Progressive mechanical and structural changes in anterior cerebral arteries with Alzheimer’s disease

Alzheimer’s disease (AD), a neurodegenerative disease characterized by the appearance of neurofibrillary tangles and accumulation of amyloid-β (Aβ) peptide in the brain [24, 40], is the most prevalent cause of dementia [1]. As AD progresses, brain atrophy occurs and patients lose brain function corresponding to the atrophy location, leading to gradual cognitive decline, personality changes, and eventual inability to sustain their activities of daily living independently [63]. Cerebrovascular changes are common neuropathologic findings in aged subjects with dementia [48]. With neurodegeneration in dementias such as AD, neurovascular coupling is impaired and results in perfusion dysfunction, which renders the brain more susceptible to high pressure [47].

Despite the close association between nerves and vasculature, compared with studies on neurons, cerebrovascular remodeling during the progression of neurological disorders and the neurovascular relationships in the brain are less delineated. Among the limited studies, using combined clinical diagnosis and noninvasive assessments such as MRI or transcranial Doppler ultrasonography, cerebrovascular alterations in AD were reported. An increase in arterial stiffness was found in patients with AD compared to the control group [56], and there was a strong association between compliance reduction in cerebral arteries and presumptive AD diagnosis [57]. However, a direct assessment relating the biomechanical behavior of cerebral arteries with AD progression is lacking.

Compliance of large cerebral arteries is critical as these arteries dampen the pulsatile pressure and protect the microcirculation and blood brain barrier (BBB) from damage [35]. Cerebrovascular dysfunction can render the brain more susceptible to pulse pressure and lead to BBB breakdown [43]. Aβ normally can be eliminated from BBB along blood vessels via artery contractions [3, 18]. With impaired BBB and vascular contractibility, Aβ is more likely to deposit in brain tissues and the cerebrovascular system [4]. Studies showed that less compliant cerebral arteries lead to ineffective clearance of Aβ and other toxic metabolites [29, 37]. Additionaly, Aβ deposition in the arterial wall was shown to be a major cause of microhemorrage in AD [38]. Therefore, compliance of large cerebrovascular arteries plays a critcal role in the clearance of Aβ and other toxic metabolites in the brain. Aβ deposition, due to loss of compliance of large cerebral arteries, results in smooth muscle cells (SMC) atrophy, causing further decrease of vascular contractability and cerebral blood flow [30, 68], creating a vicious cycle.

Cerebrovascular dysfunction can have detrimental impacts on the brain and is closely associated with cognitive impairment. The goal of this study is to provide an understanding of cerebrovascular remodeling in the progression of AD. Human anterior cerebral arteries (ACAs), which originate in the Circle of Willis and supply blood to the upper and medial cerebrum surfaces, were used in this study. To establish an understanding of the changes in mechanical properties of cerebrovascular arteries with AD progression, biaxial extension-inflation tests were performed on ACAs from the control and pathologically diagnosed AD groups, from which the stress vs. stretch relationships in the circumferential and longitudinal directions were obtained. Additionally, histology studies, multiphoton imaging, birefringence microscopy, and mass spectrometry (MS) were performed to examine structural integrity of the arterial wall. These studies showed progressive media atrophy, elastic fiber degradation, and adventitia structural disorganization in the cerebrovascular tissue that correlated with AD development.

The average age for each group is reported in Table 1 with no significant difference among the groups. The average outer diameter, inner diameter, and thickness of ACAs from the AD groups decreased slightly compared to the control group (Fig. 1a-c). The average in vivo axial stretch ratio increased slightly with AD progression, from 1.10 ± 0.01 for control to 1.11 ± 0.02 and 1.13 ± 0.02 for early AD and intermediate and advanced AD groups, respectively (Fig. 1d). However, there was no significant difference among all groups in ACA dimensions and in vivo axial stretch ratio.

