SEARCH

SEARCH BY CITATION

Keywords:

  • acetylcholinesterase;
  • Alzheimer's disease;
  • cerebrospinal fluid;
  • cholinesterase inhibitors;
  • rivastigmine;
  • tacrine

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

Protein levels of different acetylcholinesterase (AChE) splice variants were explored by a combination of immunoblot techniques, using two different antibodies, directed against the C-terminus of the AChE-R splice variant or the core domain common to all variants. Both AChE-R and AChE-S splice variants as well as several heavier AChE complexes were detected in brain homogenates from the parietal cortex of patients with or without Alzheimer's disease (AD) as well as the cerebrospinal fluid (CSF) of AD patients, compatible with the assumption that CSF AChEs might originate from CNS neurons. Long-term changes in the composition of CSF AChE variants were further pursued in AD patients treated with rivastigmine (n = 11) or tacrine (n = 17) in comparison to untreated AD patients (n = 5). In untreated patients, AChE-R was markedly reduced as compared with the baseline level (37%), whereas the medium size AChE-S complex was increased by 32%. Intriguingly, tacrine produced a general and profound up-regulation of all detected AChE variants (up to 117%), whereas rivastigmine treatment caused a mild and selective up-regulation of AChE-R (∼10%, p < 0.05). Moreover, the change in the ratio of AChE-R to AChE-S (R/S-ratio) strongly and positively correlated with sustained cognition at 12 months (p < 0.0001). Thus, evaluation of changes in the composition of CSF AChE variants may yield important information referring to the therapeutic efficacy and/or development of drug tolerance in AD patients treated with anti-cholinesterases.

Abbreviations used
AChE

acetylcholinesterase

AChE-R

the ‘read-through’ AChE splice variant

AChE-S

synaptic AChE splice variant

ChAT

choline-acetyltransferase

CMRglc

cerebral glucose metabolism

DTT

dl -dithiothreitol

Gn AChE

globular AChE molecular forms

MMSE

mini-mental state examination

PRiMA

proline-rich membrane anchor

RACK1

receptor for activated C kinase

SDS–PAGE

sodium dodecyl sulfate–polyacrylamide gel electrophoresis

Cholinesterase inhibitors (ChEIs) are currently in use for symptomatic treatment of patients with Alzheimer's disease (AD). Individual ChEIs differ from each other with respect to their pharmacological properties (Enz et al. 1991; Wagstaff and McTavish 1994; Svensson and Nordberg 1996; Svensson et al. 1996; Giacobini 1998; Nordberg and Svensson 1998). These differences may be reflected in their efficacy or safety profiles as well as in their capacity to induce feedback responses, which might be associated with drug tolerance (Giacobini 1997). ChEIs act by inhibiting the degradation of acetylcholine (ACh). Tacrine, donepezil and galantamine are reversible ChEIs, metrifonate is an irreversible ChEI, and rivastigmine is a pseudo-irreversible (slowly reversible) ChEI with an intermediate duration of action. While the primary target of these agents is acetylcholinesterase (AChE; EC 3.1.1.7), some also inhibit butyrylcholinesterase (BuChE; EC 3.1.1.8) (Nordberg and Svensson 1998).

AChE and BuChE are found in the cerebrospinal fluid (CSF) and plasma (Darreh-Shori et al. 2002). AChE molecules in the ventricular and lumbar CSF likely originate from AChE-producing neurons that secrete soluble isoenzymes into extracellular space, from where they pass into the CSF (Tomkins et al. 2001). Because of the selective loss of AChE variants in AD, an understanding of its molecular polymorphism is of considerable interest (Arendt et al. 1992; Massoulie et al. 1998).

Three different globular AChE subunits (G1) exist and arise by 3′ alternative mRNA splicing of pre-AChE mRNA (Grisaru et al. 1999; Soreq and Seidman 2001). All three mRNA spliced variants have a common core, consisting of exons 2, 3 and 4, which is sufficient for catalytic activity, but the variants possess distinct C-termini. These include the brain-abundant exon 6-encoded C-terminal peptide corresponding to the synaptic AChE (AChE-S, also called AChE-T). Another hematopoietic exon 5-encoded C-terminus (AChE-E or H) is found on the red blood cells, and the C-terminus derived from reading through the open reading frame of the pseudo-intron 4 directly into exon 5 (‘readthrough’ or AChE-R).

In CNS, various AChE-S isoforms may exist as monomer (G1) or multimers of glycosylated subunits, by assembly of two (G2) or four (G4) globular catalytic subunits with a 20-kDa structural subunit, known as PRiMA; proline-rich membrane anchor (Perrier et al. 2002). In comparison, the AChE-R splice variant accumulates in response to external stressors (Soreq and Seidman 2001). The C-terminus of this splice variant lacks the cysteine residue necessary for disulfide intersubunits binding, therefore it was initially believed that AChE-R could not be incorporated into the heavier AChE molecule, but functions as soluble monomers (Kaufer et al. 1998). However, not all of the subunits in tetrameric AChE are disulfide-bonded (Liao et al. 1993; Flores-Flores et al. 1996), suggesting that the C-terminal cysteine may not be essential for the binding of the assembly of the subunits into the heavier AChE complexes (Perrier et al. 2002). Also, AChE-R was recently found to form intraneuronal heterotrimeric complexes with protein kinase CβII, and its scaffold protein RACK1 (Birikh et al. 2003), suggesting that interprotein AChE-R complexes might exist.

Nevertheless, heavy membrane-bound AChE isoforms were shown to selectively decrease in a number of cortical and subcortical brain regions in AD patients, which consequently leads to a decline in the ratio of the G4/G1 molecular forms of AChE (Atack et al. 1983; Fishman et al. 1986; Arendt et al. 1992). It was further found that the ratio of G4/G1 AChE positively correlates with choline acetyltransferase (ChAT) activity, regarded as a presynaptic cholinergic marker (Fishman et al. 1986; Siek et al. 1990). However, as variant-selective antibodies were not available to these research groups, they could not address the changes in the composition of the AChE splice variants.

Selective up-regulation of AChE-R attenuates neurodeterioration in transgenic mice that over-express human AChE-R, suggesting that AChE-R may exert a neuroprotective effect. In contrast, excess of the human synaptic variant (hAChE-S) intensifies neurodeterioration (Sternfeld et al. 2000), and induces a feedback response of murine (m)AChE-R overproduction. This, in turn, increases antisense-suppressible susceptibility to closed-head injury (Shohami et al. 2000) and causes exaggerated response to mild stimuli such as a circadian switch (Cohen et al. 2002), suggesting that mutual excess of AChE-S and AChE-R may be detrimental.

The potential interplay between AChE variants may further be important for assessing the clinical value of AD therapeutics. Thus, various ChEIs were shown to confer mild and temporary improvements in the cognitive and behavioral functioning of AD patients, and long-term effects of such treatments suggest that they may also delay the progression of dementia (Nordberg and Svensson 1998; Giacobini 2000). The molecular and physiological processes underlying these improvements are not yet fully understood, and it is as yet unclear why the observed symptomatic improvements are temporary and which AD patients are at increased risk for developing tolerance toward these drugs. Answers to these questions can potentially lead to the development of next generation drugs, with prolonged efficacy and limited tolerance.

Recently we observed long-term changes in the composition of AChE variants in the CSF of patients treated for one year with rivastigmine. Using the immunoblot technique, it was found that long-term rivastigmine treatment caused a selective and mild up-regulation (10%) of a 50-kDa AChE-R variant (Darreh-Shori et al. 2002).

