|Home | About | Journals | Submit | Contact Us | Français|
Brain to blood transport is believed to be a major determinant of the amount of amyloid ß protein (AßP) found in brain. Impaired efflux has been suggested as a mechanism by which AßP can accumulate in the CNS and so lead to Alzheimer's disease (AD). To date, however, no study of the efflux of the form of AßP most relevant to AD, AßP1-42, has been conducted, even though a single amino acid substitution in AßP can greatly alter efflux. Here, we examined the efflux of AßP mouse1-42, mouse1-40, human1-42, and human1-40 in young CD-1, young SAMP8, and aged SAMP8 mice. The SAMP8 mouse with aging spontaneously overproduces AßP and develops cognitive impairments reversed by AßP-directed antibody or phosphorothioate antisense oligonucleotide. CD-1 mice transported all forms of AßP, although mouse1-42 and human1-40 were transporter faster than the other forms. There was a decrease in the saturable transport of mouse1-42 in SAMP8 mice regardless of age. Efflux of mouse1-40 and human1-42 was only by a non-saturable mechanism in young SAMP8 mice and their efflux was totally absent in aged SAMP8 mice. These differences in the efflux of the various forms of AßP among the three groups of mice supports the hypothesis that impaired efflux is an important factor in the accumulation of AßP in the CNS.
Amyloid beta protein (AßP) is thought to play a causal role in the development of Alzheimer's disease(Sambamurti et al., 2002) (Banks and Morley, 2003;Rosenberg, 2000). AßP is variably cleaved at the C-terminus to form peptides 39-43 amino acids in length(Rosenberg, 2000;Selkoe, 1990). The 1-40 and 1-42 peptides are the most common and the most widely studied. In general, AßP1-40 is less neurotoxic, less common in the neuritic plaques of AD, and less likely to be involved in the neuropathology of AD than AßP1-40. However, AßP1-42 is more difficult to study than AßP1-40 because of polymerization.
As is the case for any peptide, the levels of AßP are a balance between its rates of synthesis and degradation (Sambamurti et al., 2002). Both of these rates have been postulated to be altered in AD. Additionally, the level of AßP in the CNS is affected by its rate of clearance from the brain (Banks et al., 1997b;DeMattos et al., 2002;Ghersi-Egea et al., 1996;Shibata et al., 2000).
AßP is cleared from the brain by two major mechanisms. First, AßP in the cerebrospinal fluid is transferred to blood across arachnoid villi and lymphatic drainage along with reabsorption of the CSF. This is a passive mechanism and affects all substances found in the CSF. Second, AßP1-40 has been shown to be transported across brain capillaries by a saturable efflux system (Bading et al., 2002;DeMattos et al., 2002;Ghersi-Egea et al., 1996;Monro et al., 2002;Shibata et al., 2000). This transporter, tentatively identified as LDL receptor-related protein-1 (Shibata et al., 2000), has been suggested to be impaired in AD (Bading et al., 2002;Banks et al., 1999;Ghersi-Egea et al., 1996;Rosenberg, 2000). Such impairment could be a factor in the accumulation of AßP in the CNS and the onset of AD. Improved efflux may be one mechanism by which antibody directed against AßP can reverse CNS loads of AßP and lead to cognitive improvement in mice overexpressing amyloid precursor protein (Bacskai et al., 2001;Banks et al., 2002;Bard et al., 2000;DeMattos et al., 2002;Zlokovic et al., 2000).
However, the efflux studies have only been performed with the human AßP1-40, mostly in rodent models. Blood-brain barrier (BBB) transport systems can exhibit species specificity. For example, human interleukin-1∀ is transported across the mouse, but not the rat, BBB (Plotkin et al., 2000). Additionally, minor changes in peptide primary structure can alter peptide transport rate. For example, Tyr-MIF-1 is transported from brain to blood but not blood to brain, whereas MIF-1, which differs from Tyr-MIF-1 by only an N-terminal tyrosine, is transported into, but not out of, the CNS (Banks and Kastin, 1994). Some species of AßP also show such specificity. A single amino acid substitution at position 22 from glutamate to glycine reduces CNS to blood efflux of AßP1-40 by 7 fold (Monro et al., 2002). Furthermore, most of the residual transport occurred at the choroid plexus and not at the vascular BBB. Mouse1-42 and human1-40 differ by 5 amino acids (table 1).
