Review article
A neurovascular approach to clearing β-amyloid from the brain
2019 Volume 5 Pages 4-11
Details
2019 Volume 5 Pages 4-11
In the brain, arteries and veins do not run in parallel, and its perfusion system has characteristics not found in other organs. The presence of a tight blood-brain barrier and lack of authentic lymphatic vessels in the parenchyma means clearance of some waste products, such as β-amyloid (Aβ), is impeded. Aβ is thus cleared from the brain via at least four clearance pathways including 1) transcytotic delivery, 2) intramural periarterial drainage, 3) glymphatic drainage and 4) enzymatic or glial degradation. Failure in any four such pathways has been implicated in the pathophysiological processes behind Alzheimer’s disease. In clinical trials of Aβ vaccination therapy, vascular Aβ deposition was paradoxically enhanced, with encephalitis subsequently occurring in a fraction of patients. This serious side effect may be associated with insufficient clearance of solubilized Aβ through clearance systems in response to immunotherapy. Transcytotic delivery, intramural periarterial drainage, and glymphatic drainage clearance pathways depend on vascular integrity and are partly driven by vascular wall motion; therefore arteriosclerosis or perfusion pressure reduction is assumed to increase Aβ accumulation. Strategies activating clearance systems may be helpful in the treatment of intractable disease through reduction of brain Aβ, therefore aiding development of neurovascular prevention strategies for Alzheimer’s disease.
In order to maintain homeostasis and nerve function, the brain must manage ionic and amino acid concentrations through three types of strictly regulated barriers: 1) the blood-brain barrier between the blood and interstitial fluid (ISF), 2) the blood-cerebrospinal fluid (CSF) barrier between the blood and cerebroventricular CSF, and 3) the blood-CSF barrier between the blood and subarachnoid CSF [1]. The blood-brain barrier plays a major role, occupying 5000 times the area of the blood-CSF barrier [2]. Since the presence of these functions is essential for controlling ion gradients and neurotransmitter concentrations inside and outside the neuron, such systems likely possessed advantages from an evolutionary natural selection viewpoint. However, the development of these tight barriers has meant ISF containing waste products, such as amyloid β (Aβ), cannot easily return to the vascular lumen, leading to Aβ accumulation in the brain and possibly Alzheimer’s disease (AD). However, to prevent accumulation of self-aggregating and misfolded proteins, including Aβ, several processing mechanisms take place in the brain. In this article, we focus on mechanisms behind cerebral clearance and peripheral metabolism of Aβ and its relationship with dementia causing diseases, in particular AD.
Apart from juvenile AD, the main cause of sporadic AD in the elderly is thought to be a decrease in Aβ clearance [3].
The following four representative mechanisms (4-d) are thought to play a fundamental role in Aβ clearance (Figure 1) [4].
Representative pathways of Aβ clearance pathways from the brain include (1) transcytotic delivery, (2) intramural periarterial drainage (IPAD), (3) glymphatic drainage and (4) enzymatic or glial degradation. These clearance mechanisms seem to work in a complimentary manner to prevent the accumulation of Aβ. However, once the equilibrium is disrupted as a result of aging and/or arteriosclerosis accumulation of Aβ may start in the brain.
(1) Transcytotic delivery – clearance into the lumen of blood vessels by transcytosis through transporters such as lipoprotein receptor related protein-1 (LRP-1) [5] and P glycoprotein (also known as ABCB1) [6]. Aβ is transported by transcytosis from the abluminal (brain parenchyma) side to the luminal side. The receptor for advanced glycation end product (RAGE) transports Aβ from the vascular lumen to the brain parenchyma in an opposite direction to transcytosis and the RAGE inhibitor results in Aβ excretion out of the brain. Despite such theoretical evidence, a phase III trial of the RAGE inhibitor Azeliragon failed to demonstrate efficacy (TTP 488) in mild AD patients.
(2) Perivascular drainage – clearance through the perivascular lymphatic drainage pathway. Here the Aβ-containing ISF flows through the basal membrane layer of the blood vessel and is ultimately transported to the CSF space and cervical lymph nodes [7]. This route runs within the vessel wall and has thus been recently renamed ‘intramural periarterial drainage’ (IPAD) [8]. Since the breakdown of this drainage pathway is related to cerebral amyloid angiopathy pathology, therapies targeting such processes have been developed. Cilostazol, a vasoactive drug phosphodiesterase III inhibitor promotes the removal of Aβ from the vascular wall in a cerebral amyloid angiopathy model mouse [9]. Thus, an investigator-initiated clinical trial, the COMCID study, is currently being conducted in Japan in patients with mild cognitive impairment [10]. (ClinicalTrials.gov Identifier, NCT02491268)
3) Glymphatic drainage – this conduit system was originally reported in 1985 as the ‘paravascular drainage pathway’ [11] but recently renamed due to the pivotal role of glial cells in the system [12]. CSF flows into the brain parenchyma through the Virchow-Robin cavity around the artery and is transported into the parenchyma by the action of aquaporin 4 expressed in the foot processes of astrocytes. In the brain parenchyma, the CSF, together with ISF, dissolve Aβ and drain into the perivenous space [12]. They further drain into cerebral meningeal lymphatic vessels [13]. In Aβ-overexpressing mice, occlusion of the cerebral meningeal lymphatic vessels led to Aβ deposition in the lymphatic vessels, as well as the brain parenchyma, strengthening the importance of these systems in Aβ clearance [14].
