Progression Of Alzheimers Disease Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

The blood-brain barrier is involved in a wide variety of functions, one of them being the clearance of amyloid-beta peptide (38-42 amino acids in length) where Aβ (1-40) is the dominanting molecular sequence of cerebrovascular amyloid and Aβ (1-42) is the major molecular component of senile plaques [19]. It is involved with the initiation and progression of Alzheimer's disease (AD). This accumulation of Aβ is the setting for the 'amyloid cascade hypothesis' [1], which holds Aβ as the cause for senile plaques and neuron cell death [1]-detrimental to the functioning system of the BBB and thus initiating AD in the brains of late-onset AD patients. As Aβ levels increase, they are deposited on cerebral blood vessels, decreasing cerebral blood flow (CBF) and hence give rise to cerebral hypoperfusion, affecting the production of proteins required for memory, learning and cognitive function [2] and in so allowing AD to progress. This accumulation of Aβ within the walls of the leptomeningeal and parenchymal blood vessels, which lead to thickened arterial walls and microaneursym formations, is termed cerebral amyloid angiopathy (CAA), and it is associated with smooth muscle cell degeneration, contributing to the death of the endothelium making up the BBB. [1]

To prevent this build-up of Aβ within the brain, there are a multitude of receptors along the BBB that control or influence this influx and efflux of A. This essay will focus on the balance between low-density lipoprotein receptor related protein-1 (LRP) and P-glycoprotein (Pgp) with their roles as major efflux receptors of Aβ, and the receptor for advanced glycation endproducts (RAGE), the key receptor for Aβ influx, and how their interactions of Aβ entry and clearance through the BBB is cause for AD initiation and progression.


The BBB is a barrier formed by tight junctions (zonulae occludentes) of the endothelial cells of cerebral capillaries. These tight junctions, being exceptionally tight, are present only within the BBB. Its role is to regulate ion movement, as a stable environment is required for neuronal cells to function and for synaptic signals to be efficient. This is attained through the use of specific ion channels and transporters within the BBB. [3]

Transport across the BBB is mediated through five pathways. The first is passive diffusion, which allows lipid soluble non-polar molecules to diffuse through .Next are active efflux transporters, or ABC transporter efflux, which intercepts some of the passively diffused molecules and pump them out through either the luminal, or abluminal membrane of the endothelial cell. Solute carriers transport many polar molecules and may be bi-directional. Mononuclear cell migration allows leukocytes to enter the BBB through the endothelium via diapedesis, a widening of the tight junctions. Last, is receptor and adsorptive mediated transcytosis (RMT and AMT respectively); AMT is induced by a positively charged macromolecule upon the receptor and transporting it across the BBB vesicle bound. RMT operates through a ligand-receptor mechanism and its importance is the influx of Aβ via LRP-1 and the efflux of it by RAGE while Pgp operates through ABC transporters. [3]

Aβ is produced by neurons during 'interstitial fluid drainage' [4] as hypothesised by the vascular theory [19], giving rise to the amyloid precursor protein (APP) of which a 4-kD fragment of cell surface APP is the major contributor of senile plaque [29]; its preteolytic by-product formed via the amyloidogenic pathway [15] from the cleavage of β and γ-secratase is Aβ [2], and Aβ in the brain. [1] It is also produced within the circulation, where it crosses the BBB into the brain. [19] Soon after being produced, Aβ is directed toward pericytes in capillaries and smooth muscle cells before RMT through action of LRP-1 clears it from the brain as a soluble peptide. [1, 4] Aβ is also suggested to be produced by the bloodstream as RMT via LRP-1 is bi-directional. [3] The role of the ABC transporters is that of active efflux receptors, transporting lipid-soluble compounds out of the endothelium and CNS thus having a detoxifying purpose, ensuring that the neural environment is protected. Pgp as such, is found on the luminal surface of the BBB to perform this function. [3]

Build up of Aβ via RAGE and inefficient clearing can cause neuroinflammation via the activating factor-κB, secreted by proinflammatory cytokines such as tumour necrosis factor -α and interleukin 6 [2], which increase Aβ levels in the brain via more influx, less efflux and a higher Aβ production rate. [6] Thus the BBB efficiency is reduced, and neuronal functioning relating to cognitive function and memory are hampered due to ionic charge irregularities within the cerebral environment.

AD feature neurofibrillary tangles made of hyper-phosphorylated microtubule-associated protein termed tau as well as extracellular amyloid plaques which consist of mainly Aβ peptides in the brain parenchyma and blood vessel walls [1] resulting in neuronal degeneration. [2, 5] The focus of this essay is therefore not in the accumulation of Aβ but in its clearance and how a defect in clearing processes of the BBB by LRP-1 and Pgp will result in the formation of amyloid plaques reminiscent of AD.


