Introduction

The central area where nerve fibers converge is referred to as white matter due to the presence of the myelin sheath, which is rich in lipids and gives it its white appearance. Of note, white matter plays a critical role in human cognitive function1. Leukoaraiosis (LA), also referred to as white matter hyperintensities (WMHs) or white matter lesions (WMLs)2 represents a group of clinical syndromes with various etiologies. LA is primarily characterized by diffuse, spotty, or patchy ischemic changes in the periventricular and central hemispheric white matter3. LA is an imaging diagnostic term first introduced by Hachinski in 19873. Its main imaging feature is the presence of high-intensity signals on T2-weighted or FLAIR MRI images, making it one of the most significant age-related imaging changes4,5. LA can cause a range of clinical symptoms, with cognitive dysfunction being the most prominent6,7.

With the aging population and advancements in imaging technology, the incidence of LA in the elderly has steadily increased. Its primary pathological features include myelin pallor, demyelination, oligodendrocyte apoptosis, and vacuolation5. Highly sensitive MRI examinations have revealed that some areas of white matter, previously appearing normal, gradually develop visible lesions, leading to secondary degeneration in distant cortical or brainstem regions, ultimately affecting brain function8,9,10. Currently, the diagnosis of LA largely depends on cranial imaging. However, patients may not exhibit clear imaging changes in the early stages of the disease, and by the time such changes are detected, significant central nervous system damage has often already occurred10. Therefore, there is an urgent need to identify biomarkers that can assist in the early diagnosis of LA and predict disease progression.

Previous research has shown that neuronal axons and the myelin sheath form the microstructure of white matter11. The myelin sheath, also known as myelin, is composed of proteins and lipids12. Myelin basic protein (MBP) is a key protein in myelin formation13 and its absence can lead to white matter loss, which is likely associated with impaired cognitive function14. The disruption of the blood-brain barrier (BBB) is considered a potential mechanism for LA and a significant contributor to cognitive dysfunction15. The BBB is primarily composed of vascular endothelial cells (VEC), the basement membrane, pericytes, and astrocytes16. VECs regulate cerebral blood flow, transport processes, and fluid clearance by interacting with other components of the BBB, thereby influencing oxidative synthesis, metabolite transport, and the balance of interstitial fluid in the brain17. In the central nervous system (CNS), VECs form tight junctions between cells through occludin (OCLN), which is critical in modulating BBB permeability18. The cerebral hemorrhage can cause BBB damage and promote the degradation of OCLN, which is then released into the bloodstream19,20.

Chronic cerebral hypoperfusion is a key contributing factor to LA, leading to BBB integrity disruption21,22. This process upregulates matrix metalloproteinase-9 (MMP-9) expression, resulting in degradation of tight junction proteins (particularly OCLN) and their subsequent release into circulation23,24,25,26. Concurrently with BBB disruption, the myelin sheath sustains damage, leading to the shedding of MBP, which subsequently enters the bloodstream through the compromised BBB27,28. However, the characteristics of MBP and OCLN in patients with LA and LA-related cognitive impairment remain unclear. To further explore the potential of MBP and OCLN as biomarkers for LA and LA-related cognitive impairment, we collected clinical data and blood samples from healthy individuals and patients with LA. We then compared the levels of MBP and OCLN between the two groups and analyzed the correlation between these proteins and clinical features.

Methods

Sample and clinical data collection

The subjects in this study were divided into the Control group and the LA group through consecutive screening at Jiangbin Hospital of Guangxi Zhuang Autonomous Region from March 2022 to March 2025. The Control group comprised age- and sex-matched healthy volunteers with: No history of cerebrovascular disease; Normal brain magnetic resonance imaging (MRI) (Fazekas 0). LA group inclusion required: Radiologically confirmed LA (Fazekas 1–3); Absence of acute stroke on diffusion weighted MRI. Two board-certified neuroradiologists blinded to clinical data independently evaluated Fazekas scores using axial FLAIR/T2-weighted images (slice thickness = 5 mm). Discrepancies were resolved by consensus review. The severity of LA was graded according to the Fazekas score29: Fazekas 0 indicates no white matter lesions, Fazekas 1 signifies punctate foci, Fazekas 2 denotes mottled lesions that have fused into plaques, and Fazekas 3 indicates large fused lesions.

