Under the Radar: APP Uppsala Fibrils Evade Microglia, PiB, Leqembi
Quick Links
Five years ago, scientists described a deletion of six amino acids within the APP gene, reporting that three members of a Swedish family, who each carried one copy of this “Uppsala deletion,” developed Alzheimer’s disease in their early 40s. The mutation nixes residues 19-24 from the middle of the Aβ sequence, making APP more prone to β-cleavage and spewing peptides that aggregate faster than full-length Aβ (Apr 2019 conference news; Aug 2021 news).
- Researchers created mice carrying APP with the Uppsala deletion.
- These accumulate Aβ fibrils that barely evoke a response from glia.
- Fibrils also evaded PiB and lecanemab, yet bound bapineuzumab.
Now, researchers led by Dag Sehlin and Martin Ingelsson at Uppsala University have generated mice expressing the APP Uppsala gene. In the February 5 Acta Neuropathologica Communications, they reported that these mice make Aβ fibrils that barely evoke a response from microglia and astrocytes, are hard to detect with PiB PET, and bind poorly to lecanemab yet tightly to bapineuzumab. “[These mice] recapitulate several pathological features of the Uppsala APP mutation carriers,” the authors wrote.
First author María Pagnon de la Vega engineered mice to express human APP with not only the Uppsala deletion but also the Swedish mutation. More readily cleaved by β-secretase, the latter has been widely used in mice to accelerate Aβ pathology. Sehlin said they bred mice with only the Uppsala deletion, but all these rodents died for unknown reasons.
By 6 months old, the mice, dubbed tg-UppSwe, had plaques made almost entirely of AβUpp42, i.e., amino acids 1-42 sans the six deleted. Immunostaining showed small, diffuse plaques of Aβ accumulating first in the frontal cortex at 4 months, then in the hippocampus and cerebral cortex at 8 and 13 months, and finally appearing in the thalamus by 18 months (image below). Mice at any age had almost no soluble or plaque Aβ40. This mirrors the Aβ42-rich, Aβ40-poor plaques seen at autopsy of one family member with the Uppsala deletion.
Uppsala Spreading. Tg-UppSwe mice accumulated scant Aβ40 aggregates with age (left column). Plaques of Aβ42 begin in the frontal cortex, spread through the cerebral cortex, then invade the thalamus (right column). [Courtesy of Pagnon de la Vega et al., Acta Neuropathologica Communications, 2024.]
As for other pathologies, array tomography of brain tissue from 18-month-old tg-UppSwe mice detected fewer synapses surrounding Aβ fibrils. Curiously, these aggregates had barely mustered responses from astrocytes or microglia, as judged by GFAP and Iba1 immunostaining of brain tissue, respectively, and by ELISA of GFAP and soluble TREM2 in brain extracts. The few activated astrocytes and microglia the mice did marshal seemed unable to detect fibrils since they barely co-stained with plaques in brain slices. Sehlin chalked this up to the mice having very few soluble Aβ oligomers, which are what attract glia to compact Aβ into dense-cored plaques. These soluble aggregates can be found in high-speed supernatant from mouse brain extracts, but Pagnon de la Vega found very little in tg-UppSwe mice (Nov 2022 news). Indeed, Andrew Stern of Brigham and Women’s Hospital in Boston agreed that a paucity of soluble Aβ forms might explain the poor response from microglia. “Glial access to diffusible Aβ aggregates of the right size could determine inflammation, synapse loss, and tau aggregation,” he wrote (comment below). In addition to scant gliosis, tg-UppSwe mice, like most amyloid models, had no tau pathology.
Stealth Plaques? Amyloid plaques in 18-month-old UppSwe (left) and ArcSwe (right) mice bind bapineuzumab (top), but only the ArcSwe plaques bind lecanemab (bottom right). [Courtesy of Pagnon de la Vega et al., Acta Neuropathologica Communications, 2024.]