To address the more fundamental issue of therapeutic value, we have now expanded our previous study by including CSF samples from patients treated with tacrine as well as untreated AD patients. To test whether treatment with ChEIs could retrieve closer to normal composition of AChE variants in the AD patients, we specifically determined the long-term changes in the protein level of AChE-R and AChE-S, using the conventional reducing Western-blot technique. In addition, we also performed immunoblot analyses following non-reducing electrophoresis of the CSF proteins. This technique preserves the linkage between globular subunits of the heavy AChE isoforms, such as G2 and G4 molecules as well as the heteromeric complexes of AChE-R. To test whether CSF could reliably reflect the long-term changes in the CNS, we also compared the expression of AChE variants and multimeric isoforms found in CSF with those found in cortical brain homogenates taken postmortem from AD patients and controls. Our findings demonstrate changes in the composition of AChE variants with opposite directions in treated and untreated patients and distinct long-term effects for various ChEIs.

Materials

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

Recombinant human AChE-S (rAChE-S), bovine liver catalase (EC1.11.1.6), calf intestinal mucosa alkaline phosphatase (EC3.1.3.1), Triton X-100, Tris-HCl, glycerol, sodium dodecyl sulfate (SDS), dl-dithiothreitol (DTT), Tween-20, thimerosal and 2-mercaptoethanol (2-ME) were purchased from Sigma Chemical Co. (St. Louis, MO, USA). The anti-Core Ab (AChE [N-19]; sc-6431), goat anti-rabbit IgG (sc-2030); and donkey anti-goat IgG (sc-2033) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Rabbit anti AChE-R Ab was as detailed before (Sternfeld et al. 2000; Darreh-Shori et al. 2002). Pre-cast polyacrylamide mini-gels (10% or 4–20%, used for non-reducing and reducing SDS–PAGE, respectively) were purchased from Bio-Rad (Hercules, CA, USA). Chemiluminescence ECL Plus™ detection kit and Hyperfilm-ECL were from Amersham Life Science (Little Chalfont, UK). The molecular weight (MW) standards were ‘Cruz Marker’ (Santa Cruz, sc-2035) and ‘ECL DualVue’ (Amersham, RPN810).

Patients and collection of CSF samples

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

Briefly, the AD patients were outpatients of at least 50 years of age (Table 1). Only patients with mild dementia were included, as defined by a mini mental status examination (MMSE) score (Folstein et al. 1975) of 20–30 inclusive. Patients in the rivastigmine group received the drug (Exelon®, Novartis) twice daily with food. The maximum recommended dose was 12 mg/day. Patients in the tacrine group received the drug (Cognex®, Par kDavis Scandinavia) four times daily and the maximum recommended dose was 160 mg/day. Other details concerning inclusion/exclusion criteria, escalation of the dosage, side-effects of the drugs and neuropsychological assessments, measurements of cerebral glucose metabolism, CSF-tau and CSF-Aβ42 were described previously (Nordberg et al. 1999; Darreh-Shori et al. 2002; Stefanova et al. 2003). All patients and their responsible caregivers provided written informed consents to participate in the study. The study was conducted according to the Declaration of Helsinki and subsequent revisions and was approved by the Ethics Committee of Huddinge University Hospital, Sweden.

Table 1.  Demographic summary of the AD patients in the treated and untreated group
Patients groupTacrineRivastigmineUntreated
  • a

    Significant difference between the untreated and the tacrine group ( p  < 0.05).

  • b

    Significant difference between the untreated group and the rivastigmine group ( p  < 0.05).

No. of patients17115
Age (mean ± SE)66.3 ± 1.670.1 ± 1.974.6 ± 2.0a
Gender (M/F)10/77/41/4
MMSE at baseline24.3 ± 0.924.9 ± 0.824.6 ± 0.7
MMSE at 12 months23.1 ± 1.223.9 ± 1.0b20.8 ± 1.8

CSF samples from all of the treated patients were taken at baseline (pretreatment samples) and after 3 and 12 months of treatment. CSF samples from the untreated patients were collected at two occasions with a time interval of 14.7 ± 2.2 months (mean ± SEM). Collection involved lumbar puncture in the L3-4 or L4-5 interspace. All samples were immediately centrifuged at 2000 × g for 10 min and were kept frozen as 1-mL aliquots at −70°C until the assay time. CSF samples were thawed and centrifuged at 15 000 × g for 3 min at 4°C immediately before the assay.

Cortical extract from AD and control brain

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

Cortical brain homogenates (P2 membrane fraction) from parietal cortex of AD patients (n = 3, mean age 77.7 ± 4.8 years, postmortem delay 26 ± 1.5 h, non-smoking) and from parietal cortex of non-AD controls (n = 3, mean age 82.3 ± 1.5 years, postmortem delay 21 ± 3.7 h, non-smoking) were prepared as described before (Marutle et al. 1999). Pooled extract from each group was prepared and kept at −80°C until the assay.

Reducing and non-reducing SDS–PAGE

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

AChE splice variants were quantified by immunoblotting after resolution by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) under reducing conditions, as described in details before (Darreh-Shori et al. 2002), with some minor modifications as follows. Duplicates or more of each CSF sample were applied on mini-gels (4–20% precast gel). Immediately after addition of 6X-Loading buffer [0.5 m Tris-HCl, pH 6.8, containing glycerol (30% v/v), SDS (10% w/v), DTT (10% w/v), and bromphenol blue (1.2% w/v)] to each CSF sample, it was agitated for ∼5 s and incubated in a 95°C water-bath for 5–10 min.

Non-reducing SDS–PAGE was performed as described above with the exception that 10% precast gels were used instead of 4–20% gels and that the reducing agent, DTT, was omitted from the 6X-Loading buffer.

Gels were blotted for 2.5 h at constant current (10 mA/Gel) at 4°C with an ice-cooling unit as described previously (Darreh-Shori et al. 2002).

Detection of AChE variants in cortex and CSF

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

Two primary antibodies were used to detect variants and multimers of AChE in brain homogenates and CSF. The anti-Core antibody is an affinity-purified goat polyclonal antibody that detects the core domain of human AChE, common to all variants. The anti-AChE-R Ab is an affinity-purified rabbit polyclonal antibody elicited towards the C-terminal sequence of AChE-R (Shohami et al. 2000; Sternfeld et al. 2000). The specificities of the primary antibodies were tested by using recombinant AChE-S protein and AChE-R protein on the gels as control.

The immunostaining procedure was essentially the same as outlined before (Darreh-Shori et al. 2002). After probing blots with one of the primary Ab and the corresponding secondary Ab, the membranes were stripped to remove the bound antibodies, by submerging the membranes in 40 mL stripping buffer [100 mm of 2-ME and 2% SDS in 62.5 mm Tris-HCl, pH 6.7] and incubation at 55°C for 30 min with occasional agitation. To ensure the removal of the antibodies, blots were incubated with ECL Plus™ detection reagents and exposure of Hyperfilm™ECL™ for 10–15 min. The blots were then washed and reprobed by the other primary Ab and its corresponding secondary Ab as described above.

Sedimentation analysis

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

Molecular isoforms of AChE were also separated by sucrose gradient sedimentation technique by ultracentrifugation at 165 000 × g in a continuous sucrose gradient (5–20% w/v) for 18 h at 4°C in a Beckman rotor (SW 41 Ti). Pooled CSF (1.0 mL) or brain homogenate (0.30 mL) samples were applied on the top of gradients. The gradients contained 10 mL Tris-HCl (10 mm, pH 7.4) containing 1.0 m NaCl, 50 mm MgCl2 and 0.5% Triton X-100. Approximately 40 fractions were collected from the bottom of each tube. Enzymes of known sedimentation coefficient, bovine liver catalase (11.4S; ∼250 kDa) and calf intestinal alkaline phosphatase (6.1S; ∼140 kDa), were used in the gradients to estimate the sedimentation coefficient of AChE isoforms. AChE activity in the collected fractions was determined as outlined before (Darreh-Shori et al. 2002). Then, the fractions with considerable AChE activity were used to detect AChE isoforms by anti-AChE-R and anti-Core Ab immunostaining after resolution by SDS–PAGE under reducing and non-reducing conditions as described above. To remove any confounding effect due to the stripping procedure and re-probing of the blots, duplicate blots were prepared and used.

