The length of amyloid-β in hereditary cerebral hemorrhage with amyloidosis, Dutch type. Implications for the role of amyloid-β 1-42 in Alzheimer's disease

In hereditary cerebral hemorrhage with amyloidosis, Dutch type (HCHWA-D), a genetic variant (E22Q) of amyloid β (Aβ) accumulates predominantly in the small vessels of leptomeninges and cerebral cortex, leading to fatal strokes in the fifth or sixth decade of life. Aβ deposition in the neuropil occurs mainly in the form of preamyloid, Congo red negative deposits, while mature neuritic plaques and neurofibrillary tangles, hallmark lesions in Alzheimer's disease (AD), are characteristically absent. A recent hypothesis regarding the pathogenesis of AD states that Aβ extending to residues 42-43 (as opposed to shorter species) can seed amyloid formation and trigger the development of neuritic plaques followed by neuronal damage in AD. We characterized biochemically and immunohistochemically Aβ from three cases of HCHWA-D to determine its length in vascular and parenchymal deposits. Mass spectrometry of formic acid-soluble amyloid, purified by size-exclusion gel chromatography, showed that Aβ 1-40 and its carboxyl-terminal truncated derivatives were the predominant forms in leptomeningeal and cortical vessels. Aβ 1-42 was a minor component in these amyloid extracts. Immunohistochemistry with antibodies S40 and S42, specific for Aβ ending at Val-40 or Ala-42, respectively, were consistent with the biochemical data from vascular amyloid. In addition, parenchymal preamyloid lesions were specifically stained with S42 and were not labeled by S40, in agreement with the pattern reported for AD, Down's syndrome, and aged dogs. Our results suggest that in HCHWA-D the carboxyl-terminal Aβ heterogeneity is due to limited proteolysis in vivo. Moreover, they suggest that Aβ species ending at Ala-42 may not be critical for the seeding of amyloid formation and the development of AD-like neuritic changes.

Hereditary cerebral hemorrhage with amyloidosis, Dutch type (HCHWA-D), 1 is an autosomal dominant form of cerebral amyloidosis clinically defined by recurrent strokes and fatal cerebral bleeding in the fifth or sixth decade of life (1). Amyloid deposits are preferentially localized in the walls of small arteries of the leptomeninges and cerebral cortex. Parenchymal deposition is largely in the form of preamyloid lesions or diffuse plaque-like structures that are Congo red-negative and lack the dense amyloid cores commonly present in Alzheimer's disease (AD). Immature or early plaques are occasionally found surrounded by dystrophic neurites, although neurofibrillary tangles are consistently absent (2).
Biochemically, HCHWA-D amyloid fibrils are composed of a genetic variant of amyloid ␤ (A␤) of AD and Down's syndrome (3,4) in which there is a glutamine (Gln) for glutamic acid (Glu) at position 22 due to a missense mutation (cytosine for guanine) at codon 693 of the ␤ precursor protein (␤PP) 770 gene (5). This mutation is pathogenic since it segregates with the disease (6). Several in vitro studies have shown that Glu 3 Gln substitution yields an A␤ variant that has an abnormal conformation (higher content of ␤-sheet), forms amyloid-like fibrils at a faster rate than wild type A␤, and is toxic to cultured human leptomeningeal smooth muscle cells (7)(8)(9)(10)(11)(12), further supporting its pathogenic role in vivo. Our initial biochemical characterization of amyloid in HCHWA-D revealed that the main constituent of leptomeningeal fibrils was an A␤ species starting at Asp-1 and ending at Val-39 (13). Most of the studies on the composition of amyloid in AD indicate that the main constituent of cerebrovascular deposits is A␤ 1-40 (14,15), similar to the major soluble form of this peptide in biological fluids of AD and non-AD individuals (16,17). However, a more recent report, using a different method, indicates that A␤ 1-42 is also a major component of cerebrovascular amyloid deposits in AD (18). In vitro experiments with cell lines transfected with cDNA of ␤PPs containing mutations at codon 717 found in some kindreds with familial AD (FAD) show a higher proportion of secreted A␤ 1-42 in the conditioned medium when compared with cells transfected with wild type ␤PP (19). These data, together with the long known fact that senile plaques are mainly composed of A␤ 1-42 (14, 20 -21), have led to the speculation that the latter may be an essential factor in the pathogenesis of AD (22). The purpose of this study was to reexamine in more detail the biochemical composition of cerebrovascular amyloid in HCHWA-D and determine whether the longer A␤ species are important components of such deposits. In addition, we performed an immunohistochemical analysis of vascular and parenchymal amyloid deposits using antibodies specific for A␤ ending at Val-40 and Ala-42 to assess the relative distribution of these A␤ species.