RAGE is a transmembrane receptor of the immunoglobulin (Ig) superfamily which responds to a multitude of ligands, one of which is soluble Aβ. Interactions between RAGE and Aβ activate RMT pathways, inducing pathologically relevant cellular [8] effects through a prolonged state of perturbation resulting in a range of chronic diseases [10], of which AD is included. [7-10] Human RAGE protein is comprised of six distinct domains and as an unprocessed precursor, is 404 amino acids (aa) long in length, with a molecular weight of 42,803 Da: an extracellular region (aa 1-342) consisting of a signal peptide (aa 1-22), three Ig-like regions- an Ig like V-type domain which contains one N-terminal acting as the ligand binding site (aa 23-116), and two Ig-like C2-type ½ domains (aa 123-221). Following this is a single transmembrane region (aa 343-363) and lastly, a short but highly charged cytoplasmic tail (aa 364-404). [7] There is only one type of RAGE gene in humans, located in chromosome 6p21.3 [20] near the major histocompatibility complex, but there are several alternative forms derived from splicing of the full RAGE sequence which have a role in regulating ligand access to RAGE for RMT. [11]

CNS Aβ levels increase due to the action of RAGE influx through the BBB. Evidence for this was obtained through experimentation on transgenic APP mice. Mice were infused with a radiolabeled peptide containing the first 40 aa of Aβ at 4.5 nM [9], a pathophysiologically relevant concentration. These mice showed significant uptake through RAGE but when RAGE specific IgG was used, transport of Aβ through the BBB was inhibited. [9] RAGE is the dominant factor influencing influx of Aβ. The transport model, which has concluded, that in a healthy human brain, if efflux via LRP-1 were to be completely inhibited, RAGE influx alone would cause all soluble Aβ in the brain interstitial fluid (ISF) to be replaced by plasma Aβ in just approximately 40 minutes. [8] The amount of Aβ transcytosed in the presence of unlabelled competitors was recorded to determine the surface area of cerebrovascular permeability for Aβ. This was done so as to confirm the transport of human Aβ across the mouse BBB by RAGE [9] as competing RAGE-specific Ig-G and soluble RAGE (sRAGE) impaired its activity which would have resulted in undetectable transport levels in the homozygous RAGE-null mice. [11]

A more detailed experiment was performed on human brain microvascular endothelial cells (HBMEC) to investigate Aβ uptake by RAGE. HBMEC of 5 day old primary cultures were analysed at 37 °C which showed radiolabelled I-sAβ (1-40) binding to only the apical surfaces of the cells. The rate of binding was rapidly increased within the first 5 minutes and steadied from then till 60 minutes. After binding, 75-80% of the bound I-sAβ could not be removed by acid wash. This value was significantly reduced when repeated at 4°C thus showing Aβ binding to be time, and temperature-dependent. To further test this, the cell surfaces were rinsed with trypsin:proteinase K:EDTA to clear away any ligand-bound complexes attached to the surface but the difference in results using acid wash and K:EDTA were insignificant, clearly suggesting uptake into the apical membrane and its occurrence at physiological temperatures. I-sAβ bound to the apical side became saturated upon introduction of increasing amounts of the competitive unlabelled sAβ(1-40) (Fig 1a) where Scatchard analysis pointed out 2 independent sAβ (1-40) sites of different affinities to Aβ (Fig 1A). Endocytotic inhibitor PAO, 20 µM concentration, reduced I-sAβ (1-40) uptake into HBMECs. (Fig 2) C-inulin was used as a control in all experiments to monitor extracellular leakage and its rate of leakage was able to cancel out the basolateral apical transport of I-sAβ (1-40). Transendothelial transport of I-sAβ at 37°C was markedly higher than that of C-inulin (Fig 3). [19] Upon introduction of anti-RAGE antibody (20 µg/ml), 63% of I-sAβ were inhibited from binding at the apical side of the cell (Fig 4a). At physiological temperature (37°C) and the maximum inhibitory concentration of anti-RAGE antibody which was 20 µg/ml, apical basolateral movement of Aβ was inhibited by 36% (Fig 4b). This study showed that sAβ is readily bound to and undergoes uptake at the apical side of the HBMEC. [19]

Figures from [19]

The involvement of RAGE overexpression at the BBB with the internalisation of Aβ was studied again by using transgenic models which had overexpressed mutant human amloid precursor protein (mAPP) and/or presenilins. As a high percentage of the transgenic models showed subtle changes in neuronal function and pathology, analysis of RAGE began by developing the transgenic mice with an overexpression of wild type (wt) RAGE generating transgenic (Tg) RAGE mice. To observe the influence of RAGE in an Aβ-rich environment, Tg RAGE mice were crossed with animals expressing mAPP, producing 4 genotypes of single Tgs (Tg RAGE and Tg mAPP), double Tgs (Tg mAPP/RAGE) and nonTg controls. RAGE expression was determined via use of immunoblotting of the cerebral cortex with an antibody. RAGE expression was found to be lowest in the control group; the Tg mAPP mice had increased RAGE antigen levels; Tg RAGE mice had even higher levels than Tg mAPP mice and Tg mAPP/RAGE mice had a figure three times higher than that of Tg mAPP. Double Tg mice gave the same readings as Tg RAGE mice. This suggested that RAGE expression is enhanced in conditions where ligand levels such as Aβ are increased. [22] This means that high Aβ levels, produced within the circulation, do not have a fixed ratio of RAGE. Expression of RAGE is increased in the presence of high Aβ levels: overexpression of wt RAGE transgene in AD-type mice also expressing mutant human APP led to an increased level of Aβ-mediated pathological cellular effects of the neurons, causing difficulties in spatial learning, memory and severe neuropathologic changes. These signs were not seen in single transgenics, thus supporting the idea that RAGE is indeed an integral part of Aβ-induced neuronal perturbation in AD. [10] The RAGE expression enhancement comes about from a shift in localisation of RAGE from neurons to the BBB, thus allowing for an increase of Aβ into the interstitial spaces of the brain. [23] Thus the role of the BBB in the overexpression of RAGE is that of allowing for the increase Aβ influx perhaps through secondary messengers via a pathological incident, which thus causes AD to be initiated and progress rapidly.