Changes in the mechanical properties of ACAs with AD progression were studied using biaxial extension-inflation test (Fig. 2). At 80 mmHg, the axial force increased significantly from 18.78 ± 4.40 mN for the control, to 45.93 ± 18.75 mN and 51.18 ± 13.68 mN for the early and intermediate and advanced AD groups, respectively (p < 0.05, Fig. 2a). The pressure-diameter response showed a leftward shift with AD progression (Fig. 2b). Consistently, as AD developed, the initial low stiffness region of the circumferential stress-stretch curve shortened (Fig. 2c). This was also manifested by the leftward shift of the maximum circumferential stress-stretch curves indicating arterial stiffening with AD progression (Fig. 2c). The longitudinal stress-stretch curves (Fig. 2d) also demonstrated a similar arterial stiffening behavior with shortened toe region and pronounced stress elevation in the intermediate and advanced AD group. To better analyze the arterial mechanical changes, the stress-stretch curves in the circumferential direction were fitted with exponential functions and differentiated to obtain tangent stiffness (Fig. 2e, f). As AD progressed, tangent stiffness in the circumferential direction increases, with significance (p < 0.05) found between the control and intermediate and advanced AD groups (p < 0.05).

Microstructural study of ACA revealed a trend of medial atrophy and adventitial layer disorganization with AD progression. Histological images with Movat’s stain showed a well-developed internal elastic lamella (IEL) presented between the medial and the intimal layers (Figs. 3a–d). The media consists of smooth muscle cells (SMCs), sparse elastic fibers, and small amounts of proteoglycan and collagen fibers. Collagen was abundant in the adventitial layer observed in the form of wavy bundles (Fig. 3a). Media atrophy was observed, characterized by the loss of SMCs and elastic fibers starting from the media-adventitia interface (Fig. 3b). As AD progressed, the atrophy region extended towards the lumen, causing a wider empty band with loosened structure, especially in the advanced AD stage (Fig. 3c, d). The progressive degradation of elastic fibers was further observed in multiphoton images taken at the circumferential cross section of ACA (Fig. 3e–l). Although the distribution of elastic fibers in the media was sparse, the loss of elastic fibers become evident with an empty band appearing at the media-adventitia interface at the intermediate and advanced AD stages (Fig. 3g, h, k, l). This empty band separated the elastic fiber network from the media and adventitia and was bordered by a thin layer of wavy elastic fiber at the interface of media and adventitia which did not seem to degrade with AD progression. Collagen fibers in the media seemed to maintain a continuous and wavy configuration among all groups. The width of the empty band was quantified for all groups (Fig. 4). The average area fraction was 31.8 ± 8.5% in media and 47.8 ± 16.4% in adventitia. Based on this result, when an observable empty band appeared in the multiphoton image, the width of the band was determined when the area fraction value was below 14.8%. The width of the empty band in the elastic fiber network significantly increased in the intermediate and advanced AD groups (Fig. 4i). The average empty band width increased significantly to 21.0 ± 15.4 μm in the intermediate and to and 32.8 ± 9.24 μm in the advanced AD group, while the empty band was unnoticeable in most of the control and early AD samples.

The progressive changes in structural organization can be seen in the qBRM images presented in Fig. 5. For each sample, qualitative images of the entire section (Figs. 5a, d, g, j) were acquired with CCP-BRM. From each full-sample image, a region was selected for qBRM. In the relative retardance maps (Fig. 5b, e, h, k), the intensity of the media dropped off heavily with advanced AD as the extend of separation and loss of structural integrity of the SMC layer increased. The same general trend was observed in the relative retardance values of the adventitia as the extent of collagen fiber degeneration increased in advanced AD and the extent of alignment decreased. Furthermore, these structural changes were also represented in the optic-axis orientation maps (Fig. 5c, f, i, l), where the loss of structural integrity of the layers of the arterial wall resulted in a loss of its compact structure and uniform orientation. The qBRM images of the AD cases, specifically for collagen of the adventitia, showed a clear trend of increased disorder for each step further in AD progression. Collagen fibers were seen to shift from being more circumferentially aligned and wavy (Fig. 5c, f) to more disordered and separated as AD progresses (Fig. 5i, l). The qBRM images also showed evidence of structural breakdown and delamination, especially of the endothelial intimal layer and the IEL.