Quantification of AChE isoforms in CSF samples

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

Quantification of relative changes of protein levels of AChE splice variants and heavier molecular isoforms was performed only on CSF samples from the treated and untreated AD patients. To quantify each band, the films were scanned using a Sharp JX-325 scanner. The background signals were uniformly removed and optical densities (ODs) of the bands were calculated as a product of contour OD and the area of the contour using ImageMaster™ 1D software (version 1.10; Pharmacia Biotech). Duplicates or triplicates of all of CSF samples of each patient were applied on the same gel, allowing the comparison of average ODs of bands detected in samples collected at 3 and 12 months with the patients' own average baseline values.

Statistical analysis

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

Data are expressed as mean values and SEM. The changes in protein levels of CSF AChE variants with time (within-groups) were assessed by repeated measured (RM) anova on the optical density values. A significant anova (p < 0,05) followed by a post hoc analysis that tested the significance of results at each time point compared with baseline and other time points. All values expressed as the percentages were individually calculated for each patient compared to the patient's baseline value [(1–At12/At0) × 100; the At0 and At12 were the values measured for the patient A at baseline and after 12 months, respectively]. For between-groups analysis, one-factor anova or unpaired Student's t-test was performed on data expressed as percentages of baseline. Regression analysis was done as described in the text.

No statistical comparison was performed between the data derived from homogenates of the AD brain versus the control brain or between homogenates of the AD or the control brain and the CSF from treated/untreated AD patients.

Comparative analysis of AChE variants in cortex and CSF

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

Different AChE variants and multimeric complexes were identified in brain homogenates, derived from parietal cortex of AD patients and controls, and was compared with the AChE variants found in CSF (Fig. 1).

image

Figure 1. Immunodetection of AChE splice variants and multimeric complexes after resolution by reducing and non-reducing SDS–PAGE. Blot I shows the protein bands detected by anti-Core Ab, indicating that these proteins contain the core-domain common to all AChE variants. Blot II represents the same blots as in blot I that werere-probed by anti-AChE-R Ab after complete removal of anti-Core Ab as described under Methods. It shows those bands that have the C-terminal characteristic to the ‘read-through’ AChE-R splice variants. The lanes, as indicated in the figure, show the detected AChE bands in representative blots of CSF of AD patients (‘AD CSF’); purified recombinant human AChE-S protein (‘rAChE-S’); P2 membrane fraction of cortical region of postmortem AD brain tissue (‘AD brain’) and control (‘C brain’).

Download figure to PowerPoint

Immunodetection by anti-Core antibody

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

After non-reducing separation of proteins in the brain homogenate, anti-Core Ab detected four intense protein bands, two of which with migration properties corresponding to those of recombinant G1 AChE-S monomers (rAChE-S, ∼70 and ∼65 kDa) and two which migrated considerably slower that could represent the heavy G4 tetramers (> 240 kDa) and the medium size G2 dimers(∼130–150 kDa) or heteromeric AChE complexes (Fig. 1, blot I, right panel).

In the CSF, the major detected bands migrated similarly to the heavier multimers (G4 and G2) found in the cortex, but a third band migrated similarly to a protein with a MW of about 50 kDa (Fig. 1, blot I, CSF lane, right panel) corresponding to G1 AChE-R (Darreh-Shori et al. 2002). Recombinant AChE used as a positive marker for the anti-Core Ab (Fig. 1, lane ‘rAChE-S, right panel) yielded a heavy isoform corresponding to G4 and a light band with a MW close to the 65-and 70-kDa G1 isoforms (found in brain homogenate) but not the intermediate G2 isoform. The intensity of the signal corresponding to the G1 isoform was relatively weaker in CSF, indicating that the most abundant AChE in CSF occurs as a component of multimeric complexes.

Next, the proteins in the brain homogenate and CSF as well as rAChE-S were separated under fully reducing conditions, which should lead to the breakage of the putative disulfide bonds linking AChE subunits in the heavy AChE molecules.

In brain homogenate, the anti-Core Ab detected three major bands of about 132–150 (dimers), 75 and 65 kDa (monomers). The signal of the heavy G4 band disappeared, indicating that this multimer was broken down to the lighter G2 and G1 subunits (Fig. 1, blot I, left panel). The 65- and 75-kDa bands could reflect the formation of covalent PRiMA–AChE complexes.

In principle, all AChE bands detected in cortex homogenate were also found in CSF. The heavy band in the CSF samples was reduced to four lighter bands. The 132-kDa band and three bands with MW close to the monomer of AChE, a diffuse broad band with MW between 65 and 75 kDa, a sharp 55-kDa as well as the 50-kDa AChE-R band that was detected in non-reduced CSF sample (Fig. 1, compare ‘CSF’ lanes in blot I). The latter two bands could reflect monomers without the PRiMA subunit. In the lane, corresponding to the rAChE-S protein, the heavy G4 band was predictably reduced to two bands (the broad 65–75 kDa band and the 55 kDa band), but not the 50-kDa AChE-R. Furthermore, presence of the 65–75 kDa band in the ‘rAChE-S lane’ (as shown in Fig. 1, blot I, left panel) suggested that the dispersed band detected in the CSF is unlikely to be an artifact, but might be due to heterogeneity of AChE subunits in the band or differences in the degree of glycosylation of these subunits.

Immunodetection by anti-AChE-R antibody

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

To determine which of the bands had a C-terminus characteristic of AChE-R, we reprobed the same blots with the anti-AChE-R Ab. The ‘rAChE-S’ lane served as negative control, as it has a C-terminus corresponding to that of synaptic AChE and thus should not be detected by this antibody. The anti-AChE-R Ab expectedly failed to detect this protein (Fig. 1, blot II, ‘rAChE-S’ lanes). However, the anti-AChE-R Ab did detect (under fully non-reducing conditions) the heavy complexes, both in brain homogenate and CSF (Fig. 1, blot II, right panel). This indicates that at least some of the AChE subunits incorporated into this heavy molecule should have (besides the epitope recognized by the anti-Core Ab) the sequence recognized by the anti-AChE-R Ab, which is selective for the C-terminus of the AChE-R splice variant. This antibody also detected some of the lighter bands, although with much weaker signal intensity, the 70-, 65- and 50-kDa bands (compare the lanes in right panel of blots I and II in Fig. 1).

After resolution of the brain extracts and CSF protein by reducing SDS–PAGE, the anti-AChE-R Ab detected two of the three bands detected by the anti-Core Ab (the 65-kDa and the 50-kDa AChE-R, but not the 55-kDa AChE-S variant, Fig. 1, blots I and II, left panels). The 65-kDa band was quite sharper than the dispersed 65–75 kDa band detected by the anti-Core Ab, indicating that the dispersed 65–75 kDa band, as mentioned before, was a mixture of the 75-kDa AChE-S and 65-kDa AChE-R variants. It should be noted that the signal intensity of the 50-kDa band (considered to be one of the major G1 AChE-R variants (Darreh-Shori et al. 2002), was highly increased in the reduced CSF sample compared to non-reduced CSF sample (Fig. 1, blot II), indicating that the AChE-R variant was part of the heavy complexes. Overall, we found that the AChE-R variant could be one of the major AChE species normally expressed in the cortex and CSF. Furthermore, it was shown that the G1 catalytic subunit of AChE-R could have the capacity to form multicompound complexes that migrated like the G4 AChE isoform. Because this finding was unexpected, we have further separated the different molecular forms of AChE by sucrose gradient sedimentation technique and immunoblotted the AChE isoforms in the collected sucrose fractions.