Soluble RAGE, or RAGE_v1, is a variant of RAGE, with an inclusion of intron 9 and a deletion of exon 10, giving rise to a reading frameshift at position 332aa, creating a distinct C-terminus sequence. [7] It can be secreted as a 'soluble' form from cells in comparison to other variant soluble forms, RAGE_v3, v7 and v9, which are found only in cell lysates. [7] This isoform circulates in the plasma in the body and in particular, along the BBB, as the usual transmembrane area bounding it to cell surfaces is missing [7], allowing it to compete with RAGE due to it not being able to enter the CNS [11] and in doing so, hampers the progression of AD by reducing Aβ influx.

Using transgenic mice with overexpressed mutated human APP encoded by a minigene containing familial AD substitutions controlled by the PDGF B-chain promoter (PD-hAPP) [9], and controls from the same litter, a brain perfusion of I-labelled sRAGE and [99Tc] albumin was performed into their cerebral circulation. This was performed on 6-month old mice at 100µg/day of sRAGE for 3 months. Results showed that the controls had increased levels of hippocampal Aβ compared to the PD-hAPP mice after 9 months (P< 0.001) by 78% (total Aβ) and 72% (Aβ 1-42), (Fig 5a, 5b). Introducing sRAGE to the controls resulted in a 78% decrease in Aβ deposits. Serum analysis of sRAGE mice detected Aβ-sRAGE complexes. RAGE-specific IgG immunoprecipitation followed by immunoclotting with Aβ-specific IgG showed an Aβ-immunoreactive band of 4 kDa. However when the process was done with RAGE-specific IgG immunoblotting followed by Aβ-specific IgG immunoprecipitation, the result gave a RAGE-immunoreactive band of 40 kDa. Plasma Aβ analysis by ELISA at 9 months concluded that Aβ (1-40) and Aβ (1-42) were significantly higher (P<0.001) in PD-hAPP mice treated with sRAGE than with the controls, [9] thus proving that sRAGE decreases Aβ transcytosis. The influence sRAGE had on AD was projected onto Kaplan-Meier survival curves in groups of patients segregated on the context of sRAGE sixtiles [21] to determine the significance of sRAGE as a marker. With the lowest sRAGE level, only 12.5% of patients reached 65 years of age without AD symptoms compared to 57.1% from the higher sRAGE levels. Highest sRAGE levels were found in healthy controls, while the lowest were found in AD groups [21], corresponding to the cross-sectional study of 404 patients where the 152 patients with clinical AD had the lowest levels of sRAGE, followed by the vascular dementia patients and finally the controls. [31] Hence, although not immediately part of the BBB, sRAGE is involved with Aβ deposition and influx through the BBB of which its precursor RAGE is part of, influencing the mechanisms of AD.

One of the polymorphic sites of the RAGE gene, G82S, located at exon 3, replaces glycine with serine at codon 82 within the ligand-binding domain of the receptor. Endothelial cells expressing this 82S variant of RAGE have an increased affinity for ligand binding. Thus a genotype distribution and allele frequencies for the polymorphism in AD patients and controls was set up in Hardy-Weinberg equilibrium, which showed the frequency of the 82S allele dominating in AD patients. The relationship between G82S and sRAGE is one where plasma sRAGE levels were found to be significantly lower in 82S allele carriers (GS+SS) of the AD group. The fastest progress was found in patients with GS+SS genotype (RAGE 82) at 52.27% deterioration while the slowest had a reading of 29.73% (P= 0.0298). [20]

Aβ -RAGE interaction at the BBB also enhances the expression of the vasoconstrictor endothelin-1 (ET-1) which influences CBF [9], reducing pulsatile flow along the BBB, increasing Aβ deposition onto the arterial walls of AD patients, allowing for a higher influx rate into the CNS by RAGE. Introducing Aβ into the cerebral vasculature caused an increase in immunocytochemically detectable levels of ET-1. [9, 11] This effect was inhibited when RAGE-specific IgG was used instead but not so with nonimmune IgG. Aβ-induced reaction of CBF was significantly suppressed upon introduction of the ET-1A receptor blocker, BQ610. None of the other blockers- ET-1A blocker BQ788 and ET-1A antagonist had the same suppressing effect on Aβ-induced CBF. [9]

In the case of RAGE-null mice, infusing Aβ into their cerebral vasculature did not have any impact concerning the levels of immunoreactive ET-1 compared with the controls. In tissue culture, Aβ introduced to mice brain endothelial cells showed an increased expression of preproendothelin-1 (PPET-1) mRNA in a RAGE-dependent manner.[11] This was observed in cells exposed to the Aβ after 15 minutes and mRNA levels remained high at 60 minutes. Use of RAGE-specific IgG or sRAGE prevented PPET-1 increase but not nonimmune IgG. [9]

