Magnetic Resonance Imaging for Monitoring of Hepatic Disease Induced by Ebola Virus: a Nonhuman Primate Proof-of-Concept Study

ABSTRACT Severe liver impairment is a well-known hallmark of Ebola virus disease (EVD). However, the role of hepatic involvement in EVD progression is understudied. Medical imaging in established animal models of EVD (e.g., nonhuman primates [NHPs]) can be a strong complement to traditional assays to better investigate this pathophysiological process in vivo and noninvasively. In this proof-of-concept study, we used longitudinal multiparametric magnetic resonance imaging (MRI) to characterize liver morphology and function in nine rhesus monkeys after exposure to Ebola virus (EBOV). Starting 5 days postexposure, MRI assessments of liver appearance, morphology, and size were consistently compatible with the presence of hepatic edema, inflammation, and congestion, leading to significant hepatomegaly at necropsy. MRI performed after injection of a hepatobiliary contrast agent demonstrated decreased liver signal on the day of euthanasia, suggesting progressive hepatocellular dysfunction and hepatic secretory impairment associated with EBOV infection. Importantly, MRI-assessed deterioration of biliary function was acute and progressed faster than changes in serum bilirubin concentrations. These findings suggest that longitudinal quantitative in vivo imaging may be a useful addition to standard biological assays to gain additional knowledge about organ pathophysiology in animal models of EVD. IMPORTANCE Severe liver impairment is a well-known hallmark of Ebola virus disease (EVD), but the contribution of hepatic pathophysiology to EVD progression is not fully understood. Noninvasive medical imaging of liver structure and function in well-established animal models of disease may shed light on this important aspect of EVD. In this proof-of-concept study, we used longitudinal magnetic resonance imaging (MRI) to characterize liver abnormalities and dysfunction in rhesus monkeys exposed to Ebola virus. The results indicate that in vivo MRI may be used as a noninvasive readout of organ pathophysiology in EVD and may be used in future animal studies to further characterize organ-specific damage of this condition, in addition to standard biological assays.

each time point, the imaging protocol consisted of the combination of (i) precontrast T2weighted (T2W) imaging, (ii) precontrast and postcontrast T1-weighted (T1W) imaging, and (iii) precontrast mapping of T2* relaxation time to evaluate liver appearance, volume, and hepatobiliary function in vivo over time and in comparison to histopathology (Fig. 1B shows the imaging protocol and endpoints). One monkey could not be imaged at day 2, whereas precontrast T1W, T2W, and T2*W MRI on the day of euthanasia could not be performed on another monkey. Imaging sessions were tolerated well by all animals.
Qualitative MRI findings. T2W imaging was used to qualitatively assess the overall appearance of the liver parenchyma, as well as hepatomegaly, indicative of progressive hepatic impairment. Hepatomegaly was first observed on day 5 and was present in all monkeys on the days of euthanasia, especially enlargement, thickening, and extension of the left hepatic lobes ( Fig. 2A). Progressive diffuse increases in liver signals also started on day 5 in 83.33% of the monkeys and were measured for all monkeys by the days of euthanasia (Fig. 2B). Gallbladders, biliary tracts, and vessels were bright, as expected, on T2W images.  Preexposure three-dimensional (3D) T1W high-resolution isotropic volume excitation (THRIVE) images showed within-range liver volumes (Fig. 3A) and normal uniform parenchymas with intrahepatic vessels and gallbladders (Fig. 4A), similar to images from healthy humans (19,25) and other NHPs (26). After EBOV exposure, postcontrast THRIVE images demonstrated progressive hepatomegaly (Fig. 3A) and extensive hepatic abnormalities, including decreased parenchymal signal intensity, which were first observed on day 5 and became progressively more pronounced on the day of euthanasia (Fig. 4A). THRIVE images 40 min postcontrast showed rapidly decreasing hepatic signal intensity and decreased biliary secretion of gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid (Gd-EOB-DTPA) as disease progressed (Fig. 4A). On the day of euthanasia, there was conspicuous loss of parenchymal signal intensity in postcontrast THRIVE images relative to intrahepatic vessels and complete absence of contrast agent biliary excretion, indicating decreased hepatocyte uptake of Gd-EOB-DTPA possibly related to hepatocyte damage.
In this study, T2* liver maps showed an overall increased signal at day 5, but values returned to baseline levels by the day of euthanasia (Fig. 5A).