Immunodetection of AChE isoforms in sucrose gradient fractions

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

AChE isoforms in pooled predrug CSF samples of the AD patients and pooled homogenates of postmortem tissue from AD and control brains were separated by sucrose density gradient technique (Fig. 2).

image

Figure 2. Analysis of AChE isoforms by sucrose density gradient technique in pooled CSF samples of the AD patients (a) , in pooled brain extract from the control (b)  and AD brain (c).  Enzyme with known sedimentation coefficient, catalase (C; 11.4S) and alkaline phosphatase (P; 6.1S), were used to approximately determine sedimentation coefficients of AChE isoforms. It should be noted that the scales on y -axis in the graph a is different from that in graphs b and c. The open symbols indicate the fractions used in the subsequent immunoblot analysis.

Download figure to PowerPoint

Sedimentation analysis revealed that G4 AChE had a sedimentation coefficient of about 10.6S and was the major active AChE molecular form in CSF. A second peak was also present, corresponding to a mixture of G2 (5.5S) and G1 form (4.0S), with 10 folds lower activity (Fig. 2a). Similarly, a G4 and a G2 + G1 peak of AChE activity were observed in collected sucrose fractions from brain homogenate, but with comparable AChE activity between the G4 and the G2 + G1 peaks (Fig. 2b,c).

Then, the sucrose fractions with considerable AChE activity (open symbols in Fig. 2) were used for immunostaining of AChE isoforms with the anti-Core Ab and the anti-AChE-R Ab, under both reducing and non-reducing conditions.

Under non-reducing condition (Fig. 3), the anti-Core Ab detected three heavy bands with approximate MW between 250, 200 and 150 kDa (bands a–c, respectively) in the fractions 19–22 (peak G4 in Fig. 2a) from pooled CSF samples of the AD patients. These bands had migration properties and patterns similar to those of the recombinant AChE protein on the gel under non-reducing SDS–PAGE (lane ‘rAChE’ in Fig. 3, blot I). Furthermore, the anti-Core Ab detected two more bands in the sucrose fractions 26–31 (peak G2 + G1) with similar MW to band b (b*∼200 kDa) and band c (c*∼150 kDa, Fig. 3, blots I and III), perhaps reflecting as yet unclear heteromeric complexes of AChE-S. Alternatively or additionally, SDS by itself might eliminate the hydrophobic interactions (e.g. between two G2 in a G4 complex or between G1s and G3 or G2 in a G4 complex) resulting in multiple bands.

image

Figure 3. Non-reducing immunoblot analysis of AChE isoforms separated by sucrose density gradient technique. Blot I shows AChE bands detected by the anti-Core Ab in sucrose fractions 19–22 (the G4 peak) and 26–31 (the G2 + G1 peak in Fig. 2a ). Blot II is a duplicate of the blot I, but was immuno-stained by the anti-AChE-R antibody . A comparison between these two blots suggest that AChE-R variant is included in two of the three heavy bands (band a and b), whereas the third heavy band (band c) is devoid of AChE-R variant. Blot III, which was immuno-stained by the anti-Core Ab , illustrate that the heavy bands of AChE (a, b and c) could also be detected in sucrose fractions of the pooled brain homogenate from AD brain. In the blot IV, AChE protein in sucrose fractions from the pooled control brain extract was immuno-stained by the anti-AChE-R Ab and demonstrates that the heavy complexes containing the AChE-R variant are present not only in CSF but also in brain tissue. MW marker in the blot III and -IV was the Cruz Marker. The ‘rAChE’ lanes in the blots I, II and IV was purified recombinant human AChE protein (0.4 µg protein). The gels were precast 10% polyacrylamide gels. The numbers on the top of the blots specify the sucrose fractions (the open symbols in Fig. 2a– c)  used for immunostaining.

Download figure to PowerPoint

In the CSF, but not in the brain homogenate, the signal intensities of these immunostained bands were much stronger than the intensities of the bands in the lanes corresponding to the sucrose fractions 19–22. This observation indicated that the protein levels of AChEs in the G2 + G1 peak (Fig. 2a) were much higher than that in the G4 peak in CSF, contrasting the 10 times higher enzyme activity in the G4 peak.

To control for a confounding effect such as stripping and re-probing the blots by the second pair of the antibodies, we used duplicates of the blot. Under non-reducing conditions, the anti-AChE-R Ab stained bands a and b, but not the band c, in the fractions 19–22. In the fractions 26–31 the bands b* and c* were detected in both pooled CSF and brain homogenate (Fig. 3, blots II and IV). Thus, the anti-AChE-R Ab detected specifically some of the bands that were also stained by the anti-Core Ab, providing strong evidence for the specificity of the anti-AChE-R Ab and incorporation of AChE-R subunits in the heavier complexes of AChE. Furthermore, a comparison between blots I and II in Fig. 3 indicated that the band c should be composed of another AChE subunit than AChE-R variant, as this band was detected only by the anti-Core Ab. However, bands a, b, b* and c* could be a mixture of different AChE variants including the readthrough AChE.

The collected sucrose fractions were also used for immunostaining with the anti-Core Ab and the AChE-R Ab after resolution of the protein under reducing SDS–PAGE (Fig. 4). In the lanes corresponding to sucrose fractions 18–20 only one band, e′ (∼50 kDa), could be detected by the anti-Core Ab, most likely due to the detection limit of the antibody. However, the anti-Core Ab stained several other lighter bands (Fig. 4, blot I, lane 30: a′∼100 kDa; lanes 32–34: b′∼75 kDa; lanes 31–33: c′∼65 kDa, and the duplicates band in lanes 30–34: d′ and e′ bands with a MW of ∼55 and ∼50 kDa, respectively). A comparison between the ‘CSF’ lane in the left panel of blot I in Fig. 1 and the lanes 31–33 in Fig. 4 suggested that the dispersed 65–75 kDa could successfully be resolved as the b′ and c′ bands after the sucrose gradient sedimentation procedure.

image

Figure 4. Reducing immunoblot analysis of AChE isoforms isolated by sucrose density gradient technique. In blot I is shown several lighter AChE isoforms, which were detected by the anti-Core Ab , after reduction of heavier complexes of AChE in sucrose fractions of brain homogenate of the control. Five bands was stained (band a′∼100, b′∼75, c′∼65, d′∼55 and e′∼50 kDa). Blot II shows AChE protein in sucrose fractions from CSF of the AD patients. Blot II was immuno-stained by the anti-AChE-R Ab , illustrating that some of the detected bands had a c-terminal unique to readthrough AChE. MW marker in these two blots was the ‘ECL DualVue’. The gels were precast 4–20% polyacrylamide gels. The numbers on the top of the blots specify the sucrose fractions (the open symbols in Fig. 2a– c)  used for immunostaining.

Download figure to PowerPoint

Under reducing conditions, the anti-AChE-R Ab stained only three bands of the five bands detected with the anti-Core Ab: a′ (∼100 kDa, Fig. 4, blot II, lanes: 27–29), c′ (∼65 kDa, lanes: 29–31) and e′ (∼50 kDa, lanes 19–21 and 27–31). Thus, the 100 kDa band, a′, could represent a G2 or a corresponding heteromeric complex of the 50-kDa AChE-R variant, which provide further evidence indicating that AChE-R variant could produce heavier complexes.