With Aβ-RAGE interaction mediated by ET-1, cerebral blood vessels are exposed to constriction and stiffening, due to Aβ buildup on vessel walls. RAGE is an integral factor of Aβ influx through the BBB by RMT, with non-activating variants such as sRAGE, competing with RAGE for ligand binding, acting as a decoy and with the G82S variant, facilitating Aβ binding. It also contributes to CBF reduction, which ultimately leads to the accumulation of Aβ in the CNS. These factors allow the formation of senile plaques, neuronal death and AD progression. There is some evidence to show that RAGE decreases the degrading of Aβ through enzymatic activity. Reductions of SDS and formic-acid souble Aβ in the brains of 6 month old mice where they had RAGE knockout and Swedish and Arctic Aβ precursor protein mutations (arcAβ), suggested an increased activity of Aβ degrading proteases. At 6 months, enzymatic degrading activity was significantly higher in RAGE knockout arcAβ mice than in arcAβ mice and the wt animals. At 12 months, the results were not statistically significant except when in comparison to the wt animals (p= 0.043). [24] This may suggest that increased RAGE expression could result in less Aβ clearance and contribute to AD. However, to prevent the initiation of AD, there must be better effective clearance routes, such as that of the main efflux receptor of the low-density lipoprotein receptor family (LDLR), lipoprotein receptor-1 (LPR-1).

To demonstrate the rate of efflux, in vivo brain-to-blood efflux transport of Aβ (1-40) was measured via the BEI (brain efflux index) method using [3H]dextran as a reference compound. [12] A solution of radiolabelled Aβ with a concentration 18.2 nM, was injected into a rat brain which had a Aβ cerebral concentration of 0.6 nM. The Aβ was eliminated in a time-dependent manner of 60 minutes and the information gathered together with elimination rate constants and distribution volume from a brain slice uptake study derived an efflux rate of 11.0 +/- 3.4 µL/(min g brain). Efflux rates of Aβ measured in this study was calculated to be 26.9-99.0 times greater than influx rates, strongly implying a net movement of Aβ out through the BBB to protect the neuronal environment. [12] The transport model calculated that should influx be completely inhibited, LRP efflux could remove all soluble Aβ from ISF in 1 minute [8].

There are two physiological pathways of Aβ efflux. One is directly into the bloodstream through mainly receptors such as LRP-1, where it is endocytosed via RMT and degraded [14] within the endothelial cells and Pgp where Aβ influx and efflux are modulated and will be discussed later. The other is by ISF bulk flow into the CSF and cervical lymphatics. [13] It is noted that ISP bulk flow accounts for only 10-15% of Aβ efflux out of the CNS, thus it is LRP-1 that is the main route for Aβ efflux. [14] These two pathways were compared with one another on their rate of efflux- quick efflux by vascular transporters, mainly LRP-1, through the BBB into the blood and a much slower efflux rate by ISF flow. [13]

Western blot analysis was used to locate the expression of LRP-1 in rat brain and in its capillary-rich fractions along the BBB [12] by using anti-LRP-1 antibody, which binds to the β-chain of LRP-1. LRP-1 is an approximately 600 kDa cell surface receptor with an affinity for multiple ligands and is highly expressed in the brain. [14] Its extracellular ligand binding regions are rich in cysteine ligand-binding repeats, grouped in to four clusters of 2,8,10 and 11 repeats (domains 1-4) [14, 15], with a large proportion of ligands, particularly Aβ, binding to domains 2 and 4. LRP-1 is a polypeptide cleaved in the trans-Golgi network into 515 and 85 kDa subunits. [15] The 515 long aa terminal subunit allows ligands to bind, while the 85kDa subunit binds it to the cell membrane through noncovalent association. Also, LRP-1 is susceptible to binding to receptor-associated protein (RAP), a 39 kDa chaperone protein [17], which ensures correct protein folding, but also acts as an antagonist to potentially binding ligands which may disrupt the usual activity of LRP-1.[15, 26] Other ligands in interest are apo-lipoprotein E (ApoE) and α2-macroglobulin (α2m), which are associated with LRP-1 binding to facilitate Aβ efflux out of the BBB.

Using TR-BBB cells- a specially preserved endothelial cell line from rat brain [16], it was proven that Aβ was uptaken by TR-BBB cells. The study first tested if radiolabelled Aβ (1-40) would form oligomers at the concentration, which was 0.1 nM, where it would be used at, with a light-induce chemical cross-linking experiment- PICUP[16], determined by immunoblot analysis. The analysis showed no oligomers to be present, allowing the specific Aβ study concentration to be used. First, the cells were incubated with radiolabelled Aβ for 5 minutes at 37°C and then washed with acid buffer to remove any cell surface Aβ. The amount of Aβ that was uptaken by the TR-BBB cells was shown to be very rapid, steadying at 8 minutes, and was calculated as a cell/medium ratio (µl/mg) as seen in Fig 10.

