, kindly provided by Dr. M. L. Schmidt (The Center for Neurodegenerative Diseases, University of Pennsylvania School of Medicine, Philadelphia, PA), a polyclonal antibody raised against native Aβ purified from AD brain and made in our laboratory, Angela (1:250), 20 was used for double-immunofluorescent labeling of Aβ and AMY 117 on a 40-μm ethanol-fixed, cryoprotected AD brain section. Here, Aβ IR was visualized using a Texas-Red-conjugated donkey anti-rabbit secondary antibody (1:400; Jackson ImmunoResearch Laboratories, West Grove, PA) whereas AMY 117 IR was visualized using a fluorescein-isothiocyanate-conjugated donkey anti-mouse secondary antibody (1:100; Jackson ImmunoResearch Laboratories). A series of six confocal images were obtained through the section at 1-μm intervals using a Leica confocal laser scanning microscope, as previously described. 15
By optimizing the immunostaining protocols for the AMY 117 MAb and the Aβ antibodies on AD brain sections, we were able to document the clear co-occurrence of the AMY 117 and Aβ antigens within the vast majority of nondiffuse plaque lesions in AD cortex and hippocampus. The degree of AMY 117 IR within Aβ plaques varied from plaque to plaque, but importantly, the two antigens very frequently existed in the same lesion. AMY 117 IR was primarily associated with more rounded, compacted Aβ plaques, whereas its co-localization with large diffuse Aβ deposits was rare. The images in Figure 1 indicate that AMY 117 and Aβ can both co-mingle and specifically co-localize (eg, overlap) within an individual plaque. In a subset of more mature plaques, AMY 117 IR surrounds that of Aβ; in such cases, the two antigens may partially overlap or may segregate but abut each other. AMY 117 IR was not diminished by absorption of the antibody with Aβ peptide, suggesting that AMY 117 is a non-Aβ antigen. In general, the Aβ antigen was more abundant than the AMY 117 antigen in AD, aged human, and DS brains.
The discrepancy between our results and those reported earlier, 15 in which limited or sometimes no co-localization was described between AMY 117 and Aβ IR in AD brain sections, is probably due, in part, to differences in staining conditions. First, the pretreatments used in our study enhanced the Aβ IR, sometimes even in the ethanol-fixed sections, implying that some Aβ deposits may have been missed in the study by Schmidt et al. 15 Second, the Aβ42 MAb 21F12 and the Aβ polyclonal antibody R1282 are extremely sensitive at detecting multiple forms of Aβ deposits and, as such, allowed greatly increased detection of Aβ deposits relative to those detected by the polyclonal Aβ antibody 2332 used in the previous study. 15 Indeed, we have performed side-by-side comparisons of the three antibodies under the optimal staining conditions for each and have consistently found Aβ42 MAb 21F12 to detect the greatest number of Aβ deposits (C. A. Lemere and T. J. Grenfell, unpublished data). Our Aβ polyclonal antibody R1282 also detected more Aβ deposits than polyclonal antibody 2332 under optimal conditions but, in some cases, labeled fewer Aβ deposits relative to those stained by Aβ42 MAb 21F12. Third, in our study, we always single-labeled sections adjacent to the double-labeled sections so as to characterize the staining pattern for each antibody on its own to avoid the potential for steric competition between the antibodies for their respective antigens within lesions. Both in our hands, and recently in those of M. L. Schmidt (personal communication), a competition between Aβ and AMY 117 antibodies for their respective antigens in individual plaque lesions was observed and may explain why some AMY-positive plaques appeared to be Aβ negative in their study.
In full agreement with the earlier report, no AMY 117 IR was observed in Aβ-bearing blood vessels or in cerebellum, implying that there is a regional and cellular specificity for this protein to associate with Aβ. In our study, AMY 117 IR occurred only in those brain areas having Aβ immunoreactive plaques and never in regions devoid of Aβ. In contrast, low-magnification photomicrographs previously published by Schmidt and colleagues 15 (as shown in Figure 3 of their paper) illustrate an example of an AD case in which AMY 117 IR was detected in a region of parahippocampal gyrus devoid of Aβ using polyclonal antibody 2332. Upon request, the authors kindly provided us with several adjacent sections from the same block of tissue depicted in the figure. In our hands, the region previously described as being devoid of Aβ was Aβ42 plaque-rich using a different Aβ antibody (Aβ42 MAb 21F12) and pretreatment of the tissue with formic acid. Pretreatment of tissue and improved Aβ antibody sensitivity may account for the increased detection of Aβ deposits and, as such, the visualization of a much greater co-occurrence of Aβ with AMY 117 in the current study. Our conclusion is that regions of AMY IR are Aβ-rich. Careful inspection of Table 1 in the aforementioned paper confirms that in AD brain regions where both antigens were examined (positive staining listed as present (Y) or absent (N)), Aβ IR alone or Aβ and AMY 117 IR together were described, but not AMY 117 IR alone. 15 In the current study, small, punctate AMY 117 immunoreactive deposits that were not directly associated with Aβ IR were occasionally observed, but only in areas of abundant Aβ immunoreactive plaques. In general, these small, punctate AMY 117 deposits were located between and close to plaques and, as such, may represent the outer edge of a plaque that exists above or below the plane of the AMY-117-stained section, so that Aβ may be surrounded by the AMY 117 antigen (as we indeed observed in mature plaques (Figure 1, d–f , arrows)) or the two antigens partially overlap (exemplified by Figure 1, d–f , arrowheads).
