Taxon-level causal claims linking gut bacteria to pre-symptomatic amyloid are not yet carried by the rodent perturbation evidence

Question: What causal evidence from germ-free, antibiotic-treated, or fecal microbiota transplantation rodent models links specific gut bacterial taxa to cerebral amyloid-beta deposition before symptom onset?

Answer

Across the retrieved papers, the evidence supports a weaker claim than the question presumes: whole-community manipulations (germ-free status, facility sanitary status, FMT, antibiotic depletion) do alter amyloid burden, glial state, and cognition in AD mouse models, but no retrieved study closes the loop from a named bacterial taxon to quantified cerebral Aβ deposition in the pre-symptomatic window. The two FMT studies that measure Aβ report burden changes attributable to donor community or injury context rather than to an identified organism — FMT reduced Aβ deposition in sham 5xFAD mice yet the effect shifted with donor source and was absent after TBI [2], and conventionally housed donors enhanced amyloid pathology while facility-matched donors did not [4]. Germ-free work in 3xTg and 5xFAD mice reports immune and cognitive endpoints rather than plaque counts [14]. Where a single organism is dosed (Bacteroides fragilis), the reported readouts are frontal-cortex transcripts and metabolites, not Aβ [5]. Taxon names enter the literature mainly from human cohorts, and the one genetic-causal design in the set fails its own sensitivity analyses [15]. The mechanistically specific evidence is molecular, not taxonomic: bacterial functional amyloids, the conserved polysaccharide PNAG, and the metabolite imidazole propionate each modify Aβ or AD-like pathology in the absence of a defined colonizing species [1] [10] [8].

Germ-free AD models perturb immunity and cognition but leave plaque burden unmeasured

The most direct germ-free test in this corpus compares 3xTg and 5xFAD mice with and without a microbiome and reports that systemic immune responses were microbiome-modulated and sex-dependent, and that absence of a microbiome reduced cognitive deficits, primarily in females [14]. That study's own framing is that preclinical work shows the microbiome affects "deposition and/or clearance" of Aβ, but its reported outcomes are adaptive immune populations in brain-associated and systemic tissues rather than amyloid quantification [14]. Independent axenic work found no effect of microbiota on short- and long-term memory or novel object recognition when germ-free mice were compared with defined minimal or moderately diverse communities, with the reliable phenotype being anxiety-like behaviour in colonized animals [18]. The germ-free behavioural signal is therefore neither uniformly replicated nor tied to an amyloid readout [14] [18].

FMT changes amyloid burden, but the variance tracks donor environment and injury state, not donor taxa

In 5xFAD mice, sex-matched FMT from C57BL/6 donors reduced Aβ deposition and shifted plaque morphology in sham animals, reduced microgliosis in males, yet did not attenuate TBI-induced amyloid escalation, and alpha diversity was unchanged while beta diversity showed time-dependent post-injury shifts [2]. A facility-level comparison shows the same confound from the other direction: 5xFAD mice aged in a conventional facility developed enhanced amyloid pathology and a microbiota that differed from wild-type littermates — differences absent in specific-opportunistic-pathogen-free housing — and FMT from conventionally bred 5xFAD donors impaired memory in wild-type recipients while SOPF-derived FMT did not [4]. This implicates the sanitary status of the donor community as a determinant of both pathology and transmissible phenotype, and it means donor "dysbiosis" in AD models can be an artefact of husbandry rather than a disease-linked taxonomic signature [4].

Candidate-taxon dosing in amyloid models reports glial and metabolic endpoints, not plaques

Chronic intermittent dosing of 3xTg-AD mice with B. fragilis from 8 weeks of age — before the 24-week amyloid-modeled timepoint — produced earlier and more consistent colonization in 3xTg-AD than wild-type mice, lower frontal-cortex GFAP, SLC1A3 and FOXO3 expression, and restoration of amino acid derivatives and neurotransmitters toward wild-type levels [5]. That design has the correct temporal architecture for a pre-symptomatic causal test but stops at neuroinflammation and neurotransmission, which the authors themselves flag as requiring mechanistic follow-up [5]. Diet-based community remodelling in AD mice similarly reports composition shifts — increased Bifidobacterium and decreased Bacteroidetes in AD mice, with ketogenic interventions restoring over half of the altered taxa and raising Lachnospiraceae and Muribaculaceae — without an amyloid endpoint [7].

