Pathogenic Effects of D23N Iowa Mutant Amyloid β-Protein

Cerebral amyloid β-protein angiopathy (CAA) is a key pathological feature of patients with Alzheimer's disease and certain related disorders. In these conditions the CAA is characterized by the deposition of Aβ within the cerebral vessel wall and, in severe cases, hemorrhagic stroke. Several mutations have been identified within the Aβ region of the Aβ protein precursor (AβPP) gene that appear to enhance the severity of CAA. We recently described a new mutation within the Aβ region (D23N) of AβPP that is associated with severe CAA in an Iowa kindred (Grabowski, T. J., Cho, H. S., Vonsattel, J. P. G., Rebeck, G. W., and Greenberg, S. M. (2001) Ann. Neurol. 49, 697–705). In the present study, we investigated the effect of this new D23N mutation on the processing of AβPP and the pathogenic properties of Aβ. Neither the D23N Iowa mutation nor the E22Q Dutch mutation affected the amyloidogenic processing of AβPP expressed in H4 cells. The A21G Flemish mutation, in contrast, resulted in a 2.3-fold increase in secreted Aβ peptide. We also tested synthetic wild-type and mutant Aβ40 peptides for fibrillogenesis and toxicity toward cultured human cerebrovascular smooth muscle (HCSM) cells. The E22Q Dutch, D23N Iowa, and E22Q,D23N Dutch/Iowa double mutant Aβ40 peptides rapidly assembled in solution to form fibrils, whereas wild-type and A21G Flemish Aβ40 peptides exhibited little fibril formation. Similarly, the E22Q Dutch and D23N Iowa Aβ40 peptides were found to induce robust pathologic responses in cultured HCSM cells, including elevated levels of cell-associated AβPP, proteolytic breakdown of smooth muscle cell α-actin, and cell death. Double mutant E22Q,D23N Dutch/Iowa Aβ40 was more potent than either single mutant form of Aβ in causing pathologic responses in HCSM cells. These data suggest that the different CAA mutations in AβPP may exert their pathogenic effects through different mechanisms. Whereas the A21G Flemish mutation appears to enhance Aβ production, the E22Q Dutch and D23N Iowa mutations enhance fibrillogenesis and the pathogenicity of Aβ toward HCSM cells.

Cerebral amyloid angiopathy (CAA) 1 is a common pathology found at increased frequency in patients with Alzheimer's disease (AD) and related disorders such as Down's syndrome and hereditary cerebral hemorrhage with amyloidosis of the Dutch type (1)(2)(3)(4)(5)(6)(7). A␤ is a 39 -43-amino acid peptide that is proteolytically derived from its larger parent transmembrane molecule amyloid-␤ precursor protein (A␤PP) (8). Although A␤PP exhibits a variety of biological activities, its role as the parent molecule of the A␤ peptide has drawn the most interest (9). In this regard, full-length A␤PP can undergo proteolytic cleavage by ␤and ␥-secretases to liberate the A␤ peptide. Alternatively, fulllength A␤PP can be proteolytically processed by an enzyme termed ␣-secretase at position 16 of the A␤ domain, resulting in a non-amyloidogenic membrane-spanning carboxyl-terminal fragment and truncated secretory forms of A␤PP that are released into the extracellular environment.
Mutations in the A␤PP gene have been linked to several familial diseases. Substitutions at residues flanking the A␤ region of A␤PP give rise to early-onset AD, apparently through effects on the processing of A␤PP and production of A␤ (8). Other mutations have been identified that fall within the A␤ region of A␤PP, specifically at the adjacent A␤ residues 21 and 22 (10 -14). Interestingly, mutations at these residues appear to associate preferentially with CAA (10 -14). Extensive CAA can associate with several clinical syndromes, including intracerebral hemorrhagic stroke (2,15) and dementia with white matter destruction (16,17).