MATERIALS AND METHODS
The brains from three HCHWA-D patients stored at Ϫ75°C were used for this study. Formalin-fixed, paraffin-embedded brain tissue was also available for immunohistochemical analysis. The clinicopathological, and DNA characterization of these patients has been previously reported (4,5). Normal leptomeninges were obtained from a 45-year-old male with no amyloid deposition upon neuropathological examination.
Isolation of A␤ from Leptomeningeal Vessels-Leptomeningeal vascular amyloid was isolated as described (18). Briefly, leptomeninges were carefully dissected from brain coronal sections, cut with scissors into 1-3-mm pieces, and placed in 0.1 M Tris-HCl, pH 8, on ice. Tissue was then washed by resuspension in 0.1 M Tris-HCl, pH 8, containing 1 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, 0.7 g/ml pepstatin A, 20 g/ml aprotinin (buffer A), and centrifuged at 800 ϫ g for 5 min at 4°C; the procedure was repeated five times. The material was collected by filtration through a 50-m nylon mesh (Spectrum), washed with 0.1 M Tris-HCl, pH 8, without inhibitors, and resuspended in 20 volumes of 2 mM CaCl 2 in 0.1 M Tris-HCl, pH 7.5. 0.3 mg/ml collagenase CLS-3 (Worthington) and 10 g/ml DNase I (Worthington) were added, and the mixture was incubated for 18 h at 37°C. After digestion, the suspension was filtered through a 350-m nylon mesh, and the filtrate was centrifuged at 10,000 ϫ g for 12 min. The pellet was resuspended in 100 volumes of 2% SDS (Bio-Rad) in 0.1 M Tris-HCl, pH 8, and incubated for 2 h at room temperature. The SDS-insoluble material was recovered by centrifugation as above and washed three times with distilled water. The pellet was dissolved in 10 volumes of 99% formic acid (Sigma) and incubated for 1 h at room temperature. After centrifugation at 10,000 ϫ g for 5 min, an aliquot of the supernatant was dried under N 2 and analyzed by SDS-polyacrylamide gel electrophoresis (PAGE), Western blot, amino-terminal sequence, and mass spectrometry. This fraction was named fraction L. The formic acid supernatant was then loaded on a Superose 12 column (10 ϫ 300 mm) (Pharmacia Biotech Inc.). Amyloid peptides were separated in 75% formic acid at 0.2 ml/min using a Bio-Cad Sprint chromatography system. Eluent was monitored at 280 nm, collected in 1-ml fractions, and analyzed as above. A second method of leptomeningeal amyloid extraction was done to avoid the steps of collagenase/DNase digestion at 37°C and incubations at room temperature, during which extensive proteolysis may occur. Leptomeninges were dissected as described above, rinsed in buffer A, placed in 20 volumes of 99% formic acid, sonicated for 10 s with a W-380 sonicator (Heat Systems-Ultrasonics, Inc.) on ice, and incubated for 15 min at 4°C. After centrifugation at 10,000 ϫ g for 5 min, the supernatant was dried under a steam of N 2 and subjected to mass spectrometry. The fraction obtained by this short procedure was named fraction SH.
