Female sex is linked to a stronger association between sTREM2 and CSF p-tau in Alzheimer's disease

Abstract

In Alzheimer's disease (AD), Aβ triggers p-tau secretion, which drives tau aggregation. Therefore, it is critical to characterize modulators of Aβ-related p-tau increases which may alter AD trajectories. Here, we assessed whether factors known to alter tau levels in AD modulate the association between fibrillar Aβ and secreted p-tau₁₈₁ determined in the cerebrospinal fluid (CSF). To assess potentially modulating effects of female sex, younger age, and ApoE4, we included 322 ADNI participants with cross-sectional/longitudinal p-tau₁₈₁. To determine effects of microglial activation on p-tau₁₈₁, we included 454 subjects with cross-sectional CSF sTREM2. Running ANCOVAs for nominal and linear regressions for metric variables, we found that women had higher Aβ-related p-tau₁₈₁ levels. Higher sTREM2 was associated with elevated p-tau₁₈₁, with stronger associations in women. Similarly, ApoE4 was related to higher p-tau₁₈₁ levels and faster p-tau₁₈₁ increases, with stronger effects in female ApoE4 carriers. Our results show that sex alone modulates the Aβ to p-tau axis, where women show higher Aβ-dependent p-tau secretion, potentially driven by elevated sTREM2-related microglial activation and stronger effects of ApoE4 carriership in women.

Keywords: Alzheimer’s Disease; Microglia; Sex Differences; p-tau; sTREM2.

Figure 1. Main effects of potential modulators and CSF p-tau₁₈₁ across Aβ− cognitively normal plus Aβ+ ranging from cognitively normal to AD demented (longitudinal sample: N = 322; sTREM2 sample: N = 454).

Significant main effects were found between CSF sTREM2 and cross-sectional CSF p-tau₁₈₁ (C, p < 2e−16) and between ApoE4 carriership (non-carriers: n = 185; carriers: n = 137) and cross-sectional (F, p = 7.69e−11) and longitudinal (G) p-tau₁₈₁ levels. Sex (male: n = 156; female: n = 166) (A, B) and age (D, E) were not associated with higher p-tau₁₈₁ levels or faster p-tau₁₈₁ increases. F and p-values were derived for nominal variables (A, B, F, G) from ANCOVAs, and standardized beta-estimates (β) and p-values were derived for metric variables (C, D, E) from linear regressions. (A) and (B) were controlled for age, education, clinical status, and amyloid-PET (in centiloid); (C), (F), and (G) were controlled for sex, age, education, clinical status, and centiloid; (D) and (E) were controlled for sex, education, clinical status, and centiloid. The center line of each boxplot (A, B, F, G) indicates the median (50th percentile), the bounds of the box show the interquartile range (25–75th percentiles), and the whiskers extend to 1.5 times the interquartile range from the edges of the box. * = p_FDR < 0.05.

Figure 2. Interaction effects of potential modulators on the association between amyloid-PET (in centiloid) and CSF p-tau₁₈₁ across Aβ− cognitively normal plus Aβ+ ranging from cognitively normal to AD demented (longitudinal sample: N = 322; sTREM2 sample: N = 454).

A significant interaction effect was found between centiloid and sex (male: n = 156; female: n = 166) on cross-sectional (A) and longitudinal (B) p-tau₁₈₁ levels (B only borderline significant after FDR correction). No significant interaction effects were found between centiloid and sTREM2 (C), age (D, E) or ApoE4 status (non-carriers: n = 185; carriers: n = 137) (F, G) on cross-sectional/longitudinal p-tau₁₈₁ levels. Standardized beta-estimates (β) and p-values were derived from linear regressions. (A) and (B) were controlled for age, education, and clinical status; (C), (F), and (G) were controlled for sex, age, education, and clinical status, and (D) and (E) were controlled for sex, education, and clinical status. * = p_FDR < 0.05; # = p_FDR < 0.1.

Figure 3. Sex as modulator for the effects of CSF sTREM2 and ApoE4 carriership on CSF p-tau₁₈₁ across Aβ− cognitively normal plus Aβ+ ranging from cognitively normal to AD demented (ApoE4 sample: N = 322; sTREM2 sample: N = 454).

Female sex compared to male sex (male: n = 244; female: n = 210) is associated with a stronger association between sTREM2 and p-tau₁₈₁ levels (A). Stronger association between ApoE4 carriership (non-carriers: n = 185; carriers: n = 137) and p-tau₁₈₁ in women than in men (B). 3-way interaction effect of centiloid, sTREM2, and sex on p-tau₁₈₁ levels with women showing a stronger association between centiloid and p-tau₁₈₁ with increasing sTREM2 levels (β = 0.312, T = 2.316, p = 0.021, p_FDR = 0.063) (C). Standardized beta-estimates (β) and p-values were derived from linear regressions. (A) and (B) were controlled for age, education, clinical status, and centiloid. (C) was controlled for age, education, and clinical status. The center line of the boxplot (B) indicates the median (50th percentile), the bounds of the box show the interquartile range (25–75th percentiles), and the whiskers extend to 1.5 times the interquartile range from the edges of the box. * = p_FDR < 0.05.