Abstract
Both Ebola virus (EBOV) and Reston virus (RESTV) cause disease in nonhuman primates, yet only EBOV causes disease in humans. To investigate differences in viral pathogenicity, humanized mice (hu-NSG-SGM3) were inoculated with EBOV or RESTV. Consistent with differences in disease in human infection, pronounced weight loss and markers of hepatic damage and disease were observed exclusively in EBOV-infected mice. These abnormalities were associated with significantly higher EBOV replication in the liver but not in the spleen, suggesting that in this model, efficiency of viral replication in select tissues early in infection may contribute to differences in viral pathogenicity.
Ebola virus ([EBOV] species Zaire ebolavirus; family Filoviridae) causes the severe, highly lethal EBOV disease in both humans and nonhuman primates (NHPs). The closely related Reston virus ([RESTV] species Reston ebolavirus) also causes severe, lethal disease in NHPs. However, despite several accounts of RESTV infection in humans, no clinical illness has been reported [1–3].
Investigating differences in pathogenicity between RESTV and EBOV in vivo has been challenging due to a lack of suitable animal models. Humanized mice are a new rodent model for studying filoviruses [4]; these include hu-NSG-SGM3, in which EBOV-Makona caused lethal disease in up to two thirds of infected mice [5]. Using hu-NSG-SGM3 mice inoculated with EBOV-Makona or RESTV-Pennsylvania, we assessed clinical signs, clinical chemistry, viral replication, and histopathology to identify differences in pathogenesis early in infection. Humanized mice were susceptible to infection by both viruses, but only EBOV-infected mice developed acute disease. Disease was associated with higher early replication of EBOV in the liver and corresponding elevations in hepatic enzymes, suggesting that differences in tissue-specific viral replication between EBOV and RESTV may contribute to disease development in this model and may also be a factor contributing to differential pathogenicity in humans.
MATERIALS AND METHODS
Biosafety
All work with infectious virus or infected animals was conducted in a biosafety level 4 (BSL-4) laboratory at Rocky Mountain Laboratories (Hamilton, MT) following established BSL-4 standard operating procedures approved by the Institutional Biosafety Committee.
Ethics Statement
All animal experiments were approved by the Institutional Animal Care and Use Committee of the Rocky Mountain Laboratories (animal study protocol 2015-065-E) and reviewed by the Centers for Disease Control and Prevention and were performed following the guidelines of the Association for Assessment and Accreditation of Laboratory Animal Care, International (AAALAC) by certified staff in an AAALAC-approved facility.
Humanized Mice
Two cohorts of SGM3 mice [NOD.Cg-PrkdcscidIl2rgtm1Wjl Tg(CMV-IL3, CSF2, KITLG)1Eav/MloySzJ] (stock no. 013062), engrafted with CD34+ human hematopoietic cells derived from 2 separate donors (Hu-NSGTM-SGM3, stock no. 701362), were obtained from Jackson Laboratories (Sacramento, CA). Mice were group-housed in an isolator-caging system (Innovive Inc., San Diego, CA) with sterile bedding in a climate-controlled laboratory with a 12-hour day/12-hour night cycle; irradiated commercially available mouse chow and water were provided ad libitum. The mice were humanely euthanized with isoflurane vapor at the indicated time points or when clinical illness scores based on weight loss (>20%), piloerection, neurological signs, changes in mentation, ataxia, dehydration, or dyspnea indicated that the animal was in distress or in the terminal stages of disease.
Clinical Chemistry
Whole blood was analyzed on Piccolo Xpress chemistry analyzers using the General Chemistry 13 Panel (Abaxis) within 2 hours of collection.
Virus Inoculation
Mice (n = 57; 12 weeks post-engraftment) were inoculated intramuscularly in the hind limbs with sterile Dulbecco’s modified Eagle’s medium (DMEM) or with 103 FFU of RESTV-Pennsylvania Reston virus/M.fascicularis-tc/USA/1989/Philippines89-Pennsylvania (GenBank NC_004161) or EBOV Makona C07 (Ebola virus/H.sapiens-tc/GIN/2014/Makona, GenBank KP096421; first passage after isolation) diluted in DMEM.
Virus Titrations
Tissue samples were homogenized in 10× weight/volume of DMEM. Serial 10-fold dilutions of the homogenate were used to inoculate confluent monolayers of Vero-E6 cells in triplicate in 48-well plates for 1 hour. The inoculum was then replaced with DMEM containing 2% fetal bovine serum and antibiotics and incubated for 12 days. Tissue culture infectious dose 50% (TCID50) was calculated using the Spearman-Kaerber method by observing cytopathic effects.
