Biomarker changes associated with fornix deep brain stimulation in Alzheimer's disease

Abstract

Introduction: Deep brain stimulation of the fornix (fx-DBS) is being investigated for treatment of Alzheimer's disease (AD). The therapy aims at alleviating memory and cognitive circuit dysfunction. In preclinical models of AD, electrical stimulation of the memory circuit has demonstrated a possible disease-modifying potential. Here we examined changes resulting from fx-DBS in hippocampal atrophy and amyloid accumulation in AD patients with fx-DBS.

Methods: Repeated magnetic resonance imaging and positron emission tomography (PET) images acquired over the course of 12 months were used to assess changes in hippocampal volume in 36 ADvance trial patients compared to 40 matched untreated AD patients from the Alzheimer's Disease Neuroimaging Initiative, and in 10 separate patients with repeated flutemetamol PET and cerebrospinal fluid (CSF) markers.

Results: We observed a reduction of hippocampal atrophy and amyloid beta (Aβ) PET binding, and an increase in the CSF Aβ/total-tau ratio in DBS patients.

Discussion: These findings highlight the potential of fornix deep brain stimulation to modify AD biomarkers and possibly progression in some patients.

Highlights: Fornix deep brain stimulation (fx-DBS) is being investigated to treat Alzheimer's disease (AD). Results show that fx-DBS modifies imaging and cerebrospinal fluid (CSF) markers. It reduces hippocampal atrophy and increases the amyloid beta/total-tau CSF ratio. These findings highlight the potential of fx-DBS to modify AD.

Keywords: Alzheimer's disease biomarker; amyloid clearance; fornix DBS; hippocampal atrophy; neuromodulation therapy.

FIGURE 1

DBS targeting the columns of the fornix. (A) Three‐dimensional illustration of an exemplar DBS electrode targeting the fornix in MNI152 template space. The electrode contacts are shown positioned just anterior to the column of the fornix (turquoise, label taken from a high‐resolution hypothalamic region atlas ⁴⁵ ) against the backdrop of a 100‐micron resolution, FLASH 7 Tesla brain. (B) Coronal and (C) sagittal slices of fx‐DBS leads are displayed in native patient space. A, anterior; fx, fornix; DBS, deep brain stimulation; MNI, Montreal Neurological Institute; P, posterior.

FIGURE 2

Hippocampal volume changes observed over 12 months in fx‐DBS patients and ADNI controls. Both “on” and “off” patients within the fx‐DBS cohort exhibited significantly less hippocampal volume loss over 1 year compared to matched ADNI controls. (A) Boxplots of the percent hippocampal volume change over 1 year per group. (B) Line graph showing absolute hippocampal volume change in all groups across subjects, with the blue line showing the linear model within the group with confidence interval. Abbreviations: ADNI, Alzheimer's Disease Neuroimaging Initiative; DBS, deep brain stimulation; fx, fornix.

FIGURE 3

Changes in Aβ PET and CSF markers over time following fx‐DBS. (A) Single sagittal slices show the change in amyloid deposition as measured by PET tracer (¹⁸F‐Flutemetamol) binding following fx‐DBS treatment in each of the seven included Biomarkers trial patients. Higher tracer uptake is indicative of higher regional amyloid deposition. (B) Repeated PET measurements of amyloid deposition over time in the cerebral gray matter and (C) the bilateral hippocampi specifically show lower Aβ deposition following commencement of DBS treatment. (D) CSF measurements of the Aβ42/total‐tau ratio show an increase in this marker following fx‐DBS. (E) PET amyloid deposition changes within the cerebral gray matter are displayed separately for the time period between baseline and ≈6‐months follow‐up and between ≈6‐months and ≈12‐months follow‐up. (F) PET amyloid deposition changes within the bilateral hippocampi are shown within these time periods. For all graphs, group trends are shown in gray, whereas individual values and trends are also illustrated as colored dot‐dash lines. Aß42, amyloid beta 42; CSF, cerebrospinal fluid; DBS, deep brain stimulation; MNI, Montreal Neurological Institute; PET, positron emission tomography; SUVR, standardized uptake value ratio; t.tau, total tau; *p < 0.05.

FIGURE 4

Voxel‐wise changes in amyloid deposition over time following fx‐DBS. (A) Serial axial slices in MNI152 space showing voxels with a significant linear change over time on the top (increases in red‐yellow; decreases in blue‐light blue). The lower part shows, using the same anatomic reference slices, voxels with a significant quadratic change over time on the top. Positive quadratic change (in dark red‐red) describes a pattern of initial clearance followed by an amyloid increase between first and second follow‐ups; negative quadratic change (in dark blue‐blue) the opposite. (B) Plots of example voxels (a–f; location indicated in A) showing the modeled linear (purple) and quadratic (dark gray) trend of the local amyloid burden over time across the seven subjects. Individual values and trends are also provided in colored dot‐dash lines. DBS, deep brain stimulation; fx, fornix; MNI, Montreal Neurological Institute; PET, positron emission tomography; SUVR, standardized uptake value ratio.

FIGURE 5

Normative connectomic analyses elucidate relationships between stimulation location, clinical outcome, and imaging biomarkers. (A) One year hippocampal volume change: 3D rendering of streamlines associated with decreased atrophy or growth (left). The right shows that overlap of VTA with anterior commissure and fornix allows for accurate estimation of individual benefit validated using repeated 10‐fold cross‐validation. (B) One year ADAS‐Cog change: 3D rendering of streamlines associated with slowed cognitive decline or cognitive benefit. The right shows that overlap of VTA with anterior commissure and fornix allows for accurate estimation of individual benefit validated using repeated 10‐fold cross‐validation. (C) Longest follow‐up whole brain amyloid change: 3D rendering of streamlines associated with slowed amyloid deposition. The right shows that overlap of VTA with fornix allows for accurate estimation of individual benefit. Abbreviations: ADAS‐cog, Alzheimer's Disease Assessment Scale ‐ Cognitive Subscale; DBS, deep brain stimulation; VTA, volume of tissue activated.