Long-term Sudan Virus Ebola Survivors Maintain Multiple Antiviral Defense Mechanisms

Abstract

Background

The critical issues of sustained memory immunity following ebolavirus disease among long-term survivors are still unclear.

Methods

Here, we examine virus-specific immune and inflammatory responses following in vitro challengd in 12 Sudan virus (SUDV) long-term survivors from Uganda’s 2000–2001 Gulu outbreak, 15 years after recovery. Total RNA from isolated SUDV-stimulated and unstimulated peripheral blood mononuclear cells was extracted and analyzed. Matched serum samples were also collected to determine SUDV IgG levels and functionality.

Results

We detected persistent humoral (58%, 7 of 12) and cellular (33%, 4 of 12) immune responses in SUDV long-term survivors and identified critical molecular mechanisms of innate and adaptive immunity. Gene expression in immune pathways, the interferon signaling system, antiviral defense response, and activation and regulation of T- and B-cell responses were observed. SUDV long-term survivors also maintained robust virus-specific IgG antibodies capable of polyfunctional responses, including neutralizing and innate Fc effector functions.

Conclusions

Data integration identified significant correlations among humoral and cellular immune responses and pinpointed a specific innate and adaptive gene expression signature associated with long-lasting immunity. This could help identify natural and vaccine correlates of protection against ebolavirus disease.

Ebolaviruses belong to the Filoviridae family of the Mononegavirales, a large order of enveloped viruses with nonsegmented, negative-strand RNA genomes [1]. In humans, ebolaviruses cause hemorrhagic fever and ebolavirus disease (EVD), with high morbidity and mortality [2]. Of the 5 members of the genus, the Zaire ebolavirus (EBOV) and Sudan ebolavirus (SUDV) species pose the greatest threat [3].

Profiles of immunity developed in filovirus survivors have recently begun to shed light on previously understudied immune responses [4–7]. Important immune mediators of survival during acute and early convalescent stages have also been determined [5, 8, 9]. Human studies have shown that recovery largely depends on the rapid and efficient development of innate and adaptive immune responses [5, 10]. While humoral immune responses could persist decades postinfection [6, 11–14], little is known about the persistence of cell-mediated immunity [15, 16].

Recent molecular studies have provided insight into the global transcriptional changes during viral infection in humans [17, 18], including significant changes in gene expression [19, 20]. Variations may discriminate between fatal and nonfatal EVD cases [7, 21–23]. It has been suggested that survival following infection correlates with the ability of the host to mount an early and robust innate immune response, such as the interferon signaling pathways critical in the suppression of viral pathogenesis [24–28]. Although these studies have significantly advanced our understanding of the transcriptomic signature of the cellular immune response during acute infection [7, 21, 22, 29], the molecular mechanisms underlying a successful long-term cellular response remain poorly understood.

METHODS

Participants

We enrolled a subgroup of post-SUDV SN (survivors) from the Gulu cohort of the SUDV outbreak of 2000–2001 in the Gulu district, Uganda [30]. Eligible survivors with previously laboratory-confirmed EVD subsequently declared virus free were recruited by the Uganda Virus Research Institute (UVRI) in the Gulu district of Uganda in July 2017. All patients provided individual study-specific written informed consent covering sampling, storage, and use of biological samples. Additionally, each subject in this study completed a personal health questionnaire. Participants were all adults without familial relations. Twelve individual survivors and 3 healthy local community members who were not infected participated in this study, the latter serving as baseline controls. Whole blood, peripheral blood mononuclear cells (PBMCs), and serum samples were collected and stored on site. The Helsinki Committees approved the study protocols of the UVRI in Entebbe, Uganda (reference number GC/127/13/01/15), Soroka Hospital, Beer Sheva, Israel (protocol number 0263-13-SOR), and the Ugandan National Council for Science and Technology (registration number HS1332).