Quantification of protein levels of AChE isoforms in CSF

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

Densitometric analysis of reducing SDS–PAGE immunoblots revealed changes in the protein levels of CSF AChE variants. In the rivastigmine group, the protein level of the 50-kDa AChE-R (G1) was the only one to be mildly increased (10%, p < 0.05) at 12 months compared to baseline. The protein level of G1 AChE-S (55-kDa) and G1 AChE-R (65-kDa) bands remained essentially unchanged up to 12 months treatment (Darreh-Shori et al. 2002). In the tacrine group, the protein level of all bands was elevated to a much higher extent compared with the baseline (Fig. 5a and Table 2): the 50-kDa AChE-R (67–91% increase), the 65-kDa AChE-R isoform (29–62%), the 55-kDa AChE-S isoform (63% at 3, but not at 12 months) and the 70-kDa AChE-S (about 44% increase, Fig. 5a and Table 2). In the untreated group, 13–37% reduction in the protein levels of all of these bands was found at the end of the follow up (Fig. 5a and Table 2).

image

Figure 5. Densitometric quantification of long-term changes in the protein levels of different AChE isoforms in the CSF of the treated (tacrine and rivastigmine group) and untreated AD patients. (a)  Percentage changes of the protein levels of some of the AChE splice variants detected under reducing conditions at 3 and 12 months compared to the baseline levels. (b)  Percentage changes of protein levels of heavy AChE complexes detected under non-reducing conditions. The error bars represent SEM.

Download figure to PowerPoint

Table 2.  Summary of the densitometric analysis of the changes in the composition of different AChE splice variants and multimeric complexes in treated and untreated AD patients
AChE molecular variants in CSFRivastigmine groupTacrine groupUntreated group
3 months12 months3 months12 months12 months
  1. The arrows, [UPWARDS ARROW]or [DOWNWARDS ARROW], indicate direction of the changes in protein levels of AChE variants. The number of arrows indicates the significant levels (p < 0.05 and p < 0.001, respectively). The horizontal arrows, ⇆, indicate no significant changes in the protein level. All values are expressed as the percentages of changes compared to the baseline levels.

Light isoforms (G1)
 AChE-R (50 kDa)10%⇆10%[UPWARDS ARROW]67%[UPWARDS ARROW][UPWARDS ARROW]91%[UPWARDS ARROW]32%[DOWNWARDS ARROW][DOWNWARDS ARROW]
 AChE-R (65 kDa)– 8%⇆− 5%⇆29%[UPWARDS ARROW]62%[UPWARDS ARROW]37%[DOWNWARDS ARROW]
 AChE-S (55 kDa)− 8%⇆− 1%⇆63%[UPWARDS ARROW]18%⇆28%[DOWNWARDS ARROW]
 AChE-S (70 kDa)− 7%⇆− 5%⇆44%[UPWARDS ARROW]43%[UPWARDS ARROW]13%[DOWNWARDS ARROW]
Medium sized form (G2)
 AChE-S 130 kDa21%[DOWNWARDS ARROW][DOWNWARDS ARROW]14%[DOWNWARDS ARROW]98%[UPWARDS ARROW]117%[UPWARDS ARROW]32%[UPWARDS ARROW]
Heavy complexes
 AChE-R (> 240 kDa)10%[UPWARDS ARROW]8%⇆54%[UPWARDS ARROW][UPWARDS ARROW]59%[UPWARDS ARROW]37%[DOWNWARDS ARROW]
 AChE-S (> 240 kDa)5%⇆10%⇆33%[UPWARDS ARROW][UPWARDS ARROW]36%[UPWARDS ARROW]− 13%⇆

Changes in the protein levels of the multimeric AChE complexes in the CSF of the AD patients were quantified following non-reducing resolution of CSF proteins. In the rivastigmine group, the protein level of the heavy AChE-R was mildly increased (10%, p < 0.05) after 3 months of treatment, whereas the protein level of the medium sized G2 AChE-S band was reduced by 14–21% compared to baseline. No significant change was observed in the protein level of G4 AChE-S in this group (Fig. 5b and Table 2). In the tacrine group, again the protein level of all bands was elevated to a much higher extent: the heavy AChE-R (54–59%); the G4 AChE-S (33–36%; Fig. 5b and Table 2). Interestingly, in the tacrine group in contrast to the finding in the rivastigmine group the protein level of medium sized G2 AChE-S was elevated as well (98% at 3 months and 117% at 12 months, Fig. 5b and Table 2). Also, interindividual differences were most prominent in the tacrine group, compared to the other groups, as suggested by much larger SEM values (Fig. 5).

In the untreated group, there was about 37% reduction in the protein level of the heavy AChE-R complex, whereas the medium sized AChE-S complex was elevated by about 32%. No significant change was found in the protein level of the heavy G4 AChE-S complex after one year (Fig. 5b and Table 2).

Changes in MMSE and correlation analysis

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

The demographic data of the patients are summarized in Table 1. The mean age of the untreated patients was higher than the mean age of patients in the tacrine group, but not higher than that in the rivastigmine group. No significant difference was found between the three patient groups for the MMSE values at baseline.

There was a significant decline in cognition, as assessed by percent changes of MMSE, in the untreated AD group at the end of follow-up compared to their baseline values (16%, p < 0.03). Although, the percent changes of MMSE also showed a reduction in the tacrine group (∼7%, p > 0.07) and in the rivastigmine group (∼4%, p > 0.2), the reduction was statistically not significant, indicating that the AD patients in both treated groups preserved their cognitive ability after one year compared to the baseline levels.

To see whether the changes in the protein levels of the AChE isoforms could correlate to the percent changes in MMSE at 12 months, we performed ‘stepwise regression’ analysis (backward procedure, F-to-Enter = 4.000 and F-to-Remove = 3.996). We found that the percentage changes in the protein level of four AChE isoforms were correlated to percent changes in MMSE at 12 months (R = 0.7, F = 6.996 and p < 0.0006, n = 31). This analysis suggested that the percent changes in the heavy complex and the 50-kDa G1 of AChE-R bands positively correlated with percent changes in MMSE (% the heavy complex: p < 0.0007 and %50-kDa G1: p < 0.05), whereas the percent changes in the 65-kDa AChE-R and in particular medium sized AChE-S showed an inverse correlation with percent changes in MMSE at 12 months (%G1: p < 0.02 and %G2: p < 0.0003). Considering the contribution of all of these bands we calculated the ratio of AChE-R to AChE-S (R/S ratio). R/S ratio was defined as the ratio of the average percent changes in protein levels of the AChE-R [(%50-kDa G1 + %the heavy complex)/2] after 12 months to the average percent changes in protein levels of the other two AChE bands [(%65-kDa G1 + %G2)/2] after 12 months. The R/S ratio strongly correlated with percent changes in MMSE at 12 months (p < 0.0001, Fig. 6), suggesting that in general the AD patients with larger R/S ratio were quite stable in their cognitive function after one year compared to their baseline levels.

image

Figure 6. Correlation between the percentage changes in MMSE of AD patients and the ratio of percentage changes of AChE-R to AChE-S at 12 months, as was defined in the result section. It illustrates the stabilization of cognition in the majority of the treated AD patients versus worsening of cognition in the untreated AD patients as a function of long-term changes in the protein level of different AChE variants and molecular isoforms in CSF. The letters ‘T’, ‘R’ and ‘U’ indicate the AD patients in the tacrine, the rivastigmine and the untreated group, respectively. The dashed lines represent the 95% confidence interval. It should be noted that the scales on the x -axis of this graph are in reversed order.