We recently identified a mutation at a third site within the A␤ region of A␤PP that resulted in substitution of asparagine for aspartic acid at A␤ position 23 (A␤PP D694N) (18). The mutation was carried by an Iowa family with a three-generation history of autosomal dominant dementia with onset in the sixth or seventh decade and, in two patients studied radiographically, extensive white matter abnormalities and posterior cortical calcifications. Neuropathological examination of the proband revealed severe CAA with numerous small cortical hemorrhages and both cortical and subcortical infarctions. Abundant neurofibrillary tangles and dystrophic neurites were also present, a possible result of either a direct toxic effect of the mutant amyloid peptide or hypoxia due to the severe vascular narrowing. A␤ plaques were relatively sparse and generally of diffuse morphology.
In the present study, we investigated the pathogenic mechanism of the Iowa A␤PP mutation. Studies of position 22 mutations suggest that they may generate an A␤ peptide with properties of altered fibrillogenesis (19 -21) and increased toxicity to cerebrovascular cells (21)(22)(23)(24)(25)(26). The alternative possibility is that the Iowa mutation may affect the production rather than the biochemical properties of A␤, a mechanism that has been suggested for the Flemish glycine for alanine substitution at position 21 (27,28). Our data indicate that the Iowa mutation results in increased toxicity of A␤ to cerebrovascular cells in culture.

EXPERIMENTAL PROCEDURES
Materials-A␤ peptides were synthesized by solid-phase Fmoc (9fluorenylmethoxycarbonyl) amino acid chemistry, purified by reverse phase high performance liquid chromatography, and structurally characterized as previously described (29). The lyophilized A␤ peptides were first resuspended in sterile distilled water to a concentration of 250 M. Before addition to HCSM cells, the peptides were diluted to a final concentration of 25 M in serum-free culture media. The secondary structures of the resuspended peptides were determined by circular dichroism spectroscopy as previously described (30). The anti-A␤PP mouse monoclonal antibody (mAb) P2-1, which specifically recognizes a unique epitope in the amino-terminal region of human A␤PP, was prepared as previously described (31). The anti-A␤ mouse mAb 6E10 was obtained from Senetek (Napa, CA). Thioflavin T and Congo red were purchased from Sigma. Secondary peroxidase-coupled sheep antimouse IgG and peroxidase-coupled donkey anti-rabbit IgG were purchased from Amersham Pharmacia Biotech. Supersignal Dura West chemiluminescence substrate was purchased from Pierce.
Mutant APP Clones-The SmaI/ClaI fragment of cDNA of the human Dutch (E693Q) A␤PP751 (a kind gift of Dr. Konrad Beyreuther) was cloned in the BamHI/EcoRV restriction sites of the pcDNA3.1(ϩ) expression vector (Invitrogen, Carlsbad, CA). Wild-type A␤PP was generated using the QuikChange TM site-directed mutagenesis kit (Stratagene, La Jolla, CA); A␤PP Flemish (A692G) and A␤PP Iowa (D694N) were generated from the wild-type A␤PP. The oligonucleotides used for mutagenesis are listed below, with the changed codon underlined: A␤PP751 wild-type, sense primer 5Ј-GGT GTT CTT TGC AGA AGA TGT GGG TTC AAA CAA AGG TGC-3Ј and its complementary antisense primer; A␤PP751 Iowa (Asp-694), sense primer 5Ј-GGT GTT CTT TGC AGA AAA TGT GGG TTC AAA CAA AGG TGC-3Ј and its complementary antisense primer; A␤PP751 Flemish (A692G), sense primer 5Ј-GGT GTT CTT TGG AGA AGA TGT GGG TTC AAA CAA AGG TGC-3Ј and its complementary antisense primer. Sequences of all constructs were confirmed by DNA sequencing and restriction analysis.
Cell Culture and Transient Transfections-Human H4 neuroglioma cells (28,32) were grown in Opti-MEM medium (Life Technologies/ Invitrogen) supplemented with 10% (v/v) fetal bovine serum in 100-mm dishes. Nearly confluent cells were transfected with 10 g of A␤PP constructs using Superfect according to company protocol (Life Technologies, Inc./Invitrogen); experiments were performed four times.