Characterization of Cellular Immune Response

To monitor virus-specific cellular immune responses, we used irradiated SUDV complete virus (Gulu isolate) [13]. Cellular immunity was determined by stimulating PBMC with SUDV complete irradiated virus and assessing the cell’s transcriptomic signature and soluble analytes from matched supernatant cultures. PBMC stimulation assays were performed as described [11]. We collected 50 mL of whole blood from survivors and noninfected controls into CPT vacutainers (BD Biosciences) and isolated PBMCs according to the manufacturer’s protocol. Total cell yields were split between 2 culture conditions in Roswell Park Memorial Institute medium (RPMI) plus 5% fetal bovine serum: no stimulation (stimulation control), and 10 μg of irradiated SUDV. Culture volumes were 1 mL. We performed stimulation for 6 and 18 hours to identify early and delayed responses.

Unstimulated 6- and 18-hour cultured PBMCs were also prepared to serve as nonspecific stimulation background controls. Unstimulated, uncultured PBMCs served as baseline controls. Stimulations were done at 37°C in an atmosphere of 5% CO₂. PBMC-stimulated and unstimulated cultures were pelleted after incubation, and the resultant supernatants were collected.

RNA Isolation and Sequencing

Total RNA was isolated from stimulated and unstimulated PBMCs using the miRNeasy Midi Kit (Qiagen) according to the manufacturer’s protocol and then quantified using a NanoDrop 2000c Spectrophotometer (ThermoFisher Scientific). Gene expression was analyzed using 2 NanoString nCounter gene expression code sets targeting 594 general immunology genes (nCounter Immunology Panel) and 255 inflammation genes (nCounter Inflammation Panel). Probe set-target RNA hybridization reactions and normalization were performed according to the manufacturer’s protocol. Normalized counts from stimulated PBMCs were analyzed against the mean baseline (unstimulated, uncultured resting cells) and mean background (unstimulated cultured cells) readout of PBMCs of both survivors and healthy controls. Differentially expressed genes were defined as those having a > 2-fold significant change in expression. The Gorilla tool subjected data to functional enrichment analysis with gene ontology (GO) terms [31]. Significant functional enrichment terms were defined as those with a false discovery rate corrected P value ≤.05.

Quantification of PBMCs Soluble Mediators

The levels of 9 soluble mediators were analyzed as previously described [11], according to the manufacturer’s instructions, in matched culture supernatant samples from 6 and 18 hours of SUDV antigen-stimulated and unstimulated PBMCs by Q-Plex technology (Quansys Biosciences). The assay included cytokine and chemokines: interleukin 2 (IL-2), IL-6, IL-8, IL-10, interferon-γ (IFN-γ), tumor necrosis factor-α (TNF-α), monocyte chemoattractant protein 1 (MCP-1), MCP-2, and inducible protein-10 (IP-10).

Serum SUDV Antibody Analysis

The levels of circulating anti-SUDV antibodies were defined by chemiluminescence enzyme-linked immunosorbent assay (ELISA), as previously described [32]. ELISA cutoff values for immunoglobulin G (IgG)-positive immunoreactivity were set as mean >2 × SD above negative control sera.

Virus-specific IgG functionality was measured by virus neutralization in a plaque reduction neutralization test (PRNT) [13] and by antibody Fc-dependent reporter assay. Neutralization was measured using 1:10 dilutions of sera. The positive neutralization cutoff value was determined as PRNT₈₀ above 1:10 dilution by a control set of negative sera. Virus-specific antibody binding to the FcγRIII receptor was done using antibody-dependent cellular cytotoxicity (ADCC) Reporter Bioassay, Core Kit (G7018; Promega) according to manufacture instructions with minor modification. Briefly, 96-well cell culture plates were precoated with SUDV irradiated antigen at 2.5 µg/mL concentration diluted in 1 × phosphate-buffered saline and incubated overnight at 4°C. Following incubation, plates were covered by blocking solution with 1% bovine serum albumin for 1 hour. Next, sera samples from survivors and controls were diluted at 1:200 in ADCC assay buffer and incubated for 2 hours. Following incubation, plates were supplemented with ADCC bioassay effector cells 70×10³ cells/well incubated in RPMI 10% (v/v) fetal calf serum medium for 16–18 hours. Finally, 75 µL of Bio-Glo luciferase assay reagent was added to each well and incubated at room temperature for 20 minutes. An Infinite F200 pro-ELISA reader (Tecan) was then used to acquired relative light units results and converted them to pg/µL concentrations.

