Association of cognitive function with Neurofilament light chain in the aqueous humor of human eye

Abstract

Objectives: To evaluate the predictive clinical role of neurofilament light chain (NfL), amyloid-β (Aβ), glial fibrillary acidic protein (GFAP), and phosphorylated tau at threonine 181 (p-tau181) proteins in human aqueous humor (AH) and quantify the retinal macular microvascular parameters by optical coherence tomography angiography (OCTA) as early diagnostic markers of Alzheimer’s disease (AD). Methods: This prospective, single-site, cross-sectional, cohort study enrolled 55 participants, including 38 patients with neovascular age-related macular degeneration (nAMD) and 17 individuals with senile cataracts. The single-molecule array platform was used to quantitatively measure the levels of AH NfL, Aβ40, Aβ42, GFAP, and p-tau181 proteins in AH. The mini-mental state examination (MMSE) score was used to assess the global cognitive function. OCTA scan with 6 × 6 mm macular area was used to quantify the retinal thickness and microvascular densities of superficial retinal capillary plexuses and deep retinal capillary plexuses. Results: NfL, Aβ40, Aβ42, GFAP, and p-tau181 were detected in all AH samples by Simoa platform. Individuals with cataract had higher concentrations of NfL and p-tau181 but lower Aβ40 and Aβ42 and similar GFAP compared to those with nAMD. Lower MMSE scores showed a negative correlation with NfL concentration of AH not only in the nAMD group (p = 0.043), but also in the cataract group (p = 0.032). However, the MMSE scores were not associated with the levels of Aβ40, Aβ42, GFAP, or p-Tau181. Further analysis found that the Aβ40 and Aβ42 concentrations showed a strong positive correlation (p < 0.0001). In addition, the NfL concentration showed a mild positive correlation with that of GFAP in the cataract group (p = 0.021). Although it has not reached statistical significance, there was a correlation between the levels of NfL and Aβ42 in the nAMD group (p = 0.051). Moreover, the macular superficial vessel density values had a negative correlation with the concentration of NfL (p = 0.004) but a positive correlation with MMSE scores (p = 0.045). The macular deep vessel density values were negatively correlated with the concentration of p-tau181 (p = 0.031) and positively correlated with MMSE scores (p = 0.020). Conclusion: The examination of AD-related biomarkers in human AH and OCTA may improve the ocular-based AD detection methods and contribute to forestalling the progression of preclinical AD.