Shea D, Colasurdo E, Smith A, Paschall C, Jayadev S, Keene CD, Galasko D, Ko A, Li G, Peskind E, Daggett V. SOBA: Development and testing of a soluble oligomer binding assay for detection of amyloidogenic toxic oligomers. Proc Natl Acad Sci U S A. 2022 Dec 13;119(50):e2213157119. Epub 2022 Dec 9 PubMed.

Recommends

Please login to recommend the paper.

Comments

  1. This is an excellent study building on extensive molecular simulations and experiments over the last decade. Protein aggregates are highly heterogeneous in size and structure, and the researchers have developed a small peptide that binds to aggregates based on their structure rather than their protein composition. In this case, the peptide binds to α-sheet aggregates that form early in the aggregation of Aβ and which are shown to be cytotoxic.

    Using this peptide for capture, it is possible to develop a highly sensitive assay to detect low levels of these Aβ α-sheet aggregates in plasma samples. This assay is highly selective in distinguishing between control patients and patients with AD or early stages of AD.

    Since the α-sheet aggregates of Aβ are formed early in the aggregation process, the results in this study suggest that in all stages of AD there is a roughly three- to fourfold higher rate of Aβ aggregation taking place than in controls, assuming changes in the blood reflect the changes taking place in the brain. This seems plausible, since α-sheet aggregates are so small they can presumably easily enter the blood.

    This finding is in good agreement with a previous study by Colin Masters and co-workers that showed that in normal brain, 2 mg of Aβ is deposited over 19 years, while in an AD brain 6 mg is steadily deposited (Roberts et al., 2017). Taken together, it seems there is a raised level of production of Aβ aggregates and their deposition as plaques.

    Unlike an aggregation reaction in the test tube, a wide range of aggregates of different sizes and structures, including plaques, are present in people. As such, the mechanisms of toxicity will depend on the relative concentration of these different species and their potencies. The α-sheet aggregates in this work were shown to be small, cytotoxic, and disruptive to neuronal signaling.

    This is in agreement with our previous work showing that early small Aβ aggregates were effective at permeabilizing the cell membrane (De et al., 2019). These α-sheet aggregates will contribute to the toxicity observed in AD, and this work suggests the level of this toxicity is raised in AD compared to controls.

    However, aggregates with different structures are also toxic and will be present. For example, we found that β-sheet-containing Aβ protofilaments, formed at later stages, were more effective at causing inflammation, sensitively recognized by the innate immune response, and that this can cause deficits in LTP, a cellular correlate of memory loss (Hughes et al., 2020). 

    In general, the relative amount of different aggregates in the brain will depend on their rate of formation and removal. Larger aggregates, for example, may be harder to remove and hence accumulate, so the dominant toxic species in humans is likely to be different from that of aggregates formed during an aggregation reaction, and they may also change with disease progression [Acta Neuropathologica Communications volume 7, Article number: 120 (2019)].

    Given that it is now possible to sensitively detect aggregates of different structures in human samples, in particular α-sheet as well as β-sheet containing aggregates, it would be very interesting to track how both change in plasma and CSF over time to determine when they form and increase in number. Then we can start to work out their relative contributions to cellular toxicity and symptoms of AD. Such studies may help identify, in more detail, the species that need to be targeted by therapies, as well as those that might allow for early diagnosis. It will also be fascinating to see how the current antibody-based treatments for AD alter the amount of these a sheet aggregates.

    The capability to sensitively detect α-sheet protein aggregates, as described in this work, opens up many new exciting possibilities to study protein aggregates formed in AD at new levels of detail. 

    References:

    Roberts BR, Lind M, Wagen AZ, Rembach A, Frugier T, Li QX, Ryan TM, McLean CA, Doecke JD, Rowe CC, Villemagne VL, Masters CL. Biochemically-defined pools of amyloid-β in sporadic Alzheimer's disease: correlation with amyloid PET. Brain. 2017 May 1;140(5):1486-1498. PubMed.

    De S, Wirthensohn DC, Flagmeier P, Hughes C, Aprile FA, Ruggeri FS, Whiten DR, Emin D, Xia Z, Varela JA, Sormanni P, Kundel F, Knowles TP, Dobson CM, Bryant C, Vendruscolo M, Klenerman D. Different soluble aggregates of Aβ42 can give rise to cellular toxicity through different mechanisms. Nat Commun. 2019 Apr 4;10(1):1541. PubMed.

