Alzheimer's disease (AD) is a complex neurodegenerative disorder characterized by memory loss and cognitive impairment (1). The formation of neuronal amyloid-β (Aβ) plaques and tau protein hyperphosphorylation are histopathologic characteristics of AD (2–5). Many factors are involved in the development of AD, including age, heredity, obesity, diabetes, hypertension, and brain trauma (6–8). Recently, early blood–brain barrier (BBB) breakdown that contributes to AD pathogenesis has been increasingly recognized (9). The BBB, which is an important boundary between the blood and brain, is mainly formed by vascular endothelial cells of the central nervous system (CNS) and relies on continuous complexes of tight junctions (9–12). CSF/plasma albumin quotient (Qalb) is an ideal measure for the permeability of the BBB. Because albumin is only synthesized in the liver, albumin in the cerebrospinal fluid is therefore derived only from the blood and is transported to the central nervous system by passive diffusion (13). Thus, an abnormal elevation of Qalb is thought to reflect increased BBB permeability. In previous studies, Musaeus et al. similarly found that Qalb can contribute to the identification of different types of dementia, which may also be related to the impairment of BBB (14). BBB damage and vascular dysregulation can be detected before a cognitive decline in patients with AD (15, 16). Meanwhile, a large number of studies also indicated that Aβ accumulation is generally accompanied by the occurrence of cerebral small-vessel disease (CSVD) (16), indicating that there is a synergistic effect between BBB damage and CSVD in the occurrence of AD. CSVD involves a series of pathological processes, including perforation of the cerebral arterioles, venules, and capillaries (17). White matter hyperintensity (WMH) and cerebral microbleeds (CMBs), identified using magnetic resonance imaging (MRI), are diagnostic markers of CSVD (18).

Previous studies evaluating the relationship between brain Aβ (using PET imaging), CSVD, and cognition have shown conflicting results. Some studies found no relationship between brain Aβ and WMH in individuals with cognitively normal, mild cognitive impairment (MCI), subcortical vascular MCI, and AD (19), while a few studies reported that higher brain Aβ correlated with higher WMH volumes (20, 21). Another study using PET to show Aβ compliance found no interaction effect between Aβ and lacunae or CMB in AD (22). The pathogenesis of CMB includes small vascular wall damage caused by vascular risk factors and the accumulation of Aβ. Therefore, the development of cerebral microhemorrhage in patients with AD may help explain the overlap of cerebrovascular and neurodegenerative pathological mechanisms in cognitive dysfunction and dementia (23). Moreover, whether microvascular disease damage is associated with BBB damage in AD remains unclear.

In this study, we investigated the association between CMB and the BBB dysfunction measure, Qalb, as well as fluid biomarkers of Ab/tau in patients with AD. Our findings may deepen our understanding of the impact of small cerebrovascular lesions and BBB breakdown on AD pathology.

In this study, we found that the Qalb value and CSVD burden in patients with AD were significantly higher than those in the cognitively normal population, and the Qalb value positively correlated with the CSVD burden. Additionally, we explored the relationship between CMB, an important marker of CSVD, and Qalb in AD. Patients with AD were divided into two groups: those with CMB and those without CMB. The average age of patients with AD with CMB was higher, and they had a higher CSVD burden and APOE4 carrier rate. In the CMB group, Qalb increased with the increasing number of CMB. Additionally, the number of CMB negatively correlated with CSF Aβ42 concentration after adjusting for age, sex, and APOE genotype. Meanwhile, more microbleeds were accompanied by a higher Qalb after adjusting for age, sex, APOE genotype, and CSF Aβ42. This reflects that BBB damage may exacerbate cerebral microbleeds, which is independent of Aβ pathology in AD progression. Finally, we explored the differences in Qalb, CSVD burden, and CMB between APOE4 carriers and non-carriers. Except for the trend of a higher Qalb in patients with AD with APOE4, no other differences were found.