The above raises two fundamental questions which we address here. The first question is whether AßP1-42 is transported out of the brain. AßP1-42 is considered the most neurotoxic of the AßP's and should be the peptide of greatest interest in BBB studies. However, to date, only the efflux of the AßP1-40 has been investigated. The second question is whether any measurable efflux of AßP1-42 is impaired in a mouse model of spontaneous overproduction of AßP. The mouse strain (SAMP8) investigated has a natural mutation which leads to age-related overexpression of APP, accumulation of AßP, and cognitive defects (Flood et al., 1993;Flood and Morley, 1998;Morley et al., 2000;Morley et al., 2002;Nomura et al., 1996). The learning and memory deficits in SAMP8 mice are reversed by administration of AßP directed antibodies and antisense to APP mRNA (Banks et al., 2001;Kumar et al., 2000;Morley et al., 2000;Morley et al., 2002). A subset of this question is whether any such impairments in efflux would precede or follow the onset of AßP accumulation. If following, then the impaired efflux could be related to the increased AßP levels. If preceding, then impaired efflux could be a factor in the age-related accumulation of AßP.
Radioactive Labeling and Purification of AßP: Five :g of one of four carrier-free, recombinant amyloid beta proteins were used: mouse1-42, mouse1-40, human1-42, human1-40 (American Peptide Co, Sunnyvale, CA) was labeled by the chloramine-T method. This was purified on a column of G-10 Sephadex by eluting 0.1 ml fractions with protein-free, chloride-free phosphate buffer solution. Both SDS and non-denaturing gels showed that about 70% of radioactively labeled mouse1-42 eluted as monomer (figure 1).
Measure of CNS to Blood Efflux Rates: The method used to quantify the fate of centrally administered I-AßP has been described previously (Banks and Kastin, 1989) and has been used to assess brain efflux (Banks et al., 1986;Banks et al., 1990) and sequestration (Banks and Broadwell, 1994;Cashion et al., 1996). Two month old male ICR mice, two month old SAMP8 mice, or twelve month old SAMP8 mice (all from our in-house colonies) were kept on a 12/12 hour light/dark cycle with food and water freely available. They were anesthetized on day of study with 0.15 ml of 40% urethane. The scalp was removed and a hole made into the lateral ventricle, 1.0 mm lateral and 1.0 mm posterior to the bregma, with a 26 gauge needle with a tubing guard which kept the depth of the holes constant (3.0-3.5 mm). Mice received 1.0 ul intracerebroventricular (icv) injections containing 5,000 cpm (about 0.125 ng) of I-AßP in a lactated Ringer's solution with 1% bovine serum albumin. Mice were decapitated at 2, 5, 10, and 20 minutes after injection. The whole brain was removed, the pituitary and pineal were discarded, and the level of residual radioactivity in the whole brain was determined from the counts after 3 minutes in a gamma counter. The level of radioactivity in whole brain at t=0 was determined in mice overdosed with anesthetic as previously described (Banks and Kastin, 1989). The cpm remaining in the brain was divided by the cpm injected and multiplied by 100 to yield the percent of the injected dose remaining in brain (%Inj/brain). The log of this value was plotted against time, and the half-time disappearance was calculated by multiplication of the inverse of the slope of this relationship by 0.301. Regression lines were compared with the Prism 3.0 program and as described below under Statistical Analysis.
The possible presence of a saturable component was tested by giving another group of mice an icv injection of I-AßP with or without 1 :g/mouse of unlabeled AßP included in the injection. These mice were decapitated 10 min after the icv injection, the means and error terms calculated, and the two groups compared by t-test.
Statistical Analysis: Means are reported with the number of mice used (n) and the standard error of the mean (SEM). Student's t-test was used for comparison of two groups. More than two groups were compared by analysis of variance (ANOVA) followed by Newman-Keuls post test. The p values were reported for relevant statistically significant differences. Regression lines were calculated by the least squares method with the Prism 3.0 program (GraphPad, Inc., San Diego, CA) and the slope (m) with its standard deviation of the mean, the intercept (i) with its error term, the correlation coefficient (r), the number of points on which the line was based (n) and the p value reported. Regression lines were compared for statistical differences with the Prism 3.0 program, which first determines whether there are differences between slopes and, if not, whether there are differences between intercepts. More than two slopes were compared by ANOVA followed by Newman-Keuls range test with the standard deviation of the mean taken as the standard error term and, because two means (the slope and the intercept) were calculated from the data, n-1 was used as the value for n.