Quantitative MRI measurements. Hepatomegaly observed qualitatively by T2W imaging ( Fig. 2A) was confirmed by quantitative segmentation of liver volumes on T1W 3D images (THRIVE) acquired 30 to 35 min after contrast agent injection (Fig. 3A) (Fig. 4B). Liver enhancement curves (relative to precontrast images) were almost overlapping from baseline to day 2. At day 5, curves had earlier and lower peaks with signal plateaus until the end of acquisition. On the days of euthanasia, curves also showed earlier and lower peaks and a slow decrease in signal enhancement through the end of the acquisition (Fig. 4B).  Hematology, serum chemistry, and virologic assessments. Clinical observations on day 2 included mild bilateral axillary lymphadenopathy in three of nine monkeys. By day 5, the remaining six monkeys had loss of appetite, fever (average 40.6°C), and bilateral axillary lymphadenopathy. Two monkeys met endpoint criteria at day 6 and were euthanized. The remaining monkey was euthanized on day 9, per the study design (Table S2). Blood virus titers were negligible at day 2 but had significantly increased at day 5 and on the days of euthanasia (Fig. 6). Liver plaque assays performed on tissues harvested at necropsy indicated absence of replicating virus in animals euthanized at day 2, whereas live virus was detected in the livers of animals euthanized at day 5 and at the end of study (Fig. S1).
Hematologic findings included overall increasing white blood cell (WBC) counts from day 2, with day 5 values being significantly different from day 0 values (Fig. 6, P = 0.025). At day 5 and on days of euthanasia, significant neutrophilia was observed ( Fig. 6, P = 0.000 and P = 0.005, respectively) accompanied by lymphopenia ( Fig. 6, P = 0.005 at day 5). No changes in absolute numbers of monocytes were noted during the course of the study. Steady declines in red blood cells (RBCs) were also observed from day 0, which resulted in significant, albeit mild, normocytic anemia ( Fig. 6) (P = 0.002) by the days of euthanasia. Accordingly, hemoglobin concentrations and hematocrit values also steadily decreased FIG 6 Changes in serum chemistries and viremia during the course of the experiment. HGB, hemoglobin; HCT, hematocrit; PLT, platelet. D0, day 0 (exposure); D2, day 2; D5, day 5; DE, day of euthanasia. All values are medians and 25% to 75% confidence intervals. Corrected P values of ,0.05 were considered to indicate statistical significance compared to day 0 (*). throughout the study and were significantly lower than those on day 0 by the days of euthanasia ( Fig. 6, P = 0.004 for both). Platelet numbers sharply decreased after day 2, being significantly lower than those at day 0 at both day 5 ( Fig. 6, P = 0.008) and the days of euthanasia ( Fig. 6, P = 0.003).
In terms of serum chemistry, liver enzyme activities (ALT, AST, and gamma-glutamyl transferase [GGT]) were mostly stable until day 5 and then significantly increased during the study (Fig. 6). Total bilirubin concentrations also increased by the days of euthanasia (P = 0.029). Albumin concentrations started declining at day 5 ( Fig. 6) and were much lower than those on day 0 on the days of euthanasia ( Fig. 6, P = 0.045). Among other metabolic indices, the ratios of blood urea nitrogen (BUN) to creatinine concentration significantly declined after exposure ( Fig. 6, P = 0.002 at day 5 and P = 0.033 on the days of euthanasia). A selected list of assessed blood biomarkers, with medians [25% to 75%], are reported in Table 1.
Histopathology and immunohistochemistry. Histopathological evaluation of liver tissues stained with hematoxylin and eosin revealed widespread increased hepatocellular degeneration by day 5 that progressed through the days of euthanasia, with occasional intracytoplasmic eosinophilic viral inclusion bodies and areas of hepatocellular necrosis (Fig. 7). These findings corroborate contrast-enhanced MRI findings, indicating progressive hepatobiliary impairment. Immunohistochemistry targeting EBOV matrix protein (VP40) antigen was negative on day 2 and then increased in intensity as the disease progressed (Fig. 7), supporting viral dissemination within the liver.

