末梢血CHI3L1発現と早期発症アルツハイマー病におけるAPOE ε4状態との関連

YKL-40 (CHI3L1) is a glycoprotein secreted by reactive astrocytes and peripheral immune cells, implicated in inflammation and tissue remodeling in Alzheimer's disease (AD). While elevated CHI3L1 levels have been observed in cerebrospinal fluid and plasma, its expression at the transcript level in peripheral blood - and modulation by genetic risk factors such as APOE ε4 - remains poorly understood.

We analyzed peripheral blood CHI3L1 mRNA expression in a well-characterized cohort comprising individuals with biomarker-confirmed AD (n = 34), mild cognitive impairment (MCI; n = 31), and cognitively healthy controls (HC; n = 21). CHI3L1 expression levels were compared across diagnostic groups and stratified by APOE ε4 status, age at onset (early-onset < 65 years; late-onset ≥ 65), and sex. Correlations were examined between CHI3L1 and inflammatory gene transcripts (IL1B, TNF, MMP9, LRP1, and TREM2).

Peripheral CHI3L1 expression was elevated in individuals with early-onset AD (EOAD), particularly among APOE ε4 carriers (EOAD APOE ε4+, n = 13 vs. EOAD APOE ε4-, n = 8; p = 0.026). Stratified analyses revealed an exploratory association between CHI3L1 expression, APOE genotype, and sex, with the highest levels observed in female ε4 carriers with EOAD. Across diagnostic groups, CHI3L1 levels positively correlated with transcripts of IL1B, MMP9, and LRP1, with the strongest associations again in APOE ε4 + individuals. Notably, these effects were more pronounced in the MCI and AD groups than in healthy controls, indicating early immune activation in at-risk individuals.

Our exploratory findings indicate that peripheral CHI3L1 expression may reflect APOE ε4-linked immune activity, with a trend toward higher expression in EOAD and in female ε4 carriers. The observed genotype- and sex-dependent expression patterns indicate preliminary differences in peripheral immune activity that warrant replication in larger cohorts. Peripheral CHI3L1 may thus serve as a hypothesis-generating marker of genotype-linked inflammatory phenotypes rather than a validated biomarker.