In recent years, the role of BBB damage in AD has been widely explored. Generally, BBB permeability may promote the accumulation of Aβ (29–32). In previous studies, Qalb was confirmed to be associated with the BBB and CSVD burden in AD (33, 34). When the BBB is compromised, the permeability of albumin and IgG into the normal CSF is altered. Protein and other molecules in the blood may enter the brain through the enlarged BBB gap, leading to elevated albumin concentration in CSF. It is generally recognized that higher Qalb represented greater damage to the BBB in neurodegenerative dementias (16, 33). In this study, we found that patients with AD had higher Qalb. Moreover, the possibility of higher Qalb accompanied by increased Aβ deposition was also confirmed. However, the correlation between Qalb and Aβ decreased after adjusting for the APOE genotype, which may indirectly indicate that APOE4 is involved in the destruction of BBB by Aβ pathology. Therefore, we analyzed the differences in BBB function between APOE4 carriers and non-carriers. Although no significant difference was observed between the groups, there was a trend for higher Qalb in APOE4 carriers in the AD group. These results also further confirmed that the genetic polymorphism of APOE may be indirectly involved in the BBB damage in AD, by mediating Aβ pathology.

In addition to BBB damage, cerebrovascular imbalance and vascular factors also contribute to the occurrence and development of AD. In our study, patients in the AD group had a higher CSVD burden. To further explore the possible causal relationship between BBB damage and AD combined with cerebral microangiopathy, we also performed a subgroup comparison of CMB. In our study, CMB number positively correlated with Qalb in patients with AD, which was independent of age, sex, APOE genotype, and Aβ pathology. Vascular injury can lead to a prominent BBB rupture, presenting as cerebral microhemorrhage, which is commonly seen in individuals with an increased genetic risk for AD and MCI. BBB damage is more severe in patients with AD with more microbleeds, which further indicates that the occurrence of CMB is related to BBB dysfunction. Increased BBB permeability leads to the extravasation of red blood cells from small cerebral vessels, which is a major feature of CMB. BBB dysfunction allows neurotoxic blood-derived components, blood cells, and pathogens to enter the brain, leading to the deterioration of the brain environment, further aggravating brain degeneration and CMB occurrence (34). The location of the microbleeds is related to its etiology: CAA causes lobar microbleeds, while hypertensive vascular lesions cause microbleeds commonly in the basal ganglia, thalamus, cerebellum, and brainstem. Microbleeds in patients with AD are mainly observed in the lobars, especially the occipital lobes (35, 36). However, the molecular and cellular mechanisms underlying BBB dysfunction associated with CMB development remain unclear. Previous studies on CMB and AD have found that cerebral microbleeds may be a bridge between other important theoretical hypotheses about the pathogenesis of AD, the amyloid cascade hypothesis (37–40). In our cohort, we found a more severe CSVD burden and CMB in the probable AD population defined as positive for ¹⁸F-florbetapir PET. Similarly, we also observed a higher proportion of combined cortical microhemorrhage in patients with AD. Although no differences in biomarkers were found in patients with AD with or without CMB in the present study, correlation analysis revealed that CSF Aβ42 levels decreased with an increase in CMB number, which was consistent with a previous study (41). This suggests that the Aβ metabolic pathway may be directly involved in the formation of intracerebral microbleeds in patients with AD. Even after adjustment for Aβ42, the correlation between CMB and Qalb remained significant, which further indicated that BBB damage may directly contribute to the development of CMB in AD.

Because of the limited sample size, we did not distinguish the specific brain lobe and deep regions of CMB to further analyze their relationship between AD pathology and BBB damage. Further studies should include more samples and perform detailed analyses to explore the relationship between BBB and CSVD in AD.

Overall, our study provided evidence that patients with AD have a higher Qalb and CSVD burden, and Qalb value positively correlated with CSVD burden. Furthermore, Qalb was closely associated with CMB when Aβ42 levels were adjusted. This further indicates that BBB damage may be involved in the pathological mechanism of AD and may be a risk factor for CSVD pathogenesis.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Department of Ethics Committee, the First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei, China. The patients/participants provided their written informed consent to participate in this study.

XL, QW, and ZC designed this study, analyzed and interpreted the data, and wrote the manuscript. YW and CL recruited patients and collected samples. PW, MN, and CL performed MRI and PET scanned and analyzed. FG and LD performed the experiments. All authors are accountable for the contents of this research.