Detection and clearance in Alzheimer’s disease: leading with illusive chemical, structural and morphological features of the targets

Highly specific interactions between biomolecules, such as antigen-antibody, protein-ligand, or nucleic acid base pair complementary are on the basis of the organization of complex organisms. The same principles may be tentatively used in molecular medicine for diagnosis and therapeutics. A molecule can be designed to selectively bind a protease and thereby inhibit the production of a peptide that forms toxic aggregates in the brain or an antibody may be produced to bind specifically to that peptide for detection or clearance purposes. Unfortunately, interference in biological systems is not that simple. For a start there is the inhibition of the physiological role of the protease; moreover, several cleavage fragments may be produced, which may continue to diverge due to putative post-translational modification and self-assembly processes, hiding the toxic target in a “soup” of peptide species varying in size, structure and chemical composition. A perspective of the current status and challenges in targeting peptide species for diagnosis and treatment in the context of Alzheimer’s disease is given.

Detection – fluid biomarkers: Early diagnosis followed by an accurate evaluation of the therapeutic intervention is critical in neurodegenerative diseases associated with amyloid formation. Sensing protein biomarkers is crucial in this context because often an aberrant form of one protein appears years or decades before the first symptoms (Selkoe and Hardy, 2016). Thus, the target is a protein species that, contrary to nucleic acids, is not easily amplified nor selectively recognized by base-pair complementarity.

Sensors are fabricated to combine a biorecognition motif with a transducer/amplification unit. Antibodies are usually used as biorecognition elements due to the highly specific and accurate nature of antigen-antibody interactions. Several aptasensors were also produced. Aptamers are short oligonucleotides that may selectivity bind to a protein but are prone to fast degradation in biological media, hampering its application in routine analysis.

Researchers found ingenious ways to maximize analytical performance. In general, sensors employ a sandwich-type architecture: two layers decorated by a dual antibody pair bind the target analyte, which becomes sandwiched. A signal, that depends on the components used, is generated in a concentration-dependent fashion. Typically, an enzyme-driven signal amplification step (as in the classical enzyme-linked immunosorbent assays) is applied. In the context of the Alzheimer’s disease biomarkers, other approaches aimed to reduce the overall complexity and length of currently used protocols or outperform the traditional sensors employ new components such as oligonucleotide-modified Au nanoparticles (Georganopoulou et al., 2005) or electrochemiluminescence sensors based on metal-organic frameworks (reviewed in Leite et al., 2023).

In Alzheimer’s disease, the predominant amyloid beta peptides is amyloid-β (Aβ) 1–40 (subscript numbers indicate amino acid length), about 10 times more abundant than the most reactive isoform Aβ_1–42. Monomeric Aβ_1–42 is an inverse biomarker because its concentration in the cerebrospinal fluid (CSF) drops, presumably due to self-aggregation, long before clinical manifestations (Bateman et al., 2012). Because Aβ basal production varies from individual to individual, the determination of Aβ_1–42/Aβ_1–40 ratio is more reliable than that of Aβ_1–42 alone. CSF samples provide 85–90% accuracy to identify prodromal Alzheimer’s disease in the mild cognitive impairment stage (Scheltens et al., 2016). Aβ_1–42 should be combined with the determination of the total tau and the phosphorylated tau for better diagnosis.

There is growing evidence that the soluble oligomeric Aβ species are the neurotoxic factors, while the soluble monomers and the mature fibrils are relatively harmful (Sengupta et al., 2016). The oligomers are not established as biomarker because its concentration in CSF seems to be 10³ times lower, with a picomolar (or fg/mL) range, in comparison with Aβ monomers (Mroczko et al., 2018). Moreover, those oligomers constitute a heterogeneous mixture of species which hinders accurate quantification.

Other promising biomarkers correlate with alterations in the synaptic function, such as a decrease in acetylcholine synthesis, a neurotransmitter associated with brain functions such as learning and memory and the CSF concentration of neurogranin, a dendritic protein involved in memory consolidation.

Standard methods are directed to the analysis of CSF samples which hampers large-scale population screening and triggers the search for biomarkers in body fluids requiring less invasive methods to collect (e.g., blood or saliva). However, current techniques do not detect significant variations of the above-mentioned protein species in blood samples of Alzheimer’s disease patients and healthy individuals (Scheltens et al., 2016). Screening for differences in unrelated blood biomolecule profiles failed to produce consistent results. Validation of biomarkers in saliva was also unsuccessful.

Treatment – anti-amyloid approaches: Alzheimer’s disease is characterized by Aβ senile plaques and tau neurofibrillary tangles. Researchers explored strategies to avoid the production of Aβ peptides by γ-secretase and β-secretase inhibitors, to inhibit the amyloid cascade by delivery of active compounds and to promote Aβ clearance by active and passive immunization. The recent results of the promising clinical trials using antibodies were largely negative, which boosted the interest in other approaches that will not be covered here, such as tau-target immunotherapy. Aβ, in particular Aβ_1–42 and potentially shorter Aβ peptides with post-translational modifications, are reactive species that self-aggregate, coexisting in a broad range of multimeric structures. The oligomers are presumably the toxic species but not only the nature of the “toxic Aβ oligomer” but also the role of that putative species remains elusive (Benilova et al., 2012). Aβ oligomers are obtained as mixtures of interconvertible species whose distributions are potentially affected by the analytical tools used by researchers for their isolation and characterization. As a result, the conclusions of the different research groups are difficult to conciliate. The simple question, whether its toxicity is related to unspecific interactions, probably driven by hydrophobicity, between Aβ oligomers and cellular membranes and proteins, or is whether there is a specific conformer that binds to a specific target remains unanswered. Both hypotheses support the higher reactivity of smaller Aβ species; they present simultaneously higher specific surface and size similarity to putative targets. Despite all the unknowns, evidence is accumulating that size is insufficient to define toxicity and conformation plays a role (Nguyen et al., 2021).