Isolation of Cortical Vascular Amyloid-Cerebral cortex was dissected from leptomeninges and white matter and suspended in 30 volumes of 15% SDS in 0.1 M Tris-HCl, pH 8, containing protease inhibitors as described above. After stirring for 48 h at room temperature, the material was filtered through 350-, 150-, and 75-m nylon meshes. Vessels were washed five times with 20 volumes of 1% SDS in 0.1 M Tris-HCl, pH 8, filtered again through a 75-m mesh, washed with 2 mM CaCl 2 , 0.1 M Tris-HCl, pH 7.5, and digested with collagenase CLS-3 and DNase I in the same buffer. After 18 h at 37°C, the suspension was centrifuged at 10,000 ϫ g for 12 min, and the pellet was washed in 0.1% SDS in 0.1 M Tris, pH 8. Amyloid was solubilized in 99% formic acid and separated on a Superose 12 columm as described above.
Protein Sequence Analysis-Automated Edman degradation sequence analysis was carried out on a 477A protein-peptide sequencer, and the resulting phenylthiohydantoin amino acid derivatives were identified using the on-line 120A PTH analyzer (Applied Biosystems, Foster City, CA).
Mass Spectrometry-Protein and peptide samples were prepared for matrix-assisted laser desorption/ionization mass spectrometry using the dried droplet method (26). The matrix used was ␣-cyano-4-hydroxycinnamic acid (Sigma), which was purified by recrystallization. To produce the dried droplets, a saturated solution of matrix was prepared using either 2:1 aqueous 0.1% trifluoroacetic acid:acetonitrile or 1:2:3 formic acid:2-propanol:water (27) at room temperature. The sample was added to this solution so that the final sample concentration was 1-10 M. 0.5 l of the solution was placed on the mass spectrometer's probe and allowed to dry. The sample was then ready for analysis. The linear, time-of-flight mass spectrometer used was custom built at the Skirball Institute at the New York University Medical Center (28).
Synthetic Peptides-Synthetic peptide comprising the amino acid sequence 1-40 of A␤ was obtained from Sigma (Lot 34H04931). A␤ corresponding to the sequence 1-42 was synthesized in the Keck facility at Yale University by solid-phase N-tert-butyloxycarbonyl chemistry. Peptides were purified by high pressure liquid chromatography with a Vydac C18 column and a linear gradient of 0 -80% acetonitrile in 0.05% trifluoroacetic acid. The purity of the peptides was assessed by amino acid composition and mass spectrometry. Peptide concentration in the stock solutions was determined by amino acid composition.
Collagenase and DNase I Digestion of HCHWA-D Amyloid Fraction SH and Synthetic A␤ 1-40 -Fraction SH of HCHWA-D leptomeningeal amyloid was incubated with 100 volumes of the supernatant obtained after centrifugation at 10,000 ϫ g for 10 min of a normal leptomeninges homogenate in 0.1 M Tris-HCl, 2 mM CaCl, pH 7.5, to which collagenase CLS-3 and DNase I were added as described above. After 18 h at 37°C, the suspension was centrifuged at 10,000 ϫ g for 12 min, and the resultant pellet was resuspended in 99% formic acid and analyzed by mass spectrometry.
Synthetic A␤ 1-40 was dissolved in 0.1 M Tris-HCl, pH 7.5, 2 mM CaCl 2 , and digested with collagenase CLS-3 and DNase I at an enzyme: substrate dilution of 1:50 and 1:500, respectively. After 16 h at 37°C, the solution was analyzed by mass spectrometry.