Efflux of AßP Mouse1-42: The relation between log(%Inj/brain) and time was statistically significant for the CD-1 (m = -0.0190 ∀ 0.0020, i = 1.58 ∀ 0.020, r = 0.959, n = 10, p<0.001), young SAMP8 (m = -0.0079 ∀ 0.0025, i = 1.54 ∀ 0.026, r = 0.773, n = 9, p<0.05 ), and aged SAMP8 groups (m = -0.0054 ∀ 0.0016, i = 1.54 ∀ 0.017, r = 0.792, n = 9 , p<0.05 ), demonstrating a measurable efflux of mouse1-42 in all three groups of mice (Figure 2, upper panel). However, statistical comparison of the lines showed that differences existed among the slopes of the line: F(2,64) = 7.87, p<0.001. ANOVA confirmed a difference among slopes [F(2,24) = 12.7, p<0.001]. Newman-Keuls range test showed that the slope of the line for CD-1 mice was different from the slope of the line for the young SAMP8 (p<0.001) and aged SAMP8 mice (p<0.001). There was no difference between the SAMP8 young and aged mice (Figure 2, bottom panel) even by t-test. The half-time disappearances calculated from the slopes were 15.8 min for the CD-1 mice, 38.0 min for the young SAMP8 mice, and 55.9 min for the aged SAMP8 mice. Coinjection of unlabeled mouse1-42 (1 :g/mouse) with the radioactive mouse1-42 showed a saturable component to efflux in all three groups of mice (Figure 3).
Efflux of Other AßP in CD-1 mice: Each of the other AßP's had a significant relation between their log(%Inj/brain) and time, demonstrating that the CD-1 mouse could clear each of them: mouse1-40: (m = -0.0076 ∀ 0.0031, i = 1.56 ∀ 0.032, r = 0.660, n = 10, p<0.05); human1-42: (m = -0.0076 ∀ .0024, i = 1.57 ∀ 0.025, r = 0.739, n = 10, p<0.05); human1-40: (m = -0.0185 ∀ 0.0026, i = 1.51 ∀ 0.026, r = 0.931, n = 10, p<0.001). This yielded half-time disappearance rates of 39.4 min (mouse1-40), 39.7 min (human1-42), and 16.3 min (human1-40). Comparison of the slopes showed a statistical difference by 2-way ANOVA with mouse group (CD-1, young SAMP8, aged SAMP8) and peptide (mouse1-42, mouse1-40, human1-42, human1-40) being the two independent variables for interaction (p<0.001), a trend for peptide (p = 0.07) and no effect for mouse group. The Newman-Keuls range test showed that the efflux by CD-1 mice of mouse1-42 and human1-40 was faster than the efflux of mouse1-40 and human1-42 (Figure 4). Efflux of each of these peptides was inhibited by its unlabeled version, showing each was transported out of the CNS by a saturable system (Table 2). The ANOVA showed that each of the AßP were able to cross-inhibit the efflux of radioactive mouse1-42 out of the brains of CD-1 mice: F(3.21) = 20.2, p<0.001 (Figure 5). Newman-Keuls range test showed that each peptide inhibited at the p<0.001 level with no difference among the peptide groups.
Efflux of Mouse1-40, Human1-42, and Human1-40 in SAMP8: A significant relation between log(%Inj/brain) and time existed for mouse1-40 (m = -0.0047 ∀ .0019, i = 1.62 ∀ 0.020, r = 0.658, n = 10, p<0.05) and human1-42 (m = -0.0069 ∀ .0026, i = 1.59 ∀ 0.027, r = 0.683, n = 10, p<0.05) in young SAMP8 mice, demonstrating efflux. The half-time disappearance was 63.5 min for mouse1-40 and 43.3 min for human 1-42. However, 1 :g/mouse of unlabeled peptide co-injected with its radioactive version did not affect the cpm retained by brain, showing that there was not a saturable component to efflux (Table 2).
A significant relation between log(%Inj/brain) and time did not exist for mouse1-40 or human1-42 in aged SAMP8 mice, demonstrating the absence of efflux. Consistent with this, 1 :g/mouse of unlabeled peptide co-injected with its radioactive version did not affect the cpm retained by brain.