Unlike other amyloid models, including age-matched mice expressing APP with the Arctic and Swedish mutations (ArcSwe), 18-month-old tg-UppSwe mice showed no amyloid PET signal on in vivo scans using PiB, which binds dense-cored plaques. Nor did tg-UppSwe amyloid bind radiolabeled mAb158, the mouse equivalent of lecanemab, which specifically binds Aβ protofibrils (image above). These results echo the weak PiB PET signal seen in Uppsala deletion carriers and the almost undetectable soluble Aβ protofibril level in the Uppsala brain on autopsy. In contrast, the UppSwe mouse brain lit up with labeled mAb3D6, the forerunner of bapineuzumab, which, like lecanemab, binds Aβ’s N-terminus (image above). “The results … demonstrate that mutations far removed from this region can result in stable differences in structure that are reflected in the ability of antibodies to bind to this region,” concluded Charles Glabe, University of California, Irvine (comment below).
“The Uppsala amyloid deposits are detectable in vivo, but their conformation is antibody-specific,” Sehlin told Alzforum. To Glabe, the implications are that there may be no “pan” Aβ antibodies. “This may limit the effectiveness of individual monoclonal antibodies,” he wrote.
“This principle is important for drug development, especially in the age of antibody therapeutics,” Sehlin noted. He hesitated to speculate about what these results might mean for the two living carriers of APPUpp, since they also carry one normal copy of APP.
Curiously, Sehlin and colleagues found that aggregates of Aβ in soluble fractions from tg-UPPSWe mice bound lecanemab better than did plaques or aggregates released by detergent, leading them to suggest that the peptide changes conformation when it becomes insoluble. “I think the evidence is so far insufficient to make this conclusion,” wrote Stern. “I hope that high-resolution structures of aggregates from all these fractions of both tg-UppSwe tissue, and human Uppsala mutation tissue, can test this hypothesis: Is it possible to identify different polymorphs in the different fractions, or are they mostly the same structure, just different in length, compaction, and/or binding partners?” he wrote. Sehlin has partnered with Gunnar Schröder, Forschungszentrum Jülich, Germany, to resolve the structure of AβUpp fibrils from the tg-UppSwe mice and the autopsied APPUpp carrier by cryoEM.—Chelsea Weidman Burke
References
Mutation Interactive Images Citations
Mutations Citations
News Citations
- APP Upp: Mutation Nixes Six Amino Acids from Aβ, Spurs Aggregation
- Double Whammy: APP Uppsala Deletion Ups Aβ and Its Aggregation Propensity
- Short Aβ Fibrils Easily Isolated from Alzheimer's Brain Fluid
Therapeutics Citations
Further Reading
No Available Further Reading
Primary Papers
- Pagnon de la Vega M, Syvänen S, Giedraitis V, Hooley M, Konstantinidis E, Meier SR, Rokka J, Eriksson J, Aguilar X, Spires-Jones TL, Lannfelt L, Nilsson LN, Erlandsson A, Hultqvist G, Ingelsson M, Sehlin D. Altered amyloid-β structure markedly reduces gliosis in the brain of mice harboring the Uppsala APP deletion. Acta Neuropathol Commun. 2024 Feb 5;12(1):22. PubMed.
Annotate
To make an annotation you must Login or Register.
Comments
Brigham and Women's Hospital, Harvard Medical School
This was an interesting paper from Uppsala, following up their previous report of the Uppsala mutation (Pagnon de la Vega et al., 2021), now modeling it in transgenic mice. The most striking observation from the tg-UppSwe mice, compared to tg-ArcSwe or tg-Swe, is the absence of an innate immune response around amyloid plaques. This is different from the original patients with the Uppsala mutation, who had pronounced gliosis. Plaques in the tg-UppSwe mice were thioflavin S-negative and PiB-PET-negative, whereas people with the Uppsala mutation demonstrated thioflavin S positivity and weak PiB-PET positivity. I wonder if these differences mean that the conformation adopted by the AβΔ19-24 peptide in the tg-UppSwe mice was different from that which it adopts in people with the Uppsala mutation.