Transfected cells and conditioned media were collected after 48 h. The media samples were centrifuged at 14,000 ϫ g for 5 min to remove cellular debris and stored at Ϫ20°C before use. Cells were washed with 10 ml of PBS and solubilized in 50 mM Tris-HCl (pH 8.0), 0.5 M NaCl, 0.1% Triton X-100 with 400 M phenylmethane sulfonylfluoride, 2 M pepstatin, 2 M leupeptin, and 1.5 M aprotinin; DNA was sheared via a 23-gauge needle, and insoluble cellular debris was removed by centrifugation.
Immunoblotting of A␤PP and Fragments in Transfected Cells-Proteins from cell extracts and conditioned media were separated under reducing conditions by SDS Tris-Tricine 10 -20% polyacrylamide gel electrophoresis (Novex, San Diego, CA). The separated proteins were transferred onto a polyvinylidene difluoride membrane at 200 mA for 2 h and blocked with 5% nonfat dry milk. The membrane was incubated with antibodies C8 (1:500) for the full-length A␤PP and C83 fragments (antibody kindly provided by Dr. Dennis Selkoe) or mAb 6E10 (1:1000) for secreted A␤PP␣ at 4°C for overnight. The horseradish peroxidaseconjugated anti-rabbit Ig secondary antibody or anti-mouse Ig secondary antibody was visualized by ECL detection system (Amersham Pharmacia Biotech). Only faint immunoreactivity was observed in cells transfected with vector alone. A␤PP-specific bands from transfected cells were quantitated from film using a Bio-Rad GS-700 densitometer. Relative levels of immunoreactivities were compared by analysis of variance (Statview, Abacus Concepts).
A␤40 Sandwich ELISA-A␤40 levels in conditioned media were measured by ELISA using A␤ antibodies 2G3 for capture and biotinylated 3D6 for detection (33). 2G3 was developed against A␤ 33-40 and does not cross-react with A␤42; 3D6 recognizes A␤1-5 and detects only A␤ with the amino-terminal aspartic acid (33). 96-Well plates (Costar, NY) were coated overnight at room temperature with 2G3 (10 g/ml) and blocked in 0.25% human serum albumin for 1 h at room temperature. Then 100 l of A␤40 standards and samples of conditioned media (in triplicate) were applied for 1 h and washed three times with 0.05% Tween 20 in Tris-buffered saline. Then 100 l of biotinylated 3D6 (0.2 g/ml) was applied for 1 h and washed three times with 0.05% Tween 20 in Tris-buffered saline and followed by 100 l of horseradish peroxidase-avidin (1:2000) for 1 h. After three washes with Tris-buffered saline, the reaction was developed with the colorimetric substrate Slow 3,3Ј,5,5Ј-tetramethylbenzidine-ELISA (Pierce) and stopped after 15 min by addition of 2 M sulfuric acid, and the absorbance was read at 450 nm. Levels of secreted A␤40 were compared by analysis of variance (Statview).
Assembled A␤ Congo Red Binding/Precipitation Assay-Lyophylized A␤ peptides were initially resuspended in distilled water to a concentration of 200 M and then immediately diluted to a final concentration of 100 M in 100 mM Tris-HCl (pH 7.5). Triplicate samples of each peptide were continuously incubated at 37°C on a rocking platform. At the designated time points, 100-l aliquots of each sample were collected and incubated with 2 l of a sterile filtered Congo red solution (1.5 mg/ml in distilled water) for 1 h at room temperature in the dark. The samples were then centrifuged at 14,000 ϫ g for 10 min, 75 l of the resulting supernatants were placed in a 96-well microtiter plate, and the absorbance was read at 492 nm. The extent of A␤ peptide assembly was reflected in the loss of absorbance compared with buffer and Congo red-only controls.