Immunohistochemistry-Sections of temporal and frontal lobes, hippocampus, and cerebellum from three cases of HCHWA-D were deparaffinized in xylene and rehydrated in ethanol and TBS. After rehydration and treatment with 80% formic acid for 15 min, endogenous peroxidase was quenched by incubation with 0.3% hydrogen peroxide in methanol for 30 min. Sections were blocked with 10% fetal calf serum in TBS and incubated with S40 or S42 at a 1:500 dilution overnight at 4°C, biotinylated anti-rabbit (Sigma) at 1:800 for 1 h, and avidin-horse radish peroxidase (Amersham). The reaction was detected using 0.03% 3,3Ј-diaminobenzidine tetrachloride and 0.01% hydrogen peroxide, 50 mM Tris-HCl, pH 7.4. S40 and S42 were adsorbed by incubating them at a 1:500 dilution in 0.3% bovine serum albumin in TBS with 20 g of synthetic A␤ 1-40 and 1-42, respectively, for 3 h at room temperature or overnight at 4°C.
Congo Red Staining-Samples were placed on poly-lysine coated slides, air dried, and fixed in 100% ethanol for 5 min. After a 20-min incubation in a 80% ethanol, 0.1% NaOH, saturated NaCl solution, samples were stained for 2 h in 2% Congo red, destained in the above solution, and observed under polarized light.

SDS-PAGE and
Western blot with anti-A␤ 1-28 showed that formic acid-soluble amyloid (fraction L) extracted from HCHWA-D leptomeningeal vessels consisted largely of 4-kDa monomeric A␤ together with minor dimeric and polymeric species (Fig. 1, a and b). When the SDS-insoluble fraction was treated with formic acid, a small pellet was present after low speed centrifugation that was inaccessible to further analysis. Western blot of fraction L with S40, an antibody specific for A␤ ending at Val-40, showed intense immunolabeling of amyloid protein. Conversely, S42, specific for A␤ extending to Ala-42, revealed no staining with up to 5 g of amyloid loaded on the gel, although a strong positive band was obtained with 2 g of A␤ 1-42 synthetic analog (Fig. 1c). Direct amino-terminal sequence, DAEFRHDSGYEVHHQKLVFFAQDVGSNK, confirmed that A␤ started mainly at Asp-1. A minor sequence beginning at Ala-2 was also present. Fraction L was further analyzed by mass spectrometry. As shown in Fig. 2a, A␤ was composed mainly of 1-40, 1-38, and 1-37 with minor 1-36 and 1-39 species. The peak observed at 4015.3 (peak 1) can also correspond to the amino-terminal truncated species 2-38 (calculated mass, 4015.5). In one patient, a small signal corresponding to A␤ 1-42 was detected (Fig. 2a, peaks 14, 15, and  16). This isoform was not detected in the other two cases (not shown), indicating that with the method used, A␤ 1-42 was a minor component of the crude leptomeningeal amyloid preparation. To test whether the extensive carboxyl-terminal heterogeneity found in fraction L was due to proteolysis occurring during the isolation procedure, HCHWA-D amyloid was obtained from frozen leptomeningeal tissue that was thawed, rinsed in buffer A, and sonicated in 99% formic acid (fraction SH). Mass spectra of fraction SH showed a similar pattern of carboxyl-terminal truncated species as obtained from fraction L, indicating that they were not generated during digestion with collagenase/DNase I (Fig. 2b). Moreover, when fraction SH was incubated in the soluble fraction of normal leptomeninges containing collagenase and DNase I, the resultant pellet showed a similar spectrum with a conserved relative yield of the carboxyl-terminal truncated A␤ species (not shown). A similar experiment using synthetic A␤ 1-40 did not result in partial proteolysis of the peptide. Fraction L was then separated by size exclusion gel chromatography on a Superose 12 columm in 75% formic acid as described (18). A␤ eluted as two major peaks at approximately 13 and 16 ml with an additional minor peak at 18 ml, named peaks A, B, and C, respectively (Fig. 3). In the three patients examined, the gel filtration elution profile was identical (not shown). The length of the A␤ species in these peaks was determined by amino-terminal sequence and mass spectrometry. A similar pattern of carboxylterminal truncated A␤ species was obtained as with the crude fractions SH and L. The A␤ species in peak A were 1-40, 1-39, 1-38, 1-37, and 1-36, with some of these isoforms also starting at position Asp-2 (Fig. 4a). The lower signal of A␤ 1-40 obtained after Superose 12 separation as compared with the crude fractions may be due to a lower recovery of this more hydrophobic species since the concentration of formic acid was reduced from 99 to 75% for chromatography. Yet, we cannot rule out that some activity of lysosomal proteases at pH 1.4 may contribute to the lower yield of A␤ 1-40 after gel filtration. Peak B showed, in addition, a set of shorter derivatives comprising the sequences 1-25, 1-24, 1-23, 1-22, and 2-24 (Fig.  4b). Peak C displayed a very low signal with essentially the same pattern as peak B (not shown). In all cases, formylated A␤ species containing 1 or 2 formic acid molecules were found, possibly reflecting covalent modification of Ser-8 and Ser-26 (29,30).