4) Enzymatic and glial degradation – this is an Aβ degradation pathway through the action of proteases such as neprilysin, insulin-degrading enzyme, plasmin, angiotensin-converting enzyme and glia (astrocytes and microglia) [15]. Recently, it was reported that microglia expressing triggering receptors expressed on myeloid cells 2 (TREM2) surround senile plaques and have a role in protecting neurons from Aβ, consistent with clinical findings that a loss-of-function mutation of Trem2 was found in AD [16].
The presence of IPAD was suggested by Schwalbe in 1860s, who demonstrated that India ink injected into the cistern could be detected in the cervical lymph nodes within 1 minute [17]. Bradbury and Cserr et al. subsequently reported 1) tracer injected into the cerebral ventricle flowed into the cervical lymph node via the lamina cribrosa, 2) tracer injected into the caudate nucleus flowed into the cervical lymph node unrelated to CSF flow, and 3) radioisotope-labeled tracer injected into the brain parenchyma was observed along the arterial wall intracranially [18]. In the early 1990s, Weller et al. further showed that 1) India ink injected into the striatum of rats was present in close proximity to the dilated perivascular space, along the middle cerebral artery branch and the arterial circle of Willis, reaching the lamina cribrosa along the olfactory and ethmoidal arteries, and 2) further through the lamina cribrosa, extending from the nasal lymphatics to the cervical lymph nodes [19].
As a result of detailed observation by Carare et al., the capillary vessels in the brain and the basement membrane of the artery wall were shown to constitute the IPAD, with the route acting as a high-speed drainage system for ISF [20]. This lymphatic drainage is much faster than the rate at which molecules diffuse through the extracellular space and ISF is almost instantaneously cleared along the arterial wall within the brain and meninges, with a soluble tracer injected into the brain reaching the basement membrane of the meningeal blood vessel within 10 minutes. In a theoretical model [21], arterial pulsation is thought to be the driving force of IPAD, caused by centrifugal force generated by reflected waves following pulse waves originating in the heart.
Since Aβ is detected in the middle cerebral artery and the basilar artery wall but not in the cervical internal carotid artery [22], solutes flowing through the IPAD appear to leave the vessel wall at the base of the brain [7], or via the cerebral meningeal lymphatics [13], into the local lymph nodes.
The glymphatic pathway drainage system is gaining recognition in scientific literature. However, in the schematic diagram of the pathway, there are no barriers in the Virchow-Robin cavity and subarachnoid space, and CSF freely drains into the brain parenchyma [13]. This should be reconsidered as there is principally no space around arteries of the cerebral cortex [8]. Indeed, it is extremely rare to see the Virchow-Robin cavity in the cerebral cortex on MRI. Also, if ISF-containing Aβ is cleared around the vein, it is not clear why cerebral amyloid angiopathy is not observed at this site. In a study injecting fluorescence Aβ into the CSF space of mice at 6-10 months and 24-30 months of age, Aβ was found in the pial-glial basement membrane between the pia matter and astrocytes of the glia limitans, suggesting that CSF enters the brain along the pial-glial basement membrane but not through the Virchow-Robin periarterial space [8]. Nevertheless, one may not entirely exclude the possibility of the glymphatic pathway playing a role in pathological conditions or the aging process. Further elucidation of this pathway may clarify whether the permeability of the pia mater increases and the cortical Virchow-Robin cavity enlarges in the pathological conditions, leading to increased contribution of the glymphatic pathway in the clearance of waste products including Aβ.