Transmission Electron Microscopy-Aliquots of the A␤ peptide samples used in Congo red binding experiments described above were  ). B, A␤ in the brain of an individual with the A␤PP Iowa mutation. Total A␤ immunoreactivity was detected in the temporal lobe of a 68-year-old man with the A␤PPD694N mutation using 7-m paraffin-embedded sections pretreated with 99% formic acid and stained for A␤ (antibody 6E10) as described (12). Severe deposition of A␤ in the walls of cerebral blood vessels but very little parenchymal A␤ was observed. Scale bar ϭ 200 m. placed on carbon Formvar-coated 200-mesh copper grids and negatively stained with 2% uranyl acetate. The grids were viewed with a JEOL 1200 EX transmission electron microscope at 80 kV.
HCSM Cell Culture-Primary cultures of HCSM cells were established and characterized as previously described (34). The HCSM cells used in these studies were between passages 4 -7 and maintained in 24-well tissue culture dishes with Dulbecco's minimum essential medium containing 10% fetal bovine serum (Gemini Bio-Products, Calabasas, CA), non-essential amino acids, and antibiotics (Life Technologies, Inc.). For experiments, near-confluent cultures of HCSM cells were placed in serum-free medium containing 0.1% bovine serum albumin overnight before treatment. Then 25 M freshly solubilized A␤ peptides were added to the cultures in serum-free medium and incubated at 37°C for the specified lengths of time. Cells were routinely monitored for degenerative morphological changes using an Olympus IX70 phasecontrast microscope. Cell viability was quantified using a fluorescent live/dead cell assay (Molecular Probes, Eugene, OR) following the manufacturer's protocol. The cultured HCSM cells were viewed using an Olympus IX70 fluorescence microscope, and the number of live and dead cells were counted from several fields (n ϭ 4) from at least three separate wells for each experiment.
HCSM Cell Surface A␤ Immunofluorescence-HCSM cells were grown to near confluence in 24-well tissue culture plates. The cells were placed in serum-free medium containing 0.1% bovine serum albumin overnight before treatment. The cells were then incubated in serumfree medium for 6 days in the absence or presence of 25 M freshly solubilized A␤ peptides. The cells were rinsed in PBS and fixed in 2% paraformaldehyde, PBS for 20 min at room temperature. The HCSM cells were rinsed with PBS and then incubated with 200 l/well Protein Blocker solution (Research Genetics, Huntsville, AL) for 30 min. The HCSM cells were again rinsed and then incubated with 250 l of mAb6E10 diluted 1:5000 overnight at 4°C. The following day, the HCSM cells were rinsed with PBS and then incubated with 250 l of a secondary fluorescein isothiocyanate-coupled sheep anti-mouse IgG antibody (Amersham Pharmacia Biotech) diluted 1:200 for 2 h at room temperature. The HCSM cells were rinsed with PBS three times, and 250 l of PBS was then added to each well. HCSM cell-surface A␤ immunofluorescence was measured at excitation wavelength 485 nm and emission wavelength 535 nm using a Cytofluor II fluorescence plate reader (PerkinElmer Life Sciences). Each measurement was performed in triplicate, and five fields were scanned for each well.
HCSM Cell Surface Thioflavin Fluorescence Assay-HCSM cells were grown to near confluence in 24-well tissue culture plates. The cells were placed in serum-free medium containing 0.1% bovine serum albumin overnight before treatment. The cells were then incubated in serum-free medium for 6 days in the absence or presence of 25 M freshly solubilized A␤ peptides. The cells were then rinsed with PBS three times and fixed with 2% paraformaldehyde in PBS for 20 min at room temperature. The HCSM cells were rinsed with PBS after fixation, stained with 0.1% thioflavin T for 10 min at room temperature, and rinsed with 80% ethanol three times. 250 l of PBS was added to each well, and thioflavin T fluorescence was measured at excitation wavelength 440 nm and emission wavelength 485 nm using a Cytofluor II fluorescence plate reader. Each time point was performed in triplicate, and five fields were scanned for each well.