Download figure to PowerPoint

A comparison between the three groups showed that the R/S ratio was higher in both treated groups (p < 0.002 for the rivastigmine group and p < 0.05 for the tacrine group) than in the untreated group. This ratio was 1.3 ± 0.09 in the rivastigmine group and 1.0 ± 0.10 in the tacrine group after 12 months, whereas in the untreated group it was 0.8 ± 0.12 at the end of the follow-up time.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

The immunopositive AChE bands, both AChE splice variants (AChE-R and AChE-S) and all AChE bands (monomers and heavier complexes), could be detected in CSF of AD patients as well as in the brain homogenate from the parietal cortex of AD and control subjects. This finding supports the animal study indicating that the origin of AChE in CSF is from CNS neurons (Chubb et al. 1976). Furthermore, recent in situ hybridization analyses, using probes selective for AChE-S and AChE-R mRNA clearly show that both mRNA variants are expressed in neurons of frontal and temporal cortices of both AD and control brains (unpublished experiment). Thus, it is reasonable to assume that the CNS might be at least one of the major sources of AChE proteins in CSF. Changes in the protein levels of CSF AChE variants may then reliably be used as a surrogate marker of the changes in the expression of AChE isoforms in brain's cholinergic neurons. However, it should be noted that a contribution from other tissues was not investigated in the current study.

We found that AChE-R was not exclusively monomeric, as part of it was incorporated into heavier AChE complexes. This is an unexpected finding since the C-terminal peptide of this variant lacks the cysteine residue essential for the formation of the classical disulfide intersubunits linkage in the heavy AChE molecules (Li et al. 1991). Then, we performed additional analysis after separation of the different molecular forms of AChE by sedimentation analysis and the subsequent immunostaining of these molecular forms in the sucrose fractions. The results strongly supported the notion that the AChE-R variant could be incorporated into heavier complexes of AChE. In fact under non-reducing condition, two out of three heavy complexes of AChE, detected by the anti-Core Ab, were also immunostained by the anti-AChE-R antibody. Although the nature of molecular interactions resulting in inclusion of the AChE-R variants in the heavier AChE complexes is not clear, it has been shown that in tetrameric AChE, not all of the subunits are disulfide-bonded (Liao et al. 1993; Flores-Flores et al. 1996). Thus, the cysteine residues in the C-terminal domain of AChE monomers might not be essential for multimerization of AChE subunits (Perrier et al. 2002).

Nonetheless, recent findings show that the C-terminus of AChE-R interacts with the scaffold protein RACK1 and with protein kinase CβII to form triple complexes (Birikh et al. 2003). This provides an additional tentative explanation for the heavier AChE-R-containing complexes.

The result of the sedimentation analysis in the current study was in agreement with those reported by others (Saez-Valero et al. 1999; Saez-Valero et al. 2000). In the sucrose fractions from pooled CSF sample, the AChE activity in the G4 peak was about 10-fold higher than that in the G2 + G1 peak, whereas in the brain homogenate fractions both peaks had equivalent AChE activities. Interestingly, immunoblot analysis showed that in the CSF fraction, but not in the brain homogenate, the protein level of AChE were much higher in the sucrose fractions corresponding to the G2 + G1 peak than in those corresponding to the G4 peak. This finding is contrasting the 10 times higher enzyme activity in the G4 peak and highlights the necessity of accompanying sucrose density gradient analysis with immunoblotting of the CSF AChE protein level in the collected sucrose fractions. The observation might indicate that the majority of the enzymes in the G2 + G1 peak might be (i) either in an inactive state, (ii) aged enzyme, or (iii) that complex molecular formation would result in an increase in the catalytic efficiency of the subunits in the native G4 complexes.

Non-reducing immuno-blot analysis revealed that the heavier AChE-R and AChE-S complexes were the most abundant AChE species in CSF and cortex. In CSF, the medium-sized AChE-S multimers and the 50-kDa AChE-R (but not 55-kDa AChE-S) monomers were present only in traceable amounts as a free isoform (G4 >> G2 = G1). In the cortex, besides the heavy AChE complexes, we found a mixed population of monomers of both AChE-S and AChE-R splice variants. The major AChE subunits in the cortex had MW of about 65- and 75-kDa, whereas those found in CSF were about 20 kDa lighter, i.e. the 50-kDa AChE-R and 55-kDa AChE-S. The MW differences might reflect differences in glycosylation levels (Kronman et al. 2000; Chitlaru et al. 2002), and/or in vivo C-terminus cleavage (Grisaru et al. 2001). Another likely explanation for the MW differences might be the presence or absence of the 20-kDa PRiMA subunit (Perrier et al. 2002), as it has been shown that both amphiphilic and nonamphiphilic AChE isoforms are present in the CSF (Saez-Valero et al. 1999; Saez-Valero et al. 2000).

Immunoblot techniques with antibodies selective for specific AChE splice variants revealed bi-directional changes in the levels and composition of these variants in the CSF of ChEI-treated as compared with untreated AD patients during a follow-up period of one year (as is summarized in Table 2).

In untreated patients, we observed decreased levels with time of both AChE-S and AChE-R. This is compatible with previous reports, which attributed much of the AChE reduction in the CNS to cholinergic neurons (Atack et al. 1983; Fishman et al. 1986; Arendt et al. 1992). Continuous loss of such neurons as the disease progresses would expectedly be accompanied by reduction in the levels of both AChE variants.

Intriguingly, we observed an opposite process in the CSF of ChEI-treated AD patients, namely, increases in the protein levels of AChE variants, which coincided with the stabilization of their MMSE scores compared to baseline level as well as to the untreated group. The changes in the protein level of AChE splice variants and the monomers and heavy complexes in CSF of the treated and untreated AD patients may reflect the actual molecular changes in the cholinergic neurons in the brain. Therefore, our data provides neurochemical evidence, supporting the clinically observed stabilization of AD in ChEI treated patients. Both tacrine and rivastigmine reversed the AD-induced decrease in AChE protein levels into an increase. The elevation of CSF AChEs in the treated patients might partially be due to induction of the putative feedback mechanism via ACh receptors (Perry et al. 1988; Nitsch et al. 1998; von der Kammer et al. 1998). However, it could also reflect regeneration of cholinergic neurons as a result of stimulation (Swaab et al. 1994) by ChEI treatment. As most of the brain AChE is produced in cholinergic neurons, these findings might explain the stabilization of the disease in AD patients by ChEI-induced rescue of the expression capacities of AChEs. It has also been shown that ChEI therapy increases the regional cerebral blood flow, glucose metabolism and the number of nicotinic ACh receptors in AD patients treated with ChEIs (Nordberg et al. 1992; Nordberg et al. 1997; Nordberg et al. 1998). Interestingly, the assessment of cerebral glucose metabolism (CMRglc) and CSF-tau measurements in these patients also indicate distinct long-term responses in treated and untreated and show differences between rivastigmine and tacrine treated patients. A significant increase in CMRglc has been found in the rivastigmine group compared to both tacrine-treated and untreated group. While CSF-tau level has been unchanged in the rivastigmine-treated group, it is increased in both tacrine-treated and -untreated patients (Stefanova et al. 2003).

It has been reported that AChE-R accumulates in response to acute stress and treatment with ChEIs (Kaufer et al. 1998; Shapira et al. 2000). In the current study, indeed, the expression of G1 and the heavier AChE-R appeared to be more sensitive to the effect of treatment or lack of the treatment than that of AChE-S.

Another difference between these three patient groups involved the protein level of medium-sized AChE, which was reduced in the rivastigmine-treated AD patients, but greatly increased in both untreated and tacrine-treated AD patients (over 100% increase).

Animal studies have shown that increased AChE-R expression might be neuroprotective, but excess of the synaptic AChE-S intensifies neurodeterioration (Sternfeld et al. 2000). Additional studies suggest that massive accumulation of both AChE-S and AChE-R, in laboratory animals, induces erratic behavior associated with impaired working memory under mild stressors (Cohen et al. 2002) as well as neurodeterioration (Sternfeld et al. 2000). In contrast, AChE-R accumulation alone resulted in neuro-protection associated with prolonged conflict behavior (Birikh et al. 2003).