Mass spectrometry results showed that the relative amounts of collagen are altered in AD compared to controls (Table 2). The levels of type I collagen were increased in AD, and this increase was accompanied by a decrease in type IV collagen. Further, gene set enrichment analysis revealed that several groups of genes were significantly differentially expressed (FDR < 0.05) (Fig. 6). Growth factor binding (GO:0019838), ECM structural constituent (GO: 0005201), and protease binding (GO: 0002020) were significantly enriched in AD. These gene sets include type I collagen, HtrA serine peptidase 1, elastin, biglycan, and periostin. The structural constituent of muscle gene set (GO:0008307) was significantly under-expressed in AD. On an individual protein level, the abundances of several members of the collagen family were significantly increased in AD: types V, VII, XII, XVI, and XXI (FDR < 0.05) (Fig. 7). The levels of several smooth muscle cell-associated proteins decreased in AD ACA relative to controls, specifically perlecan, myosin light polypeptide 6 and myosin regulatory light polypeptide 9, and leiomodin-1 (FDR < 0.05) (Fig. 7).

This study focused on the structural and functional changes of human ACAs with AD progression. Changes in wall structure were observed including media atrophy, elastic fiber degradation, and adventitial collagen disorganization, all of which likely contributed to progressive arterial stiffening with AD development. Histological staining revealed media atrophy, characterized by the loss of SMCs and elastic fibers, in the ACAs with AD. The atrophy initiated at the media-adventitia interface and propagated towards the lumen as AD developed (Fig. 3). Multiphoton imaging provided further evidence of the degradation of elastic fibers accompanying the loss of SMCs (Fig. 4). Elastic fiber degradation appeared in early AD stage and increased in later AD stages. As a result of elastic fiber degradation, an empty band absent of SMCs and elastic fibers was observed at the media-adventitia interface that widens significantly in later AD stages (Figs. 3 and 4). Although the role of elastic fibers in cerebral arteries and its degradation with AD remains to be understood, it is important to note that the media-adventitia interface contains perivascular nerves [10]. Cerebral arteries are innervated extrinsically to control vascular contraction [17]. As AD progresses, significant perivascular innervation loss has been reported [11]. Among limited studies on perivascular innervation degradation and corresponding artery structural changes, elastin downregulation was reported with perivascular innervation degradation in femoral artery and abdominal aorta [20]. Mass decrease in both middle and posterior cerebral arteries after perivascular innervation removal, indicating loss of ECM components or SMC in large cerebral arteries, has also been reported [8, 52]. The cross-sectional area of elastin in rat cerebral arterioles decreases after sympathetic denervation [6]. Innervation loss may damage blood flow regulation [31], which may in turn lead to inadequate driving forces in Aβ drainage, causing further media atrophy [21, 39, 77]. The media atrophy at the media-adventitia interface observed in this study (Figs. 3 and 4) is especially intriguing because soluble Aβ was reported to drain along the basement membrane between SMCs in the media [14, 75], and deposited Aβ was found replacing SMC in smaller cerebral arteries at the media-adventitia interface [74]. With media atrophy and Aβ elevation in AD, vascular pulsation is weakened, leading to more insufficient Aβ drainage [75]. Evidence has also shown that when Aβ in artery wall interacts with endothelium, oxidative stress increases which further contributes to Aβ increase, neurodegradation, and possible cognitive decline [69, 70].

Adventitial collagen disorganization is another major structural change observed in this study (Fig. 5). Due to the structural anisotropy and organization of ACA, the arterial components exhibit optical birefringence. Namely, strong birefringence is seen in the collagen fibers of the adventitial layers, while the elastic fibers and SMCs that make up most of the media exhibit weaker birefringence [78]. Using polarized light microscopy, the birefringence of these arterial components can be reliably imaged in a label-free manner [65] that enables the determination of the relative retardance and local orientation of structural anisotropy of cells and ECM components for every pixel in the image. These imaging approaches have been used to characterize collagen realignment across arterial layers with varying transmural pressure [26] and to study regions of cerebral arteries susceptible to aneurysm [13, 59]. In our study, the quantitative metrics from qBRM provided insights into the disorganization or decompaction of the layer structures in vasculature. As a function of AD severity, we observed a loss of compaction and integrity of SMCs in the media, evidenced by reduced intensity, as well as degeneration of collagen structure in the adventitia, evidenced by broader and more disoriented (more colorful) layers of collagen (Fig. 5). These observations demonstrated progressive structural breakdown in cerebral arteries with pathological AD development.