Insert figure 10 here from [16]

To prove LRP-1 was integral in uptake of Aβ, the same experiment was conducted again but this time in the presence of RAP, an LRP-1 antagonist. The experiment yielded a 90% inhibitory effect here. A separate experiment concluded that human RAP at 5µM redued radiolabelled Aβ (1-40) by 20.3% [12]. An experiment in the TR-BBB scope was conducted using a neutralising antibody against LRP-1 -RRR, and co-incubation produced a 69% inhibition rate against Aβ uptake by TR-BBB cells. LRP-1's involvement was tested again by LRP-1 knockdown in TR-BBB cells by siRNA treatment using two stealth RNAis. This was a chemically modified RNA molecule that avoided the interfeuron pathway, against LRP-1 and a stealth negative control RNA. [16] TR-BBB cells were cultured with siRNAs (1 and 2) against LRP1 for 30 hours whereupon LRP-1 expression levels were reduced by the siRNAs by 34 and 29% respectively in comparison to the control siRNA. (Fig 11a) When the siRNA-treated and control cells were incubated with radiolabelled Aβ, it was determined that uptake of AB by TR-BBB cells was much reduced in the siRNA cultures by 38% in siRNA1 and 52% in siRNA2 and that Aβ binding to LRP-1 was also significantly lessened. [16]

To show that a lack of LRP-1 would cause a decrease in Aβ efflux, cell lines termed L1 (LRP-deficient), L2 (LRP deficient but with overexpression of APP) and L3 (expressing LRP and overexpression of APP), were introduced with 500 µg of RAP to inhibit the LRP. This was to allow the non-amyloid-ogenic pathway, where APP is cleaved by α-secretase, preventing Aβ from forming by way of producing soluble APP. [32] The cell lines were then seeded to a specific density low enough to ensure proper growth rates in the presence of RAP. A day before collection, the cell lines were washed and cultured for 24 hours in a serum-free, nutridoma-containing medium. A deficient of LRP showed a decrease in Aβ efflux in all cell lines upon RAP treatment, with the decrease most dramatic in L3. [26] Thus, with RAP treatment inhibiting LRP-1, there is a significant decrease in Aβ efflux.

As seen above, LRP-1 is a definite requirement in the internalisation of Aβ into brain capillary endothelial cells, and it is done so through the process of RMT, facilitated by ligands ApoE or α2m.[27] To prove uptake via the α2m-LRP pathway, the uptake of synthetic radiolabelled Aβ in cultured fibroblasts was studied. Using a fixed concentration of 0.1 nM of Aβ, the optimum ratio of radiolabelled α2m/Aβ was observed. The Aβ was incubated with different concentrations of α2m of 0, 1 ,2, and 3 nM at 37°C overnight and later mixed with LRP+/- and LRP-/- fibroblast cells for 24 hours. The resulting medium was then analysed for γ radiation via scintillation counting. The number of counts lost from the original input was taken to be the percentage of Aβ internalised. [18] Higher concentrations of α2m gave increased Aβ uptake results. There was insignificant uptake in LRP-/- cells at all α2m concentrations. Increased concentrations of unlabelled Aβ-α2m complexes competed with the radiolabelled Aβ at a concentration-dependent rate, inhibiting its removal. Also, the uptake of the radiolabelled Aβ-α2m complex was fully saturated with half-maximal-uptake at 50pM of the complex, an estimated Kd range, from 0.2 -10, noted for α2m binding to various cell types. [19] Thus a receptor -mediated process is proven by firstly, the competitiveness between the unlabelled and radiolabelled Aβ-α2m complexes and secondly, by the saturation of radiolabelled-Aβ uptake at physiological levels. [18]

As mentioned, ApoE is another ligand that binds to LRP-1. This interaction affects the movement of Aβ out of the BBB by the indirect binding through a chaperone protein- APoE. ApoE is a 34 kDa reactive lipoprotein that is largely involved with the transport of cholesterol and lipids in plasma. Human ApoE is 299 aa residues in length and has independently folded amino-terminal and carboxy-terminal domains [25], connected by a flexible hinge region. There are 3 isoforms of ApoE- differentiated by single amino acids at aa positions 112 and 158, termed ApoE2, ApoE3 and ApoE4. 40-60% of AD patients are suspect of carrying at least one ApoE4 allele, while ApoE4 homozygosity increases the risk of AD from 20 to 90%. However, ApoE2 binds to the Aβ C-terminal region, causing the interaction between Aβ and the cell membrane to be disrupted, reducing its fusogenic properties allowing it to also form sodium dodecyl sulphate-stable complexes with Aβ and hence facilitate its clearance through LRP-1 [5]. ApoE2 and E3-Aβ complexes are thus cleared at a significantly faster rate than ApoE4-Aβ complexes [25], explaining why having an E2 allele reduces the risk on AD onset.

To prove that ApoE2-Aβ had a protective function, tests were conducted on rat neuromicrovascular endothelial cells (NECs). NECs were cultured for 24 hours and then introduced to Aβ (1-40) or Aβ (1-42) of 10-7 M alone, or with ApoE2 or E4 of 0.2 µg/ml concentration. NECs were left to culture for another 3, 24 and 48 hours, during which trypan blue exclusion was used to determine cell survival. Cells that uptook the ApoE-Aβ complex through the endothelium hence became stained.