Immunoblotting Analyses of HCSM Cellular Proteins-Near-confluent cultures of HCSM cells were incubated with or without 25 M freshly solubilized A␤ peptides for 6 days. After incubation, the HCSM cells were rinsed with PBS three times, and the cells were solubilized in a lysis buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Triton X-100, 5 mM EDTA, and proteinase inhibitor mixture (Roche Molecular Biochemicals). The cell lysates were centrifuged at 14,000 ϫ g for 10 min to remove insoluble material. Protein concentrations were determined by the method of Bradford (35). Cell lysates were stored at Ϫ70°C until analysis. Aliquots of the cell lysate samples were electrophoresed in non-reducing SDS, 10% polyacrylamide gels, and the proteins were electroblotted onto Hybond nitrocellulose membranes (Amersham Pharmacia Biotech). Unoccupied sites on the membranes were blocked with 5% nonfat milk overnight in PBS with 0.05% Tween 20. The membranes were probed with mAb P2-1 (5 g/ml) or an mAb to smooth muscle cell ␣-actin (1 g/ml) and then incubated with a secondary peroxidase-coupled sheep anti-mouse IgG antibody at a dilution of 1:1000. The peroxidase activity on the membranes was detected using an enhanced chemiluminescence system and analyzed using a Fluor-S Multi-Imager (Bio-Rad) with the manufacturer's Multi-Analyst software.

Iowa Mutation at Position 23 in A␤ Is
Associated with Familial CAA-A newly described mutation at A␤PP position 694, causing an aspartate to asparagine substitution, is associated with familial CAA in a kindred in Iowa (12). This mutation is at position 23 of the A␤ peptide, immediately downstream of two other A␤PP residues with mutations associated with familial CAA, residue 21, A to G substitution (Flemish mutation), and residue 22, E to Q (Dutch), E to K (Italian), and E to G (Arctic) substitutions (Fig. 1A). Immunostaining for A␤ in a brain tissue sample from an individual with the Iowa mutation reveals extensive CAA (Fig. 1B).
Effects of CAA Mutations on A␤PP Processing and A␤ Production-We first determined whether the A␤PP Iowa mutation altered the processing of A␤PP in human H4 neuroglioma cells. These cells are derived from the human central nervous system and have been used to study the effects of A␤PP mutations on A␤PP processing (28,32). The production of full-length A␤PP, secreted APP␣ and C83 (the amino-and carboxyl-terminal products of ␣-secretase cleavage of A␤PP) were measured by quantitative immunoblotting, whereas A␤ (produced from ␤and ␥-secretase cleavage of A␤PP) was measured by ELISA ( Fig. 2). The Flemish, Dutch, and Iowa mutations did not cause a significant change in amounts of full-length A␤PP, secreted APP␣, or C83 (Fig. 2B). No significant differences in the levels of A␤PP fragments were observed after correcting for slight differences in total A␤PP expression. Because we observed a high degree of A␤40 immunoreactivity in both vessel and parenchymal A␤ deposits in the brain of the patient with the Iowa mutation (12), we hypothesized that increased A␤40 production might be associated with this mutation. Using the same transiently transfected cells, we found the Flemish mutation led to a significant 2.3-fold increase in the amount of secreted A␤40 in the conditioned culture medium, confirming earlier reports (27,28). In contrast, the Dutch and Iowa mutations did not affect A␤40 secretion (Fig. 2B).