Statistical Analysis

Statistical analyses were performed using GraphPad Prism software, version 9. Differences in immune response between study groups were assessed by analysis of variants (ANOVA) and Wilcoxon rank-sum test; P values represent 2-sided P values, and P values <.05 were considered statistically significant following correction for type I error. Correlation analyses were assessed by the Spearman correlation matrix using the R package. Principal component analysis (PCA) plots were generated using R edge packages in GraphPad Prism software and the ClustVis web tool [33].

RESULTS

Population Cohort and Study Design

We collected blood samples, 15 years postinfection, from 12 healthy confirmed SUDV SN who lived in or around the Gulu district of Uganda during the initial outbreak in 2000–2001 [30]. Five of the 12 survivors had previously shown antigen-specific positive cellular and humoral responses, and 4 had not [11]. No previous data were available for the remaining 3 survivors (Figure 1A). Three additional samples were collected from uninfected healthy donors who were local community members and used as controls. Clinical and demographic data of the study cohort are summarized in Figure 1B.

Cellular Immunity of Long-term SUDV SN

Early Gene Expression Responses in SUDV SN and Controls Are Primarily Associated With Cell Trafficking and Inflammatory Responses

We first used the gene expression data (n = 699) to evaluate resting (uncultured cells) and cultured unstimulated PBMC gene transcript baseline responses in SUDV SN and controls. We found no significant differences in the gene expression responses of unstimulated resting and cultured PBMCs from SUDV SN and controls (Supplementary Figure 1). Next, we assessed mean changes in gene expression in SUDV SN regardless of their previous immune state. At 6 hours poststimulation, 98 genes were scored as differentially expressed genes (DEGs), 88 upregulated, and 10 downregulated (Figure 2A). Of these, 18 upregulated and 7 downregulated genes were specific to the SUDV SN (Figure 2B and 2C), and 23 upregulated genes were common to SUDV SN and controls (Supplementary Figure 2). Upregulated DEGs detected in both SUDV SN and controls had primary roles in immune cell trafficking and nonspecific inflammatory responses, while DEGs essential for defense response to a virus, T-cell chemotaxis, activation, and regulation were identified primarily in SUDV SN.

GO terms of DEG observed in both SUDV SN and controls identified a statistically significant association with the following biological processes: regulation of IL-1–mediated signaling pathway, monocyte and lymphocyte chemotaxis, and positive regulation of natural killer cell chemotaxis (Supplementary Figure 2). SN-only upregulated DEGs were enriched for genes of antiviral defense responses, including negative regulation of viral genome replication and negative regulation of viral life cycle (Figure 2D). Downregulated DEGs in SUDV SN only were enriched with the biological process production of molecular mediator involved in inflammatory response.

Late Gene Expression Responses in SUDV SN Show Increased Expression of Antiviral Defense Responses and Adaptive Immunity Mediators

The same expression analyses performed 6 hours after stimulation were repeated 18 hours poststimulation. The results (Figures 3 and Supplementary Figure 3) followed the same pattern observed at 6 hours poststimulation. Compared to the 6-hour data, higher expression levels and an increase in the number of DEGs associated with antiviral defense responses and adaptive immunity were detected in SUDV SN. GO term analyses showed the same biological process enrichment as those seen at 6 hours poststimulation; other processes necessary for regulation and activation of innate and adaptive immunity, that is, response to IFN-α and IFN-γ–mediated signaling pathway, were identified only in SUDV SN 18 hours poststimulation (Figure 3D).