Histology and Immunohistochemistry
Tissues were fixed and processed as previously described [5]. Specific immunoreactivity was detected using polyclonal rabbit serum against EBOV VP40 (diluted 1:2000) or RESTV NP (diluted 1:250), followed by a Biogenex biotinylated anti-rabbit antibody (Fremont, CA).
Statistical Analyses
Weight changes were analyzed using a nonparametric Kruskal-Wallis (analysis of variance [ANOVA]) P-value model to compare donor cohorts and experimental groups. Wilcoxon rank-sum tests were used to further analyze weight change differences between RESTV-inoculated and EBOV-inoculated mice. Clinical chemistry was analyzed using one-way ANOVA followed by Tukey’s multiple comparison tests to generate adjusted P values.
RESULTS
Groups of humanized mice, representing 2 donor cohorts in equal numbers, were inoculated intramuscularly with DMEM (mock; n = 17) or with 103 FFU of EBOV (n = 18) or RESTV (n = 18). To investigate replication in primary target organs and clinical chemistry early in infection, groups of 5–6 DMEM-, 6 EBOV-, and 6 RESTV-inoculated animals were euthanized 6, 10, and 14 days postinoculation (DPI). The 3 groups of animals designated for euthanasia at 14 DPI were weighed daily over the course of the study (Figure 1A). End-point survival was not assessed; however, of the EBOV-inoculated mice scheduled for euthanasia 14 DPI, 1 was euthanized 13 DPI and 2 others reached euthanasia criteria 14 DPI due to disease signs, including decreased activity and severe weight loss. Within inoculation groups, no significant differences were observed between mice from different donor cohorts in daily weight loss or end-point weight changes. However, differences in weight loss were observed between inoculation groups; average end-point weight losses were −2.2% (±1.39 standard deviation [SD]) in DMEM-, −8.5% (±2.35 SD) in RESTV-, and −18.8% (±7.13 SD) in EBOV-inoculated animals. The DMEM-inoculated mice exhibited no significant weight loss throughout the study. Both RESTV- and EBOV-inoculated mice had end-point weight loss that was significantly different from the overall mean of DMEM-inoculated mice (Kruskal-Wallis [ANOVA]; P = 1.6 × 10–4 and P = 4.3 × 10–8, respectively). However, weight loss in EBOV-inoculated animals was more severe and significantly different compared with the loss in RESTV-inoculated mice (Wilcoxon rank-sum test; P = 2.0 × 10−10).
Weight loss and clinical chemistry values in humanized mice inoculated with Reston virus (RESTV) or Ebola virus (EBOV). (A) Weight change in humanized mice engrafted with CD34+ cells from donor 1 (DR1; n = 3 per inoculation group) and donor 2 (DR2; n = 3 per inoculation group) inoculated intramuscularly with Dulbecco’s modified Eagle’s medium (DMEM) (mock), or with 103 FFU of RESTV-Pennsylvania or EBOV-Makona at indicated days postinfection (DPI). (B) Clinical chemistry on whole blood samples collected at the indicated times of euthanasia: solid symbol, 6 DPI; open symbol, 10 DPI; and hatched symbol, 14 DPI. Statistical values are based on comparison of all groups to control or RESTV to EBOV at each time point; *, P < .5; **, P < .01. Abbreviations: ALB, albumin; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AMY, amylase; AST, aspartate aminotransferase; GLU, glucose; TP, total protein.
Alterations in clinical chemistry analytes were observed in EBOV-inoculated mice beginning 10 DPI and progressed at 14 DPI; no significant changes in analyte values were seen in RESTV-inoculated mice compared with control mice (Figure 1B). The most striking enzyme increases in EBOV-inoculated mice were in aspartate aminotransferase (AST), alkaline phosphatase (ALP), and alanine aminotransferase (ALT) levels; all of these increases were significant compared with control mice (P < .0001). Additional changes from controls in EBOV-inoculated mice at 10 and 14 DPI include the following: significant decreases in albumin ([ALB] both at P < .0001), total protein ([TP] P = .02 and P = .001), and glucose ([GLU] both at P = .008); decreases in calcium levels; and increases in amylase (both at P = .03). At 10 DPI and/or 14 DPI, significant differences were observed between RESTV- and EBOV-inoculated mice in AST, ALT, ALP, ALB, TP, and GLU levels.