Effects of CAA Mutations on in Vitro A␤ Fibril Assembly-
Since the newly described Iowa mutation in A␤PP did not affect A␤PP amyloidogenic processing, we next evaluated its affect on A␤ fibril assembly. In these experiments we used a quantitative Congo red binding/precipitation assay to determine the in vitro assembly of wild-type and CAA-mutant A␤40 peptides. We focused our studies on A␤40 peptides because of their abundance in CAA and our previous in vitro investigations on their cerebrovascular cytotoxic effects of A␤40 (22-24, 30, 36). The wild-type and Flemish A␤40 peptides exhibited little assembly over the 48-h time period, with Ͻ15% precipitation of Congo red (Fig. 3). The Dutch and Iowa mutant A␤40 peptides, in contrast, rapidly assembled in solution and precipitated nearly all of the Congo red by 6 h (Fig. 3). The E22Q Dutch and D23N Iowa mutation reside at adjacent sites within A␤, and each results in a loss of a negative charge (Fig. 1A). Therefore, we also included a double mutant Dutch/Iowa A␤40 peptide in these experiments to determine the effect of losing both negative charges on peptide assembly. The double mutant Dutch/ Iowa A␤40 also assembled rapidly, achieving near total depletion of Congo red by 3 h. To further assess the fibrillar nature of each A␤ peptide, we performed circular dichroism spectral analysis to determine the presence of ␤-sheet secondary structure in each preparation after the 48-h incubation period. Dutch, Iowa, and Dutch/Iowa A␤40 each showed strong ␤-sheet secondary structure, consistent with a fibrillar assembly of these peptides, whereas the wild-type and Flemish A␤40 peptides did not exhibit this structure (data not shown).
To confirm that actual fibril assembly occurred, we per- formed transmission electron microscopic analysis of the different A␤ peptide preparations. Consistent with the above findings, Dutch, Iowa, and Dutch/Iowa A␤40 showed abundant characteristic fibrils (Fig. 4), whereas similar structures were rare or absent in preparations of wild-type or Flemish A␤40 (data not shown). Together, these studies demonstrate that Iowa mutant A␤40, like Dutch mutant A␤40, exhibits an increased propensity to assemble into ␤-sheet containing fibrils. Furthermore, this property appears to be more enhanced in double-mutant Dutch/Iowa A␤40.
Effects of CAA Mutations on HCSM Cell Surface A␤ Binding and Fibril Assembly-Previously, we showed that forms of A␤ that are pathogenic for HCSM cells exhibit robust binding and fibril assembly on the surface of these cells (24,30,36). We therefore examined the ability of the CAA-mutant A␤40 peptides to bind to HSCM cells by performing quantitative immunofluorescence studies. As shown in Fig. 5A, after 6 days of incubation with the peptides, neither wild-type A␤40 (lane 1) nor Flemish mutant A␤40 (lane 2) exhibited appreciable binding to the HCSM cell surface. In contrast, in the same experiments, CAA mutant Dutch A␤40 (lane 3), Iowa A␤40, and double-mutant Dutch/Iowa A␤40 showed robust binding on the surface of HCSM cells.
We also used thioflavin T, a dye that fluoresces when bound to amyloid fibrils that contain ␤-sheet, to quantitatively assess the structure of the different CAA mutant A␤ peptides on the surface of HCSM cells. Similar to the above findings, Fig. 5B shows after 6 days of incubation with the peptides only Dutch A␤40, Iowa A␤40, and double mutant Dutch/Iowa A␤40 exhibited increased binding of thioflavin T relative to untreated HCSM cells. The elevated thioflavin T binding observed with this subset of CAA-mutant A␤40 peptides indicates the formation of A␤ fibrils on the surfaces of the HCSM cells. The presence of fibrillar A␤ on the HCSM cell surface with this subset of CAA-mutant A␤40 peptides was confirmed by transmission electron microscopy (data not shown).