    Hughes C, Choi ML, Yi JH, Kim SC, Drews A, George-Hyslop PS, Bryant C, Gandhi S, Cho K, Klenerman D. Beta amyloid aggregates induce sensitised TLR4 signalling causing long-term potentiation deficit and rat neuronal cell death. Commun Biol. 2020 Feb 18;3(1):79. PubMed.

    View all comments by David Klenerman

  2. SOBA, as described by Shea et al., is a technically sophisticated and, based on this proof-of-concept study, almost perfectly discriminatory plasma Aβ oligomer assay for control vs. MCI and AD individuals. In previous work, the authors theorized in molecular dynamic simulations, and then observed under certain experimental conditions, that synthetic Aβ42 transiently adopts an unconventional “α-sheet fold” before converting to the conventional β-sheet fold of amyloid fibrils. All of the cited articles reporting features of the α-sheet peptides come from the senior authors’ lab, raising the question of independent confirmation of this unusual conformer and its absolute requirement for multiple types of amyloid diseases, as the authors state.

    The authors further state that the α-sheet conformer is the “defining feature of toxic soluble oligomers in several amyloids, including toxic Aβ42 oligomers.” It is unclear from this paper how the designed α-sheet peptide is solely specific to toxic oligomers and not to non-toxic oligomers; indeed, ref. 38 speaks of a “functional amyloid formation” occurring in a Streptococcus, presumably not necessarily toxic but rather functional per se.

    The authors designed a stable, soluble, non-toxic, α-sheet, main-chain peptide that can complement and bind (capture) toxic peptides having similar, main-chain α-sheet conformations, independent of their amino acid (i.e., side chain) sequence. The authors use this α-sheet synthetic peptide to capture material from human plasma and CSF (the soluble oligomer binding assay, or SOBA), and then detect that portion of the captured proteins which comprise “toxic Aβ oligomers” using the conventional 6E10 antibody to the N-terminus of Aβ monomers as the sole detector. 6E10 also reacts with APP and its sAPP-α fragment. The concept of using a structurally complementary, designed synthetic peptide as an affinity reagent is unique and clever, and this appears to markedly decrease matrix interference in the complex milieu of human plasma, giving remarkably sensitive, low-femtomolar limits of oligomer quantification (Fig S2E).

    In Figs. 1E and F, a single CSF from an AD patient vs. one from a control are probed with the α-sheet synthetic peptide AP193 to capture oligomers followed by their detection with 6E10 (before and after SOBA). The authors report that this experiment reveals toxic oligomers of Aβ42 that are apparent trimers, hexamers, and dodecamers. It is surprising that all Aβ42 oligomers in the AD CSF fall into these three sizes as isolated by size-exclusion chromatography (SEC), and none are seen in the control CSF. Using such SEC columns, aqueously diffusible Aβ oligomers from AD brain extracts are found overwhelmingly in the void volume (i.e., very high MW), not as such low-n oligomers (Stern et al., 2022).

    For measuring oligomer neurotoxicity, Shea et al. use a single electrophysiological assay based on decreased spike number in cultured neuroblastoma cells, with low spike numbers rescued by the designed α-sheet conformer (AP193) used for capture in the SOBA assay. It is also unusual to deduce results by testing just one CSF sample from AD vs. control, particularly as many (379) plasma samples were quantified in this paper.

    There are additional questions about the assay which remain unanswered, at least within this paper. The SOBA assay appears to deplete substantial amounts of the total CSF UV-absorbing small proteins (Fig. 1F), presumably only a small portion of which is actually oligomeric Aβ. In plasma, even less of the total protein is Aβ. This suggests that much of the signal captured in the SOBA assay may derive from non-Aβ sources, and specificity for Aβ is solely conferred by the 6E10 detection, which requires further validation using multiple anti-Aβ antibodies recognizing both the N-termini and C-termini of Aβ. Even ignoring this concern about assay specificity for Aβ in plasma, the assay’s ability to discriminate perfectly between AD, MCI, and control is surprising (an AUC of 0.99).

    This was mostly a clinically defined cohort (out of 310 subjects, 62 had autopsy confirmation and 177 had known CSF Aβ42 values; Table S1). Since clinical diagnosis, even when partially combined with CSF Aβ42 values, as done here, is known to be an imperfect predictor of AD neuropathology, it is surprising that the discrimination by SOBA could be perfect when the AD diagnosis may be incorrect for at least a few and probably numerous cases. SOBA was perfect at predicting clinical diagnosis even among a sub-cohort deliberately selected to contain 50 percent of cases in which the clinical diagnosis disagreed with the CSF Aβ42 cutoff (Fig S8, sample set #2). Presumably this cohort would include even more incorrect AD diagnoses as defined by neuropathology.