Either way, the apparent inertness of tg-UppSwe plaques compared to tg-ArcSwe or tg-Swe could be a model to understand how Aβ causes inflammation. Aberrant astrocytic and microglial activity is well known to correlate with synapse loss and dementia in people with Alzheimer’s disease. In turn, the amount of diffusible Aβ in aqueously extracted ultracentrifugal supernatants of AD brain is well known to correlate much better with dementia than total plaque burden and/or amount of Aβ extracted with formic acid (McLean et al., 1999; McDonald et al., Brain, 2010; Esparza et al., 2013). This correlates with the results in Pagnon de la Vega et al. in which the tg-UppSwe mouse had five- to 10-fold less Aβ in the 100,000 g aqueous supernatant, but not other fractions, compared to tg-ArcSwe and tg-Swe (Fig. 4C). Thus, I agree with the authors’ conjecture that the low concentration of Aβ in the 100,000 g supernatant is an explanation for the absent immune response. I agree with the authors that glial access to diffusible Aβ aggregates of the right size could determine inflammation, synapse loss, and tau aggregation.
The authors also conjecture that the Aβ in the 100,000 g aqueous supernatant undergoes a conformation change when it transitions into other fractions, but I think the evidence is so far insufficient to make this conclusion. They base this on the observation that mAb158, the murine precursor of lecanemab, better (~two-fold) detected this fraction and worse (~two-fold) detected the fraction released by Triton X-100 by ELISA, and that mAb158 did not stain plaques. Direct measure of mAb158 ELISA detection of the formic acid-released fraction was not done. We have shown that lecanemab binds to fibrils in the aqueous ultracentrifugal supernatants of sporadic human AD brain with identical structure to those fibrils found in detergent-released supernatants (Stern et al., 2023). I don’t think invoking a conformational change between the 100,000 g aqueous supernatant and the Triton X-100-released or formic acid-released supernatant is yet necessary to explain the differences in inflammatory reaction; the abundance of Aβ is sufficient. I hope that high-resolution structures of aggregates from all these fractions of both tg-UppSwe tissue and human Uppsala mutation tissue can test this hypothesis: is it possible to identify different polymorphs in the different fractions, or are they mostly the same structure, just different in length, compaction, and/or binding partners?
In summary, this is a fascinating paper. I hope that further study of the Uppsala mutation can help identify the structural and biochemical determinants of Aβ aggregate-induced inflammation and toxicity.
References:
Pagnon de la Vega M, Giedraitis V, Michno W, Kilander L, Güner G, Zielinski M, Löwenmark M, Brundin R, Danfors T, Söderberg L, Alafuzoff I, Nilsson LN, Erlandsson A, Willbold D, Müller SA, Schröder GF, Hanrieder J, Lichtenthaler SF, Lannfelt L, Sehlin D, Ingelsson M. The Uppsala APP deletion causes early onset autosomal dominant Alzheimer's disease by altering APP processing and increasing amyloid β fibril formation. Sci Transl Med. 2021 Aug 11;13(606) PubMed. Correction.
McLean CA, Cherny RA, Fraser FW, Fuller SJ, Smith MJ, Beyreuther K, Bush AI, Masters CL. Soluble pool of Abeta amyloid as a determinant of severity of neurodegeneration in Alzheimer's disease. Ann Neurol. 1999 Dec;46(6):860-6. PubMed.
Esparza TJ, Zhao H, Cirrito JR, Cairns NJ, Bateman RJ, Holtzman DM, Brody DL. Amyloid-β oligomerization in Alzheimer dementia versus high-pathology controls. Ann Neurol. 2013 Jan;73(1):104-19. PubMed.
Stern AM, Yang Y, Jin S, Yamashita K, Meunier AL, Liu W, Cai Y, Ericsson M, Liu L, Goedert M, Scheres SH, Selkoe DJ. Abundant Aβ fibrils in ultracentrifugal supernatants of aqueous extracts from Alzheimer's disease brains. Neuron. 2023 Jul 5;111(13):2012-2020.e4. Epub 2023 May 10 PubMed.
University of California, Irvine
In this paper, the authors from the laboratory of Dag Sehlin, and coworkers, provide an interesting update on the properties of the Uppsala APP mutation in transgenic mice (Pagnon de la Vega, et al., 2024). A previous report described that the Uppsala mutation deletes amino acid residues 19-24 (FFAED) of the Aβ sequence and gives rise to an aggressive form of early onset familial AD, and that it results in enhanced fibrillization of the mutant Aβ (Pagnon de la Vega et al., 2021). The results described in this paper show that structural differences in the Aβ fibrils resulting from the Uppsala mutation give rise to amyloid deposits that do not bind to the “gold standard” PET radioligand [11C]PiB for amyloid imaging, nor to the canonical amyloid-specific fluorescent dye thioflavin S, nor to the “protofibril” specific monoclonal mAb158 that is the murine parent of lecanemab, recently approved by the FDA for treatment of AD.