Effects of CAA Mutations on A␤-induced Pathologic Responses in Cultured HCSM Cells-Incubation of HCSM cells with pathogenic forms of A␤ leads to increased cell-surface secreted A␤PP, degradation of vascular smooth muscle cell ␣-actin, and subsequent apoptotic cell death (24,30,36,37). We therefore tested the ability of each mutant form of A␤ associated with CAA to induce these pathologic responses in cultured HCSM cells. Near confluent cultures of HCSM cells were incubated in the absence or presence of freshly solubilized of each CAA-mutant (25 M) A␤40 peptide for 6 days at 37°C. After this time, cell lysates were prepared, and the HCSM cellassociated A␤PP or vascular smooth muscle cell ␣-actin was quantified by immunoblotting with mAbs to each respective protein as described under "Experimental Procedures." Incubation with Dutch (lanes 4), Iowa (lanes 5), or double mutant Dutch/Iowa (lanes 6) A␤40 caused a sharp increase in cellassociated A␤PP (Fig. 6) and decrease in vascular smooth muscle cell ␣-actin (Fig. 7) compared with untreated HCSM cells (lanes 1). The effects of double mutant Dutch/Iowa A␤40 on degradation of vascular smooth muscle cell ␣-actin were noticeably more robust than either single mutant form of A␤ (Fig. 7). In contrast, wild-type (lanes 2) or Flemish (lanes 3) A␤40 had no effect on the cell-associated A␤PP levels or vascular smooth muscle cell ␣-actin levels in HCSM cells. In parallel experiments, the extent of HCSM cell death was determined as described under "Experimental Procedures." Similar to the above results, Dutch and Iowa CAA mutant A␤40 peptides induced HCSM cell death, whereas wild-type and Flemish A␤40 peptides did not (Fig. 8). Again, the double mutant Dutch/Iowa A␤40 was found to be much more toxic to the HCSM cells than either single mutant form of A␤.

DISCUSSION
CAA is a significant age-related risk factor for hemorrhagic stoke and a common pathological feature of AD and certain related disorders. This condition is characterized by deposition of amyloid in the medial and adventitial layer of primarily small and medium-sized arteries and arterioles of the cerebral cortex and leptomeninges (1)(2)(3)(4)(5)(6)(7). Cerebrovascular A␤ deposition is intimately associated with smooth muscle cells in the vessel wall (2, 38 -40). It has been suggested that the subsequent degeneration of the smooth muscle cells is involved with loss of vessel wall integrity (38,39,41). CAA and intracerebral hemorrhage are the primary pathologies of a growing number of hereditary disorders that result from distinct point mutations within the A␤ domain of A␤PP. These include the A21G Flemish mutation (12), E22Q Dutch mutation (10,11), and E22K Italian mutation (14). A recent new addition to this list is the D23N Iowa mutation (18). The Iowa CAA disorder is characterized by progressive dementia with pathological findings of extensive cerebrovascular A␤ deposition (often severely narrowing or occluding the vessel lumena), small foci of hemorrhage and ischemic infarction, widespread neurofibrillary tangles, and relatively sparse senile plaques. It is unclear how the D23N Iowa mutation promotes cerebrovascular A␤ deposition. In the present study we investigated the effects of the D23N Iowa mutation on A␤PP processing and A␤ fibrillogenic and pathogenic properties toward HCSM cells.
Previous studies have reported that different A␤ CAA mutations can differentially affect amyloidogenic processing of A␤PP. To investigate the possible effects of the D23N Iowa CAA mutation on A␤PP amyloidogenic processing, we used a cell transfection paradigm to express different CAA-mutant forms of A␤PP. We found that the A21G Flemish mutation increased the production of A␤ peptide, whereas the E22Q Dutch mutation did not (Fig. 2), consistent with previous reports (27,28). Similar to the Dutch mutation, the D23N Iowa mutation did not increase A␤ production in transfected cells (Fig. 2). This suggests that the CAA observed in patients with the Iowa mutation is unlikely to result from overproduction of A␤ peptide.
