Blood-Brain Barrier Opening with MRI-guided Focused Ultrasound Elicits Meningeal Venous Permeability in Humans with Early Alzheimer Disease

Abstract

Background Opening of the blood-brain barrier (BBB) induced with MRI-guided focused ultrasound has been shown in experimental animal models to reduce amyloid-β plaque burden, improve memory performance, and facilitate delivery of therapeutic agents to the brain. However, physiologic effects of this procedure in humans with Alzheimer disease (AD) require further investigation. Purpose To assess imaging effects of focused ultrasound-induced BBB opening in the hippocampus of human participants with early AD and to evaluate fluid flow patterns after BBB opening by using serial contrast-enhanced MRI. Materials and Methods Study participants with early AD recruited to a Health Insurance Portability and Accountability Act-compliant, prospective, ongoing phase II clinical trial (ClinicalTrials.gov identifier, NCT03671889) underwent three separate focused ultrasound-induced BBB opening procedures that used a 220-kHz transducer with a concomitant intravenous microbubble contrast agent administered at 2-week intervals targeting the hippocampus and entorhinal cortex between October 2018 and May 2019. Posttreatment effects and gadolinium-based contrast agent enhancement patterns were evaluated by using 3.0-T MRI. Results Three women (aged 61, 72, and 73 years) consecutively enrolled in the trial successfully completed repeated focused ultrasound-induced BBB opening of the hippocampus and entorhinal cortex. Postprocedure contrast enhancement was clearly identified within the targeted brain volumes, indicating immediate spatially precise BBB opening. Parenchymal enhancement resolved within 24 hours after all treatments, confirming BBB closure. Transient perivenous enhancement was consistently observed during the acute phase after BBB opening. Notably, contrast enhancement reappeared in the perivenular regions after BBB closure. This imaging marker is consistent with blood-meningeal barrier permeability and persisted for 24-48 hours before spontaneous resolution. No evidence of intracranial hemorrhage or other adverse effect was identified. Conclusion MRI-guided focused ultrasound-induced blood-brain barrier opening was safely performed in the hippocampi of three participants with Alzheimer disease without any adverse effects. Posttreatment MRI reveals a unique spatiotemporal contrast enhancement pattern that suggests a perivenular immunologic healing response downstream from targeted sites. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Klibanov in this issue.

Figures

Graphical abstract

Figure 1:

Clinical trial flow diagram.

Figure 2:

MRI scans show blood-brain barrier (BBB) opening and closure within targeted brain volumes in a 73-year-old woman with Alzheimer disease. Postcontrast T1-weighted images at, A, baseline, B, immediately after treatment (Tx), and, C, 24 hours after treatment show parenchymal contrast material extravasation (arrow in B) due to BBB opening locally at the treatment sites and BBB closure with resolution of parenchymal contrast enhancement on repeat gadobutrol administration 24 hours later (arrow in C). D, Sagittal T2-weighted image shows three 5 × 5 × 7-mm target volumes demarcating the treated areas within the hippocampus and entorhinal cortex. E, Target sites shown on axial T2-weighted image. F, Axial image obtained with the T2* gradient-echo sequence immediately after treatment shows no evidence of signal dropout to suggest hemorrhage at the treated areas.

Figure 3:

Perivenular enhancement pattern observed immediately after treatment (Tx) in a 72-year-old woman with Alzheimer disease. A, Coronal postcontrast T1-weighted MRI scan at baseline and, B, coronal postcontrast T1-weighted MRI scan obtained immediately after treatment show parenchymal contrast enhancement in the hippocampus and entorhinal cortex (short arrow), as well as a more avid focus of enhancement in the hippocampal fissure along the course of the hippocampal sulcus vein (long arrow). Coronal postcontrast fluid-attenuated inversion recovery (FLAIR) MRI scans at, C, baseline and, D, immediately after treatment show perivenous enhancement (arrow) around the intrahippocampal sulcal vein. E, Sagittal postcontrast T1-weighted MRI scan obtained immediately after treatment shows parenchymal and intense linear perivenular enhancement in the hippocampal fissure. F, Zoomed-in version of E shows coronal levels I, II, and III. G, Coronal postcontrast FLAIR MRI scan at the level of the basal vein of Rosenthal shows patent venous lumen with surrounding ringlike perivenous enhancement (arrow). H, Coronal susceptibility-weighted MRI (SWI) scan shows the basal vein of Rosenthal. I, Coronal postcontrast T1-weighted MRI scan at the level of the hippocampal sulcus vein shows contrast enhancement surrounding the vein. J, Coronal SWI scan shows the hippocampal sulcus vein (arrow) as a small black dot. K, Coronal postcontrast FLAIR MRI scan at the level of the intrahippocampal sulcal vein shows comma-shaped contrast enhancement surrounding this small vein. L, Coronal SWI scan shows the intrahippocampal sulcal vein. M, Schematic image shows the perivenular enhancement pattern, which was observed on immediate posttreatment images in all three participants who underwent treatment. SWAN = T2* susceptibility-weighted angiography.