Cortical vessels were isolated from the SDS-insoluble fraction of cerebral cortex, digested with collagenase, solubilized in formic acid, and separated on Superose 12. Prior to formic acid treatment, the presence of large amounts of amyloid-laden vessels was confirmed by Congo red staining in the material retrieved by the 150-m mesh. This fraction eluted as a single major peak at 15 ml after Superose 12 separation. Western blot of this material revealed monomers, dimers, and higher molec- FIG. 2. Mass spectrometry of amyloid extracted from HCHWA-D leptomeningeal vessels. a, amyloid (fraction L) was extracted by collagenase-DNase I digestion, SDS treatment and solubilized in formic acid as described (18). b, amyloid (fraction SH) from the same patient as in a was extracted by a short incubation of leptomeninges in formic acid. Samples were analyzed by matrix-assisted laser desorption/ionization mass spectrometry. Insets, calculated and observed masses of A␤ species with their predicted sequences. The presence of one or two formic acid molecules is indicated by * or **, respectively. Note that fraction SH is less formylated than fraction L due to a shorter treatment with formic acid. ular weight polymers of A␤ that were positive with anti-A␤ 1-28 (not shown). Mass spectrometry analysis showed that the major species were 1-40 and 1-38. A minor 1-42 component as well as 1-39, 1-37, and 1-36 were also present (Fig. 5). An oxidized form of all the A␤ variants extracted from cortical vessels was observed, indicating that an important percentage of A␤ in these deposits may be chemically modified at methionine 35. The yield of cortical vessels in the 150-m filtrate was very low, and although this fraction eluted at the same position as the previous one after formic acid Superose 12 gel filtration, it gave no signal on mass spectrometry precluding its characterization.
Immunohistochemical analysis showed that S40 labeled strongly the leptomeningeal and cortical vessels. In addition, plaques distributed in the vicinity of vessels located in the superficial cortical areas were stained by this antibody (Fig.  6a). S42 labeling was less intense compared with S40 in leptomeningeal and cortical vessels (Fig. 6b), consistent with the biochemical data, indicating that A␤ 1-42 is a minor component in these amyloid deposits. A strikingly different staining pattern was observed in the neuropil. S42 labeled numerous preamyloid lesions that were uniformly distributed throughout the cortex. The S42-positive plaques were not labeled by S40 and were less numerous in the region adjacent to the leptomeninges as opposed to the S40 immunoreactive material. In the cerebellum, A␤ deposition was seen mainly in leptomeningeal vessels and was immunostained strongly with S40. These vessels reacted weakly with S42, and parenchymal deposits were minimal (Fig. 6, e and f). Immunostaining with S40 and S42 was abolished after adsorption with synthetic A␤ 1-40 and 1-42, respectively. Cross-immunoadsorption (S42 pre-incubated with A␤ 1-40) did not diminish the specific labeling of the diffuse plaque deposits (not shown). DISCUSSION These results indicate that vascular A␤ of HCHWA-D is heterogeneous, composed mainly of the 1-40 isoform and its carboxyl-terminal truncated species as determined by aminoterminal sequence and mass spectrometry. Our previous report showing that the major vascular A␤ sequence ended at Val-39, can be re-interpreted as due to a poor yield of Val-40 in the last cycle of Edman degradation. A similar low yield of Val-40 has been reported for A␤ peptides isolated from supernatants of cells transfected with ␤PPs cDNAs (19). The finding of a small signal on mass spectrometry corresponding to A␤ 1-42 and the faintly positive reaction on immunohistochemistry of vessel walls with S42 indicate that the longer A␤ species is a minor component of vascular amyloid deposits in HCHWA-D. Since a small pellet was obtained after treatment with formic acid, it is possible that highly hydrophobic species such as A␤ 1-42 were underestimated. However, using a similar procedure, Roher et Crude amyloid (fraction L) was resuspended in 75% formic acid, centrifuged, and the supernatant was applied on a 10 ϫ 300-mm Superose 12 column at 0.2 ml/min. A␤ eluted mainly at 13 ml (peak A) and 16 ml (peak B). A␤ yield in peak C was very low. A) and b (peak B), as in the chromatographic profile depicted in Fig. 3., are shown. In b, only the spectrum corresponding to the shorter A␤ carboxyl-terminal truncated derivatives is shown. Insets, predicted and observed molecular masses of A␤ species. The presence of one formic acid molecule is indicated by *.