The exclusion criteria were as follows: hemorrhagic stroke, ischemic stroke, immune deficiency diseases, malignant tumors, aphasia, and severe mental impairment. Clinical data collected included mean systolic blood pressure (MSBP) in mmHg, mean diastolic blood pressure (MDBP) in mmHg, heart rate variability (HRV) classified as normal, mild abnormality, moderate abnormality, or severe abnormality; fasting blood glucose (FBG) in mmol/L; 2-hour postprandial blood glucose (2hPG) in mmol/L; glycosylated hemoglobin (HbA1c) in %; blood lipid levels, including total cholesterol (TC) in mmol/L, triglycerides (TG) in mmol/L, high-density lipoprotein cholesterol (HDL-C) in mmol/L, and low-density lipoprotein cholesterol (LDL-C) in mmol/L; homocysteine (Hcy) in umol/L ; and neuron-specific enolase (NSE) in ng/ml. Mini-mental state examination (MMSE) and Montreal cognitive assessment (MoCA) scores were also conducted for all individuals with LA. The MMSE primarily assesses patients’ cognitive function across eight aspects, including time orientation, place orientation, immediate memory, attention and computation, delayed memory, language, and visuospatial abilities. The total score for the MMSE is 30, with the following classifications for cognitive function: 27–30 indicates normal cognition; 21–26 signifies mild cognitive dysfunction; 10–20 denotes moderate cognitive dysfunction; and 0–9 represents severe cognitive dysfunction. In contrast, the MoCA evaluates cognitive function in eight areas: memory, attention, orientation, language disorders, naming ability, visuospatial executive ability, abstract thinking, and delayed memory. The total score for the MoCA is also 30, with the following classifications: 27–30 indicates normal cognition; 18–26 signifies mild cognitive impairment; 10–17 denotes moderate impairment; and 0–10 represents severe impairment. The informed consent was obtained from all participants in both groups. This study was reviewed and approved by the ethics committee of Jiangbin Hospital of Guangxi Zhuang Autonomous Region (NO. LW-2024-016). Research involving human research participants were performed in accordance with the Declaration of Helsinki.

Detection of serum MBP and OCLN levels

The levels of MBP and OCLN were evaluated using enzyme-linked immunosorbent assay (ELISA) kits, specifically the MBP ELISA Kit (Boshen Biotechnology, BS-E4933H2) and the OCLN ELISA Kit (Boshen Biotechnology, BS-E4424H1). The measurements for MBP and OCLN were conducted using the Fully automatic enzyme free workstation (BIOBASE2000, BIOBASE). The level of neuron-specific enolase (NSE) was measured as part of the clinical routine laboratory testing at our hospital.

Statistical analysis

Due to the non-normal distribution of most continuous variables, as indicated by standard deviations exceeding half of the mean values, data are presented as median and interquartile range (IQR). Group comparisons for continuous variables were performed using the Mann-Whitney U test (for two groups) or the Kruskal-Wallis test (for multiple groups). Categorical data are presented as numbers and percentages and were compared using the Chi-square test or Fisher’s exact test, as appropriate. The Pearson correlation coefficient was calculated to analyze the relationship between the variables. The diagnostic performance of MBP, OCLN, and NSE was evaluated by constructing receiver operating characteristic (ROC) curves and calculating the area under the curve (AUC) with 95% confidence intervals (CI). Pairwise comparisons of the AUC values were performed for statistical significance using DeLong’s test. A two-tailed p-value of less than 0.05 was considered statistically significant. All statistical analyses were performed using the Statistical Package for Social Sciences (SPSS) software (version 25.0). GraphPad Prism software (version 10.0, GraphPad Software) was utilized to visualize the data.