Comparison of the DEG in the early versus late poststimulation sets (Supplementary Figure 4) found that the DEG 6-hour poststimulation set was mainly associated with positive regulation of proinflammatory and acute-phase responses. The DEG 18-hour poststimulation set was related to antiviral defense responses. The relative transcript levels of the shared upregulated DEG at 6 and 18 hours poststimulation showed a significant positive correlation (r = 0.8758 and P < .001).

Different Gene Expression Signatures Associated With Antiviral, Innate, and Adaptive Immune Responses Distinguish Subpopulations Within SUDV SN

To identify variation in survivor responses, we compared the total expression of genes tagged as DEGs (n = 375) in individual SUDV SN (Supplementary Figure 5). Expression levels were heterogeneous, and PCA and cluster analysis did not identify subgroups within SUDV SN 6 hours poststimulation. However, at 18 hours poststimulation, 2 different gene expression signatures were identified among the SUDV SN (Supplementary Figure 5D; labeled clusters 1 and 2).

We then examined the gene expression profiles of these 2 SUDV SN clusters to identify specific molecular pathways whose expression differentiated between them and controls (Figure 4 and Supplementary Figure 6). Analysis of these in stimulated versus unstimulated PBMCs revealed a high fold change ratio of genes (n = 56) associated with innate and adaptive antiviral immunological processes, mainly in SUDV SN cluster 1. In contrast, low to no detected fold changes were observed in SUDV SN cluster 2, which resembled the response observed in controls. Mappings between the significant DEGs within the 2 SUDV SN clusters were generated (Supplementary Figure 7).

GO analysis identified significant enrichment in biological processes defense response to virus, negative regulation of viral process, type I IFN signaling pathway, regulation of regulatory T-cell differentiation, and response to IFN-γ (Figure 4C). Additionally, unsupervised PCA showed discrimination between SUDV SN clusters 1 and 2 only at 18 hours poststimulation (Figure 4D).

Long-term SUDV SN Displays an Elevated Concentration of Soluble Cell Markers Essential for T-Cell Activation

We collected supernatants from matched PBMC cultures. Secreted levels of 9 soluble markers, selected based on their reported roles in managing host immune responses during filovirus infection [34–36], were measured by multiplex ELISA. Values were then grouped (according to sample inclusion in clusters 1 or 2 defined by expression data (Figure 4). SUDV SN cluster 1 showed higher secretion levels for most soluble mediators (Figures 5A and 5B). Within this cluster, IL-8 and IP-10 at 6 hours poststimulation and IL-2, IFN-γ, and MCP-2 at 18 hours poststimulation all showed significantly elevated concentrations following induction (Figure 5B). There was a significant positive correlation between fold changes in crucial functional cellular cytokine levels, that is, IL-2 and IFN-γ, and the transcript abundance of their corresponding genes (Figure 5C).

Antiviral Humoral Immunity

SUDV SN Maintains Robust SUDV-Specific Antibody Responses

To further investigate the correlates of long-term protection, we characterized humoral anti-SUDV IgG responses in survivor sera. Irradiated SUDV antigen ELISA was used to measure anti-SUDV IgG levels ELISA. Of SUDV SN 58% (SN1–SN5, SN7, and SN12) maintained heterogeneous robust IgG virus-specific antibodies to irradiated SUDV (Figure 6A). As expected, uninfected controls showed no positive IgG response.

Virus-Specific IgG Antibodies From SUDV SN Mediate Polyfunctional Responses

We evaluated the ability of SUDV SN sera to mediate 2 IgG-driven antiviral responses, neutralization and ADCC (Figure 6B). PRNT₈₀ of live viruses in vitro showed that 50% of SUDV SN (SN1–SN5 and SN12) had neutralizing sera. As expected, none of the sera of controls were neutralizing. We used an ADCC assay to measure IgG-mediated, Fc-dependent, antiviral response via the FcγRIIIA receptor. Five of 12 SUDV SN (SN1–SN4 and SN12) mediated antibody Fc-dependent response. None of the sera of controls mediated a virus-specific Fc-dependent response.