FIG. 5. Mass spectrometry of purified cortical vascular A␤.
Cortical vessels were isolated, and A␤ was purified as described under "Material and Methods." An aliquot of A␤ in formic acid after Superose-12 gel filtration was analyzed by mass spectrometry. Inset, calculated and observed molecular masses of A␤ species with their positions in the amino acid sequence. ox, oxidized.
al. (18) have shown that A␤ 1-42 from AD leptomeninges can be solubilized under these conditions. Our finding of shorter A␤ derivatives in the leptomeninges comprising the sequences 1-25, 1-24, 1-23, 1-22, and 2-24 strongly suggests that they are generated by carboxyl-terminal limited proteolysis. Whether degradation in HCHWA-D vessels occurs before or after the deposition of full-length A␤ remains to be addressed. In light of other amyloidoses in which intact as well as amino-and carboxyl-terminal truncated forms of soluble amyloid precursors are consistently present in the affected tissues, it seems likely that deposition precedes digestion (31). However, the finding of amino-and carboxyl-terminal heterogeneity in A␤ from non-AD cerebrospinal fluid (32) also suggests that it may be partially cleaved in the circulation. Regardless of the relative contribution of these two mechanisms to the presence of heterogeneity in amyloid deposits, the concept of a different  S42 (b, d, f). a and c, leptomeningeal and cortical vessels from HCHWA-D temporal lobe are strongly labeled by S40. Scattered plaques surrounding cortical vessels are also stained. Spots strongly stained with S40 (c) probably represent small vessels. b and d, adjacent sections as in a and c. S42 labels vascular deposits less intensely than S40. Preamyloid lesions are selectively labeled by S42. Note that S42-positive plaques are not labeled by S40 and vice versa. a and b were photographed at a magnification of 40ϫ and c and d at 100ϫ. e and f, cerebellum. Leptomeningeal vessels are strongly stained with S40 while S42 labeling is weak. Photographed at a magnification of 100ϫ. proteolytic processing of the A␤ carboxyl end in the vessel wall and the neuropil may explain the relative abundance of A␤ 1-42 in neuritic plaques in AD (14, 20 -21), Down's syndrome (33), and HCHWA-D preamyloid deposits, as suggested by our results in the present study. Peptide concentration, acidic pH, free radicals, metal ions (zinc and aluminum), and certain A␤ binding proteins, such as apolipoprotein E, serum amyloid P component, ␣1-antichymotrypsin, and proteoglycans, can modulate the aggregation of A␤ synthetic analogs in vitro (34), and, upon aggregation, A␤ is highly resistant to proteolysis (35,36). Relative differences in these and other factors between the neuropil and the vascular wall may account for an accelerated rate of aggregation of A␤ and inaccessibility to carboxyl-terminal degradation in the former compartment. Alternatively, A␤ 1-40 and A␤ 1-42 may be generated by different cellular processing pathways of ␤PP, expressed differentially in vessels and neuropil. The invariable presence of A␤ 1-42 as a major component in AD plaques, the early deposition of this species in Down's syndrome, and cell transfection studies with certain FAD-␤PP cDNAs (19) have led to the hypothesis that A␤ 1-42 cerebral accumulation (as opposed to shorter species) is a major determinant factor in the pathogenesis of AD (22). An in vitro explanation for this putatively crucial role of A␤ 1-42 in AD has been provided by the observation that A␤ 1-42 forms fibrils faster than A␤ 1-40 and that the former can seed amyloid formation by the more soluble shorter species in vitro (37).