DISCUSSION
In this proof-of-concept study, we demonstrated the value of longitudinal quantitative MRI to noninvasively assess liver size, composition, and function in EBOV-exposed rhesus monkeys. Clinical signs after EBOV exposure and the pattern of the serum chemistries and complete blood count profiles from the nine monkeys included in our study are consistent with the spectrum of clinical disease reported in several NHP EBOV studies (13,14,27). The animals in this study met the criteria for euthanasia 6 to 9 days postexposure, i.e., somewhat earlier than previous studies (24) but within the overall range of when these criteria were reached in similar studies (28). In addition, qualitative and quantitative assessments of liver appearance by T2W images and a series of precontrast and postcontrast T1W images indicated progressive alterations in MR signal during the course of our experiment. These imaging features are consistent with reported histological features related to pathological changes in the liver (e.g., hepatocellular degeneration, edema, fibrin deposition, and congestion). Similar findings have been reported in clinical data on liver inflammation accompanying acute hepatitis (18,29), although specific periportal signal changes (which have been detected in humans) were not observed in this monkey study.
In concordance with qualitative imaging findings, quantitative volumetric liver measurements from postcontrast T1-weighted images indicated significant hepatomegaly starting at day 5. Findings by visual inspection of contrast-enhanced scans were more conspicuous, particularly on imaging performed prior to the day of euthanasia. Decreased enhancement during the arterial and early parenchymal phases was interpreted as suggestive of hepatic necrosis or decreased hepatocellular uptake. During the late parenchymal and biliary phases, decreased contrast enhancement was instead thought to indicate decreased hepatocellular uptake and liver secretory function by reduction (early in disease) or absence (on the day of euthanasia) of contrast agent in the biliary tree (23). Quantification of contrast agent uptake showed decreasing relative liver enhancement (30) during disease progression (albeit without reaching statistical significance). As Gd-EOB-DTPA is taken up into hepatocytes through an energy-dependent pathway, the diminished liver uptake rate of Gd-EOB-DTPA (indicated by a decreased RLE) may be attributed to the progressive hepatocellular dysfunction associated with EBOV infection. Similarly, since Gd-EOB-DTPA is excreted into the bile by an ATP-dependent glutathione S-transferase (31, 32), the observed decreased and/or delayed Gd-EOB-DTPA enhancement kinetics in the liver parenchymas and decreased lack of visualization of the biliary trees suggest hepatic secretory impairment. The deterioration in biliary function assessed by Gd-EOB-DTPA-enhanced MRI was acute and progressed more quickly than changes in serum bilirubin concentrations (typically increasing only during late acute EBOV infection) (12,33). Our data are corroborated by previous studies that showed Gd-EOB-DTPA liver uptake curves in patients with severe, acute hepatitis and different degrees of inflammation or other liver diseases (19). Gd-EOB-DTPA liver uptake in EBOV-exposed NHPs on the days of euthanasia was remarkably similar to that in these patients, for whom hepatocellular degeneration and necrosis were demonstrated by histopathology.
Finally, elevation of T2* relaxation times observed in EBOV-exposed NHPs at day 5 is indicative of fluid accumulation in the livers (34). In MRI, T2* relaxation refers to the decay of the transverse magnetization due to spin-spin interactions and local magnetic field inhomogeneities. T2* relaxation time gets shorter, causing a decrease in MR signal, as the molecular composition of the examined tissue becomes more heterogeneous or as local "disturbances" (e.g., the presence of iron) are introduced. Shorter T2* relaxation times have been associated with the increased presence of calcium, iron, and hemorrhage (including liver iron overload), whereas increases in T2* relaxation have been associated with increased tissue water content. Histopathology confirmed some degree of hepatic congestion in all monkeys, thereby corroborating these results.
Despite these encouraging results, our study has several limitations. First, only nine NHPs were included in our experiment, and precontrast MRI could not be acquired on the day of euthanasia from one monkey that survived to day 9 due to its clinical condition. Additionally, qualitative interpretation of precontrast images on the days of euthanasia may have been confounded by residual Gd-EOB-DTPA uptake in the gallbladders. Gd-EOB-DTPA was also retained in the kidneys, thereby precluding the use of this organ as an internal control. However, our results are in line with data from clinical studies examining a range of liver diseases (17)(18)(19), warranting more in-depth investigations on the potential of MRI in characterizing liver involvement in NHP models of EVD. Our results present an apparent discordance between the time course of MRI parameters and the one of serum biomarkers of liver function, which peaked at a later date, i.e., day 9. The MRI methods described here examine organ function at a local level, and they may be complementary to analysis of blood biomarkers, increases of which are reflective of organ damage at a systemic level. Therefore, this apparent discrepancy may indicate a potential strength of imaging to detect pathogenetic changes in the liver at an organ level earlier than blood biomarker increases and may present an opportunity for earlier (and different) investigations of organ pathophysiology during the disease course.