Aβ peptides decreased cell survival after incubating for 1 and 3 hours but not 24 or 48 hours. Treatment with ApoE2 resulted in a decrease in cell survival at 1 hour but this increased at 3 hours with respect to the Aβ peptide values. NECs were not affected by ApoE2 after 24 hours but with reference to Aβ (1-42), there was a significant decrease in cell survival. Treatment with ApoE4 however, resulted in significant reductions in NEC survivals in all time cultures. NEC death, measured by lactate dehydrogenase (LDH), was notably increased for both Aβ peptides with ApoE2 preventing necrosis at 24 hours but not 48, and ApoE4 increasing necrosis in all the time cultures. [5] Hence ApoE2 showed itself to have an inhibitory effect on Aβ cytotoxicity on NECS, providing a route of faster clearance through LRP-1, while ApoE4 had a long lasting facilitating effect on Aβ cytotoxicity, slowing down Aβ clearance. Another similar experiment testing the two ligands showed similar results. Incubation with of 2.5nmol/L of LRP-1 with α2m, ApoE2, E3 and E4 for 3 hours at 4°C showed the BEI value of 25 nmol/L of Aβ(1-40) to be significantly reduced by 11.7%, 15.5%, 16.8% and 19.3% respectively. [27] Thus, ApoE significantly influences Aβ accumulation through LRP-1 and hence, AD pathogenesis.

LDLR overexpression would so cause a decrease in APoE levels, an increase in Aβ efflux and deposition. Tg mice overexpressing LRP-1 and 2 tg lines with the APP/PS1, PS1 being another gene along with APP that majorly influence Aβ accumulation, were created to test LDLR overexpression, of which LRP-1 will be constituting a major part, and how this would affect Aβ clearance and deposition. ApoE is one of the main ligands for LDLR and thus overexpression of LDLR was tested with ApoE levels in the brain with ApoE enzyme-linked immunosorbent assay (ELISA). The result showed a significant reduction in ApoE levels by 50-90% and that twice the amount of LDLR overexpression decreased APoE levels by an average of 50% while five times of overexpression caused a massive 80-90% reduction in ApoE levels. Thus higher levels of LDLR in the form of LRP-1, speeded up ApoE endocytosis, of which ApoE4, the main concern, would be binded to Aβ, thus facilitating Aβ efflux. This leads to its effect on Aβ deposition. APP/PS1 and APP/PS1/LDLR 7 month old mice had brains sections immunstained with biotinylated-3D6 antibody, an anti Aβ, to prove this influence. The APP/PS1/LDLR mice had significantly reduced Aβ deposition in the cortex and hippocampus compared with the App/PS1 Tg mice after use of anti-Aβ immunostaining. The brain sections were next stained with X-34 dye to determine Aβ deposits. APP/PS1/LDLR tg mice were observed to have a 40-70% decrease in the X-34 positive Aβ plaque load of the cortex and hippocampus in comparison to the APP/PS1 tg mice. Biochemical analysis also deduced insoluble Aβ(1-40) levels to drop by 50-75%, and insoluble Aβ(1-42) in the cortex and hippocampus to drop by 45-70% in the brains of the APP/PS1/LDLR tg mice. Results thus proved that higher levels of overexpression (10 fold) of LDLR caused a significant decrease in Aβ deposition and an increase in efflux. Lower levels of overexpression, two fold, also resulted in an inhibition of Aβ formation. Using the same processes as above, Aβ deposition was much lower in APP/PS1/LDLR tg mice compared to the APP/PS1 mice. APP/PS1/LDLR tg mice were observed to have a 30-55% reduction in total Aβ(1-40) levels and a 35% decrease in total Aβ91-42) levels. Thus LDLR, with its relation LRP-1, is effective in reducing Aβ deposition, formation, and increasing efflux activity through chaperone proteins APoE. [28]

The extracellular N-terminal of LRP-1 of where the binding of ligands occur at can be cleaved by β-secratase to produce soluble LRP (sLRP). Like sRAGE, it is not bound to membrane surfaces and thus circulates in the plasma, influencing Aβ efflux along the BBB. [30] Patients with AD had a 30% reduced sLRP compared to the healthy controls. Native sLRP shows significant binding affinity to Aβ in human plasma. This evidence was provided after coimmunoprecipitation of sLRP-bound Aβ showed 70% of Aβ (1-40) and 90% of Aβ (1-42) bound to plasma sLRP in controls. In AD patients however, a 30-35% drop was noticed in sLRP-bound Aβ peptides and 'a 300-400% increase in free, protein-bound' Aβ peptides. Thus, the conclusion was that native sLRP normally binds to 70-90% of Aβ in human plasma, creating an endogenous sink activity which may facilitate clearance of Aβ along the BBB. [30]

LRP-1 operates through the process of RMT and its activity of Aβ efflux is greatly influenced by the binding of α2m, ApoE- mainly the alleles ApoE2 and ApoE4, which have protective and magnifying cytotoxic Aβ effects respectively, and RAP, which acts as an antagonist to LRP-1, allowing proteins to fold properly, but in so acts as an antagonist to Aβ binding hence reducing the total amount of Aβ that can be bound and cleared out through the BBB. Its variant, sLRP, acts as a peripheral binding agent but however is still very significant in facilitating Aβ clearance along the BBB, counting for a large percentage of bound Aβ. LRP-1 activity of efflux is one of the key main receptors that are responsible for the degradation of neuronal cells and the onset of AD. However, the build up of Aβ to AD developing requires more than just one effective clearing route through the BBB. Another clearance method via the BBB is through that of the ABC transporters, primarily, Pgp.