Results

Patient characteristics

The brain MRI is shown in Fig. 1. The clinical baseline characteristics of the Control and the LA groups are summarized in Table 1. Among the included 198 individuals, there were 107 males and 91 females. In terms of LA severity, there were 99 individuals classified as Fazekas − 0, 44 as Fazekas − 1, 41 as Fazekas − 2 and 13 as Fazekas − 3. The laboratory characteristics were exhibited in Table 2. There was no significant difference in age between the Control group and the LA group. Compared to the Control group, both MSBP and MDBP were significantly higher in the LA group (p < 0.01). Moreover, the levels of FBG, TC, TG, LDL-C, and Hcy in the LA group were significantly elevated compared to the control group (p < 0.01).

Fig. 1
figure 1

The MRI image of brain. A Fazekas 0, healthy people. B individuals with LA (Fazekas 0). C individuals with LA (Fazekas 2). D individuals with LA (Fazekas 3). Arrows, white matter lesions.

Full size image
Table 1 The clinical baseline characteristics of the control and the LA groups.
Full size table
Table 2 Comparison of baseline clinical and laboratory characteristics between patients with LA and control subjects.
Full size table

The levels of MBP, OCLN, and NSE in patients with LA

The characteristics of MBP, OCLN, and NSE are presented in Table 3. Compared to the Control group, MBP and OCLN levels in the peripheral blood of the LA group were significantly elevated (p < 0.01). However, there was no significant difference in NSE levels between the two groups. To comprehensively assess the diagnostic potential of these biomarkers for LA, we performed receiver operating characteristic (ROC) curve analysis and calculated the corresponding area under the curve (AUC) values. This rigorous statistical approach allowed us to quantify the discriminative power of each biomarker in distinguishing LA patients from controls. The AUC values are presented in Fig. 2, showing that MBP achieved an AUC of 0.8761 (Fig. 2A), OCLN demonstrated superior performance with an AUC of 0.9592(Fig. 2B), while NSE showed relatively lower discriminative ability with an AUC of 0.6634(Fig. 2C). As detailed in Table 4, pairwise comparison of ROC curves revealed that OCLN possessed the highest discriminatory power (AUC = 0.9592), followed by MBP (AUC = 0.8761), with both significantly surpassing the performance of NSE (AUC = 0.6634; all p < 0.0001 vs. NSE). The difference in AUC between OCLN and MBP was also statistically significant (p = 0.0002).

Table 3 The characteristics of MBP, OCLN, and NSE.
Full size table
Fig. 2
figure 2

Diagnostic Performance of MBP, OCLN, and NSE for LA: ROC Curve Analysis. A ROC curve analysis of myelin basic protein (MBP) for LA diagnosis (AUC = 0.876); B Diagnostic performance of OCLN in LA detection (AUC = 0.959); C ROC analysis of NSE as a potential LA biomarker.

Full size image
Table 4 Diagnostic performance of serum biomarkers for LA.
Full size table

Associations between MBP, OCLN, and NSE with clinical features

Next, the correlation between the MBP, OCLN, and NSE and the clinical characteristics of the LA group (including HRV, stage of LA, MMSE, and MOCA) was analyzed. The results indicated no significant correlation between HRV and the levels of MBP, OCLN, and NSE (Figs. 3A-C). The level of MBP was positively correlated with the Fazekas score (R = 0.2888, p = 0.0018; Fig. 4A). There was no significant correlation between Fazekas score and the levels of OCLN, and NSE(p>0.05, Fig. 4B-C). In addition, there was a significant correlation between MBP and cognitive function, with MBP levels negatively correlating with MMSE scores (R=-0.6752, p < 0.0001; Fig. 5A) and MOCA scores (R=-0.7399, p < 0.0001; Fig. 6A). Nevertheless, no significant correlations were found between OCLN or NSE levels and cognitive function (Figs. 5B-C and 6B-C).

Fig. 3
figure 3

The correlation between HRV and the levels of MBP, OCLN, and NSE. A The correlation between HRV and the level of MBP. B The correlation between HRV and the level of OCLN. C The correlation between HRV and the level of NSE.

Full size image
Fig. 4
figure 4

The correlation between Fazekas score and levels of MBP, OCLN, and NSE. A The correlation between Fazekas score and the level of MBP. B The correlation between Fazekas score and the level of OCLN. C The correlation between Fazekas score and the level of NSE.