Quantifying virus in spleen and liver tissues showed that both EBOV and RESTV replicated in humanized mice (Figure 2A). Replication levels in the spleen were not significantly different between EBOV and RESTV at any time point examined. However, EBOV levels in the liver at both 6 and 10 DPI were significantly higher than RESTV levels. Despite differential antigen targets and staining characteristics, overall immunohistochemistry on spleen and liver tissues supported virus titration data (Figure 2B); both spleen and liver samples from mice infected with RESTV and EBOV showed antigen positivity at 6–14 DPI. At 6 DPI, RESTV antigen was expressed in low numbers of macrophages within the spleen, with rare expression in hepatocytes and Kupffer cells within the liver. At the same time point, EBOV antigen was seen in scattered macrophages within the spleen, with moderate expression in the hepatocytes, Kupffer cells, and endothelial cells within the liver. The RESTV antigen distribution increased mildly in both tissues 10 DPI through 14 DPI, but it was noticeably lower in liver than EBOV antigen levels at all time points. In both tissues, EBOV antigen distribution increased from numerous at 10 DPI, to almost diffuse positivity at 14 DPI.
![Viral replication and antigen detection in the spleen and liver of humanized mice infected with Reston virus (RESTV) or Ebola virus (EBOV). Virus titration (A) and immunohistochemistry (B) of spleen and liver samples taken at the indicated time points (6, 10, or 14 days postinfection [DPI]) from mice inoculated with 103 FFU of RESTV-Pennsylvania or EBOV-Makona. Magnification (×200) of tissue samples stained with a RESTV-specific anti-NP antibody or an EBOV-specific anti-VP40 antibody. Both EBOV- and RESTV-infected animals expressed antigen positivity in the spleen and liver at 6 DPI; however, EBOV-infected animals showed a marked increase in positivity at 10 and 14 DPI, whereas RESTV-infected animals showed mild increases in positivity. Black arrowheads indicate representative RESTV NP-positive cells in the liver. Abbreviations: NS, not significant; TCID50, tissue culture infectious dose 50%.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/jid/217/1/10.1093_infdis_jix562/5/m_jix56202.jpeg?Expires=1749481173&Signature=VHI-kFyiQ2hc2nk3qiPw7TsO1do~KvmVw6pus8XmFAVWIjc7vMEDT-nM6izSwNyKgy3HgiqHq0R22WhPNnLa~jdoa0L3l4zy8xfJ9zT2D7xG~WA37Q8btF0h~KfGepVDvA2n5er~oNTk768xiLzCLH100N0UFrwxhadbjZTsCDhzUrcv3iJczA~D~VKkKw1XYj32Ry-BPq~3Ydum7Xn16Ft23wNgUdf4osWIZ1JaIPle-dkDjT9vSIC-s9wFiHZu7Nr712TUy33v64K8lJYCNYWl7dEj2zcDHsidQRDz5m9zjMpT0R6LN7lDiLSETCZ9-Cuf2pfOT0XIQ8Lqq7nkUQ__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Viral replication and antigen detection in the spleen and liver of humanized mice infected with Reston virus (RESTV) or Ebola virus (EBOV). Virus titration (A) and immunohistochemistry (B) of spleen and liver samples taken at the indicated time points (6, 10, or 14 days postinfection [DPI]) from mice inoculated with 103 FFU of RESTV-Pennsylvania or EBOV-Makona. Magnification (×200) of tissue samples stained with a RESTV-specific anti-NP antibody or an EBOV-specific anti-VP40 antibody. Both EBOV- and RESTV-infected animals expressed antigen positivity in the spleen and liver at 6 DPI; however, EBOV-infected animals showed a marked increase in positivity at 10 and 14 DPI, whereas RESTV-infected animals showed mild increases in positivity. Black arrowheads indicate representative RESTV NP-positive cells in the liver. Abbreviations: NS, not significant; TCID50, tissue culture infectious dose 50%.
Hematoxylin and eosin staining and histological findings in control humanized mice were similar to previously described [5]. No notable histopathological changes from controls were observed in spleens or livers of RESTV-inoculated mice at any time points investigated. In contrast, at 10 and 14 DPI, EBOV-inoculated mice showed multifocal cellular necrosis in the splenic red pulp with mild acute inflammation (Supplementary Figure S1). However, the most significant pathological changes were observed in livers of EBOV-inoculated animals; at 10 and 14 DPI, moderate numbers of eosinophilic intracytoplasmic inclusion bodies within hepatocytes were observed, and hepatic architecture was disrupted by increasing numbers of infiltrating macrophages and lymphocytes and by multifocal hepatic cell necrosis (Supplementary Figure S2).