Pgp is a 170-180 kDa plasma membrane-associated protein produced by the multidrug resistance (MDR) gene. Pgp and multi-resistance drug proteins (MRP) belong to the ABC transporter superfamily and they prevent a variety of potentially harmful endogenous and exogenous compounds, and certain kinds of drug treatments, from entering the brain by pumping them out via active transport from within the BBB. [33, 34] There are two MDR genes that encode Pgp in humans- MDR1 and MDR2, with MDR1 the typical phenotype. Hence, Pgp is the product of the MDR1 gene. Pgp's structure is that of two homologous halves with each comprising of a single transmembrane domain of six helices and a nucleotide binding domain that encodes an ATP (adenotriphosphate) binding site. The mechanism of Pgp is still unknown although it is highly expressed along the luminal surfaces and contributes to a majority of the BBB by limiting the movement of substrates into the CNS. [36] However, the substrates and inhibitors of Pgp include a large variety, including that of Aβ. Despite Pgp's lack of concrete evidence on its mechanism and location being a new area of study, it has been shown that a deficiency in Pgp at the BBB does cause an increase in Aβ efflux. [35]

Using HEK293 cells which were transfected with APP695, and then tranfecting them another time but with human MDR (pHaMDR1/A) by use of the calcium phosphate method, interactions between Aβ and MDR1 were observed. Cells that were transfected with the human MDR showed increased Aβ secretion in comparison to the controls which had no human MDR. To confirm this observation, a different strain of APP- APP695sw was used which showed Aβ secretions increasing as well. However, MDR1 transfections did not cause an increase in expression of MDR1 and APP cellular levels were increased in response to the extracellular Aβ levels. This change in APP meant Pgp would have to be investigated in other ways although a basic relation between Aβ and Pgp through MDR1 was indicated here.

K269sw cells that were transfected with pHaMDR1/A and MDR1 inhibitors RU486 and RU49953 showed a significant decrease in Aβ secretion in comparison to vehicle treated cells after a 15 minute drug exposure. This time period was set to avoid any RU486 effects on glucocorticoid or progesterone receptors. Phorbol ester (PMA) was used as a positive control as it decreased Aβ secretion through secondary messenger cascades. RU486's introduction caused Aβ levels to fall with an EC50 of 10 nm with maximum inhibition decreasing as the drug concentration was entering the micromolar range. RU49953, which did not have steroid receptor binding affinity, reduced Aβ secretion with an EC50 of approximately 1 nm. (Fig 2)

Insert figure from 35

It is now clear that MDR1 is involved with Aβ efflux, and so the relationship between Pgp and Aβ was tested. Aβ binding to MDR1 was tested using highly purified vesicle bound hamster MDR1. Fluorescent quenching was used to detect Aβ binding to MDR1 at the two cysteine rich residues within the nucleotide binding domain after titration of MIANS-labelled MDR1 with synthetic Aβ (1-40) and Aβ (1-42) peptides. A saturable quenching of the fluorescent suggested a direct interaction between the Aβ peptides and transporter. The binding affinities calculated from two different MDR1 batches were 12.5 ± 1.0 µm and 6.7± 1.0 µm which corresponded to that of Aβ (1-40) and Aβ (1-42), amongst other different types of substrates suggesting Aβ as a substrate of MDR1. This was proven by testing Aβ's ability to affect ATP-dependent uptake of [3H]-colchicine into plasma membrane vesicles. The results were that both the Aβ peptides were effective substrates, marking a concentration, for 50% inhibition of the colchicine uptake, of 27µm for Aβ (1-40) and 22µm for Aβ (1-42). [35]

Aβ also acts as a stimulator of MDR1 ATPase activity- which affects MDR1's transporting ability. When Aβ peptides were introduced to plasma membrane vesicles derived from CHRB30 cells and the ATPase activity measured, it was observed that both Aβ peptides increased ATPase activity- with Aβ (1-40) increasing it by 100% at 50µm concentration and Aβ (1-42) increasing it by an estimated 40% at a concentration of 50µm. Thus, these results show Aβ to be a true substrate of MDR1. [35]

To confirm this evidence, Pgp's direct role in Aβ transport was observed. Radiolabelled Aβ peptides IAβ40 and IAβ42, Aβ(1-40) and (1-42) respectively, were microinjected into mice brains where Pgp was encoded by mdr1a and 1b. The mice used were mdr1a/b-/- double knockout (Pgp-null) and the wt controls. Radiolabelled inulin was used to measure bulk flow of ISF in all the mice. 30 minutes later, the mice brains were analysed. Levels of IAβ40 and 42 were significantly higher in Pgp-null mice than in the controls, thus suggesting Aβ efflux was impaired. ISF flows accounted for 20% of Aβ clearing, and was unchanged in Pgp-null mice, with similar readings across both groups, thus pinpointing Pgp. 31% of IAβ40 was cleared in the wt controls compared to 14% in the Pgp-null group. 16% of IAβ42 was cleared in the wt mice compared to 6% in the Pgp-null mice. This indicated that Aβ40 was more involved in BBB transport than Aβ42. [36] Also, LRP-1 levels dropped by 50% in the brain capillaries of 2 to 3 month old Pgp-null mice compared to the wt, suggesting an association between Pgp and LRP-1, perhaps through a secondary messenger system, or that Pgp is involved with LRP-1 structure development, aiding its process, playing a synergistic part. [36]