We found a significant decline of MMSE scores in the untreated AD patients after 1 year compared with their baseline, but no significant decline in the treated groups. The R/S ratio highly correlated with the percent changes in the MMSE at 12 months in all AD patients. This ratio was largest in the rivastigmine group and lowest in the untreated group. Davidsson et al. (2001) have reported that the percent increase of total CSF AChE activity in AD patients treated with galantamine and donepezil correlates with changes in MMSE. They have also found that the increase in total CSF AChEs is higher in the responders than in non-responders. However, their methodology could not distinguish between the changes in the activity levels or the protein levels of specific AChE isoforms and their correlation with MMSE. The calculated R/S ratio presented in our current report may offer a way to evaluate such modulations, and its apparent association with the changes in MMSE values suggests that it might be clinically relevant.

Conclusions

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

The changes in the protein levels of AChE variants were viewed in two different ways, which showed treatment-specific distinctions. Immunoblot analysis following fully reducing SDS–PAGE revealed the composition of variant polypeptides (AChE-S or AChE-R), whereas analysis following non-reducing electrophoresis demonstrated the formation of multimeric complexes including these polypeptides. Combined together, these analyses point at a complex picture whereby specific ChEIs change the transcription, alternative splicing and interprotein interactions of AChE variants in conjunction with the long-term efficacy of the therapeutic treatment. The long-term changes in the protein levels of AChE variants clearly differed between the treated and untreated groups, as well as between treatment with the reversible inhibitor, tacrine, and the pseudo- irreversible inhibitor, rivastigmine. These findings might be clinically important, suggesting different mechanisms adapted by neurons in response to tacrine and rivastigmine and controlling the stimulatory signals in untreated AD patients. The increased amounts of AChE variants point at up-regulation of AChE gene expression under ChEI treatment, in agreement with the previous reports (Nitsch et al. 1998; Kaufer et al. 1999), whereas changes in the composition of these variants reflect long-lasting alternative splicing modulation (Meshorer et al. 2002).