In summary, we have demonstrated that liver involvement in EBOV-exposed NHPs can be assessed noninvasively and longitudinally using non-contrast-enhanced and contrastenhanced MRI. We have also shown that qualitative and quantitative MRI results are in line with clinical and laboratory data indicating severe progressive liver dysfunction and histopathological findings of hepatocellular degeneration and necrosis. Importantly, the innovative use of hepatocyte-specific contrast material uniquely shows severe impairment of hepatocyte function and biliary excretion in this animal model of EVD.

MATERIALS AND METHODS
Animals. This study was conducted in nine male rhesus monkeys (Macaca mulatta (Zimmermann, 1780)) of ages 2 years 3 months to 3 years 8 months and weights of 3.00 to 3.50 kg at the start of the study (see Table S1 in the supplemental material). All monkeys were rhesus of Indian origin, sourced from an NIHapproved vendor colony within the United States. These monkeys were deemed healthy by veterinary staff at the start of the study. The monkeys were also screened for prior exposure to EBOV, simian foamy viruses, and simian lentiviruses prior to being assigned to this study.
Experimental design. This study was performed at a U.S. BSL-4 facility (35). The monkeys were each exposed to 1,000 PFU of EBOV variant Makona isolate C05 (Ebola virus/H.sapiens-tc/GIN/2014/Makona-C05) in 1-mL volumes injected into the left triceps via the intramuscular route, and inoculation doses were confirmed by plaque assay, as described previously (36). The monkeys were serially imaged before exposure (baseline; n = 9) and after exposure on day 2 (n = 8), day 5 (n = 6), and the days of euthanasia (n = 3). All monkeys were evaluated at least twice daily and then three times daily at the onset of clinical signs (24) (Fig. 1A shows an overview of the study design). The days of euthanasia were determined as the time point at which monkeys reached euthanasia criteria, based on a standardized scoring system (24), or at day 9 postexposure, as predetermined by study design (Fig. 1A). Based on these criteria, two monkeys were euthanized at day 6 and one was euthanized at day 9. At each postexposure imaging time point, three monkeys were euthanized for liver histopathology per the study design (Fig. 1A).
MRI acquisition. The monkeys were imaged on an Achieva 3.0T MR system (Philips Healthcare, Cleveland, OH, USA) using an 8-channel pediatric SENSE torso coil (Philips Healthcare, Cleveland, OH, USA). Anesthesia was induced with ketamine (15 mg/kg, intramuscular) and maintained with a ventilator using isoflurane (2 to 2.5%) inhalation. Monkeys were monitored during anesthesia.
After scout scans to localize the liver, the following evaluations were completed using a combination of pre-and postcontrast T1W and T2W MRI, as well as T2* relaxation mapping (Fig. 1B).
(i) Assessment for the presence of focal or diffuse liver abnormalities and gross assessment of liver size. Assessment for the presence of focal or diffuse liver abnormalities and gross assessment of liver size were performed using a 2D coronal multislice T2W turbo spin-echo (TSE) single-shot sequence, acquired in suspended respiration (SR). Sequence parameters were as follows: repetition time (TR), 1,375.9 ms; echo time (TE), 80.0 ms; excitation flip angle, 90°; refocusing flip angles, 120°; echo train length (ETL), 65; bandwidth, 513 Hz/pixel; slice thickness, 4.0 mm; slice gap, 1 mm; number of slices, 18; in-plane resolution, 0.8 by 0.8 mm 2 ; field of view (FOV), 140 by 120 mm 2 ; scan time, 24 s/SR.
MRI data analyses. Images were reviewed by a board-certified radiologist who is experienced in interpreting images obtained from rhesus monkeys. The radiologist was blinded to study day and the grouping of animals. To prevent bias, analyses were performed after all animals had been necropsied. Liver MRI data were analyzed to extract information about changes in (i) overall appearance, size, and presence/absence of focal and diffuse lesions, (ii) volume, and (iii) hepatobiliary function by the combination of precontrast T2W imaging, pre-or postcontrast T1W imaging (THRIVE), and T2* mapping. Image analysis for each endpoint was performed as follows.
(i) Coronal T2W images. Coronal T2W images were qualitatively evaluated by a radiologist to identify focal or diffuse abnormalities. Radiologist reports were parsed to identify acquisitions noted to have "normal liver appearance and no focal lesions" as opposed to acquisitions described to show "mild to moderate increase in liver MR signal." The ratios of the number of abnormal acquisitions at each imaging time point to the total number of acquisitions were recorded. Acquisitions were also used to qualitatively assess changes in liver size during the study.