The structural changes in SMCs and ECM resulted in compromised arterial integrity and the mechanical properties of ACAs in AD progression (Fig. 2). As a result of SMC atrophy, the proportion of vessel wall thickness occupied by the SMCs was reduced (Fig. 3), leading to compromised contractility of cerebral arteries [30]. The increased systematic arterial stiffening that AD patients experience always accompanies higher level of dementia and cognitive impairment [32, 53, 64]. Among limited studies that relate the cerebral artery wall stiffness and dementia, Rivera-Rivera et al. [56] observed increased intercranial artery stiffness in AD patients from noninvasive transcranial pulse wave velocity measurements by MRI, which aligned with our findings (Fig. 2). Higher arterial stiffness results in compromised blood perfusion, and the blood pressure increases correspondingly [42]. Such high pressure in cerebral arteries is further passed down to smaller intracranial vessels, which can cause microvascular lesion such as microhemorrhage and vascular rupture [34]. In addition, chronic hypoperfusion can be the onset of sporadic AD due to damage on brain metabolism and cognitive function caused by long term insufficient oxygen and nutrients supply [22]. Furthermore, in AD patients, brain lesions emerged with high arterial stiffness and white matter hyperintensity was reported to cause decline in cognition [34]. ACA supplies blood to the frontal lobe, where integrity changes of the white matter mainly occur when AD first develops [28].

Mass spectrometry measurements correlated with reduced SMC activity in AD. Using gene set enrichment analysis, the structural constituent of muscle gene set was shown to be negatively enriched in AD (Fig. 6). This gene set consists principally of myosin family members and its under-expression is consistent with SMC dysfunction. Loss of myosin VI, in particular, has been linked to alterations in synaptic structure [50]. In addition to being present in the structural constituent of muscle gene set, SMC components are also significantly decreased on an individual protein level (Fig. 7). Two myosin components, myosin light chains and Leiomodin-1, were observed at significantly lower levels in the ACAs from the advanced AD group (Fig. 7), which may have implications for the reduced smooth muscle contraction. Leiomodin-1, the level of which decreased in AD, has been detected in the aorta and in the vascular smooth muscles in the lung [19, 60]. The LMOD1 gene has been identified as a risk locus in coronary artery disease,its expression has been found to be reduced in SMCs within atherosclerotic lesions, and its reduced expression is associated with decreased cell contraction [49]. Decreased expression of myosin heavy chains has been associated with reduced force generation of SMCs and reduced arterial contractility [71].

Mass spectrometry results further indicate aberrant ECM structure and assembly in AD. Two major subfamilies of collagens are fibrillar collagens and fibril-associated collagens with interrupted triple helices (FACIT). Fibrillar collagens contain one triple helical domain and often provide structure to the ECM, while FACITs contain multiple triple helices and assist in conferring biological activities to collagens [55]. The ECM structural constituent gene set, which includes several fibrillar and FACIT collagens was positively enriched in AD, indicating that the structure and biological activity of the ECM is dysregulated in the AD cerebrovasculature (Fig. 6). A similar enrichment of the ECM structural constituent gene set has also been observed in Parkinson’s disease [23, 54]. In the present study, the changes in the expression of collagen type genes are likely contribute to the changes of biomechanical function in ACAs.

The relative levels of individual collagens are also affected in the AD ACA (Table 2). The most pronounced changes to overall collagen composition are an increase in type I collagen in AD and a corresponding decrease in type IV collagen. Similar changes have been observed in small vessel disease, where an increase in fibrillar type I and III collagens is accompanied by a decrease in nonfibrillar type IV collagen [45]. The increase in fibrillar collagens correlates with arterial wall thickening, while the decrease in type IV collagen is associated with basement membrane dysfunction. The abundance of collagens is also affected in the AD ACA (Fig. 7). Type XXI collagen, the level of which increased in AD, is known to contribute to the ECM assembly of the vascular network [15]. While little is understood about the function and relevance of type XXI collagen under pathological conditions, it is thought to mediate protein–protein interactions between fibrillar collagens, serving as a molecular bridge in the ECM [7]. Type XIV collagen, the level of which decreased in AD, is known to regulate collagen fibril formation, a function that is particularly important in tissues of high mechanical demand [67]. Its absence has been shown to compromise mechanical properties of tissues including skin, tendons, and the myocardium [2, 67].

