0.9)0.83T-CHO (mg/dL)189(160,222)181(150,207)0.02HDL-C (mg/dL)61(50,77)62(50,75)0.97TG (mg/dL)114(74,165)118(79,180)0.40LDL-C (mg/dL)108(82,133)96(72,120)0.01HbA1c (%)5.7(5.5,6.1)5.9(5.7,6.4)0.01NT-proBNP (pg/ml)66(27,128)75(41,174)0.04CRP (mg/dl)0.1(0.0,0.1)0.1(0.0,0.13)0.46TnI/T qualitative test, positive, n (%)9(8)3(4)0.19Thyroid dysfunction, n (%)10(7)4(4)0.26 Catheter findingsFFR0.91(0.87,0.94)0.88(0.83,0.93)0.01Spasm provocation with ACh, n (%)159(95)120(95)0.95Spasm provocation with Erg, n (%)1(0.6)2(1.6)0.41Epicardial spasm, Diffuse, n (%)57(34)44(35)0.90Epicardial spasm, Focal, n (%)30(18)23(18)0.96 MedicationsCalium channel blockers123(73)90(71)0.66ACEI or ARB42(25)53(42)0.002Beta-blockers24(14)40(32)<0.001Nitrates32(19)27(21)0.64Statins76(45)66(52)0.25SGLT2 inhibitor15(8.9)12(9.4)0.88Results are expressed as mean ± standard deviation, median (inter-quartile range), or n (%). P values for comparisons between two groups were calculated using the Mann-Whitney test for continuous variables and Pearson’s chi-square test for categorical variables. ACh, acetylcholine; ACEI, angiotensin converting enzyme inhibitor; ARB, Angiotensin II Receptor Blocker; ASO, arteriosclerosis obliterans; BMI, body mass index; BW, body weight; CAD; coronary artery disease, CMD, coronary microvascular dysfunction; CRP, C-reactive protein; E/A, early diastolic filling velocity/atrial filling velocity; E/e′, early diastolic mitral flow velocity/tissue Doppler imaging velocity; Erg, ergonovine; FFR, fractional flow reserve; Hb, hemoglobin; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; LVEF, left ventricular ejection fraction; LAD; left atrial dimension; rEF; Reduced ejection fraction, SGLT2; Sodium-glucose transporter 2, T-CHO, total cholesterol; TG, triglyceride; TnI/T, troponin I/T. Baseline clinical characteristics of patients with ANOCA (n = 295). Results are expressed as mean ± standard deviation, median (inter-quartile range), or n (%). P values for comparisons between two groups were calculated using the Mann-Whitney test for continuous variables and Pearson’s chi-square test for categorical variables. ACh, acetylcholine; ACEI, angiotensin converting enzyme inhibitor; ARB, Angiotensin II Receptor Blocker; ASO, arteriosclerosis obliterans; BMI, body mass index; BW, body weight; CAD; coronary artery disease, CMD, coronary microvascular dysfunction; CRP, C-reactive protein; E/A, early diastolic filling velocity/atrial filling velocity; E/e′, early diastolic mitral flow velocity/tissue Doppler imaging velocity; Erg, ergonovine; FFR, fractional flow reserve; Hb, hemoglobin; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; LVEF, left ventricular ejection fraction; LAD; left atrial dimension; rEF; Reduced ejection fraction, SGLT2; Sodium-glucose transporter 2, T-CHO, total cholesterol; TG, triglyceride; TnI/T, troponin I/T. Distribution of patients according to CFR and IMR values. Among 295 patients, 168 (56.9%) had both normal for CFR (CFR ≥ 2.0) and IMR (IMR ≤ 25), 61(21%) abnormal IMR (IMR > 25) alone, 28(9%) abnormal CFR (CFR < 2.0) alone, and 38 (13%) both abnormal IMR (IMR > 25) and CFR (CFR < 2.0). Serum Cu,Zn-SOD, Cav1, and eNOS levels in patients with CMD Serum Cu,Zn-SOD levels were significantly lower in the CMD group than in the non-CMD group (P = 0.001, Fig. 2A), but did not differ between VSA and non-VSA groups (P = 0.49, Fig. 2B). Compared with the normal group, patients with CMD or VSA also showed significantly lower Cu,Zn-SOD levels (P = 0.002, Fig. 2C), suggesting reduced Cu,Zn-SOD levels in CMD regardless of epicardial spasm. Additionally, patients with abnormal IMR had significantly lower Cu,Zn-SOD levels than those with normal IMR (Fig. S2), while patients with abnormal CFR had significantly lower Cav1 levels (Fig. S3). When stratified into four groups based on CFR and IMR, Cu,Zn-SOD levels were significantly lower in the abnormal IMR group and the dual abnormal group, compared with the normal group (Fig. 2D). Cav1 levels were significantly lower only in the dual abnormal group. In contrast, no significant differences in eNOS or ecSOD levels were observed between CMD vs. non-CMD or VSA vs. non-VSA groups (Fig. 2, Figs. S4 and S5). Linear regression using log-transformed values revealed a significant negative correlation between Cu,Zn-SOD and IMR (P = 0.001), and a positive correlation between Cav1 and CFR (P = 0.005) (Fig. S6). −A. Serum concentrations of Cav1, eNOS, and Cu,Zn-SOD by the presence or absence of CMD. Results are expressed as box-and-whisker plots; the central box covers the interquartile range, with the median indicated by the line within the box. Analyses of Cu,Zn-SOD