By performing immunohistochemical studies of plaque development in a unique temporal series of 29 DS brains from patients between the ages of 12 and 73 years, we were able to determine that Aβ deposition clearly precedes that of the AMY 117 antigen. Early, diffuse Aβ42 plaques in young DS brains were AMY 117 negative; it was only after the appearance of more mature, compacted Aβ plaques that AMY 117 IR was observed. This point was best exemplified by the immunostaining of adjacent sections from a single block of brain tissue containing both temporal cortex and hippocampus from a 16-year-old DS patient (see Figure 3, c–f ). Diffuse Aβ42 IR was detected throughout the temporal cortex, whereas more compacted Aβ42 IR plaques were seen in the hippocampus in the same section. In the adjacent section, AMY 117 IR was observed only in the more compacted plaques in the hippocampus; the temporal cortex in the same section was entirely AMY 117 negative. We cannot exclude the possibility that the AMY 117 antigen within diffuse plaques (but not compacted plaques) was destroyed by the harsh (long-term) fixation conditions or by the double pretreatment to expose antigens in the young DS brains. However, the lack of AMY 117 IR in the long-term-fixed young DS brains bearing exclusively diffuse Aβ42 IR is consistent with the lack of AMY 117 IR in diffuse Aβ42 IR deposits observed in briefly fixed tissues from middle-age and older DS patients (see Figure 2, e and f ) and in AD cases. Furthermore, the staining protocol for each antibody under each of the fixation conditions was optimized before use. Therefore, we believe it is very unlikely that technical factors could explain the lack of AMY 117 IR in the aforementioned young DS brains.
The pattern of AMY 117 IR seen in the young DS brains was also observed in briefly fixed aged human control brains in which only three of eight Aβ-bearing brains showed any AMY 117 IR, and then only in a small portion of all Aβ plaques. Again, it was the Aβ plaques that appeared to be in the process of compaction that showed AMY 117 IR. Neither AMY 117 nor Aβ IR were observed in two aged human control brains and in six young DS brains, lending further support to the conclusion that the AMY 117 antigen is not detectable in brain lesions before the appearance of Aβ.
Brains from two animal models of AD pathogenesis, aged monkey (17 to 24 years) and PD-APP transgenic mice (aged 8 to 20 months), were examined for Aβ and AMY 117 IR. Aβ IR deposits were observed in all of the monkey and mouse brains, and the number of Aβ deposits increased strikingly with age. AMY 117 IR was not detected in either species, regardless of Aβ plaque burden. The lack of AMY 117 IR in monkey and PD-APP transgenic mouse brain suggests two possibilities. First, a species difference in the AMY 117 antigen may make it unrecognizable by the human MAb. Second, insufficient maturation of the plaques in monkeys and transgenic mice, compared with that in AD brain, may be responsible for the lack of AMY 117 detection in the plaques of these animals.
In summary, we conclude that the AMY 117 antigen is a non-Aβ amyloid-associated protein that accrues in AD plaques after Aβ deposition, rather than existing as the subunit of a novel, Aβ-negative lesion in Alzheimer’s disease. Because AMY 117 appears at the time of compaction of Aβ plaques, it may turn out to play a significant role in the evolution of the plaque. This new information is critical to the interpretation of AMY IR plaques in AD brain and to the further search for the AMY antigen.
We are indebted to Dr. Marie Luise Schmidt (The Center for Neurodegenerative Disease Research, University of Pennsylvania School of Medicine, Philadelphia, PA) for providing the confocal image shown in Figure 1 and to Drs. Schmidt and John Trojanowski for many interesting and helpful discussions. In addition, we are grateful to Drs. Krystyna Wisniewski (Institute for Basic Research in Developmental Disabilities, Staten Island, NY) and Douglas Anthony (Department of Pathology, Harvard Medical School, Boston, MA) for generously providing brain tissues from young DS patients. Lastly, we thank Dr. Dora Games and Karen Khan for kindly providing fixed PD-APP transgenic mouse brains.