Response to injury and stress in the cerebrovasculature are also dysregulated in AD (Table 2, Fig. 7). Type V collagen, which is increased in AD, has been shown to interact with interleukins in the context of pulmonary arterial hypertension, thus pointing to a possible inflammatory role [5]. Another FACIT collagen that is dysregulated is type XII, the expression of which is increased in AD and has been demonstrated to increase with shear stress [41]. While little is known about its role in the cerebrovasculature, it has been identified as a promoter of axonal regeneration after spinal injury in a zebrafish model [73]. Also enriched in AD was type VIII collagen, which is an important component of the endothelium of blood vessels. Type VIII collagen is expressed by vascular SMCs, and its expression increases in response to vascular injury, particularly in the media and neointima,further, it has been shown to be upregulated in diseases associated with vascular remodeling and angiogenesis [33, 62]. Along with other factors, collagens V, VIII, XIV, and XVI have been shown to be upregulated after cardiac injury, altering the mechanical properties of scar tissue [79]. While the exact roles for the dysregulation of these different types of collagens in the cerebrovasculature remains to be elucidated, our study points to their potential involvement in inflammation and response to vascular injury. AD is associated with inflammation,analyses of proteomic and glycoproteomic data have revealed that inflammation-related proteins, including histocompatibility complex proteins, are over-represented in AD relative to controls [61, 82]. Further study is needed to determine whether these and other collagens have a role in mediating inflammation and injury response in the cerebrovasculature.

The study was limited to analysis of the ACA. However, the middle cerebral artery and downstream arterioles may be more relevant for the development of early medial temporal lobe pathology that occurs in AD. Future studies should examine more regions and larger numbers to separately examine associations with beta-amyloid and tau pathology.  There was no consideration of sex-dependency in AD development, which deserves further investigation. Arteries were inflated to mean arterial pressure (80 mmHg). Future studies considering the physiological pressure range may add information on the stress-stretch behavior within a broader pressure range. Arteries from tissue bank were kept frozen. Future studies could use freshly procured tissue to account for contributions from the SMCs to the mechanical response. Structural changes in the arterial wall can be location specific. However, imaging and further structural analysis was limited to a few regions due to its relatively small field of view. Future studies can expand on the imaging domain for a more comprehensive understanding of the arterial wall structures. Direct comparison between histology and multiphoton images should be made carefully as histology images were obtained using a very thin section of tissue sample (5-μm thick), while multiphoton images were maximum intensity projection images obtained within an imaging depth of about 60–-80 μm. To quantitatively evaluate the relationship between AD and the structural breakdown of vasculature using qBRM, it will be necessary to investigate a larger number of arterial sections from more patients with various stages of AD. For these future investigations, it will be important to ensure that the arterial structure is optimally preserved during fixation and slide preparation, as quantifying structural disorganization with metrics will require consistent and reliable tissue sections for imaging. The proteomics study was performed using data-dependent acquisition, which is biased toward abundant species. Thus, information about less-abundant proteins may be lost. For example, elastin is a key constituent in the maintenance of arterial mechanics, but its sequence coverage is low in this study. The current MS method does not detect amyloid beta, which can be due to its low abundance in large cerebral arteries. In future work a more targeted proteomics method [12] should be used to detect amyloid beta. Additionally, to better understand the correlation between mechanical changes and proteomic changes, analysis of individual layers of the arterial wall may be warranted.

This study integrates advanced optical imaging, mechanical characterization, and mass spectrometry to reveal the structural and mechanical changes in ACAs with AD progression. Progressive media atrophy, elastic fiber degradation, and adventitia structural disorganization were observed with AD development. Results from biaxial inflation extension tests further confirmed the gradual arterial stiffening as AD progresses. Dysregulation in the SMC and ECM gene sets points to reduced SMC activity, aberrant ECM assembly, and the potential inflammatory response to vascular injury in AD cerebrovasculature. Future studies are underway to understand how these cerebrovascular changes correlate with pathological changes in the brain and with AD progression.