Named taxa come from human association studies, and the genetic-causal design undercuts them

In asymptomatic adults, higher cerebral Aβ was associated with increased Alistipes communis and decreased Anaerotignum faecicola and Eubacterium rectale, with lower predicted GABA biosynthetic capacity [9]. In middle-aged Framingham participants, Aβ burden associated with depletion of Faecalibacterium, Ruminococcus and Butyricicoccus and enrichment of Alistipes, Bacteroides and Barnesiella, more strongly in APOE ε4 carriers [17]. Fecal propionate, isovalerate and propionate-producing bacteria were inversely associated with amyloid status in a 287-participant cohort [16]. These are cross-sectional or observational, and the bidirectional Mendelian randomization analysis in this set, which initially flagged Dialister and other traits, concluded that horizontal pleiotropy, genetic confounding or reverse causation made a causal reading untenable [15]. A 16-year population cohort likewise found only a modest compositional association, with Dorea linked to decreased dementia risk [13].

The mechanistically credible candidates are molecules, not species

Microbial amyloids FapC and CsgA from Pseudomonas aeruginosa and Escherichia coli hyperactivate microglia against Aβ fibrils and converge with Aβ in phagocytic compartments, a phenotype not seen with Aβ fibrils alone, which the authors interpret as microbial amyloid triggering progressive AD pathology through microglia-driven neuroinflammation [1]. Curli content (CsgA) is reported higher in AD patients than controls by bioinformatic analysis of gut metagenomes [3]. The conserved microbial polysaccharide PNAG colocalizes with plaques in human AD tissue, drives TLR2-dependent inflammasome signalling and Aβ accumulation in human neuronal and 3D brain models, and PNAG vaccination reduced amyloid pathology and glial activation in APP/PS1 mice [10]. Chronic imidazole propionate administration exacerbated AD-like pathology in mice via endothelial barrier impairment and GSK3β-dependent tau hyperphosphorylation, with a chromosome 12 locus linking plasma ImP to AD risk in humans [8]. Each of these is a manipulable molecular agent whose producing taxa remain putative [8] [10].

Conflicts and gaps

Directionality is unresolved. Germ-free 3xTg and 5xFAD mice show reduced cognitive deficits [14], yet germ-free male mice show heightened neuroinflammation and larger lesion volumes after brain injury while antibiotic depletion reduces post-injury neuroinflammation [6], and antibiotic depletion in aged mice rejuvenates the brain and improves hippocampal memory [12]. In Drosophila, microbiome depletion worsened Aβ42 and tau accumulation [11] — the opposite sign to a "microbiome promotes amyloid" model, albeit in an invertebrate. No retrieved rodent study reports antibiotic depletion with quantified plaque burden in an AD model, and none mono-colonizes a germ-free amyloid model with a named human-derived Aβ-associated taxon. The pre-symptomatic requirement is also largely unmet: the 8-week B. fragilis dosing precedes plaque onset but measures no plaques [5], and human pre-symptomatic taxa [9] have no rodent validation in this corpus.

What would change the answer

Gnotobiotic mono-colonization of germ-free APP/PS1 or 5xFAD mice with the taxa now only correlated with pre-symptomatic Aβ — Alistipes communis, Bacteroides spp., or curli-producing E. coli — with longitudinal Aβ-PET or CSF Aβ42:40 sampling before plaque onset, would convert association into taxon-level causation [9] [17]. Antibiotic depletion timed to weaning and continued through the pre-plaque window, with plaque and soluble/insoluble Aβ fractionation as primary endpoints, would test sufficiency of the resident community. FMT from Aβ-PET-stratified asymptomatic human donors into germ-free or antibiotic-depleted amyloid mice, with donor taxa tracked by metagenomics, would bridge the human pre-symptomatic signatures to an experimental endpoint [9]. Isogenic bacterial mutants lacking csgA or PNAG biosynthesis, tested in colonized amyloid mice, would separate the molecular mechanisms from the taxa that produce them [1] [10]. Finally, husbandry status must be a controlled factor, since facility sanitary status alone shifts amyloid pathology and transmissible memory phenotypes [4].

References

[1] Javed, I., Davis, T. P., Forgham, H., Albornoz, E. A., Pietrogrande, G., Zhu, J., Chung, K. H. K., Tahir, M. S., Ali, S. A., Qiao, R., Kakinen, A., Wolvetang, E. J., Woodruff, T. M., Aguado, J., Otzen, D. E.. "Interaction of Gut-Microbial Amyloids with Endogenous Amyloids can Drives Microglial Hyperactivity and Neuroinflammation in Alzheimers Disease". bioRxiv. 2025-01-15T00:00:00.000Z. URL: https://doi.org/10.1101/2025.01.15.633290

[2] Soriano, S., Marshall, A., Holcomb, M., Flinn, H., Burke, M., Kara, G., Scalzo, P., Villapol, S.. "Sex-specific effects of fecal microbiota transplantation on TBI-exacerbated Alzheimer's pathology in mice". bioRxiv. 2025-09-17T00:00:00.000Z. URL: https://doi.org/10.1101/2025.09.11.675717