DISCUSSION
In NSG-SGM3 humanized mice, severe weight loss and indicators of hepatic damage and disease were observed after EBOV infection, but not RESTV infection, recapitulating the contrast in disease severity resulting from EBOV- versus RESTV-infection in humans. The magnitude of weight loss and clinical chemistry abnormalities in EBOV-infected mice corresponded to higher replication levels of EBOV than RESTV in the liver but not in the spleen. Associated with viral replication were increased AST, ALT, and ALP, and decreased ALB and TP levels in EBOV-infected mice, altogether suggesting that early damage to the liver and abnormal hepatic function may be a major contributor to the severity of disease caused by EBOV infection.
In this study, we focused on early time points to directly compare EBOV and RESTV infection in humanized mice, used as surrogates for early human infection. The mice were not observed past 14 DPI; therefore, it is unknown whether the RESTV-inoculated mice would have later developed more severe disease in this model. However, RESTV-inoculated SGM3 humanized mice (20 weeks post-engraftment) in pilot studies did not develop disease up to 30 DPI, whereas a subset of EBOV-inoculated mice succumbed to infection (Spengler and Prescott, 2016, unpublished data), supporting RESTV as being attenuated in SGM3 humanized mice.
Replication of RESTV is delayed compared with EBOV in Vero-E6 cells [6], but it does not appear to significantly differ in primary human macrophages [7]. Filoviruses are hepatotropic; liver damage is associated with severe disease and poor outcome [8], and rapid normalization of liver enzyme levels is reported in patients who recovered from illness [9]. Yet, RESTV and EBOV replication in hepatic cells is not well characterized. In NHPs, which are susceptible to disease from either EBOV or RESTV, both viruses target the liver [10, 11]. However, comparison of viral levels by tissue type over time have not been reported in NHPs, nor has EBOV or RESTV replication been examined ex vivo in primary human hepatic cells.
Varying immune responses to RESTV and EBOV VP35 and VP24 proteins [12], suppression of the host response [7], and host species-specific T-cell sensitivity [13] may contribute to observed differences in pathogenicity in humans. Reston virus has been shown to be less effective at inhibiting innate immune responses than EBOV or Marburg virus [12], another highly pathogenic filovirus. These differences in immune antagonism may not be universal across organs, because regulation of interferon and cytokine responses varies by tissue, in part due to the abundance of appropriate cytokine receptors [14]. Enhanced immune suppression of RESTV replication in the liver may permit viral clearance without irreparable damage to hepatic tissue, preventing severe clinical disease. In addition, other tissue-specific, viral-mediated virulence factors may exist; for example, glycoprotein was shown to modulate hepatic inflammation and necrosis in RESTV versus EBOV infection in vivo, independent of virus replication [6]. Thus, outcome is likely a combination of reduced replication efficiency and differences in virulence factors in select tissues such as the liver.
Humanized mouse models are a complex interplay of murine and human cells. Differences in susceptibility to viral infection in cells of different origins may affect outcome. The specific virus-cell interactions in humanized mice that confer susceptibility to EBOV disease are currently not known. Future work detailing these interactions will aid in elucidating how specific cell populations, circulating and present in organs, contribute to outcome. Additional variables to consider in humanized mouse models include engraftment levels; however, in this model, we have not observed an association between differential engraftment levels and outcome [4, 5].
CONCLUSIONS
The factors contributing to the widely disparate outcome of EBOV and RESTV infections in humans are unknown. Our investigations into early events postinfection in humanized mice suggest that enhanced replication efficiency in organs associated with severe disease is an important indicator of pathogenicity in humans. These studies form a basis for more emphasis on examining cell- and organ-specific responses to filovirus infection. Ultimately, characterizing filovirus species and variant replication in primary target organs and identifying underlying molecular determinants of tissue-specific replication efficiency and immune antagonism may aid in further defining filovirus pathogenicity and provide additional targets for medical countermeasure design.
Supplementary Data
Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Acknowledgments. We gratefully acknowledge the assistance of Tim Flietstra on statistical analyses, Tatyana Klimova in editing the manuscript, and Anita Mora on the figures. We also thank Jonathan Towner for helpful discussions and Tina Thomas, Rebecca Rosenke, and Dan Long for preparing and staining tissues for histopathologic studies.
Disclaimer. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention or the National Institute of Allergy and Infectious Diseases.
Financial support. This work was partially supported by an appointment to the Research Participation Program at the Centers for Disease Control and Prevention administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the US Department of Energy and Centers for Disease Control and Prevention and by the National Institutes of Health Loan Repayment Award (to J. R. S.). This work was also supported by the Division of Intramural Research of the National Institutes of Health, National Institute of Allergy and Infectious Diseases.
Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.
References