    Moreover, the authors did not observe conversion of any subject from SOBA-negative to SOBA-positive in their small number of age-matched longitudinal samples. Presumably this must occur at some point if the assay detects an evolving disease state. These remaining questions should hopefully be addressed in other cohorts in future studies of this promising new tool. Accurately quantifying hydrophobic plasma Aβ42 oligomers across heterogeneous human subjects is a challenging task which the authors should be commended for addressing, and this test should also be compared to the very few plasma immunoassays reported for this complex and potentially pathogenic analyte.

    References:

    Stern AM, Yang Y, 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. BioRxiv, October 18, 2022 bioRxiv

    View all comments by Dennis Selkoe

  3. The α-sheet secondary structure in oligomers formed by 42 residues long Aβ and α-synuclein, which was shown to correlate with toxicity, is the star of the research reported by Valerie Daggett's group. The observation of the α-sheet, first predicted by Linus Pauling and Robert Corey in 1951, is intriguing because this atypical secondary structure is rarely observed in natural proteins and an extended α-sheet has not been experimentally identified in any known natural protein.

    Shea and collaborators provide evidence that Aβ42 oligomers incubated for ~24 hours adopt the α-sheet structure, based on the observed "null" CD spectrum, which can be explained by alternating chirality of the amino acid residues forming the α-sheet. Additional data from solution infrared spectroscopy support the presence of α-sheet structure in Aβ42 oligomers that form at ~24 hours of incubation. Moreover, the amount of α-sheet structure in Aβ42 oligomers was demonstrated to positively correlate with toxicity.

    The group developed a series of de novo designed α-sheet peptides with alternating L- and D-amino acid residues, which form α-sheet hairpins. One of these peptides, AP193, is used in this most recent study as a capture agent for the soluble oligomer binding assay (SOBA). Capturing oligomers by SOBA relies on their α-sheet structure, which is not present at early Aβ42 assembly stages (at zero hours of incubation), or at later assembly stages, when Aβ42 fibrils appear. As such, SOBA is sequence-independent and can be used for capturing oligomers of various amyloidogenic proteins. When paired with sequence-specific antibody, for example, A11 for Aβ, SOBA-AD exhibits spectacular specificity and sensitivity as a biomarker for an early detection of AD or MCI or even preclinical AD in CSF.

    Structural and mechanistic details of α-sheet formation in Aβ42 oligomers remain fuzzy, however. What drives formation of the α-sheet, which is less energetically favorable than the β-sheet, in Aβ42 oligomers but not monomers? Which peptide regions in Aβ42 adopt the α-sheet structure? The α-sheet-rich oligomers seem to be on the pathway to β-sheet formation, which requires large conformational changes when cross-β-fibrils start to emerge. What drives such structural conversion? If formation of the toxic α-sheet is sequence-independent, how can a single amino acid mutation associated with a familial form of AD exert the documented significant effects on the onset of AD pathology? How does α-sheet structure in Aβ42 oligomers (hexamers and dodecamers) exert toxicity? Clearly, more research is needed to address these questions.

    View all comments by Brigita Urbanc

  4. The tsunami of elegant amyloid and tau diagnostic chemistry flowing in from myriad experts should not befog the possibility that the root cause of AD may yet lie elsewhere. Curiously missing from these erudite reports is any mention of a role for an infectious agent, microbe, virus, prion, and the like. Two diagnostic precedents come to mind: the recent demonstration that molecular mimicry of a gut microbial peptide is likely responsible for diabetes Type 1 (Girdhar et al., 2022), and the long slog for the actual root cause of peptic ulcer that led to an unappreciated microbe (Yeomans, 2011). 

    References:

    Girdhar K, Huang Q, Chow IT, Vatanen T, Brady C, Raisingani A, Autissier P, Atkinson MA, Kwok WW, Kahn CR, Altindis E. A gut microbial peptide and molecular mimicry in the pathogenesis of type 1 diabetes. Proc Natl Acad Sci U S A. 2022 Aug 2;119(31):e2120028119. Epub 2022 Jul 25 PubMed.

    Yeomans ND. The ulcer sleuths: The search for the cause of peptic ulcers. J Gastroenterol Hepatol. 2011 Jan;26 Suppl 1:35-41. PubMed.

    View all comments by Leslie Norins

Make a Comment

To make a comment you must login or register.

This paper appears in the following:

News

  1. Soluble Oligomer Assay Rivals Other Biomarkers in Detecting Preclinical AD