Even more sceptical

My head is spinning with more questions than answers

I have some more questions about the experimental design in terms of methodology and correlation. What exactly is a double-filtration apheresis supposed to prove to me if my focus was exclusively on β-amyloid-like clots? I will subgroup this article into two rasing questions.

1. What Is the ThT-Signal Telling, and What Can It Not Tell Without Any Further Validation?

Let me first, once more, start with the Thioflavin T (ThT)-signaling issue. In this regard, I would like to quote Carrol et al. (2025)¹:

“Fluoroprobes have also been used as competitors to identify non-fluorescent compounds by displacement. Although fluoroprobes have played critical roles in studying tauopathies, they typically lack specificity for different fibril polymorphs. Indeed, their generality is often a great strength because a single fluoroprobe such as ThT has the versatility to detect a wide range of fibrils, largely independent of sequence or substructures. Yet the field would also benefit from complementary fluoroprobes that are selective for subsets of tau polymorphs.”

“Largely independent of sequence” implies a fundamental lack of proteomic specificity. In biochemistry, the primary structure of a protein, its specific amino acid sequence, is the definitive determinant of its identity, distinguishing unique entities such as Spike protein, fibrinogen, or albumin. Because ThT binding is effectively independent of this primary sequence, the probe is biochemically “blind” to the identity of the target molecule. It does not recognize specific proteins; instead, it reacts exclusively to a common physical motif: the cross-beta-sheet quaternary structure. Since numerous proteins are capable of undergoing conformational transitions into beta-sheet-rich fibrillar structures under physiological or pathological stress², ³, ⁴, ⁵, ⁶: such as acute systemic inflammation, localized protein misfolding, or shear stress induced by apheresis instrumentation–the resulting fluorescence signal in plasma represents a non-specific aggregate of various amyloidogenic and non-amyloidogenic species, rather than a quantifiable marker of a single distinct pathology.

We’ll examine the reaction of ThT a bit closer: Noormägi et al.⁷ wrote in 2012:

”(…) Recently, it has been demonstrated that several low-molecular weight compounds like Basic Blue 41, Basic Blue 12, Azure C, and Tannic acid interfere with the fluorescence of ThT bound to Alzheimers’ amyloid-b fibrils and cause false positive results during the screening of fibrillization inhibitors. In the current study, we demonstrated that the same selected substances also decrease the fluorescence signal of ThT bound to insulin fibrils already at submicromolar or micromolar concentrations. (…)”

"(…) It has been demonstrated that ThT binds to the grooves of the cross-beta structure of fibrils [4,7,9]. The binding stoichiometry of bound ThT inducing the characteristic fluorescence is about 0.09 moles of ThT per mole of insulin in fibril form [16]. Our results show that the apparent fluorescence quenching starts at very low submicromolar concentration of selected compounds and is completed at concentrations equal to 0.1 stoichiometry, which is in agreement with the binding stoichiometry for ThT determined earlier [16]. Compounds that interfere with ThT are similar to the structure of ThT (see Table 1), and, therefore, it is realistic that they compete with ThT for common binding sites on amyloid fibrils. Basic Blue 41 and Basic Blue 12 expose high, nanomolar affinity to the insulin fibrils, which is higher than affinity of ThT. This fact could be further exploited for design of novel amyloid probes or potential starting points for discovery of substances inhibiting protein fibrillization process, which are applicable as drug candidates for amyloidogenic diseases. We have to mention that Methylen Blue, which has similar structure to Basic Blue 41 and Basic Blue 12 is currently in clinical trials as disease modifying drug for Alzheimers disease [17,18]. (…)”

The authors demonstrated in 2012 that ThT and ThT-like structures possess a competitive nature, which can lead to significant interference, resulting in both false-positive and false-negative signaling. Furthermore, it has been demonstrated–at the latest since Biancalana et al. (2010)⁸ –that ThT binds non-specifically to the general hydrophobic β-sheet structures of fibrils:

“Subsequent molecular dynamics analysis of the PSAM in the presence of ThT enabled the first application of dye-binding simulations to a discrete and well-defined ThT binding site [50]. The hybrid experimental and computational approach employed to study the PSAMs has thus provided an important link between dye-binding simulations and experimentally testable hypotheses. Molecular dynamics simulations were conducted in a similar manner as for the eight-stranded KLVFFAE protofibrils described above [44, 45]. In order to maximize computational efficiency, the N- and C-terminal globular domains of the PSAMs were removed, as these have been experimentally demonstrated not to bind ThT. The initial system was composed of a periodic water box containing ~5600 water molecules, the “excised” PSAM flat β-sheet, two ThT molecules, and two neutralizing chloride ions. Simulations were conducted at 310 K over 100 ns, and the resulting populations of bound ThT molecules were clustered into predominant binding modes. These simulations demonstrated that ThT binds primarily to the designed site along the shallow groove formed by cross-strand Tyr side chains, making subtle contacts with the adjacent Leu ladder (Figure 4f-4h, Site a). The striking similarity between the Val-Phe [44] and Tyr-Leu [50] ThT-binding motifs observed in simulations offers support for the validity and potency of such aromatic-mediated ThT binding modes. These results strongly suggest that surface-exposed channels rich in aromatic amino acids are the primary binding sites of ThT. Together, the findings rationalize the broad binding spectrum of ThT toward diverse fibrils.”

In summary it can be said that the reliance on ThT fluorescence to quantify “amyloid clot burden” in patient plasma is flawed without an orthogonal confirmation due to three primary biochemical constraints:

1. Lack of Proteomic Specificity (The “Blindness” Problem) ThT is not a targeted probe; it is a structural dye that interacts exclusively with the quaternary architecture of cross-beta-sheet motifs. Because it binds largely independently of the primary amino acid sequence, it cannot distinguish between different proteins. Whether the signal originates from pathogenic amyloidogenic species or benign, functional proteins is indistinguishable to the probe. In a complex matrix like blood plasma—which contains high concentrations of diverse proteins—a ThT signal merely indicates the presence of some structure with beta-sheet characteristics, not the presence of a specific pathological clot.

2. Susceptibility to Chemical Interference (The Competition Problem) Research by Noormägi et al. (2012) demonstrates that ThT is prone to competitive inhibition. Numerous small-molecule compounds share structural similarities with ThT and compete for the same binding sites on amyloid fibrils. In a clinical or experimental setting, this leads to significant signal distortion:

• False Negatives: Competitive displacement of ThT by other substances leads to “fluorescence quenching,” which can be erroneously interpreted as the clearance of amyloid fibrils.

• False Positives: Chemical interference can alter the binding kinetics, leading to inaccurate quantifications that do not reflect the actual state of fibrillization.

3. Structural Ambiguity and Lack of Polymorph Selectivity As underscored by Carrol et al. (2025), ThT lacks the specificity to discriminate between different fibril polymorphs. Its “versatility” in detecting a wide range of fibrils is, in this context, a diagnostic liability. Furthermore, the work of Biancalana et al. (2010)³ reinforces that ThT binding is driven by broad hydrophobic interactions, specifically aromatic amino acid motifs, rather than highly specific molecular recognition.

The use of ThT-based assays to validate the efficacy of procedures like apheresis is methodologically not enough. When biomass is removed from plasma, a reduction in the non-specific ThT signal is an inevitable physical outcome, regardless of whether a disease-causing agent was targeted or if the plasma environment was simply chemically altered. Without independent, rigorous validation, such as mass spectrometry or another orthogonal structure specific fluorescence like Congo red⁹, to identify the specific protein composition of the material, ThT fluorescence remains an ambiguous indicator that conflates structural detection with specific clinical pathology.

This was further confirmed by Biancalana et al. (2009)¹⁰ who detailed the exact structural requirements for ThT binding, reinforcing its reliance on hydrophobic pockets rather than sequence-specific recognition. While Girych et al. (2016)⁴ discusses how combining probes can attempt to mitigate the shortcomings of using ThT alone. Jamali et al. (2022)¹¹ demonstrated that Z-scan optical method can complement the Thioflavin T assay for investigation of anti-Alzheimer’s impact of polyphenols. This illustrates the necessity of using complementary optical methods, as ThT-only assays are insufficient for characterizing fibrillization inhibitors. Hirata et al. (2024)¹² highlight ThT’s utility as a stress-response probe, further emphasizing its non-specificity regarding the type of amyloid structure or underlying disease mechanism. Kalitnik et al. (2025)¹³ provide a framework for the complex interplay between different amyloid species, debunking the “single-clot” pathology narrative. However, the authors summarize:

“In contrast to ThT and other dyes specific to amyloid fibrils in general, antibodies paired with fluorescence probes can selectively recognize distinct amyloid proteins. Traditional sequence-specific epitopes target defined linear regions, whereas structural epitopes additionally recognize conformational features associated with particular stages of fibrillation (Perchiacca et al., 2012). Due to their regional specificity, antibodies can elucidate the mechanisms of cross-interactions as in the case of AαSyn and lysozyme (Vaneyck et al., 2021). Additionally, conformationally sensitive antibodies map transitions through assembly states, expanding the understanding of the fibrillation kinetics as shown in the case of interactions between Aβ and TDP-43 (Shih et al., 2020).”

