Herpes simplex virus, early neuroimaging markers and incidence of Alzheimer's disease

Abstract

While previous studies suggest the implication of herpes simplex virus (HSV) in the onset of Alzheimer's disease (AD), no study has investigated its association with early neuroimaging markers of AD. In the Three-City and the AMI cohorts, the associations between HSV infection and (i) hippocampal volume (n = 349), (ii) white matter alterations in the parahippocampal cingulum and fornix using diffusion tensor imaging (n = 260), and (iii) incidence of AD (n = 1599) were assessed according to APOE4 status. Regardless of APOE4 status, infected subjects presented (i) significantly more microstructural alterations of the parahippocampal cingulum and fornix, (ii) lower hippocampal volumes only when their anti-HSV IgG level was in the highest tercile-reflecting possibly more frequent reactivations of the virus (p = 0.03 for subjects with a high anti-HSV IgG level while there was no association for all infected subjects, p = 0.19), and (iii) had no increased risk of developing AD. Nevertheless, among APOE4 carriers, infected subjects presented lower hippocampal volumes, although not significant (p = 0.09), and a two or three times higher risk of developing AD (adjusted Hazard ratio (aHR) = 2.72 [1.07-6.91] p = 0.04 for infected subjects and aHR = 3.87 [1.45-10.28] p = 0.007 for infected subjects with an anti-HSV IgG level in the highest tercile) while no association was found among APOE4 noncarriers. Our findings support an association between HSV infection and AD and a potential interaction between HSV status and APOE4. This reinforces the need to further investigate the infectious hypothesis of AD, especially the associated susceptibility factors and the possibility of preventive treatments.

Fig. 1. Association between the levels of anti-HSV IgG in terciles and hippocampal volume or white matter integrity.

Adjusted linear regression models. 3C and AMI cohorts. T1, T2, and T3: 1st, 2nd, and 3rd terciles of anti-HSV IgG, TIV: total intracranial volume, fornix and cingulum: parahippocampal fornix and cingulum. The bars show the regression coefficients with their 95% confidence interval (reference: uninfected subjects). Linear regression models are adjusted for age at serology, sex, level of education, presence of at least one allele of APOE4, hypertension, diabetes, and cohort.