Full size image
Fig. 5
figure 5

The correlation between MMSE scores and levels of MBP, OCLN, and NSE. A The correlation between MMSE score and the level of MBP. B The correlation between MMSE score and the level of OCLN. C The correlation between MMSE score and the level of NSE.

Full size image
Fig. 6
figure 6

The correlation between MOCA scores and levels of MBP, OCLN, and NSE. A The correlation between MOCA score and the level of MBP. B The correlation between MOCA score and the level of OCLN. C The correlation between MOCA score and the level of NSE.

Full size image

Discussion

LA is a common imaging finding that often leads to cognitive impairment, lower limb dysfunction, and urinary incontinence30,31,32. Previous studies have demonstrated that hypertension, dyslipidemia (particularly TC, TG, and LDL-C), and homocysteine levels were significantly associated with the incidence of LA33,34. Herein, we found that the levels of TC, TG, LDL-C, and homocysteine in the LA group were significantly elevated, aligning with earlier research findings. Brain lesions may influence the cardiovascular system by regulating the brain-heart axis35. The prefrontal cortex (PFC) in the brain, along with the insula, amygdaloid nucleus, cingulate cortex, hypothalamus, and brainstem, regulates the autonomic nervous system (ANS) through a complex network of neural connections. This regulation ultimately affects the cardiovascular system and is commonly referred to as the brain-heart axis36,37. HRV is regarded as an indicator of cardiac autonomic nerve function38. While collecting patient blood pressure data, we further analyzed and compared the HRV between the two groups. However, there were no significant differences in HRV between the control and LA groups. Moreover, the correlation between HRV and LA stage was examined, and the results indicated no significant correlation between them. These findings suggest that LA may not impact the cardiac autonomic nervous system.

Currently, the diagnosis of LA primarily relies on brain imaging examinations. However, there may be no obvious imaging changes in the early stages of the disease10. Consequently, there is an urgent need to identify biomarkers that can assist in the early diagnosis of LA and evaluate disease progression. In the present study, we found that the level of OCLN in the LA group was significantly increased. Damage to the BBB is considered a potential mechanism underlying LA15. This damage can lead to the leakage of fluids, proteins, and other plasma components into perivascular tissues, potentially resulting in interstitial fluid accumulation and thickening and hardening of arteriolar walls. Such changes can impair blood vessel dilation, oxygen, and nutrient transport and ultimately affect cognitive function39,40. VECs, crucial components of the BBB, regulate its permeability through OCLN18. In the CNS, OCLN is crucial in promoting tight junctions between endothelial cells in blood vessels18. When the BBB is compromised, OCLN is shed from VECs and enters peripheral blood circulation. Previous studies have confirmed that elevated levels of OCLN in the blood have high specificity and sensitivity in reflecting BBB damage41,42,43. NSE serves as a specific biomarker for neuronal integrity, with elevated levels in peripheral blood indicating neuronal damage and often correlating with poor prognosis44,45. Nevertheless, in our study, we did not observe significant changes in NSE levels in the LA group. White matter comprises of neuronal axons and the myelin sheath, which is primarily formed by oligodendrocytes46. LA is mainly attributed to damage to the functional structure of oligodendrocytes. The myelin sheath also referred to as myelin, consists largely of proteins and lipids, with MBP serving as a major protein involved in myelination13. The loss of MBP has been linked to the pathogenesis of LA14. Herein, we observed that MBP levels were significantly elevated in the LA group. To comprehensively assess the diagnostic potential of MBP, OCLN, and NSE in detecting LA, we performed ROC curve analysis. The ROC curve analysis demonstrated outstanding diagnostic accuracy for both OCLN and MBP in distinguishing LA patients from controls group. Notably, the diagnostic performance of OCLN was found to be statistically superior to that of MBP. In contrast, NSE, a marker of general neuronal injury, showed only moderate discriminative ability, which was significantly lower than both OCLN and MBP. Furthermore, the positive correlation was found between MBP levels and LA grading, indicating that MBP levels are associated with the severity of LA.