Neutralization and Fc-dependent responses significantly correlated with anti-SUDV IgG immunoreactivity (r value = 0.5954, P value = .0411 and r value = 0.7118, P value = .0094, respectively; Figure 6C). The antibody Fc-dependent responses and neutralization capacity of sera also showed a significant positive correlation (r value = 0.7305, P value = .0070).

Finally, we correlated IgG levels and the fold change of selected gene markers (n = 56; Figure 4A) to identify genes associated with IgG functionality (Figure 6D). A nonparametric Spearman correlation showed statistically significant correlations between anti-SUDV IgG levels and folded changes in the expression of 19 genes at 6 hours poststimulation and 38 genes at 18 hours poststimulation compared to unstimulated controls.​​

Association Between Long-term Cellular and Humoral Immune Responses in SUDV SN

The relationship between cellular and humoral immune responses between and within each immune compartment (innate and adaptive) was probed by PCA and by constructing Spearman correlation matrix. We analyzed the 56 selected gene markers (Figure 4) and classified them into 5 groups based on their gene function [31] (Supplementary Table 1). The individual SUDV SN showed differences in immune pathways and level of responses between them and compared to analysis of only DEG responses (Figure 7A and Supplementary Table 2).

The PCA analysis revealed 2 subgroups of SUDV SN, that is, SUDV SN stimulus responder subgroup 1 (SN1–SN5 and SN12) and nonresponder subgroup 2 (SN6–SN11) (Figure 7B). An overlap between the SUDV SN nonresponders and controls was seen, pointing to similar immune responses. Analysis of the contribution of those responses in SUDV SN subgroup 1 revealed that the first 2 components separated antiviral and proinflammatory transcript responses from most of the genes (Figure 7C). For SUDV SN subgroup 2, the first 2 components separated proinflammatory transcript, antiviral, and adaptive immunity.

Spearman correlation (Figure 7D) was consistent with the PCA results. At 6 hours poststimulation, SUDV SN subgroup 1 showed a positive correlation between cellular and humoral responses. SUDV SN subgroup 2 demonstrated a similar correlation profile; however, these correlations were weaker and involved fewer responses. At 18 hours poststimulation, SUDV SN subgroup 1 showed consistent results at 6 hours, with a positive correlation between immune responses. In contrast, SUDV SN subgroup 2 showed no correlation between humoral and cellular immunity. A significant correlation was established between the same immune pathways seen 6 hours poststimulation.

DISCUSSION

The recurrences of ebolavirus outbreaks and the ongoing efforts towards vaccine development with limited efficiency emphasize the critical need to understand ebolavirus memory immunity and identify predictive long-term protection biomarkers. To shed light on the components and processes responsible for the development and maintenance of persistent memory immunity following filovirus infection, we studied SUDV antigen-specific cellular and humoral responses in long-term SUDV SN of the 2000–2001 outbreak in Gulu, Uganda, in whom we have been following their immune responses longitudinally for the past decade [11–14]. We observed robust cellular and humoral memory immune responses in SUDV SN nearly 2 decades postinfection. This long-lasting immunity was associated with specific gene expression signatures and multiple antiviral immune responses.

Recent studies have provided insight into the global transcriptional changes during and shortly after ebolavirus infection [4–9, 37–39]. However, the molecular mechanisms underlying a successful long-term immune response remain poorly understood. Here, we show that upon stimulation, PBMCs from some SUDV SN exhibit clear upregulation of genes associated with activation and suppression pathways of innate and adaptive immune compartments known to play critical roles in protection against viral infection. These included essential genes involved in IFN-dependent signaling, mainly in the innate antiviral defense responses, negative regulation of viral processing, and T-cell activation.

