# MALT1 in cerebrospinal fluid: a prognostic biomarker and potential therapeutic target in Alzheimer’s disease

## TL;DR

This study identifies MALT1 in cerebrospinal fluid as a potential biomarker and therapeutic target for Alzheimer's disease.

## Contribution

The study introduces MALT1 as a novel prognostic biomarker and therapeutic target for Alzheimer’s disease.

## Key findings

- MR analysis identified 15 CSF metabolites associated with Alzheimer’s disease risk.
- ML models identified five robust predictors, with MALT1 showing strong association with adverse prognosis in AD.
- Transcriptomic analysis linked risk genes to immune activation and lipid metabolism pathways.

## Abstract

Alzheimer’s disease (AD) is a devastating neurodegenerative disorder, and early intervention remains the only reliable strategy to slow its progression. Notably, cerebrospinal fluid (CSF) metabolites play a crucial role in the early diagnosis of AD, making their investigation highly significant.

We integrated two-sample Mendelian randomization (MR), transcriptomic, and machine learning (ML) analyses to identify causal CSF metabolites and their downstream molecular mediators in AD. MR assessed the causal effects of 338 CSF metabolites on AD risk, while integrated GEO datasets (GSE4757, GSE48350, and GSE122063) were analyzed to identify 50 differentially expressed associated genes (DEAGs). Immune infiltration and correlation analyses were performed to characterize immune infiltration. Predictive ML models, including Random Forest (RF), Support Vector Machine (SVM), Generalized Linear Model (GLM), and Extreme Gradient Boosting (XGBoost), were used to screen biomarkers, construct a diagnostic nomogram, and validate findings in vivo.

The MR analysis identified 15 potential CSF metabolites associated with AD. Elevated creatine levels (OR = 0.610, 95% CI: 0.441–0.845, p = 0.003) were protective against AD, whereas increased leucine levels (OR = 1.548, 95% CI: 1.210–1.981, p < 0.001) were associated with higher AD risk. Transcriptomic analysis revealed 50 DEAGs enriched in lipid metabolism, inflammation, and neural signaling pathways. Immune infiltration analysis demonstrated altered adaptive and innate immune populations, linking risk genes to immune activation. ML analysis identified five robust predictors (PLXDC2, DTNB, ALOX5, MALT1, and F13A1), with the SVM model showing optimal performance (AUC = 0.895) and an independently validated nomogram (AUC = 0.933). MR and GEO datasets (GSE138260) further confirmed MALT1 as a potential risk biomarker, and in vivo characterization supported its association with adverse prognosis in AD.

This study demonstrates a causal association between CSF metabolites and AD risk, highlighting MALT1 as a promising biomarker and potential therapeutic target for AD.