Extensive research demonstrates that LA is intricately linked to the pathogenesis of cognitive impairment47,48. As LA progresses, patients frequently experience cognitive dysfunction6,7. Previous research has confirmed the presence of limbic system circuit nerve fibers surrounding the ventricular system, which play a crucial role in cognitive activities such as memory and emotional behavior49. Damage to the white matter can disrupt the circuit fibers around the ventricles or impair corticocortical connections. Such disruptions can adversely affect cognitive functions, particularly attention, memory, executive function, and information processing speed50. Cognitive dysfunction resulting from periventricular white matter damage reflects impairment in cholinergic nerve projections originating from the basal forebrain to the cortex50. The greater the extent of white matter damage, the more pronounced the decline in verbal memory, visuospatial memory, organization, visual scanning, and movement among patients with LA7. This damage disrupts the cortical-subcortical fiber connections and impacts the functioning of the frontal lobe-striatum system, contributing to executive dysfunction7. The frontal and parietal lobes are particularly susceptible to LA, often leading to cognitive impairment51. Beyond structural alterations, LA contributes to cognitive dysfunction through multiple pathophysiological mechanisms, including disorders of energy metabolism52synaptic dysfunction53 and neuroinflammation54. Oligodendrocytes are specialized glial cells responsible for myelinating axons in the central nervous system, forming the structural basis of white matter through their interactions with neuronal processes46. Oligodendrocytes participate in neuron-oligodendrocyte metabolic coupling, providing lactate and other energetic substrates to maintain axonal function and survival55. White matter injury disrupts oligodendrocyte integrity, impairing their critical roles in axonal metabolic support55. This impairment leads to synaptic dysfunction and ultimately contributes to cognitive deficits53,56. In LA lesions, the activated microglia release pro-inflammatory cytokines that contribute to myelin degradation, exacerbating cognitive dysfunction57,58,59. Following myelin injury, macrophage-mediated clearance of myelin debris occurs; however, excessive macrophage activation can exacerbate neuroinflammation and potentially contribute to fibrotic scar formation60. Under these pathological mechanisms, damaged myelin releases MBP, which subsequently enters peripheral circulation through the compromised BBB, ultimately leading to elevated MBP levels in peripheral blood. Here, we observed a negative correlation between the level of MBP in peripheral blood and the scores on the MMSE and MOCA. This discovery suggests that elevated MBP levels in peripheral blood may result from myelin sheath disruption, and that the severity of myelin damage correlates with the degree of cognitive dysfunction in patients. Furthermore, it implies that MBP could serve as a valuable biomarker for assessing the severity of cognitive impairment in LA patients.

Potential confounding factors influencing both OCLN and MBP levels must be carefully considered in the interpretation of results. Several variables may affect our biomarker measurements: chronic conditions, pre-analytical variables and LA-unrelated neurological conditions. Chronic conditions like hypertension and diabetes can independently affect BBB permeability through oxidative stress and advanced glycation end-products, potentially confounding our OCLN measurements61,62. Preanalytical variables including hemolysis (controlled via standardized sample processing) and circadian rhythms (addressed through morning blood collection) were carefully considered63,64. While we excluded major comorbidities and matched groups for vascular risk factors, unmeasured variables such as silent brain infarcts or subclinical neurodegenerative processes may contribute to residual confounding12,65. Future studies incorporating CSF biomarkers, neuroinflammatory markers, and detailed medication histories would help clarify these relationships.

Currently, effective treatments for LA and associated cognitive dysfunction remain elusive. Previous studies have indicated that using angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin receptor blockers (ARBs) may reduce the incidence of LA and improve both the number of lesions and the extent of white matter injury66,67. Additionally, cilostazol has been shown to alleviate gait abnormalities in LA patients by enhancing BBB permeability and inhibiting microglial activation and apoptosis68. However, these treatments lack standardized evaluation criteria for assessing their therapeutic effects on LA. The findings from our study may offer theoretical guidance for diagnosing and treating LA. Nonetheless, this study has several limitations. Firstly, the sample size was insufficient, and further research with a larger cohort is necessary to clarify the baseline levels of MBP and OCLN. Secondly, the underlying mechanisms of cognitive dysfunction related to MBP levels in LA patients require further investigation.

Conclusion

In the present study, we demonstrated that MBP and OCLN may be used as biomarkers of LA, and MBP might have potential clinical value for predicting cognitive dysfunction in patients with LA.