Neuroimaging of Acute Ebola Virus Disease in a Non-Human Primate Model
Margaret R. Lentz1, Jeffery R. Solomon2, Srikanth Yellayi1, Richard Bennett1, Dawn Traynor1, David Thomasson1, Anna Honko1, Lisa Hensley1, and Peter B. Jahrling1,3

1Integrated Research Facility, NIAID, National Institutes of Health, Frederick, MD, United States, 2Clinical Research Directorate/Clinical Monitoring Research Program, Frederick National Laboratory for Cancer Research, Leidos Biomedical Research, Inc., Frederick, MD, United States, 3Emerging Viral Pathogens Section, NIAID, National Institutes of Health, Frederick, MD, United States


The purpose of this study was to use MRI to assess alterations in the brain that occur in rhesus macaques infected with a variant of the Ebola virus (EBOV) isolated from the most recent outbreak. EBOV was found to induce signal alterations in susceptibility weighted imaging (SWI) along vasculature that correlate to venous congestion and perivascular hemorrhage. The use of SWI or other gradient echo based methods to examine vascular changes may be of interest when examining survivors of Ebola. Additionally, the identification of non-invasive imaging biomarkers of EBOV disease progression could help in development of medical countermeasures.


The purpose of this study was to utilize MRI to assess alterations in brain parenchyma and vasculature during acute Ebola infection in non-human primates.


Global neurologic symptoms have been reported both during acute Ebola infection and in survivors, yet little is understood as to the underlying cause.1-3 Signs of neurologic involvement include headaches, delirium, cognitive impairment, hypoactivity or hyperactivity, and occasionally seizures.1,3 The use of animal models to examine Ebola virus (EBOV) infection is of interest for the investigation of disease pathogenesis, correlation to human disease, and the development of medical countermeasures. To better understand the neurologic features of EBOV that present in patients, a longitudinal neuroimaging study was performed using a non-human primate (NHP) model of Ebola.


Six rhesus macaques were infected with Ebola (Makona variant) with a target dose of 1,000 plaque forming units via the intramuscular route of inoculation. Animals were imaged before infection and prior to euthanasia (7-8 days post exposure). Subjects were immobilized using ketamine and sedated with isoflurane during imaging. All NHPs were imaged on a Philips Achieva 3T MR scanner using a Head SENSE Coil and positioned in a supine fashion. MR Sequences: A magnetization-prepared rapid acquisition gradient echo (MPRAGE) sequence was preformed which had 0.5 mm isotropic resolution (5 min). Other parameters include TR/TE of 9.7/4.7 ms, NSA=2, fat suppression (ProSet), TFE factor = 125, and a FOV 96x96x67 mm3. A T2-weighted image was produced from a multi-echo sequence (TE=80 ms). This sequence had an in-plane resolution of 0.43 x 0.43 mm2, slice thickness of 2 mm, with a total of 33 slices (12 min). Other parameters include TR=3644 ms, TE= (n x 20) ms with 4 echoes, NSA=1, fat suppression (SPIR), TSE factor = 4. Susceptibility weighted images (SWI) were acquired using a 3D, T1-weighted, EPI based sequence. Other parameters include 0.55 mm isotropic resolution, TR/TE of 54/29 ms, NSA=4, and a EPI factor = 15 (4 min). Quantitative image analysis: pre and post-infection MPRAGE images were co-registered, skull stripped and automatically segmented into tissues classes using a modified version of Sienax (FSL software) which allow for assessment of NHP white and gray matter. A paired t-test was used to assess alterations in tissue volumes.

Results and Discussion

Within 7-8 days of infection, all NHPs exhibited alterations in CNS vasculature as observed by SWI (Figure 1). These hypo-intensities were found to represent enlarged vasculature, venous congestion and perivascular hemorrhage, much of which was found within the subarachnoid space (Figure 2A). T2-weighted and FLAIR images confirmed a lack of focal lesions due to ischemia or infarct. Alterations in whole brain volumes (-3.3%, p < 0.0001) due to ventricle enlargement and reduced gray matter volumes (-7.7%, p < 0.002) were observed in all animals between pre and post-infection imaging (Figure 1). While all animals exhibited varying degrees of congestion, edema, and perivascular hemorrhage located in the sub-arachnoid space, edema and signs of extravasation were found in the cerebral cortex in 3 of the six animals (Figure 2B). Although edema and congestion in both the subarachnoid space and cortex can reduce signal intensity in MPRAGE images, the alterations in ventricles indicate dehydration or viral induced hydrocephalus may also contribute to these changes. Viral encephalitis has been suspected in patients identified during the acute phase of Ebola infection4, however future studies using gadolinium-based contrast agents will be required to determine if encephalitis occurs in this model.


Our results indicate that brain parenchyma and vasculature are affected during acute EBOV infection. EBOV in NHPs was found to induce hypo-intensities in the vasculature indicative of venous congestion and perivascular hemorrhage. Alterations in whole brain volumes within 7-8 days of infection were also observed which may be induced by edema, congestion, dehydration or viral related hydrocephalus. The identification of non-invasive imaging biomarkers of EBOV-related neurologic disease progression could aid in understanding neurologic features reported in Ebola patients and help in the development of medical therapeutics.


This work was supported by NIAID Division of Intramural Research and NIAID DCR and was performed under Battelle Memorial Institute contract (No. HHSN272200700016I) with NIAID. Additional support was provided by Leidos Biomedical Research (NCI Contract No. HHSN261200800001E).


[1] Chertow DS, et al. N Engl J Med. 371 (22): 2054-2057 (2014).

[2] Clark DV, et al. Lancet Infect Dis. 15 (8): 905-912 (2015).

[3] Solbrig MV, et al. Neurological Infections and Epidemiology. 2, 5-12 (1997).

[4] Sagui E, et al. Clin Infect Dis. 61 (10): 1627-8 (2015).


Figure 1: Enlargement of the ventricles and the sulci around the cortex are often observed in MPRAGE after infection with EBOV. Alterations in SWI observed throughout the brain also indicate enlarged vessels, venous congestion and possible perivascular hemorrhage. No focal lesions or signs of stroke are observed (by either T2-weighted or FLAIR images) in this acute EBOV model.

Figure 2: A majority of venous congestion, edema and hemorrhage was located within the meninges in the sub-arachnoid space (A), although edema (B: arrows) and perivascular hemorrhage (asterisks) can be found in the cerebral cortex of some animals.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)