Further analyses of soluble mediators produced by cultures of stimulated PBMCs supported the gene expression results, demonstrating the production of high concentrations of specific cytokines and chemokines associated with T-cell activation and regulation from those SUDV SN (cluster 1) that continued to exhibit vigorous antiviral activity. Our study’s antiviral-gene expression and positive cellular responses correlated with the SUDV SN in whom we previously demonstrated a persistent cellular response [11, 12, 14, 40]. Several essential antiviral mechanisms were reported recently in EBOV patients and short-term recovered convalescent survivors [4, 7, 9, 22, 23, 39, 41]. However, our findings are particularly interesting as they provide the first indications of potential molecular mechanisms associated with long-lasting immunity.

We also examined IgG responses and markers of cellular immunity to understand antibody-mediated immunity duration better. We showed that specific SUDV SN could maintain strong antigen-specific polyfunctional IgG persistence. Our observation of sustained neutralization and Fc-dependent ability in this SUDV cohort of survivors is consistent with our previous reports [11–14] and those on EBOV survivors [42, 43]. The low IgG levels in some SUDV SN are also interesting. They indicate a seroreversion in previously positive IgG individuals [13], which should be further examined as it may reveal other protective immunity mechanisms and pathways in these survivors. This has been shown in EVD and other diseases [44].

The challenges associated with this on-site study generated unavoidable limitations on assays and analysis. These included sample size, which was restricted by the number of remaining survivors of our cohort, a lack of specific cellular functional assays that could be performed, selection of irradiate virus as compared with replication-competent virus for stimulation, and the focus on RNA gene panels rather than complete transcriptome analysis.

Nonetheless, our work adds medically relevant data from rarely studied long-term survivors of ebolavirus infections. We show associated solid polyfunctional humoral and cellular immune responses in naturally recovered SUDV SN 15 years postinfection. We pinpoint distinctive innate and adaptive gene expression signatures, which could drive the development of early and long-lasting immunity. This may help to develop new therapeutic approaches and vaccination strategies [45, 46]. These findings align with responses shown in EVD animal models and vaccine efficacy studies [44, 47], notably the observation that continuous memory immunity was absent in some survivors, thus suggesting subtypes of survivors with different qualities of immune protection pathways that need to be elucidated.

Comparative analysis of the 2 subsets of SUDV survivors did not indicate any significant differences in survivor’s age at the time of infection, gender, hospitalization period, or treatment (data not shown). As such, the immune differences in the subgroups of SUDN SN may be a product of reexposure, different viral immunological features, and carrier states, as recently suggested [48, 49]. However, this remains a hypothesis.

The capacity of the immune responses seen in SUDV SN to protect against reinfection is probable but still unclear, and whether the different subgroups of SUDV SN will benefit to the same extent from future therapies and vaccination remains unknown. However, endurance of the immune response for nearly 2 decades postinfection strongly suggests it is likely to have a fundamental role in defining ebolavirus correlate of protection and should be further explored.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.

Notes

Acknowledgment. This manuscript is dedicated to the memory of the late Dr Leslie Lobel—a great scientist and beloved friend who passed away during this work.

Author contributions. A. S., L. L., J. J. L., R. A. M., J. M. D., F. V., and G. D. E. designed experiments. A. S., P. B., and J. J. L. acquired and processed blood samples. A. S. and P. B. perform PBMC stimulation assays. S. L. and A. A. performed transcript assays, and A. S. performed secreted cytokines analysis and serological assays. A. I. K. performed the PRNT assays, and S. G. performed the Fc effector function assays. A. S., Y. M., J. C. M., C. D., F. V., and D. G. E. performed data analysis. The manuscript was written by A. S. and Y. M. and edited by C. D., F. V., and G. D. E.

Disclaimer. Opinions, interpretations, conclusions, and recommendations are those of the authors.

Financial support. This work was supported by the European Commission through the Horizon 2020 IF-EBOLA project (grant number 666102), PANDORA-ID-Net project (grant number RIA-2016E-1609), and EPIC-CROWN2 Horizon Europe (grant number 101046084). The Center for Advanced Microbial Processing and the Center for Genomic Sciences within the Institute for Molecular Medicine and Infectious Disease at Drexel University supported the gene expression studies.

References

Author notes