Review Papers Claim by Claim

Key takeaway

The study reports that the engineered tetrapeptide Ac-HAEE-NH2 (HAEE) is detectable as an endogenous plasma peptide in healthy humans (mean 360.5 pg/mL) and is significantly reduced in patients clinically diagnosed with Alzheimer disease (AD) (generalized AD mean 175.0 pg/mL), with ROC AUC 0.832 for discrimination at a 232 pg/mL cutoff — proposing HAEE deficiency as a candidate bloodbased biomarker and suggesting exogenous HAEE as a therapeutic strategy HAEE

Comprehensive critical review and analysis

Scope and main claims

The authors claim three central findings: (1) the engineered tetrapeptide Ac-His-Ala-Glu-Glu-NH2 (HAEE) is an endogenous component of human plasma; (2) plasma HAEE concentration is significantly reduced in patients clinically diagnosed with Alzheimer disease (AD) versus controls; (3) HAEE deficiency may therefore serve as a bloodbased biomarker for AD and exogenous HAEE could be a disease modifying therapy. These claims are documented by quantitative HPLC MS measurements and ROC analysis (AUC 0.832) performed on 350 subjects (150 controls, 200 AD patients) Paper

Key experimental methods (what was done)

• Sample set: plasma collected 2023–2024 from four Russian clinical centers; n=350 total (150 healthy, 200 clinically diagnosed AD) with clinical diagnosis based on ICD 10 codes and local guidelines; mean ages: controls 53.3±14.1, AD 75.6±10.3 Cohort
• Analytical assay: solid phase extraction on MAX cartridges, HPLC separation, detection on Q-TOF and QQQ MS; internal standard [13C2]-HAEE; detailed instrument settings and acceptance criteria (SNR>10, retention time ±1%, qualifier ion ratios) are provided Analytical
• Statistics: normality/homogeneity tested by Kolmogorov and Levene tests; group comparisons by ANOVA Kruskal Wallis and Mann Whitney; ROC/AUC and cutoff selection via MedCalc with Youden J reported Statistical

Primary results (numerical)

Extracted group summary (means reported by authors):

Authors report overall group statistical significance Q=9.71 p < 0.00001 and Kruskal Wallis p < 0.00001 for generalized AD vs controls and present ROC metrics: AUC 0.832 (95% CI 0.787 0.870), sensitivity 84.1%, specificity 63.5% at a 232 pg mL cutoff, accuracy 82.65% Youden J 0.4764 PPV 75.4% NPV 75.0% PLR 2.45 NLR 0.36 Diagnostic

Strengths

• Analytical rigor: MS assay includes isotopically labeled internal standard with instrument parameters, identification criteria, stability, carryover, and crossplatform validation details — this helps analytical confidence Analytical
• Reasonable sample size for an initial biomarker study (350 total) with cross platform MS confirmation.

Limitations and critical caveats

1. Patient classification without biomarker confirmation: diagnosis relied solely on clinical ICD 10 guidelines without amyloid PET CSF p tau or neuropathological confirmation, which leaves open misclassification and inclusion of AD related dementias (ADRD) rather than pathologically proven AD — a major limitation for any biomarker claim Clinical
2. Confounding by age and comorbidities: controls are much younger (mean 53.3 years) than AD group (mean 75.6 years), and the AD group has extensive comorbidities (cardiovascular disease diabetes etc) that could affect plasma peptide proteolysis clearance or MS signals; the authors do not provide age adjusted multivariable models in the excerpts provided — this raises risk that apparent HAEE differences are age or comorbidity driven rather than AD specific Cohort
3. Cross sectional design: single timepoint measurements cannot determine whether HAEE reduction is cause consequence or epiphenomenon; no longitudinal data showing HAEE decline prior to cognitive decline are presented Cross
4. Potential conflicts of interest: authors and funder (Lifemission LLC) hold patents and commercial interests linked to HAEE as biomarker and therapy; declared conflicts increase need for independent replication in unbiased cohorts Conflicts
5. Heterogeneity across AD stages and paradoxical means: reported stage means show unexpected pattern (e.g., AD severe mean 219.3 pg mL higher than generalized AD mean 175.0 and closer to control), suggesting high within group variance or sampling artefacts — the authors note no significant differences across AD stages (p=0.746) which complicates interpretation that HAEE tracks disease severity Stage
6. Preanalytical variability risk: samples shipped across centers sometimes on dry ice 1 3 days and stored at −20 C; freeze thaw and storage can affect small peptide stability — authors report stability experiments but cross center preanalytical variance remains a possible confounder Preanalytical

Interpretation and plausibility

Biologically, authors link HAEE function to anti amyloid activity and interactions with Aβ zinc dependent aggregation supported by prior preclinical work in AβPPPS1 mice and nematode models; they hypothesize HAEE may complex with Aβ and that HAEE deficiency could reduce protective complexes and contribute to pathology. These mechanistic suggestions are plausible but presently speculative from human plasma data alone — direct brain measurements functional human biochemical interaction data and demonstration of central nervous system presence in humans would be needed to support causal claims Preclinical

What would falsify the main claim

• If independent cohorts with age matched controls and biomarker verified AD fail to replicate endogenous HAEE detection or fail to show significant group differences (AUC substantially lower), the biomarker claim would be falsified — authors list similar falsifiability criteria in their dataset meta data Falsifiability
• Demonstration that HAEE signal derives from exogenous contamination or methodological artefact (e g peptide in collection tubes or reagents) would refute the endogenous assertion; the presence of labeled internal standard and crossplatform MS reduces but does not eliminate this risk Contamination

Practical value and next steps

Given the reported AUC 0.832 HAEE has potential as a screening biomarker but specificity is modest (63.5%). For clinical adoption the following are essential:

1. Independent replication in larger, geographically and ethnically diverse cohorts with age matched controls and with biomarker confirmed AD (amyloid PET CSF p tau) to test specificity to AD vs other dementias.
2. Longitudinal studies to show temporal change and predictive value prior to cognitive decline.
3. Multivariable analyses adjusting for age sex comorbidities renal function and medication that could alter peptide clearance.
4. Orthogonal confirmation of central presence (CSF brain tissue) and demonstration of functional HAEE Aβ interactions in human-derived matrices.
5. Open sharing of raw MS data and assay SOPs to allow interlaboratory reproducibility testing and external assay validation.

Visualization of reported group means

Figure source data come directly from the paper reported summary table and ROC analysis Table

Bottom line and confidence

Bottom line: the assay appears well developed and the initial dataset is encouraging that HAEE is measurable and differs between clinically diagnosed AD patients and controls; however substantial limitations (clinical diagnosis without biomarker confirmation, marked age differences between groups, potential COI) reduce confidence that HAEE is an AD specific biomarker at this stage. I estimate moderate confidence pending independent validation. Key actions for the field are replication with biomarkerverified AD and transparent interlaboratory assay comparison Authors

Author review links