(ii) Volumetric measurements of the liver. Volumetric measurements of the liver were performed on postcontrast 3D THRIVE images (30 to 35 min after injection). To minimize measurement variability, a 3D region growing liver segmentation (38,39) algorithm was used by a single observer to quantify liver volumes. Liver volumes over time and percent change in liver volume from baseline were recorded.
(iii) Hepatobiliary function. Hepatobiliary function was assessed using 3D THRIVE images acquired before and after Gd-EOB-DTPA injection. Gd-EOB-DTPA is a hepatobiliary contrast agent that is taken up by hepatocytes (up to 50% of injected dose) and excreted in the bile by organic anion carriers that transport bilirubin (19,25,40,41). Gd-EOB-DTPA hepatic contrast enhancement is lower in the presence of hepatic impairment and/or increased blood bilirubin levels (25,41,42). Before analysis, MR images were uniformly scaled by using vendor-specific image scaling factors. Images acquired at different time points were then consolidated into a single imaging series in MIM version 6.9 (MIM Software, Cleveland, OH, USA). A volume of interest (VOI) (43) was placed in the liver parenchyma on THRIVE images before and after contrast agent injection. VOIs were selected to cover the same anatomical area at all time points, encompassing a homogeneous region free of visible blood vessels. The REL of VOIs (30), a well-validated measure of Gd-EOB-DTPA uptake in the liver, was calculated using the following formula: SI liver enh HBP 2 SI liver unenh SI liver unenh Â 100 where RLE is relative liver enhancement, enh is enhanced, unenh is unenhanced, and HBP is hepatobiliary phase, taken to be 20 min after injection (when the contrast between liver parenchyma, vasculature, and biliary tree was optimal). Time-dependent enhancement curves were also generated using this same equation, with SI liver enh being the signal at each time point after contrast agent injection. (iv) Liver parenchymas were evaluated using T2* relaxation mapping before injection of Gd-EOB-DTPA. For T2* estimation, a 12-mm diameter region of interest (ROI) was carefully placed over the liver parenchymas close to the gallbladders to avoid bile ducts and blood vessels (44). The ROI was defined on the first echo and translated onto the following echoes. For each monkey, the ROI locations were kept consistent among all time points. An in-house extension for MIM v.6.7 (MIM Software, Cleveland, OH, USA) was developed to create and postprocess T2* maps and extract ROI T2* values. Hematology, serum chemistry, and assessment of viremia. Blood work was performed at each imaging time point. Hematological assessments were obtained using whole blood (XS100i; Sysmex America, Lincolnshire, IL, USA). Serum chemistries were measured with Piccolo General Chemistry 13 panels (Abaxis, Union City, CA, USA). A list of assessed blood variables can be found in Table 1. Viremia was also assessed at each imaging time point using real-time reverse transcription PCR (RT-qPCR) (45). Liver samples were collected at necropsy, weighed, and snap-frozen in a dry ice-ethanol bath. Ten percent (wt/vol) homogenates were made in phosphate-buffered saline, samples were clarified, and the supernatants were serially 10-fold diluted for titration. Titrations were performed on Vero E6 cells, overlaid with methylcellulose, incubated for 7 days, fixed and stained with a crystal violet/formalin solution for 1 h at room temperature, and enumerated.
Necropsy, tissue processing, pathology, and immunohistochemistry. During necropsies, liver tissues were collected for histopathology and immunohistochemistry. Tissues were fixed, paraffin embedded, and sectioned at a thickness of 4 to 6 mm. The sections were mounted on glass slides, stained with hematoxylin and eosin, and coverslips were added (46). Immunohistochemistry was performed using anti-EBOV VP40 antibody (IBT Bioservices, Rockville, MD, USA), diluted 1:1,500, as a primary antibody for detection of antigen (46). Tissues from an unexposed control monkey were used to validate all immunohistochemistry procedures.
Statistical analyses. All continuous outcome measurements were compared across different experiment days by using a Kruskal-Wallis test with multiple comparisons (mean rank of each time point with respect to baseline or exposure day). Dunn's correction was applied to P values to account for multiple comparisons. Corrected P values of ,0.05 were considered to indicate statistical significance. All statistical analyses were performed using GraphPad Prism 9.3.1 for Windows (GraphPad Software, San Diego, CA, USA). One monkey could not be imaged at day 2, whereas precontrast T1W, T2W, and T2*W MRI on the day of euthanasia could not be performed on another monkey.
All values in the text are medians, with first quartile and third quartile values within brackets.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 0.3 MB.