Figure 4:

MRI scans show spatiotemporal pattern of perivenular enhancement in a 73-year-old woman with Alzheimer disease. A, Sagittal T2-weighted pretreatment image shows three selected target sites within the hippocampus (rectangles). B, Sagittal postcontrast T1-weighted image immediately after treatment shows parenchymal contrast enhancement within the targeted hippocampus and entorhinal cortex, as well as more hyperintense linear enhancement within the hippocampal fissure (arrow) longitudinally along the course of the hippocampal veins. C, Sagittal precontrast T1-weighted image obtained 24 hours after treatment shows clearance of gadolinium-containing contrast material from both the interstitial and the perivenous and perivascular spaces (arrow). D, Sagittal postcontrast T1-weighted image obtained 24 hours after treatment shows perivenular enhancement within the hippocampal fissure (arrow) along the course of the veins, indicating perivenous blood–meningeal barrier permeability. There is no parenchymal enhancement at this time point because of closure of the blood-brain barrier. E, Sagittal minimal intensity projection image obtained by using axial T2* susceptibility-weighted angiographic imaging data shows the course of hippocampal veins, which appear as a linear dark stripe within the hippocampal fissure (arrow). Perivenous enhancement seen in B and D paralleled the course of these veins, as further shown in Figure 5.

Figure 5:

MRI scans show perivenular enhancement after gadobutrol clearance and blood-brain barrier (BBB) closure at 24 hours after treatment in a 61-year-old woman with Alzheimer disease. Precontrast coronal, A, T1-weighted and, B, fluid-attenuated inversion recovery (FLAIR) images, both obtained 24 hours after treatment (at same time point as Fig 3, C), show absence of contrast enhancement in the hippocampus and perivascular space after BBB closure. Postcontrast coronal, C, T1-weighted and, D, FLAIR images, both obtained 24 hours after treatment, show enhancement along the course of the hippocampal sulcal vein (arrows). This enhancement, shown on serial axial T1-weighted sections in, E–H (arrows), was present longitudinally along the course of the hippocampal veins and could be followed posteriorly to the basal vein of Rosenthal (H). I–L, T2* susceptibility-weighted angiography (SWAN) images, corresponding to the same levels as E–H, show blooming susceptibility effects of the hippocampal venous structures (arrows) from the same participant, coinciding with the region of linear enhancement. This perivenous enhancement indicates blood–meningeal barrier permeability and occurred transiently after treatment before complete resolution at all treated sites. M, Same area as in D, zoomed on the right hippocampus, shows perivenous enhancement (arrow). N–Q, Zoomed axial SWAN images (I–L) show the hippocampal venous structures (arrows).

Figure 6:

Observed tracer movement pattern after focused ultrasound-induced blood-brain barrier (BBB) opening and proposed glymphatic efflux route in humans. A, Coronal schematic view through the temporal lobes summarizes procedural effects of BBB opening. 1, Focused ultrasound energy was delivered, along with intravenous microbubbles, to the hippocampal target site (boxed area). 2, Contrast extravasation from capillaries into the interstitial space (yellow) indicates BBB opening within the targeted volume. 3, Contrast material flows downstream from the interstitial space to the perivenular space, along the proposed glymphatic efflux route, resulting in perivenous enhancement. 4, Reactive venous permeability is proposed to account for contrast material leakage into the perivenous sheath, suggesting a physiologic perivenular neuroimmune reaction. B, Normal flow of contrast material during baseline homeostatic state. With closed (intact) BBB, contrast material travels through the intravascular compartment and exits the cranial cavity via the dural venous sinuses and along the dura. C, Flow of contrast material during focused ultrasound–induced BBB opening. Because of acquired permeability of the capillary walls, gadobutrol extravasates from the intravascular compartment into the pericapillary or interstitial space and flows downstream to perivenous spaces. D, At 24 hours after treatment, precontrast images demonstrate gadolinium clearance from both the brain parenchyma and perivascular spaces due to rapid transit related to bulk fluid and solute flow. E, Postcontrast imaging at 24 hours after focused ultrasound. Notably, after repeat intravenous gadobutrol administration, no intraparenchymal contrast enhancement is seen because of closure of the BBB. However, gadolinium-based contrast agent accumulation reappears around the draining veins, suggesting permeability of the postcapillary venules and meningeal veins.