The biochemical, genetic, and clinico-pathological features of HCHWA-D have led us to propose that it is a vascular variant of FAD amyloidosis (4). The subsequent description of a kindred in which affected members carrying the substitution ␤PP692Gly presented with either early-onset FAD or stroke is consistent with our interpretation (38). The findings that cerebrovascular amyloid in HCHWA-D is largely composed of A␤ 1-40 and shorter derivatives and that most of the preamyloid lesions are labeled by antibodies specific for A␤42 raise some interrogants about the significance of the carboxyl-end of A␤ in cerebral amyloidogenesis. HCHWA-D patients present numerous preamyloid deposits in the cerebral cortex but do not develop the full-blown pathological and clinical features of earlyonset AD (1,2,4). A premature death due to stroke is an insufficient explanation for the lack of AD features, since most HCHWA-D patients die in their fifties and sixties, an age at which most of the early-onset FAD patients are severely demented with typical AD pathology. Moreover, some HCHWA-D patients can reach the eighth decade without developing ADtype dementia or neurofibrillar lesions (2).
Although it remains to be confirmed biochemically, our present immunohistochemical data strongly suggest that A␤ species ending at Ala-42 are major components of HCHWA-D preamyloid deposits in the neuropil. A recent immunohistochemical study of HCHWA-D with monoclonal antibodies specific for A␤ ending at Val-40 or Ala-42 has shown a very similar pattern of A␤ species distribution as the one reported here with polyclonal S40 and S42 (39). Biochemical data from aged dogs having brain preamyloid lesions but not neuritic plaques indicate that the preamyloid deposits are composed mainly of A␤ 1-42 and A␤ 17-42 (40). In DS cerebellums it has also been found recently that preamyloid deposits contain mainly A␤ 17-42, leading to the suggestion that preamyloid can be defined biochemically as a lesion where a major peptide is A␤ 17-42 (41). Whether A␤ 17-42 is an important component of HCHWA-D preamyloid lesions in the neuropil is currently under investigation. Although more biochemical information regarding the relative amounts of A␤ 1-42, A␤ 1-40, and A␤ 17-42 from non-AD aged brains and HCHWA-D is needed, the available evidence suggests that the presence of A␤ with a carboxyl end extending to Ala-42 may not be critical for the seeding of amyloid formation in vivo and the subsequent development of neurofibrillar lesions and dementia, as previously proposed (42). This interpretation is reinforced by a recent report showing that A␤ 42/43 predominates in senile plaques of non-demented individuals (43). Yet, it is possible that higher concentrations of A␤ 42/43 due to a defective clearance in the AD neuropil as compared with non-AD A␤ amyloidoses may contribute to accelerate neuronal damage. The strong association of the apoE 4 allele with late onset AD (44) and the involvement of the genes on chromosomes 1 and 14 encoding presenilins, associated with most of the early-onset FAD cases (45,46), illustrate the complexity of these disorders that share a common pathway leading to the accumulation of A␤ in the brain. Within this group of "␤-amyloid diseases" (47), the different phenotype associated with HCHWA-D is related to the presence of the E22Q mutation as well as other factors associated with neuronal death and dementia besides the deposition of A␤ 42 peptides in the neuropil.