With relevance to AD, Aβ depostition was monitored at the BBB but in the absence of Pgp. APPsw+/- mice were bred with mdr1a/b-/- ones to produce mice that were positive for APPsw+/- and either Pgp-null or Pgp-wt. To prevent the mutated APP from affecting Pgp basal function in vivo, [99mTc]Sestamibi- a substrate of MDRI Pgp that can detect transport processes on Pgp [37], was introduced to both Pgp-null mice and 2-4 month old wt mice, in the presence and absence of the mutated APP gene. [99mTc]Sestamibi is usually transported out of the brain, but in Pgp-null mice, its levels in the brain increased due to an inefficient clearing pathway. 5 minutes after introducing [99mTc]Sestamibi, brain content was 0.22% ± 0.02% injected dose /g (ID/g) in the wt mice compared to 0.85% ± 0.14% ID/g in the Pgp- null mice, thus indicating Pgp activity was independent of the Appsw+/- transgene. [36]

After 12 months, Appsw+/- mice showed increased Aβ deposition in the brain. This increase was marked in the APPsw, Pgp-null mice compared to the Appsw, wt mice. Fibrillar Aβ, detected by thioflavine S staining, was also increased in Pgp-null mice. The mass of IAβ42 in the hippocampus was measured in Pgp-null mice by guanidine extraction and Aβ ELISA and the figure was twice as that of the wt controls. Hence these results show Pgp to be directly involved with the efflux of Aβ and its absence or inhibition would result in an increase in Aβ deposition and the progression of AD. [36]

Although Aβ is transported by Pgp through the BBB, how it does was not mentioned in the above experiments which determined only the outcome. Using LLC cells, proximal renal tube epithelia derived from porcine to simulate the BBB [38], transcellular transport of Aβ and its direction of movement was recorded. LLC and LLC-MDR1 cells were incubated with 5µm Aβ1-40/1-42 in its apical and basal compartments for 1 hour at 37°C. Aβ in the apical compartment was measured with a fluorescence photometer and differentiating the different Aβ peptides was performed via ELISA method. Basal (B) to apical (A) transport rates for Aβ 1-40 stood at 40 µmol x mg protein-1 in LLC-MDR1 cells, B to A in LLC cells at 20 µmol x mg protein-1 whereas A to B movement for both cell types were almost the same at 20 µmol x mg protein-1 . For Aβ 1-42, B to A movement was significantly higher in LLC-MDR1 at 50 µmol x mg protein-1 , while B to A for LLC and A to B in both LLC and LLC-MDR1 cells was similar to readings of Aβ 1-40. [38] Thus these results indicated that not only is Aβ transported by Pgp in vesicles in their soluble form- as derived from THT firbillogenesis assay [38], but that Pgp increases the B to A efflux rate through the BBB. A to B movement was constant throughout the experiment at a much lower figure than B to A due to Pgp inhibiting its movement. [39] As a result, Aβ influx into the brain is decreased while efflux is increased.

RAGE is involved with the A to B movement of Aβ and thus coincides with Pgp, suggesting that Pgp modulates the activity of RAGE, acting as a modulator to control the levels of Aβ being influxed through the BBB. Like RAGE, ET-1 inhibits the activity of Pgp after 24 hours of being introduced to HBMEC cells, causing Aβ to accumulate and AD to progress but all in so to enhance drug delivery and decrease drug elimination. [40, 41, 42] Also, one would expect LRP-1 to be affected by Pgp's activity as LRP-1 levels are significantly reduced in Pgp-null mice. However, LRP-1's activity is unaffected by Pgp activity directly. [39]


There are many roles of the BBB influencing Aβ influx and efflux which hence builds up to cause AD of which three major transporters of the BBB have been discussed in this essay. Aβ is formed in the circulation and from the neurons within the CSF, and RAGE allows for a major proportion of Aβ to be uptaken into the CSF from the blood of the capillaries. RAGE's activity is altered by either its spliced form sRAGE or its genetic structure. Other factors such as ET-1 and Pgp affect its performance at the BBB indicating that RAGE is not a single isolated component but is modulated within the BBB.

LRP-1 behaves in a similar way with sLRP and its influx rate is much affected by ligands such as ApoE, α2m and RAP. Its action is in contrast to that of RAGE and together they modulate the Aβ movement in and out of the BBB. Overexpression of these two transporters has been observed which regulate the levels of Aβ, maintaining CNS homestasis. Again, LRP-1's activity is not affected by Pgp, but its concentrations along the BBB are reduced as mentioned.

Pgp is a relatively new area of study, with a vast multitude of ligands and inhibitors, of which their mechanisms of how they bind to Pgp are still unknown. However, Pgp is a very important transporter in relation to AD as not only does it reduced influx and increase efflux through its own, it modulates the activities of RAGE and the abundance of LRP-1 along the BBB. Also, Pgp is a drug efflux pump and its inhibition will allow drugs that can treat and help with AD reach their sites, however this may affect Aβ movement.

Thus, it can be seen that the BBB's many transporters work in isolation and have a single purpose. They are all intricately linked and affecting one will affect another transporter along the BBB, may it be RAGE, LRP-1, Pgp or any other. Hence, AD's initiation and progression is tied in to the activities of these transporters. The BBB is a vast, thinking structure that regulates through feedback and a network of transporters of which only with further clarification can we discover its many roles in preventing or initiating AD.