Once again, these results prove that when it comes to amyloid, you only see what you look for. Amyloids, like those formed from Aβ, tau, or α-synuclein, have the vexing property of being able to adopt multiple different β-sheet structures from the same amino acid sequence, allowing them to hide from imaging agents, fluorescent dyes, and monoclonal antibodies that bind, or not, depending on the amyloid structure. Like berserkir, the mythical shape-shifting creatures of Scandinavian folklore that violate the principles of nature, polymorphic amyloid structures are an egregious violation of Christian Anfinsen’s principle stating that “all information required to specify the structure of the protein is encoded in its amino acid sequence,” or, in shorter terms: one sequence, one structure.
Although the Uppsala mutation deletes residues 19-24, its effects on structure are also observed in the amino terminus of Aβ, where the epitopes for lecanemab and 3D6 reside. The Aβ(Uppsala) deposits stain well with 3D6, while they stain poorly with lecanemab, even though these antibodies bind to the same sequence (residues 1-5, DAEF), indicating that they interact with the same residues in a manner that varies among fibril structures. This structure is maintained even after antigen retrieval with heat and citric acid, pH 6.3, and then 70 percent formic acid, consistent with the stability and resistance to denaturation of other amyloid fibrils. This conforma-tion specific binding is typical of monoclonal antibodies that bind to the amino terminal residues 1-7 (Reyes-Ruiz et al., 2020) of Aβ and show differential binding to amyloid from human brain (Hatami et al., 2014) and other fibril structures (Walti et al., 2016).
Lecanemab is specific for amyloid “protofibrils,” and it binds 10-fold more strongly to protofibrils than fibrils (Soderberg et al., 2022). This has been interpreted as protofibrils having a unique structure that is critical for pathogenesis yet different from that of fibrils. However, recent reports on tau protofibrils revealed that they are a polymorphic collection of 23 different partially folded and alternatively assembled fibrillar intermediates on the pathway to the formation of the energetically more stable mature fibril structure (Lovestam et al., 2024), while a study of islet amyloid polypeptide intermediates identified seven different structures during the course of fibril formation (Wilkinson et al., 2023). Each displays a different molecular surface topology that would present different epitopes for antibody recognition, rather than representing a single unique structure that is distinct from the final fibril structure.
It remains to be established whether Aβ protofibrils are similarly polymorphic, but if so, it raises questions about what lecanemab and similar antibodies actually bind. Do they recognize one, or a subset, of intermediate structures? Are the structures they recognize the ones that are toxic and important for amyloid pathology in AD? Are there other important Aβ structures associated with AD pathology that lecanemab and [11C]PiB fail to recognize? Has antibody therapy so far appeared to have had a limited effectiveness because of partial recognition of a subset of polymorphic pathogenic amyloid structures?
So far, three of the 27 published structures of Aβ fibrils listed in the atlas of amyloids (UCLA Amyloid Atlas 2023) have been found in human brain (Kollmer et al., 2019; Yang et al., 2022). However, with the exception of the fibrils isolated from cerebrovascular amyloid (Kollmer et al., 2019), the structure of the “fuzzy coat” residues 1-8 is disordered in other structures and therefore unknown and this amino terminal region is where most of the monoclonal antibodies bind. The results from this paper demonstrate that mutations far removed from this region can result in stable differences in structure that are reflected in the ability of antibodies to bind to this region, consistent with previous reports of large-scale variation in the ability of monoclonal antibodies that have epitopes in this region to bind to human AD amyloid deposits (Hatami et al., 2014). The implications of this are that conformational polymorphisms may mean that there are not any “pan” Aβ antibodies that react with all structural variants. This may limit the effectiveness of individual monoclonal antibodies, and we have only just begun to evaluate the therapeutic effectiveness of monoclonal antibodies that bind selectively to polymorphic Aβ structures. A polyclonal immune response or an individualized and curated antibody mixture that recognizes many different structures may be more effective therapeutic agents than single monoclonal antibodies.