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References

The authors would like to thank Dr David Glick, Jerusalem, for critically reviewing the manuscript and for constructive comments. This research was sponsored by Swedish Research Council (project no. 05817), Stiftelsen, for Gamla Tjänarinnor, KI foundations, Stohne's foundation, Swedish Alzheimer Foundation (to A.N) and Ester Neuroscience (to HS). Flores-Flores C. was supported by a long-term FEBS fellowship during his postdoctoral period in Jerusalem.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Materials
  5. Patients and collection of CSF samples
  6. Cortical extract from AD and control brain
  7. Reducing and non-reducing SDS–PAGE
  8. Detection of AChE variants in cortex and CSF
  9. Sedimentation analysis
  10. Quantification of AChE isoforms in CSF samples
  11. Statistical analysis
  12. Results
  13. Comparative analysis of AChE variants in cortex and CSF
  14. Immunodetection by anti-Core antibody
  15. Immunodetection by anti-AChE-R antibody
  16. Immunodetection of AChE isoforms in sucrose gradient fractions
  17. Quantification of protein levels of AChE isoforms in CSF
  18. Changes in MMSE and correlation analysis
  19. Discussion
  20. Conclusions
  21. Acknowledgements
  22. References
  • Arendt T., Bruckner M. K., Lange M. and Bigl V. (1992) Changes in acetylcholinesterase and butyrylcholinesterase in Alzheimer's disease resemble embryonic development – a study of molecular forms. Neurochem. Int. 21, 381396.
  • Atack J. R., Perry E. K., Bonham J. R., Perry R. H., Tomlinson B. E., Blessed G. and Fairbairn A. (1983) Molecular forms of acetylcholinesterase in senile dementia of Alzheimer type: selective loss of the intermediate (10S) form. Neurosci. Lett. 40, 199204.
  • Birikh K. R., Sklan E. H., Shoham S. and Soreq H. (2003) Interaction of ‘readthrough’ acetylcholinesterase with RACK1 and PKCbeta II correlates with intensified fear-induced conflict behavior. Proc. Natl Acad. Sci. USA 100, 283288.
  • Chitlaru T., Kronman C., Velan B. and Shafferman A. (2002) Overloading and removal of N-glycosylation targets on human acetylcholinesterase: effects on glycan composition and circulatory residence time. Biochem. J. 363, 619631.
  • Chubb I. W., Goodman S. and Smith A. D. (1976) Is acetylcholinesterase secreted from central neurons into the cerebral fluid? Neuroscience. 1, 5762.
  • Cohen O., Erb C., Ginzberg D., Pollak Y., Seidman S., Shoham S., Yirmiya R. and Soreq H. (2002) Neuronal overexpression of ‘readthrough’ acetylcholinesterase is associated with antisense-suppressible behavioral impairments. Mol.Psychiatry 7, 874885.
  • Darreh-Shori T., Almkvist O., Guan Z. Z., Garlind A., Strandberg B., Svensson A. L., Soreq H., Hellstrom-Lindahl E. and Nordberg A. (2002) Sustained cholinesterase inhibition in AD patients receiving rivastigmine for 12 months. Neurology 59, 563572.
  • Davidsson P., Blennow K., Andreasen N., Eriksson B., Minthon L. and Hesse C. (2001) Differential increase in cerebrospinal fluid-acetylcholinesterase after treatment with acetylcholinesterase inhibitors in patients with Alzheimer's disease. Neurosci. Lett. 300, 157160.
  • Enz A., Boddeke H., Gray J. and Spiegel R. (1991) Pharmacologic and clinicopharmacologic properties of SDZ ENA 713, a centrally selective acetylcholinesterase inhibitor. Ann. NY Acad. Sci. 640, 272275.
  • Fishman E. B., Siek G. C., MacCallum R. D., Bird E. D., Volicer L. and Marquis J. K. (1986) Distribution of the molecular forms of acetylcholinesterase in human brain: alterations in dementia of the Alzheimer type. Ann. Neurol. 19, 246252.
  • Flores-Flores C., Martinez-Martinez A., Munoz-Delgado E. and Vidal C. J. (1996) Conversion of acetylcholinesterase hydrophilic tetramers into amphiphilic dimers and monomers. Biochem. Biophys. Res. Commun. 219, 5358.
  • Folstein M. F., Folstein S. E. and McHugh P. R. (1975) ‘Mini-mental state’: a practical method for grading the cognitive state of patients for the clinician. J. Psychiatric Res. 12, 189198.
  • Giacobini E. (1997) From molecular structure to Alzheimer therapy. Jpn J. Pharmacol. 74, 225241.
  • Giacobini E. (1998) Invited review: cholinesterase inhibitors for Alzheimer's disease therapy: from tacrine to future applications. Neurochem. Int. 32, 413419.
  • Giacobini E. (2000) Cholinesterase inhibitors stabilize Alzheimer's disease. Ann. NY Acad. Sci. 920, 321327.
  • Grisaru D., Deutsch V., Shapira M., Pick M., Sternfeld M., Melamed-Book N., Kaufer D., Galyam N., Gait M. J., Owen D. et al. (2001) ARP, a peptide derived from the stress-associated acetylcholinesterase variant, has hematopoietic growth promoting activities. Mol. Med. 7, 93105.
  • Grisaru D., Sternfeld M., Eldor A., Glick D. and Soreq H. (1999) Structural roles of acetylcholinesterase variants in biology and pathology. Eur. J. Biochem. 264, 672686.
  • Von Der Kammer H., Mayhaus M., Albrecht C., Enderich J., Wegner M. and Nitsch R. M. (1998) Muscarinic acetylcholine receptors activate expression of the EGR gene family of transcription factors. J. Biol. Chem. 273, 1453814544.
  • Kaufer D., Friedman A., Seidman S. and Soreq H. (1998) Acute stress facilitates long-lasting changes in cholinergic gene expression. Nature 393, 373377.
  • Kaufer D., Friedman A., Seidman S. and Soreq H. (1999) Anticholinesterases induce multigenic transcriptional feedback response suppressing cholinergic neurotransmission. Chem. Biol. Interact. 119–120, 349360.
  • Kronman C., Chitlaru T., Elhanany E., Velan B. and Shafferman A. (2000) Hierarchy of post-translational modifications involved in the circulatory longevity of glycoproteins: demonstration of concerted contributions of glycan sialylation and subunit assembly to the pharmacokinetic behavior of bovine acetylcholinesterase. J. Biol. Chem. 275, 2948829502.
  • Li Y., Camp S., Rachinsky T. L., Getman D. and Taylor P. (1991) Gene structure of mammalian acetylcholinesterase: alternative exons dictate tissue-specific expression. J. Biol. Chem. 266, 2308323090.
  • Liao J., Norgaard-Pedersen B. and Brodbeck U. (1993) Subunit association and glycosylation of acetylcholinesterase from monkey brain. J. Neurochem. 61, 11271134.
  • Marutle A., Warpman U., Bogdanovic N., Lannfelt L. and Nordberg A. (1999) Neuronal nicotinic receptor deficits in Alzheimer patients with the Swedish amyloid precursor protein 670/671 mutation. J. Neurochem. 72, 11611169.
  • Massoulie J., Anselmet A., Bon S., Krejci E., Legay C., Morel N. and Simon S. (1998) Acetylcholinesterase: C-terminal domains, molecular forms and functional localization. J. Physiol. Paris 92, 183190.
  • Meshorer E., Erb C., Gazit R., Pavlovsky L., Kaufer D., Friedman A., Glick D., Ben-Arie N. and Soreq H. (2002) Alternative splicing and neuritic mRNA translocation under long-term neuronal hypersensitivity. Science 295, 508512.
  • Nitsch R. M., Rossner S., Albrecht C., Mayhaus M., Enderich J., Schliebs R., Wegner M., Arendt T. and Von Der Kammer H. (1998) Muscarinic acetylcholine receptors activate the acetylcholinesterase gene promoter. J. Physiol. Paris 92, 257264.
  • Nordberg A. and Svensson A. L. (1998) Cholinesterase inhibitors in the treatment of Alzheimer's disease: a comparison of tolerability and pharmacology. Drug Saf. 19, 465480.
  • Nordberg A., Lilja A., Lundqvist H., et al. (1992) Tacrine restores cholinergic nicotinic receptors and glucose metabolism in Alzheimer patients as visualized by positron emission tomography. Neurobiol. Aging 13, 747758.
  • Nordberg A., Lundqvist H., Hartvig P., Andersson J., Johansson M., Hellstrom-Lindahi E. and Langstrom B. (1997) Imaging of nicotinic and muscarinic receptors in Alzheimer's disease: effect of tacrine treatment. Dement. Geriatr. Cogn. Disord. 8, 7884.
  • Nordberg A., Amberla K., Shigeta M., et al. (1998) Long-term tacrine treatment in three mild Alzheimer patients: effects on nicotinic receptors, cerebral blood flow, glucose metabolism, EEG, and cognitive abilities. Alzheimer Dis. Assoc. Disord. 12, 228237.
  • Nordberg A., Hellström-Lindahl E., Almkvist O. and Meurling L. (1999) Activity of acetylcholinesterase in CSF increases in Alzheimer's patients after treatment with tacrine. Alzheimer's Rep. 2, 347352.
  • Perrier A. L., Massoulie J. and Krejci E. (2002) PRiMA: the membrane anchor of acetylcholinesterase in the brain. Neuron 33, 275285.
  • Perry E. K., Smith C. J., Court J. A., Bonham J. R., Rodway M. and Atack J. R. (1988) Interaction of 9-amino-1,2,3,4-tetrahydroaminoacridine (THA) with human cortical nicotinic and muscarinic receptor binding in vitro. Neurosci. Lett. 91, 211216.
  • Saez-Valero J., Sberna G., McLean C. A. and Small D. H. (1999) Molecular isoform distribution and glycosylation of acetylcholinesterase are altered in brain and cerebrospinal fluid of patients with Alzheimer's disease. J. Neurochem. 72, 16001608.
  • Saez-Valero J., Barquero M. S., Marcos A., McLean C. A. and Small D. H. (2000) Altered glycosylation of acetylcholinesterase in lumbar cerebrospinal fluid of patients with Alzheimer's disease. J. Neurol. Neurosurg. Psychiatry 69, 664667.
  • Shapira M., Tur-Kaspa I., Bosgraaf L., Livni N., Grant A. D., Grisaru D., Korner M., Ebstein R. P. and Soreq H. (2000) A transcription-activating polymorphism in the ACHE promoter associated with acute sensitivity to anti-acetylcholinesterases. Hum. Mol. Genet. 9, 12731281.
  • Shohami E., Kaufer D., Chen Y., Seidman S., Cohen O., Ginzberg D., Melamed-Book N., Yirmiya R. and Soreq H. (2000) Antisense prevention of neuronal damages following head injury in mice. J. Mol. Med. 78, 228236.
  • Siek G. C., Katz L. S., Fishman E. B., Korosi T. S. and Marquis J. K. (1990) Molecular forms of acetylcholinesterase in subcortical areas of normal and Alzheimer disease brain. Biol. Psychiatry 27, 573580.
  • Soreq H. and Seidman S. (2001) Acetylcholinesterase – new roles for an old actor. Nat. Rev. Neurosci. 2, 294302.
  • Stefanova E., Blennow K., Almkvist O., Hellstrom-Lindahl E. and Nordberg A. (2003) Cerebral glucose metabolism, cerebrospinal fluid-beta-amyloid1–42 (CSF- Abeta42), tau and apolipoprotein E genotype in long-term rivastigmine and tacrine treated Alzheimer disease (AD) patients. Neurosci. Lett. 338, 159163.
  • Sternfeld M., Shoham S., Klein O., Flores-Flores C., Evron T., Idelson G. H., Kitsberg D., Patrick J. W. and Soreq H. (2000) Excess ‘read-through’ acetylcholinesterase attenuates but the ‘synaptic’ variant intensifies neurodeterioration correlates. Proc. Natl Acad. Sci. USA 97, 86478652.
  • Svensson A. L. and Nordberg A. (1996) Tacrine interacts with an allosteric activator site on alpha 4 beta 2 nAChRs in M10 cells. Neuroreport 7, 22012205.
  • Svensson A. L., Zhang X. and Nordberg A. (1996) Biphasic effect of tacrine on acetylcholine release in rat brain via M1 and M2 receptors. Brain Res. 726, 207212.
  • Swaab D. F., Hofman M. A., Lucassen P. J., Salehi A. and Uylings H. B. (1994) Neuronal atrophy, not cell death, is the main hallmark of Alzheimer's disease. Neurobiol. Aging. 15, 369371; discussion 379–380.
  • Tomkins O., Kaufer D., Korn A., Shelef I., Golan H., Reichenthal E., Soreq H. and Friedman A. (2001) Frequent blood–brain barrier disruption in the human cerebral cortex. Cell. Mol. Neurobiol. 21, 675691.
  • Wagstaff A. J. and McTavish D. (1994) Tacrine: a review of its pharmacodynamic and pharmacokinetic properties, and therapeutic efficacy in Alzheimer's disease. Drugs Aging 4, 510540.