and Cav1 were performed in 295 cases, whereas eNOS was analyzed in the 188 cases for which measurements were available.P value indicates the difference between the 2 groups by Mann-Whitney test. 2-B. Serum concentrations of Cav1, eNOS, and Cu,Zn-SOD by the presence or absence of VSA. Results are expressed as box-and-whisker plots; the central box covers the interquartile range, with the median indicated by the line within the box. Analyses of Cu,Zn-SOD and Cav1 were performed in 295 cases, whereas eNOS was analyzed in the 188 cases for which measurements were available. P value indicates the difference between the 2 groups by Mann-Whitney test. C. Serum concentrations of Cav1, eNOS, and Cu,Zn-SOD by the presence or absence of VSA or CMD. Results are expressed as box-and-whisker plots; the central box covers the interquartile range, with the median indicated by the line within the box. Analyses of Cu,Zn-SOD and Cav1 were performed in 295 cases, whereas eNOS was analyzed in the 188 cases for which measurements were available. P value indicates the difference between the 2 groups by Mann-Whitney test. D. Serum concentrations of Cav1, eNOS, and Cu,Zn-SOD of the 4 groups by the CFR and IMR values. Results are expressed as box-and-whisker plots; the central box covers the interquartile range, with the median indicated by the line within the box. Analyses of Cu,Zn-SOD and Cav1 were performed in 295 cases, whereas eNOS was analyzed in the 188 cases for which measurements were available. P value indicates the difference among the 4 groups by Kruskal–Wallis test followed by Steel Dwass test for multiple comparison. Correlated factors for CMD in ANOCA CART analysis identified serum Cu,Zn-SOD < 97.2 ng/ml and Cav1 < 6.40 ng/ml as the first discriminating points for CMD. For eNOS, analysis in 188 cases with available measurements revealed a cutoff value of 0.158 ng/ml through three levels of branching (Fig. S7). In multivariable logistic regression, Cu,Zn-SOD below 97.2 ng/ml was an independent predictor of CMD [OR 2.71, 95% CI: 1.56–4.74; P < 0.001], along with hypertension and low Cav1 (<6.40 ng/ml) (Table 2). Even after adjustment for medications and CRP, which may influence oxidative stress, Cu,Zn-SOD remained an independent predictor of CMD (Table S1-5). The ROC curve for Cu,Zn-SOD alone showed modest predictive ability (AUC 0.611, Fig. S8A), while a 3-factor model incorporating hypertension, low Cav1, and low Cu,Zn-SOD demonstrated better discrimination (AUC 0.697, Fig. S8B). Table 2Factors correlated for CMD in ANOCA patients (n = 295).Univariable analysisMultivariable analysisOR95%CIP-valueHR95%CIP-valueAge1.021.00–1.040.02Male sex1.340.85–2.140.21Hypertension2.401.48–3.92<0.0012.611.48–4.660.001Dyslipidemia1.210.76–1.920.42Diabetes mellitus NT-proBNP1.75 1.001.01–3.041.00–1.000.0460.28Low Cu,Zn-SOD2.961.82–4.84<0.0012.711.56–4.74<0.001Low Cav12.111.23–3.630.0071.851.00–3.420.049The CART analysis was performed in order to identify the cut-off value of serum concentrations of Cu,Zn-SOD and Cav1. Low Cu,Zn-SOD = Cu,Zn-SOD < 97.2 ng/ml, Low Cav1= Cav1 < 6.40 ng/ml. Univariable and multivariable logistic regression analyses were performed to assess the association between each variable and the outcome, and P values were calculated. Factors correlated for CMD in ANOCA patients (n = 295). The CART analysis was performed in order to identify the cut-off value of serum concentrations of Cu,Zn-SOD and Cav1. Low Cu,Zn-SOD = Cu,Zn-SOD < 97.2 ng/ml, Low Cav1= Cav1 < 6.40 ng/ml. Univariable and multivariable logistic regression analyses were performed to assess the association between each variable and the outcome, and P values were calculated. Discussion The present study provides new mechanistic and clinical insights into the pathophysiology of CMD in patients with ANOCA. The principal findings are: (1) patients with CMD had significantly lower serum Cu,Zn-SOD levels compared with non-CMD patients, (2) within the CMD cohort, those with abnormal IMR exhibited the most pronounced reduction in Cu,Zn-SOD, (3) patients with both impaired CFR and elevated IMR had significantly Cav1 levels, and (4) multivariable analysis identified reduced Cu,Zn-SOD, reduced Cav1, and hypertension as independent factors associated with CMD. Collectively, these results provide, for the first time to our knowledge, supportive evidence for a potential link between systemic microvascular endothelial dysfunction and enhanced oxidative stress in patients with CMD, although its discriminatory performance was modest. Clinical characteristics of ANOCA and CMD Among 295 patients who underwent comprehensive invasive coronary function