The residual concentrations of the oligomers associated with their presumably high toxicity made these species the most promising therapeutic targets. Passive immunotherapy using anti-Aβ monoclonal antibodies has been deeply attempted. These antibodies display different Aβ selectivity profiles. Some antibodies recognize N-terminal epitopes in Aβ (aducanumab, gantenerumab, BAN2401) and bind preferentially to aggregated forms (soluble or fibrillar), while solanezumab and ponezumab, that recognize the core and the C-terminal of Aβ are primarily selective for soluble monomers (Arndt et al., 2018). The results of the clinical trials were mostly disappointing. Nevertheless, it was reported that aducanumab, approved in a controversial decision for medical use in the United States by the Food and Drug Administration, shows a less pronounced energetic preference for a specific conformation of the N-terminal of Aβ, which favors its preferential binding towards aggregated species (Arndt et al., 2018). A possible reason for the generally negative results is that patients engaged of the clinical trials are on an advanced stage of the disease. Given that there is a balance between the several amyloid species which may be interconvertible, even if the antibody recognizes the toxic species, the overall accumulation of Aβ may trigger the fast re-establishment of the initial balance and of the levels of the toxic species. Moreover, passive immunization against Aβ may initiate inflammation in the brain, a pathological feature already associated with the disease. A more general concern is that monotherapy can be destined to fail in such a multifactorial disease and a combination of strategies, targeting Aβ, tau, and inflammation processes is needed.

Future perspectives: Diagnosis using the already described CSF biomarkers has a good predicted value. Future developments will probably focus on automation and improvement of reproducibility, combined with the definition of guidelines and cut-off values. Further research in sensing synaptic biomarkers has intrinsic value, as additional information about loss of synaptic function, that is not obtained from the currently used biomarkers, can be collected.

On the other hand, it is conceivable that protein species selected as CSF biomarkers may be used as blood biomarkers. This could not be confirmed up to now; if it is a matter of technical limitations or the lack of real statistically significant differences between the patients and the healthy controls is not clear yet.

Sensing is important not only in diagnostic but also in the evaluation of therapeutic strategies. In this context, it is crucial to find ways to detect Aβ oligomers, preferentially the toxic conformers, as they constitute the most promising therapeutic target.

Putting forward new directions for therapeutic strategies in the context of Alzheimer’s disease constitutes an embarrassing and somewhat pretensions exercise. Nevertheless, it seems most likely that Aβ oligomer toxicity results from unspecific hydrophobic interactions that a range of oligomeric species can establish with membranes and other proteins. Contact with the targets should trigger the exposition of Aβ hydrophobic surfaces that otherwise remain protected by the peptide N-terminals. If this is the case, the toxic oligomers remain invisible or indistinguishable from other oligomers in the solution. In fact, antibodies that recognize Aβ hydrophobic epitopes bind preferentially to monomers, not to aggregated forms. The search for an antibody that binds to a putative toxic conformer seems rather difficult to succeed.

Antibodies targeting a region of the peptide sequence barely distinguish the Aβ oligomeric and the fibrillar aggregated forms, because both display peptide N-terminals at their surface (Sengupta et al., 2016) (a scheme of the amyloid cascade is depicted in Figure 1). Probably conformational antibodies are better for targeting the oligomers specifically. These antibodies recognize an amyloid conformation but do not discriminate between protein species (Rasool et al., 2012), which would likely be an advantage in pathologies involving the aggregation of more than one protein, such as Alzheimer’s disease. Maybe assembling antibodies in pure protein nanoparticles with precise valency and symmetry would enhance their selectivity towards amyloid oligomers. Prediction computational tools have recently boosted the design of such multivalent nanoparticles (Divine et al., 2021).

Figure 1.

Scheme of the amyloid cascade associated with Alzheimer’s disease.

Highlighted in red are target Aβ species for detection and clearance. Monomeric Aβ_1–42 (on the left), together with total tau and the phosphorylated tau are the CSF biomarkers that provide the best predictive value. Immunosensors, typically based on ELISA methods, deliver good diagnostic performance. Toxic Aβ oligomers (in the middle) are the most promising targets for immunotherapy. However, antibodies that target a specific region of the peptide sequence hardly discriminate between all the aggregated forms (toxic and non-toxic oligomers and fibrils) because their surface is similarly decorated with the peptide N-terminals. Conformational antibodies are suitable for selectively targeting toxic oligomers. Created with Microsoft Office drawing tools. Aβ: Amyloid-beta; CSF: cerebral spinal fluid; ELISA: enzyme-linked immunosorbent assay.