Interestingly, IPAD lymphatic drainage aligns with the site of Aβ deposition in early cerebral amyloid angiopathy [23]. Accumulation of Aβ starts from the smooth muscle cell basement membrane, corresponding to the IPAD pathway. Thus, stagnation of this pathway may be closely related to cerebral amyloid angiopathy pathology. Indeed, ISF flow has been impaired in aged mice, as well as those with cerebral amyloid angiopathy [24]. Furthermore, artificial constriction of the bilateral common carotid arteries in a cerebral amyloid angiopathy model mouse results in worsening of the cerebrovascular Aβ deposition [25]. Therefore, a decrease in vascular pulsation due to aging, arteriosclerosis or Aβ deposition itself appears to act as an exacerbating factor of cerebral amyloid angiopathy. When dextran corresponding to the molecular weight of Aβ was injected into the brain and visualized by two-photon microscopy in real time, the efficiency of dextran clearance was shown to be impaired due to vascular occlusion and cerebral amyloid angiopathy [26]. The glymphatic pathway is promoted by vascular pulsation by dobutamine [12]. Furthermore, as levels of neprilysin or LRP-1 decrease, clearance of Aβ is impaired, causing cerebral amyloid angiopathy [4]. These Aβ clearance systems function in a complimentary manner to prevent accumulation of Aβ.
Disease conditions where clearance mechanisms fail and protein accumulates on the blood vessel wall may be collectively called protein-elimination failure arteriopathies (PEFA). Besides Aβ, gelsolin, cystatin C and transthyretin cause cerebral hemorrhage and dementia in familial cerebral amyloid angiopathy, and especially the latter two proteins accumulate in blood vessels of organs other than the brain. Also, in cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), a hereditary vascular dementia caused by Notch3 mutation, NOTCH3 produced from vascular smooth muscle cells accumulates in the blood vessel wall as granular osmiophilic material (GOM) [27], which is also thought to be a type of PEFA. Since GOM is also observed in peripheral blood vessels, including the skin, the breakdown of such a drainage route may be a phenomenon which can also occur in organs other than the brain.
According to the sink hypothesis, Aβ is withdrawn from the brain parenchyma to the luminal space by the action of LRP-1 and P-glycoprotein, which bind to serum proteins such as sRAGE, sLRP-1, and ApoE. More than 60% of Aβ in blood is degraded in the liver: hepatocytes take up more than 90%, and phagocytic Kupffer cells less than 2% [28]. Aβ incorporated into hepatocytes is degraded and metabolized by proteases, such as neprilysin, insulin-degrading enzyme, plasmin and angiotensin-converting enzyme, and finally excreted into the bile (Fig. 2). The medicinal herb Ashwagandha, used In Indian traditional medicine, has been reported to significantly decrease the amount of Aβ in the brain by increasing the expression of LRP-1 and neprilysin in the liver in animal experiments, and may thus possess therapeutic properties to promote Aβ withdrawal [29]. Similar to this mechanism, solanezumab, an antibody against Aβ, forms a complex with Aβ in the blood, resulting in a central-to-peripheral gradient in Aβ concentration, potentially removing Aβ from the brain parenchyma. However, a clinical trial using solanezumab has failed to show efficacy against mild AD [30].
According to the sink hypothesis, Aβ is withdrawn from the brain parenchyma to the vascular lumen by transporters, such as LRP-1 and P glycoprotein (P-gp), and transported by various carriers in the blood. After being transported to the liver, LRP-1 in hepatocytes incorporate Aβ, which is degraded by Aβ-degrading enzymes (sRAGE, sLRP-1, gelsolin, GM1, α2-macroglobulin, and ApoE) and excreted into the biliary system. Therefore, by accelerating such peripheral degradation, the brain-blood equilibrium of Aβ shifts, leading to increased withdrawal of Aβ from the brain parenchyma into the blood. Since RAGE is involved in the transport of Aβ from the vascular lumen to the brain parenchyma, inhibition of RAGE may result in Aβ clearance.
Abbreviations: ApoE, apolipoprotein E; GM1, ganglioside GM1; LRP-1, low density lipoprotein receptor-related protein 1; P-gp, P-glycolipid; RAGE, receptor for advanced glycation end product; sLRP-1, soluble low density lipoprotein receptor-related protein 1
This article has outlined the mechanisms behind Aβ clearance. When vascular degeneration occurs with aging and/or cerebral amyloid angiopathy, Aβ clearance is impaired and accumulation accelerated. Therefore, prevention of vascular degeneration is also important for AD treatment. In the past, cerebrovascular disease was generally accepted to be distinct from AD resulting from a neurodegenerative process. However, it is clear that this simple dichotomy needs revision in light of the apparent vessel-dependent process of Aβ clearance. Therefore, further elucidation of the physiologically and pathologically important Aβ clearance mechanism is warranted. AD research, when viewed from a neurodegenerative standpoint, has made considerable breakthroughs. However, it is apparent that this focus only represents part of the disease trajectory, which may explain the lack of efficacy in recent clinical trials. It is our hope that a more vascular approach may offer new avenues and treatments for AD in the near future.
The author would like to thank Dr. Ahmad Khundakar (Teesside University, UK) for his editorial assistance.