testing across experienced centers, CMD was identified in 43%, consistent with prior meta-analytic evidence reporting a prevalence of 41% in ANOCA cohorts [20]. The relatively high frequency of CMD underscores its importance and its potential to explain persistent symptoms in a substantial proportion of patients without epicardial stenosis. Furthermore, approximately 20% of patients exhibited overlapping CMD and vasospastic angina, illustrating the heterogeneity and complexity of coronary vasomotor disorders. As outlined in Table 1, CMD patients were older and more likely to present with hypertension, diabetes, diastolic dysfunction, and higher NT-proBNP levels compared with non-CMD patients. These findings indicate that lifestyle-related diseases, such as hypertension and diabetes, play a key role in the pathogenesis of CMD and that CMD is associated with increased LV overload. Since these are modifiable factors, early and proactive management of comorbidities may help alleviate future cardiac load and potentially prevent the development of CMD. Serum Cu,Zn-SOD, Cav1, and eNOS in CMD CFR is an integrated index reflecting both epicardial and microvascular function, while IMR provides specific information on the coronary microvasculature and is less influenced by epicardial stenosis than CFR [21]. This distinction enables separation of structural CMD, defined by increased IMR, from functional CMD, characterized by impaired vasomotor balance with preserved IMR but reduced CFR [22]. Our results demonstrate that Cu,Zn-SOD levels were reduced in CMD irrespective of epicardial vasospasm, but the reduction was most pronounced in those with abnormal IMR, implicating a role in structural CMD. In the coronary microcirculation, Cu,Zn-SOD generates H2O2, which acts as an EDHF critical for coronary autoregulation and metabolic vasodilatation [9]. Reduced Cu,Zn-SOD activity may impair these pathways, predisposing to arteriolar hypertrophy, fibrosis, and increased microvascular resistance. Additionally, two other SOD isoforms exist; ecSOD and mitochondrial SOD (MnSOD) [14]. We measured serum ecSOD levels and found no difference between CMD and non-CMD patients, highlighting the specific association of Cu,Zn-SOD with CMD pathogenesis. Among CMD patients with both abnormal CFR and IMR, Cav1 levels were also reduced, suggesting a relationship between microvascular vasomotor dysfunction and low Cav1 levels. Cav1 is involved in various signaling pathways, including angiogenesis [16], contributing to EDHF/H2O2-mediated vasodilatation [15], and its reduced activity may compromise microvascular homeostasis. Importantly, Cu,Zn-SOD correlated inversely with IMR, whereas Cav1 correlated positively with CFR, supporting their complementary involvement in structural and functional CMD, respectively. In contrast, serum eNOS levels did not differ between CMD and non-CMD groups or among CMD subtypes, implying limited utility as a circulating marker. These findings highlight Cu,Zn-SOD and Cav1 as more relevant contributors to CMD pathogenesis. Although reduced Cu,Zn-SOD was independently associated with CMD, its ability to discriminate between CMD and non-CMD patients was limited (AUC 0.611; Fig. S8A), with substantial overlap in biomarker levels between groups (Fig. S6). Therefore, Cu,Zn-SOD cannot be considered a standalone diagnostic biomarker for CMD in its current form. The primary value of Cu,Zn-SOD lies in pathophysiological insight rather than clinical diagnosis. Reduced Cu,Zn-SOD and oxidative stress in CMD Oxidative stress, defined as an imbalance between ROS generation and antioxidant defenses, is implicated in a broad spectrum of cardiovascular pathologies [23], [24], [25]. Superoxide radicals damage lipids, proteins, and DNA, leading to cell death and vascular dysfunction [26], [27]. Cu,Zn-SOD is a key defense enzyme, which is found widely in human cells and fluids, detoxifying O2− while generating H2O2 [28]. Importantly, the generated H2O2 acts as an EDHF, supporting endothelial-dependent vasodilation in endothelial cells of both mice and human microvasculature [14], [29]. Genetic studies provide compelling evidence for Cu,Zn-SOD’s role in vascular homeostasis. Mice lacking Cu,Zn-SOD exhibit accelerated age-related pathology, oxidative damage, and impaired vasodilation [30], [31]. In our study, serum Cu,Zn-SOD was reduced in CMD, especially in structural CMD defined by elevated IMR. Our findings may provide supportive clinical evidence for a potential link between impaired antioxidant capacity and EDHF/H2O2-related microvascular dysfunction. CMD has increasingly been recognized as a systemic disorder. Our previous