And this finally leads me to an answer, shocking me a little bit:

”specificity of ThT signal in plasma matrices

reproducibility across samples

correlation with symptom burden

modulation with intervention”

Is it obvious or do I need to spell it out? The loss of a signal (or a reduction in biomass) following a massive intervention like a double filtrated apharesis (including Immusorba TR) does not imply that the removed material was the causative agent of the disease. This is a classic methodological trap: equating the clearance of an aggregate with the resolution of a pathology. Because the ThT signal is essentially a readout of non-specific structural motifs in a fluctuating plasma environment, any significant physical filtration process will inevitably lower the fluorescence intensity. This reduction is a physical artifact of biomass removal, not a validated clinical indicator of “amyloid clearance.”

By relying on such an imprecise proxy, the entire premise of “clot clearance” as a therapeutic endpoint collapses. Without establishing a direct, sequence-specific correlation between the removed proteins and the patient’s symptomatic burden, these protocols remain exploratory at best, and medically reckless at worst. We are seeing a confusion of “cleaning the fluid” with “curing the biology”.

So much for my first question, which Maria Gutschi recently expanded upon from a pharmacological perspective with several extremely important aspects.

2. Are we dancing on an active volcano?

The study by Sheik et al.¹⁴ from 2017 investigated a new approach to neutralizing SEVI (Semen-derived Enhancer of Virus Infection) amyloid fibrils, which can massively increase HIV infectivity. This demonstrates that amyloid fibrils occur naturally and can facilitate certain viral infections. The authors used hydrophobic nanoparticles to drastically reduce this effect.

Here are the key findings of the work:

The authors demonstrate that polymers with hydrophobic side chains are capable of interacting with SEVI and significantly disrupting its β-sheet structure. The β-sheet content was reduced by approximately 45% through the use of these hydrophobic materials compared to SEVI in the presence of hydrophilic control polymers.

“Herein, we show that polymers containing hydrophobic side chains can interact with SEVI and reduce its β-sheet content by ~45% compared to the β-sheet content of SEVI in the presence of polymers with hydrophilic side chains, as estimated by Polarization Modulation-Infrared Reflectance Absorption Spectroscopy Measurements.”

This structural change led to a measurable reduction in SEVI-mediated HIV infection in TZM-bl cells. A nanoparticle formulation of the hydrophobic polymer reduced SEVI-mediated HIV infection by 60% compared to control treatment.

“A nanoparticle formulation of this hydrophobic polymer reduced SEVI-mediated HIV infection in TMZ-bl cells by 60% compared to control treatment.”

A crucial point is that these nanoparticles did not require specific amyloid-targeting groups. Although this required high concentrations to observe biological activity, this approach highlights the potential of using hydrophobic interactions to specifically modify the secondary structure of amyloids.

“While these nanoparticles lacked specific amyloid-targeting groups, thus, requiring high concentrations to observe biological activity, the use of hydrophobic interactions to alter the secondary structure of amyloids represents a useful approach to neutralizing SEVI function.”

Interestingly, the authors used hydrophobic polymer nanoparticles with a nearly neutral charge and covalent bonding, as they emphasized in their study, to demonstrate that the hydrophobic properties of nanoparticles alone are sufficient to actively interact with β-sheet structures.

This raises a huge question: How on earth would actually used LNPs act with fibril-structures? However, as demonstrated by the authors, β-amyloid sheet structures occur naturally and are not necessarily an indicator of amyloid-driven pathogenicity. As allways in systems biology terms: Context, space, place and threshholds matter.

And here is the real dance on the volcano.

