Broad cross-system evidence chain
Human biofluids and brain tissue are connected to two AD mouse models, cultured cells, and C. elegans, providing multiple tests of the ULK1 hypothesis.
↳ Results, Figs. 1–7
Abstract
ULK1 (Atg1) initiates macroautophagy and mitophagy, which support neuronal growth and survival, yet how this pathway is disrupted in aging and Alzheimer's disease (AD) remains unclear. Here we report reduced ULK1 in serum and cerebrospinal fluid during aging in cognitively unimpaired participants from the COGNORM study (n = 75) and in patients with AD from the NorCog Memory Clinic Cohort (n = 316). In AD mice, ULK1 overexpression stimulates autophagic flux, reduces AD pathology and delays cognitive decline alongside increased phagocytic degradation of amyloid-β, reduced tauopathy and improved mitochondrial quality. Mechanistically, ULK1 upregulation increases autophagy and PINK1-, FUNDC1- and AMBRA1-associated mitophagy; higher autophagy and mitophagy increase cellular NAD+, which in turn deacetylates acetylated-Tau174 via the NAD+-SIRT1 axis, leading to reduced tauopathy. Using in vitro tau seeding assays and a Caenorhabditis elegans tau model, we validate the efficacy of ULK1 activators in inhibiting tauopathy. We propose that age-related decline in ULK1 leads to autophagy and mitophagy impairment and increases the progression of AD and identify ULK1 as a potential therapeutic target.
Key Claims
Primary findingSupported
In AD mice, ULK1 overexpression stimulates autophagic flux, reduces AD pathology and delays cognitive decline alongside increased phagocytic degradation of amyloid-β, reduced tauopathy and improved mitochondrial quality.
Fragile — the claim's weakest requirement is that it works through the path described. Experimental. Every point below is traceable to a quoted passage, but this forecast has not yet been validated against known replication outcomes.
two transgenic models support protective effects, but single-age cognition and small mechanistic samples limit the breadth of inference
Abstract; Results, Figs. 3–6
SecondaryTentative
reduced ULK1 in serum and cerebrospinal fluid during aging in cognitively unimpaired participants from the COGNORM study (n = 75) and in patients with AD from the NorCog Memory Clinic Cohort (n = 316)
small paired follow-up and confounded cohort comparisons conflict with null and positive age associations in regression analyses
Abstract; Results, Fig. 1a–b, g–h
SecondaryTentative
higher autophagy and mitophagy increase cellular NAD+, which in turn deacetylates acetylated-Tau174 via the NAD+-SIRT1 axis, leading to reduced tauopathy
correlated pathway readouts and one SIRT1-inhibitor experiment do not directly identify the full NAD+-mediated causal chain
Abstract; Results, Fig. 6
SecondaryTentative
Using in vitro tau seeding assays and a Caenorhabditis elegans tau model, we validate the efficacy of ULK1 activators in inhibiting tauopathy.
pooled fields and mixed activator results support proof of concept but not robust efficacy or mammalian translation
Abstract; Results, Fig. 7
SecondarySupported
higher baseline CSF ULK1 was associated with slower progression to AD
an adjusted longitudinal mixed model supports association, not a protective causal effect
Results, Fig. 1l
Derived from the full evaluation — not a separate score.
Strengths · Limitations
Strengths
Genetic testing across AD pathologies
ULK1 overexpression is evaluated in both 5xFAD amyloid and hTau.P301S tau models, with behavioral and pathological outcomes reported in each.
↳ Results, Figs. 3–6
Adjusted longitudinal progression analysis
The CDR-SB analysis models baseline CSF ULK1 by time while adjusting for demographic, clinical, Aβ, and p-tau variables.
↳ Results, Fig. 1l; Statistics and Reproducibility
Limitations
Non-independent observations drive inference
Several central cell and imaging analyses use neurons, fields, plaques, or technical repeats as inferential observations despite only three biological replicates or a small number of animals.
↳ Figs. 2j–k, 3q–s, 4e–f, 6i–l and 7a–e
Clinical comparisons retain major confounding
CU controls and AD participants differ in recruitment setting, age, sex, education, APOE status, sampling context, and biomarker laboratory, without a demonstrated adjusted disease-group ULK1 comparison.