The above results suggested that the D23N Iowa mutation instead might cause CAA by altering the properties of the A␤ peptide. In this regard, previous studies have demonstrated that E22Q Dutch mutant A␤ peptides exhibit enhanced rates of fibril assembly compared with wild-type A␤ peptides (19 -21). Since the D23N Iowa mutation results in the loss of a negative charge similar to the adjacent E22Q Dutch mutation, we speculated that Iowa mutant A␤ may also exhibit enhanced rates of fibrillogenesis. Indeed we found that D23N Iowa mutant A␤40 showed enhanced assembly that was nearly identical to E22Q Dutch mutant A␤40 (Fig. 3). Double-mutant Dutch/Iowa A␤40 also assembled rapidly, further implicating the importance of the loss of charge of Glu-22 and Asp-23 in augmenting the fibrillogenic properties of the peptide. Transmission electron microscopic analysis confirmed the assembly of Dutch, Iowa, and Dutch/Iowa A␤40 into fibrils (Fig. 4). These in vitro findings contrast with the in vivo evidence that the Dutch and Iowa mutations do not lead to dramatic A␤ deposition in plaques. It is possible that the localization of these altered forms of A␤ to cerebral vessels may reduce their levels in the brain parenchyma. It is noteworthy that A21G Flemish A␤40 exhibited a much slower assembly rate, comparable with wild-type A␤40 (Fig. 3). This finding suggests that the A21G Flemish mutation does not cause CAA by altering the fibrillogenic properties of A␤.
We previously demonstrated that E22Q Dutch mutant A␤40 robustly binds and assembles into fibrils on the surfaces of cultured HCSM cells, a cell type intimately associated with the pathology of CAA (24,30,36). Since D23N Iowa mutant A␤40 possesses potent fibrillogenic properties like E22Q Dutch mutant A␤40, it was not surprising that Iowa mutant and double Dutch/Iowa mutant A␤40 peptides also accumulated and assembled into fibrils on the HCSM cell surface (Fig. 5). Other studies of the A␤ peptide have suggested that the D23N substitution may alter the kinetics of ␤-sheet formation. 2 The finding that Flemish A␤40, like wild-type A␤40, does not readily assemble on HCSM cells further suggests that Flemish A␤ elicits a CAA phenotype by a distinct mechanism.
We have also previously shown that cell surface fibril assembly of pathogenic A␤ is required for inducing downstream pathologic responses in HCSM cells, including cell surface accumulation of secreted A␤PP, degradation of vascular smooth muscle cell ␣-actin, and ultimately an apoptotic cell death (24,30,36,37). The accumulation of secreted A␤PP is mediated by its high affinity binding to the A␤ fibrils that assemble on the HCSM cell surface and coincides with the induction of smooth muscle cell ␣-actin degradation and cell death (36). Following the same pattern for HCSM cell surface A␤ fibril assembly observed in Fig. 5, the Dutch, Iowa, and double Dutch/Iowa mutant A␤40 peptides potently induced each of these pathologic responses, whereas the wild-type and Flemish mutant A␤40 peptides did not. It is noteworthy that double Dutch/Iowa mutant A␤40 was noticeably more robust than either single mutant form of A␤40 in both the degradation of smooth muscle cell ␣-actin and loss in cell viability ( Figs. 7 and 8, respectively). These findings underscore the importance of cell surface A␤ fibril assembly as the key event in the induction of pathologic responses in HCSM cells. It should be emphasized that A␤ toxicity to HCSM cells may not be the only pathogenic effect of these peptides in familial CAA. Other mechanisms, such as effects on endothelial cell function or vessel narrowing, may also underlie the cause of dementia in these individuals.
Although CAA is commonly observed in patients with AD, the reason why specific mutations within A␤ manifest primarily as CAA remains unclear. The exception is the Flemish mutation, which presents as a mixed AD/strong CAA phenotype (12). On the other hand, patients with either the Dutch, Italian, or newly described Iowa mutations present severe cerebrovascular A␤ deposition with lesser extents of AD neuropathology. Brains from these patients demonstrate numerous diffuse, early stage A␤ deposits in the parenchyma but few of the mature, dense-cored senile plaques characteristic of AD (5,6,14,18,21). This observation suggests that this particular group of CAA mutations may alter A␤ in a manner that specifically enhances pathogenic interactions with the cerebral vasculature.