[3] Wojciechowska, A. W., Wojciechowski, J. W., Zielinska, K., Soeding, J., Kosciolek, T., Kotulska, M.. "Aggregating gut: on the link between neurodegeneration and bacterial functional amyloids". bioRxiv. 2025-03-07T00:00:00.000Z. URL: https://doi.org/10.1101/2024.11.26.624671

[4] Caroline Ismeurt-Walmsley, Patrizia Giannoni, Florence Servant, Linda-Nora Mekki, Kevin Baranger, Santiago Rivera, Philippe Marin, Benjamin Lelouvier, Sylvie Claeysen. "The same but different: impact of animal facility sanitary status on a transgenic mouse model of Alzheimer's disease". arxiv. 2024-12-25T00:00:00.000Z. URL: https://arxiv.org/abs/2412.18258

[5] Conn, K. A., Schwartz, O., Dikshit, S., Barroso-Montalvo, D. L., Hattan, D., Herman, C., Wood, C. V., Borsom, E. M., Caporaso, J. G., Aron, A., Cope, E. K.. "Bacteroides fragilis modulates gut microbiome community composition, frontal cortex gene expression, and fecal and frontal cortex metabolites in a mouse model of Alzheimers disease pathologies". bioRxiv. 2026-05-30T00:00:00.000Z. URL: https://doi.org/10.64898/2026.05.29.728805

[6] Flinn, H., Marshall, A., Holcomb, M., Cruz-Pineda, L., Soriano, S., Treangen, T., Villapol, S.. "Antibiotic-induced changes to the gut microbiome attenuate neuroinflammation in a mouse model of TBI". bioRxiv. 2025-08-15T00:00:00.000Z. URL: https://doi.org/10.1101/2024.05.11.593405

[7] M'Bra, P. E. H., Avino, M., Lim, E. N. K., Mayhue, M., Balthazar, P., Aumont, A., Prevost, K., Masse, E., Fernandes, K.. "Ketogenic interventions prevent alterations of the gut microbiome in transgenic Alzheimer Disease mice". bioRxiv. 2025-02-25T00:00:00.000Z. URL: https://doi.org/10.1101/2025.02.25.640129

[8] Vemuganti, V., Kang, J. W., Zhang, Q., McGregor, E. R., Hilser, J. R., Martinez, R. A., Harding, S. J., Harpt, J. L., Beck, K. R., Bussan, H. E., Kuehn, J. F., Deming, Y., Johnson, S. C., Asthana, S., Zetterberg, H., Blennow, K., Engelman, C., Allayee, H., Anderson, R., Ulland, T. K., Bäckhed, F., Bendlin, B. B., Rey, F. E.. "Gut bacterial metabolite imidazole propionate potentiates Alzheimer's disease pathology". bioRxiv. 2026-04-20T00:00:00.000Z. URL: https://doi.org/10.1101/2025.06.08.657719

[9] Shoubridge, A. P., Manning, S. K., Tan, S., Kidson, S., Richard, A., Trinder, J., Miller, S., Laver, K., Rahja, M., Robertson, J., Crotty, M., Taylor, S. L., Rogers, G. B.. "Change in gut microbiome precedes symptomatic Alzheimer's disease and is correlated with amyloid-beta accumulation in the brain". medRxiv. 2026-10-19T00:00:00.000Z. URL: https://doi.org/10.64898/2026.10.15.26365677

[10] Marcet, E. C., Vinacur, M., Zaidi, T., Nguyen, N., Augart-Welwood, A., Su, G., Landau, D., LaVancher, E., Arkhangelskiy, A., Power, L. H., Kelly, M. E., Stein, C. R., Carolan, E., Caldarone, B., Lemere, C. A., Wasen, C., Cox, L. M., Cairns, D. M., Kaplan, D., Pier, G. B., Cywes-Bentley, C.. "Persistent microbial material contributes to Alzheimer disease and is targetable by vaccination". bioRxiv. 2026-07-23T00:00:00.000Z. URL: https://doi.org/10.64898/2026.07.20.739171

[11] Madamanchi, K., Gurrala, S., Watson, J., Melkani, G. C.. "Microbiome Integrity Protects Against Glial-Mediated Tau and Amyloid Pathology Through Circadian and Autophagy Homeostasis". bioRxiv. 2026-05-22T00:00:00.000Z. URL: https://doi.org/10.64898/2026.05.20.726549