studies in patients with MVA have demonstrated impaired EDHF-mediated vasodilation in peripheral microcirculation despite preserved NO-mediated dilation, consistent with systemic microvascular endothelial dysfunction [32]. Moreover, CMD has been linked to cerebral small-vessel disease and cognitive decline, suggesting broader vascular implications beyond the heart [33]. The reduction in serum Cu,Zn-SOD observed here supports the concept that CMD reflects a systemic oxidative stress state rather than a purely localized coronary phenomenon. However, as no direct measurements of tissue-level oxidative stress or EDHF activity were performed in this study, these mechanistic interpretations should be considered hypothesis-generating and require further validation. Furthermore, Cu,Zn-SOD is not currently included in ESC guidelines or consensus documents for the diagnosis or management of CMD. Thus, the primary value of this biomarker at present lies in improving our understanding of CMD pathophysiology rather than immediate clinical application. Future studies are warranted to evaluate whether combining Cu,Zn-SOD with other biomarkers or clinical parameters could improve diagnostic performance. ROS generating enzymes, such as NADPH oxidase and xanthine oxidase, may further exacerbate oxidative stress in CMD [27]. This raises the possibility that therapeutic interventions aimed at enhancing antioxidant defenses or inhibiting ROS sources could mitigate CMD and its systemic vascular manifestations. Role of the NOS system in CMD Although eNOS has been regarded as a primary source of endothelial-derived H2O2 [9], our study did not detect significant differences in serum eNOS between groups. One possible explanation lies in redundancy among NOS isoforms. Our previous experimental work has shown that in the absence of eNOS, neuronal NOS (nNOS) and inducible NOS (iNOS) can maintain EDHF/H2O2 production [34]. In eNOS/nNOS-double-knockout mice, coronary microvascular dilation was preserved by iNOS-derived H2O2, supporting compensatory roles of alternative isoforms [35]. In human, hypertension and diabetes, prevalent among CMD patients, are associated with upregulation of nNOS and iNOS [36], [37]. It is therefore conceivable that multiple NOS isoforms contribute to H2O2 generation in CMD, with Cu,Zn-SOD acting as the final common mediator converting superoxide into vasoactive H2O2. This may explain why reduced Cu,Zn-SOD, but not altered eNOS, emerged as a significant factor in CMD. Additional ROS sources, including NADPH oxidase and xanthine oxidase, may also play a role in superoxide generation [27]. A comprehensive assessment of NOS isoforms and non-NOS ROS pathways will be necessary to fully delineate the mechanistic role of oxidative stress in CMD. Therapeutic implications: oxidative stress as a target Previous studies showed that oxidative stress plays a key role in the development and progression of coronary artery disease and heart failure [27], [38], [39]. Given the central role of oxidative stress, therapeutic strategies to enhance antioxidant capacity or reduce oxidative stress represent a rational approach for CMD [40], [41]. In animal models, antioxidant supplementation (e.g., vitamin E, folic acid, nicotinamide riboside, N-acetylcysteine) attenuates myocardial fibrosis, improves LV function, and prevents ischemic injury [42]. In Cu,Zn-SOD-deficient mice, antioxidants improve not only vascular function but also systemic aging phenotypes, underscoring their broad effects [30]. However, large clinical trials of antioxidants in coronary artery disease and heart failure have largely failed to improve outcomes [27], [42]. These disappointing results may reflect the nonspecific nature of antioxidants, insufficient bioavailability in target tissues, or failure to address the specific molecular mechanisms underlying oxidative stress [42]. In CMD, the absence of targeted antioxidant therapy remains a major unmet need. Pharmacological therapies with pleiotropic antioxidant properties, such as statins, angiotensin-converting enzyme inhibitors, and β-blockers (e.g., carvedilol), may provide benefit in CMD [43], [44]. The success of stratified therapy in the CorMicA trial [45], which tailored treatment based on invasive diagnostic testing, may partly reflect antioxidant effects of these agents. Beyond pharmacology, exercise therapy is a particularly attractive strategy. Exercise enhances endogenous antioxidant defenses, reduces oxidative stress, and improves CFR in clinical studies [46], [47]. Its accessibility, cost-effectiveness, and systemic effects make it an especially promising intervention for CMD. Future