In their paper, Hammarström and Nyström (2023)¹⁵ assumed the dangerous interplay between viral infections, amyloid formation, and neurodegeneration. The authors claim that many viral proteins are inherently amyloidogenic and that direct cooperation or “cross-seeding” can occur between human and viral proteins. A key example in their work is SARS-CoV-2: The abnormal blood clotting observed in severe COVID-19 cases or Long COVID was directly associated to the amyloid formation of human fibrin as well as the viral spike protein. The study concludes that virus and amyloid research must collaborate more closely to develop drugs against long-term neurological and systemic damage caused by such viral infections.

What is more, the central reference that Hammarström & Nyström rely on in their 2023 review for the spike-amyloid claim is their own preceding paper from 2022 (JACS). And what exactly did they do there? In the 2022 JACS paper, amyloidogenic sequences in the spike protein were identified using ThT kinetics, Congo red staining, and TEM ultrastructure. What initially sounds valid here turns out to be a methodological extrapolation: Because they did this on isolated synthetic 20-amino acid peptides, incubated at 37°C in vitro.¹⁶

This represents a methodologically enormous leap: they demonstrate that isolated peptide fragments can be amyloidogenic under artificial conditions in vitro, and then extrapolate in the 2023 review that this constitutes a mechanism for Long COVID coagulation pathology.

While the former paper emphasizes the pathogenic dangers of naturally occurring amyloids, the study by Dodd-o et al. (2024)¹⁷ demonstrates how this very architectural principle can be specifically utilized to combat viruses. The researchers developed a tunable antiviral platform made of synthetic peptides that self-assemble into functionalized β-sheet fibrils. These specific peptides possess two distinct domains: one forces the molecule to self-assemble into a stable fibril structure, while the other domain specifically binds to viral proteins (such as the SARS-CoV-2 spike protein). This fibril structure creates a “multivalent” binding surface that neutralizes viral complexes extremely efficiently and, due to its structure, exhibits particularly high resistance to viral mutations.

To understand the contradiction, one must examine the methodologies in detail:

• Hammarström & Nyström(2023): Their work serves as a comprehensive review and a conceptual hypothesis formulation (the "Vicious Liaison"). They frequently rely on literature data derived primarily from ThT assays, high dosage in vitro experiments with no physiological comparable parameters (e.g.: Christ et al.¹⁸; 150µl of 4 mM ThT¹⁹), and, for another example, case reports n=1. While such data are useful for identifying correlations between viral load and amyloid presence, they do not constitute mechanistic proof of the structural integrity of the amyloids, nor does the study provide direct evidence of neurodegenerative clinical signs.

• Sheik et al. (2017): In contrast to the weak evidence given by Hammarström & Nystöm, they used Polarization Modulation-Infrared Reflectance Absorption Spectroscopy (PM-IRRAS) to measure β-sheet content directly. This is a much more robust line of evidence than a simple fluorescence signal, as it analyzes the molecular vibration of peptide bonds (Amide I band), thereby allowing for a direct assessment of secondary structure.²⁰

• Dodd et al. (2024): The researchers use “tunable compositions,” a modular approach. They separate the physical property of fibril formation (self-assembly via a self-assembly domain) from the biological function of viral neutralization (via a specific targeting domain). This is a decisive difference: they leave nothing to chance (or a vague ThT signal) but instead design the amyloid structure intentionally. Unlike smaller molecules that might only block a single site, this platform exploits the multivalent nature of the fibril. The fibril serves as a “scaffold” presenting numerous antiviral units. This ensures enormous binding density and robustness. To support their claims, the group does not rely solely on ThT. They characterize their artificial fibrils using electron microscopy (TEM/AFM) and CD-spectroscopy (circular dichroism) to provide genuine proof of the structural integrity of their β-sheet constructs.

Summary: The Methodological Dubiousness of ThT-Based Apheresis

The current use of Thioflavin T (ThT) fluorescence to quantify “amyloid burden” in apheresis therapy is methodologically insufficient and clinically misleading. Because ThT reacts solely to quaternary structure (cross-beta-sheets) and is blind to primary sequences, it cannot differentiate between pathological protein aggregates and functional, systemically relevant protein structures.