References:
Pagnon de la Vega M, Syvänen S, Giedraitis V, Hooley M, Konstantinidis E, Meier SR, Rokka J, Eriksson J, Aguilar X, Spires-Jones TL, Lannfelt L, Nilsson LN, Erlandsson A, Hultqvist G, Ingelsson M, Sehlin D. Altered amyloid-β structure markedly reduces gliosis in the brain of mice harboring the Uppsala APP deletion. Acta Neuropathol Commun. 2024 Feb 5;12(1):22. PubMed.
Pagnon de la Vega M, Giedraitis V, Michno W, Kilander L, Güner G, Zielinski M, Löwenmark M, Brundin R, Danfors T, Söderberg L, Alafuzoff I, Nilsson LN, Erlandsson A, Willbold D, Müller SA, Schröder GF, Hanrieder J, Lichtenthaler SF, Lannfelt L, Sehlin D, Ingelsson M. The Uppsala APP deletion causes early onset autosomal dominant Alzheimer's disease by altering APP processing and increasing amyloid β fibril formation. Sci Transl Med. 2021 Aug 11;13(606) PubMed. Correction.
Reyes-Ruiz JM, Nakajima R, Baghallab I, Goldschmidt L, Sosna J, Ho PN, Kumosani T, Felgner PL, Glabe CG. An "epitomic" analysis of the specificity of conformation-dependent, anti-Aß amyloid monoclonal antibodies. J Biol Chem. 2020 Dec 9; PubMed.
Hatami A, Albay R 3rd, Monjazeb S, Milton S, Glabe C. Monoclonal antibodies against Aβ42 fibrils distinguish multiple aggregation state polymorphisms in vitro and in Alzheimer disease brain. J Biol Chem. 2014 Nov 14;289(46):32131-43. Epub 2014 Oct 3 PubMed.
Wälti MA, Ravotti F, Arai H, Glabe CG, Wall JS, Böckmann A, Güntert P, Meier BH, Riek R. Atomic-resolution structure of a disease-relevant Aβ(1-42) amyloid fibril. Proc Natl Acad Sci U S A. 2016 Aug 23;113(34):E4976-84. Epub 2016 Jul 28 PubMed.
Söderberg L, Johannesson M, Nygren P, Laudon H, Eriksson F, Osswald G, Möller C, Lannfelt L. Lecanemab, Aducanumab, and Gantenerumab - Binding Profiles to Different Forms of Amyloid-Beta Might Explain Efficacy and Side Effects in Clinical Trials for Alzheimer's Disease. Neurotherapeutics. 2022 Oct 17; PubMed.
Lövestam S, Li D, Wagstaff JL, Kotecha A, Kimanius D, McLaughlin SH, Murzin AG, Freund SM, Goedert M, Scheres SH. Disease-specific tau filaments assemble via polymorphic intermediates. Nature. 2024 Jan;625(7993):119-125. Epub 2023 Nov 29 PubMed.
Wilkinson M, Xu Y, Thacker D, Taylor AI, Fisher DG, Gallardo RU, Radford SE, Ranson NA. Structural evolution of fibril polymorphs during amyloid assembly. Cell. 2023 Dec 21;186(26):5798-5811.e26. PubMed.
Kollmer M, Close W, Funk L, Rasmussen J, Bsoul A, Schierhorn A, Schmidt M, Sigurdson CJ, Jucker M, Fändrich M. Cryo-EM structure and polymorphism of Aβ amyloid fibrils purified from Alzheimer's brain tissue. Nat Commun. 2019 Oct 29;10(1):4760. PubMed.
Yang Y, Arseni D, Zhang W, Huang M, Lövestam S, Schweighauser M, Kotecha A, Murzin AG, Peak-Chew SY, Macdonald J, Lavenir I, Garringer HJ, Gelpi E, Newell KL, Kovacs GG, Vidal R, Ghetti B, Ryskeldi-Falcon B, Scheres SH, Goedert M. Cryo-EM structures of amyloid-β 42 filaments from human brains. Science. 2022 Jan 14;375(6577):167-172. Epub 2022 Jan 13 PubMed.
Make a Comment
To make a comment you must login or register.