Human1-40 had a significant relation between log(%Inj/brain) and time in both young (m = -0.0047 ∀ .0018, i = 1.62 ∀ 0.018, r = 0.683, n = 10, p<0.05) and aged SAMP8 mice (m = -0.0139 ∀ .0019, i = 1.66 ∀ 0.020, r = 0.931, n = 10, p<0.001). The slopes yielded half-time disappearance rates of 64.6 min (young) and 21.6 min (aged) mice. Unlabeled human1-40 inhibited the efflux or radioactive human1-40 in both young (p<0.001) and aged (p<0.05) mice.
Several original observations arose from this first study to examine the CNS to blood efflux of AßP1-42 in vivo. We found that mouse1-42 was transported out of the CNS by a saturable transport system. All three groups of mice studied (young CD-1, young SAMP8, and aged SAMP8) were able to transport AßP1-42 out of the CNS by a saturable mechanism. However, the SAMP8 strain regardless of age was much less efficient in transporting AßP out of the CNS. We also found that CD-1 mice transported other forms of AßP at varying rates. Species specificity existed, with mouse1-42 transported much more quickly than mouse1-40 or human1-42. Evidence suggests that a single transporter or a family of transporters with overlapping affinities are responsible for the efflux of the AßP. However, the SAMP8 transport system is less universal and is shifted in its efficiency of transport away from mouse1-42 towards human1-40.
Mouse1-42 was transported out of the CNS by all three groups of mice by a saturable mechanism. The method used to examine efflux has several advantages over other efflux methods (Banks et al., 1997a;Banks and Kastin, 1989). First, it can use mice, as opposed to rats, sheep, or larger animals. Second, the icv injection version used here, as opposed to the brain parenchymal injection version or other brain parenchymal injection methods, measures global efflux; that is, it measures efflux at both the vascular and choroid plexus barriers. Finally, it quantifies actual efflux as opposed to producing an index or ratio relative to a reference substance. Therefore, the method can be used in systems where underlying assumptions of bulk flow or CNS distribution may not hold.
Efflux of mouse1-42 varied between the CD-1 and SAMP8 strains of mice. CD-1 mice were much more efficient at transporting mouse1-42 than either young or aged SAMP8 mice. The CD-1 also transported by a saturable system each of the other 3 AßP's. Transport in CD-1 was faster for mouse1-42 and human1-40 than for mouse1-40 or human1-42. The reason why CD-1 mice transport the two most different peptides equally well is unclear. If a single transporter exists for all forms of AßP, then it may be that these two forms are most similar in a critical tertiary structure important to AßP efflux. Alternatively, there may be more than one efflux system for AßP.
Although SAMP8 mice could also transport mouse1-42 by a saturable system, they did so at a slower rate. Both young and aged SAMP8 mice were equally impaired, showing that the transporter deficiency predates the accumulation of AßP and plaque formation found in the aged, cognitively impaired SAMP8. This suggests that impaired efflux could be an early contributor to accumulation of AßP in the CNS. This clearly supports the idea that a similar process in humans could be critical to the development of AD.
Efflux of the other forms of AßP were also affected in the SAMP8 mouse. The efflux of mouse1-40 and human1-42 in young CD-1 had no saturable component. Non-saturable reabsorption of CSF occurs at the arachnoid villi and the primitive lymphatics of the brain. Any substance dissolved in the CSF will exit the CNS by these mechanisms. Human1-40, although transported out of the CNS by a saturable system in the SAMP8, was effluxed at a slower rate in the SAMP8 mouse. These findings support the above observations for mouse1-42 and again suggest that impaired brain efflux may be a contributing factor to the accumulation of AßP in the SAMP8.
An age effect was observed for the transport of AßP. No statistically significant efflux, either saturable or non-saturable, could be found for mouse1-40 or human1-42. Given that any substance remaining in the CSF should have a measurable, non-saturable efflux, this suggests that the aged SAMP8 somehow disposes of these AßP's differently than the young SAMP8. One possibility is that the AßP is retained by brain parenchyma, possibly in the formation of plaques.
In conclusion, we found that the mouse1-42 version of AßP is transported out of the CNS in CD-1, young SAMP8, and aged SAMP8 mice. However variations in the efflux of AßP among these three groups of mice and among the various forms of AßP support the hypothesis that an impaired transport could be a significant contributor to AßP accumulation within the brain and to the development of AD.
Supported by VA Merit Review and R01 NS41863 and R01 AA12743.