A significant reduction in the ThT signal following double-filtration apheresis is, therefore, not evidence of selective “amyloid clearance,” but merely a physical artifact of biomass reduction. Without the use of orthogonal, sequence-specific analytics (such as mass spectrometry or structure-specific antibodies), this therapeutic approach fails to address the underlying causal pathology. Furthermore, the non-specific removal of amyloids ignores systems biological complexity: amyloid structures are not inherently toxic, but rather operate in a dynamic equilibrium between physiological function (e.g., antiviral defense) and pathological aggregation. Even more, post-viral syndrom is no new mechanism²¹, ²², ²³ and up today not fully understood²⁴, ²⁵, ²⁶, ²⁷, ²⁸. Furthermore, it has not been conclusively clarified whether amyloid proteins exclusively promote pathogenic states.²⁹, ³⁰

Conclusion

The current trend of “cleaning the fluid” without validating the proteomic identity of the target structures represents a medically risky oversimplification that substitutes scientific evidence with a non-specific fluorescence marker.

1. Carroll, E.C., Yang, H., Powell, W.C. et al. High-throughput discovery of fluoroprobes that recognize amyloid fibril polymorphs. Nat. Chem. 17,1565–1575. doi:10.1038/s41557-025-01889-7

2. Uversky VN. Amyloidogenesis of natively unfolded proteins. Curr Alzheimer Res. 2008;5(3):260-87.doi:10.2174/156720508784533312. PMID: 18537543; PMCID: PMC2719783.

3. Marin-Argany M, Candel AM, Murciano-Calles J, Martinez JC, Villegas S. The interconversion between a flexible β-sheet and a fibril β-arrangement constitutes the main conformational event during misfolding of PSD95-PDZ3 domain. Biophys J. 2012;103(4):738-47. doi: 10.1016/j.bpj.2012.07.029. PMID: 22947935; PMCID: PMC3443772.

4. Mani, K., Fransson, L.-Å., Ruscher, K., & Cheng, F. . Non-toxic amyloid beta formed in the presence of glypican-1 or its deaminatively generated heparan sulfate degradation products. Glycobiology.2013; 23(12):1510–1519. doi:10.1093/glycob/cwt079

5. Deepak Kumar Vijaya Kumar et al. Amyloid-β peptide protects against microbial infection in mouse and worm models of Alzheimer’s disease. Sci. Transl. Med. 2016; 8, 340ra72-340ra7. doi:10.1126/scitranslmed.aaf1059

6. Shewmaker F, McGlinchey RP, Wickner RB. Structural insights into functional and pathological amyloid. J Biol Chem. 2011; 286(19):16533-40. doi: 10.1074/jbc.R111.227108. Epub 2011 Mar 25. PMID: 21454545; PMCID: PMC3089495.

7. Noormägi A, Primar K, Tõugu V, Palumaa P. Interference of low-molecular substances with the thioflavin-T fluorescence assay of amyloid fibrils. J Pept Sci.18(1):59–64. doi:10.1002/psc.1416.

8. Biancalana M, Koide S. Molecular mechanism of Thioflavin-T binding to amyloid fibrils. Biochim Biophys Acta. 2010;1804(7):1405-12. doi: 10.1016/j.bbapap.2010.04.001. Epub 2010 Apr 22. PMID: 20399286; PMCID: PMC2880406

9. Girych M, Gorbenko G, Maliyov I, et al. Combined thioflavin T-Congo red fluorescence assay for amyloid fibril detection. Methods Appl Fluoresc. 2016;4(3):034010. Published 2016 Sep 6. doi:10.1088/2050-6120/4/3/034010

10. Biancalana M, Makabe K, Koide A, Koide S. Molecular mechanism of thioflavin-T binding to the surface of beta-rich peptide self-assemblies. J Mol Biol. 2009;385(4):1052-1063. doi:10.1016/j.jmb.2008.11.006

11. Jamali M, Mohajer S, Sheikhlary S, Ara MHM. Z-scan optical method complements the Thioflavin T assay for investigation of anti-Alzheimer’s impact of polyphenols. Photodiagnosis Photodyn Ther. 2022;39:102914. doi:10.1016/j.pdpdt.2022.102914

12. Hirata Y, Takemori H, Furuta K, Kamatari YO, Sawada M. Ferroptosis induces nucleolar stress as revealed by live-cell imaging using thioflavin T. Curr Res Pharmacol Drug Discov. 2024;7:100196. Published 2024 Jul 8. doi:10.1016/j.crphar.2024.100196

13. Kalitnik A, Lassota A, Polańska O, et al. Experimental methods for studying amyloid cross-interactions. Protein Sci. 2025;34(6):e70151. doi:10.1002/pro.70151

14. Sheik DA, Chamberlain JM, Brooks L, Clark M, Kim YH, Leriche G, et al. Hydrophobic Nanoparticles Reduce the β-Sheet Content of SEVI Amyloid Fibrils and Inhibit SEVI-Enhanced HIV Infectivity. Langmuir. 2017;33(10):2596-2602. doi:10.1021/acs.langmuir.6b04295

15. Hammarström P, Nyström S. Viruses and amyloids - a vicious liaison. Prion.;17(1):82-104. doi:10.1080/19336896.2023.2194212. PMID: 36998202; PMCID: PMC10072076.