↳ Table 1; Methods, Study Participants; Results, Fig. 1g–h
Aging narrative conflicts with regressions
The Discussion reports a null age association in CU participants and a positive ULK1–age association in AD, yet the Abstract and later Discussion assert age-related decline and a causative role.
↳ Abstract; Discussion, paragraphs 1 and 4; Supplementary Tables 5–6
Nabu Commentary
The paper earns its strongest score for contribution because it links human observations to amyloid and tau mouse models, cell assays, and a worm model. Methodological breadth does not offset the use of fields, neurons, and technical repeats as inferential units in several central experiments or the absence of an adequate confounding strategy for the cross-cohort ULK1 comparison. The adjusted CDR-SB model and mouse-level behavioral outcomes remain informative, but the NAD+–SIRT1 mechanism and pharmacological claims are less secure. The Abstract and later Discussion also state age-related decline and causal disease progression more firmly than the discordant human regressions support.
Check another paper
Author Note / Field Feedback
Nabu’s assessment, alongside the field’s view.
Are you an author of this paper?
Dimension breakdown
Study Quality
Limited2.7
Confidence highContribution
3.9
The paper meaningfully extends prior ULK1 work through human biofluids and brain tissue, two AD mouse models, cell assays, and a tau-worm model. The advance is substantial, although the human and drug-development claims remain less mature than the genetic model findings.
Abstract; Results, Figs. 1–7
In AD mice, ULK1 overexpression stimulates autophagic flux, reduces AD pathology and delays cognitive decline
Methodological Rigour
1.8
The design includes multiple perturbations, blinded analyses, two mouse models, and an adjusted longitudinal CDR-SB model. Methodological Rigour is substantially reduced because central cell experiments infer from fields or technical repeats, clinical groups differ without a demonstrated adjusted disease-group comparison, and randomisation was not used.
Table 1; Figs. 2j–k, 6i–l and 7a–e; Statistics and Reproducibility
No method of randomization was used to assign samples to experimental groups.
Reporting
3.2
The human-to-model-system argument is logically organized and methods and legends are generally detailed. The Abstract nevertheless presents the age-decline and NAD+–SIRT1 causal chains more firmly than the observational and perturbational evidence supports.
Abstract; Discussion, paragraphs 1 and 8
age-related decline in ULK1 leads to autophagy and mitophagy impairment and increases the progression of AD
Positioning
2.8
The Discussion engages Beclin1, AMPK–ULK1, anti-Aβ therapy, and the study's translational limitations. Its handling of the aging limitation is inconsistent because the null regression is acknowledged before later assertions of dramatic age-related decline and a causative role.
Discussion, paragraphs 1, 3–4, 6 and 8
Our data show a dramatic reduction of ULK1 in the cognitively normal older human population
Lower confidence on Methodological Rigour, Positioning — domain match limited.
Reliability
Caveats3 of 4 checks
Internal CoherenceNoted
The reported experimental results are broadly traceable, but central aging and causal interpretations do not consistently reflect the paper's discordant human associations. Several Methods–figure mismatches also affect the precision of experimental reporting.
Research ConductClean
Human consent and ethics approvals and animal IACUC approval are declared, and figure-level source data and uncropped blots are listed. No paper-grounded research-conduct concern was identified.
Flags: 2 declared / 5 total
Reference IntegrityClean
112 of 112 checkable references verified
113 references in manuscript 1 have no canonical index record — counted, but not index-checkable 3 references confirmed by manual review
Post-Publication RecordNot checked
Post-publication record check unavailable at this time.
Last checked: 08 Oct 2026
Sources: Retraction Watch ✓
Readiness
Where this paper’s evidence sits on the path from initial observation to real-world use.
Mechanistic
Genetic overexpression provides mammalian proof of biology, but Rac-BL-918 is tested only in cells and C. elegans. The paper itself identifies mouse and human iPSC-neuron validation as necessary next steps.
Results, Fig. 7; Discussion, limitations paragraph
further validation studies in mice and in human induced pluripotent stem cell-derived neurons
AI-generated, human-governed. Something look off? Contact us to request a review.