[12] Gasperini, C., Holton, K. M., Limone, F., Juttu, M., DeMeo, C. C., Kekrtova, K., Patankar, S., Wells, K. M., Giadone, R., Ben Driss, L., Wei, G., Kiem, A., Xu, Q., Lee, R., Friedlander, M., Scadden, D. T., Rubin, L. L.. "Microbiome depletion rejuvenates the aging brain". bioRxiv. 2026-02-15T00:00:00.000Z. URL: https://doi.org/10.64898/2026.02.13.705770

[13] Tynkkynen, J., Kambur, Oleg, O., Niiranen, T., Lahti, L., Ruuskanen, M. O., McDonald, D., Jousilahti, P., Gazolla Volpiano, C., Meric, G., Inouye, M., Liu, Y., Khatib, L., Patel, L., Salomaa, V., Knight, R., Havulinna, A.. "Gut Microbiome and Risk of Dementia - a Prospective, Population-Based Study". medRxiv. 2026-02-22T00:00:00.000Z. URL: https://doi.org/10.64898/2026.02.15.26345196

[14] Bostick, J. W., Connerly, T. J., Thron, T., Needham, B. D., de Castro Fonseca, M., Kaddurah-Daouk, R., Knight, R., Mazmanian, S. K.. "The microbiome shapes immunity in a sex-specific manner in mouse models of Alzheimer's disease". bioRxiv. 2024-10-24T00:00:00.000Z. URL: https://doi.org/10.1101/2024.05.07.593011

[15] Monaghan, G., Hatcher, C., Anderson, E. L., Wade, K. H.. "The causal relationship between gut microbiome composition and Alzheimer's disease: A two-sample Mendelian randomization analysis". medRxiv. 2025-08-21T00:00:00.000Z. URL: https://doi.org/10.1101/2025.08.20.25333769

[16] Kuehn, J. F., Zhang, Q., Heston, M. B., Kang, J. W., Harding, S. J., Davenport-Sis, N. J., Peter, D. C., Kerby, R. L., Vemuganti, V., Schiffmann, E. C., Tallon, M. M., Harpt, J. L., Hajra, A., Wheeler, J. L., Shankar, S., Mickol, A., Zemberi, J., Chow, H., Zhang, E., Clements, E., Noughani, H., Forst, A., Everitt, G., Kollmorgen, G., Quijano-Rubio, C., Christian, B. T., Carlsson, C. M., Johnson, S. C., Asthana, S., Zetterberg, H., Blennow, K., Ulland, T., Rey, F. E., Bendlin, B. B.. "Fecal Short-Chain Fatty Acids Vary by Sex and Amyloid Status". medRxiv. 2025-08-01T00:00:00.000Z. URL: https://doi.org/10.1101/2025.07.31.25332523

[17] Ngouongo, Y. W., Bernal, R., NJAMNSHI, W. Y., Satizabal, C. L., Beiser, A., Ruiz, A., NJAMNSHI, A. K., Vasan, R. S., Seshadri, S., Himali, J. J., Fongang, B.. "Altered Gut Microbiota Mediates the Association between APOE Genotype and Amyloid-β Accumulation in Middle-Aged Adults". bioRxiv. 2025-07-31T00:00:00.000Z. URL: https://doi.org/10.1101/2025.07.27.667048

[18] Rutsch, A., Iachizzi, M., Kirundi, J., Kantsjoe, J. B., Planchette, A. L., Mueller, T., Schmitz, W., Radenovic, A., Doenlen, R., Gomez de Aguero, M., Ronchi, F.. "Assessment of neurobehavioural traits under axenic conditions: an approach for multiple longitudinal analyses in the same mouse.". bioRxiv. 2024-11-21T00:00:00.000Z. URL: https://doi.org/10.1101/2024.11.21.623170

How this survey was made

This survey answers one question: “What causal evidence from germ-free, antibiotic-treated, or fecal microbiota transplantation rodent models links specific gut bacterial taxa to cerebral amyloid-beta deposition before symptom onset?”

We keep our own collection of recent preprints and refresh it every day from bioRxiv, medRxiv, arXiv, PsyArXiv, SocArXiv and ChemRxiv. For each survey, we search that collection for recent work and keep the strongest matches: papers that use the same terms as the question, and papers on the same topic in different words.

The text was written from the titles and abstracts of those papers; the full texts were not read. Statements that draw on a particular paper are marked with a number in brackets. The reference list just above is the 18 papers whose work this text draws on, chosen from the 62 papers those searches returned; the rest of the shortlist is not printed. It is not a complete review of everything published on the question.

Treat this as a starting point and a reading list, not a peer-reviewed review. The searches behind this survey ran on October 22, 2026 and covered work from the previous two years.