research should focus on defining the molecular pathways linking oxidative stress to CMD, enabling development of more specific therapies. For example, modulating Cu,Zn-SOD activity could provide novel therapeutic avenues. Study limitations Several limitations must be acknowledged. First, the cross-sectional design precludes any causal inference regarding the relationship between reduced Cu,Zn-SOD and CMD. It remains unclear whether decreased Cu,Zn-SOD is a cause, a consequence, or an epiphenomenon of microvascular dysfunction. Second, the absence of a healthy control group limits the ability to define normal reference ranges and further constrains the interpretation of the biomarker findings. Third, coronary physiological assessment was performed in the left anterior descending artery alone, which may not fully reflect global coronary microvascular function. Although evaluating all three major coronary arteries would likely allow for a more comprehensive assessment of global CMD, this remains a subject for future investigation. Fourth, we did not directly assess local coronary oxidative stress, for example through coronary sinus sampling or tissue-level analyses. Therefore, mechanistic conclusions regarding EDHF/H2O2 signaling remain indirect. Fifth, although ecSOD was measured and found not to differ, MnSOD was not assessed. Previous work suggests MnSOD does not directly participate in EDHF/H2O2 signaling [14], [29], but its contribution to oxidative stress cannot be excluded. Sixth, this study lacks longitudinal follow-up data, and thus we were unable to evaluate the association between reduced Cu,Zn-SOD and clinical outcomes such as major adverse cardiovascular events or heart failure with preserved ejection fraction. This represents an important area for future research. Finally, although patients with overt acute inflammatory or infectious conditions were not included based on clinical assessment, we are unable to completely exclude the influence of subclinical inflammation on circulating oxidative stress markers. Conclusions In conclusion, reduced circulating Cu,Zn-SOD levels are associated with CMD in patients with ANOCA. Reduced Cav1 further suggests disruption of microvascular homeostasis. These findings suggest a potential role of systemic oxidative stress in the pathophysiology of CMD. However, given the cross-sectional design, modest discriminatory performance, and lack of direct mechanistic and prognostic data, further studies are needed to clarify the causal relationships and clinical relevance of this association. Disclosures: J-CMD is supported by the research grants from the Japan Heart Foundation. M.T. received speaker fees from Abbott Vascular. All other authors have no disclosures. Data availability. The data underlying this article will be shared on reasonable request to the corresponding author. Declaration of generative AI and AI-assisted technologies in the manuscript preparation process. During the preparation of this work the author(s) used ChatGPT (OpenAI, San Francisco, CA, USA) in order to improve the language of the manuscript. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article. CRediT authorship contribution statement Koichi Sato: Writing – original draft. Morihiko Takeda: Writing – review & editing, Supervision. Shohei Ikeda: Writing – review & editing. Yoshihisa Kanaji: Writing – review & editing. Masahiro Hada: Writing – review & editing. Tsunekazu Kakuta: Writing – review & editing. Takahiko Kiyooka: Writing – review & editing. Yuichi Ozaki: Writing – review & editing. Atsushi Tanaka: Writing – review & editing. Yoshiaki Kawase: Writing – review & editing. Hitoshi Matsuo: Writing – review & editing. Yoshiaki Mitsutake: Writing – review & editing. Yoshihiro Fukumoto: Writing – review & editing. Satoshi Miyata: Writing – review & editing, Formal analysis. Hiroki Aoki: Writing – review & editing, Formal analysis. Hiroaki Shimokawa: Writing – review & editing, Supervision, Conceptualization. Funding This work was supported by the research grants from the Japan Heart Foundation. Grant ID:No 210004 Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: M.T. received speaker fees from Abbott Vascular. All other authors have no disclosures. Acknowledgement We thank SRIA Corporation (Sendai, Japan) for constructing and maintaining the electric data capture system for J-CMD. Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.ijcha.2026.101943. Supplementary data The following are the Supplementary data to this article: Supplementary Data 1 Supplementary Data 2 References Associated Data Supplementary Materials Supplementary Data 1 Supplementary Data 2