16. Nyström S, Hammarström P. Amyloidogenesis of SARS-CoV-2 Spike Protein. J Am Chem Soc.;144(20):8945-8950. doi:10.1021/jacs.2c03925

17. Dodd-O J, Roy A, Siddiqui Z, et al. Antiviral fibrils of self-assembled peptides with tunable compositions. Nat Commun. 2024;15(1):1142. Published 2024 Feb 7. doi:10.1038/s41467-024-45193-3

18. Christ W, Kapell S, Sobkowiak MJ, Mermelekas G, Evertsson B, Sork H, et al. SARS-CoV-2 and HSV-1 induce amyloid aggregation in human CSF resulting in drastic soluble protein depletion. ACS Chem Neurosci.;15:4095–4104. doi:10.1021/acschemneuro.4c00636.

19. Xue C, Lin TY, Chang D, Guo Z. Thioflavin T as an amyloid dye: fibril quantification, optimal concentration and effect on aggregation. R Soc Open Sci. 2017;4(1):160696. doi: 10.1098/rsos.160696

20. Desroches MJ, Chaudhary N, Omanovic S. PM-IRRAS investigation of the interaction of serum albumin and fibrinogen with a biomedical-grade stainless steel 316LVM surface. Biomacromolecules. 9(4):1257–1266. doi:10.1021/bm070289d.

21. Behan PO, Behan WM. Postviral fatigue syndrome. Crit Rev Neurobiol. 1988;4(2):157-178.

22. Hickie I, Davenport T, Wakefield D, Vollmer-Conna U, Cameron B, Vernon SD, Reeves WC, Lloyd A; Dubbo Infection Outcomes Study Group. Post-infective and chronic fatigue syndromes precipitated by viral and non-viral pathogens: prospective cohort study. BMJ. 2006;333(7568):575. doi: 10.1136/bmj.38933.585764.AE. Epub 2006 Sep 1. PMID: 16950834; PMCID: PMC1569956.

23. Bassetti CL, Pelayo R, Guilleminault C. Idiopathic hypersomnia. In: Kryger MH, Roth T, Dement WC, editors. Principles and practice of sleep medicine. 4th ed. Philadelphia (PA): Elsevier Saunders; 2005. p. 656–664. doi:10.1016/B0-72-160797-7/50073-2.

24. DiNardo AR, Netea MG, Musher DM. Postinfectious Epigenetic Immune Modifications - A Double-Edged Sword. N Engl J Med. 2021;384(3):261-270. doi:10.1056/NEJMra2028358. PMID: 33471978; PMCID: PMC8053819.

25. Liu WL, Pi Z, Luo P, Mao W. Epigenetic remodeling after viral pneumonia accelerates lung tumorigenesis. Trends in Molecular Medicine.2026;04:003.doi: 10.1016/j.molmed.

26. Xiao Y. and Vermund SH. DNA methylation in long COVID. Front.Virol.(2024);4:1371683. doi:10.3389/fviro.2024.1371683

27. Theoharides TC, Stewart JM, Hatziagelaki E, et al. Brain “fog,” inflammation and obesity: key aspects of neuropsychiatric disorders improved by luteolin. Front. Neurosci.2015;9:225. doi:10.3389/fnins.2015.00225

28. VanElzakker MB, Brumfield SA, Lara Mejia PS. Neuroinflammation and cytokines in myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS): a critical review of research methods. Frontiers in Neurology 2019; 9:1033. doi:10.3389/fneur.2018.01033.

29. Azargoonjahromi A. The duality of amyloid-β: its role in normal and Alzheimer’s disease states. Mol Brain. 2024;17(1):44. Published 2024 Jul 17. doi:10.1186/s13041-024-01118-1

30. Cheng S, Duan D, Cui H, et al. Serum amyloid A: multifaceted roles in inflammation and cellular interactions. Immunol Res. 2025;73(1):148. Published 2025 Oct 20. doi:10.1007/s12026-025-09704-8